The Roots of Modern Biochemistry [Reprint 2011 ed.] 3110115859, 9783110115857

Citation preview

The Roots of Modern Biochemistry

The Roots of Modern Biochemistry Fritz Lipmann's Squiggle and its Consequences Editors Horst Kleinkauf · Hans von Döhren Lothar Jaenicke

W DE

G

Walter de Gruyter Berlin · New York 1988

Editors Horst Kleinkauf, Professor Dr. Hans von Döhren, Priv.Doz., Dr. Institut für Biochemie und Molekulare Biologie Technische Universität Berlin Franklinstraße 29 D-1000 Berlin 10 Federal Republic of Germany Lothar Jaenicke, Professor Dr. Institut für Biochemie Universität Köln An der Bottmühle 2 D-5000 Köln 1 Federal Republic of Germany

Library of Congress Cataloging in Publication Data The roots of modern biochemistry : Fritz Lipmann's squiggle and its consequences / editors, Horst Kleinkauf, Hans von Döhren, Lothar Jaenicke. p. cm. "Fritz Lipmann: bibliography, 1924-1985": Includes bibliographies and indexes. ISBN 0-89925-489-6 (U.S.) 1. Biochemistry-Congresses. 2. Biochemistry-History-Congresses. 3. Lipmann, Fritz Albert, 1899-1986. I. Lipmann, Fritz Albert, 1899-1986. II. Kleinkauf, Horst, 1930III. Döhren, Hans von, 1948. IV. Jaenicke, L. (Lothar, 1923-

CIP-Titelaufnahme der Deutschen Bibliothek The roots of modern biochemistry : Fritz Lipmann's squiggle and its consequences / ed. Horst K l e i n k a u f . . . - Berlin ; New York : de Gruyter, 1988 ISBN 3-11-011585-9 NE: Kleinkauf, Horst [Hrsg.]

Copyright © 1988 by Walter de Gruyter & Co., Berlin 30. All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced in any form - by photoprint, microfilm or any other means nor transmitted nor translated into a machine language without written permission from the publisher. Typesetting and printing: Tutte Druckerei GmbH, Salzweg/Passau. Binding: Lüderitz & Bauer Buchgewerbe GmbH, Berlin. Printed in Germany.

Preface

In response to the call for the LIPMANN Memorial Meeting in Berlin, there has been considerable enthusiasm for this fourth and perhaps final reunion of the LIPMANN community, being held in memory of one of the greatest, if not the greatest biochemist (Severo OCHOA). Looking back on the first "Squiggle Meeting" in 1966 in honour of Fritz LIPMANN, or on the meeting in 1 9 7 5 in honour of Severo OCHOA, which gave rise to Reflections on Biochemistry, and now on the 90 essays assembled here, one recognizes how quick the development has been, and the feeling of pleasure at being a participant in this interactive recognition process called science has increased. Participating in this process has surely been a delightful experience to everyone who worked with Fritz LIPMANN, the artistic type of a scientist, who had an eye for the unusual pattern ...he found four leaf clovers everywhere (Freda HALLLIPMANN). In addition to Paul EHRLICH'S four G's, Geduld, Geschick, Glück, Geld (patience, skill, good luck, means) Koscak MARUYAMA credits him with Geist und Gesundheit (intelligence and health), and with these six G's remained loyal to his simple, naive and persistant desire to clarify what interested him: the molecular mechanism of life (MARUYAMA). The study of such mechanisms, the assembling of biochemical building blocks, guiding and teaching his collaborators and associates, developing the form and style which a young investigator may take was one facet of his life; the secondfacet was creating (Nathan KAPLAN) an environment where there were stimulating colleagues, postdoctoral fellows and students, ( and to ) the establishment of a congenial atmosphere in which science (could) flourish as a consequence of free thought, unguarded exchange of ideas, critical discussion and respectful interaction among all of its personnel (Earl STADTMAN). As editors of this tremendous amount of material, most of which can be understood as a tribute to a life spent in biochemistry, a life connected with its roots and active at many frontiers in our initial understanding of basic mechanisms, and which has contributed heavily to the present basis of our research, we understand that this collection of essays will be of value from several perspectives. It provides insight into historical connections and into the impact of political conditions, as, for example, can be seen clearly in the Symposium on Respiratory Enzymes held in 1942 in Madison, Wisconsin, where American researchers and the distinguished émigrés (Van POTTER), Otto MEYERHOF, Fritz LIPMANN, Fritz SCHLENK, Kurt STERN, Erwin HAAS, and Herman KALCKAR met. Keeping this background in mind as an obligation, Dahlem had been selected for the Memorial Meeting. The Meeting provides insight into the triggering of advancements, the unpredictable moments of success in the experimental setups, and due to its distance in time from the events, offers a survey of long-range developments. This gathering has much in common with a piece of art created by many hands, and it contains many delightful moments. We again thank our colleagues for their lectures, their contributions, and the many helpful comments and suggestions. The editors

VII

Contents

1 Fritz Lipmann 1899-1986

1

Hall-Lipmann, F Life with Fritz

3

Lipmann, F. A Long Life in Times of Great Upheaval

9

De Duve, C. Fritz Lipmann: In Memoriam

37

Maruyama, K. Lipmann's Remarkably Fulfilled Life as a Researcher

43

Zachau, H. G. Fritz Lipmann: June 12, 1899-July 24, 1986

63

Kalckar, H.M. Fritz Lipmann Molding the Design of Molecular Bioenergetics

67

Barker, H.A. Recollections of Fritz Lipmann, 1941-1945 Ratner, S. In Celebration of the Scientific Genius of Fritz Lipmann

73 81

Bennett, T.P. Lipmann and "Not Strictly Biochemistry" Gevers, W. Communication in Metabolic Control. Intuition and Method in Biochemistry: Four Years each with Krebs and Lipmann Petrack, B. Fritz Lipmann: Squiggle to Protein Sulfation

105

Parthier, B. Fritz Lipmann (1899-1986), Honorary Member of the Leopoldina Academy

109

Krebs, H.A. and Lipmann, F. Dahlem in the Late Nineteen Twenties

Ill

Blaschko, H. Our Apprenticeship

125

Karlson, P. The Kaiser-Wilhelm-Institutes in Berlin-Dahlem in the Late 1930ies and Early 1940ies: Reminiscences of a Student of Biochemistry

129

Srere, P.A. On the Origin of the Squiggle ( ~ )

135

85

95

Vili

Contents

2 Biochemistry Comes of Age Moulton, L.J. and Akers, H.A. The History of Metabolites Isolated from Urine Horecker, B.L. The Pentose Phosphate Pathway Rapoport, S.M. Glycolysis and the Dawn of Modern Biochemistry Fruton, J. S. Energy-Rich Bonds and Enzymatic Peptide Synthesis Krampitz, L. O. A Nostalgic View of the TCA Cycle in Bacteria Dimroth, P. The Role of Vitamins and their Carrier Proteins in Citrate Fermentation McElroy, W.D. Lipmann's Influence on Firefly Luminescence Wieland, Th. Sulfur in Biomime tic Peptide Syntheses Baddiley, J. The Function of Teichoic Acids in Walls and Membranes of Bacteria Buchanan, J. M. The Amido transferases: Origins of the Concept of Affinity Labeling of Enzymes Grisolla, S., Knecht, E. and Hernández-Yago, J. Intracellular Protein Degradation: Past, Present and Future Herrmann, H. and Hiskes, A.L. Lipmann's Squiggle and the Unification of Cellular Structure and Function . Siekevitz, P. The Historical Intermingling of Biochemistry and Cell Biology Racker, E. Regulation of Function of Membrane Proteins by Phosphorylation and Dephosphorylation Richter, D., Meyerhof.W., Morley,S.D., Mohr,E., Fehr,S. and Schmale, H. Molecular Biology of Brain Peptides and their Cognate Receptors Najjar, V.A. The Biological Activity of Tuftsin, Thr-Lys-Pro-Arg Kubota, K. Biosynthesis of Linear Gramicidin, Pentadeca Peptide, is Tight Linked to Serine Metabolism and to Membranous Phosphoglyceride Kurahashi, K., Ikeuchi, T. and Kudoh, J. From Phosphoenolpyruvate Carboxykinase to Sporulation: Personal Reflections on Dr. F. Lipmann Döhren, H. von, and Kleinkauf, H. Research on Nonribosomal Systems: Biosynthesis of Peptide Antibiotics . . . .

139 141 147 157 165 181 191 205 213 223

231 251 261 285

295 305 323

331

339 355

Contents

IX

Bauer, Κ. Metabolism of Carnosine and Related Peptides

369

3 Molecular Biology Sharpens its Tools

377

Schweiger, M., Schneider, R., Hirsch-Kaujfmann, M., Auer, Β., Klocker, Η., Scherzer, E., Thurnher, M., Herzog, H., Vosberg, H.-P. and Wagner, E.F. DNA Repair in Human Cells: Molecular Cloning of cDNAs Coding for Enzymes Related to Repair Goldberg, I.H. Acyl~ Phosphate Intermediates in Oxidative DNA Sugar Damage by Antibiotics Kaji, Α., Okawa, N., Tanaka, M., Mori, K. and Finver, S. Rtsl: A Multiphenotypic, Unusual Temperature Sensitive Drug Resistance Factor Hierowski, M. T., McDonald, M. W., Dunn, L. and Sullivan, J. W. The Partial Dependency of Human Prostatic Growth Factor on Steroid Hormones in Stimulating Thymidine Incorporation into DNA Lara, F.J.S. Developmentally Regulated Gene Amplification in Rhynchosciara Mommaerts, W.F.H.M. Fritz Lipmann, a Few Personal Memories, and: What Else Came Out of the High-Energy Phosphate Bond? Herrlich, P. and Karin, M. Regulation of Gene Expression by Posttranslational Modification of Transcription Factors Maas, W.K. How Does the Arginine Repressor Regulate the Synthesis of Arginine Biosynthetic Enzymes? Shoeman, R.L., Maxon, M.E., Coleman, T., Redfield, B., Brot, Ν. and Weissbach, Η. The Biochemistry and Molecular Biology of the Terminal Reactions of Methionine Biosynthesis in Escherichia coli Lane, B. G. The Wheat Embryo, Then and Now Suiko, M. Mechanism of Cytotoxic Action and Structures of Thiadiazolo-pyrimidines . Wittmann, H. G. and Yonath, A. Architecture of Ribosomal Particles as Investigated by Image Reconstruction and X-Ray Crystallographic Studies Nomura, M. Initiation of Protein Synthesis: Early Participation and Recent Revisit Hadjiolov, A.A. Structure of Ribosomal RNA Genes of Eukaryotes: Some Solved and Unsolved Questions

379

389

397

407 417

423

431

441

447 457 477

481 493

505

χ

Contents

Spiriti, A. S. Energetics and Dynamics of the Protein-Synthesizing Machinery Fox, J. L. Expression Ganoza, M. C. Punctuation in the Genetic Code: A Plausible Basis for the Degeneracy of the Code to Initiate Translation Ganoza, M.C., Baxter, R.M. and J.L. Fox Reconstruction of Translation: Role of EF-P in Regulation of Peptide Bond Formation Kucan, Z. On the Role of Spermine in Protein Synthesis

511 535

539

551 555

4 Functional Dynamics

565

4.1 The Squiggle-Symbol of Bioenergetics

565

Jencks, W.P. Energy-Rich Compounds and Work Wood, H. G. Squiggle Phosphate of Inorganic Pyrophosphate and Polyphosphates Mukai, J.-I. Enzymology of 3'-Squiggled Nucleotides Hilz, H. Pyridine Nucleotides as Group Transfering Coenzymes Slater, E. C. The Nature of Squiggle in Oxidative Phosphorylation Harting Park, J., Moore, T.K., Anderson, B. and Park, C.R. Motional Dynamics of Fatty Acids: Advantages of 1 5 N and Deuterium Substituted Fatty Acid Spin Labels for Studies of Lipid-Protein Interactions and Motion in Membrane Bilayers Hoch, F.L. Thyroid Hormones and Oxidative Phosphorylation Anke, T., Schramm, G., Steglich, W. and von Jagow, G. Structure-Activity Relationships of Natural and Synthetic E-ß-Methoxyacrylates of the Strobilurin and Oudemansin Series Bessman, S. P. and Mohan, C. The Intracellular Mechanism of Insulin Action Duine, J. A. Unity and Diversity in Biological Redox Catalysis: Comparative Enzymology of Some Microbial Oxidoreductases Showing Variation in Cofactor Identity Liu, M.-C. and Peck, H.D., Jr. Ammonia-Forming, Dissimilatory Nitrite Reductases as a Homologous Group of Hexaheme C-Type Cytochromes in Metabolically Diverse Bacteria

569 581 603 609 625

631 645

657 663

671

683

Contents

XI

Lanyi, J.K. Halorhodopsin Bakker, Ε. P. Control of Futile Transmembrane Potassium Cycling in Escherichia coli . . . . Pfanner, Ν. and Neupert, W. Bioenergetics of Protein Transport into Mitochondria: Role of Δ ψ and of Nucleoside Triphosphates Witt, Η. T. Some Recent Functional and Structural Contributions to the Molecular Mechanism of Photosynthesis Kaiser, W.M., Kaiser, G., Martinoia, E. and Heber, U. Salt Toxicity and Mineral Deficiency in Plants: Cytoplasmic Ion Homeostasis, a Necessity for Growth and Survival Under Stress

721

4.2 Molecular Recognition and Communication

735

693 699

707

713

Ebashi, S. Dawn of Ca Research: Regulation of Muscle Contraction Ogawa, Y. Comparative Aspects of the Mechanisms of Energy Transduction in Sarcoplasmic Reticulum Between Rabbit and Frog Skeletal Muscle

747

Lardy, H., San Agustín, J. and Coronel, C. Caltrin: A Versatile Regulator of Calcium Transport in Spermatozoa

759

Ogita, K., Kikkawa, U., Ase, Κ, Shearman, M.S., Nishizuka, Y., Ono, Y., Fujii, T., Kurokawa, T., Igarashi, K., Saito, Ν. and Tanaka, C. Protein Kinase C, the Structural Heterogeneity and Differential Expression in Rat Brain Tao, M. Regulation of Erythrocyte Membrane Cytoskeletal Protein Interactions by Phosphorylation Shaltiel, S., Seger, R. and Goldblatt, D. A Kinase Splitting Membranal Proteinase: Use in the Study of Receptors Involved in the Cellular Response to Hormones Roskoski, R., Jr. Fritz Lipmann, Phosphoproteins and Regulation of Aromatic Amino Acid Hydroxylase Activity Kaufman, S. The Regulation of Hepatic Phenylalanine Hydroxylase by PhosphorylationDephosphorylation Sy, J., Tamai, Y. and Toy oda, Y. Catabolite Inactivation and Adenylate Cyclase in Yeast Suiko, M. and Liu, M.-C. Protein Modification by Tyrosine-Sulfation: Possible Functional Implications

737

765

771

781

791

805 813

817

XII

Contents

Delbrück, A. and Gurr, E. Proteoglycans and Connective Tissue Pathobiochemistry Lees, M.B. Acylation of Myelin Proteolipid Protein: A Link to the Past Allende, J. E. The Unusual Regulation of the Adenylyl Cyclase of Amphibian Oocytes by Progesterone. - A Review Bloch, Κ. Sterol Synergism, A Tool for Studies on Sterol Function Le, F., Zhang, Z.-G. and Zhou, T.-C. (Chou, T. C.) Chemical Modification of Benzodiazepine Receptors of Cortical P 2 Membranes Kaufmann, E. E. Dual Pathways for the Catabolism of y-Hydroxy-butyrate: Cytosolic and Mitochondrial Mechanisms

5 Evolution De Duve, C. Prebiotic Syntheses and the Mechanism of Early Chemical Evolution Fox, S. W. Prebiotic Roots of Informed Protein Synthesis: Nature of the Lipmann Connection Zubay, G. A Case for an Additional RNA Base Pair in Early Evolution Baltscheffsky, H., Baltscheffsky, M., Lundin, M. and Nyrén, P. Inorganic Pyrophosphate in Cellular Energetics and Evolution Babloyantz, A. Selforganization in Biosystems

827 835

841 853

857

867

879 881

895 911 917 923

List of Contributors

935

Acknowledgements

943

Lipmann's Coworkers at Massachusetts General Hospital, Boston, and the Rockefeller University, New York Fritz Lipmann: Bibliography 1924-1985

949

Index

979

945

1 Fritz Lipmann (1899-1986)

Fritz LIPMANN, at the age of 8 7 , was preparing a lecture to be delivered at the Meeting of the Federal European Biochemical Society in Berlin in 1986 when he passed away. His last words were "I can't function anymore". He lived a remarkably fulfilled life, to use the very words he found for Hans KREBS, with whom he shared the Nobel Prize in 1953. At a memorial concert in New York in December 1986, the plan was born to hold a memorial meeting, and it seemed that Berlin, because of its memories, would be a good location. Berlin again, the historical grounds of Dahlem, where LIPMANN began his wanderings in the twenties, just before the dark times took over and many German scientists had to leave, taking abroad with them some seeds of biochemistry. Those attending the LIPMANN Memorial in October 1 9 8 7 will especially remember the recollections of his contemporaries - of Freda H A L L - L I P M A N N reviving vividly the thirties and forties, Berlin as it was and as it became in 1933, the years in Copenhagen, the first difficulties in the States; and of Herman K A L C K A R , recalling the Abteilung MEYERHOF, LIPMANN'S masterful experiments, and the CARLSBERG laboratory. Between these recollections we have included A Long Life in Times of Great Upheaval, a favourite of many LIPMANN friends, and three fascinating tributes to Fritz LIPMANN, the talk Christian DE D U V E delivered at the memorial concert in 1 9 8 6 , a biographical sketch by Koscak MARUYAMA, a devoted researcher of LIPMANN'S life, and Hans ZACHAU'S notes on the great biochemist, reminding us that, although he never returned to Germany, he opened his laboratory to German researchers, which was an important service to biochemistry in Germany after the war. Many of his coworkers have taken the opportunity to share experiences, moments of discovery, and reminiscences of the laboratories. In this first section we have placed H.A. BARKER'S recollections of the early Boston period in the letters on acetylation research; Sarah RATNER'S retracing of a historical experiment on citrate synthesis which brought together LIPMANN, LYNEN, O C H O A , STADTMAN, and STERN in 1950, and depiction of the enzyme club at Rockefeller; Thomas BENNETT'S description of his sufferings (and not only his) in the sad Thomas TRACTOR case in the early sixties; Wieland GEVERS' thoughtful investigation into the research methods of both his teachers, LIPMANN and KREBS, comparing them to Mozart and Bach, respectively, as to their styles, which leads to his essay on Communication in Metabolic Control. Barbara PETRACK'S survey from the discovery of activated sulfate via sulfotransferase to the last studies carried out in Lipmann's laboratory on sulfation in malignancy. The reader can find equally interesting remembrances around Fritz and quite often Freda in contributions by Hugh AKERS, Jorge ALLENDE, Samuel BESSMAN, Joseph F R U T O N , Kiyoshi KURAHASHI, Marjorie LEES, Wilfried MOMMAERTS, Barbara PETRACK, Robert ROSKOSKI, Philip SIEKEVITZ, Harland WOOD, and Ting-chong ZHOU, in other sections of this book. Benno PARTHIER tells the story of how LIPMANN finally became an honorary member of

2

Fritz Lipmann

Leopoldina (Deutsche Akademie der Naturforscher). To achieve a more complete picture of the historical background, we have chosen to reprint the reminiscences of Hans KREBS and Fritz LIPMANN presented at the first LIPMANN Meeting held on the occasion of the latter's 75th birthday in Dahlem in 1974. This is complemented by Herman BLASCHKO'S reflections on the time with MEYERHOF, and Peter KARLSON'S impressions of the Kaiser- Wilhelm-Institut up to the beginnings of the Second World War. Finally, as a tribute to the "squiggle" itself, Paul SRERE closes the first chapter by presenting his insights into the relationship between the I Ching symbolism and the squiggle, the Krebs cycle, and the code.

Life with Fritz Freda Hall

Lipmann

I met Fritz in the bitter winter of 1929. Among the high points of a Berlin winter in those days were the great costume balls: the Academy ball, the Reimann ball, and others. They were wonderful events, beautifully decorated; American jazz bands played. It was at the Academy ball "Südseekitsch" that we noticed Josephine Baker - a sliver of a girl - dancing in the arms of a big bearded Berlin painter. Fritz and I met at the "Sozialisten Ball", a much more tame affair. Nothing socialistic about it either; it was a society of doctors, lawyers, journalists, and one had to be sponsored to get in. Well, I sauntered in there and someone grabbed my wrist. That was Fritz. The Meyerhof Laboratory was there for an outing. I remember only David Nachmansohn and Ken Iwasaki. Ken wore a tall blue sorcerer's hat with moons and stars, and that hat must have worked magic. It was because of Ken's friendship with Chobei Takeda that Chobei invited us in the fall of 1960 to a princely tour of Japan. Chobei's son Masao and Ken's son Kentaro were our guides and bodyguards. Fritz, always capricious, sometimes disappeared, causing them worry, and they had to search for him. And this tour, I think, was the root of Masao's coming to Fritz's laboratory in New York in 1964, staying for four years, and also a whole string of Japanese scientists coming to him for training. Shortly after the ball Fritz asked me to go with him to the "Drei Groschen Oper" im Theater am Schiffbauerdamm - the sensation of the season. And I was late. Very late. But there was in a thin coat a shivering Fritz, waiting. We missed the beginning but it was an unforgettable performance with Lotte Lenya as Jenny and Carola Neher as Polly. Fritz was an unusual young man. There was an aura of quietness about him. He seemed to be certain of a goal. He had no position, no prospects, and it did not seem to trouble him. He loved what he was doing and he called it play. At no time was Fritz the obsessed scientist without other interests. He always had time for fun. Fritz went late in the mornings to Dahlem and often stayed until late in the evening. I, working for newspapers, had frequent deadlines; but that did not keep us from sitting late into the night at Schlichter's or Stöckler's, two restaurants popular with artists, often with his brother Heinz or his painter friend Sebba, sometimes Lenya and Kurt Weill joining us. Weekends, in the summer, we went with Ken to tennis matches. We saw all the legendary players, Tilden and Boro tra, Suzanne Lenglen and Helen Wills. Or at midnight we dropped in at the six-day bicycle races. And, of course, we went to the theater, Max Reinhardt, Barnowsky, the Staatstheater; we saw exquisite opera performances, sometimes in a small castle outside Berlin. Not to mention the cabarets and art galleries. Several of the latter opened after 1933 on 57th Street in New York. Those were the years of "Schmach und Schande", great years to be young in Berlin. I thought I would never live anywhere else, but Fritz startled me, stating there was no

4

Freda Hall Lipmann

Fritz and his brother Heinz. Fritz already there looks into the distance, into his own world, as he did all his life.

future for him in Germany. On the desk in his room Fritz kept a picture of Otto Warburg clipped from a newspaper. He admired Warburg, not least for his elegant attire. In the fall of 1929 Meyerhof moved his laboratory to Heidelberg and Fritz went with him. During the following months we met a few times in Frankfurt and in Paris where I worked for a while. It became clear that Fritz wanted to get back to Berlin. This impelled him to bestir himself and write letters which led to the contact with Albert Fischer, who spent some time in Dahlem while the Carlsberg Fondets Biologiske Institut was being built for him in Copenhagen. He offered Fritz a job as his assistant and so in June 1930 Fritz returned to Berlin. We were so innocent of what was to come. There were rumblings, people in boots and khaki pants marched around; we had an unpleasant encounter in a restaurant, and one night on his way home Fritz dropped into a Bierstube and was beaten up. We didn't take is seriously. Albert Fischer had just married for the second time. His wife, Ragna, a young woman of striking looks, had without any previous training become an expert at tissue culture. Fischer wanted her to work as his technician at the Rockefeller Institute in New York where he intended to spend further waiting time. He suggested that Fritz and I come too, the four of us going together. He proposed Fritz for a Rockefeller Foundation fellowship, and in those days it was necessary for

Life with Fritz

5

us to get married. That was when I learned that because of my birth in Ohio I was legally an American citizen. Well, Albert had not counted on the rules of the Rockefeller Institute; they would not let Ragna work in the laboratory. Albert was furious, he refused to go at all, and so in October 1931 we sailed alone to the States. I remember with great emotion the Manhattan skyline emerging from the morning mist. Oh, so different now! New York seemed a quiet town then; this was still Depression and Prohibition. A man named Roosevelt was running for President; not much was expected from him. The Rockefeller Institute was a small and quiet place. I remember dinner at the Levene's, meeting Alfred E. Cohn, getting acquainted with Alfred and Reba Mirsky. In those days Mirsky could park his little car before a Loft's on Fifth Avenue while we went in for ice cream at the soda counter. Without air conditioning the laboratories closed down during the summer and everybody went to Woods Hole. There, among others, we met Linderström Lang who ended a tour through the States in a car loaded with paintings he had done out west. He was a very talented painter. In September 1932 we went across a placid Atlantic and a very stormy North Sea to Copenhagen. The Institute on Tagensvej was just ready, a small square building, the laboratories on the lower floors, the Fischers' grand apartment the floor through on the third, and the assistant's apartment and some guest quarters on the top floor. All this connected by a wide staircase and a small elevator. There we found ourselves in the empty rooms of the top floor. I ordered the walls painted in deep, strong colors and bought on credit a few pieces of unpainted furniture, which became black and yellow. The little elevator began working very hard taking Fritz up and down, often past midnight. When home and laboratories are in the same building, it is difficult not to check on the experiments. I am afraid Albert Fischer did not get the assistant he had expected but their relations remained cordial. Fritz began at once to work on his own ideas. On the windowsill of his laboratory he kept a cigar box labeled "Mein Geheimnis". I never asked what the secret was and the box stayed behind in Denmark. I did not take well to being without an income of my own, and on a visit to Berlin when Ullstein asked me if I cared to work again for them, I said enthusiastically "Yes", and so began a commuting existence between Berlin and Copenhagen; difficult but interesting. In this way it happened that I was in Berlin the night the Reichstag burned. That was the night when Berlin seemed to empty itself of most people I had known. Life quickly lost all charm. Everything culturally worthwhile closed down. Ugly buildings sprang up. Hitler's voice boomed forever through the streets from loudspeakers. People became suspicious and closed up. Arrests were made. Ullstein began to crumble. As time went on, the last of my Jewish friends could no longer stay in their homes, they slept two or three nights in the apartment of a friend and then moved on to another hiding place until the knock on the door caught up with them and they were taken away. After the Juden boycott, when I saw Nazi soldiers stationed before every Jewish-owned shop, I decided it was time to leave Germany and look for work in Copenhagen. I found work very quickly for Berlingske Tidende, one of Copenhagen's two great papers. The Danes had a long tradition of humorous graphics and many of my

6

Freda Hall Lipmann

drawings were funny. In the early mornings I often met the King on horseback in a pale blue tunic, riding all alone. One waved and he lifted his hand to his cap. Life in Copenhagen was very pleasant. In the summer there was the Tivoli where one always took visitors. Or Dyrehavsbakken, an amusement park with a wonderful roller coaster and a small theater. In winter one had concerts, the ballet, cozy bars and restaurants. Sweden was easy to reach. We once attended a riotous Studenter Carnival at Lund. Einar and Helle Lundsgaard - Fritz had become friendly with them in Heidelberg - lived very near; we spent many evenings together. Einar was a fabulous host, staging grand parties. There were parties too at Lang's where Lang loved to sing Holberg songs. Fritz of course went frequently to Lang's famous lunch seminars at the Carlsberg Institute. Heinz Holter and Herman Kalckar and his brother, who was connected with Niels Bohr, were there, also Erik Warburg who later became a sort of hero eluding a Nazi search party which invaded his home. Heinz Herman was a neighbour in the guest room for a long time. Gerti Perlmann turned up. So did Dean Burk, Ralph Gerard, and Bill and Phoebe Stein on their honeymoon. Fischer's second marriage foundered and a third Mrs. Fischer moved in. But dark clouds began to rise on the horizon. More and more people urged us to leave Denmark. They said Hitler intended to invade Denmark and England. Fritz wrote to Dean Burk and there - luck again! - was an opening with Burk and DuVigneaud at Cornell. We shipped a few good things and sold the accumulation of seven years at a hilarious sale where everything to the last can opener and lonely saucer was carried away. Taking leave of so many friendly people was sad - in June 1939 we sailed again for New York. Barely two months later on a summer evening, sitting with Carl and Gerti Cori on a porch at Cold Spring Harbor, we heard that Hitler had invaded Poland. Lucky again! Others later had to flee from Denmark to Sweden. Some committed suicide. We settled into an apartment on the East Side near where the United Nations now stands. On those grounds were the slaughterhouses. Barges brought the animals; we could hear the bleating of sheep on the 12th floor. Fritz often stopped there in the morning to buy what he needed for his work. In the summer of 1940 we rented a cottage bare of any comforts on Hinesburg Pond - now called Lake Iroquois - in Vermont. Fritz wrote one of his key papers there. Both of us loving jazz, we often went evenings to Cafe Society Downtown, listening to Teddy Wilson playing the piano, and to young black singers who later became stars. Or we sat in the sidewalk garden of the Hotel Brevoort together with people displaced from France, among them the shining Louis Rapkine who spent all his time and energy saving people. I had a splendid time in New York, but for Fritz these two years were not happy. As the time neared the end, things really looked black. There were some offers - totally unworthy - and Fritz turned them down. Suddenly there came the chance to go to Boston. Oliver Cope had the idea of creating a science laboratory in Churchill's department of surgery at the Massachusetts General Hospital, and asked Fritz to man it, an idea Churchill seemed to treat with doubtful tolerance. We moved into an attic apartment on Beacon Hill, that had an almost Parisian flavor. Fritz, at the M.G. H., squeezed into a room the size of a walk-in closet. It was not many weeks before visitors from many countries appeared to see Fritz, much to the astonishment

Life with Fritz

7

of the department. So he cannot have been as unknown as he thought. Our apartment was within walking distance of the M. G. H. and Fritz brought so many of his visitors for lunch that I sometimes felt like a short-order cook. It was exciting for Fritz to get Constance Tuttle as his first assistant. Soon came Kaplan, Novelli, Barker, Nose. Fritz enjoyed the contact with the clinicians of the hospital, especially Joe Aub and Walter Bauer - a colorful character - who supported Fritz. So did Conant, who, when it came about that Fritz was to be given an appointment, the objection was that the students would not understand what he said, is reported to have remarked: "They will hear him when he has something significant to say". A certain measure of financial security now achieved, we decided to realize our intention not to remain a childless couple. Stephen was born in August 1945, just about V. J. day. Fritz was a joyous father, helping in raising and entertaining his son in spite of the fact that these were the most intense years at the laboratory. His attention span was always short, every so often he had to run away from the laboratory, going for a walk on the Charles River shore, often stretching out on a bench. In 1951 we bought a beautiful house on Beacon Hill, No. 100 Revere Street, one of those narrow townhouses where one lived on four floors. This house quickly seemed to take on the character of a railway station so many people passing through. I remember Herbert Fischer, a hungry-looking Benno Hess, the Baddileys, Jacques Monod, Feodor Lynen, Christian and Janine de Duve, a young man named Watson bringing a pretty girl for dinner. Members of other disciplines appeared, Szilard, Schwinger, and frequently Gamov who entertained Stephen with wonderful tricks. There were, of course, the local friends, Jordi Folch, George Wald, Gerhard Schmidt, Enders, Woodward, Jack Buchanan, Eugene Knox. A great celebration evolved during a night in the fall of 1953 when the news from Stockholm came through. More than a hundred people roamed around in the little house - the police came three times for disturbing the peace. But Fritz grew restless. He wanted better opportunities for his co-workers and he was nervous because of Harvard's retirement rules. And there again at the right moment came the invitation from Detlev Bronk to join the Rockefeller faculty. It was hard to give up the house, but in the fall of 1957 we moved to New York. Fritz was happy to come to the Rockefeller Institute, which soon became Rockefeller University. He also liked to live again in a metropolis. He enjoyed the concerts and art galleries. When, as before, he was compelled to run away from the laboratory, he simply walked over to the Whitney Museum. There is little sense for me to mention the many men and women who came to him. It's all part of the "History". It was fun for me to see many of them at the picnics we held bi-annually at our country place in Dutchess County which we bought shortly after coming to New York. Although essentially a city person, Fritz came to love the country very much. He was not botanically inclined, but through his walks in the woods with Boomerang or Pogo (little terriers) he became interested in plants and mushrooms and always returned with unusual specimens. He read avidly, everything from Hermann Hesse to Camus and Sartre, Darwin and Freeman Dyson to Ann Beattie and Joyce Carol Oates.

8

Freda Hall Lipmann

I would like to end this saying I think Fritz was a very lucky person. He left Germany unwitting of the big trouble to come. He got a chance to leave Denmark just before World War II. A door opened at Harvard to wonderful happenings, and he could complete his work at Rockefeller, a place he loved and respected. For a man without the talent or inclination to promote himself, to have not only survived through such difficult times but to have prevailed, was really wonderful. It was a life full of surprises and excitement and I am lucky to have shared it with him. He also had the satisfaction of being feted by his scientific sons and daughters at his 75th birthday in a reborn Berlin, his 80th at Grignon, and his 85th at Lake Kawaguchi at the foot of Mt. Fuji. Fritz had an eye for the unusual pattern - he found four-leaf clovers everywhere. There were four-leaf clovers pressed in books, manuscripts, wallets. He could look at a lawn containing clover and spot the one that had four leaves. This ability did not leave him even after age and operations had affected his sight. In the summer of 1953 we were vacationing on Block Island. Werner Maas came visiting and the three of us went for a walk on the beach to the South end. There we decided to climb the rather steep cliffs. Werner and I reached the top first and sat down waiting for Fritz. We waited a long time until finally Fritz clambered up holding in his hand victoriously a beautiful four-leaf clover. Werner said: Now he'll get the Nobel Prize. During the last year of his life when his strength began to fail his fortitude was simply amazing. And he took it with humor. He insisted on taking the bus uptown. He worked more than ever on his papers, spoke of new ideas, new approaches; I heard him humming in his study, certain of another day. Luck held out; he did not have to be ill in a hospital and we could be together to the end. One evening I heard him say: I can't function any more - and that was it.

A Long Life in Times of Great Upheaval 1 Fritz

Lipmann

Contents Childhood School Years, and Study of Medicine Turn from medicine to biochemistry Apprenticeship in Meyerhofs Laboratory Use of phosphate bond energy in muscle Tissue culturing with Albert Fischer in Berlin First time in America 1931-1932 Work at Rockefeller Institute Woods hole in summer of 1932 Return to Europe: Copenhagen Pyruvic acid oxidation: acetyl phosphate Back to America Massachusetts General Hospital: Coenzyme A Role of Phosphorylation in Other Group Activities Carbamyl phosphate Sulfate activation Move to Rockefeller University Protein synthesis Polypeptide chain elongation Separation of the three elongation factors, Tu, Ts, and G Increase of phosphoryl potential of serine-O-posphate and tyrosine-O-phosphate by binding into protein Bacterial Production of Polypeptide Antibiotics by a Thiol-Linked Activation and Polymerization

Childhood, School Years, and Study of Medicine I was born in the middle of the last year of the last century, in the city of Königsberg, the capital of East Prussia. It is now Russian and its name is Kaliningrad. I lived through the two world wars and was lucky to leave Germany in 1931 to live in Denmark, where I worked for seven years. In 1939, shortly before the outbreak of the second world war, I got the opportunity to come and settle in the United States when I was already 40 years old. My work was not unknown, but also not well known. 1

Reprinted with permisson from Ann. Rev. Biochem. 1984. 53. 1 - 3 3 Copyright © by Annual Reviews Inc. All rights reserved

10

Fritz Lipmann

Königsberg, when I was born, was the largest city near the Russian border, and a minor harbor in the row of Hanseatic cities; it is connected by a sound of the Baltic Sea. It was only a short train ride to the seashore, where we spent all our summer vacations. As Königsberg was a large border town, the German government saw to it that its university and cultural life were of high quality. The university had a particularly good medical faculty. The head of medicine, Professor Mathes, was quite outstanding; he had written a fine book on differential diagnosis, which was translated into English. Later, while studying in Berlin and Munich for some semesters, I listened to great surgeons and psychiatrists, and (in Munich) to Friedrich von Müller, who pioneered in connecting medicine to biochemistry. This happened, however, after the end of the first world war, and I would like to fill in here on my early life and school years to the beginning of my study of medicine. I can still remember the peaceful years in the beginning of the 20th century, when carriages were horsedrawn, including the taxis, called droschkes, a word borrowed from the Russian. When I was four or five, during visits of the Kaiser, the nurse took me to town, where I was fascinated seeing him driven through the streets in a four horse-drawn carriage by a coachman with a plumbed helmet. However, during my teens the droschke and the Kaiser's coach turned into automobiles. By then I had started to go to a classical gymnasium where for ten years I learned Latin; however, it was only toward the end that I started to appreciate it for its precision and beauty. For six years we were also coached in Greek, which, I am somewhat ashamed to say, I liked much less than the Latin. I was never very good in school, nor later at the university - just average. After finishing the gymnasium I chose, for various reasons, medicine. The youngest brother of my mother was a very lovable man, and a wellliked pediatrician. He died young from a burst appendix long before the age of antibiotics. He had been one of my heros. Many of my father's friends were prominent physicians. My father, whom I was very fond of, was a lawyer. He once confessed to me that he was not enough of a crook to be an outstanding lawyer. Owing to his charming personality, nevertheless, he did quite well. I am grateful for his support through my student years and the early period after finishing medical school when, in Germany, one could not expect to be paid. Before leaving my early history, I have to introduce two who had great influence on me during my formative years. One was my brother, Heinz, who was two years older than I. We two were the only children. We were rather different; he was blond, popular, and outgoing. I was dark and somewhat clammed up. He was early interested in the theater, and he composed poems. I rather played with toy trains and building blocks. After graduation from the gymnasium he went to Munich to study literature; later I joined him there in 1919 to spend a semester studying medicine. In Germany, one was permitted to change universities. Munich was at that time artistically an important center, close in rank to Berlin. During my semester in Munich, I lived in its version of Greenwich Village, called the Schwabing. I was admitted to my brother's circle of artist friends. The memory of this time has remained with me. My other friend, whom I met when still in school, was Siegfried (Friedel) Sebba, a gifted painter who later became much interested in the theater. He and I remained in

A Long Life in Times of Great Upheaval

11

close contact all our lives; he died only a few years ago. I shall tell more about these two later. Soon after I had started studying medicine in 1917, in the last year of World War I, about May 1918,1 was called up to serve in the army. As a medical student, I was lucky enough to join the medical service; it became my first adventure far away from home. After a brief indoctrination with about 20 medics, I went on a long train trip, not knowing whereto. We ended up in Sedan, a nice, small French town on the Marne, far from the front. We were distributed among the hospitals there and had a rather easy life. Only at the very end did I get a brief taste of war. I had to serve in an improvised field lazarett in a church not far from the fighting. I could hear cannon fire. They were short of help there, and I had the duty to superivse about 40 seriously wounded men. I had to learn to exert authority. It was not an easy job, but a grim experience, with freshly wounded men badly taken care of. Returning to Königsberg after the war's end, I met there the murderous influenza epidemic that, I understand, killed a similar number of people as had been lost in the war. I was not yet dismissed then from the army, because there were hotels requistioned in my hometown for taking care of soldiers with this disease. I was assigned to one of them. I witnessed many people dying of the dreaded pneumonia, not only the soldier patients, but also the personnel - the head nurse and one of our doctors. In retrospect, it is a miracle that I did not get it, being for months constantly in contact with the patients. Finally, in March 1919, I got out of the army and could go back to medicine. Medical study in Germany was then divided into two distinct phases: two years preclinical work with a lot of anatomy, introductory physics, chemistry, zoology, and botany, followed by examinations in these disciplines, and two years of clinical work. Since I had finished gymnasium early (mid-1917) I already had one year of preclinical before induction and, by returning to medicine in March, I could get my first part finished in the summer of 1919, with some shortcuts allowed for service. It was then that I had my glorious experience outside and in medicine for a semester in Munich, and the text half-year in Berlin. I had, eventually, to return for finals to Königsberg and finished medicine in 1920. The required practical year, half of which I spent in the municipal hospital of Königsberg, was most interesting; when not serving on the wards I reviewed the stream of cases of alcoholics, accidents, and derelicts flowing through the emergency and out-patient departments. While studying medicine. I was quite absorbed with learning what is going on inside the human body. But I had a glimpse into chemistry through our teacher, Professor Klinger, who presented us, in his preclinical lectures, with a dramatized view of the field.

Turn from medicine to biochemistry Toward the end of my studies in medicine, doubts appeared in my mind if I really wanted to become a practicing physician. I was uneasy with the prospect of charging people money for trying to make them healthy. Yet, I spent three months in the pathology department of a hospital in Berlin-Friedrichshain with a rather well

12

Fritz Lipmann

known pathologist, Ludwig Pick. Then, during the last part of the practical year, I took a very popular three-month course in modern biochemistry in Berlin, taught by Peter Rona, a collaborator of the famous Leonor Michaelis. After the course, I continued working with Rona on my first published paper (1), which was on an aspect of the then very popular colloid chemistry. This became my required M. D. thesis. A fortunate opportunity then came along to work on a stipend for a half-year in pharmacology at the University of Amsterdam. I was happy to escape from the inflation that was then rampant in Germany. In Amsterdam, I became acquainted with biochemical problems and with being in a biological laboratory. This decided me to turn to biochemistry; I felt, however, that I needed more chemistry. Meanwhile, Hans Meerwein, an excellent organic chemist, had been appointed the new head of the chemistry department in the university at Königsberg. This made me decide to return home to live cheaply with my parents and study chemistry with Meerwein. This I did for three years. Most instructive were Meerwein's lectures, which every member of the department and all students had to attend. I spent a good deal of time in the practice of inorganic chemistry. It was very instructive to realize the individualities of the important atoms by learning to analyse the mixtures of their salts. I also learned the basics of organic chemistry, analysis as well as some synthesis. Meerwein's lectures were exciting and the repetition and progress during the three-year period gave me the feeling for chemistry, which I thought I needed for going into biochemistry.

Apprenticeship in Meyerhof s laboratory After the three years, I passed the Staatsexamen, and now had to find for my chemical thesis a good biochemical laboratory where work went on that seemed to deal with the moving edge of biochemical understanding in intermediary metabolism. I chose Otto Meyerhofs laboratory at one of the Kaiser Wilhelm Institutes (now Max-Planck Institutes) in Berlin-Dahlem. When I entered it, it was only about 25 years since the cell-free fermentation of sugar to alcohol had been discovered in yeast by the Büchner brothers, as described by Kohler (2), and Harden (3) had started work on its mechanism by finding a fixation of phosphate in intermediaries on the way to yield two moles of ethanol and two of C 0 2 from one mole of glucose. Meyerhof and Karl Lohmann studied the glycolysis of glucose to form two moles of lactic acid in frog muscle extacts. Except for the last steps, it was quite analogous to alcoholic fermentation. Here, pyruvic acid is not decarboxylated as in yeast to C 0 2 and acetaldehyde, which then is reduced to ethanol by NADH, Harden's coenzyme. Instead, in muscle extract, the pyruvic acid is directly reduced by NADH to lactic acid, which is here the end product. Work by Hill and Meyerhof on heat production in an anaerobically working muscle indicated a rather exact equivalence between lactic acid formation and contraction. This connection of intermediary metabolism with energy production for the first time began to be understood in mechanistic terms. It was this kind of approach that made me choose Meyerhofs laboratory to get my introduction into intermediary metabolism.

A Long Life in Times of Great Upheaval

13

My early work there, however, was not too successful. When I asked him, in 1927, if I could work in his laboratory, he agreed to take me; probably he liked that I had spent some time in chemistry. Fortunately, I could use the work I did there for my chemical thesis. The first two tasks he gave me were interesting shots in the dark that did not yield a much wanted understanding of the role of a new compound, creatine phosphate (CrP), which had been found in muscle. It was chemically identified as CrP by Fiske & Subbarow (4), and Eggleton & Eggleton (5) had simultaneously observed the presence of an acid-unstable phosphate, which they had not characterized, and called it phosphagen. The work indicated some relation of CrP to muscle contraction. However, such a connection was then difficult to define. Meyerhof s fishing for some use of CrP was not without justification because normal contraction definitely showed a rather irregular break-down of CrP. Then I started a somewhat more successful piece of work on metabolic effects of fluoride, a known inhibitor of muscle contraction. I did not resolve the mechanism of its inhibition of contraction, but observed that it inhibited phosphatases in muscle and yeast extracts. I confirmed inhibition of glycolysis and found, to a lesser degree, also inhibition of respiration. I noticed that fluoride had been reported to react with methemoglobin to form a specific fluoromethemoglobin and thought it worthwile to look into this rather interesting interaction. Eventually, the combined

Figure 1

Portrait of Fritz and Heinz Lipmann, 1926, by Siegfried Sebba.

14

Fritz Lipmann

papers on metabolic fluoride effects were used as my chemical doctorate thesis. Of those connected with the Dahlem Institutes, only Karl Neuberg, also a professor at the Charlottenburg Institute of Technology, could accept students for promotion. Thus, he was kind enough to arrange for me to be accepted as his doctorand, based on the work done in Meyerhof s laboratory, as he was working on parallel problems in yeast fermentation. Therefore, in 1929 I obtained the additional degree of Ph. D. in chemistry. Before expanding on a more detailed description of the dramatic developments in muscle contraction at that time, I pause here briefly to describe the unusually active artistic and scientific life in Berlin in the 1920s, before Hitler's breaking it up. It was indeed the cultural center of Europe in the arts and in the sciences. My brother by then had moved to Berlin to become the dramaturge with the intendant of the Staatstheater, Leopold Jessner. There, a group of great actors presented exciting performances. As in Munich, I again had even closer contact with the people of the theater; they are a clan, closed up and agitated by their problems and intrigues. Among them I could mix with unusual characters and beautiful women, often astonishingly intelligent. They are a caste not unsimilar to the scientists. Also, my friend Friedel Sebba had then come to Berlin. He had meanwhile concentrated on oil painting. I saw Friedel often and learned from him about the difficulty with achieving the expression in his pictures as he wanted them. He had a good understanding of the kind of work I did. He knew that I worked with frogs then and once asked me, "Did the frog speak today?" That was a good question, because that is what we want to do, make the frog "speak." Later, only did the living cell start to tell me something; it took me quite a while to get there. I show here an example of Friedel's paintings, in a general way representative of the paintings as the time in Germany. It is a black-and-white reproduction of his 1926 portrait of my brother and me; he smuggled himself in as the mask my brother is holding. He put us into costumes for a masked ball, which was a prominent part of the night entertainment in those Berlin years. At one of them I met my wife-to-be, Freda Hall. That was early in the winter of 1929, when I had just finished the quite strenuous Ph. D. examinations.

Use of Phosphate bond energy in muscle Hill and Meyerhof had earlier successfully determined a constant ratio in living muscle between glycolysis, measured by lactic acid formation, and heat production in the anaerobically contracting muscle. It seemed, thus, to be quite a surprise when Einar Lundsgaard, from the University of Copenhagen, reported a very curious effect of the injection of monoiodoacetate into rats in amounts he found completely to inhibit glycolysis. These rats ran around quite normally for 5 - 1 0 minutes, but then collapsed with their muscles going into a rigor similar to the rigor mortis. However, these muscles in contrast to normal ones had been exhausted without lactic acid formation. Inhibition of glycolysis by the iodoacetate caused, rather, a complete decomposition of available creatine phosphate. Now, in Meyerhofs

A Long Life in Times of Great Upheaval

15

laboratory, it had been shown that CrP hydrolysis developed an unusually high amount of heat. Thus, Lundgaard's experiments could only indicate that in absence of glycolysis, the breakdown energy of CrP could be used by the muscle for contraction. Lundsgaard cautiously concluded that the use of this phosphate bond energy was nearer to contraction than the production of lactic acid. The final explanation was that CrP was in equilibrium with ATP; the terminal phosphate bond of ATP had, by its heat production, already been found in Meyerhof s laboratory to be on an equal energy level as the one in CrP. On the other hand, the energy yield of glycolysis, we now know, appears as the phosphate bond energy of ATP, which in turn is formed from the energy-rich phosphate of Ρ ~ enol pyruvate and phosphoglyceryl ~ phosphate. In Lundsgaard's experiment with blocked glycolysis by iodacetate, CrP acted in the short normal period of the injected rat as energy donor to ADP, acting thus as an energy buffer, substituting for glycolysis. In this respect Lohmann's finding is important that in a muscle extract the adenylic acid system acted as a coenzyme accepting the ~ Ρ of CrP. I use here the squiggle ~ to indicate an energy-rich bond. This was introduced in my review on phosphate bond energy (21). It is discussed later as indicating a free energy of AG°' of 10 + 3 to 5 kcal of hydrolysis at physiological conditions. All this revealed that the end product of glycolysis, lactic acid, did not play a role in the contraction mechanism, and it removed the previous temptation to connect the mechanism of contraction with acidification caused by the formation of lactic acid. The evolving evidence for ATP as the driving force of the muscle is discussed later. Since fluoride, like iodoacetate, inhibits muscle contraction, I injected frogs with fluoride and after stimulation observed Cr ~ Ρ and some ATP breakdown replacing the inhibited glycolysis, just as in Lundsgaard's iodoacetate injection, ending in a rigor of the muscle after exhaustion of the Cr ~ P. Toward mid-1930 the Meyerhof laboratory moved into what was called the Physiology Section of a new, beautiful Kaiser Wilhelm Institute in Heidelberg. After settling there, Meyerhof suggested that I do experiments with frog muscle on CrP breakdown at acidic pH, which was known to inhibit glycolysis. One of the drawbacks of previous experiments on CrP breakdown during contraction was that the determinations of metabolic reactions had been done by workup of the muscle after the experiments, e. g. by pulverizing the deep-frozen muscle and deproteinizing with ice-cold trichloracetic acid. The supernatant was used for lactic acid and CrP determination. It was known, however, that particularly CrP breakdown occurred during the workup. In the experiments reported with the new procedure, both lactic and CrP determinations were done manometrically with frog sartorius muscle anaerobically suspended in Ringer bicarbonate buffer between pH 6 and 8. About 100 mg frog sartorious muscle was used; the diameter was less than 1 mm. This dimension allowed an easy equilibration with the suspension fluid. Comparing the electrotitration curves of free and split CrP, due to the liberation of the strong base creatine on split of CrP, a sizable alkalinization takes place. It amounts in base equivalents per mole of CrP split depending on pH to 0.8 at pH 6 and nearly 0 at 8; the alkalinization declined almost linearily between these two pH values. The measurements at 15 °C showed curves that, on appropriate stimulation by way of platinum threads melted into a Warburg-manometer vessel, indicated in the first

16

Fritz Lipmann

hour a nearly linear C 0 2 absorption bending into C 0 2 liberation for the next hour. The levels of the turnover points depend on the pH varied anaerobically in mixtures of N 2 with CO2 of 1.1%-11.5% with constant 0.02 m bicarbonate as mentioned. The C 0 2 absorption and evolution correspond to an inital CrP breakdown and subsequently produced lactic acid. The chemical determination of CrP and lactate confirmed the manóme trie results. In an experiment with iodoacetate-poisoned muscle at pH 6.45, only C 0 2 absorption and of similar magnitude as in a pHequivalent of normal muscle was observed. However, it was not followed by the C 0 2 evolution by lactic acid due to poisoned glycolysis. This furnished confirmation of the validity of our manometry. It indicated at the low pH range an initial CrPlinked contraction period as long as CrP was available, followed by a turn to lactic acid formation, the degree of alkalinization of CrP utilization depending on pH. These results confirm the early observation with muscle, by D. Nachmansohn, that by a single two-second tetanus contraction, the extent of CrP breakdown was four times that of the formation of lactate, while by continued stimulation the use of CrP was gradually replaced by lactate formation. Our experiments now removed any doubts about a use of CrP as source of energy for the contraction in a normal muscle and also removed acidification as such, as a possible reaction connected with contraction (6).

Tissue culturing with Albert Fischer in Berlin Since the time I was still with Meyerhof in Berlin it became clear that after a halfyear in Heidelberg I had to get another job. I had been in Meyerhof's laboratory during the last two years on a fellowship that could not be renewed. I was very eager to find a place in Berlin after this interval in Heidelberg, during which Freda Hall and I saw each other only a few times in places between Heidelberg and Berlin. Therefore, I was very happy to get a job in Berlin with Albert Fischer, and to join her again. Fischer's laboratory was located within the Kaiser Wilhelm Institute for Biology, where Meyerhof had worked; the return to Berlin was, therefore, not a great change for me laboratory-wise. Tissue culture was a rather interesting new field, in which Fischer had become a leader. He wanted me to use metabolism as a method to measure cell growth. It turned out to be a feasible approach by using the fairly sizable aerobic glycolysis found in the tissue cultures of fibroblasts. The measurement of actual growth in tissue cultures is not easy. A true measure could, at that time, only be obtained by tediously counting under the microscope the cells in the state of mitosis. To use metabolism as a measure looked promising, and I worked out an easier manometric method showing essentially linear increase of aerobic glycolysis or respiration in Warburg vessels for tissue cultures. In view of the long periods of several days, differential manometers had to be used to avoid interference by the daily change in air pressure. The culture of fibroblasts was kept in plasma and appropriate fluid laid over it; the thickness of tissue and the plasma layer were kept at dimensions appropriate for a saturation with oxygen. The compensatory vessel contained the same plasma level without tissue overlaid with the same fluid layer.

A Long Life in Times of Great Upheaval

17

First experiments were made using respiratory 0 2 uptake by heart fibroblasts in phosphate buffer in presence of KOH to absorb the respiratory C 0 2 in a side vessel, separated from the tissue compartment. The 0 2 absorption was low, but constant 0.91 cmm per hour. A better increase of rather high aerobic glycolysis was found with osteoblast cultures using bicarbonate-C0 2 buffer aerobically and measuring C 0 2 evolution manometrically. As background, omitting the essential growth-promoting embryonic extract, in 37 hours a barely measurable increase of 0.12 ccm of hourly C 0 2 production compared to a quite constant hourly increase of 2.3 ccm over 47 hours with embryonal extracts. Yet, as a routine determination this method under aseptic conditions was too complex. During the first year I spent in Fischer's laboratory, the new field interested me a good deal. Particularly the results I had obtained prompted me to do experiments on the large aerobic glycolysis in normal embryonic fibroblasts. Such a glycolytic activity was in the foreground of interest because of Otto Warburg's discovery of high aerobic glycolysis in malignant tissues. This high glycolysis in normal fast growing tissues may be indicative that in both cases a large supply of anaerobic energy was needed as a safeguard against the possibility of limited oxygen supply. After I had been with Fischer for less than a year, he was offered a new institute in Copenhagen and wanted to close the laboratory in Berlin. He wanted to spend the following year supervising the construction of his new institute funded jointly by the Rockefeller and the Carlsberg Foundations. The latter was a great force in Danish science in that the income of the prosperous Carlsberg brewery went through a Carlsberg Foundation into a generous support of science as well as the arts. Fischer wanted to take me with him to Copenhagen, which, with the threat of Hitler's influence growing rapidly in Germany, was a great stroke of luck. He arranged for me to get a Rockefeller fellowship for a year, which I could spend wherever I wanted and in any way I chose. I decided to join Dr. P. A. Levene at the Rockefeller Institute for Medical Research in New York. He had published a paper on the phosphate content of egg yolk protein, which was amazingly high. The protein isolated from egg yolk, which he called vitellinic acid, contained 10% bound phosphate. In Meyerhof's laboratory I had become familiar with hydroxyester phosphates of glucose and fructose and gained an intimate experience with the energized N-P link in creatine phosphate. Since proteins contain a lot of nitrogen, I first guessed it to be such a N-P link. Levene had not made attempts to identify the linkage of the phosphate in this vitellinic acid. I therefore wrote to him and he accepted me to work on this identification in his laboratory. Freda Hall and I had become quite closely attached to each other and before leaving for New York we decided to marry and go together to America.

First time in America, 1931-1932 We took a boat, as there were no airplanes yet, and arrived in September 1931. We were told that we should look for an apartment in Long Island. We regretted that later, after renting a rather dingy place in Jackson Heights. To get to the city, I had

18

Fritz Lipmann

every morning a half-hour subway ride and then a 10-15 minute walk to the institute. There was then in the United States a depression, which I must confess, we did not notice much. Levene was a cultured and interesting person. The atmosphere at the institute, however, was quite restrained. A few Europeans were rather friendly, and of the Americans, Alfred Mirsky and his wife, Riba, and M. L. Anson took a little care of us. Anson took me to the Princeton section of the institute, now long abolished.

Work at Rockefeller Institute There were two prominent phosphorylated proteins known at that time. One in egg yolk, and the other one in the casein in the milk of mammals. The latter contained only 3% phosphorus; it was a mixture of various proteins. Therefore, in the first attempt to identify the phosphate linkage, a protein containing such high amounts of phosphorus was most promising with the 10% of Ρ in vitellinic acid. It could, moreover, be isolated easily in pure form. These compounds interested me altogether because both served as food for the tissues of growing animals with which I had started to work. Becoming a little better acquainted with the literature, I soon found out that Pimmer & Bayliss (7) had worked on many phosphorylated proteins and found as a general property a considerable lability of the phosphate toward alkaline reaction. This more or less excluded an N-P link because the N-P link is stable to alkali, but unstable to acid. They had also already found that the phosphate in egg yolk and casein was, in contrast, quite stable to hydrolysis by acid. It was therefore decided to use limited acid hydrolysis in order to isolate the phosphorylated amino acid. I soon learned to obtain the vitellinic acid; it was relatively easy to purify. My preparation contained approximately 10% phosphate, and I started with its hydrolysis in a boiling water bath for 10 hr with 2.5 molar hydrochloric acid. I used the method of Lohmann and enclosed equal samples in a number of soft glass tubes and sealed the tubes over the flame to keep the concentration of HCl constant. After 10 hr a sample was analyzed and a good deal of inorganic phosphate, short of 50%, was found. By further hydrolysis for 10 hr, I seemed to have reached the stage where most of amino acid phosphate seemed to be in free form. From this hydrolysate I tried now to isolate the amino acid phosphate. This I precipitated with barium and after some reprecipitation, analysis showed its composition to be quite consistent with the barium salt of serine phosphate. In the meantime, I had found that S. and T. Posternak had published an isolation of what they thought was a tetrapeptide of serine phosphate as it contained equal amounts of serine and phosphate. However, C. Remington had worked with casein, which on hydrolysis behaved, according to Plimmer & Bayliss, very similarly to the phosphate in vitellinic acid, but he claimed it to be bound there to different amino acids. It was, however, most likely that serine was also a carrier of phosphate in casein.

A Long Life in Times of Great Upheaval

19

Woods Hole in Summer of 1932 In June the institute emptied out and everybody went to Woods Hole on Cape Cod, in Massachusetts. Leonor Michaelis, then a number of the institute, offered me space in the rooms he regularly occupied at the well-known marine laboratory of Woods Hole. We went there by boat; it took almost half a day to New Bedford, Massachusetts, and from there, I don't remember how, to Woods Hole. I met many interesting people; almost everybody of consequence showed up. I met Guzman Barron, an interesting man with whom I often met later on. John Runnstrom, a Swede, who worked with Michaelis at Rockefeller, was there. Later, during our Danish period, we became good friends with him and his wife. It was particularly exciting to meet with Linderstrem-Lang, who had worked with Morgan in Pasadena, and had driven across the country to Woods Hole. He was one of the Danes we later got to know well - one of the most exciting persons in Copenhagen, as a scientist and as a human being. Recently, I was pleased to be invited to write about him (8). There, I expanded on my memory of this extraordinarily gifted man. He died, sadly, when he was still in his prime. At the end of the summer we returned to New York by train, and I finished the work on vitellinic acid (9). We then returned to Europe.

Return to Europe: Copenhagen Stop in London. On a visit to London, while traveling to Copenhagen, I went to the Lister Institute and saw Robert Robison. He was quite interested in this "new hexosediphosphate" Lohmann and I (10) had investigated as being formed from fructose diphosphate on incubation in muscle extract. I was then quite unsure about the nature of this product, feeling that we had too little information on it, apart from the considerable increase in acid stability of the ester phosphate. My uneasiness proved to be well founded. Shortly afterward, G. Embden became interested in this "transformation" of ours and connected it with Ragnar Nilson's phosphoglycerate formation from hexose-diphosphate + acetaldehyde with yeast, because the phosphate of phosphoglyceric acid was very difficult to hydrolyze. He confirmed the formation if the expected phosphoglyceric acid and found in addition, phosphoglycerol. Embden guessed rightly that these two three-carbon compounds, both very difficult to hydrolyze, were the products of our reaction, rather than a new hexosediphosphate. He postulated they were formed by a dismutation of the trioses phosphoglyceraldehyde and phosphodihydroxyacetone, to which fructosediphosphate was split enzymatically. This ingenious interpretation, soon fully confirmed, earned him rightly the companionship in the name of the EmbdenMeyerhof pathway for the glycolytic cycle. I described this development much later in a contribution to a volume in honor of Gerhard Schmidt, a pupil of Embden with whom we became quite friendly during the time we were in Boston (11). From England we took a boat to Jutland and the train. In Copenhagen, we found the new institute ready and, as it was there the custom for the first assistant, they had prepared for us an apartment on top of the institute to be supplied with some

20

Fritz Lipmann

furniture. There was ample space to live in, and a guest room where we could house visitors, among them Friedel Sebba, who had moved to Sweden, and Dean Burk, a colleague of mine in Meyerhof's laboratory. He had become interested in my work on the Pasteur effect, to be described later. I had ample laboratory space, and eventually such co-workers as Jorgen Lehman from Sweden, Brecke, a young student from Norway, Gertrud Perlmann from Austria, and others. I was very eager first to continue with the analysis of casein for the phosphorylated amino acid. Using the same hydrolysis procedure with 2.5 η HCl, I easily isolated again a quite pure preparation of barium and silver phosphoserine. Pasteur Effect. Through the above described finding of high aerobic glycolysis in chicken fibroblasts, I became interested in the aerobic repression of the substratewasteful energy supply by glycolysis through the highly economical respiratory energy production. Warburg suggested that the high aerobic glycolysis in tumors was due to abolition of the generally observed inhibitory effect of respiration on glycolysis in normal tissue. He called it the Pasteur effect, because Pasteur had first described it in facultative anaerobic yeast. For testing in an in vitro system for an oxidative inhibition of glycolysis and fermentation. I added positive oxidation reduction indicators of the indophenol blue type to particle-free-fermentatively active yeast extracts with negligible oxygen uptake, which in air remained colored, but by exclusion of air were easily decolorized by reduction. It appeared that they lost the activity to ferment glucose in air, but not in anaerobiosis. A glycolyzing muscle extract behaved similarly. These extracts did not respire. I thought this to be an "artificial" Pasteur effect due to oxidation of an enzyme. At that time I had played around a great deal with measuring the redox potential in fermenting yeast and glycolyzing muscle extracts. I used first quinone and iodine as oxidants, and measured, upon gradual addition, the disappearance of the nitroprusside reaction for SH-, presumably that of glutathione. The disappearance of glycolysis in these extracts coincided with the disappearance of SH-groups. Even a slight remainder of SH-groups permitted almost normal glycolysis. Today, one would argue that this oxidation of the SH-group in glutathione opens the access of the oxidant to the essential -SH in phosphoglyceraldehyde dehydrogenase. Thereto, the phosphoglyceric acid has to attach intermediately in the oxidation process to accept the phosphate for eventual phosphoryl transfer to ADP. With a strong oxidant, this reaction seems irreversible. I then became generally interested in the energy metabolism of embryonic cells. This was prompted by the finding of high glycolysis in normal chick embryo fibroblasts similar to the one in malignant tissue. I proposed that in both cases the high glycolysis may be used to overcome the possibility of oxygen deficiency (12). Recent work suggests that the respiratory inhibition of glycolysis and fermentation is due to a feedback inhibition by high concentration of ATP produced by respiration. The phosphorylation of fructose-6-phosphate by phosphofructokinase is inhibited (13). My early experiments were largely designed to emphasize the unlikelihood that the Pasteur effect is due to a cycle of lactate, formed glycolytically, being resynthesized by respiration to glucose. Such resynthesis of glucose from lactate takes place in liver, but not in muscle.

A Long Life in Times of Great Upheaval

21

Pyruvic acid oxidation: Acetyl phosphate Since in a respiring cell, the fate of pyruvate is changed from a reduction to lactate to an oxidation to acetate and C 0 2 . I wanted to study the mechanism of pyruvate oxidation. At that time, in 1937, the only case of cell-free oxidation of pyruvate had been reported with an acetone powder of a Lactobacillus delbrueckii acidificane longissimus, which I obtained from Germany. Our laboratory was equipped for growing tissue, but not bacteria. However, I managed to get a good growth of this organism, which likes a temperature of about 40 °C by growing it near the water heater in the cellar, and centrifuged it off in a milk centrifuge. A desiccator-dried powdered preparation was found best for extraction. When purified by ammonium sulfate fractionation, the enzyme lost two cofactors, thiamin pyrophosphate and flavin adenine dinucleotide, the restoration of which was needed. The most important, rather accidental, observation was made when the medium was on one occasion switched from phosphate, generally used as buffer, to bicarbonate-C0 2 . Then I did not get the oxidation of pyruvate. It returned on addition of inorganic phosphate (P¡). Suspecting an energy-rich intermediary, I added radioactive P¡ and adenylic acid in a crude extract and found pyruvate oxidation yielded ATP. After finding an old method for synthesizing acetyl phosphate, a crude sample of it was synthesized. With a dry preparation of L. delbrueckii shaken with adenylate and acetylphosphate (AcP), and approximately similar amount of AcP disappeared and ATP appeared. AcP, like CrP assays in the P¡ determination as inorganic P. The description of all these experiments, done while still in Copenhagen, was presented at the 7th Cold Spring Harbor Symposium during 1939, in the USA (14).

Back to America Toward the end of 1938, Hitler's fascism was slowly expanding into Denmark. Our Danish friends, foreseeing the danger of a war and the occupation of Denmark by Hitler's Germany, urged us to try to leave. I got in touch with Dean Burk, my American colleague in Meyerhof's laboratory, who had visited us in Denmark. I thought he might be helpful in view of our common interest in the mechanism of the Pasteur effect and I knew that he was moving in 1939 to a laboratory in DuVigneaud's department at Cornell Medical School, at the New York Hospital, and he could choose two assistants. I wrote to him that I needed a job in the United States, because Denmark had become unsafe due to the developments in Hitler's Germany. Linderstram-Lang, who knew DuVigneaud well, was most helpful in recommending me to him, Luckily, my invitation to become Burk's assistant in DuVigneaud's laboratory was arranged, which was necessary for me to enter the United States. It turned out to be helpful that my wife, who had been born in Ohio, already had an American passport. When we came to New York, a small apartment was rented for us in Tudor City. But very soon we went to Cold Spring Harbor on Long Island, where I was to give, as mentioned, a talk on all I had found with pyruvic oxidation (14). It was a great meeting with the Cori's, the Needham's and many American and other European

22

Fritz Lipmann

scientists. In the middle of this meeting, the second world war started and the Europeans were eager to hurry home. We stayed on in Cold Spring Harbor and rested a little after all the excitement. Returning to New York, we moved into one of the two large laboratories reserved for Dean Burk, Richard Winzler, and myself. Our work was then concerned with cancer. At that time, the surprising claim by F. Kögl, a well known Dutch biochemist, that cancer tissue contained D-amino acids, was causing much excitement. While the evidence was not very convincing, it was difficult to disprove. I proposed to assay hydrolysates of cancer tissue with the D-amino acid oxidase to obtain an undisputable result. I prepared a quite active enzyme and the result was definitely negative (15,16). This was seemingly essential, but not very interesting work. I met here again Rollin Hotchkiss, whom I had first encountered in Denmark. He was now working nearby at the Rockefeller Institute with René Dubos. He had isolated the two earliest bacterial polypeptide antibiotics, tyrocidin and gramicidin, from a Bacillus brevis as a quite pure mixture. Having the D-amino acid oxidase handy, we thought the hydrolysate of these antibiotics was worth analyzing for a possible presence of D-amino acids. I was quite surprised and pleased to obtain a vigorous 0 2 -uptake, which clearly indicated their presence in sizable amounts (17). After finishing, these refreshingly successful experiments during the spring of 1940, we went for a vacation to a nice place on Lake Iroquois in Vermont, not far from Burlington. I had time there to mediate on the metabolic appearance of a compound like acetyl phosphate. It had become clear from Lundsgaard's discovery (18) that the energy for the muscle contraction apparatus was by way of ATP (19,20). On the other hand, Schoenheimer and colleagues had found deuterated acetate to be transferred by, presumably, biosynthetic pathways into amino acids, fatty acids, and steroids. In the bacterial system, I had observed that acetate, by action of an acetokinase, was in equilibrium with ATR Therefore, acetate could be activated by pyruvate oxidation or by transfer of phosphoryl from metabolically formed ATP and used for anabolic purposes. Here it appeared then that acetyl phosphate might in animal tissues be a prospective acetyl donor in the biosynthesis of essential metabolites. I proposed the generalized use of ATP as energy carrier in an essay I started in 1940 in Vermont. It was called "Metabolic Generation and Utilization of Phosphate Bond Energy" (21). In this paper, I introduced a special sign, the socalled squiggle ~ for energy-rich phosphate, ~ P. It is now much used for other energy-rich linkages, e. g. thioester. It has been helpful to define group activation of the C ~ Ρ link either to donate Ρ to ADP or of the carboxyl group for acetyl donation. I argued that generally ATP was active in group activation, as of amino acids in protein synthesis; I used as presumptive examples the activation of acetate in fat, steroid, and amino acid synthesis, as indicated by the mentioned experiments by Schoenheimer and colleagues. Thus, a general use of ATP in biosynthesis of macromolecules was foreseen. The transfer of the phosphoryl potential from ATP, or indirectly by ~ Ρ donation from creatine phosphate, to energize muscle to contract was one of the prime examples of the role of ATP in energy transmission. During the last part of my stay with Burk, I had returned to work on the acetylation problem. In the middle of 1941, my two years with Burk at DuVigneaud's

A Long Life in Times of Great Upheaval

23

laboratory ended and Burk then moved to Washington. After many unsuccessful trials, I eventually landed with a Ciba fellowship at the Department of Surgery of the Massachusetts General Hospital (M.G.H.) in Boston, one of the hospitals connected with Harvard Medical School. The grant was given to Oliver Cope, who was interested in endocrinology. Therewith, I entered the unusual Department of Surgery led by Edward Churchill. It had a quite potent research floor and became the forerunner of a development of great expansion into research at the M.G.H. Oliver Cope, the second in command, gave me a spacious room. The appearance of the much discussed essay on phosphate bond energy increased my reputation in the biological sciences. Quite a few biologists and chemists recognized the essay as an illumination of the relationship between metabolic energy production and its utilization. I was very pleased to find the main idea in this essay paraphrased in Bernal's delightful little book, The Physical Basis of Life (22), in the following manner: Bernal wished to avoid troubles with defining life by limiting it partially to a common material, the proteins, and one common physicochemical process, a stepwise catalysis of organic compounds carried out practically isothermally by quantum jumps between 3 and 16 kilocalories/M, comparing it with the jumps of ca 300 in laboratory chemistry. In Figure 2,1 like to express differently the essentially same energetic cycle. My paper on phosphate bond energy also induced Dr. H.A. Barker, the well known Berkeley bacteriologist, to ask in 1942, right after my arrival at the M.G.H., to work for a year with me. Bacteriologists started then to find acetyl phosphate formed in other organisms, i.e. Utter & Werkman (23), and Koepsell & Johnson (24) in Clostridia. Werkman invited me for a few months to the

Group Potential Calories

12000

9000

Energy rich

6000

3000

0 Figure 2

HBond OxyEster Inorganic

Scheme for group activation in group potentials between 3 and 12 kcal.

24

Fritz Lipmann

Iowa Agricultural Experiment station at Ames, where I increased my knowledge of bacteriology. Most important was the increasing support of my work by the Commonwealth Fund, which allowed me to get technical assistance. I was lucky to find Constance Tuttle, who helped me a great deal for many good years. Fortunately, I had brought from Copenhagen a good supply of Lactobacillus pyruvic oxidase and had in the second year in New York found time to work out a method to determine differentially acetyl phosphate in presence of inorganic phosphate (P¡). With this now, I definitely identified the acetyl phosphate as the product of pyruvic acid oxidation in the lactobacillus.

Massachusetts General Hospital: Coenzyme A I learned more easily to prepare acetyl phosphate (Ac ~ P) from silver monophosphate and phosphorus oxychloride, and then most urgently needed an assay system for acetyl transfer in animal tissue. For this purpose I isolated a very potent enzyme from pigeon liver extract suitable for easy colorimetric determination of the known acetylation of sulfonamides and aromatic amides in general (25). With this assay, I first tried ATP + acetate, which I expected to form acetyl phosphate as in microorganisms, and found it active. But then I tried acetyl phosphate itself, which unexpectedly in animal tissues was almost explosively hydrolyzed by a heat-stable enzyme. Yet, if Ac ~ Ρ would have been active in transacetylation, since it could be used under conditions where some reaction should have been obtained, it was rather convincingly nonreactive. It appeared, however, that on autolysis the liver extract easily lost activity with ATP + acetate. This indicated the activity of an acetyl carrier, since boiled liver extract caused full reactivation. Our group, with increasing help from Commonwealth funds, had been fortified by the addition of two excellent co-workers, Drs. Nate Kaplan and Dave Novelli. With further help from the Upjohn Company, we obtained a pure enough preparation of the coenzyme to carry out a preliminary analysis. We observed the presence of adenylic acid and of a sulphydryl-containing compound. All of the known coenzymes were inactive. We were most eager for an analysis to check on the presence of a vitamin in this new coenzyme preparation. From two pharmaceutical companies we obtained negative answers. However, encouraged that Roger Williams, the discoverer of pantothenic acid, had considered it likely to be involved in intermediary metabolism, I sent him a sample of the coenzyme. He gave it to Dr. Beverley Guirard, who was familiar with the difficulties of finding pantothenic acid in crude tissue extracts, and instead hydrolyzed it for assay of the /J-alanine part of pantothenic acid. She found quantities of it in our coenzyme equivalent to its content of adenylic acid. We found, then, an enzyme in liver extract together with alkaline phosphatease to liberate all pantothenic acid from CoA (26). Concurrently, a cofactor for acetylation of choline was found by Nachmansohn & John (27) and Feldberg & Mann (28) in their work on acetylation of choline with brain extract plus acetate and ATP. We tested our purified coenzyme on this acetylation and found it to be active in choline acetylation with dialyzed brain extracts (29). Therefore, we called it CoA, with "A" for

A Long Life in Times of Great Upheaval

25

activation of acetate. Of great importance, however, would have been the identification of the SH-derivative in CoA, which we ovserved, but neglected. Eventually, it was identified in Snell's laboratory as part of the Lactobacillus bulgaricus factor (LBF) needed by this organism instead of pantothenic acid. They found it to be a peptide of pantothenic acid with thioethylamine and called it pantetheine. In view of our observation of an SH-derivative in highly purified CoA, they considered panthetheine to be a precursor of CoA. Meanwhile, Lynen had attempted to isolate not free, but rather acetyl CoA. Our observations had convincingly shown that acetyl CoA carried an activated, or let us call it energy-rich, acetyl, but we had not identified the energy-rich link between CoA and acetyl. Lynen & Reichert (30) used our sulfonamide acetylation system for assay on purification of acetyl-CoA and eventually found that it was linked to CoA by a thioester link. This important observation introduced the thioester bond as a new energy-rich bond. With the identification of pantetheine (pa) as part of CoA, the link from it to adenylic acid, presumably by a pyrophosphate bridge, was still to be defined. For this purpose, Baddily & Thain had synthesized the likely 4'-phosphate of pantetheine on the terminal of its four hydroxyl groups. A pure preparation of D-Lpantetheine 4'-phosphate was synthesized in Baddily's laboratory using a pigeon liver enzyme fraction isolated by Levintow & Novelli, and called pantetheinekinase. This was condensed with ATP by Baddily et al. (31) yielding a 45% conversion into CoA. Then the same experiment was carried out with pure D + 4'phosphopantetheine yielding an 82% conversion. This confirmed the complete structure of CoA with the third then unsure phosphate on the ribose of adenylic acid identified in Kaplan's laboratory (32) as being in 3'-position by use of 3'-specific Rye grass phosphatase. Chemical structure of CoA. This now well-known structure was presented in a review that also surveyed the CoA linked acyl-transfer reactions elaborated by Novelli, Kaplan, Soodak, Stadtman, Klein, and others in this laboratory (33). The CoAlinked reactions included fatty acid and steroid synthesis and, most importantly, citrate synthesis from oxaloacetate. The latter was the definition of the role of "active acetate" in the initiation of the citric acid cycle of H.A. Krebs. We recognized further connections with CoA-linked reactions after Stadtman joined us. The Stadtman-Barker bacterial transacetylase that had been identified by Ac ~ P^--32P exchange and by arsenate-induced hydrolysis of acetyl phosphate was found to be CoA dependent (34). Thus, this transacetylase was a bacterial type of acetyl CoA synthesis. By efforts largely of Jones, Black, Hoagland, Novelli, and Chou, the complex CoA-ATP-acetate reaction had been found to result in what we called a pyrophosphate split of ATP, which Paul Berg later showed to yield enzymebound acetyl-adenylate (35). In very informative experiments, Chou (36) separated by combined acetone and protamine precipitation the donor ATP-CoA-acetate reaction and the acceptor enzymes for arylamine acetylation, citrate and acetoacetate synthesis. There also could be shown that the bacterial transacetylase with CoA and acetyl phosphate could with pigeon acceptor enzymes replace the ATP-CoA-acetate reaction. Thus, it was shown that the activation of acetate, as well as its utilization, was due to

26

Fritz Lipmann

independent enzyme systems and that a mixture of bacterial activation and animal utilization enzymes could easily interact. When Doudoroff came to the laboratory, he and Stadtman (37) made use of such a mixed system to show that with [ 14 C]acetyl phosphate as donor, by way of transacetylase, the synthesis of acetoacetate is effected by a liver enzyme, and two [ 14 C]acetate residues appeared in acetoacetate. This result made us conclude that in this condensation both acetyl residues had to be activated. In the meantime, many laboratories had become interested in CoA activity and we were ready to leave the field with the observation, with Srere (38), of a citrate split by ATP and CoA, yielding acetyl CoA and oxaloacetate in pigeon liver extract we made as our last contribution. The complex problem of mechanism of this reaction was resolved by Walsh & Spector (39). The overall split of ATP here is to ADP and P¡. Srere has since also elaborated extensively on its mechanism. In the fall of 1953, something extraordinary happened: I received notice that I was awarded the Nobel Prize for Medicine and Physiology for my work on Coenzyme A. I was paired with Hans Krebs presumably because with acetyl CoA, the "active acetate" that starts the citric acid cycle had been defined. I have enlarged on that somewhat in a book, Wanderings of a Biochemist (40), which includes a photograph of my wife and Stephen, our son who was born in 1945, and myself, all smiling on our departure to Stockholm. That is now over 30 years ago, and although the event has lost some of its glamour, I still look back fondly on it.

Role of Phosphorylation in Other Group Activities The metabolic appearance of acetyl phosphate had led me to a consideration of the general use of group activation by way of phosphorylation as a common intermediary reaction in biosynthesis. To document this, we have been searching for such cases of group activation.

Carbamyl phosphate A phosphorylation seemed to be involved in the first step of the Krebs' urea cycle, the conversion of ornithine and ammonium carbonate τ-» carbamate and then to citrulline. I attended a Gordon conference, where there were a number of groups reporting on the phosphorolysis of citrulline ( + P ¡ + ADP) yielding ATP. This indicated a presumably unstable carbamyl phosphate as intermediary. We obtained, therefore, a culture of Streptococcus faecalis R, active in the phosphorolysis of citrulline. Mary Ellen Jones incubated its extract with ammonium carbonate and Penolpyruvate + a pure pyruvate kinase, and also the same mixture with addition of ornithine. In the mixture without ornithine we found an unstable Ρ that was not too quickly hydrolyzed in the Fiske-Subbarow Ρ determination, but was destroyed by one minute at 100°C in 0.01 Ν HCl. This differential was used for determining the unstable intermediary, and after 30 min incubation 0.75 μΜ of such phosphate was formed, which was presumed to be carbamyl phosphate. On addition to the same

A Long Life in Times of Great Upheaval

27

mixture and ornithine, citrulline in amounts nearly equivalent to the P¡ liberated from phosphoenol pyruvate was found. Thereafter, realizing the difficulty of obtaining enough unstable compound to analyze it, we approached Leonard Spector, a clever chemist with the Huntington Laboratory, one floor below us. He had the brilliant idea to try cyanate as a phosphate acceptor in the hope that isocyanate by incubation for 30 min at 30 °C would condense with phosphate. This proved to be very successful (41,42) and gave the opportunity to synthesize quite pure lithium carbamyl ~ Ρ by repeated alcohol precipitation. We also succeeded with the carbamylation of aspartic acid in rat liver extract with carbamyl phosphate as the donor. This is the first intermediary in the synthesis of pyrimidine and its nucleotides. Cohen & Grisolia (43) working with animal tissue, had obtained an unidentified ATP-linked precursor that had been found active in citrulline and carbamylaspartate synthesis.

Sulfate activation It is not surprising that in the 1950s ATP was applied in many cases of group transfer, including sulfate activation. In my review in Science (44) I summed up the early work by DeMeio, Bernstein, and McGilvery, and others who obtained sulfate transfer by addition of ATP in a cell-free system. It was also recognized that the activation and transfer using phenol as acceptor was due to a two-step reaction. Bandurski had gone further and isolated from yeast two fractions that had to be combined for activation. Hilz, in our laboratory, using 35 S, marked sulfate and found that the primary reaction yielded adenylsulfate (APS) and pyrophosphate (PP¡) analogously to the ATP-CoA-acetate reaction. After Hilz left, Phil Robbins took over and succeeded in complete analysis of what we now call the active sulfate. He found the second step to be a second phosphorylation of the APS to form PAPS, which using Kaplan's 3'-specific Rye grass phosphatase, was identified as in the 3'position of the ribose of A. Robbins describes the methodology he worked out for preparation of sulfate activating enzymes from liver and Baker's yeast. Therefrom he also isolated PAP[ 35 S]. Most of the work done on this problem, by us and others, is included with references in a review by Gregory & Robbins (45). Chick embryonic cartilage was introduced for the assay of enzymatic transfer of sulfate into chondroitin sulfate (CHS) with d'Abramo (46). A survey of our work on sulfate activation and transfer was described in detail in the mentioned review in Science (44). This includes work by Irving Goldberg on the ceribroside sulfate in brain. The latter is of clinical interest as, in man, a debilitating deficiency disease is known where the lysosomal hydrolase is missing and in all organs large amounts of this ceribroside sulfate accumulates. We found another aberration in a chondrosarcoma that was observed. There are normally three CHS—A, B, and C-differing in the manner of sulfate esterfication. We obtained a chondrosarcoma from SloanKettering Cancer Institute that had sizably increased activity to produce CHS—C only (47). Recently, a frequent binding of sulfate to tyrosine in protein has been found that earlier was observed only in rare cases (48). We are now starting to explore possible differences in this process between normal and malignant cell types.

28

Fritz Lipmann

Move to Rockefeller University In the mid-1950s I became restless at the Massachusetts General Hospital. There was, as I have experienced, a certain friction between the teaching and research departments in the same university; a certain envy by the former of the latter. Luckily, just at that time Dr. Detlev Bronk had become the president of the Rockefeller Institute, which he converted into a graduate university, with relatively low demand for teaching duties. He invited me to join the university as a member, or as it was then already called, a professor. The transition was very easy because John Gregory, who had been at the Rockefeller Institute and had joined me at Massachusetts General Hospital, moved back with me and excellently organized the laboratory facilities there by moving a half-year ahead of us early in 1957. When we joined him a few months later, we were able, without interruption, to start work there. Quite a few of my co-workers did move with me and the laboratory was generously supported by N.I.H., and our support was transferred with us with one exception that I regretted. Dr. Warren Weaver, director of the Rockefeller Foundation, had a few years earlier spontaneously offered me a personal grant of $ 5,000 a year to make use of however I wanted, without need for any report. This had been the nicest money ever given to me; unfortunately, it could not be transferred between these two Rockefeller institutions.

Protein synthesis After moving to Rockefeller we took up almost exclusively the analysis of protein biosynthesis. Already in my review of 1941, in the discussion of group activation for polymerization, I had proposed that amino acids were likely to be activated by phosphorylation of the carboxyl group. We had learned then that in animal tissue acetate activation was more complex, and was due to a pyrophosphate split in ATP, rather than a phosporylation by ATP to Ac ~ Ρ as in bacteria. Its reversibility was shown by pyrophosphate exchange in what we called the ATP-CoA-acetate reaction. Hoagland then had been working in our laboratory on acetate activation. After returning to Zamecnik's laboratory, he assayed the supernatant of liver homogenates, which supplanted polypeptide synthesis for pyrophosphate exchange with ATP in presence of amino acids. As in the ATP-CoA-acetate reaction he found a pyrophosphate (PP¡) exchange with ATP and amino acid. This indicated an activation of amino acids by an AMP-PP¡ split. Later, for both acetate and amino acid activation, Paul Berg showed that it involved an enzyme-bound acyl-adenylate intermediary. The Hoagland-Zamecnik group then made rapid progress in the field of amino acid activation by confirming Crick's prediction that short RNAs containing presumably amino acid-specific triplets of nucleotide codons would join the amino acids. With this clever formulation, Crick foresaw that positioning of amino acids (aa) in a prescribed order in proteins would be like the construction of the double helix by polynucleotide synthesis, a function of a pyrimidine-purine hydrogen bonding between the specific sequence in or attached to the units. The need for such a

A Long Life in Times of Great Upheaval

29

hydrogen bonding between triplets of pyrimidine and purine nucleotides was deduced from the number 20 for the amino acids, for which doublets were insufficient. These ideas were in the air as it became clear that the ribosome could not be responsible for sequencing the aa's. There was found in the soluble supplement to ribosomes a soluble RNA first called sRNA, soon to be changed to tRNA (t for transfer). They contain about 70 nucleotides and are specific for particular amino acids (49). It appeared to have a common O-terminal of two cytidylic acids (C's) terminating with adenylic acid (A), and that the amino acids were all linked to the O-terminal adenylic acid. Zachau et al. (50) isolated adenosinelinked amino acid from the ribonuclease digest of liver amino acyl transfer RNAs. We use the now commonly preferred tRNA. In summary, the ATP-linked activation of amino acids, with the amino acyl adenylates as intermediary, and the discovery of the amino acid binding eventually to a tRNA in an energy-rich linkage by way of the glycolic 3'-OH or either 2'- or 3'OH on the ribose (51,52). Furthermore, we had confirmed the expected amino acid specificity of activating enzymes and proposed the base pairing by hydrogen bonding. Crick's prediction was further confirmed by isolation of a messenger RNA (mRNA), discovered in three laboratories simultaneously; this includes Sam Weiss, who, when he left us for the University of Chicago, told me he would look for it. The further development to 1962 is reviewed by Nathans et al. (53), by examining the protein synthesis from amino acyl tRNA. They discuss the function of what is then still called a supernatant factor. Its bacterial and eukaryotic specificity is described and contrasted with an apparent nonspecificity of amino acid transfer RNAs. There was emphasized the specific participation of GTP discovered by Keller & Zamecnik as an important observation although the role of it was then still unknown. Furthermore, the increasing evidence for a messenger RNA serving as a template for protein synthesis is mentioned, referring to the discovery of Nierenberg & Matthei of synthetic polynucleotides to serve as templates for polypeptide synthesis on Escherichia coli ribosomes. They first discovered that polyphenylalanine is synthesized on poly U.

Polypeptide chain elongation The understanding of the mechanism of polypeptide synthesis had proceeded then to distinguish three phases of the process, eventually found in all polymerization reactions producing chains with specific subunit sequence: (a) initiation, (b) polymerization, and (c) termination. In protein synthesis the 20 amino acids found in most proteins are linked on a ribosome in an initially straight chain. This eventually folds up into the final three-dimensional structure. We worked mostly on the middle phase of chain elongation in E. coli homogenates. Choosing a simple system, we worked predominantly on polymerization of phenylalanine (phe) on a template of polyuridylic acid. There the mechanism of elongation to polyphenylalanine chains on ribosomes could be studied rather independently of the complex initiation and termination periods.

30

Fritz Lipmann

Separation of the three elongation factors, Tu, Ts, and G The supernatant of the E. coli homogenate contained, as shown by Nathans in our laboratory, the elongation complements that could be separated by Allende et al. into subunits by DEAE chromatography. They obtained two fractions: a heatstable factor Τ and a heat-unstable G. These were more definitely identified by Nishizuka by using DEAE Sephadex and a potassium gradient. This yielded three peaks, (a) a now unstable factor T, (b) a mixture of factors G and T, and (c) a peak that contained the now stable factor G, so named for its inducing a ribosome-linked GTP split. In this chromatography, in contrast to the earlier one where heat instability was found in the late appearing factor G, it had now shifted to the early T. We interpreted this change as indicating that our now isolated factor Τ contained, in addition to the earlier stable Τ we now called Ts, an unstable Tu, which in the earlier chromatogram had joined with the now stable G. This proposition was confirmed by experiments with Lucas-Lenard (54) who separated Τ into Tu and Ts. Using mostly our simplified poly-U-linked poly-phe synthesis, the addition of a new amino acid is catalyzed in four phases on the ribosome. A phe-tRNA may initiate the process instead of the complex initiation involving special I-factors and in bacteria as a peptidyl substitute, the formyl methionyl-tRNA having been elaborated in other laboratories. Lucas-Lenard, in this laboratory, found, however, with the simplified phe-system that TV-acetyl-phe-tRNA was a better initiator than plain pheRNA (55), and this was used in most experiments. In Phase 1, the start of the polymerization cycle, the acetyl-phe-tRNA is bonded to the left, the "donor" or Dsection of the upper 50s part of the ribosome and with an A-triplet on tRNA to a Utriplet on poly-U on the 30s part of the ribosome. The initiating acetyl-phe-tRNA serves in Phase 2 with its activated carboxyl group linked to tRNA from D as donor to the amino group of added phe-tRNA to Α-site. Thereby, the carboxyl link of acetyl-phe to tRNA gives way to a new peptide bond as polymerization progresses by the addition of the new phe-tRNA. It is followed by Phase 3, the translocation from A to the D site of the newly elongated peptidyl-tRNA, which was still sitting on the "wrong", or the Α-site. This translocation is catalyzed by the G-factor with GTP splitting, promoting the translocation after a new addition from acceptor to donor site. The translocation and removal of the free tRNA released as discussed from acetyl-phe-tRNA by transpeptidation simultaneously with translocation is completed in transition to Phase 4. Phase 4 is identical with Phase 1, except that the peptide chain is elongated by one phe. I have described here in an abbreviated form the work here and elsewhere between 1962-1969 on protein biosynthesis, which is discussed in detail in my 1969 paper in Science (56). The most elaborate description from our laboratory of protein biosynthesis in 1971 was presented by Lucas-Lenard and myself (57). We attempted there not only to describe, but also to rationalize all the knowledge available up to 1971. Much has been added since and much additional information is found in the book Molecular Mechanism of Protein Biosynthesis (58). Most recently, an illuminating paper by Kaziro (59) on this topic has appeared, entitled "Molecular Mechanism of Protein Synthesis and an Approach to the

A Long Life in Times of Great Upheaval

31

Mechanism of Energy Transduction." Kaziro describes and discusses new results that led him to conclude that the GTP-linked prokaryote elongation steps involving both the Tu-Ts pair binding the amino acyl-tRNA to the ribosome, and the Gpromoted translocation of the newly elongated polypeptidyl-tRNA are due to protein transformations caused by binding of the terminal ~ Ρ in GTP. The latter, involving apparently a mechanical move of a ligand on the ribosome, is proposed to be comparable to the role of ATP in muscle contraction. This proposition, founded on Kaziro's new results, goes far to explain the role of GTP in protein synthesis. Also, in the book mentioned before, isolation of elongation factors analogous to TuTs and G is described in the eukaryote supernatant fractions. They are, however, not interchangeable as mentioned before (58).

Increase of phosphoryl potential of serine-O-phosphate and tyrosine-O-phosphate by binding into protein Serine phosphate. In earlier work, I had become rather familiar with the serine phosphate in phosvitin and casein. Thus, we expected this phosphorylation to be originally catalyzed by a phosphate transfer from ATP. We tried with Murray Rabinowitz to reproduce this reaction with the egg yolk protein (60). In this protein, nearly every second amino acid is serine, and serine and phosphate are present in nearly equal amounts. The compound used had a phosphate content of 10%, similar to the one I had used in earlier experiments. In order to obtain a phosphorylation of phosvitin with 32 P, one had to remove part of the phosphate. For this purpose, we used a spleen phosphatase and for rephosphorylation we obtained good results with purified protein phosphokinases prepared either from brewer's yeast or calf brain. Dephosphorylation with alkali destroyed the serine in protein and, therefore, the enzymatic method had to be used. The results of the reverse transfer to ADP from 2 0 % - 4 0 % enzymatically dephosphorylated phosvitin may be summarized as follows: A strong affinity was already indicated by near equality of rephosphorylation from ADP to ATP and A D P + hexokinase + glucose to glucosephosphate, while, generally, the latter with A D P as catalyst is more effective. Furthermore, the specific activity of rephosphorylated phosvitin and the phosphate transferred back to A D P differed widely in favor of that of the terminal Ρ in ATP thus formed. This made it clear that the protein Phosphokinase discriminated for a transfer from special sites in phosvitin. The results actually indicated that the newly transferred phosphates may be the ones preferred for return. This is also confirmed by the finding that dephosphorylation of the Ρ in phosvitin of over 70% abolished the reverse transfer of phosphate by protein Phosphokinase. We were then reminded of the finding in Sanger's laboratory (61), both in phosvitin and casein, of serine phosphate sequences of up to six and that the phosphoryl potential of phosphate in these blocks might be higher. This suggestion was recently confirmed by Fisher et al. (62), who extended our experiments by using in addition to phosvitin a tryptic digest of casein and isolated the ser-leu-ser-ser-ser peptide. This they showed, in contrast to free serine phosphate, to transfer easily in presence

32

Fritz Lipmann

of phosphokinase its phosphate to ADP. Therefore, it appears that in the serine in the blocks in phosvitin and casein the phosphate becomes energy rich. Tyrosine phosphate. Dr. J. M. Sturtevant of Yale University and Dr. R. Epand of McMaster University, with samples of tyrosine (tyr-P) phosphate, prepared by Fukami (63), found the Δ Η of tyr-P-hydrolysis to be — 2.8 kcal mole, which is considerably below the value of Δ H° for y Ρ in ATP that Sturtevant earlier determined. Tyrosine phosphate had become of great interest because Erikson & Brugge (64) had discovered that the translation product of the transforming gene of Rous sarcoma virus (RSV) R N A was a protein phosphokinase specific for transfer of phosphate (P) to tyrosine in proteins (65). The transphosphorylation had been found first in transfer of Ρ to an immunoglobulin (IgG) to sarcoma-bearing rabbits (TBR). In this paper the TBR-IgG was to be used as acceptor of Ρ by use of the transferase isolated by Fukami from tumor tissue. Fukami & Lipmann (66) could show first that the phosphorylation of specific IgG was easily revertible to ADP, using this system for determining the equilibrium constant, Keq, of the reaction between ATP + IgG •• A D P + IgGP as described in the paper. The value of K t q at optimal pH was used to determine the change of free energy, Δ G°' of hydrolysis of Δ G°' enzyme-bound tyr-P. It was found to be — 9.48 kcal/mole. This is very near to the Δ G°' of the hydrolysis for y Ρ of ATP at analogous conditions, which is approximately 10 kcal/mole. The result indicates that the tyr-P is bound in this protein by an energy-rich bond in contrast to free tyr-P, to judge from its low heat of hydrolysis of tyr-P.

Bacterial Production of Polypeptide Antibiotics by a Thiol-Linked Activation and Polymerization In 1963,1 was invited to participate in a conference on the Origins of Prebiological Systems. I presented there a paper entitled "Projecting Backward from the Present State of Evolution of Biosynthesis" (67). There was much discussion then about the priority of proteins or polynucleotides in the earliest prebiotic evolution. In any case, the production of proteins appears to be the final goal in the polynucleotide translation of the four-letter codes of D N A to messenger R N A , which is to be expressed into the 20-letter language of the proteins. Only by this expression does the translation product become usable as living matter. Yet, it remains likely that the polynucleotide system and the polyamino acid system should have to develop in parallel. Nevertheless, I found it worthwhile to search for an alternate, less complex mechanism for the production of polypeptides without the need of nucleic acids. When attempting to find such a mechanism, I noticed the recent report on an apparently nonribosomal polypeptide synthesis. Extracts exhaustively treated with RNase had been found in several laboratories (68) to synthesize bacterial antibiotic polypeptides. Thus, with Wieland Gevers and Horst Kleinkauf (69) we started the synthesis with the smallest of a group of related antibiotics, the gramicidin S (GS) in

A Long Life in Times of Great Upheaval

33

extracts of a B. brevis. GS is composed of two identical pentapeptides that cyclize head to tail. The amino acids (aa) are activated by the reaction of ATP + aa = A M P ~ a a + PP¡. The synthesis of the half molecule, the pentapeptide sequence of D-Phe-Pro-ValOrn-Leu is carried out on two proteins. The L-Phe is ATP-activated and linked to a thiol on a protein of Mr 100,000 that we call the light enzyme. The four following amino acids are thiol-linked to a polyenzyme that contains one mole pantetheine (Pan) bound to it by a protein of M r ~ 17,000- 20,000 first isolated from tyrocidinsynthesizing polyenzymes (70). The MT 100,000 light enzyme contains a racemase and when charged with L-Phe initiates the synthesis of GS with the D-Phe. This initiation process had been further analyzed (71) and it begins by the Pan of the polyenzyme displacing by higher affinity the Pro from its -SH and carrying S ~ Pro to the initiating Phe that then reacts with its thioactivated carboxyl group with the amino group of proline to form the first peptide link Phe-Pro carried on the Pro + thio-linked to pantetheine. The Pan-Phe-Pro then reacts with Pro's neighbor Val's N H 2 to form the tripeptide temporarily hanging by the following protein and transthiolates to Pan: O II S ~ C—Val—Pro—Phe This alternation of transthiolation and transpeptidation yields the pentapeptide half of GS and cyclizes head to tail by the S-linked carboxyl of leucine reacting with the free-NH 2 of Phe. The alternation with transpeptidation is presented in Figure 3. The process is the same in GS, the later-analyzed tyrocidin (Ty), and linear gramicidin (LG), which was partially synthesized on polyenzymes. Actually, the details were most strongly recognized with Ty analysis. We obtained a more complex, but analogous biosynthesis of Ty with its decapeptide cyclyzing much more slowly. Tabulating the enzyme structure of both the GS and Ty we get: Antibiotic

No.

M, of enzymes

Amino acids activated and fixed in sequence

Subunit M, per amino acid

Pan

GS

1 2

100 280

D-Phe Pro-Val-Orn-Leu

100 70

None 1

Ty

1 2

100 230

D-Phe Pro-Phe-D-Phe

100 76

None 1

3

440

Arn-Gln-Phe-Val-Orn-Leu

Decacycle (slow)

74

1

Pantetheine is a rather long molecule of ~ 20 Â; according to the clever proposal by Lynen it transacts a specific bond formation as a swinging arm as shown in Figure 3. This process is comparable to the translocation of the peptidyl from a donor to acceptor site on the ribosome. Linear gramicidin is not yet quite resolved. Kurahashi's laboratory (72) discovered the manner of its initiation by finding that the initiating valine is formylated. The

Fritz Lipmann

34

L G structure is: jV-formyl - Val - Gly - Ala - D-Leu - A l a - D-Val - Val - D-Val Ethanolamine — Trp — D-Leu — Trp — D-Leu — Trp — D-Leu—Trp Bauer et al. (73) carried out its synthesis in a biosynthesis with thioethanolamine synthesized onto it by using a high concentration of it reacting chemically with the following structure and releasing the peptide from the enzyme: 0 II

Enz—S—C—tryptophan—terminal I would like to add here that, with the GS system, it has been possible to react the light enzyme with preactivated phenylalanine by its chemical thioesterfication without adding A T P . To recognize the synthesis of GS we combined the thioester of cold phenylanine transferred to light enzyme with the heavy enzyme that had been charged with radioactive amino acids. It appeared, thus, that the Phe-thioester was able to transthiolate to the enzyme-SH and to initiate synthesis. Trials with other amino acyl thioesters of Val or Leu to charge the heavy enzyme were unsuccessful. We began with Wieland Gevers, and soon afterwards Horst Kleinkauf joined us. Roskowski came somewhat later, and Lee has done a great deal of very fine work during the last few years. Bauer worked quite successfully on the chemicalenzymatic synthesis of linear gramicidin, as did Akers et al. (74) on the analysis of charging its two first polyenzymes with amino acids.

^

S-FC· A*· A2· A'-NH

Figure 3 A schematic representation of transpeptidation and transthiolation, mediated by enzyme-bound pantetheine in antibiotic peptide biosynthesis.

A Long Life in Times of Great Upheaval

35

We initiated these experiments with the aim o f finding a less complex m e t h o d for peptide synthesis. In prebiological evolution this might have preceded the present complex protein synthesis. I a m n o longer t o o happy with this proposition, but recently I w a s encouraged not to forget about it. I received a note from Dr. Onsager (cited in 75) indicating that since his retirement f r o m Yale he had m o v e d to M i a m i and become productively interested in the origin o f life. In a talk at a N o b e l meeting in Lindau in 1973, he apologized in his note for having discussed a thioesterification o f carboxylic acids for activation as an early event in prebiology. H e had gotten this idea without k n o w i n g about our work on antibiotic synthesis. A t a later meeting, I had the opportunity to talk with him. I was pleased with this contact, since he h a d spent a longer period at Rockefeller, where I had observed and liked him, but had never spoken with him.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35.

Rona, P., Lipmann, F. 1924. Biochem. Z. 147, 163-73 Kohler, R. 1971. J. Hist. Biol. 4, 35-61 Harden, A. 1932. Alcoholic Fermentation. London, Longman Group Fiske, C., Subbarow, Y. 1927. Science 65, 401-3 Eggleton, P., Eggleton, P. 1927. Biochem. J. 21, 190-95 Lipmann, F., Meyerhof, O. 1930. Biochem. Ζ. 227, 84-109 Plimmer, R.H.A., Bayliss, W.M. 1906. J. Physiol. 33, 439-61 Lipmann, F. 1980. Trends Biol. Sci. 5, R 3 - 4 Lipmann, F., Levene, P.A. 1932. J. Biol. Chem. 98, 109-14 Lipmann, F., Lohmann, Κ. 1930. Biochem. Ζ. 222, 389-403 Lipmann, F. 1975. Mol. Cell. Biochem. 6, 171-75 Lippman, F. 1942. In Symp. Respiratory Enzymes, pp. 48-73. Madison, Wis., Univ. Wis. Press Bloxham, D., Lardy, H. 1973. Enzymes 3, 240-74 Lipmann, F. 1939. Cold Spring Harbor Symp. Quant. Biol. 7, 248-59 Lipmann, F., Behrens, O., Rabat, E., Burk, D. 1940. Science 91, 21-33 Behrens, O., Lipmann, F., Cohn, M., Burk, D. 1940. Science 92, 32-34 Hotchkiss, L„ Dubos, R. 1941. J. Biol. Chem. 141, 163-69 Lundsgaard, E. 1930. Biochem. Z. 217, 162-209 Engelhardt, W. 1942-43. Yale J. Biol. Med. 21, 38 Banga, T.E. 1941-42. In Myosin and Muscular Contraction, ed. Α. Szent-Györgyi, 1, 563-73 Lipmann, F. 1941. Adv. Enzymol. 1, 99-162 Bemal, T.D. 1952. The Physical Basis of Life. London, Rutledge Utter, M. F., Werkman, C.H. 1943. Arch. Biochem. 2, 491-97 Koepsell, H.J., Johnson, M.J. 1942. J. Biol. Chem. 145, 379-85 Lipmann, F. 1945. J. Biol. Chem. 160, 173-90 Lipmann, F., Kaplan, N.O., Novelli, G.D., Tuttle, L.G., Guirard, B.M. 1947. J. Biol. Chem. 167, 869-70 Nachmansohn, D„ John, H.M. 1945. J. Biol. Chem. 158, 157-71 Feldberg, W., Mann, T. 1945. J. Physiol. 104, 8 - 2 0 Lipmann, F., Kaplan, N.O. 1946. J. Biol. Chem. 162, 743-44 Lynen, F., Reichert, E. 1951. Angew. Chem. 63, 47-48 Baddily, T., Thain, E.M., Novelli, G.D., Lipmann, F. 1953. Nature 171, 76-77 Wang, T.P., Shuster, L„ Kaplan, N.O. 1952. J. Am. Chem. Soc. 74, 3204-5 Lipmann, F. 1953. Bacterid. Rev. 17, 1 - 1 6 Stadtman, E.R., Novelli, G.D., Lipmann, F. 1951. J. Biol. Chem. 191, 365-76 Berg, P. 1956. J. Biol. Chem. 222, 1015-23

36

Fritz Lipmann

36. Chou, T.C., Lipmann, F. 1952. J. Biol. Chem. 196, 89-103 37. Stadtman, E.R., Doudoroff, M., Lipmann, F. 1951. J. Biol. Chem. 191, 377-82 38. Srere, P., Lipmann, F. 1953. J. Am. Chem. Soc. 75, 4874 39. Walsh, C.T., Jr., Spector, L. 1969. J. Biol. Chem. 244, 4366-74 40. Lipmann, F. 1971. Wanderings of a Biochemist. New York, Wiley 41. Jones, M.E., Spector, L. 1955. J. Am. Chem. Soc. 77, 819-20 42. Spector, L., Jones, M.E., Lipmann, F. 1957. Methods Enzymol. 3, 653-55 43. Grisolia, S., Cohen, P.P. 1952. J. Biol. Chem. 198, 189-202 44. Lipmann, F. 1958. Science 128, 575-80 45. Gregory, J.D., Robbins, P.W. 1960. Ann. Rev. Biochem. 29, 347-64 46. d'Abramo, F., Lipmann, F. 1959. In Int. Cong. Biochem. IV, 8, 36-42. London, Pergamon 47. Lipmann, F. 1966. Rheumatismus 38, 1 - 9 48. Huttner, W. 1982. Nature 299, 273-376 49. Davie, E.W., Koningsberger, V.V., Lipmann, F. 1956. Arch. Bioch. Biophys. 65, 21-28 50. Zachau, H., Acs, G., Lipmann, F. 1958. Proc. Natl. Acad. Sci. USA 44, 885-89 51. Lipmann, F., Hülsmann, W., Hartmann, G., Boman, H., Acs, G. 1959. J. Cell. Comp. Physiol. 54 (Suppl. 1), 75-88 52. Wolfenden, R., Rammler, D., Lipmann, F. 1964. Biochemistry 3, 329-38 53. Nathans, D., von Ehrentstein, G., Monro, R., Lipmann, F. 1962. Fed. Proc. 21, 127-33 54. Lucas-Lenard, J., Lipmann, F. 1966. Proc. Natl. Acad. Sci. USA 55, 1562-66 55. Lucas-Lenard, J., Lipmann, F. 1967. Proc. Natl. Acad. Sci. USA 57, 1050-57 56. Lipmann, F. 1969. Science 164, 1024-31 57. Lucas-Lenard, J., Lipmann, F. 1971. Ann. Rev. Biochem. 40, 409-48 58. Weissbach, H., Pestka, S., ed. 1977. Molecular Mechanism of Protein Biosynthesis. New York, Academic 59. Kaziro, Y. 1980. Molecular Biology and Biophysics, ed. F. Chappeville, A.L. Haenni, pp. 333-344. Berlin, Springer 60. Rabinowitz, M., Lipmann, F. 1960. J. Biol. Chem. 235, 1043-50 61. Williams, T.W., Sanger, F. 1959. Biochim. Biophys. Acta 33, 294-96 62. Lerch, K., Muir, L., Fisher, F. 1975. Biochemistry 14, 2015-23 63. Fukami, Y„ Lipmann, F. 1982. Proc. Natl. Acad. Sci. USA 79, 1693 - 97 64. Brugge, T., Erikson, R. 1977. Nature 269, 346-48 65. Hunter, T., Sefton, B. 1980. Proc. Natl. Acad. Sci. USA 77, 1311-15 66. Fukami, Y., Lipmann, F. 1983. Proc. Natl. Acad. Sci. USA 80, 1872-86 67. Lipmann, F. 1964. In The Origins of Prebiological Systems, pp. 259-80. New York, Academic 68. Lipmann, F. 1971. Science 173, 875 - 84 69. Gevers, W„ Kleinkauf, H„ Lipmann, F. 1969. Proc. Natl. Acad. Sci. USA 63, 1335-42 70. Lee, S., Lipmann, F. 1975. Proc. Natl. Acad. Sci. USA 71, 603-11 71. Lee, S„ Lipmann, F. 1977. Proc. Natl. Acad. Sci. USA 74, 2343 - 47 72. Akashi, I., Kurahashi, K. 1977. Biochem. Biophys. Res. Commun. 77, 259-67 73. Bauer, K., Roskoski, R., Jr., Kleinkauf, H., Lipmann, F. 1972. Biochemistry 11, 3266-79 74. Akers, H., Lee, S., Lipmann, F. 1977. Biochemistry 16, 5722-29 75. Lipmann, F. 1977. Biosystems 8, 231

Fritz Lipmann: In Memoriam1 Christian de Duve

Most of us are bricklayers. We are happy to add a stone to the edifice of science and we consider ourselves fortunate to contribute a cornerstone, or the base of a column, or the keystone of an arch. A rare few have the vision of an architect. They somehow see the whole building long before it is completed. Fritz Lipmann was such a visionary. Another was Albert Claude, his almost exact contemporary, who died 3 years ago. Like many biochemists of his generation, Lipmann came to science by way of medicine. He was born in Königsberg, the capital of East Prussia, at the turn of the last century. The son of a successful lawyer, he grew up in a happy and cultured family environment. His admiration for an uncle, "a very lovable man and a wellliked pediatrician," 2 led him to choose medicine as a career. His studies, performed partly in Königsberg and partly in Munich and Berlin, were interrupted by a 1-year spell in the army medical service. He saw the end of World War I on the French front and was later a witness to the murderous influenza epidemic that devastated Europe after the war. In spite of these grim experiences, he thoroughly enjoyed his studies and his early clinical activities. As he was to write 50 years later: "The biological education to which the observant student is exposed in medicine is a superior preparation for any career." 3 But when the time came to take up the practice of medicine, he became uneasy, troubled, as he said, "by the prospect of charging people money for trying to make them healthy." 2 He drifted to pathology; then, having enough of cadavers, he enrolled in a biochemistry course given by Peter Rona, a former associate of Leonor Michaelis. This course and a small piece of research in the then-fashionable area of colloid chemistry, followed by a short stay in the pharmacology laboratory of Laqueur, in Amsterdam, sealed his fate. He was going to become a biochemist. Conscious that the little chemistry he had been taught during his medical studies provided an insufficient basis for a career in biochemistry, he went back to school and spent 3 additional years learning chemistry while staying with his family in Königsberg. In 1927, at the age of 28, he finally felt ready, "eager to do real biochemistry." 3 Looking for a suitable laboratory, he chose that of Otto Meyerhof at the Kaiser-Wilhelm Institutes in Berlin-Dahlem, then the best center in Germany for the kind of biologically oriented biochemistry he wanted to engage in. He applied to Meyerhof and was accepted - an indication, no doubt, that his qualities were already apparent at that time.

1

2 3

This talk was delivered at a memorial concert held at The Rockefeller University on December 12, 1986. Reprinted with permission from FASEB J. 1, 3 - 5 (1987). See LIPMANN, F. 1984. A long life in times of great upheaval. Annu. Rev. Biochem. 53, 1 - 3 3 . See LIPMANN, F. 1971. Wanderings of a biochemist. Wiley, New York.

38

Christian de Duve

Fritz Lipmann stayed only 3 years with Meyerhof, first in Berlin and later in Heidelberg. His work, which dealt largely with the inhibition of glycolysis by fluoride, was good enough for a Ph. D. degree, but hardly earth-shattering. Yet those 3 years played a key role in his personal development and may well have shaped the whole of his scientific output. As he mentions in his autobiography, "In the Freudian sense all that I did later was subconsciously mapped out there; it started to mature between 1930 and 1940 and was more elaborately realized from then on." 3 He does not elaborate, but we can try to guess. First, he met many of the great biochemists of that time, the first pioneers of cell metabolism: Meyerhof himself, Warburg, Neuberg, Embden, and some of the gifted younger scientists, such as Hans Krebs and Hans Gaffron, who worked with them. He was particularly influenced by Karl Lohmann, the discoverer of ATP, who taught him the phosphate ester chemistry that he was to use extensively in later life. A major preoccupation in those days was mapping the glycolytic pathway. But Meyerhof's interest in this problem was not just that of a chemist who wants to know how a reaction takes place. As a biologist, he also wanted to know what for. His most important contribution, that for which he shared the Nobel Prize in medicine with A.V. Hill in 1922, had concerned the role of glycolysis in muscle. Pursuing earlier observations by Fletcher and Hopkins, he had demonstrated that the conversion of glycogen to lactic acid is the source of energy for muscle contraction under anaerobic conditions, and established the quantitative relationship between lactic acid produced and work performed. Lipmann was actually in Meyerhof's laboratory when the devastating news came from Copenhagen that a muscle poisoned with monoiodoacetate could continue to contract and perform work for a while without producing any lactic acid. He was there when Einar Lundsgaard, who had made this discovery and found the phenomenon to be correlated with the hydrolysis of creatine phosphate, came to Heidelberg in 1930 to verify this correlation quantitatively. He was there also when Lohmann discovered that creatine phosphate provides the muscle with energy by way of ATP. The key concept of phosphate-bond enery must have germinated in his mind at that time. Then, we may speculate, is when his luminously perspective vision of life started to take shape. No doubt he had the unique kind of mind needed to conceive such a vision, but it was in Meyerhof's laboratory that he found the right climate for the seed to be planted and to develop. Only 10 years later, having, in the meantime, studied cell culture with Albert Fischer in Berlin, isolated serine phosphate from phosphoproteins in the laboratory of Phoebus Levene at the Rockefeller Institute for Medical Research in New York, discovered acetyl phosphate as an intermediate of pyruvate oxidation by bacteria at the Carlsberg Institute in Copenhagen, and finally joined Dean Burk in du Vigneaud's laboratory at Cornell University Medical School just before the outbreak of World War II, he wrote his celebrated review "Metabolic generation and utilization of phosphate bond energy", published in 1941 in volume 1 of Advances in enzymology. This paper should be required reading for any student of biology. It is a landmark. I remember the impression it made on me when I first read it. I was familiar with

Fritz Lipmann: in memoriam

39

Meyerhof's measurements and had just plodded through a painstakingly detailed review of the energetics of glycolysis by Dorothy Needham. It contained all the facts and it was about as exciting as an accountant's report. Then came Lipmann's paper. It was a revelation and everything fell into place. The notion of group potential, the "squiggle", the conversion of electron-linked to group-linked energy, the key role of group transfer in biosynthesis. It is all there, only to be elaborated on further in subsequent reviews in 1946, 1948, 1960, and 1968. From here on, the line of his research is perfectly straight. He himself has referred to it with characteristic modesty as "wandering, following one's instinct without knowing exactly where it will lead." 3 It could strike the superficial observer that way, as it meanders from acetylation to uncoupling of oxidative phosphorylation and the action of thyroxin, carbamyl phosphate, the synthesis of sulfate esters, polypeptide chain elongation, the formation of bacterial antibiotics, tyrosine phosphorylation. But it was no wandering. Fritz Lipmann knew exactly where he was going. He knew intuitively what to look for because he saw the broader picture. Occasionally the results did not come out exactly as he had anticipated. He left adenylyl acetate for Paul Berg to discover, the thioester bond of acetyl-coenzyme A to Fitzi Lynen, the prediction of tRNA to Francis Crick and its demonstration to Mahlon Hoagland and Paul Zamecnik and to Bob Holley. He has bemoaned this, exclaiming "how difficult it is to see the new because it is new and how badly one may be handicapped by preconceived notions," 3 and again "how a pointer in the right direction may be neglected if one's mind is made up to look in a different direction." 3 In fact, his mind was always in the right direction, but he sometimes started on a wrong alternative before discovering his mistake and switching over to the correct one. His choices of topics, likewise, were never haphazard. Ever curious for new facts, he read voraciously and attended lectures with an eagerness and enthusiasm that remained undiminished almost up to his last day. He would then take the facts home with him and mull over them, trying to fit them within some sort of logical framework born from his deep insight. Finally, he would come up with the decisive and original approach to the problem and go back to the lab. And so we have an uninterrupted succession of seminal investigations, each addressing a different topic, each skimming the cream of the topic before moving on elsewhere. It all adds up to an impressive number of original discoveries. More important, in my opinion, is the view that has inspired them, the first coherent picture of how living machines actually operate. I cannot conclude this brief survey of Lipmann's scientific contributions without mentioning another founding father of modern biochemistry, Carl Cori, who was born three years before Lipmann and died 2 years ago. The discovery of glycogen phosphorolysis by Carl and Gerty Cori in 1936 was a key step in the development of bioenergetics. The reversal of this reaction, leading in 1940 to the first in vitro synthesis of a macromolecule - the so-called blue glycogen - remains a landmark in our understanding of biosynthetic mechanisms, even though it turned out to be biologically irrelevant. It has inspired the work of a whole generation of biochemists, including the discovery, in 1948, by Arthur Kornberg of what was then

40

Christian de Duve

known as DPN pyrophosphorylase, the paradigm of all nucleotidyl transferases, including DNA and RNA polymerases. Fritz Lipmann projected a misleadingly diffident image to the world. "I did not have", he confesses in his autobiography, "and still lack, the gift for making an impression." 3 Elsewhere, however, he tells us that he "unsuitably presented a certain self-reliance that often impressed others as arrogant." 3 He was not a very good lecturer, even though his topics were always fascinating. More often than not, he would look at his audience with candid eyes, a gentle smile on his face, and slowly his voice, never very strong, would trail away into inaudibility while he pursued some inner track of his own. It would surge up again from time to time, but the trend was not easy to follow without access to the inner dream. Such traits do not go down very well with academic circles and, despite a rising scientific reputation, Lipmann had some trouble finding a job after he decided not to follow Dean Burk to the National Institutes of Health in 1941. Eventually he landed a position in the Department of Surgery at the Massachusetts General Hospital. He spent 16 fruitful years in this somewhat unlikely environment. This is where, among others, he discovered coenzyme A, for which he received the Nobel Prize in medicine in 1953, sharing it with his old friend of Berlin years, Hans Krebs. In 1957, Det Bronk invited him to Rockefeller. He accepted with joy. He had spent a year at Rockefeller in 1931-1932 and, to him, returning to the Institute 25 years later was like coming home. He loved our "very special place" with its tradition of excellence, scholarship, and freedom - a "20th century monastery", as Bronk liked to put it. Fritz Lipmann remained with us almost 30 years, during which he continued to work in the laboratory, to inspire others, and to deliver his unique brand of original results, up to the very end of his long and richly productive career. Although utterly committed to science, Lipmann had many other interests. In his youth, he had been very close to his brother Heinz, his senior by 2 years, and a very different personality, who wrote poems and made a career as a theatrical producer. Through his brother he was introduced to the world of performers and of artists, and shared for a while their exciting night life. He was a dashing young man in those days, we are told. His suit - he had only one - was always very well cut, and he sported Borsalino hats, losing them periodically in the subway. He loved life and beauty as he loved science, and he derived from his work the authentic joy of the artist, with no concern for either fame or money. Love of beauty created a deep bond with his wife Freda, an artist in her own right, whom he had met in Berlin in 1929 and had missed greatly after Meyerhof's laboratory moved to Heidelberg. It was to be near her, much more than because he wanted to learn tissue culture, that he left Meyerhof for the laboratory of Albert Fischer in Berlin. They were married in 1931 and for 55 years were a devoted and endearing couple, succeeding wonderfully in combining mutual love with respect of each other's independence. Gentle, whimsical, unpretentious, and tolerant Lipmann could at times be impressively firm and scathingly scornful, when confronted with moral shabbiness, vanity, or duplicity. Social injustice and discrimination made him very angry. He had a great affection for young people and is remembered fondly by all those who worked

Fritz Lipmann: in memoriam

41

with him and now maintain his tradition all over the world. He had a special regard for women in science and felt strongly that they should be given better opportunities. I will always remember my first meeting with him. It was in the beginning of 1948, on my way back from St. Louis where I had spent 6 exciting months working with Earl Sutherland in Cori's laboratory. I passed through Boston for the sole purpose of paying a visit to Lipmann, as I did to Claude in New York a couple of days later. Their work was quite unrelated to what I was doing at that time, but I just wanted to meet them. They were my heroes. I am sure Lipmann had no idea who I was and no interest in insulin or glucagon, my obsessions in those days. Yet he received me with great kindness and even invited my wife and myself to dinner afterwards, in a Chinese restaurant, as I remember, where we were joined by Freda Lipmann. We spoke of Berlin, the Berlin of before the horror, the Berlin of Berthold Brecht, of Kurt Weil, of Fritz's brother Heinz, and of their friend, the painter Friedel Sebba - a Berlin that I had not known, but of which I knew through an uncle who lived there and had shown me in 1939 how the Nazis had butchered the city. It was an unforgettable evening. I never worked with Fritz Lipmann and I was not even very close to him. I always stood a little in awe of him, in spite of his great friendliness and undisguised friendship. I was not his pupil, but I feel myself truly his disciple. I am deeply honored and touched to have been given this opportunity to pay homage to this wonderful man and to express, on behalf of all his friends and admirers, our deep sense of sorrow at his passing and our feelings of heartfelt sympathy to Freda Lipmann, their son Stephen, and their family.

Lipmann's Remarkably Fulfilled Life as a Researcher Koscak

Maruyama

Introduction I first met Dr. Fritz Lipmann at the Rockefeller University in the autumn of 1959, when I visited Professor Setsuro Ebashi who had just discovered Ca uptake of sarcoplasmic reticulum. At that time, I was working with muscle contractile proteins in Dr. John Gergely's laboratory at the Massachusetts General Hospital as a Helen Hay Whitney Foundation fellow (François Chapeville and the late Günster von Ehrenstein were also the Whitney fellows in Lipmann's laboratory). Although I talked with Dr. Lipmann only for a minute or so, I was most deeply impressed with his overwhelming personality: here was a unique human being. Ever since then I have had much interest in his life and work. I published his biography in Japanese in 1972 (1), and had a chance to interview him when he came to Japan last in 1984 (2). More recently, I spent a few days in Mrs. Freda Hall Lipmann's wonderful country home at Rhinebeck, NY, and enjoyed hearing her reminisce about her husband.

With the Age of Biochemistry Modern biochemistry began when Eduard Buchner (1860-1917) first solubilized zymase from yeast in 1897. Fritz Lipmann was born two years later on June 12,1899 in Königsberg, Germany (now Kaliningrad, USSR). In 1902 Franz Hofmeister (1850-1922) coined the term biochemistry when he founded Zeitschrift für die gesamte Biochemie (3). In the USA, the Journal of Biological Chemistry was first issued by Christian Herter (1865-1910) in 1905. During the next year, Biochemische Zeitschrift and Biochemical Journal were founded by Carl Neuberg (1877-1956) and Benjamin Moore (1867-1922), respectively. In the 1920's, when Lipmann became interested in biochemistry, his native country was the world center of this science, as clearly shown by the greater number of pages in the Biochem. Z. than in either the J. Biol. Chem. or the Biochem. J. (4). This was largely due to the activities of such eminent biochemists as Carl Neuberg, Gustav Embden (1877-1933), Otto Warburg (1883-1970), and Otto Meyerhof (1884-1951). It was fortunate for Lipmann that he belonged to the Kaiser Wilhelm Institute for Biology, where the two Ottos worked. Around 1933, the rate of biochemical publications suddenly declined due to the establishment of Hitler's Third Reich and the start of the expulsion of Jewish scientists from Germany. Lipmann, then staying in Denmark, had to flee to the

44

Koscak Maruyama

USA in 1939 as did many German Jewish biochemists including Meyerhof and Neuberg. After the Second World War, biochemistry progressed rapidly worldwide, especially in the U S A where Jewish refugees played a vital role. Figure 1 shows the changes in annually printed pages of the J. Biol. Chem. from its beginning to date. Unlike European biochemical journals, publication of this organ was not affected by the war time situation (Figure 1; see Ref. 4). It is to be emphasized that Lipmann worked during the rising and the golden ages of biochemistry. In the 1970's molecular biology abruptly developed while biochemistry had reached its plateau. This has been described by Slater for Biochimica Biophysica Acta (4). The J. Biol. Chem. has included papers on molecular biology, and the size of this publication is continuing to increase in the 1980's (Figure 1).

15

m

O "O Φ

10 ΟΟ ο

-Ω

=3 α.

(0 ω Ο) (0 α

D

1910

o 0 0 coOo

η ° ° ° CO ο 0 οο° οοοΡ „Tto ο ο οο ο ο Ο

ο°οοθ°0°

I

1

1

I

I

I

L-

20

30

40

50

60

70

1980

Figure 1 Printed pages of the Journal of Biological Chemistry by years. Normalization made for pages earlier than 1958 when page size was changed.

Although Lipmann contributed much to the new molecular biology, he remained fundamentally a biochemist. His death in 1986 and that of Hans Krebs (1900-1981), and Albert Szent-Györgyi (1893-1986) symbolize the end of the dynamic biochemistry initiated by Frederick Gowland Hopkins (1861-1947).

Lipmann's Remarkably Fulfilled Life as a Researcher

45

Mentors, Rivals, and Companion The late David Novelli (1918-1982), codiscoverer of CoA, wrote in his autobiographical essay (5): "Start of art (science) was accomplished by personal teaching from Lipmann in the laboratory and daily discussions with Lipmann and Kaplan, sitting at the circular lunch table in the doctor's cafeteria at "Mass. General". Another codiscoverer of CoA, the late Nathan Kaplan (1917-1986) generalized on "how to be a good scientist" (6): "One of the most important aspects, in my opinion, is the role of research teachers. These individuals play an important part in developing the form and style which a young investigator may take. A second factor, in my view, is to work in an environment where there are stimulating colleagues, postdoctoral fellows and students". Certainly these had been the case with their respected mentor, Lipmann indeed. While a senior medical student in Berlin, Lipmann took a three-month laboratory course in biochemistry and physical chemistry in Peter Rona's laboratory at the Pathological Institute, Charité (Berlin University Medical School). Rona (1875-1945), a Hungarian biochemist, worked with Leonard Michaelis (1885-1949) and jointly published some 40 papers from 1906 to 1920 (7). Rona was a superb teacher and trained a number of physicians in basic life sciences. "FRS" (former Rona students) included three Nobel laureates: E.B. Chain, Krebs and Lipmann; others were R. Schoenheimer, H.H. Weber, K. Meyer, G. Schmidt, D. Nachmansohn, H. Blaschko, H. Kosterliz, R. Ammon and K. Iwasaki. Lipmann studied physical chemistry in relation to biology and medicine and methods of biochemical analyses, and remained in that laboratory for a time to conduct research. He studied electrophoretic behavior of iron colloids and was mainly concerned with the reversal of the positive charge to negative in the presence of citrate (8,9). His first paper on this subject was printed in Biochem. Z. 147, 163 (1924) and this became a part of his MD thesis. While in Rona's laboratory, Lipmann became acquainted with a number of biochemistry-oriented scientists such as Hans H. Weber (1896 1974) (10). He had then the fortunate opportunity to work on a fellowship for a half year in Laquer's Pharmacological laboratory at Amsterdam in 1923, thus escaping from the tremendous inflation prevailing in Germany (9). Ernst Laqueur (1880-1947), one of the discoverers of testosterone and corticosterone, was associated with the University of Königsberg from 1902 to 1920 where Lipmann finished medical course in 1920 (11). It was there that he first conducted biochemical research on the changes in blood sugar levels of rabbits following injections of glucose, glycogen and starch. With J. Planelles he published two articles on this work in 1924 and 25 (Biochem. Z., 151, 98; 163, 406). The work at Amsterdam convinced Lipmann to turn away from medicine in favor of biochemistry. However, to do this he felt he needed a stronger background in chemistry. He therefore, returned to the city of his birth and studied chemistry for three years unter Professor Hans Meerwein (1879-1965), an excellent organic chemist (12). This proved a decision which greatly affected Lipmann's future as a

46

Koscak Maruyama

biochemist. He would otherwise, not have been able to propose his high energy phosphate bond concept. After three years, Lipmann wanted to find a suitable place to complete his Ph. D. in chemistry. Two biochemical laboratories at Berlin were particularly attractive: those of Otto Meyerhof and Carl Neuberg. The former was more biologically oriented and met his desire more closely. He was accepted as an unpaid research fellow in this laboratory in 1927, because Meyerhof usually selected his collaborators through the recommendation of Peter Rona, e. g. Nachmansohn, Iwasaki, Blaschko, and others (13). Meyerhof must have warned Lipmann that he would never be paid a salary in his laboratory, because Meyerhof had only one salaried position (for Lohmann) at that time (14). Lipmann's father, Leopold, a not-so-successful Jewish lawyer agreed to continue supporting his son even though he already had an MD. Meyerhof was 43 years old in 1927 when the 28-year Lipmann joined his laboratory. He had been awarded the Nobel prize in 1922 for his lactic acid theory of muscular contraction. Lipmann's first task was to inject rabbit with a yeast autolysate containing hexokinase to see if this lowered the blood sugar level; this did not achieve the desired result. His subsequent attempts to relate breakdown of newly discovered creatine phosphate to muscle contraction were also unsuccessful. Sharing a small laboratory with Ken Iwasaki (13), he then chose a less adventurous and proportional-to-effort type of experiment: the effect of fluoride on enzyme reactions. Three papers resulted from this work, earning him a Ph. D. under the sponsorhip of Carl Neuberg in 1929. As Lipmann himself recalled (8), Meyerhof was not a person he found easy to know and the two did not became very friendly. This was in sharp contrast to his relationship with David Nachmansohn (15). According to Iwasaki (16), Meyerhof understood immediately what was told him about a scientific matter and could not tolerate a lengthy explanation or a slow response to his questions, repeating "Ich verstehe schon". A young colleague like Iwasaki, however, could understand now deeply penetrating Lipmann's way of thinking was (16). In addition, Lipmann rather shunned the unwieldy muscle energetics of which Meyerhof was so fond (8), though he later admitted that Meyerhof's energetics approach profoundly influenced in his life work (8). Meyerhof regarded Ochoa the most brilliant of his active collaborators, and they stimulated each other as good rivals (15). Under the same roof at the Kaiser Wilhelm Institute of Biology. Berlin-Dahlem, there was another giant of biochemistry, Otto Warburg. Hans Krebs was his unpromising assistant (17). Both Meyerhof and Warburg were broadly important mentors to Lipmann, and their younger coworkers were later to be leading scientists. Although Lipmann moved to Heidelberg at the end of 1929 when Meyerhof became director of the Kaiser Wilhelm Institute for Physiology in that city, he was destined to return to Berlin because his wife-to-be was there (18). Elfried (Freda) Hall was born in Defiance, Ohio in September 19, 1906 as the third daughter of Hermann and Gertrud (Kuntz) Hall. Her father was an administrative businessman from Germany. When she was a child, the Halls returned to Europe and lived in West-Preussen. Then the family moved to Berlin. Freda was interested

Lipmann's Remarkably Fulfilled Life as a Researcher

Figure 2

47

Mrs. Freda Hall Lipmann (1934).

in drawing and went to art school. Later she worked as a fashion illustrator for newspapers published by Ullstein. In early 1929, for "Genesungsfeier" to Ken Iwasaki who had just recovered from appendicitis, Lipmann, Nachmansohn, and Rothschild went to the Sozialistenball, and it was here that Lipmann happened to meet Freda Hall (Figure 2) (13,18,19). Lipmann got a job as an assistant in Albert Fischer's tissue culture laboratory of the Kaiser Wilhelm Institute for Biology, Berlin in 1930, replacing Hans Laser, a former classmate at Königsberg, who was leaving to join Meyerhof's school (8). This was a fortuitous choice for Lipmann, because he was given a chance to work in Phoebus Levene's laboratory at the Rockefeller Institute, in New York, while a new institute for Fischer (1891-1956) was completed in Copenhagen. Fischer himself had originally intended to go to the Rockefeller Institute, but they did not offer acceptable conditions (18). In September, 1931, the Lipmanns, who had been married on June 23, arrived at New York.

48

Figure 3

Koscak Maruyama

Fritz Lipmann, holding a four-leaf clover in hand at Yacatianing in Switzerland (1938).

Levene (1869-1940), who first elucidated the structure of D N A and RNA, was a very productive biochemist, publishing more than seven hundred papers in his life time (20). Lipmann isolated serine phosphate from an egg protein, utilizing a differential hydrolysis method that he had learned from Karl Lohmann (21). Did Lipmann foresee then that he would follow the same path as Levene in that Institute in future? Together with Fischer, Lipmann spent seven happy and fruitful years from 1932 to 39 in Copenhagen (Figure 3) (6,18). Fischer was very generous in allowing the younger man to pursue his own resarch direction. In the Carlsberg Laboratories, Lipmann became a closed friend of K. Linderstrom-Lang (1896-1959), E. Lundsgaard (1883-1930), and H . M . Kalckar (*1908). These remarkable scientists greatly stimulated Lipmann's research interest. His essay on Lang was a splendid one and full of his warm friendship for this eminent protein chemist (22). While there he also met other leading biochemists such as Paul Zamecnik and Rollin Hotchkiss. Under the great threat of Nazis, the Lipmanns had to escape Denmark for the United States in 1939. This was motivated by Mrs. Lipmann who forsaw disaster for the Jews under Nazi power (19). Lipmann had asked Dean Burk (*1904), a companion in Meyerhof's laboratory, to accept him as an assistant in his laboratory at Vincent du Vigneaud's Department of the Cornell Medical School. LinderstremLang wrote a strong recommendation letter to du Vigneaud, who was then mainly concerned with penicillin production. Du Vigneaud (1901-1978) was later to elucidate the structure of oxytocin and vasopressin and to receive the 1955 Nobel prize for chemistry (23).

Lipmann's Remarkably Fulfilled Life as a Researcher

49

At Cornell, Lipmann met Hotchkiss (*1911) whom he had known at Copenhagen and who had isolated two bacterial peptide antibiotics, tyrocidin and gramicidin. With D-Amino acid oxidase that Lipmann prepared, they were able to show the presence of D-amino acids in natural peptides for the first time. This experience later led Lipmann to non-ribosomal synthesis of peptides. In the summer of 1940 he began to write a review of ~ Ρ in a rented cottage hear Lake Iroquois, Vermont (18). A crisis faced Lipmann in 1941 when his two-year term with Burk and du Vigneaud ended. Where was he to find a job? Since he was not well known, he was not accepted for any really good position (8). He turned down a technician's job offered by Siegfried Thannhauser, Krebs' boss at Freiburg who had settled at Tufts University Medical School in Boston (19). Ernst Oppenheimer (1888-1962), then Chief Pharmacologist of Ciba Pharmaceutical Products, Inc. (24), who knew of Lipmann from German contacts mentioned him to a close friend of his, Oliver Cope, the surgeon of Massachusetts General Hospital who was looking for an endocrinologist in his laboratory (25,26). Eventually, Lipmann became a Ciba fellow in the Department of Surgery, Massachusetts General Hospital. It was Lipmann's pleasure that Zamecnik just joined Huntington Biochemical Laboratory, MGH, as an instructor. Lipmann had met him in Copenhagen. This odd situation of "a biochemist among surgeons" resulted in the discovery of CoA.

Way of Research As Lipmann himself admitted, what he published during his apprenticeship in Meyerhof's laboratory was not particularly significant (8), however, he learned there very important aspects of dynamic biochemistry, both theoretical and practical. He also became familiar with the use of bacteria for experimental materials when mammalian tissues did not work. This led him to his 1939 discovery of acetyl phosphate in Copenhagen. In answer to my question in 1984 "What was your most impressive work?", he replied that at Copenhagen (2). After unsuccessful trials with pigeon liver preparations for pyruvate oxidation, he found that acetone powder suspension of lactic acid bacteria actively reduced methylene blue in the presence of pyruvate. In order to employ Warburg's manometry, he had to replace phosphate with a bicarbonate buffer. To his surprise, the bacterial system became inactive in bicarbonate, although methylene blue was rapidly reduced in phosphate. He immediately recognized that acetyl phosphate was the intermediate of pyruvate oxidation. Recalling those excitement years at the age of 85, he said it still was motivation for his then current research (2). While at Heidelberg, he witnessed the collapse of Meyerhof's lactic acid theory as the energy source of muscle contraction. At the end of 1929 Einar Lundsgaard visited the laboratory to demonstrate "alactacid" contraction accompanied by creatine phosphate breakdown. Looking at this drastic situation, and taking high heat generation due to hydrolysis of creatine phosphate or newly found ATP into consideration, Lipmann slowly but steadily developed the concept of high energy

50

Koscak Maruyama

phosphate bonds in C r ~ P , ATP, acetyl ~ P , etc. His famous review in 1941 introduced the concept of group transfer: phosphate, acetate, amino, etc. This concept was his leitmotif in later work: CoA, sulfate transfer, protein synthesis and protein kinase. It is, therefore, understandable that he would rather have been awarded the Nobel prize for the proposal of the group transfer concept than for his discovery of CoA (19). It is to be noted that his 1941 review attracted a number of biochemists and H.A. Barker, Kaplan, and Novelli joined Lipmann's laboratory at M G H in this order. The latter two were the strongest contributors to the isolation and characterization of CoA and described the CoA story in detail (5,6,27). Lipmann had discovered it as a heat-stable cofactor for acetylation of sulfonamides in pigeon liver preparations (1945). During this CoA study, there was one great misfortune about which Lipmann felt much chagrined: he was not able to identify "active acetate" in mammalian tissues simply because he concentrated too much on acetyl phosphate (8,9). Fedor Lynen (1911-1979) discovered acetyl CoA in 1951 (28,29). When Novelli revealed the role of acetyl phosphate to be active acetate in a bacterial extract, Lipmann beamed with joy and he danced a little jig in the laboratory at M G H (5). One summer day in Woods Hole, Lipmann heard George H . A . Clowes (1877-1958), the research director of Eli Lilly and Company, who had been the prime mover in the commercial production of insulin and helped its first clinical use in Toronto (30). Clowes had observed that dinitrophenol (DNP) retarded embryonic development of sea urchin eggs in spite of its augmentation of oxygen uptake (31). It occurred to Lipmann that D N P might shut off oxidation and phosphorylation coupling in mitochondria. One member of Lipmann's laboratory was W. F. Loomis who had worked on mitochondrial respiration at Wisconsin under the supervision of D . E . Green (1910-1983). Loomis was able to show the decrease in Ρ : O ratio from 2.2 to 0.2 in the presence of DNP, using rabbit kidney mitochondria preparations. His two pages note in 1948 (J. Biol. Chem. 173, 807) marks a milestone in the history of oxidative phosphorylation research, mainly due to the invention of the term "uncoupling". However, Lipmann did not go further in this attractive field except for similar experiments with thyroxine. He must have instinctively realized that the mechanism of oxidative phosphorylation was beyond his scope of research. Lipmann's finding of the role of carbamyl phosphate in Krebs' urea cycle directly derived from his group transfer concept (8). He had thought of possible carbamyl transfer and had synthesized citrullin which he had long kept in a drawer. When he learned of the phosphorolysis of citrullin with A D P and P¡ at a Gordon Research Conference, he immediately suspected carbamylphosphate as an intermediate compound in that reaction. He suggested to Mary Ellen Jones that she use a bacterial extract rather than liver preparations containing whole urea cycle enzymes. Jones obtained an unstable phosphate compound by incubating a bacterial extract with ammonium carbonate and the phosphoenolpyruvate system. Addition of Ornithin decreased the unstable compound but produced citrullin. Leonard Spector, then working in Zamecnik's laboratory one floor below Lipmann's, chemically synthesized carbamylphosphate in an elegant manner. Thus, carbamylphosphate was proven to be the intermediate compound (J. Am. Chem. Soc. 77, 819 (1955)).

Lipmann's Remarkably Fulfilled Life as a Researcher

51

Since the 1941 review, Lipman had the idea of amino acid activation by way of phosphorylation of the carboxyl group as the intermediate step of protein synthesis. Mahlon Hoagland who had worked on the acetate-CoA-ATP system in Lipmann's laboratory, in Zamecnik's laboratory showed that pyrophosphorolysis of ATP was involved in the amino acid activation. In 1954 Lipmann proposed that certain amino acid specific sites on a temperate were activated by transfer of PP from ATP and then PP was exchanged with the carboxyl group of the amino acid specified by the temperate. Zamecnik's group discovered the participation of small RNA (now tRNA) in the amino acid activation. The intermediate was identified as enzymebound aminoacyl adenylate, the same as in acetyl CoA formation. This was first discovered by Paul Berg (*1926) in 1955 and had also been missed by Lipmann. On the other hand, Lipmann and his associates at the Rockefeller University contributed much to the mechanism of the elongation step of protein synthesis (5,8). In 1960, Harold Bates and Lipmann reported that -/-glutamyl cysteinyl RNA was the intermediary in synthesis of the tripeptide glutathione (J. Biol. Chem. 235, PC22). The glutathione synthesis had been attributed to the action of its synthesizing enzyme without participation of RNA. Unfortunately, it turned out that the requirement of RNA was an artifact forged by Bates who had been responsible for a similar fabrication at Yale University. The above paper is recorded as "not to be distributed" on Lipmann's publication list available from the Rockefeller University Archives. Lane and Lipmann corrected the fraudulent paper in 1961 (J. Biol. Chem. 236, PC80). This sort of forgery in science is caused, not seldom, by an ambitious but mentally aberrated associate in a world famous laboratory, as also happened to Zamecnik later (32). Lipmann endured the problems caused by Bates, and challenged RNA-independent synthetic pathway of polypeptides stimulated by evolutional point of view (8). With Wieland Gevers and Horst Kleinkauf, Lipmann resumed work on the biosynthesis of gramicidin S on which he had worked before. Each amino acid was activated to form aminoacyl adenylic acid without tRNA and was shown to be linked to a thiol on an enzyme. Peptide synthesis is mediated by a phosphopantheine arm present on the multifunctional enzyme components and stepwise addition of amino acids occurs (9). This non-ribosomal peptide synthesis pathway has also been established by work independently carried out in Kiyoshi Kurahashi's laboratory in Osaka (33). Even after the age of 80, Lipmann continued to publish original articles with high standard from his "Emeritus Professor's" laboratory at Rockefeller. He became interested in Rous sarcoma or transformed cells from the viewpoint of phosphorylation and sulfation of tyrosine residues in proteins. On the occasion of the Lipmann Symposium held near Mt. Fuji in 1984 to celebrate his 85th birthday, he presented his new results on the decreased sulfation of proteins in transformed cells (34).

Productivity and Quality Scientist's contribution to science is evaluated by the extent to which his achievements promote his own area of research. If introductions of entirely new concepts or

52

Koscak Maruyama

unique techniques significantly and broadly stimulate the development of the field, the contribution becomes much more distinguished. Recently, a researcher's work has been appreciated both by productivity, i.e. the number of original articles published per year, and by the citation index, i.e. the number of citations made of a particular paper.

Table 1

Productivity Total publication Total

Warburg Meyerhof Lipmann Krebs Ochoa

530/1905-70 406/1910-53 516/1924-85 356/1923-81 538/1928-86

Nobelist* "Average" *

136 62

8.2 9.4 8.5 6.1 9.2

per year Original 4.0 4.7 4.4 3.5 3.9 3.2 1.5

* After Zuckerman (35).

First, we shall examine Lipmann's prolificacy during his life. Table 1 summarizes his productivity as compared with closely related scientists. His total publications between 1924 and 1985 number about 516. Annually, he was involved in an average of 8.5 published works, 4.4 of which were published in refereed journals. This is not including articles printed in book form. As shown in Figure 4, much more were published after receiving the Nobel prize (1953) than before. There were two peak periods, in 1953 and 1969. Lipmann's productivity was almost at the same level as his mentor Meyerhof and also Krebs' mentor Warburg. Professor Severo Ochoa (*1905), his companion in Meyerhof's laboratory has been producing scientific publications at a similar rate. Hans Krebs, colaureate of the 1953 Nobel prize, was somewhat less productive so far as scientific publications were concerned. All of these men, however, were much more productive than average Nobel laureates (35). Any scientist, of course, owes much credit to his or her collaborators except for the young apprenticeship days. Lipmann worked alone until he had his own laboratory at the Massachusetts General Hospital (MGH), where Constance Tuttle joined him as secretary-technician. From 1950 to 1972 there were more than ten collaborators in Lipmann's laboratory at MGH or at Rockefeller University (Figure 5). The numbers of associates in Figure 5 were estimated from those coauthoring his publications, and therefore may differ from the actual numbers; in 1964, there were as many as 29 collaborators (8). Among over 200 Lipmann alumni, Daniel Nathans received the Nobel prize in 1978. Most of the others have become leaders in various fields of biochemistry and molecular biology and have shown their esteem for their mentor by organizing four Lipmann Symposia during the past twenty years including this Memorial.

53

Lipmann's Remarkably Fulfilled Life as a Researcher

1920 Figure 4

Figure 5 Korean).

30

40

50

60

70

1980

Lipmann's annual scientific publication.

Collaborators in Lipmann's laboratory. Dotted: orientals (Chinese, Japanese and

In the later years Lipmann's coworkers numbered from three to five. There were always several oriental fellows, Chinese, Japanese, and Korean who supported Lipmann's work mainly in the last stages. The oriental inclination to respect the wisdom of age showed itself clearly here.

Koscak Maruyama

54

Several examples of citations of Lipmann's work are indicated in Table 2, among which his 1941 famous review article on ~ P is noteworthy. Between 1961 and 1986, 45 years after its publication, 165 citations have been noted, and the total number of references to this work were probably several times greater.

Table 2

Most cited papers of Lipmann (1961-1986) No. cited

Article (1941) (1944) (1945) (1945) (1958) (1969) (1971) (1973)

~P Acetyl Ρ Acetyl Ρ Sulfonamide Sulfate activation Protein synthesis Peptide synthesis Non-ribosomal peptide synthesis

Adv. Enz. 1, 99 JBC 153, 571 JBC 158, 505 JBC 160, 173 Science 128, 575 Science 164, 1024 Science 173, 875 Accts. Chem. Res. 6, 361

165 786 76 63 86 106 120 68

The 1944 paper on acetyl phosphate was selected as a citation classic (36). Other publications on acetylation of sulfonamide (1945), sulfate activation (1958), protein synthesis (1969,1971), and non-ribosomal peptide synthesis (1973) have been cited more than 50 times as of 1986. The papers listed in Table 2 are a portion of those of which Lipmann was first author. His co-authored work also was well recognized: a paper on the active Ca uptake of sarcoplasmic reticulum by Ebashi and Lipmann (J. Cell Biol. 14, 389 (1962)), for example, has been cited 305 times. Apart from his most distinctive contribution to biochemistry, discovery of CoA for which he was awarded the Nobel prize, Lipmann's concepts on high energy phosphate bond " ~ P " and group transfer were most important to this field. These may well be in the category of paradigms defined by Thomas Kuhn (37). Since Lipmann's proposals, all biochemistry textbooks have been revised to some degree. In this respect, Lipmann is compared with Frederick Gowland Hopkins, leader of the Cambridge school, who recognized the concept of dynamic biochemistry.

Comment on Lipmann In his autobiography, "Wanderings of a Biochemist" (8), Lipmann described his scientific journey in which he followed his own instinct without knowing exactly where it could lead. His response was not quick, but he thought deeply, with frequent interruptions, until he found a significant subject, no matter how much time it required. On the other hand, Lipmann depended on personal relationships when trying to find out a place to work as described in Mentors, Rivals and Companion. He was the type of person who needed others to understand himself. However, once settled, he went on his own way, and this sometimes resulted in inharmony with the people around him (19).

Lipmann's Remarkably Fulfilled Life as a Researcher

Figure 6

55

Lipmann's country home at Rhinebeck, Ν. Y. Illustrated by Mrs. Freda Hall Lipmann.

56

Figure 8

Koscak Maruyama

Lipmann's deck chair in the garden of his country home (1987).

Lipmann's Remarkably Fulfilled Life as a Researcher

Figure 9

57

Mrs. Lipmann and Pogo at rest in the walk way of their country home (1987).

Lipmann used to leave New York on Friday afternoon for Rhinebeck, N Y to spend weekends at the country home bought in 1959 with this Nobel prize money (Figures 6 and 7). He liked to meditate in a deck chair in the wild garden filled with maple, oak and black walnut, and tulip trees (Figure 8). He liked to stroll in the 20 acres of woods together with his Australian terrier, Pogo who was named after the hero of a mid 1970's political cartoon strip (Figure 9), which Lipmann and his son Stephen enjoyed. Lipmann avoided involvement in political and social activities, only twice signing Nobel laureate public appeal letters: for prohibition of the hydrogen bomb and endorsing freedom for the Polish Worker's Union (19). He appreciated nature's beauty as expressed in flowers, trees and birds, and loved to chat with Mrs. Lipmann in detail on these things. He spent many hours reading books, and during his last days, for example, read with great interest Martin Kamen's "Radiant Science, Dark Politics" (19). However, writing took him a great deal of time and was only accomplished with much pain and several rewritings. He had a talent for writing as revealed in his autobiography (8,9) and wrote quite a few essays, such as those on Lang (22) and Embden (38). Even his scientific essays such as on ATP (39) are easy to follow and very informative. This ability he shared with his brother Heinz (1887-1931) (Figure 10), a dramaturge with the famous intendant of the Staatstheater, Leopold Jessner who was also a Königsberg-born Jewish (1878-1945) (40). His manner of speaking, on the contrary, was not as fluent. One often had to concentrate on his words in order to follow his thought. Lipmann was often an absent-minded professor. At a scientific meeting at a large auditorium, Lipmann asked a question, and the chairman said to him to come to a microphone as he could

58

Koscak Maruyama

Figure 10

Heinz and Fritz Lipmann in Könisberg.

not be heard. After he did so, he tried to seek out his original seat in vain despite he could sit anywhere. He was in his stockings and was looking for his shoes (25). When I interviewed him at the Mt. Fuji View Hotel in 1984, he said that as early as 1930 he had thought that ATP must be a direct energy source for muscle contraction. The discovery of ATP was announced independently by Fiske and Subbarrow of Harvard Medical School and Karl Lohmann in 1929 (30). To my question about why he had not mentioned this in any of his writings at that time, he hesitatingly replied "You know, Meyerhof did not like that idea". I felt he always tried to be honest with himself. On July 17th, 1986, just five days after his 87th birthday Lipmann suffered from a mild stroke at his country home which affected his walking. The next day he wanted to call his laboratory people to learn how the research was developing. He took lunch outdoors and talked pleasantly with his collaborators, Spector, Suiko, Liu,

Lipmann's Remarkably Fulfilled Life as a Researcher

59

and Decker for several hours after lunch. In the evening he had no appetite and was unable to climb the stairs to his bedroom, remaining on a bed in his study. "I can't function any more", were his last words before slipping into a coma (19). Since his Boston days Lipmann had spoken English even at home. He was hospitalized in St. Francis Hospital in Poughkeepsie, NY and died on July 24. He did not believe in any religion, and accepted death just as such. His ashes were scattered around the path in the woods he liked so much on the grounds of his home. There is no tomb for him - he has simply returned to Nature. Lipmann had asked his wife to arrange a concert of music he loved in place of a funeral, and the Lipmann Memorial Concert was held at the Caspar Hall of Rockefeller University on December 12,1986. In accordance with Stephen's choice, Beethoven's Quartet in Ε-flat Major, Opus 127, and Mozart's Quintet for Strings in G Minor, K. 516 were played by the Muir String Quartet for the roughly 200 people who gathered. It is of much interest that Lipmann had great respect for Warburg during his BerlinDahlem days (19), although Warburg later said he did not remember him at that time (41). Later, however, when Chapeville told Warburg that he had worked in Lipmann's laboratory, Warburg said "Then, you are my grandson, because Lipmann's mentor Meyerhof is my academic son" (8,42). Lipmann must have been very honored when Warburg called him "one of the few who may be able to solve the great problem of differentiation of life" in the Fritz Lipmann Dedicatory Volume (43). Paul Ehrlich (1854-1915) once said that there are four G's - Geduld, Geschick, Glück, and Geld - needed to be a researcher. Two more should be added further: Geist and Gesundheit. Fritz Lipmann possessed all six of these qualities. In his memorable essay on Gustav Embden (38), Lipmann referred to him as a romantic scientist and to Meyerhof as a classic scientist. It seems to me that Lipmann's work has already stood the test of time, and therefore is classic, but his way of research was certainly romantic. A rather slow starter; Lipmann accomplished his first eminent work on acetyl phosphate when he was 40 years old. But during the following 45 years, his scientific activity never flagged. This is rather surprising because of the general tendency for most mature scientists to become either administrators or to turn to more speculative research. In addition to the Nobel prize, Lipmann was awarded the National Science Medal in 1966, and nominated as a foreign member of the Royal Society, London in 1962. Yet, he did not want to be an institutional director, but remained a professor (he did not even like to be called professor emeritus, a title he was given in 1970.). What was the Lipmann's secret? He remained loyal to his simple, naive and persistent desire to clarify what interested him: the molecular mechanism of life. Writing about Hans Krebs with whom he shared the Nobel prize, Lipmann commented on his "remarkably fulfilled life" (44). I should like to say the same of Dr. Fritz Lipmann's 87 years. Finally, this great scientist's life cannot be spoken of without reference to his great debt to his wife of 55 years. According to the late Ken Iwasaki, more than half of what Fritz did was due to Freda's support (13). A woman of personal strength and

60

Koscak Maruyama

fortitude, artist herself, she provided her husband with an ever-lasting harmonious and pleasant atmosphere in which to live and develope both personally and professionally.

Acknowledgement The author expresses his m o s t sincere gratitude to Mrs. Freda Hall Lipmann for her kindest responses to my inquires about her husband and also for her providing m e with invaluable photographs. Acknowledgement is also due to the following for their valuable information; Professor S. Ebashi, Dr. M . H o a g l a n d , Dr. K. Iwasaki, Professor H. M . Kalckar, Professor Y. Kajiro, Professor K. Kurahashi, Professor H. Kleinkauf, Professor S. Ochoa, and Dr. L. Spector.

References 1. Maruyama, K. 1972. Pursuit for Life (in Japanese), Chuokoronsha, Tokyo. 2. Maruyama, K. 1984. Gendaikagaku, 150, 56-57 (in Japanese). 3. Kohler, R.E. 1982. From Medical Chemistry to Biochemistry. Cambridge University Press, Cambridge. 4. Slater, E.C. 1986. Biochimica et Biophysica Acta: The Story of a Biochemical Journal. Elsevier, Amsterdam. 5. Novelli, G.D. 1980. In: Chemical Recognition in Biology (F. Chapeville, and A.L. Haenni, eds.). Springer, Berlin, pp. 415-430. 6. Kaplan, N.0.1986. In: Selected Topics in the History of Biochemistry. Pernonal Recollections. II. (G. Semenza, ed.). Elsevier, Amsterdam, pp. 255-296. 7. Ammon, R. 1960. Arzneimitt. 10, 321-327. 8. Lipmann, F. 1971. Wanderings of a Biochemist. Wiley, New York. 9. Lipmann, F. 1984. Ann. Rev. Biochem. 53, 1-33. 10. Krebs, H. and F. Lipmann. 1974. In: Energy, Biosynthesis and Regulation in Molecular Biology (D. Richter, ed.). de Gruyter, Berlin, pp. 7-27. 11. Riesser, O. 1949. Arch. exp. Pathol. Pharmacol. 206, 117-123. 12. Hesse, G. 1949. Ζ. angew. Chem. 61, 161-168. 13. Iwasaki, K. and N. Shimazono. 1976. Seikagaku 18, 59-71 (in Japanese). 14. Blaschko, H. 1980. Ann. Rev. Pharmacol. Toxicol. 20, 1-14. 15. Nachmansohn. D. 1979. German-Jewish Pioneers in Science, 1900-1933. Springer, Berlin. 16. Iwasaki, N. 1952. Kagaku no Ryoiki, 5, 52-53 (in Japanese). 17. Krebs, Η. 1981. Reminiscences and Reflections. Clarendon Press, Oxford. 18. Lipmann, F.H. 1988. In this volume. 19. Lipmann, F.H. 1987. Oral communication. 20. Van Slyke, D.D. and Jacobs, W. A. 1944. Biogr. Mem. Natl. Acad. Sci. 23, 75-126. 21. Lipmann, F. 1983. TIBS Sept., 334-335. 22. Lipmann, F. 1980. TIBS June, 111-112. 23. Plane, R. A. 1976. J. Chem. Educ. 53, 8-12. 24. Yonkman, F.F. 1962. Pharmacologist 4, 128-129. 25. Hoagland, M. 1987. Personal communication. 26. Kalckar, H.M. 1987. Personal communication. 27. Novelli, G.D. 1966. In: Current Aspects of Biochemical Energetics (N.O. Kaplan, and E.P. Kennedy, eds.). Academic Press, New York, pp. 183-197. 28. Lynen, F. 1969. Persp. Biol. Med. 12, 204-218. 29. Lipmann, F. 1980. Naturwiss. Rundschau 33, 221-223. 30. Bliss, M. 1982. Discovery of Insulin. Univ. Chicago Press, Chicago.

Lipmann's Remarkably Fulfilled Life as a Researcher

61

31. Clowes, G . H . A . and Krahl, M . E . 1936. J. Gen. Physiol. 30, 315-330. 32. Kohn, A. 1986. In: False Prophets. Blackwell, Oxford, pp. 9 3 - 9 7 . 33. Kurahashi, K., C. Nishio, K. Babasaki, J. K u d o h and T. Ikeuchi, 1985. In: Cellular Regulation and Malignant G r o w t h (S. Ebashi, ed.). Japan Sci. Soc. Press, Tokyo, pp. 177-186. 34. Lipmann, F. and Liu, H . C . 1985. ibid. pp. 393-400. 35. Zuckerman, H. 1977. Scientific Elite. Free Press, New York. 36. Barrett, J.T., ed. 1986. Contemporary Classics in Life Sciences, Vol.2. Molecules of Life, ISI Press, Philadelphia. 37. Kuhn, T. S. 1971. The Structure of Scientific Revolutions. 2nd ed., Univ. Chicago Press, Chicago. 38. Lipmann, F. 1975. Mol. Cell. Biochem. 6, 171-175. 39. Lipmann, F. 1979. TIBS Jan. 2 2 - 2 4 . 40. Rühle, G. 1980. Theater in unserer Zeit. Suhrkamp, F r a n k f u r t . 41. Kamen, M . D . 1985. In: Radiant Science, D a r k Politics, Univ. Calif. Press, Berkley, pp. 100-101. 42. Chapeville, F. 1984. Personal communication. 43. Warburg, O. 1966. In: Current Aspects of Biochemical Energetics ( N . O . Kaplan and E.P. Kennedy, eds.), Academic Press, New York, p. 108. 44. Lipmann, F. 1982. TIBS Sept. 337.

Obituary New York Times, July 25, 1986 Jencks, W.P. (1986) N a t u r e 323, 672 Zachau, H . G . (1986) Biol. Chem. Hoppe-Seyler, 367, 1183-1184 Mukai, J. (1986) Kagaku to Seibutsu 24, 8 1 4 - 8 2 1 Ebashi, S. (1987) Tampukushitsu-Kakusan-Koso, 32, 2 5 1 - 2 5 3 Roskoski, R. Jr. (1987) TIBS, 12, 136-138 de Duve, C. (1987) Faseb J. 1, 3 - 5 Ochoa, S. (1987) Eur. J. Cell. Biol, in press

Fritz Lipmann: June 12, 1899 - July 24, 1986 Hans G. Zachau

Fritz Lipmann died in Poughkeepsie, New York, after a short illness at the age of 87. He was one of the truely great biochemists of our century. We owe to him the understanding of the energetics of metabolism at the molecular level. He created the concept of the energy-rich bond which is one of the cornerstones of today's biochemistry. Lipmann's father was a lawyer in the city of Königsberg, then the capital of East Prussia. Königsberg is now Kaliningrad. Fritz Lipmann remembered the times before the advent of automobiles, when all traffic was by horse-drawn cars and the Kaiser was driven through the streets in a four horse-drawn carriage by a coachman with a plumed helmet. All vacations were spent at the baltic coast near Königsberg. After a classical gymnasium education Lipmann started to study medicine in 1917. He was soon drafted and had the grim experience of serving as a medic in a field hospital. After the war he continued his studies in Munich, Berlin, and Königsberg, but he never practiced medicine. He recounted that he felt "uneasy with the prospect of charging people money for trying to make them healthy"(l). Lipmann came to biochemistry by taking a three month course given by Peter Rona, a collaborator of Leonor Michaelis. In order to have a firm basis for future biochemical work Lipmann decided, however, to study chemistry first, which he did in Königsberg under Meerwein. For his Ph.D. thesis and first research work Lipmann then went to what was probably the best lab available at the time, the one of Otto Meyerhof at the Kaiser-Wilhelm-Institute here in Berlin. He started work on creatine phosphate, on the fluoride inhibition of glycolysis, and on other topics related to glycolysis. Although Lipmann must have been a very dedicated research worker already at that time, he still liked to tell how much he enjoyed the rich cultural life both during his one semester in Munich and later in the middle and late twenties in Berlin. His brother was working as a dramatic producer at Leopold Jessner's Staatstheater. A close friend of Fritz Lipmanns was a painter. In this circle of artist friends Lipmann also met his wife-to-be, Freda Hall. When Meyerhof moved to Heidelberg in 1930 Lipmann joined him there for half a year, but then returned to Berlin to work with Albert Fischer in, what was very rare at that time, a tissue culture lab. This was his first bona fide job after several years either without pay or with only very modest fellowships. Let me quote a few sentences from Lipmann's autobiography (2). Adapted from H.G. Zachau "Memorial Address on Dr. F. Lipmann, delivered at the 17th FEBS Meeting Berlin on 26th August 1986", Biol. Chem. Hoppe-Seyler 367, 1183-1184, 1986, and from Orden Pour le mérite für Wissenschaften und Künste, Reden und Gedenkworte; Verlag Lambert Schneider, Heidelberg, in press.

64

Hans G. Zachau

"Finding a job was a new and harassing experience. I did not have much of a reputation, and I did not have, and still lack, the gift for making an impression. In addition, however, I discovered that being Jewish in Germany of 1930 was already a great handicap if one was looking for a university position. Even liberal professors were reluctant to put us on their staffs; they expected trouble." During a year at the Rockefeller Institute in New York in 1931 Lipmann started to work on the phosphorylated proteins of egg yolk and milk, a topic which he was going to take up again many years later. In view of the upcoming Nazi rule Lipmann did not return to Germany in 1932, but instead went to Copenhagen were Albert Fischer had started a new lab. Lipmann's main interests during his six years in Copenhagen were the Pasteur effect and the study of acetyl phosphate as a product of pyruvate oxidation in Lactobacillus. In 1939 when war and the German occupation of Denmark became likely events Lipmann managed to secure a job in the U. S. with Dean Burk in Du Vigneaud's department at Cornell Medical School. It was the time when Kögl's claim of Damino acids in tumor tissue was an important issue. With the help of D-amino-acid oxidase, Lipmann was able to show that this hypothesis was very unlikely. There was an interesting side-aspect to this work. Together with Rollin Hotchkiss, Lipmann could show that the newly isolated bacterial polypeptide antibiotics tyrocidin and gramicidin contain D-amino acids. Starting with his thesis, Lipmann always had a particular interest in biologically active phosphate compounds. In 1940/41 he wrote his famous article, "Metabolic Generation and Utilization of Phosphate Bond Energy", which appeared in Volume 1 of Advances in Enzymology (3). In this article he clarified the relationship between various metabolites with high energy phosphate bonds, defined the term "group potential", and introduced the well known squiggle for an energy-rich bond. It is remarkable that up to the age of 40 Lipmann worked mostly by himself. A large number of experimental papers with him as the only author were the result. There was of course also a number of joint publications with colleagues. It was only relatively late, however, that Lipmann had the means to take on coworkers and build a group of his own. In 1941 Lipmann moved to the Massachusetts General Hospital in Boston and continued to work on active acetyl, then concentrating on what was to become coenzyme A. This work brought him the Nobel Prize in 1953. After the Nobel Prize, Lipmann rather than retiring, made the fullest possible use of the research money then more easily available, and gathered together a fair sized research group. Major contributions in the following years were on carbamoyl phosphate and on sulfate activation. Still in Boston Lipmann's group had started to work on protein biosynthesis. This work was then continued on a large scale after 1957 at the Rockefeller University in New York. Work on tRNA and its energy-rich amino-acid bond and work on elongation factors may be mentioned from this period. Since the middle sixties non-ribosomal peptide synthesis was in the center of interest. One of the last pieces of research done in Lipmann's lab, when he was over 80, combines his old interests in group transfer and energy-rich bonds with his keen

Fritz Lipmann: June 12, 1899-July 24, 1986

65

sense for timely topics. It was the time of excitement about the tyrosine phosphorylation in oncogene-coded phosphokinases. Fukami and Lipmann (4) could show, by careful equilibrium measurements, that the tyrosine phosphate within a protein has an energy-rich bond with a AG close to the one of the y-P of ATP. Lipmann talked about this result with his friends with the same enthusiasm that he had shown before when he recounted the beautiful oxygen uptake in the D-amino-acid oxidase reaction of tyrocidin or when he recalled the experiments with those highly efficient mitochondria of pigeon breast muscle. It is almost impossible to do justice to Lipmann and his research in such a brief account. With an intuition we can fully appreciate only in retrospect he devoted himself to the basic problems of biochemistry and he always aimed at clear and simple answers. To his many coworkers and students, altogether more than 140, Lipmann was a good boss, to some he became a good friend. He enjoyed much the regular reunions of his former coworkers, the last ones being held at the occasions of his 80th and 85th birthdays. Soon after the war Lipmann opened his laboratory also to German coworkers. This was an important service to German biochemistry that had suffered severely from the expulsion of the Jewish scientists. It is not surprising that Lipmann's achievements have been widely recognized: several honorary degrees and memberships in Academies, the Nobel Prize, the National Medal of Science. We can be proud that he was a member of the Orden Pour le mérite and an honorary member of Gesellschaft für Biologische Chemie. I will not end this memorial address in an overly laudatory way because I know that Lipmann would not have liked that. However, I want to express my admiration for him. Lipmann was not only a very clear-sighted scientist; I think we can call him a wise man. He took the big decisions of his life right and he took the decisions at the right times. It is quite something if one can say that in looking back on such a long life - Lipmann chose his teachers right; (we all know that one cannot be careful enough in choosing one's teachers). - He left Germany well before the Nazis took over. And he left Europe in time before the outbreak of World War II. - He chose big, important topics for his research work, choosing them when they were just becoming attackable. - And he was wise also in the one big personal decision of his life, that is when he married Freda Lipmann, his wife of 55 years. Lipmann was one of the great scientists of this century and he was a person whom we will keep in fond memory.

66

Hans G. Zachau

References 1. 2. 3. 4.

Lipmann, F. 1984. Annu. Rev. Biochem. 53, 1-53. Lipmann, F. 1971. Wanderings of a Biochemist. Wiley, New York. Lipmann, F. 1941. Adv. Enzymol. 1, 99-162. Fukami, Y. and Lipmann, F. 1983. Proc. Natl. Acad. Sci. USA 80, 1872-1876.

Fritz Lipmann Molding the Design of Molecular Bioenergetics Herman M.

Kalckar

When Fritz Lipmann in 1927 selected as a place for research, the Kaiser Wilhelm Institute in Berlin-Dahlem, "Abteilung MeyerhoP' was his choice. Otto Meyerhof and A.V. Hill won a joint Nobel Prize in 1922 for their bold attempts to relate mechanical muscle work during contraction to lactic acid generation. This multidimensional approach appealed to Lipmann's "Faustian" temperment, as he expressed it in a recent autobiography. After some biochemical studies of glycolysis in intermediary reactions in muscle extracts in the lab, with Karl Lohmann as his teacher and coworker, Lipmann tried his luck on studies on intact muscle. He soon felt that his experiments in this area did not render the strong confirmation of the Meyerhof-Hill doctrine of constancy between lactic acid production, heat generation and mechanical work which he had anticipated. This probably gave the young researcher some uneasy feelings. Young Lipmann learned, however, that Gustav Embden, the outstanding muscle physiologist at University of Frankfurt am Main, was having much the same type of difficulties. Moreover, Embden found that the time relation between mechanical work and lactic acid formation was "loose", since he often found a considerable amount of lactic acid generated after a tetanus. Then, in 1930, a brilliant young Danish physiologist, Einar Lundsgaard made an unanticipated and revealing discovery. He found that monoiodoacetate injected into rabbits or into frogs brought about within one hour a general state of muscle contracture, yet no detectable amount of lactic acid was formed. If the gastrocnemia were isolated long before the rigor occured. Lundsgaard found the isolated muscle, upon electric stimulation, was able to perform a limited number of twitches, yet no lactate was formed. Lundsgaard found, however, the phosphocreatine, newly discovered by C. H. Fiske and Y. Subbarow in Boston, was undergoing rapid fission into creatine and inorganic phosphate. Lundsgaard made the bold proposal that this fission represented an energy source closer to the restoration of mechanical muscle energy than did glycolysis. This proposal was greeted with enthusiasm in various laboratories in Germany, foremost by Embden who knew Cyrus Fiske from the International Congress of Physiology in Boston 1929. In contrast, Meyerhof who had just finished a monograph on the pivetol role of glycolysis in muscle contraction and did not recognize the abberant data from Embden's lab, felt first embarrassed by the Lundsgaard proposal. Although Lundsgaard had been invited by several universities in Germany, including Frankfurt, Lundsgaard thought it more provocative to accept an invitation from Meyerhof at the new KWI in Heidelberg. This was indeed a fortunate choice, because here Lundsgaard met Lipmann for the first time and

68

Herman M . Kalckar

they established right away a "rapport" of mutuality. Lipmann had already studied the two Lundsgaard papers and sensed the enormous perspectives of the coupling between phosphoryl fission and mechanical energy. Lipmann contacted Meyerhof and asked if he could perform a monitoring experiment in normal muscle during contraction. Lipmann inserted a thin frog muscle into a Warburg manometer vessel in bicaronate C 0 2 buffer and stimulated the muscle by way of a platinum thread melted into the Warburg manometer vessel. If phosphocreatine undergoes fission into creatine and inorganic phosphate, the creatine being a base would bring about C 0 2 absorption during the first series of twitches. Lactic acid formation which would ensue later, would then give rise to carbon dioxide liberation from the bicarbonate buffer. This is precisely what happened, i.e., the carbon dioxide absorption and evolution corresponded to an initial phosphocreatine fission and a subsequent lactic acid production. (1, 2). This masterful experiment, a holistic type of approach, finally convinced the physiologist Otto Meyerhof that the Lundsgaard hypothesis was correct. Even before this event, Lipmann had been coauthor with Karl Lohmann, on a more chemically oriented program on phosphoric esters in muscle extracts. They had used Embden's device for arresting glycolysis in muscle "pulp" and now described the accumulation of some novel types of phosphoric esters highly resistent to acid hydrolysis. Lipmann later expressed his unreserved admiration for Embden for identifying the acid-resistant phosphoric esters as a mixture of phosphoglycerate und glycerophosphate. In 1934, Lohmann and Meyerhof described phospho-enolpyruvate (PEP) formation from phospho-glycerate, and the well known EmbdenMeyerhof pathway could be formulated (1). All these events fired Lipmann's passion for experimental biochemistry, and this was further stimulated by his brief meeting with Otto Warburg, the great designer of biochemical experimentation. I should mention briefly that Lipmann in 1931 as a Rockefeller Research Fellow in P. A. Levene's department showed his own power as an analytical chemist in protein chemistry by identifying the phosphorus residue in phosphoproteins as phosphoserine. In 1932 it was clear to everybody that the Weimar Republic was being undermined by the "Third Reich" and no scholar of Jewish extraction considered returning. Fortunately, in Lipmann's case, scientists with great influence in the development the sciences in Denmark were fully prepared to offer an attractive research position to Fritz Lipmann. Einar Lundsgaard's mentor, Valdemar Henriques, senior member of "Carlsbergfondet", as well as the director of the new Biological Institute, Albert Fischer, concentrated their efforts to create an attractive research associateship for Lipmann in Copenhagen. Lipmann was indeed pleased to accept this invitation.

Fritz Lipmann Molding the Design of Molecular Bioenergetics

69

Lipmann's scientific approach in Copenhagen Lipmann's early experimental work to Fischer's tissue cultures, already compelled him to introduce important revisions in formulation of the metabolic biology of cancer cells. Moreover, he also felt that the Warburg postulate of 1925 needed an amendment. According to Warburg normally growing animal cells with access to oxygen manifest pure respiration without glycolysis; in contrast cancer cells although respiring as well, showed marked aerobic glycolysis. However, Lipmann found that normally growing avian and mammalian tissue cultures well provided with oxygen, not only showed high respiration but quite often generated considerable amounts of lactic acid. The difference between normally growing animal cell cultures and cultures of cancer cells, in regard to aerobic glycolysis, seemed therefore more subtle than anticipated. Lipmann also changed the formulation of the so-called Pasteur-Meyerhof cycle, paving the way for a more direct chemical type of interference. Thus he was able to demonstrate that oxidation of sulfhydryl groups by oxidants like quinone or iodine, greatly suppressed the glycolysis in cell extracts. (1, 2). An other important initiative was Lipmann's reexamination of a postulated fermentative pathway, which he unmasked as a respiratory pathway for glucose-6phosphate, generating carbon dioxide. In short, Lipmann was the "father" of the "G6P pentose shunt" pathway. This was published in 1936 (3). Lipmann had several coworkers during his stay in Copenhagen including Scandinavian post doctoráis and his friendship with Lundsgaard was resumed, and that is where I came to know him, and learn from him.

Well Over Fifty Years Ago Lipmann Became my Mentor, a Very Special Mentor Let me pick a few of his musings which revealed his constant struggle against accepted ideas which sensed were becoming "passe". Otto Warburg's early cancer research, just mentioned, was already making a great impact in scores of biochemistry and cancer laboratories in the USA. I remember a response in 1935 when I asked Lipmann about a sound planning for my own research, citing with great enthusiasm the new field "metabolic biology of tumors or tumor cultures". Lipmann was pondering for some time, and then he said in his hesitant "sostenuto" tempo, that Warburg's genius was certainly the introduction of the new tissue manometric methods, but that cancer research as of 1935 was becoming confusing. I have just mentioned in this Berlin introduction, Lipmann's own observations on certain embryonic tissue cultures, manifesting tumor-like aerobic glycolysis. Lipmann's suggestions to me were forward-looking - "why not use these flexible manometric techniques for new basic problems in metabolic biology?" "Read articles," he suggested "by Hans Krebs or Albert Szent-Györgi or Engelhardt, who had published about their observations on metabolic events in cell dispersions or in

70

Herman M. Kalckar

nucleated avian erythrocytes, and then do not forget the Fiske-LohmannLundsgaard methods on analysis of phosphate and the phosphoric esters." Lipmann's advice turned out to be great advice. In this manner, I found my way into areas like oxidative phosphorylations in extracts of kidney cortex (fresh kidneys obtained at the onset of Lundsgaard's liver perfusing experiment on rabbits or cats). In 1938,1 remember Lipmann's special compliments to me, when I reported to him about my identification of phospho-enol-pyruvate (PEP) formation from malate and inorganic phosphate in well oxygenated cortex dispersions. These observations I had first published in Nature in 1938 (4). As you all know, Merle Utter and Kiyoshi Kurahashi expanded this field greatly about 10 years later, discovering GTP as a highly active participant. During the late thirties, Lipmann developed a close friendship with the profound and brilliant protein chemist at the Carlsberg Laboratory, Kaj Linderstr0m-Lang. Lang was very impressed with Lipmann's bold and new ideas about biosynthesis of peptide bonds, invoking ATP and transphosphorylations. The model in the thirties, centering around enzymic peptide synthesis by catalysis in the reverse by peptidases, seemed inadequate. A young brilliant American scholar, Rollin Hotchkiss, who was close to Lang, felt much the same way. In 1939 Lipmann published three important papers which were to change our concepts not only of carbohydrate metabolism but also of protein and lipid synthesis. He found that an acetone powder extract from Lactobacillus delbrückii was able to catalyze the oxidation of pyruvate when three soluble factors were added. Two of the cofactors were thiamine pyrophosphate (described by Lohmann) and flavinadenine nucleotide (described by Warburg). However, the absolute requirement for inorganic phosphate was an exciting surprise. Lipmann was particularly alert, because the Warburg group had found, at about the same time, that glyceraldehyde dehydrogenase from yeast required inorganic phosphate as well as a pyridine nucleotide (later called NAD). The product formed 1,3-diphosphoglyceric acid, was the first enzymatically generated acylphosphate described. Lipmann had already suspected that "Acetyl phosphate" might be formed in the oxidative pyruvate decarboxylation reaction and he took bold steps to try to prove his case. Moreover he also suspected the reaction would potentially be able to convert AMP to ATP. All this was unravelled by an "unrivalled solo performance" within the year 1939. Lipmann prepared acetyl phosphate by an old "prescription" in organic chemistry (Liebig's Annalen 1864), using tri-silver phosphate, phosphoric acid and acetyl chloride. In this way he obtained a mix with significant amounts of monoacetyl phosphate. This mix was incubated with a crude dialyzed extract from L. delbrückii and upon addition of 5'-AMP, appreciable amounts of ATP were formed after 1 hour incubation. It had not escaped Lipmann's attention either that acetyl phosphate may not only serve as a phosphoryl donor, but also as an acetyl donor. Less than a month after my arrival to CalTech in early March 1939, I received a special letter from Fritz Lipmann in New York, telling me about the positive outcome of the bold acetyl phosphate experiment and the favorable response from Nature giving it high priority for publication.

Fritz Lipmann Molding the Design of Molecular Bioenergetics

71

Lipmann's letter provided me with an immense stimulation in my presentations at the small workshops which I had been invited to present at CalTech Biology in the spring of 1959. The participants counted Dave and Jim Bonner, Borsook, Charlie Coryell, Delbrück, Norm Horowitz and others. From the beginning, my main topic centered around the possible function of phosphorylations in macromolecular biosyntheses, not forgetting Carl and Gerty Cori's discovery of glucose-1-phosphate in 1936. Let me finish with an old saying which seems to fit the development of molecular biology: "Life is what's happening while you are making other plans." Fritz Lipmann always kept this in mind on his winding trail to new discoveries.

References 1. 2. 3. 4.

Lipmann, F. 1971. Wanderings of a Biochemist. Wiley, New York (cf. pp. 33-111). Lipmann, F. 1984. A Long Life in Times of Great Upheaval. Ann. Rev. Biochem. 1 - 3 3 . Lipmann, F. 1936. Fermentation of Phosphogluconic Acid. Nature (London) 138, 588. Kalckar, H. M. 1938. Formation of a New Phosphate Ester in Kidney Extracts. Nature (London) 142, 871.

Recollections of Fritz Lipmann, 1941-1945 H.A. Barker

In 1941-42, I was scheduled to have a first sabbatical leave from my position as Assistant Professor of Soil Microbiology at Berkeley. The year before I had come across Lipmann's paper on the stimulation of pyruvate oxidation by orthophosphate in extracts of Lactobacillus delbriickii and the probable formation of acetylphosphate. A little later I read his review on "The Metabolic Generation and Utilization of Phosphate Bond Energy" and decided that I would like to spend part of my sabbatical leave with him. Although I was impressed by his ideas about the role of high energy phosphate compounds in metabolism, I was particularly desirous of learning how to prepare and handle cell-free preparations since until then all of my studies of bacterial metabolism had been done with growing cultures or with suspensions of washed cells - so called "resting cells". I wrote to Lipmann, probably in March 1941, asking if I could spend part of my sabbatical with him. He replied that that would be "very pleasant" but said he was still uncertain where he would be located in the autumn. He admitted to being "quite dilettantic in bacteriological matters" but was hopeful that "simplified" bacterial systems could be used to clarify mechanisms more difficult to study in animal systems. He said his main interest was in following pyruvate degradation further. He hoped to "localize the action of the vitamins". And he expressed an interest in a reaction I had studied, the oxidation of ethanol to acetate coupled with the reduction of C 0 2 to methane, catalyzed by Methanobacterium omelianskii. In mid June, Lipmann informed me that he would soon be located in Boston with an appointment "as advising biochemist to the Surgical Laboratories of Harvard Medical School in Massachusetts General Hospital" and would have a laboratory to do the work he chose. He mentioned that Dr. Hastings had arranged this appointment. In the meantime I had suggested that I should work on metabolism of dried propionic acid bacteria that both degraded a variety of substrates and were known, from the work of Harland Wood and C. H. Werkman, to use C 0 2 in the formation of succinate. Lipmann replied that the proposal to work on propionic acid bacteria was "an extremely good one." Since Lipmann's space in MGH had to be renovated and would not be ready until November, I arranged to spend September and October in the Bacteriology Department of the Yale Medical School in New Haven. I first met Lipmann when he stopped off" briefly in New Haven on his way to Boston in the latter part of October. He was traveling with Dr. Nachmansohn a friend from his days in Meyerhofs laboratory, I think. There was not enough time to talk much science, but we made a start at getting acquainted. I had brought along my camera and took a picture of Lipmann & Nachmansohn (Figure 1).

74

Figure 1

H . A . Barker

Fritz Lipmann and David Nachmansohn, October 1941 (Photo by H.A. Barker).

Early in November I moved the family to Cambridge where we lived in an apartment on the north bank of the river, about a half mile west of the MGH. The Surgical Laboratories was apparently a research unit of the Harvard Medical School and was headed by a Dr. Churchill. For me he was a rather remote figure whom I talked with only a very few times during 6 months in the Laboratories. Dr. Oliver Cope, a pleasant young Assistant Professor of Surgery, appeared to be in more immediate charge of our facilities and funds. Other members of the medical staff were Dr. Champ Lyons, Associate Professor of Surgery, who was concerned with bacterial infections and had two technicians isolating and identifying bacteria, and Dr. Beecher, Professor of Anaesthesia who was always carrying out elaborate and to me, rather gruesome anaesthesia experiments on dogs. Lipmann and I were fitted into this group although we really did not belong to it. Lipmann had two small rooms for laboratory work, one of which had a writing desk. I was assigned one side of a laboratory bench, possible 8 ft. in length in a room otherwise occupied by one of Dr. Lyons technicians. There were several pieces of apparatus that we could share with the technicians, a centrifuge, pH meter, manual colorimeter, which we always used for phosphate determinations, and a photoelectric colorimeter, which was seldom used. Lipmann also had a good Warburg-type respirometer. During the first month or two after my arrival Lipmann was involved in experiments to isolate and characterize enzymatically produced acetylphosphate. Since acetyl-

Recollections of Fritz Lipmann, 1941-1945

75

phosphate reacted like orthophosphate in the usual colorimetric methods of estimating orthophosphate, he had had to develop a more specific method for distinguishing between the two compounds. He had found that orthophosphate could be quantitatively precipitated as the calcium salt from neutral alcoholic solution, whereas acetylphosphate remained fully soluble; the colorimetric assay could then be applied to each fraction. With the aid of this differential precipitation method he was able to isolate fairly pure acetylphosphate from reaction solutions. The product was characterized by determining the content of phosphate and acetic acid after hydrolysis. Lipmann determined the phosphate and I determined the acetic acid by steam distillation and titration. The results agreed well with expectation. I prepared to investigate some aspect of the metabolism of the propionic acid bacteria by growing up a batch of Propionibacterium pentosaceum and drying the washed cells by the "sloppy drying" method used by Lipmann to obtain active preparations of Lactobacillus. When stored in a refrigerator these dry cells retained the ability to catalyze a variety of enzymatic reactions for several months. Initially we concentrated on the utilization of lactate and pyruvate because Lipmann had learned about the experiments of Chaix-Audemard, working in Fromageot's laboratory, who reported that pyruvate is fermented by propionic acid bacteria in the presence of sufficient sodium fluoride to inhibit lactate fermentation completely. This seemed worth checking since lactate was generally thought to be an intermediate in the conversion of pyruvate to propionate. Our experiments (Barker & Lipmann 1944) confirmed the above observation and showed further that fluoride inhibited the reduction of lactate but not its oxidation to pyruvate. Furthermore in the presence of fluoride, pyruvate was converted to propionate and acetate without any accumulation of lactate. This led to the postulate that lactate and pyruvate are converted to an unknown precursor of propionate by different enzymatic reactions, only that from lactate being inhibited by fluoride. This conclusion seemed justified, but did little to clarify the path of propionate formation. Little did we realize how complicated the chemistry of the propionic acid fermentation would turn out to be! Even now it is not obvious how fluoride interferes with lactate utilization, though it probably interrupts electron flow from lactate to fumarate. Lipmann and I also made some observations on the phosphate metabolism of propionic acid bacteria utilizing various sugars and polyalcohols. We observed the transfer of phosphate from ATP or phosphoenolpyruvate to various acceptors, and oxidative phosphorylation with sugars and polyalcohols in the presence of fluoride and a suitable oxidant. Without going into details, the results showed that oxidative generation of phosphate esters and transphosphorylations are important reactions in propionic acid bacteria as in yeast and muscle. Several of the experiments on the oxidative phosphorylation of sugars and polyalcohols, in particular those with erythritol, sorbitol, inositol and arabinose, were done by Lipmann. Although most of the experiments were carried out in early 1942. The results were not published until 1949 because other problems in both our laboratories were more interesting. However, the experience I obtained with Lipmann proved to be very valuable in my later scientific career.

76

H . A . Barker

I was Lipmann's only close associate, other than his technician, during the six months I was in his laboratory. We were somewhat isolated intellectually since we had little in common with the M.D.s in the Surgical Laboratories, and Lipmann was seldom called upon to provide advise to others on biochemical problems. During the time I was there he gave one poorly attended seminar for the professional staff at the Mass. General Hospital and perhaps one seminar at Harvard. We generally ate bag lunches together in the laboratory and were occasionally joined by one of Lipmann's friends. These included Gerhard Schmidt, Gertrude Perlmann, David Nachmansohn and Birgit Vennesland. Aside from Vennesland, I do not recall anyone from the Harvard biochemistry group who visited Lipmann during the time I was there. Lunch conversation was commonly devoted to some aspect of microbiology or biochemistry, although we touched on other topics as well. In the spring the American Society of Biological Chemists met in Boston. Lipmann gave a short talk on his characterization of acetylphosphate and his unsuccessful attempts to demonstrate a function of the compound in animal tissues. He was disappointed to be unable to show that it could serve neither as a phosphoryl or acetyl donor in animals. Shortly before I left Boston near the end of April, we went to a performance of Porgy and Bess with the Lipmanns. While we were waiting for the curtain to rise, he told me with some excitement and evident relief that his appointment at MGH had just been renewed for another year. In this connection I recall that some weeks earlier I had been asked by Dr. Churchill, Director of the Surgical Laboratories, what I thought of Lipmann's ability as a scientist. I gave him a good recommendation. I think my presence as a Fellow of the Guggenheim Foundation helped to bolster Lipmann's standing with the medical people, who seemed to have little or no personal appreciation of his ability and scientific background. Of course, the discovery of acetylphosphate was largely of theoretical interest at that time, and since Lipmann was unable to demonstrate its formation or utilization in animal systems, the relevancy to medicine was not obvious. Furthermore Lipmann's style of communication in his seminars was not such as to easily impress those on the periphery of biochemical thought with the significance of his ideas. I did not see Lipmann again until well after the end of the war, but we corresponded from time to time and his letters throw some light on the gradual development of his work on acetylation reactions and on his personality. In what follows I shall quote from Lipmann's letters. On June 27, 1942, Lipmann sent me some data on some of the experiments on oxidative phosphorylation of sugars and polyalcohols by propionic acid bacteria that he had carried out and added: "Time has gone without much having been done. Got some money to buy new gadgets and spent much time finding out if and how to get them. Hope to get a spectrophotometer for U.V. but am doubtful if I will succeed." „In the meantime F. F. Nord has asked for a paper for a new journal, Arch, of Biochem., for August and for that I got out some old experiments on ero tonic acid in liver. I thought to mention there - discussing the difficulties to imagine presently

Recollections of Fritz Lipmann, 1941-1945

77

the reaction fatty -*• keto acid —• something like 'In recent not yet published work by Barker and Lipmann a very similar problem was met with regarding the interrelation of propionic-lactic-pyruvic acids'. - if you have nothing against it. The writing of this paper has greatly upset my schedule and I have nothing done with either acetylphosphate or other things. I should like to hear how you think now about the way to publish our data after having given a seminar about it. I think your suggestion to make it into two reports was very good. And you should take the writting of the lactic acid story and I take the phosphorylation side. It sounded as if you had a good time there [vacationing in the mountains]. We are leaving next week for Vermont. I have somewhat bad conscience because most people don't take vacation so seriously this year. Nevertheless I am happy to get out for a time." August 31,1942. "We had a pleasant stay in Vermont, being quite enchanted with it as ever. I had hoped to do, but did not much writting and look with horror into the future, am not even through with a final report on acetylphosphate. I received a letter from Koepsell-Johnson and they sent me a little enzyme. [Note: The KoepsellJohnson enzyme preparation from Clostridium butylicum was reported (J. Biol. Chem. 145, 379-386, 1942) to convert pyruvate anaerobically to acetylphosphate, carbon dioxide and Η 2 ·] I have not tried it yet, but it seems as if they get acetylphosphate. " "Regarding Ruben's plans to try the reversibility [of the conversion of pyruvate and phosphate and 0 2 to acetylphosphate and carbon dioxide]: I would be glad if that could be definitely settled. I believe the reaction is reversible, it is only difficult to find the right reductant. I think I told you that Vennesland, who was with Hastings, once tried a breakdown experiment (only one!) which was not clearly negative, a trace of *C0 2 in the pyruvate fraction which she could not decide if true or carried along. Shortly afterward she left Boston. Now, I wonder if I should not ask Hastings and her if they are interested still and would like to do some more experiments. I had, in fact, already mapped some experiments with Vennesland. She is coming to Boston very soon and I shall talk to her about it." October 13, 1942 "I just talked with Hastings and he says that they have shut down the cyclotron and are not able therefore to do anything. So - I would be glad if Dr. Ruben does this experiment. I wonder if it will be possible to get the bacteria. I asked recently Werkman, but he does not have them. Hope Davis will help." "Am very glad you are writing on the propionic bacteria paper; have done little myself and have not much hope to get serious writing done before the review on O/R is off my hands." "I worked recently on a simplification of a preparation of acetylphosphate and I hope soon to have a very easy recipe ready..." "With the enzyme from Koepsell-Johnson I made some experiments and was glad to find the same as they found, and am fairly convinced that there is acetylphosphate formed. It is also possible to get pretty good phosphate transfer with their preparation, anaerobic with pyruvate. This time it seems to be real."

78

H . A . Barker

December 1, 1942. "I am quite afraid to confess that I am not ready with the paper you have sent me. Now Annual Review is coming increasingly near and this year will be lost. I have not even written my paper on acetylphosphate. But I have done some experiments which may give some further clue on acetylphosphate. In the first place the method of preparation of acetylphosphate could be very simplified. If you or Ruben are interested, I can send you the recipe. Then the Wisconsin people sent me some of their enzyme and I have played around with it a good deal. I feel sure that there is acetylphosphate formed. I could confirm that a labile phosphate compound is formed, found also a coupling with phosphorylation of glucose, and finally a rapid exchange between inorganic (radioactive) Ρ and acetyl-P in the presence of the butylicum enzyme. This exchange, I am looking into now. Am also going to try other enzyme systems. I don't know yet what it means. I have not even written to Wisconsin about it." April 5, 1943. "I enjoyed hearing from you and the interesting things you tell me about work, your own and DoudorofTs. It's fine that he got the phosphorolysis of sucrose, it will help to make people understand the universality of the reaction." "When your letter arrived I was in the midst of writing this Review (for Ann. Rev. Biochem.) and near crazy, because I was far from ready (that was Jan. 15). It took me into the second half of February to finish it, and I hope I shall never again do it." In the meantime your culture [of Lactobacillus delbrückii] had arrived. Fortunately it kept alive until late February and I made a dry preparation, with good yield and the expected properties... With the suspension there was only little phosphorylation, as I had seen before with fresh preparations.... This preparation did not fix Ρ in absence of fluoride; earlier experiment had given fixation without fluoride in older preparation. I think it is good for experiments to age the dry preparation at 0° in desiccator. This bacillus is therefore the same I worked with. Thank you for all the trouble. I am very glad it becomes available again in this country." "I have not done much since I finished the Review. Started some experiments on acetylation in liver, of sulfanilamide which have not yet gone very far. Isn't there an organism called something like B. acetylcholinusl I wonder if, and if why it is called so." "I saw Ruben's article in JACS about photosynthesis and phosphorylation. I have the suspicion that you brought the phosphorylation virus to the West Coast. With this I am sending a recipe for preparation of acetylphosphate, somewhat too elaborate - I had it written down by my technician as we do it now. The propionic acid papers are on my mind..." June 21,1943. "It is well a few months ago that I wrote last about me and here. Our fourth floor is becoming more and more deserted. Dr. Churchill became a colonel quite a while ago and is now somewhere overseas. Lyons left recently for a military hospital in this country first. And Beecher has just signed up and will leave very soon. Dr. Cope is, for the duration, acting chief of the surgery. He and I are now soon the only remaining, although the laboratories are rather busy, Beecher keeping his laboratory going and part of the Bacteriology crew remaining."

Recollections of Fritz Lipmann, 1941-1945

79

"Work: I have tried hard to find out how acetylphosphate behaves with animal tissues. Have not succeeded to make it very clear. There is - at least in pigeon liver a transfer of Ρ from ac-P to glucose. The great difficulty in all such experiments is the very rapid breakdown of ac-P, 1 - 2 min. half life, in all tissues, the possible significance and mechanism of which is not yet clear. - Did some experiments on acetylation of sulfonamides and choline. Pigeon liver acetylates sulfonamides easily, also as homogenate. Interesting is that oxygen is needed although acetate stimulates considerably. Acetyl-P did not give encouraging results, nor did it acetylate choline in brain homogenate." "Saw DoudorofTs paper on phosphorolysis of sucrose. Fine! - Write sometime about you and things up there. Some time ago Kamen was here and we talked a bit about photosynthesis and phosphorylation. - This summer I hope I get some writing done. I am so far behind that it makes me more and more unhappy and something radical has now soon to be done about it. I simply have to give up laboratory work for some months and that is so difficult to do if there is no publisher somewhere whe sends you telegrams." November 19, 1943. "The immediate reason for my writing is that I remember you once mentioned that you know a source of powdered glass. I got some from Werkman, but may need more and don't like to ask him for more.... Anyway, I would be happy if I could get about a kilogram or two. Would like best if we could buy it. I have started with E. coli. Got about as far as to repeat Werkman's experiments and convinced myself that acetylphosphate is formed and transfers to glucose. Would like to purify the enzyme because the crude extracts, when they react with pyruvate at all, tend to break down acetylphosphate rather rapidly." "Hasn't much happened in the meantime. I mean to me personally. We remembered our common theater going when we recently saw the film 'Watch on the Rhine' (and thought the play better) and re-saw 'Porgy and Bess' - this time together with the Walds." "I have not found anything exciting in the meantime. Am planning some work correlated with shock, there are some interesting metabolic aspects." December 3, 1943. "This is to let you know that I am hard at work on our papers and to prevent you from taking steps to put me into prison, where certainly I belong for my negligence." January 24, 1944. "Received the manuscript and consider it now in a practically final shape. The other one is progressing although slowed up by various interferences. So far the general plan and tables are ready and now to be filled. I see already that I will have difficulties to write the pentose part and shall do so only rather sketchy." "One of the interferences was a recent 'professional' visit to New York. There Green told me that he is planning to isolate my Delbrückii enzyme and I said I would be delighted and - being not too sure about the reliability of the Type Culture Collection - would rather write to you and ask you to send him the Davis strain."

80

H . A . Barker

June 18, 1944. "The proof of the manuscript arrived suddenly here yesterday apparently to speed up publication. I sent it back already and ordered 300 reprints (enough?) half to be sent to you. Don't bother about pay, goes on hospital account." "The large scale sucrose synthesis was a fine and effective finale. Made quite a sensation - Ν. Y. Times two articles, News Edition of A.C.S. Didn't it in a way start at a lunch discussion of ours about the reason for the preference of some organisms for a disaccharide being on the energy level of a phosphate ester - it gave free entrance to the phosphate." "I had a rather interesting spring. When Werkman's people found in Coli extracts pyruvate split into acetylphosphate and formate, my hopes to get at the essential feasibility of reductive carboxylation with acetylphosphate as partner were revived. I got also so far as to show a definite - but very small - formation of keto acid with the dinitro-phenylhydrazine method, from acetylphosphate + formate. Then, Werkman very generously invited me to Ames to do with them the isotope test. That we did during May and using C 13 -formate (made enzymatically, H 2 + C0 2 ) in a hydrogenylase-free coli extract (no C 1 3 going into C 0 2 ) the carboxyl in pyruvate becomes pretty 'heavy', up to 1.60 in our best experiments with added acetylphosphate. Am somewhat excited about it." August 25, 1945. "I have been busy lately mostly with sulfanilamide acetylation in liver extracts, an enzymatic condensation effected by adenylpyrophosphate but without formation of a distinct phosphorylated intermediate. I frequently have pangs of bad conscience for not having done more on our paper. The data are assembled into ten tables so far and I hope to be able to send these on to you for comments, for you to translate into real language, but not before OctoberNovember." "The most important event in our life is the birth of a son, just two weeks old By the way, Kaplan is going to be with me from September and I hope we will get along nicely together."

In Celebration of the Scientific Genius of Fritz Lipmann Sarah Ratner

Unlike many contributors to this volume, I came to know Fritz Lipmann through our mutual scientific interests catalyzed by encounters at lectures and meetings. When Rockefeller Institute (later University), brought him to New York in 1957, our paths crossed much more often: at our monthly Enzyme Club meetings held at Rockefeller, and occasionally at the Rockefeller chamber music concerts. Performances by soloists failed to draw him there. Since I passed his apartment on my way home, I often dropped him off. I was much in awe of him at first but soon lost that feeling when we chatted about carbamyl phosphate. ATP, and his concepts of high energy phosphate ( ~ P). Our rambling conversations usually dealt with biochemistry and music and the performances we had heard or witnessed and sometimes how to write scientific prose with clarity. This great, uniquely endowed man I write of, had over the years become a close and dear friend, thoughtful, kind, and delightfully broad in his interests. I treasure his friendship immeasurably. His wife, Freda, and I also became good friends during the several months each year she joined Fritz in New York. Despite Fritz's reluctance to write reviews or give talks, in 1971 he wrote an unusual scientific biography "Wanderings of a Biochemist" (1) in which he covered the main problems he had dealt with over a span of forty years and the new basic concepts that emerged from his work. Except for a phrase here and there, he left it to the reader to grasp the underlying significance of his achievements. The sad sounding title seemed to carry the shadow of his several transatlantic crossings many years earlier to find safe haven from the Nazis. Actually his outlook was realistic and he regarded science and scientists with an international perspective. One look at the book's explanatory and very clear preface quickly dispells past shadows. He explains how the great interest of his students caused him to change an early draft of the book's structure and organization because his students kept asking "how" and "when" and "where" questions about his novel concepts in relation to what was generally thought at the time he conceived them. He did indeed comply and the book includes many insights into his way of thinking. Although he did not seem aware of it, it was especially notable to me on reading the preface, that the atmosphere in his laboratory was unusual in being pervaded by so lively a spirit of interchange and inquiry. Fritz had the impression, as did many others, that the interest of most postdocs in the history of biochemical developments was limited to the most recent 5 or 10 years. The unique atmosphere pervading his laboratory was generated and nurtured by Fritz Lipmann quite naturally because of the need to clarify and illuminate each problem he had given so much thought to. His laboratory group of carefully

82

Sarah Ratner

selected, gifted individuals were working on different problems. Each felt the time spent in association with him to be the most fortunate and influential period in their lives. I quote below from an essay written by Earl Stadtman (2) referring to a period spent in Fritz Lipmann's lab around 1949-50. It eloquently reflects the laboratory atmosphere that prevailed both in Boston at Massachusetts General Hospital and as I imply, also in New York, years later, when Fritz was restructuring his scientific autobiography. "Although we had first thought that "x" might be the transacetylase itself, Barker and I discussed the possibility that it could in fact be CoA. An opportunity to test this hypothesis presented itself when, after completion of my thesis research in late 1949, I moved to Boston to continue my studies as a postdoctoral fellow in the laboratory of Dr. Fritz Lipmann. The year in Lipmann's laboratory was an unusually rich and satisfying experience for me, both from the personal and the scientific point of view. Barker had assured me that in Lipmann's laboratory I would find more than an opportunity to do stimulating research. Here there was an intangible benefit to be gained by close association with a man of unusual imagination and perception and one who possessed a gentle warmth of character and concern for others that endeared him to all his associates. It was here I discovered that productive laboratories are not merely the reflection of good scientific discipline and expert direction but depend almost as much on the establishment of a congenial atmosphere in which science can flourish as a consequence of free thought, unguarded exchange of ideas, critical discussion and a respectful interaction among all of its personnel. Such was Lipmann's laboratory." Fritz was crowned Nobel Laureate in 1953 for the discovery of CoA, and its function in acetyl activation, and transfer, a few years after his findings were published. His discoveries were pivotal, touching on, and illuminating, several interrelated metabolic pathways under investigation in other laboratories. Severo Ochoa, Chairman of Pharmacology at New York University School of Medicine, was particularly interested because he and J. Stern had shown that the "condensing enzyme" which catalyzes the formation of citrate from oxaloacetate required an acetyl group as the two-carbon donor. Earl Stadtman's work with Dr. Lipmann in Boston was successful in showing that the enzyme transacetylase was necessary for acetyl activation in addition to CoA; they were extremely excited about it. Some time in 1950 Severo Ochoa invited Earl Stadtman to join him in an experiment to test the transacetylase enzyme on his citrate synthesizing system. I was a member of Ochoa's department at the time and awaited the outcome with the keenest anticipation. Even though I was thoroughly immersed in urea cycle enzymes, I regarded the experiment as an occasion of great historical significance. Dr. Lipmann had previously given much thought and successful experimentation to: acetyl activation, the presence of pantothenic acid in CoA, and the role of CoA in acetyl transfer. He had generated and provided proof for the basic concepts but he was not present to witness this experiment, perhaps because he was confident of the successful outcome. The experiment clearly showed

In Celebration of the Scientific Genius of Fritz Lipmann

83

that acetyl phosphate, plus CoA, plus transacetylase and oxaloacetate and condensing enzyme leads to citrate synthesis. Transacetylase was indeed essential (3). Shortly thereafter Feodor Lynen published a brief report of a discovery on one aspect of the structure of acetyl CoA showing that it contained a thioester bond. Lynen came to New York from Munich to participate with Severo Ochoa in an experiment directly concerned with acetyl CoA in citrate synthesis bringing with him a preparation of acetyl CoA. The experiment planned was to test the interaction of acetyl CoA directly with oxaloacetate in the presence of the condensing enzyme. Fritz Lipmann came down from Boston to witness this experiment. It turned out that Lynen's findings as well as Lipmann's were confirmed by the new data. A free -SH group was liberated in stoichiometric proportion to the amount of citrateformed. All the pieces now fell into place (4). The three biochemists simultaneously present, Fritz Lipmann, Severo Ochoa, and Feodor Lynen, on this second historical and auspicious occasion were future Nobel Laureates. Their eagerness to know the answers to questions they were asking was irrepressible and this question was addressed in short order. I doubt if such high level collaboration on a crucial problem could ever occur again. In the 1950's most biochemists knew what was going on in other labs for they read the journals as soon as they appeared. Mutual respect permitted swift collaboration and immediate publication of new results in the Journal of Biological Chemistry. However, on the first occasion Stadtman and Ochoa were the authors (3); on the second occasion Ochoa and Lynen were the authors (4). Fritz Lipmann's name appeared on neither one. He came down from Boston to witness the second occasion possibly because he had learned something from Lynen and perhaps to remind Lynen, by his presence, what Lynen and Ochoa had learned from Fritz Lipmann. The circumstances surrounding these two historically significant occasions would have remained forever buried in journals if I had not lifted them out of my memory motivated by this opportunity to write about Fritz Lipmann. I wish to express in this way his scientific vision, his extraordinary accomplishments, and his unique personal qualities that caused him to become endeared to all who came to know him. The image of Fritz Lipmann drawn and framed here stems from my great admiration for Fritz and his achievements although my personal exposure was somewhat limited. His image will be greatly enhanced by the sensitive and perceptive obituary written by a close associate, Robert Roskoski (5). I quote portions that illuminate the genius of Fritz Lipmann and the depth of his understanding. These lines will also add to the pleasure and appreciation biochemists will have on reading Fritz Lipmann's "Wanderings of a Biochemist." "On the basis of his achievements, he was unquestionably, a biochemical genius. How these achievements were made, however, differ from the expected characteristics of a scientist of such stature. To begin with, he was either unwilling or unable to follow a complicated argument. It was not unusual for him to state that he could not follow the logic of a seminar or a personal conversation. When he wanted to understand a train of thought he asked question after question to get the speaker to make smaller and

84

Sarah Ratner

smaller steps between an observation and conclusion. As a result of this much closer scrutiny, Lipmann might find that a conclusion was untenable. I believe that this constant striving to simplify contributed to his success. His curiosity and desire to learn also prompted him to ask many questions of a general nature, that might seem elementary for a first year graduate student. This curiosity and enthusiasm never waned. After understanding the background of a problem, however, he was often able to make incisive comments or pose questions that helped to clarify ideas and concepts. Nearly everyone who came into contact with him held high regard for his scientific intuition. The source of this astuteness, however, might defy or transcend apparent logic. Lipmann wrote that he lacked the ability to make a good impression. Although he spent considerable time preparing for various symposium talks, he was not a gifted speaker. He was also ill at ease in answering questions after his seminars. After contemplating a question for several days (or months), he might formulate a new line of experimentation which went further than any implications that the questioner might have originally had in mind. Although his thought processes may have lacked speed, they certainly possessed undeniable power."

References 1. Lipmann, F. 1971. Wanderings of a Biochemist, Wiley, New York. 2. Stadtman, E. 1976. in Reflections on Biochemistry In Honour of Severo Ochoa, 161-172, Pergamon Press, Oxford, New York. 3. Stern, J.R., Shapiro, B., Stadtman, E.R. and Ochoa, S. 1951. J. Biol. Chem. 193, 703-720. 4. Stern, J.R., Ochoa, S. and Lynen, F. 1952. J. Biol. Chem. 198, 313-321. 5. Roskoski, R., Jr. 1987. Trends in Biochemical Sciences, 12, 136-139.

Lipmann and "Not Strictly Biochemistry" Thomas Peter

Bennett

Introduction Lipmann agreed to speak with me during the summer of 1959 when I was a first year graduate student at the Rockefeller University. It had taken weeks of attempted introductions before I finally approached Chris Gillespie, Lipmann's secretary and assistant, directly. I proposed a summer research project in biochemistry that would combine notions about D-amino acids introduced to me by my professor Sidney Fox, Lipmann's early interest in D-amino acids in cancer, and the current "hot topic" of activating enzymes. She saw merit in my request and suggested that I come to tea/coffee in the Lab Library the following afternoon and stay to explain my idea to Dr. Lipmann. After thoughtfully listening to me, Lipmann seemingly admonished me in his warm but stern manner: "Ideas are a dime a dozen in this field; the important thing is to get in and do the experiments to support or reject your notions." I agreed, but pointed out that I needed a lab bench to do the experiments. At that moment I became a pre-doctoral student in the Lipmann Lab! I was later to explore whether there were D-amino acid activating enzymes (later termed amino acyl tRNA synthetases) in bacteria that produced D-amino acid containing polypeptides such as gramicidin and bacitracin. My idea secured me a lab bench in the Lipmann Lab, but not the solution to the polypeptide biosynthetic problem (as other papers in this volume document). Yet from that bench I was able to make several contributions and six years later receive my Ph. D. under Lipmann's tutelage. Some two years after our initial encounter I would be admonished by Lipmann, again in a warm and stern manner to "think like a scientist" as we in the Lipmann Lab faced what has been characterized by William Broad and Nicholas Wade in Betrayers of the Truth as the "case {that} rocked the world of biochemistry" (1). Lipmann's candid and decisive handling of this situation impressed me then and with a twenty-five year perspective I am even more impressed. Over the years, the issue of a moral crisis in scientific research, as evidenced by instances of scientific fraud, reached such proportions that in 1981 a sub-committee held Congressional hearings on the subject. This was followed in 1982 by publication of the controversial volume, Betrayers of the Truth by Broad and Wade. Many scientists who were convinced, unlike Broad and Wade, that it was not the barrel of science that was rotten but some apples within the barrel, began speaking out at meetings, conferences and writing about this subject. By invitation I took a message to students on campuses and to civic leaders (2,3,4). It was based on Broad and Wade's writings, Lipmann's admonition to me to "think like a scientist", and my perception of Lipmann at the cutting edge of what has become a serious

86

Thomas Peter Bennett

pragmatic and ethical issue in science. It was grounded in my Rockefeller years in the Lipmann lab. In this task I began with two basic questions: What happens when in some fashion a scientist does not fulfill his or her responsibility to colleagues and to society to strive for excellence, to be truthful, and to be ethical in this work? Does the honor code of science operate effectively? I used as an example a major case with which I am personally familiar. It is the first contemporary case Broad and Wade mention.

Character and "Doing Science": The Traction Case Broad and Wade call this the "Thomas Traction case." It occurred at the Rockefeller University in the Lipmann laboratories, during my second year as a graduate student. It is entirely correct to say, as Broad and Wade contend, that this case "shook the world of biochemistry." And I look back on this case as an object lesson in the necessity for integrity in science, as well as a demonstration of the effectiveness of the honor code of science and the character and integrity of three practicing scientists. The fellowship of science was clearly evident in the Lipmann laboratory, which at my arrival in 1959 housed laboratory benches for 20 to 30 researchers (post-doctoral fellows, research associates, several graduate students such as myself, technicians, and clerical staff) spread out over ten laboratory work rooms. As all of you remember, Fritz Lipmann's laboratory was international, attracting over half its personnel from outside the U. S. The laboratories were in operation 24 hours a day, seven days a week, for we all had experiments ongoing in the study of protein biosynthesis at that time - the eve, if you will, of the "breaking of the genetic code". One of the most important of the many Americanized European rituals which prevailed there was that at about 3:15 each weekday afternoon everyone drifted into the laboratory library for tea, coffee, and professional chat by all, as equals, on the progress of work. (This meant Lipmann, the washroom staff, everyone.) When someone was having problems with an experiment, the group would offer comments and suggestions; the coffee/tea gathering was the source of a daily communal exchange of ideas and insights. All of us recall these gatherings. The individual whom Broad and Wade call Thomas Traction (an apparent play on the word "retraction" by Broad and Wade) joined the Lipmann laboratory as a post-doctoral fellow shortly before I arrived. He was riding the crest of a wave of success based on his doctoral studies at Yale. At a time when the synthesis of DNA, RNA and proteins was just beginning to be understood, he and Melvin Simpson (his major professor at Yale) had made "significant contributions" to the biosynthesis of a key protein molecule, cytochrome c, in cell-free extracts. Their findings had been published to acclaim. Traction was accepted in the Lipmann laboratory because of his success and promise, while Simpson began a sabbatical in Europe to lecture and write about their findings. Shortly after his arrival, Traction began work on a research problem which coincidentally paralleled my own project. He promptly made a discovery about the biosynthesis of a small protein-like molecule (a peptide called glutathione) which

Lipmann and "Not Strictly Biochemistry"

87

was soon published in a biochemical journal with Lipmann as co-author. In a laboratory filled with an international roster of bright, young researchers, Thomas Traction became the "numero uno post-doc." In addition to his scientific "golden touch," he was affable, helpful, filled with boundless energy and with a seemingly endless fund of stories about himself and his exploits which he aired at tea or late in the evenings when we were weary of working at our benches. When Melvin Simpson returned to Yale, reassembled his laboratory began trying to extend the successful experiments with cytochrome c, he difficulties. In the Lipmann laboratory, a new post-doctoral research will call him Dr. L) began working on the synthesis of a third (ophthalmic acid).

group, and experienced associate (I polypeptide

It was at this point that Dr. L failed in attempting to parallel Traction's success in the biosynthesis of glutathione with his biosynthesis of ophthalmic acid, a similar peptide. Dr. L then decided to duplicate precisely Traction's original glutathione experiments in the hope of seeing where his own research might have gone off track. And then strange things began to happen in the Lipmann laboratory. It was as if we had a poltergeist. Substances were mislabelled, the freezer was turned off, distilled water was contaminated. The consequences for our research studies were disastrous; the effect on morale, equally so. Several of us completely halted our laboratory work and moved over to the library; Dr. L, the poltergeist's frequent target, moved to another laboratory at Rockefeller in the hope of sorting out his findings; others, convinced they could accomplish nothing in the midst of this chaos, decided to "see America." Thomas Traction took charge, as the "ranking post-doc," to get the matter in hand, for Lipmann was abroad on a brief trip. Traction had, he said, friends in the FBI and the CIA who would help him get these problems cleared up. When Lipmann returned, at a coffee session, Traction suggested that we all be given lie-detector tests to determine if any of us was the mischief-maker. He said that some friends of his with the police department were willing and prepared to assist. Three of us at this gathering said we would refuse to take such a test as a matter of principle, that it was unthinkable; Lipmann shared our concern and overruled Traction on this. Some days later, a special coffee/tea gathering was called by Lipmann "to discuss the mischief problem further." Traction was unable to be present. The meeting was not a comfortable one for me: I didn't like the proceedings, especially when Traction, whom I genuinely liked, emerged as the focus of negative comments by several others. When I took issue with the process, Lipmann rose like a lion and gave me a verbal thrashing, ending with the words, "Why don't you think like a scientist?" Lipmann concluded the discussion with the promise, "I will talk with him (Traction) later." There were no more "incidents" in the laboratory. Traction came in several times, then disappeared. Retraction papers were published in the Journal of Biological Chemistry by Lipmann and the Yale laboratory some 12 and 18 months respectively after the initial papers had appeared. However, most scientists in the field had already learned informally what had occurred as early as the day following our

88

Thomas Peter Bennett

afternoon meeting with Lipmann. The world of international biochemistry was indeed stunned, as Broad and Wade comment. Over the next few years, I learned that Lipmann at the time of our meeting had knowledge of which the rest of us were unaware. Melvin Simpson had already summoned Traction back to Yale to repeat critical experiments under constant surveillance. He could not repeat the key cytochrome c experiments. Dr. L, even when working under poltergeist-proof conditions, could not repeat Traction's work on the peptide glutathione. Here is Broad and Wade's account: Simpson remembers receiving a call from Lipmann after Traction's return to Yale: "I understand Traction has been down with you," Lipmann said. "Yes, we're having a little trouble in repeating experiments," Simpson replied cagily, not wishing to give too much away at that point. "You put your cards on the table and I'll put mine on the table," Lipmann suggested. Simpson agreed. "We can't repeat anything Traction has done." "Neither can we," Simpson replied. Simpson had people watch Traction in shifts as he repeated the experiments. This time, none of them worked... Simpson was desolated by what had happened. "The pain is all gone now," he says, "but it lasted a long time. I had to take a summer off from work, and I rebuilt a little sailboat I had. I couldn't go into the lab for about three months. It was so painful. Thomas was a confidence trickster. He'd do favors for everybody. If somebody wanted tickets to something, Traction would get them. If you needed to borrow a car, Traction would leave right on the spot. I don't know why he did it. Except that we felt he was really off his rocker. He was smart enough so that he didn't have to do it." Traction's undergraduate college in Massachusetts has no record of his ever having received a degree. Traction left research and to this day denies any wrongdoing... Broad and Wade see the Traction case as an example of what's wrong with today's research. To them, it is the tip of the iceberg. In contrast, my personal knowledge of the circumstances of this case leads me to conclude that unquestionably it demonstrates two things: the inherent strength and effectiveness of the scientists' code of honor, and the character and integrity of the practicing scientists involved. Let me cite three examples. First, Dr. L: The importance of professional integrity, strength of character and selfdiscipline, are apparent when we consider his actions in a series of circumstances culminating in events which might well have caused him considerable professional embarrassment at the onset of his promising scientific career. Dr. L could easily have abandoned his research project and shifted to another when he had difficulties with the original experiments. Instead, he persevered. At a later stage, the poltergeist represented a threat of potential physical harm to Dr. L himself; he discussed this with several of us and we often stayed at our benches late in the evening to keep him company. However, instead of halting the project, he moved his experimental work

Lipmann and "Not Strictly Biochemistry"

89

elsewhere and persevered. At the third stage, despite the possibility of "guilt by association" because of his involvement in this research, he again persevered, and coauthored with Lipmann one of the retraction papers. Second, Melvin Simpson. Simpson's integrity as a scientist, at a maturing stage in his career, led him to share information with Fritz Lipmann which would be professionally damaging to them both. To be open and candid in this instance, rather than to withhold information or deny any knowledge of the problems, placed Simpson, his laboratory, and his institution at jeopardy. Nonetheless he acknowledged his obligation to his colleagues and to society; first, Simpson shared his information with Lipmann, damaging though this was, and then he published his retraction. Third, Fritz Lipmann. Lipmann's integrity as a scientist led him not just to endure, but to invite, considerable professional embarrassment in the course of carrying out what he regarded as his duty - his obligation to science, to his colleagues, to society, to himself. As a Nobelist, he could probably have glossed over this, or invoked the principle of immunity from scrutiny. Instead he chose to deal with it in a vigorous and public way. For Lipmann to publish a retraction paper - for the first time in his illustrious career - was indeed an event which rocked the international world of biochemistry. Yet to Lipmann the possibility of not doing so was repugnant. His admonition to me ("think like a scientist!") in retrospect went beyond issues of logic or method to reflect on the totality of the situation, the honor code, and the values involved in scientific research. It was the scientific process itself - the process of constant exchange of ideas among researchers; of science as an additive process that builds on earlier work; of science as a process of continuous monitoring of research results and speedy disclosure of error or fraud - and the positive character values inherent in this process which assured that the Traction incident would be discovered and then guaranteed its public disclosure. I thus return to my original premises: that character is inextricably entwined in doing good science and integral to the very nature of the scientific process itself; that an implicit and effective honor code operates within science; that for Fritz Lipmann, and for all of us who knew him, "thinking like a scientist" embodied a life long intellectual and ethical commitment. And Traction: Was he ever in fact a scientist at all?

Lipmann and Subjects "Not Strictly Biochemistry" From the time I joined his lab in 1959 until our last meeting in 1985, Lipmann often startled me by his characteristic candor in encouraging me to explore my interests in teaching, in popular aspects of science and in the interface of science and the humanities. For me, as a student thinking only of science, Lipmann's ever increasing collections of science related to art, his humanistic interests and concerns, along with his encouragement of my nascent interest in those areas, were a great source of personal support. His enthusiastic discussion of topics that he termed "not strictly

90

Thomas Peter Bennett

biochemistry" put me at ease in my developing interests. During the years that followed I enjoyed his sage counsel, conversations and encouragement in areas of teaching, writing, art and the humanities that he continued to term "not strictly biochemistry". I would like to acknowledge my personal and professional gratitude to Lipmann for his warm, stern and expansive encouragement of my efforts "not strictly biochemistry" as I share some examples with his international students and colleagues.

Graphic Biochemistry and Graphic Words Lipmann's artistic and graphic translations of biochemical concepts were often included as integral parts of his papers (5,6,7). The most significant of these symbols was the squiggle ( ~ ), which he invented to denote energy carrying bonds and incorporated into the first major graphic concept in biochemistry: the metabolic dynamo. Published in his classic "Advances in Enzymology" article in 1941, it was a graphic that Freda Lipmann had helped him to develop and execute. By 1965 the metabolic dynamo had metamorphosed in the biochemical and textbook literature to be hardly recognizable as Lipmann's original graphic. For this reason and with his encouragement, Lipmann's classical diagram (Figure 1) was reintroduced to a new generation of international students in Modern Topics of Biochemistry (8), which I co-authored with my earlier mentor Earl Frieden. Lipmann's incisive summary of his Harvey Lecture (7) about biosynthetic mechanisms introduced our Chapter 10 (Figure 2). In my subsequent writing efforts (9,10,11) Lipmann offered his encouragement and gentle but decisive criticism: "I don't think I would say it that way"; "Maybe you should consider adding this to the story"; "This diagram could be made sexier!"

Figure 6 - 7 . The metabolic dynamo generates ~ P current. This is brushed off by adenylic acid, which likewise functions a s the wiring system, distributing the current. Creatine — P. when present, serves as Ρ accumulator. Classical diagram by Lipmann. 1941.

Figure l a

91

Lipmann and "Not Strictly Biochemistry"

Figura 6-7. La dinamo metabòlica genera corriente de —P. El á c i d o adentlico actúa c o m o colector y distribuidor de esta corriente. La creatina Ρ c u a n d o está presente actúa como acumulador de — P. Diagrama clásico de Liprnann. 1941.

Alimentalo

l e

Figura 6-7 - O dinamo metabòlico géra urna corrente de - φ Esta é captada por ácido adenilico, que funciona como o sistema de Tíos, distríbuindo a corrente. Fosfato de creatina, quando presente, funciona como acumulador de - Ρ Diagrama clàssico de Lipmann. 1941.

g l 6-7 I d

ftlodti

IÍ«*itl!«5SB5rti.

-Ρ

- Ρ * » » * * ο *·>*»< F Lipmann, 19J1

Figure I The metabolic generator - Lipmann's classical diagram (Ref. 8; (a) English, (b) Spanish, (c) Portuguese and (d) Japanese edition).

Thomas Peter Bennett

92

2a

"It seems that in the field of biosynthesis we have a rare example of progress leading to simplification." F. Lipmann, 1949

2b

"Parece como si en el campo de la bioslntesis tuviéramos un raro ejemplo de progreso que conduce a la simplificación". F. Lipmann, 1949.

2c

2d

F. Lipmann, 1949

Figure 2 Lipmann's summary of his Harvey Lecture (7).

Always hiding my poetry under a basket so as not to jeopardize my hopefully growing scientific reputation, I was encouraged by Lipmann to publish "Reductionism: Seven Lessons" (12), which in its fourth canto, attempts to encapsulate bioenergetics and Lipmann's role (Figure 3). 4. ATP T h e solar furnace; Life-giving energy. Light converted; Lipmann bonds formed. Metabolic dynamos; ~Phosphate generated. Water splits ATP; Life is moved. Figure 3 Fourth canto in "Reductionism: Seven Lessons" (12).

Lipmann's words to me — "Think like a scientist" — ring in my memory. They will always evoke his overwhelming intellectual powers and his absolute integrity as a scientist. As well, however, Lipmann's "not strictly biochemistry" attitudes and insights have been an enduring resource for me and I hope they illuminate yet other facets of Fritz Lipmann for you.

Lipmann and "Not Strictly Biochemistry"

93

References 1. Broad, W. and N. Wade. 1982. In: Betrayers of the Truth: Fraud and Deceit in the Halls of Science. Simon and Schuster, New York, pp. 73-76, 228-229. 2. Bennett, T.P. 1983. The Use and Abuse of Science. Lecture presented at the Annual Meeting of the Franklin Inn Club. Philadelphia. 3. Bennett, T.P. 1983. Science and Scientists: The Character Education Connection. Wayne, Pa: Valley Forge Military Academy. 4. Bennett, T.P. 1985. New Days and Events, and the Character of Scientists. Phi Beta Kappa Lecture delivered at Haverford College, Haverford, PA. 5. Lipmann, F. 1941. Advances in Enzymology 1, 99. (Reprinted by Bobbs-Merrill, Reprint Series in the Life Sciences, No. B-184. 6. Lipmann, F. 1954. In: The Mechanism of Enzyme Action (W. D. McElroy and B. Glass, eds.) Hopkins, Baltimore, p. 599. 7. Lipmann, F. 1948-1949. The Harvey Lectures, Series XLIV, Copyright 1950. Courtesy of Charles C. Thomas, Publisher, Springfield, 111. 8. Bennett, T.P. and E. Frieden. 1966. Modern Topics in Biochemistry: Structure and Function of Biological Molecules. Macmillan. 9th Printing, 1970. Spanish Edition, 1967. Japanese Edition, 1969. Portuguese Edition, 1969. 9. Bennett, T.P. 1968. Graphic Biochemistry. Volume 1: The Chemistry of Biological Molecules. Volume 2: The Metabolism of Biological Molecules. Macmillan. 2nd Printing, 1969. Japanese Edition, 1970; 2nd Edition, 1971. 10. Bennett, T.P. 1969. Elements of Protein Synthesis. Freeman, San Francisco. 40 Page TextPamphlet. 11. Armstrong, F.D. and T.P. Bennett. 1979. Biochemistry. Oxford University Press, New York. 12. Bennett, T.P. 1973. "Reductionism: 7 Lessons." In: Perspectives in Biology and Medicine. Summer.

Communication in Metabolic Control Intuition and Method in Biochemistry: Four Years each with Krebs and Lipmann

Wieland G ever s In his characteristically understated and whimsical autobiography Wanderings of a Biochemist, Fritz Lipmann writes of the demanding oral examinations required of him for the Ph.D. degree: "What has remained with me from all these strenuous efforts is not the facts I had to learn but the contacts with impressive personalities" (1). There can be no doubt that a total of eight years spent in Oxford and New York, four with Hans A. Krebs and the other four with Lipmann himself, qualifies me to express the same sentiment with feeling and gratitude. The situation is compounded by the strong contrasts afforded by these two "tap-roots of modern biochemistry", contrasts which close contact over the two generous periods permitted me to savour in a unique manner. Modern biochemistry students would probably class both Krebs and Lipmann as having been "metabolic biochemists". This would be a facile view, since it is really the era of biochemistry in which these two played so large a part, that was "metabolic": I suspect that a hypothetically reincarnated Krebs would again be drawn to a career of preoccupation with intermediary metabolism in 1987, while Lipmann would take on some problem of molecular technology or engineering: he says in his book that his older brother "was early interested in the theatre, and he composed poems. I rather played with toy trains and building blocks" (2). The child indeed is father of the man. Krebs struck me throughout as a highly civilized and cultured man, interested in the elegant use of language (3). He was dedicated to Method: the Method of analysing the problem to be tackled, the Method of gathering the necessary data, the Method of publishing and consolidating the resulting picture. Methodically he sought out apparently unconnected pieces of information and generated the momentum of new projects, quick to see the potential implications of results reported to him, and constantly calculating out ratios and equivalences like a kind of biochemical land surveyor. He put intense pressure on students or associates who were lucky (or unlucky) enough to be discovering interesting things. He took pains to explain his analysis and conclusions and was a fine teacher in the particular as well as the general. He was an overpoweringly intelligent man whose method of doing things worked very well. On joining Lipmann in New York during the last few years of his full Membership and Professorship in the Rockefeller University, I was taken into a different world. Here was a shy and quiet, almost other-worldly man whose daily activities were apparently unplanned and certainly unpredictable. Long, rambling conversations, punctuated by silences sometimes awkward, were held at the workbench with total

96

Wieland Gevers

disregard for the time of day or the need to carry out an experimental manipulation: here was a dreamy preoccupation with the essence and the interior logic of scientific problems, especially as they related to "building blocks" and what binds to what and why. Lipmann's striking ability to pick on the crucial issues in every laboratory seminar and to ask the first few questions, was a source both of anxiety and satisfaction to those who benefited from his probing analysis of the presented data. He had a feeling for, rather than detailed knowledge of, organic and physical chemistry and nothing delighted him as much as when his intuitions were confirmed by experimental studies. Lipmann was also a man of wide culture, but he had the soul of an artist who had strayed into the molecular world of biochemistry. His interest in every facet of this rapidly expanding subject was genuine and deep, his powers of concentration remarkable. Toughness of intellect lurked behind the gentle exterior, and Lipmann was quite capable of leading a research group in a hot competitive race to the solution of a problem. My years with Krebs had inculcated in me a deep interest in metabolic regulation, the field which briefly in the mid 1960's shared centre stage with molecular biology, only to be unceremoniously edged off by the DNA restriction enzymes in the next decade. The studies on the molecular biotechnology required for the biosynthesis of peptide antibiotics, initiated under Lipmann when I started in New York, and worked out with Horst Kleinkauf and Robert Roskoski over the period 1966-1970, could not have been further from the themes of metabolic control (4). Yet I continued to follow the literature on regulation and taught this subject to what by now must be almost an entire generation of South African biochemistry and medical students. The theoretical treatment by Newsholme and Crabtree, drawn from a long involvement with enzymes and a Krebsian approach to metabolism (5), came into conflict with a different formalism based on a curious marriage of classical genetics and mathematical biology (associated with the names of Kaczer, Burns and Porteous (6), Westerhoff, Groen and Wanders (7) and Heinrich and Rapaport (8). I have previously analysed the situation from my own independent vantage point and have found myself responding in an intuitive manner which I probably can, with some humility, ascribe to the impressionable years I spent with Lipmann (9). Metabolic regulation is primarily a problem of communication between very many molecules in systems. Lipmann probably shrank from the approaches which are conventional in this field, because one encounters so often in it a gross simplification of cellular structure and the subcellular functional division of labour required to make the mathematics work. The brick-building molecular biotechnology of which he was so fond is often ignored or wished away. Channeling of substrates between enzymes (10) would have appealed to Lipmann, but not the notion of "elasticity coefficients" involving in practice, at least, direct extrapolations of in vitro kinetic data to in situ metabolism (11). In any case, these are good reasons for an intuitive search in metabolism for what is probably right, based on experience of systems in a variety of ways, as opposed to a hard-nosed, specialist or "tunnel-vision" approach to the analysis of such problems.

Communication in Metabolic Control

97

Steady-state-Metabolic Systems Generally speaking, it is easier to talk about a metabolic system if one assumes that it is in a steady state for a finite time interval. Such a steady-state metabolic system in an intact animal starts with a series of inputs which generate a steady flow of material. The inputs are either exogenous (e.g. intravenous nutriment) or, more typically, they represent release of material from pools which are either very large (high concentrations or large poolvolumes) or in a non-aqueous physical phase. A common feature of many inputs in steady-state metabolism is the fact that a constant flow of material is initiated because the steady depletion of input material does not slow down the flow, at least for a finite time (5). This condition can be satisfied by "full saturation with substrate" of an input enzyme (zero-order kinetics) or by partial saturation of an enzyme in a large volume of fluid. It also applies to lipids being "mobilized" from oily vacuoles or droplets, or small molecules which are being removed by enzymes from insoluble granules or precipitates (glycogen) or from insoluble or soluble proteins making up a "protein reserve". (Not all steady inputs conform to this pattern - see below). In each of the above cases, the rate of the process representing the input is the input flux rate: material is supplied at this rate to "downstream" steps which transmit the flux and usually divide it up into branches so that a number of outputs are eventually formed. Outputs are defined either as materials which disappear into "sinks" (large pools of substances in a different physical phase, such as oil droplets, granules, structural proteins or expired gases), or as substances which are excreted from the system or which accumulate in the system without causing changes in any input fluxes. The combined low chemical potential of the output materials in the metabolic system is, of course, what provides the negative free-energy change permitting the flux from the inputs (combined high chemical potential) to the outputs to occur. It is an important feature of metabolic systems that very few individual outputs are ever derived from a single input; this is a consequence of the highly branched and circuitous nature of metabolic flow maps. In fact, most outputs represent material derived from more than one input. The many steps that link inputs and outputs ("downstream steps") actually encômpass the activities of virtually all the enzymes, permeases and "factors" which occur either inside cells or in the extracellular compartment, including the plasma with its special property of moving substances rapidly around the body in the circulation. Although exceptions may occur (see below), "downstream" steps generally respond passively to the flow of material derived from inputs; in the sense of enzymology, they are all "under-saturated" and thus have the property of "settling down" when presented with a particular rate of supply of their pathway substrate(s): If the enzyme of permease has a high capacity (e. g. there is a high concentration of catalyst), the incoming substrate(s) will be used up, slowing down the step as depletion occurs, until the step has "settled down" to the transmission of the flux and the maintenance of a particular concentration of pathway-substrate(s). In a very general way, relatively higher steady-state concentrations of intermediates will be found before "slower" steps, and lower concentrations before "faster" steps.

98

Wieland Gevers

The pattern of "downstream" pathways may be organized into simple or complex cycles, convergent loops, or nonconvergent single or multiple branches: This is what creates the multiple and shared outputs discussed previously. Clearly, the "competition" for fluxes which occurs at all branch points is critically important for the distribution of the steady-state flux from a given input to all the outputs derived from it. The relative activities of the competing catalysts and their "downstream" enzymes, i.e. the relative abilities of the divergent subflux pathways to move material from the shared substrate in their particular directions, are the key to this competition. "Channeling" of substrates between enzymes in multi-enzyme systems of various kinds (10) provides a real difficulty in understanding the kinetic basis of flux through a sequence of reactions. Rough equivalence between the concentrations of substrates and enzymes, and the notion, for example, of one-to-one-to-one complexes between two enzymes sharing a common substrate/product (in a solution largely devoid of free substrates), raises the possibility that "downstream" reactions may depend on the concentrations of complexes and not of the substrates themselves. In principle, the difficulty can be overcome by simplifying the pathway to one where complexes, however multimolecular, are considered to be "intermediates" in the same sense as freely diffusible substances. Srere has summarized the advantages of multienzyme complexes in terms of cell metabolism, emphasizing the evolutionary development of complementary surfaces required for enzyme-enzyme interactions, some of which may even involve allosteric phenomena (12). Such notions impart an entirely different meaning to the notion of branched metabolic pathways. Setting aside these complexities (as Krebs would do) in favour of diffusible metabolites, metabolic systems which are in a steady state can be perceived as consisting, on the one hand, of inputs from various sources (where the concentrations of input materials either steadily fall or remain constant, but where the actual rate of input is constant), and, on the other hand, of outputs into various "sinks" (where concentrations may or may not rise but this does not affect the rate of any input). The rest of the system consists of a myriad of steps (mediated or unmediated transport events from one part of the body or across membranes, or covalent transformations catalysed by enzymes which transmit and distribute the input fluxes to the various output fluxes). The concentrations of all intermediates are constant and their individual levels are a general inverse function of the capacities of the step catalysing their removal in each case. A number of important metabolic realities need to be considered at this point: (1) Enzymes, more often than not, have two or more substrates/products. This fact emphasizes the need to consider fluxes as representing several kinds of material transfer which occur simultaneously during metabolism, e. g. carbon skeletons, pairs of hydrogen atoms and phosphate groups going their ways independently (5). Another consideration is that the actual levels of metabolic intermediates in steady states are the "settling down" end-points of more complicated situations than would be the case for single-substrate enzymes.

Communication in Metabolic Control

99

(2) Many enzymes are sufficiently active (high capacities) to catalyse something approaching an equilibrium situation in respect to all their reactants. Such enzymes effectively create a branch point in metabolism by competing with succeeding enzymes for their products. As a result, they deprive their successors of the facility of a simple "settling down" to achieve flux transmission at a particular substrate concentration adjusted to the capacities of these steps, as previously discussed. Ultimately, the enzymes concerned have to "settle down" together and so the implications of the near-to-equilibrium status of many enzymes are mainly centred on the determination of the steady-state concentrations of metabolites and any bearing that this may have on the control of flux distribution if branch points are involved. The enzymes that are "far-fromequilibrium" are important in giving directionality to metabolic sequences. As a crude generalization, they tend to be enzymes that have a lower capacity (enzyme concentration times turnover number) than others, so that they tend to be somewhat more saturated with their substrates (while usually not being fully saturated). They are also often subject to the influences of multiple allosteric modifiers and often undergo inter-conversion cycles between less active and more active forms. These features are extremely important in defining the "set" of an entire metabolic network in a given steady state, and again when there is a shift to a new steady state in a system. For example, an external influence which causes the doubling of the activity of an input step may cause a "downstream" step which is far-from-equilibrium and fairly highly saturated with substrate, actually to become fully saturated, imposing a ceiling on flux through the particular branch of the "downstream" system. Alternatively, it is common to find that the same stimulus that alters an input step is associated with activation of one or more "downstream" step(s) by positive allosteric modifier(s) or by conversion to more active forms, in order to transmit the new, higher flux through the relevant part of the system without becoming "bottle-necked" at that step. (3) There are known cases where the enzymes catalysing input steps are influenced by the concentrations of "downstream" intermediates, for example, by allosteric feedback interactions. When this happens, the input flux rate, which in any case for a given steady state has a set of determinants which are "external" to the system (e.g. the effective concentration of the active input enzyme protein), acquires also an "internal" determinant which plays a "settling down" role in the fixing of the steady-state input flux. It is in this kind of situation that "steadiness" in an input flux is imposed by a regulatory loop rather than by substrate independence of an input step. (4) It must thus not be thought that "regulation" or "flux-generation" in a metabolic system is exerted only by input steps. There are many "downstream" steps within the system which determine the distribution of input fluxes in respect of flow to the output fluxes, apart from the steps which communicate with input steps. Most of the individual steps are in any case the principal determinants of the individual metabolite concentrations in the system: such metabolite concentrations in themselves are often crucially important for the physiological function of cells, tissues and organs. In a mechanistic sense, the

Wieland Gevers

100

through-put fluxes are actually mostly irrelevant to the functioning system; they are just needed to keep it going.

Changes from a Steady State The above description of a steady state as consisting essentially of a set of steady fluxes through a very complex network of reaction steps, with every intermediate in the system maintained at a steady concentration, implicitly recognizes a potential role for each component in determining any or all the fluxes and concentrations that prevail. It is, nevertheless, quite obvious that the importance of an individual component for any given flux or intermediate concentration cannot be assessed in a single steady state. A shift from the steady state is needed to enable one to approach the quantitative problem of the regulatory role played by each component in "determining" a given flux or concentration of an intermediate in the original steady state. The behaviour of the system as it changes brings out the hidden "power structure" behind its de facto operation; a new steady state may or may not have the same "power structure", depending on the extent to which the properties of steps in the system are adapted to the new fluxes and metabolite concentrations. As a first approximation, one can consider the total control of a defined flux or a particular metabolite concentration as being equal to unity: the various contributors to that control each have a fractional part of the control which may vary from near zero to near unity, while the sum of all the fractions is expected to be equal to one. This concept is more subtle than might at first appear, since some components may have negative controlling influences (for example, a defined flux diminishes when the activity of the component is increased), while others have a positive controlling influence in a more obvious way. The summation to unity of all control fractions thus includes negative and positive fractions: Branches which do not lead to reconvergence of pathways are potent sources of negative controlling influences. (Some authors believe that such non-reconvergent pathways are rare (6), but they are probably more common than reconvergent pathways). There would appear to be no problems in accepting the following basic tenets of metabolic regulation: (1) Enzymes and carrier molecules (parameters) are embedded, in cells and organisms, in a functionally inter-dependent network, which is a system and therefore has systemic properties (variables) such as various defined fluxes and concentration profiles which can exist in a set of steady-states. (2) Because of the complexity of the pathways followed by chemicals through the metabolic network, the control of defined fluxes and concentrations (variables) is shared between different components (parameters) in such a manner that some make large and others small or virtually insignificant contributions to a particular variable, in a particular steady state: The situation can also be different or the same when the system shifts to a new steady state. The behaviour of a system in a steady state is thus determined by the way in which control is distributed comprehensively amongst all the components (parameters).

Communication in Metabolic Control

101

(3) The consequences of a change in one or more parameters of a system, which is the usual cause of a shift to a new steady-state, are related to the original control situation only if the changes are infinitesimally small. A step change(s) is much more complicated because the responses of all the other parameters (and the resulting changes in the variables) cannot usually be extrapolated from the steady-state which existed before the change. Recognition of the broad generality of the above concepts has led three prominent groups to develop a coherent system of metabolic control analysis (6-8). This approach restricts itself to infinitesimally small modulations of single steps in steady-state systems: It distinguishes clearly between the parameters of such a system and the different variables dependent on them. The formal nomenclature, proposed by a large cohort of workers in this field (13), classes coefficients into those which are global·. Cp (Control coefficient) =

/

(where V represents a variable such as a flux or a metabolite concentration and Ρ denotes a parameter such as an enzyme activity or an input flux) and those which are local·. ôv ÔS evs (Elasticity coefficient) = — / — (where ν represents the actual rate of a "step" in the whole system and S is a metabolite substrate, product or effector, that interacts with the catalyst of the particular "step"). The approach of metabolic control analysis (6) is really nothing more than an attempt to provide a convenient and experimentally observable measure of the "importance" of individual steps in metabolic systems as a whole. The simplest (and ideal) way to determine a "flux control coefficient" (input or output flux) in respect of a metabolic step is to plot the effects of different concentrations of a specific irreversible inhibitor (or activator) on the particular flux in the system under steadystate conditions, and then to determine the initial slope of the inhibition or activation curve. The coefficient can be as high as 1.0 when the step is completely controlling or it can be much lower, tending to zero; it can even be negative when the tested step is part of a branch which permanently drains flux-material away from the particular output being measured. It is algebraically correct that the sum of the control coefficients of all steps in a metabolic system in respect of a particular flux is unity ("Summation property") (6). This means in effect that the enzymes of the system compete for control of the flux. Since it is very difficult in most systems to determine all the relevant coefficients, and since many of them are negative, the theorem probably has less application in practice than has been claimed (6). Attempts to "shortcut" the direct experimental determination of flux control coefficients by making use of a further theorem, the "Connectivity property", are very ingenious but, with the exception of modelled systems, are dependent on assumptions in which some of the rigour of the overall quantitative approach is sacrificed (11).

102

Wieland Gevers

Shared Control in Metabolism The new paradigm that has emerged from metabolic control analysis is that control of metabolic fluxes and the concentrations of intermediates is not vested uniquely in so-called "rate-limiting" or "pacemaker steps". Kaczer and Burns have gone so far as to postulate that no single enzyme can have more than a token importance in controlling a flux since the sum of very many flux control coefficients can only be one (14). This idea has been incorporated into a purported molecular explanation of Mendelian recessiveness in heterozygotes: the removal of 50% of the total capacity of any enzyme causes almost no diminution of flux through the pathway in which the enzyme is embedded (14). What was once thought to be a "safety factor" is seen to be nothing more than the natural consequence of shared control between many enzymes. A number of considerations apply to the notion of communal control of metabolism. Firstly, the analysis already provided above of input and output divergences in steady-state metabolic systems, clearly predicts the existence of multiple control points of significance in respect of most defined fluxes. After all, even a single nonconvergent branch in a pathway already provides a minimum requirement of three steps that are likely or certain to exert significant control on the flux from an input to the minimum of two outputs. Where an intermediate feeds back to an earlier input step, two further elements, namely those forming and removing the effector, are given further shares of the sum of instantaneous control in the steady state. In complex systems in which several fluxes occur independently, the stage is set for the kinds of multiple controls, shifting from one steady state to another, which have been described for various processes in respiring and phosphorylating mitochondria (15). In the absence of network arrangements that increase the multiplicity of control steps, there may be much less sharing, perhaps giving a true pacemaker function to single enzymes or steps. The example has been given earlier of an input flux catalysed by an enzyme insensitive to a change in the concentration of its substrate(s) and to that of any downstream metabolite: Such an enzyme has a control coefficient of 1.00 for the flux as defined and is therefore wholly controlling in this sense (a flux-generating enzyme or step (16)). In one very complex system, gluconeogenesis from 3-carbon precursors in rat hepatocytes, provided with glucagon, the single step catalysed by pyruvate carboxylase in the direction of flux has an apparent flux control coefficient of no less than 0.89, hardly trivial for one enzyme amongst the 50 or so that have a bearing more or less directly on the pathway in question (11). The frequency of phenotypic recessiveness in heterozygotes afflicted by the silencing of autosomal structural genes is really only evidence for the low flux or metabolite control coefficients of many enzymes in the range steady-state flux rates that usually occur in the cells of such individuals. Other enzymes or functional proteins may indeed have higher coefficients even under physiological conditions, and the heterozygous state may then be dominant or co-dominant: the LDL receptor is a good example (17).

Communication in Metabolic Control

103

The Level of Sensitivity One of the frustrating complexities of metabolic control analysis is the sheer difficulty of predicting the extent of change in actual catalysis through a single step that will result from a change in the conditions at that step. The global control coefficient, Cp, singly evokes an infinitesimally small change in the catalysis at the step catalysed by the parameter Ρ as having a particular measurable effect on the flux rate (V) through the pathway. The coefficient will be higher in cases where one or more "downstream" steps shows ultrasensitivity (see below), than in cases where such steps show normal or subnormal sensitivity, for reasons of their own. Koshland has analysed several distinctive causes of ultrasensitivity in metabolic systems (18). Normal sensitivity is defined as that afforded by the rectangular hyperbola of Michaelis-Menten-Henri kinetics; clearly, this is greatest in the first order domain of the curve which applies when enzymes are highly undersaturated: above this degree of saturation the sensitivity dwindles to zero. Ultrasensitivity is defined as being greater than the maximum afforded by hyperbolic sensitivity and a well-known cause is homotropic or heterotropic cooperativity at maximum slope. Other causes are "multi-step" systems where a single effector simultaneously influences several steps in a sequence, and "zero order" activity of one or more converter enzymes acting on a metabolic step. A very interesting and significant cause of high sensitivity is "branch point ultrasensitivity" where particular kinetic properties of the enzymes competing for a changing input flux are associated with dramatic alterations in the output fluxes (19).

Conclusion The complexity of cellular chemistry should not tempt us into a resurgence of vitalism; it should rather be a challenge to enhance our understanding of "how it works" (Krebs) when the "building blocks are thrown together" (Lipmann), using intuition and method in a fruitful combination.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

Lipmann, F. 1971. Wanderings of a Biochemist. Wiley-Interscience, N e w York., p. 9. Lipmann, F. 1984. Ann. Rev. Biochem. 53, 1 - 3 3 . Krebs, H . A . 1981. Reminiscences and Reflections. Clarendon Press, Oxford. Lipmann, F., W. Gevers, H. Kleinkauf and R. Roskoski. 1971. Adv. in Enzymology 35, 1 - 3 4 . Newsholme, E. A. and B. Crabtree. 1979. J. Molec. Cell. Cardiol. 11, 8 3 9 - 8 5 6 . Kaczer, H. and J.W. Porteous. 1987. Trends in Biochem. Sci. 12, 5 - 1 4 . Westerhof^ H.V., A . K . Groen and R . J . A . Wanders. 1984. Bioscience Reports 4, 1 - 2 2 . Heinrich, R. and T. A. Rapaport. 1974. Eur. J. Biochem. 42, 1 0 7 - 1 2 0 . Gevers, W. 1987. S. Afr. Med. J. 72, 7 8 3 - 7 8 7 . Srivastava, D . K . and S.A. Bernhard. 1986. Science 234, 1081-1086. Groen, Α . Κ . , C.W.T. van Roermond, R.C. Vervoorn and J.M. Tager. 1986. Biochem. J. 237, 379-389. 12. Srere, P.A. 1987. Ann. Rev. Biochem. 56, 8 9 - 1 2 4 . 1 1 . 13. Burns, J. Α., A. Cornish-Bowden, A . K . Groen et al. 1985. Trends in Biochem. Sci. 10, 16.

104

Wieland Gevers

14. Kaczer, H. and J.A. Burns. 1981. Genetics 97, 639-666. 15. Groen, Α. Κ., R. J. Α. Wanders, Η. V. Westerhoff, R. van der Meer and J. M. Tager. 1982. J. Biol. Chem. 257, 2754-2757. 16. Crabtree, B. and E.A. Newsholme. 1987. Trends in Biochem. Sci. 12, 4-12. 17. Brown, M.S. and J.L. Goldstein. 1977. Ann. Rev. Biochem. 46, 897-930. 18. Koshland, D.E. 1987. Trends in Biochem. Sci. 12, 225-229. 19. LaPorte, D.C., K. Walsh and D.E. Koshland. 1984. J. Biol. Chem. 289, 14068-14075.

Fritz Lipmann: Squiggle to Protein Sulfation Barbara Petrack

"Fritz Lipmann." I was dismayed to learn that a young (otherwise knowledgeable) biochemist did not know the name. Surprised, I asked, "Have you never heard of ATP as the mediator of metabolic energy?" "Of course," he said, "I'm not ignorant." "Well, the concept that metabolic energy is captured in the squiggle Ρ bonds of ATP, was not written in stone, but emerged from Dr. Lipmann's creative mind." The young biochemist was impressed. Clearly, he had never considered the origin of such basic information. I went on, "Did you ever hear of coenzyme A and its role in fatty acid metabolism? Fritz Lipmann received the Nobel Prize in Medicine in 1953 for the discovery of coenzyme A. How about the discovery of PAPS as activated sulfate?" "Lipmann discovered that, too?" "Yes, and the discovery that peptide chain elongation involves a GTP-dependent translocation on the ribosome. And PKA activation involves liberation of the catalytic subunit by cAMP binding to the regulatory subunit and removing it. And the discovery of carbamyl phosphate as the intermediate in citrulline and pyrimidine biosynthesis and" By now the student, as impressed with Dr. Lipmann as I, was eager to borrow my copy of The Wanderings (1) and his 1984 autobiography (2) and took the reference for the Advances in Enzymology paper (3) so that he could read, first hand, about the discovery of ~ P. This experience with minor variations, has often been repeated during past years. Despite not always knowing the name Fritz Lipmann, young scientists have incorporated his vast contributions into the core of basic biochemistry, thereby perpetuating the essence of the man. Fritz Lipmann was my hero, not only for his scientific brilliance, but also for his kindness and sensitivity to the needs and wishes of others. I am forever grateful that he accepted me as a post-doc in his laboratory although he knew that I was pregnant. In an era when working mothers were generally viewed with strong disapproval - demonstrating empathy and social conscience ahead of his time - he understood that my scientific and maternal needs could both be satisfied. Dr. Lipmann had an unusual ability to derive great pleasure from life. His enthusiasm was not limited to science, but enveloped creative interests in music and fine arts. I can still see the joy radiating from his face as he reveled in the music of the Tokyo String Quartet.

106

Barbara Petrack

At the last Enzyme Club meeting attended by Dr. Lipmann, he and I spoke of sulfated proteins, a subject that interested him in his last years, evolving from his earlier discoveries of activated sulfate. I, too, have recently been interested in a sulfated peptide, cholecystokinin. Protein sulfation, currently under study in many laboratories, originates from Fritz Lipmann's Squiggle; his last scientific publications add new possibilities to elucidiating the physiological significance of protein sulfation. This work began in the 1950s, when Robbins and Lipmann demonstrated (4, 5) that biological sulfations involve the ATP-dependent generation of an activated sulfate, which is then enzymatically transferred to various acceptor molecules. Every biochemistry text describes the 2-step activation of sulfate: ATP + SO|" -> Adenos¡ne-5'-phosphosulfate + PP (APS) APS + ATP

-> 3-Phosphoadenosine-5'-phosphosulfate + ADP (PAPS)

PAPS is the universal sulfate donor in the esterification of phenolic substituents on a wide variety of endogenous and exogenous cellular components. During early studies, sulfate transfer was demonstrated for the syntheses of chondroitin sulfate and sulfated cerebrosides. Also, many drugs and the catecholamine neurotransmitters are excreted following sulfoconjugation by PAPS (6). Sulfation usually results either in activation or inactivation of acceptor molecules. Specificity is derived from the sulfotransferase enzyme as well as the acceptor molecule used as substrate. Some of the enzymes involved have not yet been fully characterized. More recently, the transfer of sulfate from PAPS to proteins and peptides has been reported.

Protein and Peptide Tyrosine Sulfation Huttner, using a [ 35 S]sulfate radioactive-labeling technique, demonstrated that many proteins are sulfated via post-translational modification, always on Tyr residues (7). Tyrosyl-protein sulfotransferase catalyzes sulfate transfer from PAPS to the hydroxyl moiety of Tyr. Sulfated proteins appear either to be secretory or membrane proteins that are synthesized in the rough endoplasmic reticulum and sulfated in the Golgi complex (8) using a transmembrane carrier system for PAPS (9). Sulfation might mark a protein for sorting as a secretory protein. The presence of acidic amino acids near a Tyr, with few hydrophobic or basic amino acids, seems to identify the specific Tyr that is recognized by the enzyme (10). Cholecystokinin (CCK) is a sulfated peptide that functions as a gut hormone, stimulating enzyme release from the pancreas and gallbladder contractions. These actions of CCK are dependent on sulfation of the Tyr residue in position 27 on the 33-amino acid peptide (11), consistent with its role as a secretory protein. High concentrations of CCK are also present in the brain, primarily as the sulfated carboxy-terminal octapeptide (CCK-8); it appears to function as a neurotransmitter or neuromodulator (11). In contrast to peripheral CCK receptors, most neuronal

Fritz Lipmann: Squiggle to Protein Sulfation

107

CCK receptors bind sulfated and non sulfated CCK with equal affinities (12). Further, some behavioral effects of CCK are specific for sulfated-CCK, whereas other effects are seen with either peptide (13). Sulfation might provide a mechanism for obtaining more than one phenotype from a single translation product (8). A microsomal sulfotransferase from rat brain sulfates Tyr of CCK-8, using [ 3 5 S]PAPS (14). The pH optimum of 5.8 enables the enzyme to operate in secretory vesicles. CCK sulfotransferase does not catalyze sulfation of either tyrosine methylester or p-nitrophenol, distinguishing it from the soluble phenolsulfotransferase. The enzyme favors substrates with acidic amino acids surrounding Tyr, as does tyrosylprotein sulfotransferase. It does not catalyze sulfate transfer to Tyr residues in either met-enkephalin or angiotensin. Further studies are needed to determine if these two sulfotransferase enzymes are the same and also to identify the CCK fragment for post-translational sulfation during processing of the preproCCK gene product. The amino acid sequence of prepro-CCK derived from cDNA sequence data (15) shows that CCK-8 is linked, via Gly-Arg-Arg, to a 9-amino acid peptide (called CAP-9) containing two Tyr residues. Processing at Arg-Arg (followed by amidation of the carboxy terminal Phe by alpha-amidating enzyme) simultaneously liberates CCK-8 and CAP-9. In addition to sulfation of Tyr in CCK8, both Tyr residues in CAP-9 isolated from pig brain are adjacent to acidic amino acids and are fully sulfated (16). Sulfated-CAP-9 might yet have a physiological function.

Sulfation in Malignancy In 1984 Liu and Lipmann discovered that sulfation was markedly reduced in transformed cells (17,18). [ 35 S]sulfate uptake, as well as sulfation, also was decreased and these effects were most pronounced in fibroblasts that were permanently transformed with Rous sarcoma virus (RSV). The decrease in the level of Tyrsulfated compounds was apparently due to their secretion from the transformed cells. It had previously been reported that the concentration of fibronectin, a cell surface protein, also was reduced in transformed cells (19). Liu and Lipmann demonstrated that their transformed cells secreted fibronectin, from which Tyrsulfate was liberated by pronase proteolysis (20). Membrane-bound fibronectin functions in the adhesion of cells to some surfaces, allowing contact inhibition, which stops further growth (21). Decreased fibronectin in the cell, could diminish contact inhibition, enabling continued cell growth and proliferation. In addition to sulfation on Tyr, the carbohydrate moieties of some glycoproteins also are sulfated, e.g., N-CAM, neural cell-adhesion molecule and other glycoproteins that appear to be involved in development.

Comparison of Sulfation and Phosphorylation Since creating the squiggle P, Dr. Lipmann contributed most importantly to elucidating the role of protein phosphorylation in controlling metabolic processes.

108

Barbara Petrack

The discovery in 1979 of the induction of tyrosine-specific protein kinases as gene products of virus-transformed cells (22) again captured Dr. Lipmann's attention. In 1983, Fukami and Lipmann reported that protein bound, but not free, Tyrphosphate bonds are energy-rich (23). Furthermore, viral transformation by inducing Tyr-specific protein kinase, increases protein Tyr phosphorylation in transformed cells. (24). These opposite effects of malignancy on tyrosylprotein phosphorylation and sulfation, combined with the unique selection of Tyr in mediating these effects suggests that tyrosylprotein phosphorylation and sulfation could possibly be involved in the development of malignancy. Interestingly, although the phosphorylation of Tyr in proteins is reversible by specific phosphatases, current evidence suggests that sulfation of protein Tyr residues is irreversible in vivo·, a comparable sulfatase has not yet been found. Clearly, Dr. Lipmann's creative thinking remained to his last days. He lived a rich life and enriched the lives around him. It was a joy to know him.

References 1. Lipmann, F. 1971. Wanderings of a Biochemist. Wiley, New York. 2. Lipmann, F. 1984. Annu. Rev. Biochem. 53, 1. 3. Lipmann, F. 1941. Adv. Enzymol. 1, 99. 4. Robbins, P.W. and F. Lipmann. 1957. J. Biol. Chem. 229, 837. 5. Lipmann, F. 1958. Science. 128, 575. 6. Roth, J. A. 1986. Trends Pharmaceut. Sci. 7, 404. 7. Huttner, W.B. 1984. Methods Enzymol. 107, 200. 8. Huttner, W.B. 1987. Trends Biol. Sci. 12, 361. 9. Schwarz, J.K., J.M. Capaso and C.B. Hirschberg. 1984. J. Biol. Chem. 259, 3554. 10. Lee, R.W.H. and W.B. Huttner, 1985. Proc. Natl. Acad. Sci. USA 82, 6143. 11. Dockray, G.J. 1982. Brit. Med. Bull. 38, 253. 12. Wennogle L.P., D.J. Steel and B. Petrack. 1985. Life Sci. 36, 1485. 13. Crawley, J.N., J. A. Stivers, L.K. Blumstein and S.M. Paul 1985. J. Neurosci. 5, 1982. 14. Vargas, F., O. Frérot, M. Dan Tung Tuong and J.C. Schwartz. 1985. Biochemistry. 24, 5938. 15. R.J. Deschenes, R.S. Haun, C.L. Funckes and J.E. Dixon. 1985. J. Biol. Chem. 260, 1280. 16. Eng J., U. Gubler, J.-P. Raufman, M. Chang, J. D. Hulmes, Y.-C. E. Pan and R. S. Yalow. 1986. Proc. Natl. Acad. Sci. USA 83, 2832. 17 Lipmann F. and M.-C. Liu. 1985. In: Cellular Regulation and Malignant Growth (S. Ebashi, ed.). Japan Sci. Soc. Press/Springer, Berlin, p. 393. 18. Liu, M.-C. and F. Lipmann. 1984. Proc. Natl. Acad. Sci. USA 81, 3695. 19. Vaheri A. and D.F. Mosher. 1978. Biochim. Biophys. Acta. 516, 1. 20. Liu, M.-C. and F. Lipmann. 1985. Proc. Natl. Acad. Sci. USA 82, 34. 21. Lipmann, F. and M.C. Liu. 1985. Current Topics in Cell. Regul. 27, 193. 22. Erikson, R. L., M. S. Collett, E. Erikson and A. E. Pirchio. 1979. Proc. Natl. Acad. Sci. USA 76, 6260.

23. Fukami, Y. and F. Lipmann. 1983. Proc. Natl. Acad. Sci. USA 80, 1272. 24. Hunter, T. and B.M. Sefton. Proc. Natl. Acad. Sci. USA 77, 1311.

Fritz Lipmann (1899-1986), Honorary Member of the Leopoldina Academy B. Parthier

The Fritz Lipmann Memorial Symposium is a welcome occasion to express our reverence for Professor Lipmann, the outstanding biochemist and molecular biologist, who was associated with the Deutsche Akademie der Naturforscher Leopoldina from 1969 until his death. This address also gives me the opportunity to transmit the respect of this Academy and of its president, Professor Heinz Bethge, to the audience and to thank the organizers, especially Professor Horst Kleinkauf, for the invitation. The Leopoldina Academy is, at least in Europe, the oldest among the still active scientific Societies; it was founded 1652 in Schweinfurt. Although of international character, its fundaments are the German speaking countries with the headquarter in Halle since more than hundred years. Its goal is expressed in the still modern motto of the founders: "Die Natur zu erforschen zum Wohle der Menschheit" (exploration of Nature for the benefits of mankind). The genuine treasure of the Leopoldina are the almost one thousand members, outstanding and highly reputed personalities of natural and medical sciences from all over the world. Many of them assemble regulary at various scientific symposia and meetings, because the prestige and credit of Leopoldina depends on the activity, originality, and intellectual creativity of its members. Professor Fritz Lipman became one of the rare Honorary Members. This represents the highest distinction Leopoldina awards to a member. It was bestowed on F. Lipmann together with Sir Hans Krebs. Most remarkably, he was elected without previously having been an ordinary member. This exceptional event occurred so far only twice, namely, in the case of the great Otto Warburg and in Fritz Lipmann's, the "artistic type of scientist" - as he was called by his colleagues. Some sentences of the laudation on occasion of his election read as follows: "Fritz Lipmann contributed basically to the marvellous progress of our knowledge in biochemistry which likewise influences the intellectual culture of our epoche. The results of his research have lead to the explanation of a general cellular intermediate metabolism and to the elucidation of the rules and the mechanisms of these processes... In a fundamental manner he influenced current biology and deepened the understanding of life..." (translated by B. P.). This was communicated to the public audience at the biannual assembly of the Leopoldina in 1969, however, the following background story was omitted. In 1954, when the late Professor Kurt Mothes became president of the Deutsche Akademie der Naturforscher Leopoldina, he made great effort in reorganizing the scientific sections by electing outstanding scientists as Leopoldina members. One of them proposed by several members was Fritz Lipmann. He had been chosen for

110

Β. Parthier

membership at April 28, 1956, and president Mothes congratulated him kindly, respectfully, and in anticipation of his joining. But unfortunately, Lipmann rejected. In a letter of May 25, 1956, he wrote to Kurt Mothes: "It is of course always pleasant to get recognition of this type. However, I have made a rule not to accept a membership in cases where I could only be on an inactive status." - This was certainly an honest and honorable opinion worth to be kept in mind by some scientists. Nevertheless, the decision was made not by Professor X or Y but by the Nobel laureate Fritz Lipmann, and it was (almost) unique in the more than three hundred years old history of the Leopoldina. The presidium thought about this case for more than twelve years. Then, on occasion of his 70th birthday, Fritz Lipmann was elected as one of the five Honorary Members. Now, perhaps after getting older, probably being aware that an Honorary Member is delivered on account of continuous activities, possibly because the reputation of Leopoldina had radiated throughout the world in the meantime, he accepted the award in a very kind letter to President Mothes dated July 16, 1969: "I am greatly pleased to have been offered an honorary membership in your Society and I gladly accept it... It is particularly pleasant for me to get this honour together with Hans Krebs, a good old friend. I am happy to have been chosen to be one of the few and so be in the company of such outstanding men...". Outstanding personalities like Fritz Lipmann, who were forced to leave Germany in the thirties and became famous in the United States or other foreign countries not only contributed to the international reputation of this Academy. They belong to that generation of scientists which promoted the connections and the mutual understanding between people of the Old and the New World. Can we be sure that future missing of this generation do not cause unpredictable divergencies in our scientific world? Professor Lipmann visited Halle and Leopoldina in 1972, and he gave a public lecture on "Non-ribosomal biosynthesis of antibiotic polypeptides and its analogy to the synthesis of fatty acids". Some hundreds of professional and student scientists of the whole country celebrated enthusiastically the speaker of a splendid presentation. I myself remember vividly this visit, because Professor Mothes asked me to take care of the guest and to show him some of the medieval cultural attractions in and around Halle. I was profoundly impressed by the sincerity of his character, his modesty in behavior and by the wonderful clearness of his scientific ideas.

Dahlem in the Late Nineteen Twenties1 Hans A. Krebs and Fritz

Lipmann

Hans A. Krebs

The Kaiser Wilhelm Gesellschaft Dahlem was at that time the main campus of the Kaiser Wilhelm Gesellschaft (in 1948 renamed Max-Planck-Gesellschaft). This Society was initiated in 1910 with the intention of providing outstanding scientists with first-rate research facilities. The attitude of the founders was clearly expressed by Emil Fischer when he tried to persuade - successfully - Richard Willstätter to abandon his professorship at Zürich and to join the Society. Fischer, according to Willstätter (1), described the attitude in these words: "You will be completely independent. No-one will ever trouble you. No-one will ever interfere. You may walk in the woods for a few years, if you like; you may ponder over something beautiful". On the whole this policy (based on utmost care and competence in selecting the right people) has paid magnificent dividends: Otto Warburg, Otto Meyerhof, Albert Einstein, Max von Laue, Fritz Haber, Otto Hahn, Lise Meitner, Carl Erich Correns, Richard Goldschmidt, Michael Pqlanyi, Carl Neuberg and many others made the fullest use of the opportunities. By the late 1920's, within 15 years of its foundation, and despite the upset caused by World War I, Dahlem had become one of the world centres of scientific research. Not only did it attract many of the best scientists in Germany but also young people from all over the world.

Collaborators of Warburg and Meyerhof, 1926-1930 During my stay at Dahlem the people working in the laboratory of Warburg included Erwin Negelein, Hans Gaffron, Robert Emerson, Fritz Kubowitz, Werner Cremer, Erwin Haas, Walter Christian, Walter Kempner, Akiji Fujita and several other Japanese. MeyerhoPs laboratory, accommodated in the same bulding, a few steps away, included Karl Lohmann, Karl Meyer, Fritz Lipmann, Hermann Blaschko, Severo Ochoa, Frank Schmitt, Ralph Gerard, Dean Burk, David Nachmansohn, Louis Génévois, Ken Iwasaki. Other young biologists working in the same building were Victor Hamburger, Curt Stern and Joachim Hämmerling. Many of these left their mark on later scientific developments.

1

From: Lipmann Symposium: Energy, Regulation and Biosynthesis in Molecular Biology (D. Richter, ed.) de Gruyter, Berlin · New York, 1974

112

Hans A. Krebs and Fritz Lipmann

Achievements of the Laboratories of Warburg and Meyerhof in the Nineteen Twenties The discoveries made in the middle and later part of the 1920's in Warburg's laboratory included the discovery of the aerobic glycolysis of tumours, the general occurrence of the Pasteur effect, the accurate quantitative measurements of cell respiration and cell glycolysis, the carbon monoxide inhibition of cell respiration and the light sensitivity of this inhibition which made it possible to measure the action spectrum of the oxygen transferring enzyme in respiration (now referred to as cytochrome a 3 ) and to identify the catalyst as an iron porphyrin, the development of spectrophotometric methods of analysis (twenty years later commercially incorporated by Beckman into his black box), the discovery of copper in blood serum and the fall of its concentration in anaemias. Meyerhofs laboratory made decisive contributions to what is now called the Embden-Meyerhof pathway of gycolysis. It laid the groundwork leading to the discovery of hexokinase, aldolase and other enzymes. Monumental discoveries by Lohmann were those of ATP, first identified as a cofactor of glycolysis and of the "Lohmann reaction" - the interaction between ATP and creatine. One of the secrets of these outstanding achievements in both laboratories was the creation of new methods, such as the tissue slice technique, manometry and spectrophotometry by Warburg, and Lohmann's method of distinguishing between the many different phosphate esters by measuring their rate of hydrolysis at 100°C in 2Ν HCl. Now, some 45 years later, we can assess the achievements of Warburg and Meyerhof in proper perspective. Many scientific papers may seem to be very important at the time of their appearance but as the field develops it is appreciated that they were less significant than was at first thought; as time goes on the really significant contributions stand out as lasting landmarks. An amazing feature of the terms of Warburg and Meyerhof was their smallness. Altogether there were hardly more than two or three dozen people who participated in these great developments I have listed. Meyerhof had four or five small rooms and the total number of his collaborators at any one time was not more than five. There was only one trained technician, Walter Schulz, and a part-time typist. The technician was Meyerhofs personal assistant and together with a "Diener" he looked after the general laboratory affairs such as maintenance of apparatus and ordering of materials. When I joined Warburg there was one large room for six people in all, plus several instrument rooms, plus one Diener. Warburg had no secretarial help - we all typed ourselves. There was no technician in the ordinary sense of helpers. It is true Warburg's long-standing collaborators - Negelein, Kubowitz, Christian, Haas - were originally technicians but not in the ordinary sense. They were research assistants who had been primarily trained as instrument mechanics in work shops in the factories of the Siemens Company in Berlin. They knew how to handle instruments and how to make accurate measurements, and Warburg taught them all the chemistry they needed. In 1928 Warburg obtained one more room for four extra people and when he moved into his new Institute in 1931

Dahlem in the Late Nineteen Twenties

113

there were a few more places but the total number remained deliberately small. Meyerhof of course also had more space and more staff when he moved to Heidelberg in 1930. To those who participated what we did seemed to us quite normal and natural, a matter of course. Warburg, however, was always fully conscious of the monumental nature of his contributions. He has stated this in writing more than once, for intellectual modesty was not his strongest point. He considered himself to be in direct line with the giants of biology and in particular the chemically orientated biologists a direct successor of Pasteur, with no-one of comparable calibre between him and Pasteur (1822-1895). Meyerhof appeared much more humble, and not concerned with his own assessment of his position in the history of science. He was content to leave this to his peers and to posterity. And not only did the genius of leadership by Warburg and Meyerhof make outstanding contribution to the subject and inspire, the small band of deeply motivated and committed young collaborators. In the process Warburg and Meyerhof also educated (I do not say "train" because "educate" means more than train; educating includes the transmission of an outlock, not merely of technicalities) a future generation of leading scientists. This happened without actually aiming at doing it, without any policy of postgraduate training programmes. It happened naturally, and I believe something of this sort will always happen naturally. Born leaders attract born followers who develop into leaders - as long as bureaucracy and the erroneous concepts of equality do not interfere.

Day-to-Day Life in the Laboratories We all worked very hard and intensively, though the atmosphere was relaxed. In Warburg's laboratory the working hours were from 8 a.m. to 6 p.m. for six days a week. Most of the reading and most of the writing had to be done at home in the evenings, at weekends and during the summer vacation. Warburg and Meyerhof were in attendance more or less all the time and always accessible to their collaborators. There were no committee meetings and hardly any academic tourism. There was a brief luncheon interval where the younger people from different departments (especially from the laboratories of Warburg and Meyerhof) met in a common room for a simple snack consisting usually of eggs, sandwiches and milk. Coffee and tea breaks were unknown. The main vacation was rather long. Warburg closed his laboratory for 8 weeks during August and September but during this time he wrote most of his papers while on his estate in the Island of Rügen. Warburg liked to point out that the working hours were much less than they had been in his younger days. When we worked in Heidelberg in Krehl's Department of Medicine, Krehl often made a round of the laboratories on Sunday evenings and expected most of the workers to be in attendance. In Meyerhofs laboratory working hours were less rigid but hardly shorter. Warburg's control of the laboratory was very autocratic but we never questioned the justification of his authoritarian rule because we thought he was entitled to this on account of his outstanding intellect, his achievements and his integrity, qualities

114

Hans A. Krebs and Fritz Lipmann

which we admired enormously. On the whole his rule was benevolent but it could also be fierce. On one occasion he dismissed a research worker instantaneously after an incident in which Warburg thought he had not shown proper respect and courtesy. For Warburg autocratic control was essential in the interest of high standards of the work as well as of personal conduct. His was autocratic rule at its best. He never exploited the junior, as does autocratic rule at its worst. Democratic rule may at best make full use of the pooled resources but at worst it may create a situation where ignorance and obstruction prevails over competence and efficiency. Warburg was most generous in giving credit to his collaborators. Many pieces of research to which he had made the main contribution and which he had written were published without his name, except perhaps in an acknowledgement by the author. A review of his work (2) which established the oxygen transferring catalyst of cell respiration as an iron porphyrin ended with the passage "In concluding I wish to emphasize that the results which I have presented are largely due to the work of my collaborators, Drs. Negelein and Krebs". Yet the whole work was conceived by Warburg himself and the greater part of the critical experiments were carried out by him with his own hands. I know of at least one specific incident where Meyerhof was also very fair. In 1929 Lipmann had discovered (after Einar Lundsgaard had reported muscular contractions without lactate production in the presence of iodoacetate) that on anaerobic contraction muscle becomes initially alkaline even in the absence of iodoacetate, the rise of pH being due to the hydrolysis of creatine phosphate. Lipmann measured the pH change manometrically by the uptake of C 0 2 which reacts with the OH ~ ions formed. This was an important finding because it helped to establish the now generally accepted concept that creatine-P, through the Lohmann reaction, can energise contraction. As it was very important to him to find a job he was anxious that he should get proper recognition and he said in a somewhat resigned spirit to one of his colleagues, Hermann Blaschko. "This will be just another paper by Meyerhof and Lipmann". Blaschko then encouraged Lipmann to ask Meyerhof whether he may not be the first author. Meyerhof s immediate reply was "But of course" (3). And Fritz was the sole author of a second paper printed directly after the joint one, in which he showed that fluoride can act similarly to iodoacetate (4).

Financial Position of Young Research Workers Our financial position was very restricted and quite a few people in the laboratory received no salary or grant at all. Hermann Blaschko tells me that when he asked Meyerhof to be accepted in his laboratory Meyerhof eventually agreed to have him but he told him "I cannot give you any payment" whereupon Blaschko replied "I did not expect one". Fritz Lipmann was not paid during the first 2 years of his stay with Meyerhof, nor was Severo Ochoa paid. I was lucky and received a starting salary of 300 marks which rose to 400 after one year. It is difficult to equate this with present prices but it meant that we had to live frugally and count every penny. If we were careful we could afford one modest holiday a year; we could afford concerts

Dahlem in the Late Nineteen Twenties

115

and theatres in the cheapest seats. There were no travel grants for attending meetings. This did not mean that we were isolated because there were plenty of opportunities for learning something about new scientific developments within Berlin itself, through the colloquia at Dahlem and through the Berlin Chemical and Medical Societies. Who then, financed our maintenance? As far as I know our parents, even though they could ill afford it, for inflation had devalued the pre-war Mark by a factor of 10 12 (a million million) by the end of 1923. Parents were willing to make sacrifices for a good training of their sons. Of course, we felt very uncomfortable to be a burden to our parents when we were between 25 and 30 years of age. Franz Knoop impressed me in 1920 by saying during his lectures to medical students that he had earned nothing until he was 37 years of age, although he had made an outstanding discovery - that of /^-oxidation - when he was 30. It was understood that an academic career meant willingness to put up with very modest material standards of living. We were motivated by a keen dedication to our work and we were maintained by the hope that when we had received a thorough training - which we expected would last until we were about 30 - we would eventually get a worthwile job - satisfying professionally as well as financially. The sacrifices needed for a long period of training meant also that only the keenest did not give up. Most of us were medical graduates and could, if we wanted to, at any time branch off into a relatively lucrative medical career - but we preferred research. The financial and some other aspects of the scene at Dahlem were of course not unique. Erwin Chargaff - my contemporary working at that time in Vienna recently wrote (5) "No-one who entered science within the past 30 years or so can imagine how small the scientific establishment then was. The selection process operated mainly through a form of an initial vow of poverty. Apart from industrial employment in a few scientific disciplines, such as chemistry, there were few university posts, and they were mostly ill-paid". Nowadays there are many undergraduate students who insist on their "rights" to be fully supported by the state. We expected no rights even at the post-graduate and postdoctoral level. We were satisfied if we could work hard under reasonable conditions and learn. We did not feel entitled to expect much, let alone make demands, before we had learned a lot. In spite of our restricted economic circumstances we were on the whole very happy because we felt that we were receiving a first rate training and were doing something worthwhile. We had no undue worries about our long-term future although this looked very uncertain in view of the economic and political difficulties in Germany. I am not aware that any of us anticipated a particularly successful career. Being close to giants of science we felt very small and Warburg himself did not do much to encourage our self-confidence. In fact when I had to leave his laboratory he told me that he considered my chances in biochemistry as slight and advised me to return to clinical medicine (which I did).

116

Hans A. Krebs and Fritz Lipmann

Relations between Dahlem and German Universities Of a peculiar sort were the relations of Warburg and Meyerhof to the official representatives of German physiology and biochemistry, i. e. the German university departments. Warburg regarded himself as an outsider, and Meyerhof too, but perhaps less so. It is remarkable that German universities had not appreciated officially Meyerhof s qualities. By the time he got the Nobel Prize in 1923, at the age of 39, he held the post of an assistant in the Physiological Institute of Kiel University; he had been passed over for a junior professorial appointment (Professor Extraordinary) in favour of a man called Pütter, whose claim to distinction remained slight. After the award of the Nobel Prize to Meyerhof, Warburg succeeded in persuading the Kaiser Wilhelm Gesellschaft to offer Meyerhof a post. This sense of being outsiders had to do with the Cinderella treatment of biochemistry by the German universities. The number of chairs and departments of biochemistry or physiological chemistry was very small in Germany. Independent departments existed in four universities only, in Frankfurt with Embden at its head, at Freiburg, Tübingen and Leipzig. In other universities biochemistry was a subsection of physiology and the heads of these sub-sections did not have the rank of full professor. In some universities the professor of physiology was essentially a biochemist. This applied, for instance, to Heidelberg where Kossel was the Professor of Physiology and where first-rate work was done on protein chemistry. Thus it came about that Leonor Michaelis, one of the brightest biochemists of his time could not be absorbed into the German university system and therefore left Germany in 1921 for Japan, to move later to Johns Hopkins University and the Rockefeller Institute. While in Germany he had to earn his living as a clinical biochemist in one of the municipal hospitals in Berlin. Hopkins, in a general address at the International Congress held in Stockholm in 1926 commented on the neglect of biochemistry in German universities and spoke about the importance of "specialised institutes of general biochemistry". He referred to the appeal by Hoppe-Seyler in Volume 1 of his Zeitschrift für Physiologische Chemie in 1877 that institutes of biochemistry should be generally set up in the universities. In 1877 the only institute of this kind was that in Strasbourg with Hoppe-Seyler at its head. Hoppe-Seyler's appeal was at once opposed by the physiologist E. Pflüger in his Journal, and 49 years later Hopkins remarked that in fact Hoppe-Seyler's appeal for the recognition of biochemistry as an independent subject had still not yet found proper response in his own country and he added "It is difficult to see how Germany can continue the lead along the path which for a long time she has almost trod alone". He emphasised that academic centres in general in Europe are in this respect behind those in America. These remarks, incidentally, were made at the suggestion of F. Knoop for the benefit of the German readership and they were reprinted in translation in the Münchener Medizinische Wochenschrift (6).

Dahlem in the Late Nineteen Twenties

117

Now and Then Today the scene of scientific research seems very different. In the 1920's pure research was still widely looked upon as a luxury or extravagance that did not deserve major support from the State. The Kaiser Wilhelm Gesellschaft, after all, was founded in 1910 as a private organisation (while the German Universities were all state-controlled). Now, 50 years later, scientific research is regarded a necessity for the survival of a nation, and large sums are provided by Governments for training people and for pursuing research. But although the scene has changed some fundamental principles governing successful research are still the same, and will always remain the same - the recognition of leadership in research, the value of long training, the need for hard work and for dedication, an attitude of humility. What has gone, among other things, are the biting polemics in science in which Warburg liked to indulge, hurting and ridiculing the opponents - there are many examples of these in Warburg's books and in the Biochemische Zeitschrift (7). Gone, also, has the autocratic rule which Warburg, his teachers and his contemporaries practiced. This kind of benevolent (or mostly benevolent) dictatorship at its best as I already stated, helped to maintain high standards - but it could easily degenerate into arbitrary injustices, exploitation and mismanagement. Today a different basis of the relations between seniors and juniors has evolved. The master may still rule, and rule firmly, but the basis of his authority is now a natural respect, a natural mixture of admiration and affection which he has earned by his work and conduct; in a good laboratory authority is no longer based on the power invested in a head of a laboratory. I think this very occasion here is an illustration of the kind of human relations between the master and his followers which nowadays exist in a really ideal team. It was the desire of the followers to express a sense of loyality, gratitude and affection which has brought this symposium into being. This assembly of so many people from so far away, motivated by mutual goodwill, I find deeply gratifying and moving. The motive which brought us together remind me of a remark which Warburg made to me during a casual chat in the laboratory. For two years I shared an island bench with him, working opposite to each other in close proximity. Although both of us were not exactly talkative while preoccupied with our experiments there were occasional conversations touching on many aspects of life. One day he remarked "The worst defeat a scientist can suffer is to die early because the fruits of his labour mature very slowly". Although these words - like many other of his casual sayings are still deeply ingrained in my memory - I did not properly appreciate their full meaning at the time. But today their meaning is clear to me. The fruits of a scientist's labour are of several kinds. Promotion to a good post and a Nobel award are some; a very important one is the long term response from students and collaborators. Much of a research scientist's time is spent in helping to shape the outlook, career and life of his juniors, and the seeds which he plants take a long time to grow. Today we see the rich harvest that has come to Fritz from helping and guiding his younger

118

Hans A. Krebs and Fritz Lipmann

associates - a harvest not only in the form of gratitude and affection but also in the form of the intensely pleasing knowledge that there is a new generation willing to carry on the work and to uphold the standards and ideals which motivated the master. Such kinds of thoughts, I suspect, must have been in Warburg's mind when he suddenly spoke about the importance of living to old age. Perhaps some experiences of his father (8) (a founder of a large and devoted school of physicists), had impressed him 1 .

References 1. Willstätter, R. 1958. Aus meinem Leben. Zweite Auflage. Weinheim Verlag Chemie, p. 200. 2. Warburg, O. 1928. Über die chemische Konstitution des Atmungsfermentes. Naturwissenschaften 16, 345-350. 3. Lipmann, F., Meyerhof, O. 1930. Uber die Reaktionsänderung des tätigen Muskels. Biochem. Z. 227, 84-109. 4. Lipmann, F. 1930. Über den TätigkeitsstofTwechsel des fluorid-vergifteten Muskels. Biochem. Z. 227, 110-115. 5. Chargaff, E. 1974. Building the Tower of Babble. Nature 248, 776-779. 6. Hopkins, F.G. 1926. Über die Notwendigkeit von Instituten für physiologische Chemie. Münch. Med. Woch. 73, 1586-1587. 7. Krebs, H . A . 1972. Otto Heinrich Warburg. Biographical Memoirs of Fellows of the Royal Society. 18, 629-699. 8. Franck, J. 1931. Emil Warburg zum Gedächtnis. Naturwissenschaften 19, 993-997. 9. Weber, H . H . 1972. Otto Meyerhof - Werk und Persönlichkeit, in Molecular Bioenergetics and Macromolecular Biochemistry pp. 2 - 1 3 , Springer-Verlag, Berlin, Heidelberg, New York. 10. Peters, R. A. 1954. (with a contribution by H. Blaschko): Otto Meyerhof. Obituary Notices of Fellows of the Royal Society. 9, 175-200. 11 Murait, A. von 1952. Otto Meyerhof. Ergebn. Physiol. 47, I - X X . 12. Nachmansohn, D., Ochoa, S., Lipmann, F. 1952. Otto Meyerhof 1884-1951. Science 115, 363-369. 13. Nachmansohn, D. 1972. Biochemistry as part of my life. Ann. Rev. Biochem. 41, 1 - 2 8 .

Fritz Lipmann I am going now to describe my relationship to Dahlem when I was in Meyerhof s laboratory, and here I would also like to include Berlin. From my early youth Berlin, the great city, had been for me an magnet. I was born in a small town, Koenigsberg, then in East Prussia, and my first contact with Berlin was after absolvation of my abiturium, as we called the final examination ending gymnasium time and giving the right to enter into University. Then, my parents gave me as a gift, a week all on my own in Berlin to experience the theater and to experience the great city. I spent more time in Berlin later on studying medicine, and again I was impressed by the experience of what happens in that city.

1

Further material on Dahlem in the 1920's and the personalities of Warburg and Meyerhof will be found in references (6), (9), (10), (11), (12) and (13).

Dahlem in the Late Nineteen Twenties

119

And then I came back to Berlin for quite a long period after I had finished my medical studies. Still under the influence of medical friends, I spent the latter half of my practical year there, and the first three months with Ludwig Pick, an excellent pathologist, because it was thought that to become a physican one had to do some pathology. Then I heard about a biochemistry course which was given at the Charité, the Medical School of Berlin University, where a large number of physicians who became very good biochemists, including Hans Krebs and David Nachmansohn, were trained. That was Rona's laboratory. When I went to Ludwig Pick to tell him that I would spend the last three months of my practical year taking the biochemistry course of Peter Rona, he threw his hands up in amazement. Biochemistry then was still an unknown entity in Germany. Rona's name is probably known to very few of you; for quite a while he was a collaborator of Leonor Michaelis. I think it was an extraordinary course. I learned there the latest advances in the biochemistry of that time: manometry, pH measurement, electrophoresis, and so on. Actually, I stayed on with Rona and did a medical doctorate thesis which was obligatory in Germany. It did not need to be very important. Mine was on the electrophoretic behavior of iron oxide colloids, mainly concerned with the reversal of the positive charge to negative in the presence of citrate. Colloid chemistry was very modern in those days - many people used it to describe the protoplasm. It seemed enough then to call it a colloid to imply one understood something about it. We have learned better. That was in 1921-1922. Then I decided to go back to my home town to learn chemistry since I was lucky enough to be able to do this with Hans Meerwein who was the professor of chemistry at the University. I should guess that his name is known to those of you conversant with organic chemists. He was a superior chemist, and later moved from Koenigsberg to Marburg. In three years I learned a great deal, particularly from lectures; all during those years all students had to attend his lectures and it was a tremendous pleasure. He gave them all himself; there was no substitution by assistants, and that gave us students a contact with his personality. Then I got my Verbands-examen, which enabled me to start on a thesis. After that first step was finished I became somewhat restless. I felt it was now time for me to find a place where I could do biochemistry, for which I had been preparing myself all this time. It was not without other reasons that I chose Berlin; but I was mostly motivated by the existence in Berlin-Dahlem of the two institutions which at that time seemed to me to do work in the field I had begun to become interested in, intermediary metabolism. These were the laboratories of Carl Neuberg and Otto Meyerhof, and I debated for a time whether I should join Meyerhof or Neuberg. I eventually decided for Meyerhof because of his much more physiological leaning, and just as Hans Krebs has told you, I likewise didn't expect any salary and for the first two years I didn't get anything. I just asked him if I could work there and as I had studied chemistry and had some experience I was lucky enough to be accepted; he asked me, interestingly enough, if I had any problem to work on and I was ashamed to say that I hadn't. I had to get a problem from him. The problems I worked on in the early days there were not very important. I did some work on fluoride inhibition of glycolysis and fermentation, which was published in Biochemische Zeitschrift. These papers I could put together and use as

120

Hans A. Krebs and Fritz Lipmann

my chemical doctor's thesis. It reflects interestingly on the status of the heads of laboratories at the Kaiser Wilhelm Institutes that Meyerhof was unable to be my doctor "father" because he couldn't accept graduate students although he was a titular professor. However, Neuberg, whose institute was next door, could; he was a professor at the Technische Hochschule and I actually had to farm out my thesis, so to say, to Neuberg, who became my doctoral "stepfather" in a way. He always treated me very kindly. I will now say a little about the Meyerhof laboratory. While I was sitting here just now and thinking about my choice of laboratory, I am almost surprised to discover that although most impressed by Warburg, I never even dared to think of going to work with him. One of the important aspects of Meyerhof was that he was not as stern and was much looser than Warburg. But he was Warburg's pupil and there is a paper by Warburg and Meyerhof which came from the Naples Laboratory. The Marine Laboratory in Naples was one of the meeting grounds for biochemists in the same sense that Woods Hole is or was in the U.S.A. In his earlier years Meyerhof tended to be very interested in philosophy, had joined a school of philosophers, and wrote several papers of a philosophical character. It was the influence of Warburg, I understand, that decided him to become a biochemist; and he took all the traits of Warburg, the feeling that to do good work one needs the most exact methodology and has to have full confidence in one's results. The work in his laboratory, as you have heard already, centered about the muscle and I worked during that time largely with muscle or muscle extracts, mostly related to glycolysis. It is only in the later period after the laboratory had moved to Heidelberg that I did a fairly nice piece of work on a determination of creatine phosphate breakdown in living muscle which I measured manometrically. These manometers that I used were somewhat difficult to construct because I wanted to stimulate the muscle in the manometer and one had to seal in platinum electrodes, which was very hard to do without a leak. This work started when Meyerhof suggested that I should try to see what happens in muscle contraction at relatively high acidity, that is, in a bicarbonate solution with C 0 2 in the gas phase. It was then that I found, during the early phase of a series of contractions, an alkalinization, i.e., manometrically, C 0 2 absorption instead of the expected C 0 2 liberation from lactic acid formation. Chemical analysis of the muscle showed that the alkali that formed early corresponded very nicely to a creatine phosphate breakdown; this had been found to yield alkali because of the strong alkalinity of the guanidinium base liberated. This was an early proof that without inhibitor under these acid conditions creatine phosphate breakdown could cause contraction. It is now time to show you Meyerhof. This picture (Figure 1) is rather typical because it shows that it was not easy to approach Meyerhof. We actually talked very little and what I learned from him was largely by diffusion. But this was to influence me all through my life. This picture of us together was taken in 1941 at a conference in Madison, and you might say we look a little uneasy, which was because I had given a lecture there on the Pasteur effect and had shown disagreement with his interpretation, and he wasn't too happy about 1. Our working hours were much more relaxed than those at Warburg. I remember that we took pretty long lunch intermissions and went rather often to a little

D a h l e m in the Late Nineteen Twenties

Figure 2

121

122

Hans A. Krebs and Fritz Lipmann

restaurant, which was next to a further-up subway station, and sat there and happily talked in the garden. When I say we, in the next picture (Figure 2) you can see some of the people who were "we". You can see, from left to right, Ken Iwasaki, Karl Lohmann, Walter Schultz, and Schroeder, who was something in between a scientific assistant and a diener; then David Nachmansohn and Paul Rothschild; and again me, this time with a long tie. Ken Iwasaki, Nachmansohn, Paul Rothschild, and I were the ones who often went to that little restaurant, and not only that, we even went together to masquerade balls which were very fashionable and much attended at that time; but they were very good entertainment and as free in spirit as in present-day terms. At one of them, the socialist ball, which had nothing to do with socialists, I met Freda Hall who was to become my wife. So that was an important event during my Berlin days. Actually, at this particular ball, David Nachmansohn danced more with Freda than I did, but he was already married. Now to return to the laboratory. I had some contact with Ralph Gerard - 1 think we shared a laboratory when I entered the Meyerhof Laboratory; we met again in later life and I was rather fond of him. Then I moved into another laboratory and worked very close to Ken Iwasaki who spent a good deal of time in Berlin and I am sure had a very good time there; he was not married then. I am told that Mr. Takeda, Masao's father, became a very good friend of Ken Iwasaki; they spent much time together in Berlin and are still very good friends. Ken Iwasaki is now retired from his biochemistry professorship and has a laboratory in the Takeda Company, which is one of the largest pharmaceutical companies in Japan. The topics on which the work was done in Meyerhof s laboratory were not too varied; it was mainly concentrated on the muscle, but it also included nitrogen fixation on which Dean Burk and Ken Iwasaki worked. The status of our understanding at that time may be shown by what Karl Meyer did. He was trying, in parallel to what had been called zymase by Harden, to isolate the "glycolytic enzyme"; it wasn't quite realized then that the glycolytic enzyme was of course composed, as we now know, of numerous enzymes and that these enzymes could eventually be separated. That came not much later, but in 1928 it was really surprising that one could even aim at thinking of isolating as a unit something like a glycolytic enzyme. Then, shortly before we moved away from Berlin, I worked for a little while with Lohmann, and I learned much from him. As you heard already, he really was an artist in determining by acid hydrolysis the different phosphorylated compounds that were in a mixture. For example, in this way he discovered the equilibration of glucose 6-phosphate with fructose 6-phosphate. The latter has a much faster hydrolysis time than glucose 6-phosphate, which is one of the most difficult to hydrolyze phosphate estes. On the other hand, fructose diphosphate is completely hydrolyzed within three hours. Lohmann used the acid hydrolysis of phosphate esters very effectively and, as I said, he discovered many new compounds. I am very eager to show once in a while an experiment from which the wrong conclusions were drawn. Table 1 shows such an experiment that I did with Lohmann and which eventually became of great consequence. In this experiment fructose

Dahlem in the Late Nineteen Twenties

123

diphosphate was incubated in muscle extract, and I came in to try and see if this conversion of fructose diphosphate would also go without fluoride that was added in Lohmann's earlier experiments. The table shows the change of fructose diphosphate without fluoride to what was called a difficult to hydrolyze hexose diphosphate. One can see that by the incubation in muscle extract phosphate is not released. However, the hydrolysis time of the added fructose diphosphate goes up if one uses the three hours mentioned above as a standard. One can see that with time it becomes converted, and eventually 70% of it is present as what was thought to be a different hexose diphosphate which was much more difficult to acid-hydrolyze.

Table 1 Conversion of F D P into acid-stable phosphate ester in muscle extract of winter frogs after incubation at 20 °C Incubation

Phosphate bound (mg P¡)

Phosphate acidhydrolyzed in 3 hr (mg P¡)

Converted phosphate (mg Ρ,)

(min)

(%)

0 20 60 120

0.48 0.50 0.50 0.49

0.48 0.39 0.25 0.19

0.12 0.27 0.33

25 53 68

Biochem. Z. 222, 389 (1930)

To turn a little more to the history of this part of biochemistry as I did in the meeting at the Ciba conference last week where I also mentioned these experiments, Nilsson had shown a little earlier in Euler's laboratory that in the presence of fluoride, the same fructose diphosphate, with a yeast preparation, when paired with acetaldehyde, gave phosphoglyceric acid as the oxidation product parallel with reduction of acetaldehyde to ethanol. This was the first appearance of phosphoglyceric acid in the picture of fermentation and glycolysis and nobody at that time could appreciate why this compound was formed as an oxidation product of fructose diphosphate. Nilsson came to the Meyerhof laboratory and we discussed this strange compound, phosphoglyceric acid, without seeing the light. The right idea was Embden's who essentially repeated our experiment. We had just made barium precipitates of what we considered a hexose diphosphate. He found that it was surely difficult to hydrolyze but that it was not hexose diphosphate but rather a mixture of phosphoglycerol and phosphoglyceric acid which he assumed to be formed by a dismutation reaction. Thus, we had misinterpreted the conversion of fructose diphosphate. However, in the hands of Embden it became the reason why we now talk about the Embden-Meyerhof cycle; from this reaction he then concluded the disruption of fructose diphosphate into an equilibrium mixture of two trióse phosphates and mapped out the foundation of our present scheme, recognizing phosphoglyceric acid to be the oxidation product of phosphoglyceraldehyde, the biochemically dominant of the two trióse phosphates formed. That's just a sidelight on Lohmann's artistry with hydrolysis. As I said, I learned much from him to handle what I did soon afterwards. When I went to New York to

124

Hans A. Krebs and Fritz Lipmann

work with Levene, I chose the phosphoproteins as an object of investigation. You heard in Dr. Helmreich's talk that I there isolated serine phosphate from the egg yolk phosphoprotein. Lohmann was the one who suggested that I work with Levene. He had in mind, I think, that I should work on nucleotides, but I chose the mentioned topic because Levene had isolated a very phosphate-rich protein from egg yolk which attracted my interest. The methods I used were actually borrowed from what I had learned from Lohmann. It appeared that the phosphate in the yolk protein is alkali-labile but very acid-stable, and choosing acid hydrolysis as a means of degradation of the protein, I thus isolated the serine phosphate. Much later, I turned to what was the prominent interest in Meyerhof s laboratory, namely, bioenergetics. This came to be an underground well, so to say, that eventually opened up after I discovered acetyl phosphate and led to my writing the paper on generation and utilization of phosphate bond energy. I would like to close by saying a little more about Berlin in those days. Before going to New York, Freda Hall and I were married, and we were amazed to see that there was such an enormous difference in the way of life between Berlin and New York, particularly among young women and young men. We read in the "Saturday Evening Post" three articles by Hergesheimer, who was then a rather fashionable novelist, in which he described Berlin as the center of Europe: people didn't go to Paris, they went, rather, to Berlin. There was the theater, there was the music, there was the dance; you could have everything. There was also a great freedom in Berlin in the late twenties, a similar freedom to that which has developed in America in recent years. It was due to the breakdown of the family ties by which young people were held because they had to depend on their families, and it had the effect that the young men and women interacted much more freely with each other. After I had moved away from Berlin and from Germany, it took a long time to forget the way of life we had experienced there.

Our Apprenticeship H. Blaschko

Much has been written in recent years about the period during which Fritz Lipmann worked with Otto Meyerhof both in Dahlem and Heidelberg. This period belongs to the golden age of biochemistry. To those who have shared this experience it has been a decisive one, and I am proud to be one of them. My time with Meyerhof coincided with Lipmann's. I entered the laboratory at Dahlem in the beginning of 1925, and I left Heidelberg late in 1932. This period was interrupted, twice by illness, and once I had a year as an assistant in the University of Jena. And in 1929-1930 Meyerhof sent me for a year to University College, London, to work with his friend, Α. V. Hill. A few years later I came back to England, which became my home. It was only then that I was able to embark on the work that I have continued throughout my active life. It may be worthwhile to point out that Meyerhof had a great influence on a number of workers whose studies had some relevance to the development of pharmacology and neurobiology. One of the reasons why I was first attracted to Meyerhof s laboratory was what I heard about it from my friend Rolf Meier. Meier had been with Meyerhof during the latter's stay at Kiel, before he came to Dahlem. In later years Rolf Meier became the founder of the pharmacological laboratories of the Ciba Company in Basel. These laboratories acquired a world-wide reputation during his time as director. Already at Kiel, Harold E. Himwich was a visitor to Meyerhof s laboratory. After his return to the United States he made some early measurements of the oxygen consumption of the mammalian brain in vivo. He eventually became director of a neurochemistry institute in Galesburg, Illinois. An early visitor to Meyerhof's laboratory was R. W. Gerard, who came to Dahlem after a few years with Α. V. Hill in London. After his return to the United States he became a pioneer in the use of intracellular electrodes, a method that has become widely used in both physiology and pharmacology. Another visitor from the United States was Chalmers Gemmili, who made important early observations on the role of insulin in the carbohydrate metabolism of isolated muscle. Also Frank O. Schmitt came to Dahlem from the United States. He made many important contributions to neurochemistry. Eventually he joined the Massachusetts Institute of Technology, where he initiated the Neurosciences Research Program. Later, in Heidelberg, there was Eric Boyland whose subsequent work on the metabolism of aromatic hydrocarbons has been of great importance for our knowledge of carcinogenic substances. In Heidelberg there was also A. Grollman, who subsequently developed important new methods for measuring the minute volume of the human heart. Hermann Lehmann joined Meyerhof s laboratory in Heidelberg only after I had left it, but I saw much of him in later years when he worked with Joseph Needham in

126

H. Blaschko

Cambridge. His main area of research was in the study of abnormal hemoglobins, but the pharmacologists remember him as the discoverer of a genetically determined inability to hydrolyse succinyl-choline, an observation that led subsequently to the opening of the new field of pharmacogenetics. A few years ago we commemorated here in Berlin the life and work of David Nachmansohn, friend and biographer of Otto Meyerhof. All workers in the field of biochemical pharmacology are familiar with his studies of the enzymes acetylcholine esterase and choline acetyl transferase. He did not live to see the spectacular developments that in recent years the molecular biologists have brought to the study of the acetylcholine receptor. Many of his ideas on the involvement of specific proteins in nerve function have been vindicated by recent findings, although some of his theories are still not generally accepted. My own interests in research have centred around the biochemical pharmacology of the biogenic amines, and particularly of adrenaline and the catecholamines. This work began almost by accident. I took it up because of my earlier experience in Meyerhof's laboratory. Adrenaline was a substance classed as "autoxidizable". On Meyerhof's suggestion I had carried out in Dahlem a study of the reversibility of the cyanide inhibition of autoxidations. In Cambridge, soon after my arrival there, I was able to show that the three catecholamines, adrenaline, noradrenaline and dopamine were substances of the enzyme monoamine oxidase. This was made possible when by the use of cyanide we were able to suppress "autoxidation" and to isolate the enzyme-catalysed action of monoamine oxidase. It so happened that this enzyme was cyanide-insensitive! Much of the early work on the oxidase was carried out by the use of manometry. This was a method that I had first encountered in Meyerhof's laboratory, an experience reinforced early in 1933, when as a convalescent in the Medical Clinic at Freiburg i. Br. I spent much of my time out of bed in the laboratory of my friend Hans Krebs, who was also one of my doctors. Manometry was a method that I found useful for many years. At first it served in the investigation of the amine oxidases, and it remained in use for this until my retirement twenty years ago. Another application of the manometric technique was in the study of the aminoacid decarboxylases. Peter Holtz, the discoverer of L-Dopa decarboxylase, had already used manometry, and when we demonstrated the importance of this enzyme in catecholamine biosynthesis, it became a useful tool for the study of the substrate specificity of this enzyme. The enzyme L-cysteic acid decarboxylase was also discovered by the use of manometry. The early suggestion that the enzyme was involved in the formation of taurine in the mammalian body has since been amply confirmed. The enzyme has also proved of interest to the neurochemists, since it is now known that taurine appears in a number of neurones, e. g. in the retina and in the cerebellum. These neurones contain the decarboxylase. I might mention here that manometry also served for many years as the method of choice in the studies of acetylcholine esterase. As elsewhere, the method has been superceded by more sensitive techniques.

Our Apprenticeship

127

In 1935, when we studied the monoamine oxidase of mammalian liver, the preparation that we first used, was the "Körnchensuspension" described by Otto Warburg many years earlier. I first encountered these in Dahlem. After the War, by the use of differential centrifugation it was established that monoamine oxidase was a mitochondrial enzyme. More recently the enzyme has often been used as a marker for the outer mitochondrial membrane. Our experience with differential centrifugation proved of great usefulness in later years when this method helped us to describe and study the chromaffin granules of the adrenal medulla. We helped to develop and refine the method of density gradient centrifugation, and the procedure developed in 1967 by two of my colleagues, David Smith and Hans Winkler (1) is the method of choice for obtaining satisfactorily pure samples of chromaffin granules. These cell organelles have become of renewed interest in recent years: their soluble proteins, the chromogranins, have been discovered in a number of other secretory granules, e.g. in the pancreas and the parathyroid. The distribution of chromogranins in the central nervous system has also been studied. This catalogue of methodical gains that I carried away from my period in Dahlem and Heidelberg gives a very incomplete picture of the debt that I owe to the privilege of having had my apprenticeship in that exciting environment. We were apprentices, in the true meaning of the word, content with working on problems that had been formulated by our teachers. We were content because we believed that this was worthwhile. Otto Meyerhof was not always very communicative and the not all too frequent occasions when we talked over our problems with him were occasions that we do remember. But we did not learn only from Meyerhof, we also learnt from each other. Karl Lohmann, in his quiet and unobtrusive way, was a great influence. And there were also many other exciting and stimulating contacts that we made there. These experiences also included the seminars and meetings that we were able to attend. I have already described some of these in a recent essay (2). This Symposium bears witness to the fact that the influence and experiences to which we were exposed in our early days are still alive, and that through the work of Fritz Lipmann and his contemporaries this tradition is being handed on to a following generation.

References 1. Smith, A.D. and Winkler, H. 1967. A simple method for the isolation of adrenal chromaffin granules on a large scale. Biochem. J. 103, 480-482. 2. Blaschko, H. 1983. A Biochemists Approach to Autopharmacology. In: Comprehensive Biochemistry 35, 189-231. Elsevier.

The Kaiser-Wilhelm-Institutes in Berlin-Dahlem in the Late 1930ies and Early 1940ies: Reminiscences of a Student of Biochemistry Peter Karlson

I The "Kaiser-Wilhelm-Gesellschaft zur Förderung der Wissenschaften" was founded in January 1911. The aim was to erect scientific institutes devoted to research only, in contrast to the university institutes which always had teaching responsibilities. It was a private enterprise, the money came from rich men and from the industry, and it remained an institution independent from the state (at least de jure) even under the regime of the Nazis. Its philosophy was first to look for an outstanding scientist, get him, and build an institute according to his needs. Then leave him alone - and with his inspired research he will eventually win the Nobel Prize. The Kaiser-Wilhelm-Gesellschaft was refounded 1948 as "Max-Planck-Gesellschaft zur Förderung der Wissenschaften", again as private institution, and is still legally independent, though more than 90% of its funds are government money. The Prussian government assigned a large area of the state-owned Royal demesne, the "Domäne Dahlem", to the Kaiser-Wilhelm-Gesellschaft to build the scientific institutes there. Among the first institutes to be built, several were concerned with biochemistry: The Kaiser-Wilhelm-Institut für Chemie, where Richard Willstätter worked from 1912-1916; the Kaiser-Wilhelm-Institut für experimentelle Therapie with August Wassermann as director and Carl Neuberg as head of the division of biochemistry; and the Kaiser-Wilhelm-Institut für Biologie with the directors Carl Correns, Hans Spemann, Richard Goldschmidt, Max Hartmann and, last not least, Otto Warburg. In spite of the difficulties during the years of the first World War 1914-1918, scientific life at the institutes flourished. However, the great years of biochemistry was the decade between 1925 and 1935. When Otto Meyerhof won the Nobel Prize in 1922, the directors of the Kaiser-Wilhelm-Institut für Biologie sacrificed some laboratory space to house him, and he became co-director at the institute of biology. There he continued his studies about glycolysis and muscle contraction which finally led to the elucidation of the Embden-Meyerhof-pathway of breakdown of glucose. With his famous pupil Lohmann, he discovered that high energy phosphates (creatine-phosphate and ATP) were formed in muscle and were the source of energy for muscle contraction. Meanwhile, Otto Warburg had discovered his "Atmungsferment", now known as cytochrome oxidase, and around 1935 he found the first "yellow enzyme". He also identified nicotinamide as part of a coenzyme which he called "triphosphopyridine

130

Peter Karlson

nucleotide". For more than two decades, the abbreviations DPN and TPN were standards in the formulation of metabolic reactions and pathways. Finally, mention must be made of Carl Neuberg who worked on alcoholic fermentation in yeast and discovered a number of intermediates. By use of inhibitors, he could manipulate the fermentation so that not alcohol but other reduced end products appeared, among them glycerol. Regrettably, he adhered far too long to the idea that methylglyoxal was an intermediate in alcoholic fermentation. Due to these developments, Dahlem was the Mekka of postdoctoral students of chemistry or medicine who wanted to learn biochemistry. Many of them stayed several years in one of the laboratories mentioned. Among these was also Fritz Lipmann who worked with Otto Meyerhof from 1927 to 1930, when Meyerhof moved to Heidelberg to become director at the Kaiser-Wilhelm-Institute for Medical Research. He later worked with Albert Fischer in Berlin before he went, in 1931, to Kopenhagen. Among other scientists who spent their years of apprenticeship in biochemistry in one of the Kaiser-Wilhelm-Institutes in Berlin-Dahlem in these years were Hermann Blaschko, Hans Krebs, David Nachmansohn and Severo Ochoa, to name only a few very important ones.

II

In these years - between 1925 and 1933 - , my affiliation to biochemistry was very feeble. I was a school-boy, and though interested in nature, as well as in chemistry and physics, I had no profound knowledge of any of the fields mentioned. However, it so happened that our family lived not very far from Berlin-Dahlem, and I used to take long walks in the streets and often passed the Kaiser-Wilhelm-Institutes, wondering what might be going on behind these walls. Terrible things happened after Hitler came to power in January 1933. I was then fourteen years old. The Nazies insisted that all Jewish scientists were expelled, not only from the universities but also from the institutes of the Kaiser-WilhelmGesellschaft. Due to the fact that the institutes of the Kaiser-Wilhelm-Gesellschaft did not belong to civil service, as the universities, it took a few years longer. But from Dahlem, Carl Neuberg had to leave in 1934, Meyerhof, who worked in Heidelberg, had to leave Germany in 1938, and Richard Goldschmidt went to Berkeley, USA, in 1935. Only Otto Warburg, being quarter Jewish, remained director of his institutes during the years of the war. The house of my parents, especially my father, was strongly against the Nazis. We used to have Jewish guests, before 1933 as well as afterwards. I soon heard that many outstanding Jewish scientists and artists - musicians, actors and others - had to leave Germany in the years 1933 and 1934. We regretted this; we could not forsee that those who left were lucky, those who could stay for one year or another were in most cases very unlucky.

The Kaiser-Wilhelm-Institutes in Berlin-Dahlem in the Late 1930ies and Early 1940ies

131

In 1937,1 took my final highschool exam (Abitur) and wanted to enrol in the faculty of science in Spring 1937. At that time, the Nazi regime was fully established, and a second prerequisite for admission to the university besides the abitur was a membership in one of the Nazi organizations. Though there was some freedom for the elder generation - not everyone was forced to join one of the NS-organizations - , children and teenagers were totally "organized". It started at the age of 6 - 1 0 with "Jungvolk", followed by the "HitlerJugend" and, for the girls, the "Bund Deutscher Mädchen". The students were organized in the "NS-Studentenbund". Thus I joined not only the university but also this NS-organization. It should be said that these organizations were, in practice, more like social than political organizations. It is true that occasionally, e.g. at the large demonstrations at Labour Day (May 1), we had to parade in brown uniform. But, since virtually all students belonged to the NSDStB (with the exception of foreign and half-Jewish students), it is obvious that they were not all politically engaged. My subdivision or "Kameradschaft", for example, continued the tradition of one of the classical student's associations (Burschenschaften). A foreigner may ask how this was possible that a whole generation was treated in such a way. Well - it was a totalitarian regime. There was, in this respect, little individual freedom. If you wish to be admitted to the university, please join this NSorganization. If you refuse, choose some other profession. Though I had also interests in other sciences, particularly biology, I had chosen chemistry as my field. This I had practised at home in the bathroom with various succuss. I absolved my curriculum in the proper way. It was interesting but not very exciting. As mentioned above, I was interested in biology as well as in chemistry; but biochemistry was not a field that could be studied at the university of Berlin in the late 1930ies. Of course, in the faculty of medicine, physiological chemistry was taught. The Institute of Physiological Chemistry happened to be next door to the chemical institutes in Berlin, but there was virtually no contact, and it was difficult for students of sciences to enrol in medical courses. At least, no one informed me that this was possible. After I had finished my diploma at the chemical institute of the university in Berlin, I looked for another place to study for my Ph. D. I must admit that my elder brother gave me the advice that I should do so, and he also suggested that I should ask Adolf Butenandt if I could work with him for a thesis. Butenandt accepted me, and in May 1940 I began my thesis work at the Kaiser-Wilhelm-Institut für Biochemie.

III I have outlined above what main subjects of biochemistry were studied in the various laboratories in Berlin-Dahlem in the late 20ies and early 30ies. As a matter of fact, these were not the biochemical problems that made headlines between 1925 and 1935. The interesting fields were the vitamins and the hormones. Vitamin Bj

132

Peter Karlson

was isolated in 1926, vitamin D 3 in 1932. Strange enough, in Berlin-Dahlem there was no research on vitamins, one of the modern fields in biochemistry. However, the important discovery that vitamins are part of coenzymes was made in Dahlem by Warburg and Theorell. The hormones were another field successfully explored. Insulin was isolated by Banting and Best 1922, and the nature of thyroxine was elucidated in 1926. The first steroid hormone, oestrone, was isolated in 1929, followed by androsterone in 1931. These discoveries gained wide publicity. Of course, it is easier for a journalist to describe diseases and their cure by vitamin or hormone treatment, let alone the sex hormones, than to report on the elucidation of the Embden-Meyerhof-pathway! Adolf Butenandt had been called to Berlin-Dahlem to succeed Carl Neuburg who had been forced to resign as director of the Kaiser-Wilhelm-Institut für Biochemie. Butenandt was one of the exponents of hormone chemistry of these years. As I mentioned, Butenandt had accepted me as a Ph.D. student, and he gave me a stereochemical problem, the elucidation of the structure of lumi-estrone, an isomer arising from irradiation of estrone with ultraviolet light. It was chemical work; I had to carry out chemical reactions with estrone and lumiestrone, which should eliminate the stereochemical differences and lead to identical compounds from estrone as well as lumiestrone. Professor Butenandt visited the laboratories twice a week and expected a progress report. At such visits, he also discussed the next steps and possibly new experimental routes. My experiments went with varying success. I once had to change my plans, but finally solved my problem showing that lumiestrone was, in fact, 13-epiestrone. This was the gist of my thesis, and in 1942 I obtained my Ph.D. I had learned a lot in these two years. The experimental skill was of some importance. But under Butenandt's guidance, I was introduced to science at its best: free scientific discussions. Once a week, all scientists of the institute came together for a literature seminar to review and discuss current papers. It was the custom that everyone had one journal on which he reported. These seminars were hampered by the lack of foreign, especially American, literature due to the war (The Biochemical Journal remained available for quite some time). But we still had German and Swiss literature, especially the Helvetica Chimica Acta with its many steroid papers which were of interest. Emphasis at these seminars was on the chemistry of natural products, especially steroids: but other fields were also discussed, e.g. endocrinology, cancer problems, biochemical genetics, virus research, and biochemical problems of fundamental interest. Another part of may education as scientist was the free exchange of current results and ideas with fellow scientists. In Butenandt's institute, there were no locked doors or colleagues who would discuss their results only when the paper was already in print; everybody talked freely about current results or problems of his or her research. This free discussion was very profitable; it often led to suggestions for technical or methodical improvement or even a new approach to a given problem. It also included criticism - important for any student who wanted to become a scientist. Moreover, from the discussions with the people in the laboratory working on different subjects, I learned much about many biochemical problems apart from my own work.

The Kaiser-Wilhelm-Institutes in Berlin-Dahlem in the Late 1930ies and Early 1940ies

133

I became interested in the substances that are produced under the influence of genes and are responsible for eye-pigment formation in insects. It was Wolfhard Weidel, another Ph.D. student, who worked on this subject and had just made a breakthrough: he had discovered that kynurenine, a metabolite of tryptophan, could substitute the natural compound responsible for the pigment formation in Drosophila v + strain and Ephestia a + strain. The next step was the elucidation of the structure, since it was soon realized that the structure given by Japanese authors was incorrect. He solved this problem within six months and got his Ph.D. in October 1940. Immediately afterwards, he was drafted to the army. Another field of general biological interest under study at the Kaiser-WilhelmInstitut für Biochemie was the structure of tobacco mosaic virus. In 1938, Butenandt had sent Gerhard Schramm to the laboratory of The Svedberg to learn the methods of protein characterization. Back in Germany, Schramm constructed his own type of ultracentrifuge, an air-driven design. In 1943, Schramm discovered the dissociation of tobacco mosaic virus into subunits of protein and the nucleic acid, and the spontaneous reaggregation of the components to virus-like particles by subtle shifts of the pH value of the solution. This gave a new insight into the structure of tobacco mosaic virus and was, to my knowledge, the first time that a capsid protein of a virus was isolated as monomer. This important result was obtained during the war and under difficult conditions, and it was not generally acknowledged until the 1950ies. James Watson gave the reason in his famous book "The Double Helix": "Virtually no one outside Germany, however, thought that Schramm's story was right. This was because of the war. It was inconceivable to most people that the German beasts would have permitted the extensive experiments underlying his claims to be routinely carried out during the last years of a war they were so badly losing. It was all too easy to imagine that the work had direct Nazi support and that his experiments were incorrectly analyzed. Wasting time to disproye Schramm was not to most biochemists' liking". I can confirm that virtually all the work done in the Kaiser-Wilhelm-Institut für Biochemie during the war was fundamental research. There were no projects of military importance under investigation. After my Ph. D., I continued to work in the institute and I should have noticed such activities. My own subject were studies towards the isolation of the moulting hormone of insects, now known as ecdysone. Thus, even during World War II, at least in the early 40ies, Dahlem remained a place of fundamental research, and remained open (at least in most cases) for exchange of ideas in the old tradition of the 1920ies and 1930ies.

On the Origin of the Squiggle ( ~ ) Paul A. Srere

Of the many great contributions to biochemistry made by Lipmann, it was his article on bioenergetics in 1941 that had the greatest influence on me. I did not read the article until 1948 when I was a second year graduate student. I was struck by the clarity of his thinking and his ability to reduce the complexity of a myriad of metabolic observations into the simple and still relevant "metabolic wheel". The piece de resistance of that sketch was the introduction of the elegant notation of a squiggle ( ~ ) to denote a "high-energy" bond (biologically speaking). In retrospect it took not only extraordinary scientific insight for the development of the ideas but superb artistic intuition on the universal appeal of that mark ~ . The symbol itself suggests simplicity, symmetry, and power. It was not until years later after being frightened by George Santayana's dictum, "Those who cannot remember the past are condemned to repeat it", that I became interested in the early history of scientific thought. Imagine my surprise when I accidentally discovered that as long ago as 2900 B. C. there was knowledge not only of ~ but of the Krebs cycle and several fundamental concepts of what we now know as molecular biology. This happened one day when in a book store my eye was caught by the cover of a book which promised the following things: 1. I Ching — is the first written book of wisdom, philosophy, and oracle and had acted as a guide for leaders and scholars throughout history. 2. I Ching — is the guide that tells you what to do about everything from your love life to your business affairs. 3. I Ching — is an advisor which reveals the best action to be taken in every decision and event in life and can be used every day both for divination and selfexploration. 4. I Ching — is the vehicle for understanding the patterns of change that govern all life. While the first three points were very good selling points, it was the final statement that convinced me that I needed the book. I was really interested in the promise that through the I Ching, I could understand the patterns of change that govern all life. I bought the book, and my first quick perusal indicated it was probably of no value in helping me to understand metabolism or its regulation. However, a more careful reading showed me there was much to be learned from it. First, let me point out that I Ching is a system of fortune telling first invented by the Emperor Fu Hsi who lived 2953-2838 B.C. (Figure 1). To him is attributed, in addition to the I Ching, the teaching of hunting, fishing, sheep herding, and the invention of musical instruments. The I Ching is a method of

Paul A. Srere

136

Figure 1

The emperor Fu Hsi in fortune telling.

fortune telling that depends on ordering a series of events which are randomly arrived at, such as the toss of coins or the dropping of straws. I could not help but wonder how a series of random events used predictively and explanatorially could have been sufficiently good to last 5000 years and have sufficient power to outlast many more scientifically based theories. Despite this skepticism, my attention was drawn to the basic symbols of I Ching, the trigrams, which were arranged as shown below. The trigrams each have a special name and predictive interpretation and could, for example, be constructed in the following way. One would designate one side of a coin as a yin ( —) and the other side as yang (--). One would pose a question about an action to be taken, toss the coin three times, and construct the trigram, starting at the bottom in the order the two sides of the coin appeared. Eight possible trigrams are possible, and they appear in front of Fu Hsi. I was struck firstly that they were arranged as a cycle. Secondly, it seemed that the trigrams, especially ( ξ ) , looked like a shorthand for citric acid. I was startled to realize that old Fu Hsi had predicted the Krebs cycle. The eight trigrams indicative of the eight carboxylic acids of the cycle showing citrate, cisaconitate, isocitrate, α-ketoglutarate, succinate, fumarate, malate, and oxaloacetate. Had we known this previously, citrate could never have been excluded from the cycle as it was briefly some years ago. A further development of the I Ching was to construct hexagrams, each hexagram having a special meaning. I was even more amazed when I turned the page to find the arrangement of the hexagrams, the complete development of I Ching (Figure 2).

On the Origin of the Squiggle ( ~ )

Figure 2

137

Arrangement of hexagrams - the complete development of I Ching.

I am sure you will be as astounded as I was that the I Ching had presaged the circular chromosome of E. coli. In addition, the number of hexagrams totals to 64, which of course corresponds to the number of codons that properly arranged is the basis for the information in all living cells. When I went back to study the trigram I found that in my excitement of finding the presaging of the Krebs tricarboxylic acid cycle, I had overlooked the central feature that usually accompanies the symbols for the eight elements of the universe (Figure 3). That symbol, a divided circle separating heaven ( Ξ ) from earth (ΞΞ), the symbol fqr yin and yang. What was amazing was that the circle was divided with a squiggle ~ . What could have been more prophetic that the juxtaposition of ~ withift the Krebs cycle? Here was the prediction that the generation of the energy of life by a cyclic process involving conversion of trifunctional entities. I was amazed to realize that just as we now know that what initially is generated is energy ( ~ ) , a chemiosmotic force and not a high energy phosphate compound. Lipmann also showed in his 1941 article that ~ is the main product of the metabolic wheel, and ~ Ρ is formed later.

138

Figure 3

Paul A. Srere

The symbol for yin and yang surrounded by the symbols for the eight elements.

Like the emperor Fu Hsi, Fritz Lipmann had divined the secrets of the universe. It was said of Fu Hsi, "Before his time, the people were like beasts, clothing themselves in skins and feeding on raw flesh, knowing their mothers but not their fathers". It can also be said of Fritz Lipmann, "Before his time, the biochemists were like beasts, clothing themselves with phenomenology and feeding on raw data, knowing their biology but not their physical chemistry".

2 Biochemistry Comes of Age

The history of biochemistry has been told profoundly and from a contemporary perspective by J . S . FRUTON (Molecules and Life, Wiley-Interscience, New York 1972), and the development of concepts and methods has been sketched in numerous essays in Comprehensive Biochemistry (Vol. 30 f, Elsevier Amsterdam 1972f.) and in the memoirs of many of its protagonists in Annual Reviews of Biochemistry (Palo Alto CA.). The preparation of enzyme active extracts of cells was a major methodological achievement that gave birth to biochemistry out of chemistry and biology. Soon the outlines of glycolysis in yeast and muscle were worked out, still the pièce de résistance of - or paradigm for - any biochemical education; physical chemistry in the form of colloid chemistry, kinetics and thermodynamics showed that biochemical reactions are governed by physico-chemical laws. Otto MEYERHOF was highly influential in the development of biochemistry by creating a school of critical experimentation, which was trained in the exact argumentation of its Master. In the early 1930's his institute became the nerve center of quantitative biochemistry, and much that is reported in the following pages had its beginnings directly or indirectly, there. If, in addition, the coworkers were well-trained in modern chemistry, as F. LIPMANN had been by Hans MEERWEIN in Königsberg, new concepts emerged: one such paradigm was the mechanism of reactions of energy-rich mixed anhydrides explained as their dissociation into (crypto)ionic species depending on the "solvent". As always - genius has to come at the right time. The tools had already been worked out; the concepts had been developed; as a result the strategies could be planned to tackle the gap between the structure and function of molecules in biochemistry-just as we now hope to bridge the gap between the structure and function of macromolecular aggregates on the one hand and organization in cellular biology, differentiation and evolution on the other. It is said in Chapter 1 that Fritz LIPMANN was essentially a lucky man, taking up at just the right time being in harmony with Nature's stratagems, and with an open receptive mind - well-trained in basic chemistry and enzymology at a time when mechanistic chemistry became amalgamated with enzymology. Already in 1930 he postulated that the "Pasteur effect" is the result of blocking the "glycolytic ferment" by electron withdrawal through a cytochrome. Ten years later he found that oxydative decarboxylation of pyruvic acid requires inorganic phosphate, yielding ATP, and he proposed as the intermediate an activated acetate then assumed to be acetyl phosphate - but shown, after the discovery of coenzyme A and its action, to be Feodor LYNEN'S acetyl CoA. Herman KALCKAR'S and Fritz LIPMANN'S influential reviews of 1941 discussed the energy coupling against the background of the discoveries of that time. He conceived the idea of "group potential" and "group transfer" and applied it not only to phosphate groups but to

140

2 Biochemistry Comes of Age

the transfer of other groups as well. About the same time C . CORYETT introduced the terms exergonic/endergonic for systems in which the chemical potential is the parameter, and S. RUBEN conjectured that photosynthesis is the EMBDEN-MEYERHOF cycle in reverse driven by photogenerated NADPH and ATP. In 1955 Fritz LIPMANN completed the KREBS urea cycle by showing that it starts with a carbamoyl group transfer to form citrulline from the regenerated ornithine. Thus, his contributions span the evolution of biochemistry over the past 60 years. It is with this background in mind that the exemplary essays in this chapter have been arranged. It begins by discussing how much biochemistry owes to the medical urinophilia of early times ( L . J . MOULTON and H . A . AKERS) and how its basic concepts were shaped by the enzymologists who discovered, using their specific methods and (pre)conceptions, the chains and cycles of glycolytic and pastglycolytic energy shuffling ( S . M . RAPOPORT, B . L . HORECKER); how much even misconceptions were essential for the development of its concepts (J. S. FRUTON), and how by comparative studies on higher and lower forms of life general molecular principles of metabolism ( L . O . KRAMPITZ) and of cofactor function (P. DIMROTH) were derived. The results were then applied to connect unknown forms of energy conversion to the basic patterns, such as the transformation of electron into photon fluxes (W.D.MCELROY). There was mutual influence: Enzyme chemistry also stimulated synthetic (thiol)chemistry (T. WIELAND) and natural products chemistry (J. BADDILEY); enzyme kinetics set the stage for unraveling by affinity labeling the mechanisms of complex biosynthetic reactions involving group activations and transfer (J. M. BUCHANAN); enzyme regulation proceeded from simple Michaelian control to protein turnover (S. GRISOLIA) and special cyclic processes were discovered that govern ω-peptidyl turnover triggered by ATP-energy ( K . BAUER). Fritz LIPMANN was not only pivotal in "classic" biochemistry, but also one of the leaders into what is now called "molecular biology". Its advances have overcome the boundaries between the fields of biochemistry, histology and cell physiology, to form one broad area. A main contribution was his work on protein kinases, unifying the understanding of the influence of site and time on the high energy state (H. HERRMANN and A . L . HISKES), linking structure to function (P. SIEKEVITZ) from muscle to receptors, from respiration to chromosomes, from morphogenesis to cancer (E. RACKER). He had the vision to proceed from the simple to the complex, and to reduce the complex to the simple. Ribosomal synthesis or large polypeptides and their splitting to small neuro- ( D . RICHTER) and activator peptides ( V . A . NAJJAR) has its origins in his research, as well as do studies on the thiotemplate mechanism (H. v. DÖHREN and H. KLEINKAUF) for specific microbial functionalized and functional oligopeptide factors ( K . KUBOTA) involved in sporulation - a process regulated by catabolic indicators through a chain of genes and their products (K. KURAHASHI).

There are to be sure several contributions included in this chapter that might as well have been included in one of the others, and the selection and assignment in some cases might seem a bit arbitrary. But this simplyshows all the more the broad scope of Fritz LIPMANN'S vision, the fertility of his mind, and his ability to traverse borders between different fields, in short: his genius.

The History of Metabolites Isolated from Urine L.J. Moulton and H.A.

Akers

During an association with Dr. Lipmann numerous occasions were available to question him on various aspects of biochemical history. Dr. Lipmann was quite aware of his role in the development of biochemistry and the fact that his life had overlapped those of all the other major biochemistry "heroes". He did not consider it his task, however, to become the historian of biochemistry, other than to record his own personal contributions. On several occasions he expressed his sadness to me in the lack of interest and awareness most fledgling biochemists had in the history of their own discipline. During afternoon tea Dr. Lipmann could be readily "tapped" for historical information. He not only enjoyed talking about an event or person but he also held in high regard the interest shown in the evolution of biochemistry. In one of these sessions the discussion topic was the fact that many metabolites were originally isolated from urine and that this feature had often been fossilized in the common names of these substances. After some discussion the group reached the consensus that: historically once the true origin of urine was known the next logical step must have been that urine was a reasonable place to look for metabolites. As he regularly said following such a discussion, "look into that". As will be seen below, the search for metabolites was not the genesis force behind the discovery of compounds in urine. In fact the error made in our discussion was the concept of a deliberate search for what we now call metabolites. This topic has always been of personal curiosity, resurfacing whenever a different uro-metabolite was encountered. Upon regularly teaching a yearlong introductory biochemistry course a different situation occurred. Students when confronted with the literal meaning of the names of substances such as kynurenine, urocanic acid, glucuronic acid, or hippuric acid were not fascinated or even curious about the names' origin. Rather they made statements to the effect of "another perverted biochemist". These ideas may be brought on in part by an unawareness of chamber pots and cloth diapers and the ample opportunities these once made for legitimate observations on excreta. Once a few comments are made on the history of plumbing and development of porcelain fixtures those erroneous ideas can be flushed. As Dr. Lipmann suggested the topic was "looked into". Urophilia, the love of urine, appears to have been common among early chemists. A superficial examination of their records reveals many experiments and observations which were made on urine and other bodily excretions. However, a closer look at the circumstances surrounding these experiments reveals that the pioneer chemists were not depraved men, but rather products of their times. The use of urine was a logical way to meet medical and economic needs.

142

L.J. Moulton and H.A. Akers

Since time immemorial, one of man's primary motivations has been the accumulation of wealth. Greed is one of the historical reasons for urine analysis. This is illustrated in the alchemist's work with urine, which was often used as a hopeful precursor to gold. "Gold is the most perfect of all metals," wrote the alchemist Bernard Trevisan (1). "In gold God has completed His work with the stones and rocks of the earth. And since man is nature's noblest creature, out of man must come the secret of gold". Thus urine was a common reagent in the laboratories of the neophite "chemists". Although the alchemists failed to produce gold from urine, they did accidentally convert urine to phosphorus (2). In 1669, Henning Brand (or Brandt) prepared phosphorus by "simply" allowing 50 to 60 pails of urine to "putrify" for 14 to 15 days. The liquid was then heated intensely for two or three days, after which water, Caput Mortuum, wine, and sand were added with more heating, and eventually phosphorus was recovered. Brand's is believed to be the first isolation of a chemical substance from urine. While the alchemists used urine in an attempt to gain economic power, others used it to obtain military power. Gunpowder at this time was a mixture of saltpeter, sulfur, and charcoal. Sulfur and charcoal were easily obtainable, but saltpeter was more difficult to procure. Saltpeter was often obtained by the oxidation of urine, with the oxidation of the nitrogen in the organic material carried out by soil microbes (3). Thus, during this time period the utilization of a country's urine in a large way determined its military prowess. Medicine has also historically been involved with the chemical use of urine, and it was the most common medical practice until the time of Harvey (4). Illness was believed to be caused by an imbalance of the humors-phlegm, blood, yellow bile, and black bile. Because urine was considered to be filtered blood, the medical diagnosis was made almost solely upon an examination of the patient's urine. Diseases were diagnosed according to the color, smell, quantity, and translucency of the urine. Because uroscopy was the major occupation of the physician, the urine flask was his pictorial symbol (5). Uroscopy alone is obviously an inadequate method of diagnosis for most disease. However, some genetic disorders can be correctly identified by primitive urine analysis. Of particular interest is porphyria, a metabolic disorder in which porphyrins are overproduced and build up in the brain. The abundance of porphyrins results in a wine-red urine. A recent examination of the medical records of King George III reveals that he suffered from porphyria, and the Mad King's insanity was caused by this metabolic disorder. The king's doctors noted that the illness was caused by "the force of a humor," and attacked his legs, bowels, and brain. The historians Ida MacAlpine and Richard Hunter wrote that: "Of all the king's symptoms, the most revealing one, which has now led us to a discovery of the true nature of his illness, is the color of his urine. At least half a dozen times the doctors who examined him noted that the king's urine was dark, red, and discolored (6)." Porphyria was evidently common in the royal houses of Europe, and even today many of King George Ill's descendants suffer from the disorder.

The History of Metabolites Isolated from Urine

143

Another metabolic malfunction which can be recognized because of the color of the patient's urine is alkaptonuria. It is characterized by the excretion of homogentisic acid, and results in black urine. Medical reports from as early as the sixteenth and seventeenth centuries made mention of people who produced dark urine (7). The first case of alkaptonuria which was identified with certainty occurred in 1859 (8). It was recognized that the patient's urine had unusual reducing properties, and the urine darkened when alkali was added. Because of the urine's ability to absorb oxygen in alkaline solution, the disorder was given the name alkaptonuria. If urine containing homogentisic acid is left to stand, it is oxidized and darkens, which is why the disease seems to have had a much higher occurrence before the invention of toilets. During the mid 1500's, Theophrastus Paracelsus, the Martin Luther of medicine, began advocating the use of chemistry in medicine. Paracelsus taught that "the changes which take place in the body are chemical, and the ills of the body must be treated by chemicals (1)." The analysis of urine remained the primary method of medical diagnosis, but for the first time it was examined by primitive chemical means, such as distillation. Another merging of chemistry and medicine came about in response to the urgent need of finding a cure for "the stone". Bladder stones had been a cause of suffering and death for centuries, but during the eighteenth century the affliction had reached "epidemic" proportions in Europe. Old surgical writings abound with accounts of victims of the stone; it afflicted all classes of society and age groups. Bladder stones were so prevalent that they were considered to be one of the most common reasons that infants cried (9). Several methods of "treating" the stone had evolved over the centuries. Both the Greeks and the "alchemical physicians" used alkaline substances internally on their patients in an effort to dissolve the stones in situ (10). Eighteenth century physicians also used alkaline solutions to treat the stone, and their recipes included the injection of a limewater and soap solution directly into the bladder. Other remedies that were taken orally were made from a combination of egg shells, garden snails, and vegetables. It is doubtful that these antidotes were beneficial, and it is likely that such remedies caused more suffering than they alleviated (11). In addition to the alkaline treatments, primitive methods of surgical removal of the stones existed, but the eighteenth century patient submitted to the surgery only as a last resort. There were no adequate anesthetics available, and the mortality rate was extremely high (12). Another alternative was lithotrity, in which the stones were crushed in situ by introducing instruments into the bladder without surgery. "Electrolysis" was also used in an effort to "dissolve" the stone (13). None of these methods produced satisfactory results, and it was believed that the most likely means of curing the bladder stone rested with chemists. As stated by Wollaston: "If in any case a chemical knowledge of the effects of disease will assist us in the cure of them, in none does it seem more likely to be of service than in the removal of the several concretions that are formed in various parts of the body (14)."

144

L.J. Moulton and H.A. Akers

As the fields of chemistry and medicine merged, chemists were commissioned by both Societies and individuals to find a method of dissolving the stone in vivo. It was assumed that once the techniques were found to dissolve the excreted calculi, the same methods could then be utilized to dissolve the stones remaining in the bladder. The chemical study of the bladder stone was also expected to produce a knowledge of the cause of the stones as well as lead to prevention and cure. The methods used to analyze the stones were primitive. George Pearson (15) recorded some of the experiments performed to "investigate the properties of the constituent parts of urinary concretion." In one of his experiments, five hundred calculi were pulverized. They were then "triturated" with five ounces by weight of "lye of caustic soda." The substance was then digested, boiled, and filtered. Pearson recovered a tasteless, white, heavy powder and noted that the filtered liquid tasted weakly of soda, and was both heavy and soapy to the feel. Taste, smell, solubility in the mouth, weight, color, and consistency were primary means of chemical classification. The results of work by Pearson and others, however, was not a cure for the bladder stone. Rather, it led to the isolation of several new compounds. The first discovery was made by Scheele in 1776 who reported finding a "new acid principle" in urinary calculi. Scheele noted that the acid was sparingly soluble in water and it dissolved rapidly in alkaline solutions. It turned pink or purple upon the addition of nitric acid. Scheele found this same compound in human urine, and he believed that all urinary calculi were composed of the salts of this acid. Scheele's compound was named lithic (stone) acid. The name was later changed by Fourcroy to uric acid (16). Another result of the chemist's work with bladder stones was the isolation of cystine. Cystine stones were excreted by victims of the metabolic disorder cystinuria. The illness is caused by the impaired renal absorption of several amino acids including cystine, which are excreted in the urine. The stones were first identified by Wollaston in 1810 (17). He believed that the stones were made of an oxide because they reacted with both acids and bases. Because the stones occurred in the bladder, Wollaston named the compound cystic oxide ("kystus" is Greek for bladder). Later Berzelius stated that the substance was not an oxide, and he suggested the name cystin. It is interesting to note that the cystine utilized for early metabolism studies was provided by cystinuria patients (18). Because they recognized that urine was the matrix of stone formation, chemists began analyzing animal stones and urine samples and compared them to samples obtained from humans. The following table lists compounds which were first isolated from either bladder stones or urine as a result of the search to find a way to dissolve the stones in vivo. These compounds were isolated using techniques similar to those used to obtain uric acid and cystine. This is not intended to be a comprehensive list. In 1804, Fourcroy and Vauquelin began a classification of calculi and urine from various animals (24). It was noted that herbivores had stones which consisted mainly of carbonate of lime. Carnivores such as lions and tigers were shown to have urine that was made up almost entirely of urea, with very little uric acid. It was also observed that snakes excreted large amounts of uric acid, but no urea. From this

The History of- Metabolites Isolated from Urine

145

Table 1 Substances originally isolated from urine or stones during searches made to find ways to dissolve stones Chemical name Original name

Source

Discoverer/Date

Ref.

Urea

native salt of urine

human urine

(19)

Uric acid Cystine Zanthine Murexane Alloxane Murexide

lithic acid cystic oxide zanthic oxide purpuric acid * erythric acid ammonium purpurate *

bladder bladder bladder bladder bladder bladder

Boerhaave, 1727 Rouelle, 1773 Scheele, 1776 Wollaston, 1810 Marcet, before 1817 Prout, 1818 Brugnatelli, 1818 Prout, 1820

stone stone stone stone stone stone

(16) (17) (20) (21) (22) (23)

* Derived from a uric-nitric acid mixture.

comparative data originated the concept that urea is predominantly the endpoint of amino acid metabolism in mammals, and uric acid the final product in bird and snake amino acid metabolism. This illustrates that the movement to find a cure for the bladder stone not only stimulated an interest in the components of urine, but it also led to work in other areas of biochemistry. Hippuric acid is an example of a compound which, after its isolation from urine, yielded results in the area of metabolic pathways, leading eventually to the discovery of substances such as ornithine or processes such as beta-oxidation. Hippuric acid contains a benzoic acid moiety, and the early researcher, using mainly solubility information, could not diiferentiate between benzoic and hippuric acids. In 1773, Rouelle noted the presence of "benzoic acid" in cow urine and later in camel urine (25). Fourcroy and Vauquelin in 1799 also observed "benzoic acid" in the urine of herbivorous animals (26). The first researcher to determine the fate of an ingested foreign substance was Wohler in 1824 (27). He fed a dog benzoic acid and isolated from its urine a compound which had the solubility properties of benzoic acid. In 1829, Liebig isolated hippuric acid from horse urine (28). The compound was named "hippuric acid" because it was first recognized in the urine of a horse, and "hippos" is the Greek word for horse. Liebig was the first to notice that "benzoic acid" derived from urine contained nitrogen. The first documentation of the biotransformation of a foreign compound was made by Uri in 1841 (29). After administering benzoic acid to himself and to other human volunteers, he recovered large quantities of hippuric acid in the urine collections. Because he also observed a decrease in uric acid production, he postulated that uric acid was a precursor of hippuric acid. As a result, he suggested the use of benzoic acid to treat gout, which is caused by the chalky deposits of uric acid. Thus, the study of urine led to both the first knowledge of a biotransformation, and the first knowledge of the fate of a foreign substance. The study of urine also served as a primitive means of medical diagnosis, a hopeful cure for the bladder stone, and the discovery of new compounds and elements. In addition, urine was used as a source of military strength, and was a legendary means of attaining instant wealth.

146

L.J. Moulton and H.A. Akers

Urine was studied as a logical source of answers to problems of the particular time period. The pioneer scientists who analyzed urine were not urophiliacs, but instead were products of their time. They were men, like the man honored by this volume, to whom we owe a great deal.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29.

Jaffe, B. 1930. In: Crucibles. Tudor, New York. Weeks, M.E. 1945. In: Discovery of the Elements. Published by J. Chem. Educ., Easton, PA. Peterson, H.L. 1964. In: Encyclopedia of Firearms. Connoisseur, London. Debus, A.G. 1966. In: The English Paracelsians. Franklin Wats, New York. Copeman, W.S.C. 1960. In: Doctors and Diseases in Tudor Times. Dawson's of Pall Mall. MacAlpine, I., R. Hunter. 1969. Sci. Am. 221, 38. Garrod, A.E. 1908. Lancet 2, 73. Boedeker, C. 1859. Z. Rationelle Medicin 7, 130. Ellis, H. 1969. In: History of the Bladder Stone. Blackwell, Oxford. Guerlac, H. 1956. Isis 48, 124 and 433; Viesltear, A.J. 1969. Bull. Hist. Med. 43, 477. Coxe, W. 1816. In. Memoirs of the Life and Administration of Sir Robert Walpole, Earl of Oxford. Vol. IV, London. Dickinson, W.B. 1821. Med. Chir. Trans. 11, 61. Stark, W. 1815. Ann.Chim. 5, 614. Wollaston, W.H. 1797. Phil.Trans. 87, 386. Pearson, G. 1798. Phil. Trans. 88, 15. de Fourcroy, A.F. 1798. Ann. Chim. Phy. 27, 225. Wollaston, W.H. 1810. Trans. Royal Soc. London 100, 223. Loeury, A. and C. Neuberg. 1904. Physiol. Chem. 43, 338. Kurzer, F. and P.M. Sanderson. 1956. J. Chem.Educ. 33, 452. Marcet, A. 1817. In: An Essay on the Chemical History and Medical Treatment of Calculous Disorders. London. Prout, W. 1818. Phil.Trans. 108, 420. Brugnatelli, G. 1818. Ann.Chim. 8, 201. Prout, W. 1819. Ann. Phil. 14, 363. de Fourcroy, A.F. and N. Vauquelin. 1804. Ann.Mus. 4, 329. Berzelius, J. 1840. Lehrbuch der Chemie, Vol.9, 3rd ed., Dresden; Hallwachs, W. 1858. Ann. Chem. Pharm. 105, 207. de Fourcroy, A.F. and L.N. Vauquelin. 1799. Ann.Chim.Phys. 31, 63. Wöhler, F. 1824. Tiedemann's Z. Physiol. 1, 142. Liebig, J. 1829. Poggendorffs Ann.Phys. Chem. 17, 389. Ure, A. 1841. Pharm. J. Transact. 1, 24.

The Pentose Phosphate Pathway B.L. Horecker

Introduction Fifty years ago most of the key enzymatic reactions accounting for the conversion of glucose to ethanol and C 0 2 in yeast and to lactic acid in muscle had been described, primarily in the laboratories of Otto Warburg in Berlin-Dahlem and of Otto Meyerhof in Heidelberg. Respiration, the uptake of oxygen by cells and tissue extracts, was also under intensive investigation, notably by Keilin at the Molteno Institute in Cambridge, by Warburg in Dahlem, and by Theorell in Stockholm. The role of cytochrome C and cytochrome oxidase in oxidative metabolism was firmly established. The coenzyme of fermentation, cozymase, had been discovered and identified as a dinucleotide containing one equivalent of nicotinamide mononucleotide and one of adenylic acid linked together by a pyrophosphate bond. Cozymase was clearly the coenzyme of fermentation in yeast and glycolysis in muscle. It carried electrons from the oxidation of glyceraldehyde 3-phosphate for the reduction of acetylaldehyde to ethanol in yeast or pyruvate to lactic acid in muscle. Cozymase was later named diphosphopyridine nucleotide, or DPN, when its structure was established. A second related coenzyme, isolated and characterized in Warburg's laboratory (1), contained an additional phosphate group on the 2' position of the adenylate moiety; this coenzyme, called triphosphopyridine nucleotide, or TPN 1 , would not replace DPN in glycolysis or fermentation but specifically catalyzed the oxidation of one of the intermediates of glycolysis, glycose 6-phosphate, to 6-phosphogluconate (1). Warburg also demonstrated that this product was further metabolized ("fermented") by yeast extracts in a process that required the same coenzyme and resulted in the production of C 0 2 . Lipmann's attention was attracted to this pathway because of his interest in the phenomenon, described by Pasteur seventy years earlier (2), of the inhibition of fermentation in air, the so-called Pasteur Effect. Lipmann had found that oxidizing agents such as iodine, quinone, and 2,6-dichloro-phenolindophenol plus oxygen reduced the rate of lactic acid production from glucose by muscle extracts (3) and also inhibited the fermentation of glucose in yeast maceration juice (4). In his search for an alternate oxidative pathway for the metabolism of carbohydrate, Lipmann was attracted to Warburg's observation of the oxidation of glucose 6-phosphate to 6-phosphogluconate. He demonstrated that the production of C 0 2 from 6-

1

In 1964 the IUB Enzyme Nomenclature Commission proposed the new names "nicotinamide adenine dinucleotide" (NAD) and "nicotinamide adenine dinucleotide phosphate" (NADP) as more appropriate for these coenzymes.

148

B.L. Horecker

phosphogluconate by yeast macerate could not be attributed to fermentation, since it occurred only in the presence of oxygen (5). This process was not inhibited by bromoacetate, which inhibited alcoholic fermentation, suggesting that "phosphogluconic acid might be a first product of carbohydrate oxidation occurring in a manner different from fermentative breakdown" (5). Lipmann suggested that oxidation of 6-phosphogluconate to the a-keto acid followed by decarboxylation of the latter would yield the 5-carbon product, arabinose phosphate. A few years later, Dickens demonstrated that the products formed in the oxidation of 6-phosphogluconate by N A D P included a phosphate ester giving the orcinol reaction for pentoses (6), and proposed that the oxidation of carbohydrate might proceed by a series of oxidations and decarboxylations leading stepwise from hexose phosphate through pentose phosphate and tetrose phosphate to trióse phosphate. At this time it appeared that N A D was indeed the coenzyme of fermentation, whereas N A D P served as the coenzyme for an alternative oxidative pathway for the metabolism of carbohydrate, the "hexose monophosphate shunt" of Engelhardt and Barkash (7). Only later, from the work of Lipmann, Krebs, Lehninger, and others, did it become clear that the coenzyme for mitochondrial respiration and for the production of high energy phosphate was N A D , and not NADP, which is now considered to function as the coenzyme for reductive biosynthetic processes. N A D P H , generated in the oxidation of glucose 6-phosphate and also in illuminated plant chloroplasts (8), provides a major source of reducting power, particularly in animals and higher plants. Thus, the hexose monophosphate shunt proved to be not the major pathway for the oxidation of carbohydrate, but rather a mechanism for the production of 5carbon sugars for the synthesis of nucleotides and nucleic acids and of N A D P H for reductive biosynthesis, e.g., for the formation of lipids and steroids, and also for detoxifications via the cytochrome P 4 5 0 system.

The Oxidative Branch of the Pentose Phosphate Pathway As indicated earlier, Lipmann's suggestion that a pentose phosphate would be the first product formed in the oxidative decarboxylation of 6-.phosphogluconate was supported by experiments carried out by Dickens with partially purified yeast extracts. Among the products formed in the oxidation of 6-phosphogluconate by N A D P was a phosphate ester giving the orcinol test for pentose. Dickens observed, however, that with these same extracts ribose 5-phosphate was much more rapidly metabolized than was arabinose 5-phosphate, making it unlikely that arabinose 5phosphate was an intermediate in hexose monophosphate oxidation (6). The proposal that ribose 5-phosphate was a product of 6-phosphogluconate oxidation by yeast extracts, later confirmed by Scott and Cohen (9) using paper chromatography and a ribose-requiring mutant strain of E. coli, posed a new problem, namely how to account for the inversion of configuration at the 3-carbon atom of glucose. The solution came in 1951, when P.Z. Smyrniotis, J.E. Seegmiller,

The Pentose Phosphate Pathway

149

and I, using a purified preparation of 6-phosphogluconate dehydrogenase from brewers' yeast which yielded pentose phosphate from 6-phosphogluconate and NADP in stoichiometric quantities (10), identified ribulose 5-phosphate as the primary product (11). This initial product of the oxidative decarboxylation of 6phosphogluconate was slowly converted by traces of ribose 5-phosphate isomerase, shown to be present in the' enzyme preparation, to an equilibrium mixture containing 25% of ribulose 5-phosphate and 75% of ribose 5-phosphate. The same stoichiometric conversion was observed by Seegmiller in my laboratory with an enzyme preparation from bone marrow (12). The discovery of ribulose 5-phosphate as the primary product of the oxidative decarboxylation of 6-phosphogluconate suggested that oxidation occurred in the 3-position rather than in the 2-position of the 6-carbon substrate, and fit the pattern developed in Ochoa's laboratory for other decarboxylations by "double-headed" enzymes such as isocitric dehydrogenase and malic dehydrogenase (13 and references cited therein). The fact that all of these oxidative decarboxylations were reversible, resulting in the fixation of C 0 2 , led us for a brief while to wonder whether the reversal of the oxidative decarboxylation of 6-phosphogluconate might not be the long-sought mechanism for C 0 2 fixation of photosynthesis. We were indeed able to demonstrate the formation of 6-phosphogluconate from pentose phosphate and C 0 2 (14). However, the reduction of 6phosphogluconate to glucose 6-phosphate posed a problem which was solved by the demonstration by Cori and Lipmann (15) that the first product of glucose 6phosphate oxidation was not the free carboxylic acid, but rather the 6phosphoglucono-delta-lactone. We were then able to demonstrate the reversibility of this product as well (16). For this mechanism of fixation of C 0 2 in glucose 6phosphate to be feasible required the action of another enzyme discovered soon thereafter by Lipmann and Brodie (17), namely a lactonase, that catalyzed the interconversion of the ¿-lactone and the free acid. However, this pathway for the fixation of C 0 2 in photosynthesis was not consistent with the observations emerging from Calvin's laboratory, which established 3-phosphoglycerate as the early product of photosynthetic C 0 2 fixation (18). At this point the following reactions appeared to account for the conversion of glucose 6-phosphate to pentose phosphate: Glucose 6-P + NADP ^ á-gluconolactone 6-P + NADPH (5-Gluconolactone 6-P ^ 6-P-gluconate 6-P-Gluconate + NADP ribulose 5-P + C 0 2 Ribulose 5-P ^ ribose 5-P

The Non-oxidative Branch of the Pentose Phosphate Pathway Dische (19), and later Valvogel and Schlenk (20), had demonstrated that ribose 5phosphate on incubation with erythrocyte lysates or liver extracts was converted to trióse and hexose phosphates, suggesting that the oxidation of glucose 6-phosphate

150

B.L. Horecker

might be part of a cyclic pathway. In our laboratory Seegmiller had indeed observed that the pentose phosphate formed in the oxidation of 6 phosphogluconate by a bone marrow preparation was further metabolized with the appearance of glucose 6-phosphate. Thus, in the presence of a mechanism for regenerating NADP from NADPH (we used pyruvate and lactic dehydrogenase), the pentose phosphate pathway could indeed function as a cyclic mechanism. The presence of trióse phosphates and fructose 1,6-bisphosphate among the products of ribose 5-phosphate metabolism (19) suggested that the first step in the conversion of pentose phosphate to hexose phosphate would be cleavage of the 5carbon chain to yield a 2-carbon fragment at the oxidation level or glycolaldehyde, with the remaining 3-carbon fragment, probably a mixture of glyceraldehyde 3phosphate and dihydroxyacetone phosphate, giving rise to hexose monophosphate by way of fructose 1,6-bisphosphate. This mechanism was, however, rendered unlikely by the observation of Dische (20) that the red cell hemolysates that catalyzed the conversion of pentose phosphate to hexose phosphate did not convert fructose 1,6-bisphosphate to fructose 6-phosphate. The formation of hexose monophosphate from trióse phosphate alone was also difficult to reconcile with the observation of Glock (22), working in Dickens' laboratory, that the yields of hexose monophosphate from ribose 5-phosphate exceeded the 60% predicted if only 3 of the 5 pentose carbon atoms were converted to hexose. In addition, despite attempts in many laboratories, it proved impossible to isolate the 2-carbon fragment or a product formed from that fragment. In order to characterize the primary product of metabolism of the pentose phosphates, Smyrniotis, Klenow and I purified the so-called "cleaving enzyme" from liver, following the cleavage reaction by measuring the disappearance of pentose phosphate using the orcinol reaction. Using the same reaction, we detected a new orcinol-reactive product which we identified as the 7-carbon sugar ester, sedoheptulose 7-phosphate (23), which Calvin and his coworkers had identified as an early product of C 0 2 fixation in photosynthesis (24). We found that spinach leaves were an excellent source of the pentose phosphate cleaving enzyme. Significantly, with crude extracts of both liver and spinach, sedoheptulose 7phosphate was formed as an early product but later disappeared with the accumulation of hexose monophosphate (25). Clearly, sedoheptulose 7-phosphate was an intermediate in the conversion of pentose phosphate to hexose monophosphate. The formation of sedoheptulose 7-phosphate from pentose phosphate was the result of the transfer of a C 2 group from one pentose phosphate to another, and the other product of the reaction was identified as glyceraldehyde 3'-phosphate (23). This C 2 transfer mechanism was also proposed by Racker and his colleagues (26) when they found that a crystalline enzyme preparation from yeast would catalyze the transfer of a C 2 -fragment from hydroxypyruvate to glyceraldehyde 3-phosphate. They named this enzyme transketolase because it catalyzed the transfer of a ketol-linkage from one molecule to another. The cofactor for the transfer of the C 2 -fragment was shown to be thiamine pyrophosphate (26,27). With a purified rat liver transketolase preparation, using a mixture of ribose 5phosphate and ribulose 5-phosphate as the substrate, we had observed that the

The Pentose Phosphate Pathway

151

reaction required the presence of crystalline muscle aldolase (23). At the time our explanation was that the transfer was freely reversible and that in order to accumulate sedoheptulose 7-phosphate it was necessary to remove the other product, glyceraldehyde phosphate, which was the function provided by aldolase. The true explanation emerged after a new ketopentose ester, identified as xylulose 5phosphate, was isolated by Ashwell and Hickman (28), as a product of ribose 5phosphate metabolism with a spleen preparation. Xylulose 5-phosphate was shown by Racker and his coworkers (29) to be the true substrate for transketolase. In our laboratory the enzyme that catalyzed the interconversion of xylulose 5-phosphate and ribulose 5-phosphate was purified by Paul Stumpf while he was on a sabbatical leave in my laboratory (30), and this enzyme was found by us to be a contaminant of the usual crystalline preparations of rabbit muscle aldolase (31). Thus, the requirement for aldolase in the conversion of pentose phosphate to sedoheptulose phosphate was explained by Racker's discovery that the true substrate for transketolase was xylulose 5-phosphate and not ribulose 5-phosphate, and our finding of xylulose 5-phosphate 3-epimerase activity as a contaminant of crystalline aldolase. Finally, the finding that fructose 6-phosphate was also a substrate for transketolase (32) confirmed that all of the substrates for this enzyme possessed the same steric configuration at carbon atoms 3 and 4. This last observation also paved the way for the elucidation of the final step in the conversion of pentose phosphates to hexose monophosphates (see below). Experiments reported in 1952 at the Second Symposium on Phosphorus Metabolism at Johns Hopkins University (25), carried out with a crude rat liver preparation, provided the first evidence that sedoheptulose 7-phosphate was indeed an intermediate in the pathway from pentose phosphate to hexose monophosphate. The disappearance of pentose phosphate was accompanied by the accumulation of the 7carbon sugar ester, which then slowly decreased in quantity as the hexose monophosphates accumulated. Significantly, however, when sedoheptulose 7phosphate alone was added to these extracts, it was not consumed and the formation of monophosphates was not detected. The explanation came when it was discovered that a source of trióse phosphate was also required for the conversion of sedoheptulose 7-phosphate to fructose 6-phosphate; this source could be either pentose phosphate or fructose 1,6-bisphosphate (25,33). Experiments with labeled substrates established that the conversion involved the transfer of a C 3 group from the sedoheptulose 7-phosphate to glyceraldehyde 3phosphate, yielding fructose 6-phosphate and leaving the remaining four carbon atoms of sedoheptulose 7-phosphate as the free tetrose ester, erythrose 4-phosphate. The enzyme that catalyzed the transfer of the dihydroxyacetone group from sedoheptulose 7-phosphate to glyceraldehyde phosphate was named transaldolase. In these early experiments erythrose 4-phosphate was identified on the basis of its conversion to sedoheptulose 1,7-bisphosphate catalyzed by aldolase present in the rat liver enzyme preparations (34). The 4-carbon sugar phosphate was later synthesized by Ballou and coworkers (35) and confirmed as a substrate for transketolase. As indicated earlier, Racker and his coworkers (26,36) had already demonstrated the cleavage of fructose 6-phosphate by transketolase and obtained indirect evidence for the formation of erythrose 4-phosphate as one of the products.

B.L. Horecker

152

The reversal of this reaction, namely the transfer of a C 2 group from xylulose 5phosphate to erythrose 4-phosphate, would yield a second equivalent of fructose 6phosphate. The conversion of pentose to hexose phosphates can therefore be expressed by the following series of reactions: transketolas

1.

2 Pentose Ρ

2.

Sedoheptulose Ρ + glyceraldehyde Ρ

3.

Pentose Ρ + erythrose Ρ 3 Pentose Ρ ^

^ sedoheptulose Ρ + glyceraldehyde Ρ

transketola

transaldolas

- fructose Ρ + erythrose Ρ

^ fructose Ρ + glyceraldehyde Ρ

2 fructose Ρ + glyceraldehyde Ρ

This series of reactions, constituting the non-oxidative pentose phosphate pathway, is readily reversible, and it could be shown that when fructose 6-phosphate is added to rat liver extracts sedoheptulose 7-phosphate is produced (37,38). The primary functions of this non-oxidative branch of the pentose phosphate pathway are: 1. to generate pentose phosphates from hexose monophosphates, or 2. to convert any excess of pentose phosphate produced in the oxidative branch of the cycle, resulting from a requirement for large quantities of reducing power (NADPH), back to hexose monophosphate. The non-oxidative mechanism for the interconversion of pentose phosphate and hexose phosphate also serves to regenerate the ribulose 1,5bisphosphate which is the primary substrate for C 0 2 fixation in photosynthesis (for a review see (39).

Schiff-base Mechanisms As indicated earlier, in the reactions catalyzed by transketolase, enzyme-bond thiamine pyrophosphate serves as the coenzyme; the C 2 group is transferred from the donor substrate to the enzyme-bond cofactor, thiamine pyrophosphate, and from this complex to an acceptor aldehyde to form the new product (reviewed in 40). The structure of the C 2 -thiamine pyrophosphate adduct has been established by the work of Breslow, Krampitz and Holzer and their associates as a dihydroxyethyl group covalently linked to the 2-carbon atom of the thiazole ring, analagous to the hydroxyethyl group formed in the oxidation of pyruvate (41,42). For the C 3 transfer catalyzed by transaldolase no corresponding cofactor could be detected. The fact that the reaction required the presence of both donor and acceptor substrates suggested that the reaction mechanism involved transfer of the dihydroxyacetone group from the substrate to the enzyme and then to the donor. Evidence for the formation of such a complex was obtained by Venkataraman and Racker (43) with crystalline rat liver transaldolase, and by Sandro Pontremoli and Carlo Ricci in my laboratory with crystalline transaldolase from Candida utilis (44). In each case the substrate was 14C-labeled fructose 6-phosphate, and the enzyme complex was separated from the free substrate either by chromatography (43) or by ammonium sulfate precipitation (44). In the latter case it could be demonstrated

The Pentose Phosphate Pathway

153

that a true immediate was formed, because on addition of erythrose 4-phosphate to the enzyme dihydroxyacetone complex the reaction was completed and radioactive sedoheptulose 7-phosphate was isolated and identified (45). Because of the lability of the dihydroxyacetone-enzyme adduct, elucidation of its structure proved difficult. The key came from experiments in which the adduct was reduced, at slightly acid pH, with sodium borohydride, which stabilized the enzyme complex. Significantly also, the catalytic activity was abolished, suggesting that the reduced dihydroxyacetone residue was linked to an amino acid at the active site (45). A similar enzyme-linked complex was formed when crystalline rabbit muscle aldolase was treated with sodium borohydride in the presence of dihydroxyacetone phosphate (46,47). Early model experiments carried out in Westheimer's laboratory (48) and later by Speck and Forist (49) had shown that dealdolization reactions were catalyzed by primary and secondary amines, probably by way of ketamine of Schiffbase intermediates, and in 1962 the first evidence was obtained for formation of a Schiff-base derivative involving the carbonyl group of dihydroxyacetone phosphate and the ε-amino group of a lysine residue. The definitive evidence for formation of Schiff-base intermediates with both aldolase and transaldolase was published in the Festschrift dedicated to Otto Warburg on the occasion of his 80th birthday (50). In each case after acid hydrolysis of the sodium borohydride-reduced complex the amino acid derivative recovered was identified as N6-/?-glyceryllysine. The Schiff-base mechanism has been extended to other aldolases and was employed by Rutter (51) to discriminate between the Class I aldolases, which formed the Schiff-base intermediates, and Class II aldolases, which contained metal cofactors and which were not inactivated by reduction with borohydride in the presence of substrate. The identification of lysine as the reactive group at the active sites of both transaldolase and fructose 1,6-bisphosphate aldolase left unanswered two important questions: 1. why was one particular lysine residue among the many lysine residues present in each enzyme subunit peculiarly reactive, and 2. why was the dihydroxyacetone phosphate Schiff-base formed with aldolase readily dissociable, in contrast to the dihydroxyacetone phosphate Schiff-base formed with transaldolase, which did not spontaneously dissociate and which could not be formed with free dihydroxyacetone, but only by transfer from a donor substrate? The participation of other residues in binding of the substrates to the enzyme and in the aldolizationdealdolization has been established for both types of enzymes. For aldolase, these additional residues may include the COOH-terminal tyrosine, cysteine sulfhydryl groups, an additional lysine residue more than 100 residues removed from the Schiff-base forming lysine residue, and a histidine residue (reviewed in 52). Significantly, photooxidation of a histidine residue converts aldolase to an enzyme that catalyzes the transfer of the dihydroxy-acetone phosphate group (e.g., a transaldolase), but not the dealdolization reaction (53). The x-ray crystallographic structure of rabbit muscle aldolase has now been reported (54) and residues at the active center identified. Finally, for the transaldolase of Candida utilis, a histidine residue appears to stabilize the dihydroxyacetone Schiff-base through hydrogen binding between the imidazole nitrogen and the Schiff-base carbanion (55).

154

B.L. Horecker

Acknowledgement I o w e a great personal debt to the German school o f biochemistry that fluorished during the 1920's and early 1930's. M y chief mentor was Erwin Haas, w h o came to the University of Chicago from Otto Warburg's laboratory in 1939. H e taught m e respect for science, " h o w to work" in the laboratory, to carefully design the experiments, and to record in detail the results o f each day's activities. F r o m Lipmann came the inspiration and valuable clues that led me to initiate the studies o n the pentose phosphate pathway. O t t o Meyerhof discovered aldolase and provided important information pertaining to its specificity requirements. A s s o m e o n e has said: "We can see so far because w e stand o n the shoulders of giants".

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.

Warburg, O., W. Christian and A. Griese. 1935. Biochem. Z. 282, 157. Pasteur, L. 1861. Comptes Rendu 52, 1260. Lipmann, F. 1933. Biochem. Z. 265, 133. Lipmann, F. 1933. Biochem. Z. 265, 205. Lipmann, F. 1936. Nature 138, 588. Dickens, F. 1938. Biochem. J. 32, 1626. Engelhardt, W.A. and A. P. Barkash. 1938. Biokhimiya 3, 500. Jagendorf, A. 1956. Arch. Biochem. Biophys. 62, 141. Scott, D.B.M. and S.S. Cohen. 1951. J. Biol. Chem. 188, 509. Horecker, B.L. and P.Z. Smyrniotis. 1951. J. Biol. Chem. 193, 371. Horecker, B.L., P.Z. Smyrniotis and J.E. Seegmiller. 1951. J. Biol. Chem. 193, 383. Seegmiller, J.E. and B.L. Horecker. 1952. J. Biol. Chem. 194, 261. Ochoa, S., A.H. Mehler and A. Romberg. 1948. J. Biol. Chem. 174, 979. Horecker, B.L. and P.Z. Smyrniotis. 1952. J. Biol. Chem. 196, 135. Cori, O. and F. Lipmann. 1952. J. Biol. Chem. 194, 417. Horecker, B.L. and P.Z. Smyrniotis. 1953. Biochim. Biophys. Acta 12, 98. Brodie, A.F. and F. Lipmann. 1955. J. Biol. Chem. 212, 677. Calvin, M. and P. Massini. 1953. Experientia 8, 445. Dische, Z. 1938. Naturwiss. 26, 252. Waldvogel, M.J. and F. Schlenk. 1947. Arch. Biochem. Biophys. 14, 484. Dische, Z. 1949. Abstr., 1st Internati. Cong. Biochem., Cambridge. Glock, G.E. 1952. Biochem. J. 52, 575. Horecker, B.L., P.Z. Smyrniotis and H. Klenow. 1953. J. Biol. Chem. 205, 661. Benson, A.A., J.A. Bassham and M. Calvin. 1951. J. Am. Chem. Soc. 73, 2970. Horecker, B.L. 1952. In: Phosphorus Metabolism II (W.D. McElroy and B. Glass, eds.) Hopkins, Baltimore, pp. 460-461. Racker, E., G. de la Haba and I. G. Leder. 1953. J. Am. Chem. Soc. 75, 1010. Horecker, B.L. and P.Z. Smyrniotis. 1953. J. Am. Chem. Soc. 75, 1009. Ashwell, G. and J. Hickman. 1954. J. Am. Chem. Soc. 76, 5889. Srere, P.A., J. Cooper, V. Klybas and E. Racker. 1955. Arch. Biochem. Biophys. 59, 535. Stumpf, P.K. and B.L. Horecker. 1956. J. Biol. Chem. 218, 753. Horecker, B.L., J. Hurwitz and P.Z. Smyrniotis. 1956. J. Am. Chem. Soc. 78, 692. Racker, E., G. de la Haba and I. G. Leder. 1954. Arch. Biochem. Biophys. 48, 238. Horecker, B.L. and P.Z. Smyrniotis. 1953. J. Am. Chem. Soc. 75, 2021. Horecker, B.L., P.Z. Smyrniotis, H. Hiatt and P. Marks. 1955. J. Biol. Chem. 212, 827. Ballou, C.E., H.O.L. Fischer and D.L. McDonald. 1955. J. Am. Chem. Soc. 77, 2658. de la Haba, G., I.G. Leder and E. Racker. 1955. J. Biol. Chem. 214, 409. Bonsignore, Α., S. Pontremoli, A. de Flora and B.L. Horecker. 1961. Ital. J. Biochem. 10,106.

The Pentose Phosphate Pathway

155

38. Pontremoli, S., A. Bonsignore, E. Grazi and B.L. Horecker. 1960. J. Biol. Chem. 235, 1881. 39. Vischniac, W., B.L. Horecker and S. Ochoa. 1957. In: Advances in Enzymology, Vol. XIX. Academic Press, New York, p. 1. 40. Horecker, B.L. 1959. J. Cell. Comp. Physiol. 54, Supp. 1, 137. 41. Krampitz, L.O., I. Suzuki and G. Gruell. 1961. Fed. Proc. 20, 971. 42. Holzer, H., R. Katterman and D. Busch. 1962. Biochem. Biophys. Res. Commun. 7, 167. 43. Venkataraman, R. and E. Racker. 1961. J. Biol. Chem. 236, 1883. 44. Pontremoli, S., B.D. Prandini, A. Bonsignore and B.L. Horecker. 1961. Proc. Natl. Acad. Sci. USA 47, 1942. 45. Horecker, B.L., S. Pontremoli, C. Ricci and T. Cheng. 1961. Proc. Natl. Acad. Sci. USA 47, 1949. 46. Grazi, Ε., T. Cheng and B.L. Horecker. 1962. Biochem. Biophys. Res. Commun. 7, 250. 47. Grazi, Ε., P.T. Rowley, T. Cheng, O. Tchola and B.L. Horecker. 1962. Biochem. Biophys. Res. Commun. 9, 38. 48. Westheimer, F.H. and H. Cohen. 1938. J. Am. Chem. Soc. 60, 90. 49. Speck, J.C. and A.A. Forist. 1957. J. Am. Chem. Soc. 79, 4659. 50. Horecker, B.L., P.T. Rowley, E. Grazi, T. Cheng and O. Tchola. 1963. Biochem. Z. 338, 36. 51. Rutter, W.J. 1964. Fed. Proc. 23, 1248. 52. Lai, C.Y., N. Nakai and D. Chang. 1974. Science 183, 1204. 53. Hoffee, P., C. Y. Lai, E.L. Pugh and B.L. Horecker 1967. Proc. Natl. Acad. Sci. USA 57, 107. 54. Sygusch, Beaudry and Allaire. 1988. Proc. Natl. Acad. Sci. USA. In Press. 55. Brand, K., O. Tsolas and B.L. Horecker. 1969. Arch. Biochem. Biophys. 130, 521.

Glycolysis and the Dawn of Modern Biochemistry S.M.

Rapoport

Introduction Nowadays it has become customary to date the revolution in modern biology from the discovery of the DNA double helix. To my mind this is an one-sided view which overemphasizes structure and neglects the importance of dynamic aspects. Synthesis, repair and modification of DNA as well as bioengineering cannot be understood or manipulated without enzymes. The same holds for the machinery of gene expression. In addition to these considerations one should appreciate the fact that the pioneers of modern molecular biology, among them Lipmann, Ochoa and Kornberg, grew to scientific fame in the study of glycolysis and allied metabolic pathways. Let me elaborate upon my claim that the elucidation of glycolysis has been one of the turning points, i.e. the first revolution in the development of modern biochemistry.

Essential Steps in Elucidation of Glycolysis First, I should like to review briefly the main stages on the way to elucidation of glycolysis (see 1). The first step dating back to the beginning of the century was the discovery by the Buchners that glycolysis could proceed even after the destruction of the integrity of the cell. From then on for many years the main objects of metabolic research were yeast juices or muscle pulp and its extracts. The second step was the insight by Harden and Young that phosphate plays an essential role in glycolysis. More than 20 years later the pace of discovery quickened. The identification of pìiosphocreatine by Fiske and Subbarow and of ATP by Lohmann combined with the introduction of thermochemistry by Meyerhof led to the conception of energyrich bonds and to the discovery of transphosphorylation. The adenylate system was conceived as a coenzyme of glycolysis as well as a transmitter of energy. Dische proposed oxydoreduction as a key event and postulated a glycolytic cycle. The accidental discovery of 3 PG by Nilsson and the work of Lohmann and Lipmann on the conversion of fructose 1,6-bis-phosphate to more hydrolysis-resistant phosphate esters was the final trigger for G. Embden to propose the unified scheme of glycolysis basically as we know it now. With one stroke everything fell into place and the road was free to elucidate all the missing steps within a few years. The methylglyoxal theory of Neuberg, dominant for 20 years, was finally put aside. The idea of circular processes with many intermediates, that finally add up to a simple balance equation, became a paradigm - actually a philosophy of viewing metabolic processes

158

S. M. Rapoport

in a subtle and dynamic manner. One has to admit that about that same time a metabolic cycle for the formation of urea had been proposed by Krebs - however without furnishing an understanding of the energetic of the process and of the connections between material and energy exchange. He wrote even in 1934 (2): "These reactions take place only in the intact liver cell. Both the chemical mechanism used in the cell to bring about these reactions and the physical mechanisms for transmitting the energy required for the urea synthesis are unknown, and there is yet little hope of progress along these lines. Our lack of knowledge of the chemical nature of enzymes makes impossible an understanding of the chemical mechanism even of simple enzyme reactions, and even more remote must be the possibility of penetrating into the complex system of urea synthesis." The epoch-making review of F. Lipmann (3), in which the squiggle was introduced, represents the acme of conceptional generalization stemming from research in glycolysis. Here the idea of energy-rich phosphate bonds was extended far beyond Lohmann to other compounds, the term "group potential" was introduced, the "phosphate cycle" was conceived with ATP as the universal energy currency; perspicaciously the essential features of activated fatty and aminoacids as acyl compounds were predicted. The situation preceeding the unified theory has been concisely described by A. Szent-Györgyi in 1937 (4): "Every year during the last decade a new and different theory of fermentation claimed our adherence. New ways in which the cell can decompose glucose into lactic acid were constantly being discovered. In the last two years, a new theory was again proposed by Dische, Meyerhof and Parnas, and this time I think it really describes the main road of fermentation." The study of glycolysis was the most powerful stimulus for the detailed study of enzymes, guided by a strategic concept, i.e. the elucidation of a whole pathway. Methods for the isolation and crystallization of the glycolytic enzymes were established, primarily in the laboratories of Warburg and Cori, to become a standard for all enzymologists and the mechanistic and kinetic studies on them paved the way to a growing understanding of the enigmatic catalytic mechanisms of enzymes up to the present day. The course of the subsequent developments in biochemistry was largely conditioned by the experiences with glycolysis. Among others one may cite the photosynthesis of ATP and the Calvin cycle, the discovery of cAMP and other cyclic nucleotides as co-factors of phosphorylation and second messengers, the regulation of enzyme activities by phosphorylation and the unravelling of cell receptors. However, one should also note the confines of the approach based on glycolysis. I shall mention just two examples: one, the failure to elucidate oxydative phosphorylation with the futile search for covalently bound phosphorylated intermediates; the other one the limitations in the understanding of protein synthesis imposed by the restriction of the question to that of activating of amino acids without regard for the problem of coding.

Glycolysis and the Dawn of Modern Biochemistry

159

Obstacles; Misconceptions, Methods To appreciate the revolution in biochemistry, however, it is not enough to view the consequences of the discoveries around glycolysis. I think that is even more instructive to consider the state of biochemistry preceding its Golden Age, and to delineate the intellectual and methodical obstacles, which had to be overcome. As Florkin has rightly stressed, colloidology was the greatest deleterious influence. I suppose that hardly any biochemist nowadays knows the tenets of the bio-colloidal approach and I will mention just two: for one, according to the colloidal doctrine, proteins - or enzymes for that matter - did not have defined molecular weights; secondly, a cell was considered to be a colloidal system without structure passing from a sol to a gel state or vice versa under the influence of ions under various circumstances. All biological phenomena including energy conversions were supposed to be caused by changes on the surfaces of micelles. It is obvious that such an approach obstructed rigorous physical and chemical research and promoted mystifications of various kinds. For instance Willstätter came to the conclusion that all enzymes are colloidal systems consisting of an unspecified large-sized carrier and a low-molecular active component. Warburg wasted years with studies on the catalytic properties of charcoal containing iron as the essential impurity on its surface. The strong influence of bio-colloidal conceptions is clearly visible in the first publications of Lipmann from the laboratory of Meyerhof, e.g. in that on the mechanism of the action of fluoride to inhibit the liberation of inorganic phosphate (5). In the discussion of the results he considered first a colloid-chemical interpretation, but he himself apparently was not comfortable with it. He preferred the analogy with a defined substance, fluor-met-Hb and suggested that fluoride may act on an iron-containing component of phosphate-cleaving enzymes - obviously under the influence of Warburg. The vague notions of the properties of enzymes and of cellular structure are clearly apparent in the article of Karl Lohmann in 1930, on the formation and cleavage of phosphate esters in muscles in the presence of fluoride and oxalate (6). Here the proposal was advanced, that fluoride may move the activities of enzymes into new directions combined with a more complete destruction of colloidal cell structure. A second obstacle to progress was the idée fixe, that the key intermediate molecules would have to be intrinsically labile. This misconception may be traced back to Harden's view that the phosphate esters he discovered were not on the main pathway of the conversion of glucose. Meyerhof and others postulated reactive forms of glucose with the phosphate esters as "stabilization products". In other words the specific metabolic properties were ascribed to the low molecular compounds rather than to the enzymes acting on them. Closely allied was the concept of lactacidogen proposed by Embden. He was convinced of the importance of phosphate for muscle activity and metabolism and postulated that a hypothetical substance, the lactacidogen, served both as the source of lactate and inorganic phosphate, but also as the donor of energy for muscular contraction.

160

S. M. Rapoport

I need not expand on the decisive role of methodology for the progress of science. What were the current methodologies at that time? Titrimetry was still very common, at least in Europe. Colorimetry was in ascendence, but photometry was practically unknown; gasometric and manometric methods were in restricted use. The isolation of substances was cumbersome; it required large amounts to start with and their identification was fraught with uncertainties. Precipitation methods with alkaloids and salts of heavy metals predominated. Filtration was the prevailing procedure for separations; the centrifuges were dangerous monsters, anchored in concrete or suspended by chains with much breakage of glass. The determination of phosphate illustrates well the positive and negative importance of methodology. Careful observation of the development of color in the phosphomolybdate blue reaction led Fiske and Subbarow as well as the Eggletons to the discovery of phosphocreatine, whereas the version of the method used by Lohmann made him overlook the compound. The method derived by Lohmann for the analysis of phosphate esters, based on their differences in hydrolysis rate, led him to the discovery of ATP; but the same method because of its lack of specificity misguided him and Lipmann to miss the formation of phosphoglyceric acid. All in all the situation before the establishment of the unified theory of glycolysis may be likened to a landscape in grey dawn, with only a few unconnected peaks visible amidst unpenetrable darkness.

Glycolysis and Regulation of Metabolic Pathways Let me return to the developments inaugurated in the field of metabolic pathways with emphasis on their regulation. The knowledge of the glycolytic pathway paved the way for the elucidation of other metabolic routes, so that now the main components of the metabolic network of cells could be identified, even though many lacunae remain. The stoichiometric relations of the enzymatic reactions and their intermediate products as well as branching and merging points and some bypasses became known. Many enzymes have been identified, purified and characterized kinetically, and their effectors recognized. However, the compartmentation of enzymes and metabolites still presents difficulties and is an area of contention. Nevertheless, one can state that with the identification of the fundamental pathways of intermediate metabolism the stage has been set for the study of their regulation. A multitude of experimental investigations have yielded insight into various principles of regulation such as induction and repression of enzymes, modulation of their activity by covalent modification and by allosteric effectors, as well as the distributary function of co-enzymes and metabolites. These lines of research have yielded an understanding of the mechanisms of transmission of signals to and within the cell, as well as the knowledge of the structure and function of receptors. At present we are faced with an overwhelming plethora of options for possible regulatory mechanisms. The existing difficulties are illustrated by the fact, that for some enzymes more than 20 allosteric effectors have been described. It is necessary to distinguish between biologically relevant effectors and the multitude of phenomena, some of which are only found under non-physiological conditions in vitro. It

Glycolysis and the Dawn of Modern Biochemistry

161

is obviously impossible to assess their relevance intuitively and it appears necessary to choose a more theoretical approach to clarify our thinking. The statement of Lord Kelvin (7) seems to be appropriate in this context: "... when you can measure what you are speaking about, and express it in numbers, you know something about it;... when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind...". To answer the need for a rigorous and quantifiable description of the regulatory properties of metabolic pathways a theory of control was developed (see 8 for a short review and literature) which allows the evaluation of the role of individual enzymes in the determination of fluxes and metabolite concentrations, as well as the contributions of effectors.

Table 1

Knowledge required for the understanding of the metabolic regulation of a pathway

Kinds of reactions and their thermodynamic constants Kinds of enzymes Kinds of metabolites Kinds of effectors Loops and inner couplings Cellular localization Interaction with other pathways merging and branching Protein-protein interactions Concentrations and kinetic properties of enzymes under cellular conditions Effective concentrations of significant species of metabolites and effectors Enzyme activation or inhibition by covalent modification Interaction with cellular membranes The influence of extracellular signals, ini. hormones

+ + + + + + 0 0 + + 0 ? +

+ indicates, that the knowledge exists for the erythrocyte 0 indicates, that the criterion does not apply

Another complementary approach to systematize and evaluate the mass of empirical data is mathematical modelling of metabolic pathways, keeping in mind that the regulatory properties of metabolic pathways as entities cannot be derived from the knowledge of the structure and function of the individual enzymes alone. A good model should offer an unambiguous analysis of the regulatory properties of the pathway and should identify the functionally important enzymes. Furthermore it represents an independent test of the experimental data bringing forth their inconsistencies. By means of models it is possible to disentangle the complexities of a metabolic pathway, which render the verbal analysis of the cause-effect relations impossible. Finally models may be used to establish general principles of the structural and functional design of metabolic pathways as they developed during evolution. However, the knowledge required for the understanding of the regulation even if only a single pathway is considered, is not available for most cells. Table 1 presents a selective list of the data required. While there may be more or less adequate knowledge of the first 6 requirements for some metabolic pathways, matters are

162

S. M. Rapoport

quite different with respect to the other requirements. We are reduced to simplifications and uncertain extrapolations. These basic uncertainties are the source of a variety of opinions on the general feasibility to establish rigorous models of metabolic regulation. Some scientists go so far as to doubt the possibility of any application of data obtained on isolated enzymes to cellular conditions, while other have underestimated the difficulties of the task. For this reason a satisfactory experimental and theoretical analysis of any pathway assumes general importance. The glycolysis of erythrocytes was an ideal experimental basis for the theoretical analysis and modelling for several reasons: 1. Firstly in the mature erythrocyte the organelles have disappeared so that only the interplay of cytosol and the plasma membrane has to be considered. 2. Secondly its metabolism is reduced practically solely to glycolysis, uncontaminated by other pathways; neither glycogen synthesis and breakdown nor gluconeogenesis interfere. 3. Thirdly the data on enzyme kinetics and metabolite concentrations are probably more complete than those for any other cell. 4. A number of external factors influences the metabolism in a reproducible manner. 5. A variety of mutants causing anemia in man permit the study of the metabolic consequences produced by deviations of quantity and properties of enzymes. Nevertheless, even the simple setting of the mature erythrocyte offers a number of complexities caused by the large number of enzymes, their complicated kinetics, the existence of the D P G by-pass, of feed-back inhibitions, and the inner couplings mediated by the N A D and the adenylate systems. By means of models it was possible to characterize glycolysis both under steady- and non-steady-state conditions by a minimal number of data, to identify the sites of effectors, to assess the relative importance of the various enzymes and to understand the differences between in vivo and in vitro conditions. The models predict the correct steady state and quasi-steady-state concentrations of metabolites and explain the relative constancy of ATP and 2,3-bisphosphoglycerate under various experimental conditions. Furthermore the time-dependent changes of metabolite concentrations under blood preservation conditions and the changes in pyruvate kinase deficient erythrocytes are described. Further extensions of the models include the reactions of synthesis and breakdown of adenine nucleotides, the active and passive fluxes of ions across the cell membrane and osmotic as well as electric effects. Thus a metabolic-osmotic model of the erythrocyte was constructed. I should like to turn to an aspect, which has become controversial in the last years. Research on glycolysis was from its beginning on, mainly concerned with the study of aqueous extracts, which represent for the most part cytosolic constituents. This circumstance was fortunate since it helped to circumvent the conceptional obstacle presented by the existing ignorance of cell structure. One should remember that mitochondria were first recognized and used as biochemical objects no earlier than the late forties. The use of unstructured systems has conditioned the view of glycolysis as the interplay of independent soluble enzymes connected by the

Glycolysis and the Dawn of Modern Biochemistry

163

diffusion of intermediate products. This conception was strengthened by the great successes to isolate all enzymes participating in glycolysis and to study each of their reactions separately, as well as by the achievements of modelling. Recently the existence of topological complexes has been proposed with respect to glycolysis, even in erythrocytes (9). Thus the validity of the achievements of the golden age of research in glycolysis is challenged. The evidence for complexes of glycolytic enzymes in vivo to my mind is weak and inconsistent and characterized by lack of functional criteria, which would permit the rejection of artefacts. The proponents leave quantitative considerations out of their sight in at least three aspects: 1. The large differences in the molarity of glycolytic enzymes within one cell type, amounting to as much as three orders of magnitude, can yield only a minute proportion of a stoichiometric complex among them. 2. Reasonable association constants among the enzymes in the complexes postulated have neither been discussed nor measured. 3. The kinetic relevance of the complexes has not been scrutinized. A consideration of glycolysis solely as an ATP-producing pathway is one-sided. Production and consumption of ATP are closely geared to each other and from a functional point of view consumption of ATP governs its production. This is true both for respiration and glycolysis. In essence this insight was gained by Meserhof in his demonstration, that the lack of ATPase is the basic cause for the accumulation of hexose phosphates as found by Harden and Young in yeast juice. In view of the overriding role of ATP-consuming processes for the control of ATP-production it appears an important task to draw up a bioenergetic budget in which the various ATP-consuming processes of the cell are identified, quantified and their interrelations clarified. Even for the mature red cell we were not able to apportion the ATPconsuming processes, which are mostly located in the cell membrane, despite our knowledge of the amount and the regulation of ATP-production. We were more successful in our attempts to balance ATP-formation and consumption with reticulocytes (10) and ascites tumor cells (11). In these cells we could account for about three fourth of the ATP-budget. In both types of cells protein metabolism takes first place, with synthesis accounting for 25-30%, closely balanced in the steady state by protein degradation, which consumes 10% of the ATP in keeping with the ration of 3 : 1 of ATP needed for making and breaking of a peptide bond. Obviously proteolysis proceeds mainly via ATP-dependent systems. Ion transport takes second place, mostly by way of Na + Κ + -ATPase but also to a sizeable extent via Ca 2 + transport in ascites cells. Synthesis of ribonucleic acids, only measurable in ascites tumor cells, accounts for another 10% and is not strictly coupled to protein synthesis. What about the unaccounted ATP-consuming processes? Among the many possibilities one deserves special consideration. According to recent estimates one out of 6 proteins of the cell are thought to be reversible phosphorylated, presumably serving regulatory functions at a considerable expense of energy (12). The main conclusion from these studies, aside from furnishing a synoptic view of the energy budget of a cell has been the realization that each of the major ATPconsuming processes appears to be regulated independently and that they do not

164

S. M. Rapoport

compete with each other for ATP; in other words the cell keeps separate accounts for each ATP-consuming process. How this is achieved we do not know. In closing I should like to discuss two points: one pertains to the epistemological problems, which are caused by the ongoing revolution in methodology; and the second concerns the need for synoptic integrative research. We have been witnessing a stupendous development of methods. Their sensitivity has increased by 15-18 orders of magnitude. The output of data per time has grown by perhaps up to 6 orders of magnitude. The specificity has reached in many cases the theoretical limit. The number of scientists has grown by two orders of magnitude. The results of these developments is an overwhelming mass of data which continue to accumulate at an increasing rate. The great bulk of the results are phenomenological. Without a corresponding development of theory we are in danger of drowning in this cataract despite frantic efforts to master the situation with the help of computers. The results presented on the modelling and on the ATP budget of some types of cells are just a modest beginning in the quest for a quantitative approach to the regulation of metabolism. New methods such as N M R spectroscopy which permit a synoptic view of the metabolites in situ of cells, organs and organisms without disturbing their integrity, promise further advances, with glycolysis again in the forefront of these developments. In closing I want to plea the case of an integrative approach to biology. It was A. Szent-Györgyi who stated that the pursuit of the research of life carries us to ever lower levels of the organization of matter until we end up with molecules, electrons and quanta, but somehow have lost underway the essence of life. We should never forget that living systems are complex and hierarchical entities encompassing an as yet unmanageable multitude of interactions within and among their subsystems. To my mind the understanding of this interplay is a central task of future biological research. The insight of Lipmann will remain a beacon to guide us in this endeavour.

References 1. Florkin, M. 1975. Comprehensive Biochemistry, Vol.31 (M. Florkin and E.H. Stotz, eds.). Elsevier, Amsterdam. 2. Krebs, H.A. 1934. Erg. Enzymol. 3, 247. 3. Lipmann, F. 1941. Advan. Enzymol. 1, 99. 4. Szent-Györgyi, A. 1937. In: Perspectives in Biochemistry (J. Needham and D.E. Green, eds.). Cambridge University Press, p. 165. 5. Lipmann, F. 1928. Biochem. Zt. 196, 1. 6. Lohmann, K. 1930. Biochem. Zt. 222, 324. 7. Kelvin, Lord (Thomson, W.). 1889. Popular Lectures and Addresses, Vol. 1, p. 73. Mac Millan. 8. Heinrich, R., S. Rapoport. 1983. Biochem. Soc. Transact. 11, 31. 9. Srere, P.A. 1987. Ann. Rev. Biochem. 56, 89. 10. Siems, W., W. Dubiel, R. Dumdey, M. Müller, S. M. Rapoport. 1984. Eur. J. Biochem. 139, 101. 11. Müller, M., W. Siems, F. Buttgereit, R. Dumdey, S.M. Rapoport. 1986. Eur. J. Biochem. 161, 701. 12. Goldbeter, Α., D.E. Koshland. 1987. J. Biol. Chem. 262, 4460.

Energy-Rich Bonds and Enzymatic Peptide Synthesis Joseph S. Fruton

I count it an honor to participate in this symposium and, as its title suggests, this essay deals with some historical aspects of the relation of Fritz Lipmann's "squiggle" to the problem of enzyme-catalyzed formation of peptide bonds. I must begin by confessing that, in 1941, when I studied the oft-celebrated article in the first volume of Advances in Enzymology (1), I did not become a Lipmann disciple. Indeed, I found Hermann Kalckar's review, published in the same year (2), to be a sounder contribution. On reflection, and after re-reading both articles, I believe that this bias reflected my early scientific education. To explain this inclination, I beg the reader's indulgence if I summarize briefly my scientific life in the 1930s. During the latter part of my undergraduate studies at Columbia College, in 1929-1931, I came under the influence of three outstanding teachers. The first was John M. Nelson (1876 1965), who taught elementary organic chemistry and did research on enzymes (3); his most famous Ph. D. student had been John Howard Northrop (1891-1987), and I first heard about Northrop's crystallization of pepsin from Nelson. I also came to know Roger Herriott, who was doing his graduate work in Nelson's laboratory. My second teacher was the physical chemist Louis P. Hammett (1894-1987), who introduced me to the book on thermodynamics by Lewis and Randall (4), and who was then beginning his notable studies on the kinetics and mechanisms of organic reactions. The third teacher was Selig Hecht (1892-1947), whose course in general physiology and whose personal guidance led me to try to become a biochemist. From Hecht I first learned of the exciting developments in the study of biological oxidations, and I began to read avidly the articles of Keilin and Warburg. Hecht's research on vision was then underway, and George Wald was one of his students (5). From 1931 to 1934, I worked for my doctorate in the Department of Biological Chemistry at the Columbia College of Physicians and Surgeons, with the organic chemist Hans T. Clarke (1887-1972) as my research advisor (6). In this work, I synthesized various TV-substituted cystine and cysteine derivatives, including peptides, and studied the relative lability of the disulfide compounds in alkali and the relative rates of the auto-oxidation of the thiol compounds. I also attempted to determine the oxidation-reduction potentials of the RSSR-2RSH systems by colorimetric estimation of the equilibria established with some of the electromotively-active dyes developed by William M. Clark and by Leonor Michaelis (7); for a recent discussion of the potentials of disulfide-thiol systems, see Gorin et al. (8). As a graduate student, I benefited greatly from the help and advice of the physical chemist Crawford F. Failey (1900-1981) and the organic chemist Oscar Wintersteiner (1898-1971), who were then junior members of Clarke's department.

166

Joseph S. Fruton

In 1934,1 was given the opportunity to work at the Rockefeller Institute for Medical Research, in the newly established laboratory of Max Bergmann (1886-1944), who had just come to the United States from Dresden, where he had been dismissed as Director of the Kaiser Wilhelm Institute for Leather Research. Bergmann was a distinguished organic chemist who had made notable contributions in both carbohydrate and protein chemistry. Two years earlier, he and his associate Leonidas Zervas (1902-1980) had published their famous paper on the so-called carbobenzoxy method of peptide synthesis (9). Soon after I came to the Rockefeller Institute, Zervas arrived there, and it was my good fortune to share a laboratory with him during his two-year stay in New York. Whatever skill I acquired in the art of peptide synthesis came from this association with Zervas. In the years that followed, my research dealt largely with proteolytic enzymes, and the most rewarding results came from the study of the action of well-defined proteinases on synthetic peptide substrates. In the course of this work, I made repeated visits to Northrop's laboratory at the Princeton branch of the Rockefeller Institute, and learned from Moses Kunitz, Mortimer Anson and Roger Herriott something of the art of crystallizing enzymes. The finding of the first synthetic substrates for chymotrypsin, for trypsin, and especially for pepsin, gave support to the peptide theory of protein structure at a time when that theory was still a matter of debate. During the late 1930s, serious attention was being given to such theories as the Wrinch cyclol hypothesis, and Linderstram-Lang had suggested that whereas denatured proteins may contain peptide bonds that are cleaved by proteinases such as trypsin, the finding of synthetic substrates for this enzyme did not prove the existence of such bonds in "genuine globular proteins" (10). In the years before 1940, there were two other developments in Bergmann's laboratory that influenced my later scientific thought. The first was the observation by Heinz Fraenkel-Conrat that, in the presence of aniline (or Phenylhydrazine), the proteinase papain catalyzes the conversion of carbobenzoxyglycinamide into the sparingly-soluble anilide (or phenylhydrazide), and that this reaction is more rapid than the condensation of carbobenzoxyglycine with either of the two nucleophiles (11). Shortly afterward, I showed that chymotrypsin catalyzes the synthesis of benzoyl-L-tyrosylglycinanilide from benzoyl-L-tyrosine (or benzoyl-L-tyrosinamide) and glycinanilide (12). These experiments provided the first unequivocal demonstration of the ability of a proteinase to catalyze the synthesis of peptide bonds, either by reversal of the hydrolytic process, or by effecting what Bergmann called "replacement" reactions. He also seized upon Fraenkel-Conrat's discovery to elaborate on the long-standing idea that the biological synthesis of proteins involves the reversal of proteolysis. This idea was based principally on the "plastein" reaction, which had been studied actively during the 1920s by Wasteneys and Borsook (13). The first opportunity I had to speak publicly about enzyme-catalyzed peptide bond formation was at the 1938 Cold Spring Harbor Symposium, where I discussed briefly the energy relations, and the necessity of coupling endergonic peptide synthesis with exergonic processes. Admittedly, my suggestions regarding such processes were feeble: "One mechanism may be mechanical removal by the circulatory system; more complex interrelationships with other chemical systems such as those concerned with oxidation-reduction might, under proper metabolic conditions, favor protein synthesis" (14).

Energy-Rich Bonds and Enzymatic Peptide Synthesis

167

The other major development in Bergmann's laboratory during the 1930s did not involve me as a participant, but as a critical observer. After Zervas had left in 1936, his place in the laboratory in which I worked was taken by Carl Niemann (1908-1964). At that time, Bergmann's personal experimental effort was devoted to the search for new specific reagents for the amino acid analysis of protein hydrolysates. Niemann's assignment was to use these and other available methods to determine the amino acid composition of several protein preparations. From the data which he collected, there emerged the Bergmann-Niemann periodicity hypothesis of protein structure, which stated that the total number of amino acid residues per presumed protein molecule falls into a series of multiples of 288, that the content of each amino acid can be denoted by 2" χ 3m, and that every amino acid recurs at a regular periodic interval in the polypeptide chain of a protein (15). The fame of this hypothesis was brief, for it was soon disproved, and Charles Chibnall rightly ascribed it to "the hypnotic power of numerology" (16); its principal contribution was to encourage Gordon, Martin and Synge to develop a chromatographic technique for the more accurate amino acid analysis of protein hydrolysates (17). As Niemann's laboratory partner, I ventured to express to him my disbelief in the conclusions which he and Bergmann had drawn from data obtained on protein preparations of doubtful homogeneity by means of analytical methods of uncertain accuracy. I do not recall having ever spoken to Bergmann about my scepticism, for in those years out relationship was rather formal; it thawed greatly after 1939, and by that time he had abandoned the hypothesis. In my education as a scientist, the imprint of this experience has been lasting. I have been less ready than others to accept biochemical theories which seemed to be founded on experimental data obtained with methods of uncertain reliability, and I also learned that even the most renowned scientists can occasionally go astray. To this sketch of my scientific education at the Rockefeller Institute, I must add that I had the good fortune there to gain inspiration and valuable counsel from several senior members, notably Leonor Michaelis, Donald Van Slyke, and Karl Landsteiner. I also learned much from junior members of various research groups, in particular Albert Claude, who had just begun his studies on the differential centrifugation of subcellular elements, and Colin MacLeod, who was working on the purification of the pneumococcal transforming factor. Of course, everyone at the Institute knew about the isolation of tobacco mosaic virus by Wendell Stanley at the Princeton branch, and the subsequent demonstration by Bawden and Pirie that it is a nucleoprotein. For a youngster in his twenties, the Rockefeller was an exciting place, although as an avid reader of the biochemical literature I soon realized that many important currents were flowing elsewhere. Apart from my personal contacts with Rudolf Schoenheimer and his group in Clarke's department at Columbia, and the knowledge they provided about their pioneer isotope work, I depended largely on my reading of German and British journals to learn of the exciting advances being made in the study of metabolic processes. The 1930s were the time when enzymes began to occupy the center of the stage, and were finally admitted to be catalytic proteins. Various substances, ambiguously named coenzymes, were identified and, through the rapid determination of their structure, found to be related to vitamins. Hans Krebs's elucidation of the pathway of urea synthesis and his completion of the citric acid cycle were high points in the excitement. So were the

168

Joseph S. Fruton

discovery by Alexander Braunstein of the transamination reaction, and that by Carl and Gerty Cori of the reversible phosphorolysis of glycogen to produce glucose-1phosphate. In following the tortuous emergence of what came to be called the Embden-Meyerhof pathway of anaerobic glycolysis, and the role of ATP in this process, I began to appreciate the insights of Gustav Embden and to learn about the work of Jacob Parnas and of Dorothy Needham on coupled oxidation-reduction reactions. Also, I read with much interest the papers by Zacharias Dische, Vladimir Belitser and Hermann Kalckar, who provided evidence for the coupling of phosphorylation to aerobic oxidation. The greatest excitement of all was evoked by the publications from Otto Warburg's laboratory, especially those on the intermediate formation of 1,3-diphosphoglycerate in the oxidation of glyceraldehyde-3phosphate to 3-phosphoglycerate, and the transfer of the 1-phosphoryl group to ADP. This achievement opened a new vista on the problem of the coupling of exergonic oxidations to endergonic phosphorylations, and brought home to many biochemists the importance of working with highly-purified crystalline enzyme proteins. Here, not only science, but also craftsmanship was needed, and after the successes of Sumner and Northrop, the greatest craftsman was Moses Kunitz (1887-1978). To these recollections of my euphoria during the 1930s I must add that in preparing this essay, I re-read some of the published tributes to Fritz Lipmann, and in one such tribute, which appeared in 1966, I found a short chapter by Hubert Chantrenne, who wrote: "Biochemistry in 1938 looked somewhat depressing to a student eager to learn about the processes of life." He attributed this feeling to the contents of the textbooks available to him at that time. Ater summarizing the high points of Lipmann's famous 1941 review article, Chantrenne concluded that "Biochemistry is no longer the chemistry of death and decay; it is the chemistry of the living cell, with its essentially irreversible, oriented processes, admirably organized and controlled" (18). There can be no doubt that the articles by Lipmann and by Kalckar in 1941 rendered inestimable service in collecting and organizing the available knowledge about the coupling of biological oxidations to phosphorylations. Lipmann, however, went further in his generalizations, by introducing the concept of "group potential" in transfer reactions and by making a sharp distinction between "energyrich" and "energy-poor" bonds. I did not find his approach to be a revelation, possibly because during my student days I had learned to use the catch-phrase "From thermodynamic considerations it follows that..." when discussing biochemical problems with my contemporaries. It seemed to me that his novel use of the term "bond energy" and the symbol he introduced to denote energy-rich bonds were merely the kind of jargon and shorthand that biochemists had often employed to simplify complex chemical relationships. From my previous reading of the biochemical literature, it was clear that Lipmann's ideas were extensions of those advanced by Otto Meyerhof, and I was ready to accept Lipmann's emphasis on the central role of ATP in carbohydrate metabolism. However, I was less convinced by his generalization that ATP is a unique funnelling agent of chemical energy in biochemical synthesis. His briefly stated hypothesis, based on the properties of acetyl phosphate, that protein synthesis involves the polymerization of aminoacyl

Energy-Rich Bonds and Enzymatic Peptide Synthesis

169

phosphates, was only of mild interest, as it did not mention the possible enzymatic catalysts that might be involved or the need for specificity in the mechanism whereby the amino acid sequence of proteins is determined. A few years earlier, I had begun to study the specificity of proteinases (cathepsins) extracted from various animal tissues and, by the use of synthetic peptide substrates, had found counterparts of the well-defined digestive enzymes. In my report of the results at the 1941 Cold Spring Harbor Symposium, I noted that "...while there is no direct experimental proof for the statement that the proteolytic enzymes are responsible for the biological synthesis of proteins, it must be admitted that at the present time these enzymes are the only agents that are known to catalyze the synthesis of peptide linkages" (19). Because of the war, and my move to Yale University afterwards, I did not resume concerted work on the problem until 1949. By that time, there had been several important new developments which I will discuss below. My main difficulties in reading Lipmann's 1941 article were, as best as I can now recall, threefold. 1. Although the arithmetical summation of the appropriate freeenergy data for the hydrolysis of the individual reactants in a transfer reaction was justified on theoretical grounds, the numerical value calculated for the over-all change in free-energy in that reaction could be only as reliable as the experimental data upon which the calculation was based. In particular, I was doubtful about the reliability of the values based on thermochemical data for heats of formation and heat capacities, since these calculations involved differences between large numbers of variable precision. Also, I had doubts about the estimates of free-energy values derived solely from calorimetric determinations of changes in enthalpy. I felt, therefore, that the possible influence of the "hypnotic power of numerology" could not be disregarded, and I was not prepared to accept the evidence for the sharp distinction which Lipmann made between energy-rich and energy-poor bonds. 2. I could not see how the free energy values Lipmann cited in his 1941 article, applicable to closed systems, could be used in connection with metabolic processes operating in open systems, where the reactants might be poised at concentrations far from equilibrium ratios; it was only in a later paper (20) that Lipmann considered this question. 3. Finally, and most importantly, I could not unlearn the dicta that thermodynamic data alone cannot predict the pathway or the mechanism of a chemical process, and that a strongly endergonic (and hence apparently irreversible) reaction can be made to proceed if a chemical mechanism is available for coupling the reaction to an appropriate exergonic process. At this point I wish to insert a historical note about the concept of the energetic coupling of chemical reactions. It was discussed a long time ago, in 1900, by Wilhelm Ostwald (1853-1932), one of the founders and the chief protagonist of the then-new physical chemistry. In making a distinction between the transfer of chemical energy and the transfer of heat or electrical energy, he wrote: A direct reciprocal transformation of chemical energies is only possible to the extent that chemical energies can be set in connection with each other, that is, within such processes that are represented by a stoichiometric equation. Coupled reactions of this kind may be distinguished from those that proceed independently of each other; their characteristic lies in the fact that they can be represented by a single chemical equation with definite

170

Joseph S. Fruton

integral coefficients. On the other hand, chemical processes that are not coupled cannot transfer energy to each other (21). To this I would add that the term "energetic coupling" has had a somewhat ambiguous status in the biochemical literature. In 1930, while writing about the relation between oxygen uptake and glycogen resynthesis in muscle, Otto Meyerhof stated that "...oxidation and resynthesis do not represent a chemically-coupled process, for which one can give a stoichiometric equation, but an energetically coupled one" (22). Four years later, Jacob Parnas wrote: ... the resynthesis of phosphocreatine and adenosine triphosphate is not linked to glycolysis as a whole, but to definite partial processes; and this leads further to the conclusion that this resynthesis does not involve a relationship that might be termed "energetic coupling", but more probably involves a transfer of phosphate residues from molecule to molecule (23). The remainder of this essay will be devoted to some historical aspects of the development, after the Second World War, of the problem of peptide-bond formation in biological systems. It is impossible, in a few pages, to do justice to the great contributions made after 1945 by the many gifted investigators, including Lipmann, to our present knowledge about this process. Therefore, in keeping with the spirit of this symposium, I will say most about those aspects of the problem in which I have had a personal interest. They may be stated in the form of three questions: 1. What are the free-energy changes in the synthesis of a peptide bond which joins two amino acid units? 2. What are the biochemical reactions in the union of amino acids to form the peptide bonds of proteins and natural smaller polypeptides? 3. What enzymes, if any, catalyze this union? I begin the consideration of the energy changes in peptide-bond formation by quoting from Lipmann's Harvey lecture in December 1948; it was entitled „Biosynthetic Mechanisms:" When we look at the great variety of cell constituents it seems an overwhelming task to find the processes by which all these compounds are manufactured. However, if we start to focus our attention on primary linkages, we discover that the links in different classes of compounds are quite similar. As an endlessly repeating process in the building up of cell material, we find elimination of water between groupings. One of the most common reactants in these condensations through intermolecular dehydration is the carboxyl group. Quite generally, ester and peptidic links appear to be formed by carboxyl activation. The energy input necessary to perform these linkings as they occur in fat, protein and elsewhere is very similar, around 3,000 calories. Phosphate bonding is used... in the metabolic machinery as a means to parcel and transfer energy. Available energy is processed initially into energy-rich phosphate bonds of approximately 15,000 calories each and distributed through the adenylic-adenyl pyrophosphate and probably other systems. A key, or at least one of the keys, to its very general applicability appears to be the acid anhydride nature of the energy-rich

Energy-Rich Bonds and Enzymatic Peptide Synthesis

171

phosphate bond. This makes it possible for the ATP to act as a common reagent for all these condensations involving the removal of water. In those reactions involving the carboxyl groups, this general statement can be qualified more precisely. It appears that here the primary reaction is the formation of an acid anhydride between carboxyl and phosphate (24). Lipmann's use of + 3 kcal per mole for the free-energy change in the formation of the peptide bonds of proteins referred to the condensation of two dipolar amino acid molecules to form a dipolar dipeptide. This value came from Henry Borsook's calculations from the thermochemical data his colleague Hugh M. Huffman had obtained for several amino acids and and their derivatives (25). Later recalculations by Linderstrom-Lang (26) and by Carpenter (27) from Huffman's data gave a value of about -I- 4 kcal per mole for this process. It was obvious, however, that the freeenergy change in the formation of a free dipeptide could not be used in connection with the formation of the peptide bonds of long polypeptides with only one α-amino group and only one a-carboxyl group. Indeed, calculations for benzoylglycine and benzoylglycylglycine suggested that in the process: RCO—Ν H CH (R) COO" + + NH 3 CH(R)COCr ^ RCO—NHCH(R)CO—NHCH(R)COO~ + H 2 0 the free-energy change might be about + 2 kcal per mole. I have already indicated my reservations about free-energy values derived solely from thermochemical data. The measurement of equilibrium constants was needed, and the availability of isotopes made such determinations possible even for supposedly "irreversible" reactions. In 1951 I performed the first direct experimental determination of the equilibium constant for the enzyme-catalyzed synthesis of a peptide bond comparable to those present in the interior of long polypeptide chains. The reaction was the chymotrypsin-catalyzed condensation of benzoyl-Ltyrosine and 15 N-labeled glycinamide to form Bz-Tyr-Gly-NH 2 , and the equilibrium constant at pH 7.9 and 25 °C turned out to be 0.49 M~ giving a free-energy value of -I- 0.4 kcal per mole for the synthetic process. This result was reported in a paper (28) with Julian Sturtevant, my colleague in the Yale Department of Chemistry, and his graduate student Alan Dobry. They made calorimetric measurements for the chymotrypsin-catalyzed hydrolysis of Bz-Tyr-Gly-NH 2 and found the enthalpy change to be — 1.55 kcal per mole. In the same paper, we also gave an equation for the pH dependence of the equilibrium constant in such a reversible reaction in terms of the ionization constants of the carboxyl and ammonium groups of the two reactants. It was clear that the large difference in the apparent free-energy change (at pH values near 7) for the hydrolysis of the peptide bond in a free dipeptide and in a blocked dipeptide is a consequence of the difference between the apparent pAT values of free amino acids (ca. 2.4 and 9.6), and those for acylamino acids (ca. 3.8) and amino acid amides (ca. 8.0). I now jump ahead a quarter of a century, because during the 1970s Michael Laskowski Jr. took up the question of the thermodynamics of peptide bond synthesis in connection with his studies on the interaction of trypsin with the protein inhibitor isolated from soybeans by Kunitz. Laskowski and his associates demon-

172

Joseph S. Fruton

strated that the enzyme specifically hydrolyzes the Arg 63 -Ile 64 bond in the inhibitor and also found that in presence of 60 per cent glycerol the equilibrium constant for the hydrolysis of this bond is reduced by a factor of 8. This observation led them to investigate systematically the effect of the addition of organic solvents on the equilibrium constant (Ksyn) for the chymotrypsin-catalyzed condensation of benzyloxycarbonyl-L-tryptophan and glycinamide to form Z-Trp-Gly-NH 2 . By means of HPLC analysis, they followed the attainment of equilibrium in both the hydrolytic and condensation reactions, and found striking increases in the value of KsyD as the proportion of the cosolvent was raised (29). Extrapolation of the plot of Ksyn versus volume per cent of organic solvent to zero cosolvent gave a value of 0.45 M - 1 , in excellent agreement with my result for the synthesis of Bz-Tyr-GlyNH 2 . Moreover, a plot of Ksyn versus pH accorded well with the curve predicted by the equation in my 1951 paper with Dobry and Sturtevant. Laskowski's beautiful experiments show beyond doubt that the favorable effect of the organic cosolvents on peptide-bond formation in such systems is a consequence of the shift in the ρ Κ of the carboxyl component in the condensation reaction to higher pH values. Laskowski's work has awakened interest in the use of proteolytic enzymes for the chemical synthesis of peptides. I mentioned this possibility in a 1949 review article (30), but at that time we knew much less about proteolytic enzymes, and peptides did not occupy as prominent a place in the aspirations of pharmaceutical companies as in later years. One of the first successes was achieved by my former postdoctoral student Ken Inouye who, with several associates, effected the conversion of porcine insulin to human insulin by means of a trypsin-catalyzed condensation reaction (31). I shall not dwell on this practical aspect of peptide-bond synthesis, except to note that while condensation reactions may be useful in special cases, a more fruitful method has been the enzyme-catalyzed aminolysis of peptide esters and amides (32). The roots of this approach lie in the 1940s, with the demonstration by Hans Neurath that trypsin and chymotrypsin catalyze the hydrolysis of esters analogous to the amides that I had found to be substrates for these enzymes (33). Subsequently, ester analogues of amide and peptide substrates were shown to be cleaved by other proteinases (papain, subtilisin, pepsin, etc.). After 1949, an extensive series of papers appeared from my laboratory on the proteinase-catalyzed reaction of substituted amino acid amides (e.g., Bz-Tyr-NH 2 ) with amino acid derivatives (e.g., H-Gly-NH 2 ) to form substituted peptides (e.g., Bz-Tyr-Gly-NH 2 ). In the initial paper of the series (34), published in January 1950,1 called attention to Fraenkel-Conrat 's earlier work, and termed such acyl-transfer processes "transpeptidation" reactions, by analogy to the transglycosylations catalyzed by sucrose Phosphorylase. From the data available at that time, it appeared likely that in the action of enzymes such as papain or chymotrypsin, water and the receptor amine competed for reaction with an acyl-enzyme intermediate, as had been suggested in 1937 by Weiss (35) for papain. Moreover, in 1949 we knew from the reports by Elliott and by Speck (36) on the biosynthesis of glutamine, and that of Konrad Bloch (37) on the biosynthesis of glutathione, that ATP is an obligatory participant in the formation of the y-glutamyl bonds of these two substances. I therefore offered the suggestion that the energy needed for the synthesis of the peptide bonds of proteins from free amino acids might be provided

Energy-Rich Bonds and Enzymatic Peptide Synthesis

173

by the cleavage of ATP in the formation of these y-glutamyl bonds, and that transpeptidation reactions that involve glutamine and glutathione might be operative in protein biosynthesis. A few months later, there appeared a paper (38) by my friend Charles Hanes, who had entirely independently come to propose essentially the same hypothesis. Hanes and his associates demonstrated, by means of paper chromatography, the enzyme-catalyzed transfer of y-glutamyl groups from glutathione and related peptides to various amino acids and peptides. Later work, notably that by Alton Meister, showed that the enzyme (y-glutamyl transpeptidase) also catalyzes the hydrolysis of glutathione (39), and other investigators found that some glutaminases preferentially transfer the y-glutamyl group to amine acceptors (40). Obviously, the speculations that Hanes and I offered about the possible role of glutamine and glutathione in the formation of polypeptides soon became entirely superfluous when the mechanisms in the ribosomal synthesis of proteins began to be unraveled. During the 1950s, the research in my laboratory on enzyme-catalyzed transpeptidation was largely devoted to quantitative studies on the transfer efficiency of welldefined proteinases such as papain, chymotrypsin and trypsin in the reaction of amide substrates with amino acid derivatives as receptors of the acyl group. We also examined the transferase activity of two intracellular enzymes, cathepsins Β and C, which we had purified to an appreciable extent, and whose hydrolytic specificity we had studied. Among the predoctoral students engaged in this research was Mary Ellen Jones, better known for her later work in Lipmann's laboratory, and the postdoctoral associates included Robert Johnston, who had just received his Ph.D. for his work with Konrad Bloch on the biosynthesis of glutathione. It soon became evident that in the competition between water and an amine acceptor for the enzyme-activated acyl group, both the reaction mechanism of the enzyme and its specificity with respect to the amino compound were important in determining the efficiency of a proteinase as a transferase in aqueous solutions. Thus, the greater transfer efficiency of papain, as compared with that of trypsin, in the catalysis of the same transpeptidation reaction (41) could be ascribed to the fact that with papain the acylenzyme is a thiol ester, whereas with trypsin it is an oxygen ester. Also, whereas in the papain-catalyzed reaction of Z-Gly-NH 2 with H-Leu-Gly-OH, the extent of transamidation to generate Z-Gly-Leu-Gly-OH greatly exceeded the extent of hydrolysis of the amide substrate, the reverse was the case when H-GlyGly-OH was the amine acceptor (42). This result revealed a previously unrecognized feature of the specificity of papain. The concurrent hydrolysis of both the amide substrate and the peptide product made a direct determination of the equilibrium constants for such aminolysis reactions impossible. Moreover, my attempt to estimate the equilibrium constant for the chymotrypsin-catalyzed formation of Bz-Tyr-NH 2 from benzoyl-L-tyrosine and 15 N-labeled ammonium ions was unsuccessful, because the extent of synthesis was too small. To my knowledge, no reliable values are yet available for the freeenergy change in the hydrolysis of the alpha-amide bond in TV-substituted amino acid amides. After I determined the value for the hydrolysis of Bz-Tyr-Gly-NH 2 , such data would have permitted an unambiguous calculation of the free-energy change in the formation of this peptide from Bz-Tyr-NH 2 and glycinamide.

174

Joseph S. Fruton

However, since Dobry and Sturtevant (43) had found the enthalpy change in the chymotrypsin-catalyzed hydrolysis of Bz-Tyr-NH 2 to be — 5.8 kcal per mole under the conditions of our experiments, the free-energy change in the transpeptidation reaction could be estimated to be about — 4 kcal per mole (— 5.8 + 1.6), on the assumption that the entropy changes for the hydrolysis of Bz-Tyr-NH 2 and of BzTyr-Gly-NH 2 are not very different. At that time, it seemed appropriate to conclude, therefore, that such exergonic transpeptidations could not be excluded, on thermodynamic grounds alone, from consideration as possible reactions in the biological formation of the peptide bonds in proteins and smaller natural polypeptides. Concurrently with our initial studies on the enzyme-catalyzed aminolysis of amide substrates, Max Brenner (44) showed that chymotrypsin catalyzes the conversion of L-methionine isopropyl ester to H-Met-Met-OH and H-Met-Met-Met-OH. As in the case of the amide substrates, there are many difficulties in the estimation of the free-energy change in the aminolysis of amino acid or peptide esters. Various estimates had been made of the value for the hydrolysis of ethyl acetate and α-amino acid esters (27), and it appeared likely that the free-energy change in the hydrolysis of the latter is about — 6 kcal per mole at pH 7 and 25 °C suggesting that proteinasecatalyzed transpeptidation reactions which involve ester substrates are somewhat more exergonic than those with amide substrates. Such a conclusion would not surprise a peptide chemist; around 1904 Emil Fischer and Theodor Curtius had prepared peptides by the aminolysis of esters. Thus, on thermodynamic grounds alone, such transpeptidation reactions also could not be dismissed from the consideration of the enzymatic mechanisms in protein biosynthesis. I should add, however, that there are important kinetic differences in the action of enzymes such as chymotrypsin on its amide and ester substrates. With amides, the rate-limiting step is the formation of the acyl-enzyme, whereas with esters it is the cleavage of the acyl-enzyme (45). One may therefore account for the greater efficiency of acyltransfer from esters as reflecting a higher concentration of the acyl-enzyme in the steady state of chymotrypsin catalysis. At this point, I should note that, during the early 1950s, the response to the hypothesis that protein synthesis involves the catalysis of transpeptidation reactions was rather unfavorable. In his Lane Lectures, given in 1951, Linderstrom-Lang mentioned approvingly Borsook's criticism of the results we had published up to that time, on the grounds that the extent of hydrolysis greatly exceeded the extent of transpeptidation. This led Linderstrom-Lang to write: ... I therefore propose to look for specific transpeptidases, possibly a class of enzymes related to the proteinases but still sufficiently different from these to enhance [aminolysis] at the expense of [hydrolysis] (26). Because this symposium is devoted to the memory of Fritz Lipmann, it is perhaps appropriate to note that during the 1950s he did not appear to consider the catalysis of transpeptidation reactions by enzymes related to the proteinases as playing any part in protein biosynthesis since, to my knowledge, he did not refer in his writings to any papers from my laboratory. I was interested, therefore, to read in his autobiography, published in 1971, Lipmann's recollections of his early thoughts on protein synthesis. There, he wrote:

Energy-Rich Bonds and Enzymatic Peptide Synthesis

175

At that time much effort was made to show that under special conditions, proteolytic enzymes could perform a peptide condensation (Figure 1). Although the direct evidence was missing, I was much encouraged when, in a spirited discussion of the 1951 situation, Linderstram-Lang sided with my earlier proposed carboxyl activation rather than reverse proteolysis. The primitive understanding of the mechanisms involved may be gathered from a review presented at the 1949 Federation Meetings where I summed up reactions in which synthesis of a peptide bond depended on the split of ATP; for instance, the synthesis of glutamine and of glutathione had been found to link with ATP... I naively thought then that protein synthesis could be more or less solved if one understood the mechanism of amino acid activation. It took me a long time to realize that, in contrast to most other biosyntheses in the making of a protein, this was just a premise. Still it was a step one needed to understand before turning to the bigger problem of producing a predetermined sequence of amino acids (46). As a not entirely objective historian of biochemistry, I must add several comments to this quotation. Figure 1 is a photograph of two test tubes labeled A and B, and is entitled "Proteinlike material from peptide (sic) digest of egg albumin. A. Without chymotrypsin. B. After 45-min incubation with chymotrypsin solid gel has formed. [Reprinted from the Journal of the American Society, Vol.73, 1288 (1951). Copyright 1951 by the American Chemical Society. Reprinted by permission of the copyright owner.]." The author of the cited paper was not named; it was Henry Tauber, and his article dealt with the formation of plastein-like products obtained by the action of chymotrypsin on peptic digests of egg albumin. Indeed, Lipmann's use of the term "reverse proteolysis" suggests that, in writing his autobiography, he did not recall, or chose to ignore, the published literature of the 1950s on proteinasecatalyzed transpeptidation reactions. Since reference was made to a discussion in 1951 with Linderstrom-Lang, I must add that I had many conversations with him on this subject, beginning in 1948, during an extended stay at the Carlsberg Laboratory, and continuing until shortly before his death in 1959. If we disagreed on some aspects of the problem of enzymatic peptide-bond formation, these differences of opinion spurred me to try to get better data on the thermodynamics of the process and the transfer efficiency of proteinases. All who knew Kaj Linderstrom-Lang loved him, not only for the warmth and generosity of his friendship, but also for his eagerness to consider the views of his younger colleagues. As is the case for many of my contemporaries, he exerted a decisive influence on my scientific education. Lipmann began his experimental studies on peptide bond formation during the 1940s, and chose the acetylation of sulfanilamide by liver preparations as a model system for the ATP-dependent synthesis of what he called a "peptidic" bond (24). In connection with this work, he introduced into biochemical methodology the formation of hydroxamates upon the treatment of reactive acyl compounds with hydroxy lamine. At first Lipmann believed this test to be specific for acyl phosphates and that ATP is cleaved to ADP and phosphate; later studies required a revision of both these views. Of course, the most important outcome of these studies was Lipmann's discovery and isolation of the cofactor he named coenzyme A. After the

176

Joseph S. Fruton

work of Esmond Snell and Feodor Lynen showing that the action of coenzyme A involves a thiol group (47), we learned during the 1950s that thiol esters play a central role in biological acyl transfer. We also learned that in the catalytic activity of enzymes as different as papain and glyceraldehyde-3-phosphate dehydrogenase, cysteine thiol groups are involved in the intermediate formation of thiol esters, as surmised for papain by Weiss in 1937. In the thermodynamic scale of what Lipmann defined as energy-rich bonds, the values for the hydrolysis of thiol esters such as acetyl-CoA rank as high as the currently-accepted values for the hydrolysis of the pyrophosphate bonds of ATP (48). Although reliable free-energy values for the aminolysis of thiol esters in well-defined model systems do not appear to be available, there can be no doubt that such transpeptidation reactions are even more exergonic than the aminolysis of ordinary esters. By the 1950s, therefore, it was clear that, in the biological synthesis of proteins or natural polypeptides from free amino acids, transpeptidation reactions involving amino acid amides, esters, or thiol esters were all as plausible as the aminolysis of the amino acyl phosphates postulated by Lipmann in 1941. Consequently, at that time, the question of the nature of the so-called activated amino acids was still open, although it was generally agreed that, in the formation of at least some naturallyoccurring amide and peptide bonds, ATP is an obligatory participant. During the succeeding years, however, the free-energy value assigned to the hydrolysis of the pyrophosphate bonds of ATP dropped from that originally estimated by Lipmann to about — 7 kcal per mole (49). Although his hypothesis that the activation of amino acids is coupled to the cleavage of ATP had received strong experimental support, it was also evident that thermodynamic considerations alone could contribute little to the identification of the chemical processes in the biological formation of peptide bonds. During the 1950s there were many speculations about the biochemical mechanisms for the utilization of ATP in the formation of proteins from free amino acids, and for the translation of genetic information into the unique amino acid sequence of proteins. I will not attempt here to retell a story that forms one of the most dramatic chapters in the history of biochemistry, since it has been told by many writers, often in different ways. Rather, I will begin with a quotation from a lecture given by Francis Crick in 1957. He proposed that protein synthesis occurs on ... an RNA template in the cytoplasm. The obvious place to locate this is in the microsomal particles, because their uniformity of size suggests that they have a regular structure. It also follows that the synthesis of at least some of the microsomal RNA must be under the control of the DNA of the nucleus... Granted that the RNA of the microsomal particles, regularly arranged, is the template, how does it direct the amino acids into the correct order?... It is... a natural hypothesis that the amino acid is carried to the template by an "adaptor" molecule, and that the adaptor is the part which actually fits on to the RNA. In its simplest form one would require twenty adaptors, one for each amino acid. What sort of molecules such adaptors might be is anybody's guess... [T]here is one possibility which seems inherently more likely than any other - that they might contain nucleotides. This would enable them to join on to the RNA template by the same "pairing" of bases as is found in DNA (50).

Energy-Rich Bonds and Enzymatic Peptide Synthesis

177

In the same year, Mahlon Hoagland, Paul Zamecnik and Mary Stephenson (51) reported the discovery of what we now call tRNA (transfer RNA). Earlier work had shown that the activation of amino acids for the RNA-dependent synthesis of polypeptides involves the cleavage of ATP to AMP and pyrophosphate, with the intemediate formation of aminoacyl-AMP compounds. After the identification of tRNAs as Crick's proposed "adaptors", aminoacyl-tRNAs were recognized to be reactants in the formation of peptide bonds (52). In relation to the theme of this essay, three subsequent developments are of special interest. First, the aminoacyl group was found to be linked to a hydroxyl group of the ribosyl portion of the terminal adenosine unit of tRNA by means of an ester linkage. Then, the elegant studies of Paul Berg (53) indicated that the equilibrium constant in the over-all process is near 1, so that the free-energy change for the hydrolysis of this ester bond appeared to be similar to that for the hydrolysis of a pyrophosphate bond of ATP (ca. — 7 kcal per mole), and not too different from the estimated value for the hydrolysis of ordinary amino acid esters. Finally, some of the enzymes which catalyze the over-all process (aminoacyl-tRNA synthetases) were later reported to effect the hydrolytic cleavage of aminoacyl-tRNA when AMP and pyrophosphate are absent (54). In chemical terms, therefore, the picture that emerged around 1960 was that the activation by ATP of amino acids for protein synthesis had produced a special kind of aminoacyl ester, whose polynucleotide portion provided for the specificity of the entry of the aminoacyl unit into a genetically determined sequence, and which participated in aminolysis reactions during the course of polypeptide formation. The excitement generated in the late 1950s was heightened in the succeeding decade by the deciphering of the genetic code and by the dissection of much of the ribosomal apparatus for the formation of polypeptide chains. Lipmann's laboratory made important contributions, which he summarized in an article published in 1969 (55). I read the article with great admiration, and was interested to note that Lipmann had incorporated the word "transpeptidation" into his scientific vocabulary. For obvious reasons, the section that interested me most was the one entitled "peptidyl transferase" but Lipmann had little to say about this enzyme, except to mention the work of Munro on the use of puromycin in the study of its action. During the 1970s, and to a lesser degree afterwards, I read fairly carefully the published literature on the peptidyl transferase activity of ribosomes, in the hope that someone had succeeded in extracting the purported enzyme and studied its catalytic properties in aqueous solutions. To my knowledge, this has not yet been achieved; indeed, some investigators have surmised that the aminolysis reaction in peptide bond formation occurs spontaneously (56). Such non-enzymatic "zipper" actions were a feature of some of the early template hypotheses during the 1950s, including the one that Lipmann offered in 1953 (57). My interest in peptidyl transferase was twofold. First, here surely was the kind of transferase to which Linderstram-Lang referred in the quotation I cited before - an enzyme, possibly related to the proteolytic enzymes, that preferentially catalyzes transpeptidation reactions. The other reason for my interest was that Thomas Krenitsky (58) had shown that an esterase associated with the microsomal fraction of liver homogenates preferentially catalyzes aminoacyl-transfer reactions involving L-phenyl-

178

Joseph S. Fruton

alanine methyl ester, and that this activity is inhibited by chloramphenicol; these studies in my laboratory were continued and extended by Michael Goldberg (59). I do not wish to suggest the identity of what is officially classified as a liver esterase with the peptidyl transferase activity of liver ribosomes, but I do wish to call attention to the possibility that an enzyme firmly imbedded in a subcellular structure may act almost exclusively as a transferase, but may reveal its capacities as a hydrolase when released into aqueous solutions. As an aside, I would add that I have been surprised by the reluctance of some leading biochemists to consider any revision of the currently-enshrined distinction between transferases and hydrolases, especially for the many enzymes that are known to catalyze the transfer of acyl, glycosyl and phosphoryl groups to water as well as to specific acceptors. Should the presumed physiological role of an enzyme override our knowledge of its chemical action? Indeed, the word hydrolysis appears to be abhorred and, in the case of acyltransfer enzymes, the word deacylation is frequently preferred. Can it be that present-day biochemists reflect a tradition that goes back into the last century, with attitudes such as that of Claude Bernard, who distinguished sharply between the synthetic processes which he believed to be linked to life and the chemical reactions in the breakdown of body constituents? As regards the role of transpeptidation reactions in the biosynthesis of relatively small naturally-occuring polypeptides, larger than glutathione, but smaller than most proteins, we now know that many of the peptide hormones are formed by the cleavage of large proteins by specific proteinases, thus indicating that these enzymes are not only agents of "death and decay". We also know that, apart from such selective proteolysis of proteins made in RNA-dependent systems, there is at least one other biological mechanism for the biosynthesis of polypeptides that does not involve the aminolysis of peptidyl-tRNA by aminoacyl-tRNA. This mechanism is operative in the biosynthesis of some microbial antibiotics, and has been variously named the "protein template mechanism," the "protein thiotemplate mechanism," or the "multienzyme thiotemplate mechanism". In addition to the important contributions of many investigators in Japan, Norway, and the United States (60), this mechanism has been studied intensively by Horst Kleinkauf, first in Lipmann's laboratory in New York (61) and then here in Berlin. From their achievements it is now evident, for example, that a particle-free extract from a strain of Bacillus brevis can synthesize the cyclic decapeptide gramicidin S in a process that does not involve the participation of RNA. Instead, an assembly of protein subunits, each having a 4phosphopantotheine unit to which an aminoacyl group is attached, effects a series of transpeptidation reactions which lead to the production of the repeating pentapeptide unit of the antibiotic. The formation of the thiol ester bonds is coupled to the cleavage of ATP to AMP and pyrophosphate, and the amino acid sequence of the product is determined by the spatial arrangement of the subunits in the catalytic assembly. With variations, this mechanism has been found to be operative in the biosynthesis of the tyrocidines and the linear gramicidins made by other strains of Bacillus brevis. In this essay, I have attempted to summarize some aspects of the recent historical development of a biochemical problem that Lipmann illuminated greatly through his important experimental contributions. In telling the story from my own

Energy-Rich Bonds and Enzymatic Peptide Synthesis

179

perspective, I have not concealed the fact that there were tensions derived in large part from differences in scientific outlook. Since I am no longer a working biochemist, but a historian, I may perhaps be permitted the observation that such differences, no matter how unpleasant they may be to scientists as persons, are frequently valuable for the development of the field of their common interest. In an article (62) entitled "The Emergence of Biochemistry," I wrote: There is little evidence of a linear historical progression within a single scientific discipline toward the so-called mature biochemistry of today, and the continuity of the biochemical enterprise may be seen rather in the competition among attitudes and approaches derived from different parts of chemistry and biology. Inevitably, such competition is attended by tensions among the participants. I venture to suggest that this competition and these tensions are the principal source of the vitality of biochemistry and are likely to lead to unexpected and exciting novelties in the future, as they have in the past.

Acknowledgement The preparation of this paper was aided by grant RH-20739-86 from the National Endowment for the Humanities. I am also grateful to Drs. Sidney Altman, Frederic L. Holmes, Ann Körner, Sofia Simmonds and Dieter G. Söll for their valuable comments on early drafts of this paper.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

Lipmann, F. 1941. Adv. Enzymol. 1, 99-162. Kalckar, H.M. 1941. Chem. Rev. 28, 71-178. Herriott, R.M. 1955. J. Chem. Ed. 32, 513-517 (1955). Lewis, G . N . and M. Randall. 1923. Thermodynamics and the Free Energy of Chemical Substances. McGraw-Hill, New York. Wald, G. 1948. J. Gen. Physiol. 32, 1 - 1 6 . Vickery, H.B. 1975. Biog. Mem. Natl. Acad. Sci. U.S. 46, 3 - 2 0 . Fruton, J.S. and H.T. Clarke. 1934. J. Biol. Chem. 106, 667-691. Gorin, G., A. Esfandi, and G.B. Guthrie. 1975. Arch. Biochem. Biophys. 168, 450-454. Fruton, J.S. 1982. TIBS 7, 37-39. Fruton, J.S. 1979. Ann. N. Y. Acad. Sci. 325, 1 - 1 5 . Bergmann, M. and H. Fraenkel-Conrat. 1937. J. Biol. Chem. 119, 707-720. Bergmann, M. and J.S. Fruton. 1938. J. Biol. Chem. 124, 321-329. Wasteneys, H. and H. Borsook. 1930. Physiol. Rev. 10, 110-145. Fruton, J.S. 1938. Cold Spring Harbor Symp. Quant. Biol. 6, 50-57. Bergmann, M. and C. Niemann. 1937. J. Biol. Chem. 118, 301-314. Chibnall, A.C. 1942. Proc. Roy. Soc. B113, 136-160. Gordon, A.H., A.J.P. Martin, and R.L.M. Synge. 1941. Biochem. J. 35, 1369-1387. Chantrenne, H. 1966. In: Current Aspects of Biochemical Energetics (N.O. Kaplan and E.G. Kennedy, eds.). Academic Press, New York, pp. 33-37. Fruton, J.S. 1941. Cold Spring Harbor Symp. Quant. Biol., 9, 211-216. Lipmann, F. 1960. In: Molecular Biology (D. Nachmansohn, ed.). Academic Press, pp. 37-47.

180 21. 22. 23. 24. 25. 26.

Joseph S. Fruton

Ostwald, W. 1900. Ζ. physik. Chem. 34, 248-252. Meyerhof, O. 1930. Die Chemischen Vorgänge im Muskel. Springer, Berlin, p. 38. Pamas, J., P. Ostern, and T. Mann. 1934. Biochem. Z. 272, 64- 70. Lipmann, F. 1950. Harvey Lectures 44, 99-123. Huffman, Η.M. 1942. J. Phys. Chem. 46, 885-890. Linderstrem-Lang, K. 1952. Lane Medical Lectures: Proteins and Enzymes. Stanford Univ. Press, pp. 93-115. 27. Carpenter, F.H. 1960. J. Am. Chem. Soc. 82, 1111-1122. 28. Dobry, Α., J.S. Fruton and J.M. Sturtevant. 1952. J. Biol. Chem. 195, 149-154. 29. Homandberg, G.A., J. A. Mattis and M. Laskowski, Jr., 1978. Biochemistry 17, 5220-5227. 30. Fruton, J.S. 1949. Adv. Protein Chem. 5, 1-82. 31. Inouye, K., K. Watanabe, K. Morihara, Y. Tochino, T. Kanaya, J. Emura and S. Sakakibara. 1979. J. Am. Chem. Soc. 101, 751-752. 32. Fruton, J.S. 1982. Adv. Enzymol. 53, 239-306. 33. Neurath, Η. and G.W. Schwert. 1950. Chem. Rev. 46, 69-153. 34. Fruton, J.S. 1950. Yale J. Biol. Med. 22, 263-271. 35. Weiss, J. 1937. Chem. Ind., pp. 685-686. 36. Elliott, W.H. 1948. Nature 161, 128-129; Speck, J.R., 1949. J. Biol. Chem. 179, 1405-1426. 37. Bloch, Κ. 1949. J. Biol. Chem. 179, 1245-1254. 38. Hanes, C.R., F.J.R. Hird and F.A. Isherwood. 1950. Nature 166, 288-292. 39. Meister, A. and S. Tate. 1976. Ann. Rev. Biochem. 45, 559-604. 40. Hartman, S.C. 1971. In: The Enzymes, 3rd Ed. (P.D. Boyer, ed.). Academic Press, New York, vol.4, pp.79-100. 41. Durell, J. and J.S. Fruton, 1954. J. Biol. Chem. 207, 487-500. 42. Dowmont, Y.P. and J.S. Fruton. 1952. J. Biol. Chem. 197, 271-283. 43. Dobry, A. and J.M. Sturtevant. 1952. J. Biol. Chem. 195, 141-147. 44. Brenner, M., H.R. Müller and R.W. Pfister. 1950. Helv. Chim. Acta 33, 568-591. 45. Bender, M.L. and F.J. Kezdy. 1965. Ann. Rev. Biochem. 34, 49-76. 46. Lipmann, F. 1971. Wanderings of a Biochemist. Wiley, New York, p. 80. 47. Lynen, F., E. Reichert and L. Rueff. 1951. Ann. Chem. 574, 1-32. 48. Burton, K. 1955. Biochem. J. 59, 44-46. 49. Benzinger, T.H., C. Kitzinger, R. Hems and K. Burton. 1959. Biochem. J. 71, 400-407. 50. Crick, F.H.C. 1957. Symp. Soc. Exp. Biol. 12, 138-163. 51. Hoagland, M.B., P.C. Zamecnik and M.L. Stephenson. 1957. Biochim. Biophys. Acta 24, 215-216. 52. Zamecnik, P.C. 1979. Ann. Ν. Y. Acad. Sci. 325, 269-301. 53. Berg, P., F.H. Bergmann, E.J. Ofengand and M. Dieckmann. 1961. J. Biol. Chem. 236, 1726-1734. 54. Söll, D. and P.R. Schimmel. 1974. In: The Enzymes, 3rd Ed. (P.D. Boyer, ed.). Academic Press, New York, vol. 10, pp. 489-538. 55. Lipmann, F. 1969. Science 164, 1024-1031. 56. Krayevsky, A.A. and M.K. Kukhanova. Prog. Nucleic Acid Res. 23, 1-51. 57. Lipmann, F. 1954. In: Mechanism of Enzyme Action (W.D. McElroy and B. Glass, eds.). Hopkins, pp. 599-604. 58. Krenitsky, T. and J.S. Fruton. 1966. J. Biol. Chem. 241, 3347-3353. 59. Goldberg, M. and J.S. Fruton. 1969. Biochemistry 8, 86-97. 60. Kurahashi, K. 1981. In: Antibiotics IV. Biosynthesis (J. Corcoran, ed.). Springer, Berlin, Heidelberg, New York, pp. 325-352. 61 Lipmann, F. 1973. Acc. Chem. Res. 6, 361-367. 62 Fruton, J.S. 1976. Science 192, 327-334.

A Nostalgic View of the TCA Cycle in Bacteria Lester O. Krampitz

The classical investigations by Krebs and his coworkers (1) firmly established the tricarboxylic acid cycle in animal tissues. The criteria used to formulate the cycle were more easily met by employing animal tissue. These criteria were: 1. rates of reaction, i.e. for a proposed compound to be an intermediate in a sequence of metabolic reactions it must be converted to the same end products as the parent compound and at an equal or greater rate. 2. Use of specific inhibitors of the metabolism of intermediates to accumulate those intermediates from the parent compound. These criteria and others enabled Krebs and coworkers to establish the cycle in minced animal tissues. When these same criteria were applied to microorganisms, evidence for the cycle was inconclusive. For example, with intact cells of Escherichia coli, Micrococcus lysodeikticus and yeast, citrate was not or only slowly oxidized. Furthermore, with a minimal medium, E. coli will not grow with citrate as the sole carbon source for energy. In addition, malonate inhibition which was used with minced animal tissues so successfully to accumulate succinate during oxidation of substrates by the TCA cycle did not give similar results with cells of the microorganisms. Although many of the individual reactions involved in the TCA cycle were demonstrated in various microorganisms, their order of sequence could not be established. A concept emerged that the cycle operated to a limited extent for the purpose of forming carbon skeletons for cell constituents but that other mechanisms were involved for oxidation of food stuffs for energy purposes. Therefore, other approaches were employed to determine if the cycle or other mechanisms were present for energy yielding purposes. When the isotopes of carbon became available, it was thought that one could readily determine the mechanism of oxidation in microorganisms and do so quantitatively. The following simple expedient was developed by several interested in the problem (2,3). An isotopically labeled substrate was oxidized by the organism in question in the presence of a pool of non-isotopic compound suspected of being an intermediate in the oxidation process. The supposition being that an isotopie intermediate derived from the isotopie substrate would equilibrate with the non-isotopic pool which was large enough to permit its isolation and determination of its isotope content. Table 1 presents the results of such and experiment (2). On examination of the specific radioactivities of the compounds involved in Table 1 for M. lysodeikticus there are serious inadequacies in the experimental approach. For example, the specific activities of the carbon atoms of a-ketoglutarate are lower than the corresponding atoms of succinate. If a-ketoglutarate is a precursor of succinate, as it is in the TCA cycle, the specific activity of the corresponding carbon atoms of the former cannot be lower, but must be equal or greater than those of succinate. If the specific activity in the a-ketoglutarate is greater, one can explain the results by postulating a source of succinate from some other nonisotopic sources

182

Lester O. Krampitz

Table 1 Oxidation of acetate-2- 1 4 C by M. lysodeikticus ketoglutarate and succinate Succinate COOH

cpm/mmol

COOH HLCJ

5,218

H

2C J COOH

a-Ketoglutarate

2,943

H.Ç1 H

in the presence of nonisotopic α-

2,943

661

2,241

290%

6.3 3.9 136.0 10.2 9.9 1,765 3,290 -

* For particulars of experiments, see ref. (4).

Inasmuch as the results from carrier type experiments with E. coli were interpreted as showing that the oxidation of acetate occurred by means of the C 4 dicarboxylic acid cycle, it is interesting to analyze the results obtained with this organism by use

A Nostalgic View of the TCA Cycle in Bacteria

187

of the noncarrier technique. In contrast to carrier type experiments where isotopie acetate was incorporated into carrier succinate to a high degree, but into aketoglutarate or citrate in insignificant amounts, acetate-2- 14 C has completely equilibrated with citrate and a-ketoglutarate as well as the C 4 dicarboxylic acids (see Table 5). Actually, the residual acetate has a lower specific activity than the other acids, indicating sources of nonisotopic acetate from endogenous substances which are oxidized to acetate. The data depicting the distribution pattern of the isotope in the carbon atoms of the various acids are not shown, but follow precisely the pattern expected from the oxidation of acetate-2- 14 C by the TCA cycle. Furthermore, the total recovery of 14 C from all of the isolated compounds is greater than 90%, indicating the quantitative importance of the cycle. It can be concluded that M. lysodeikticus and E. coli, the former a strict aerobe and the latter a facultative anaerobe, both oxidize acetate by the TCA cycle. The reader is also referred to the work by Ajl and Wong (8) for an appraisal of the TCA cycle in E. coli.

Quantitative Aspects of the TCA Cycle in E. coli Krebs (9) demonstrated that acetate and several other substrates were oxidized under anaerobic conditions by non-proliferating cells of E. coli when fumarate was added as the oxidant. The following equation illustrates the stoichometries: Acetate + 4 fumarate + 2 H 2 0

2 C 0 2 + 4 succinate

Krebs was concerned at the time with the role of fumarate in respiration and no consideration was given to the mechanism of oxidation of acetate. The equation shows that 2 moles of C 0 2 are produced which are equivalent to the carbon atoms of acetate and Krebs concluded that the acetate was oxidized to C 0 2 by some unknown mechanism with fumarate as the oxidant. Swim and Krampitz (10) repeated these experiments substituting CH 3 — 1 4 COOH for nonisotoptic acetate. The results are shown in Table 6. Table 6

Oxidation of acetate-l- 14 C by E. coli in presence of fumarate

Compound Acetate (final) Succinate Fumarate Malate Respiratory C 0 2 Initial CH 3 14 COOH Initial fumarate

Quantitity (μηιοί)

Specific activity (cpm/μηιοΐ)

Added

24.3 370 31.8 57.6 162 100 500

866 362 41 46 23 1,650

12.7 81.2 0.8 1.6 2.2

-

14

C

(%)

-

If the carbon dioxide arose from the carbon atoms of acetate, the carbon dioxide would contain all of the radioactivity of the initial acetate. In fact, the specific activity of the carbon dioxide was very low, indicating that only a small fraction of the initial acetate was oxidized to C 0 2 . Upon closer examination these results can be

Lester O. Krampitz

188

interpreted to be in agreement with the TCA cycle. To initiate the TCA cycle, the following steps would occur. (1) Two molecules of fumarate would undergo a dismutation forming one molecule of oxalacetate and one molecule of succinate. (2) The acetate (acetyl CoA) would condense with the oxalacetate to form citrate, with carbons 1 and 2 of the citric acid molecule containing carbon atoms originally present in the acetate. Carbon number 1 (carboxyl group) would contain the 14 C from CH 3 · 14 COOH). (3) Transformation to a-ketoglutarate would then occur and during the process nonisotopic carbon number 6 of citric acid would be evolved as carbon dioxide. Fumarate would act as the electron acceptor for the oxidation of isocitrate to oxalosuccinate, forming another molecule of succinate. (4) The aketoglutarate would be oxidatively decarboxylated to succinate, evolving nonisotopic carbon atom 5 as carbon dioxide. The electrons in the oxidative step would reduce another molecule of fumarate to succinate. In summary, the observed data would be obtained: i.e., the evolution of two molecules of carbon dioxide, which are nonradioactive, and the formation of four molecules of succinate. Three molecules of succinate would result from the reduction of fumarate; one would arise through the operation of the TCA cycle, and would contain the isotope in one carboxyl group from the original isotopie acetate. Succinate would not be oxidized further because of the anaerobic conditions of the experiment, and a dismutation between it and a molecule of fumarate does not occur. The data in Table 6 show that the succinate present at the conclusion of the experiment contained the major portion of the isotope and that the carbon dioxide evolved contained insignificant amounts. Variations from the exact theoretical stoichiometry are due to endogenous ractions taking place in the whole cells. Although these results showed that the carbon atoms of acetate were not oxidized to carbon dioxide and that they could be accounted for by the TCA cycle, it is possible that part of the acetate was utilized by a Thunberg condensation (oxidative condensation of two molecules of acetate to form succinate) and that in some manner the oxidation of fumarate to carbon dioxide was linked with the utilization of acetate. It was possible, however, to determine the extent of the Thunberg condensation by mass analysis of the succinate formed from acetate-2- 13 C ( 13 CH 3 —COOH). By the same technique the extent of the occurence of the TCA cycle can also be ascertained by the mass analysis of the succinate- 13 C. If an experiment similar to the one referred to above is performed, except that acetate-213 C ( 1 3 CH 3 —COOH) and fumarate are the substrates, and the succinate is isolated and examined for 13 C content, the following alternative results could be expected, depending upon whether the Thunberg type of condensation or the TCA cycle occurred. A detailed description of the fate of acetic acid under these conditions was given previously. If 1 3 CH 3 —COOH is substituted for CH 3 — 1 4 COOH it will be seen that the molecular species of the succinate formed by way of the TCA cycle will be: H00 1 2 C— 1 3 CH 2 — 1 2 CH 2 — 1 2 C00H Emphasis should be placed upon the fact that the carbon atom which will contain the isotope of 13 C is one of the methylene carbons. On the other hand, if the Thunberg condensation occurs, the molecular species of the succinate formed will be:

A Nostalgic View of the TCA Cycle in Bacteria

H H I I HOOC—13C—H + H—13C—COOH I I H H

189

HOOC—13CH2—13CH2—COOH + 2 H

Both methylene groups of the succinate will contain the isotope, 13 C. In order to obtain a gaseous product for mass analysis in the mass spectrometer succinate was degraded to ethylene containing only the two methylene carbons of succinate. The normal complement of 13 C in carbon is 1.1%; therefore, any value above that is enriched in 13 C. For illustration purpose the normal complement of 1.1 % 1 3 C is not taken into consideration, however, in the results presented calculations were made taking it into consideration. The molecular species of ethylene are thus: 13

Mass 30 CH 2 = 1 3 CH 2

13

Mass 29 CH 2 = 1 2 CH 2

12

Mass 28 CH 2 = 1 2 CH 2

In the previous discussion, CH 3 1 4 COOH oxidation with fumarate as the oxidant, was shown that, of the four succinates formed, three were derived from the reduction of fumarate; therefore, nonisotopic and one was derived from the oxidation of acetate via the TCA cycle containing isotope. The results of mass spectrometric analysis of the species of ethylene obtained from the isolated succinate in an experiment with 1 3 CH 3 —COOH and fumarate are presented in Table 7. Table 7 Mass spectrometric analysis of ethylenes obtained from fumarate

Molecular species Percent

13

CH 3 —COOH oxidation with

Mass 30

Mass 29

Mass 28

13 CH2=13CH2 0.18

13 CH2=12CH2 21.1

12 CH2=12CH2 78.7

During the oxidation of 1 3 CH 3 —COOH with fumarate as the oxidant three moles of succinate are formed from the reduction of three moles of nonisotopic fumarate and one mole of succinate contains 13 C from the acetate or theoretically 75% of the ethylene should be of the species 1 2 C H 2 = 1 2 C H 2 , mass 28. 25% of the ethylene should contain 13 C. The results depicted in Table 7 reveal that 78.7% of the ethylene had amass of 28,21.1% amass of 29 and only 0.18% amass of 30. If corrections are made for a small recycling of the succinate, the value of an ethylene of mass 30 is so negligible that it becomes in the realm of experimental error. It can therefore be concluded that under these conditions the TCA cycle accounts quantitatively for the metabolism of acetate and fumarate and that the Thunberg condensation of acetate does not occur. Additional evidence of the existence of the TCA cycle in E. coli was provided for by Gilvarg and Davis (11) who employed the microbial mutant analysis approach. The wild-type strain of E. coli can grow on a synthetic medium consisting of minerals, ammonia, and a simple carbon source. They obtained several mutant strains of E.

Lester O. Krampitz

190

coli which would grow on glucose, lactate, or succinate, provided that glutamate was also present. Some stage in the synthesis of glutamate was blocked in these mutant strains, since the wild-type strain would grow on the substrates without the addition of glutamate. The mutant strains would not grow on acetate with glutamate present. Apparently glucose, lactate, or succinate could serve as a source of carbon for the mutant strains, whereas acetate could not. The possibility existed that different mutations had occurred, one pertaining to the utilization of acetate and a second for the synthesis of glutamate. Experience has shown that the occurence of a double mutation is a rare event. Furthermore, Gilvarg and Davis irradiated one of their mutant strains with ultraviolet light to increase the rate of reversion back to the wild type. They selected for reverse mutant strains which had lost the glutamate requirement, and also for ones which had lost the acetate block. Several revertants of each were isolated; every one proved to have lost both blocks. These results showed clearly that the glutamate requirement and the inability to utilize acetate by the mutant strains were related phenomena. After a systematic survey for the location of the enzymic block, it was discovered that the organisms were lacking or very deficient in the enzyme which condenses acetyl CoA and oxalacetate to form citrate, i.e., condensing enzyme. All the other enzymes required for the activation of acetate and the TCA cycle were present in amounts comparable to those in the wild-type strain. The loss of the ability of the mutant strains to synthesize the condensing enzyme readily explains why the organism cannot utilize acetate as an energy source for growth, and also explains the requirement for glutamate. α-Ketoglutarate is the precursor of the carbon skeleton for glutamate synthesis by E. coli. Since a-ketoglutarate is obtained indirectly from citrate by reactions of the TCA cycle and the synthesis of citrate by the mutant cannot occur because of the absence of condensing enzyme, the glutamate requirement is obvious. The fact that the loss of ability of the mutant strain to synthesize one enzyme of the TCA cycle has created conditions under which the organism cannot survive unless special nutritional conditions are satisfied indicates the importance of the TCA cycle to the cell. In summary, the discovery of coenzyme A and its acyl derivatives by Fritz Lipmann contributed immensely to the elucidation of the TCA cycle in bacteria.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

Krebs, H . A . und Johnson, W. A. 1937. Enzymologia 4, 148. Saz, H.J. and Krampitz, L.O. 1954. J. Bacteriol. 67, 409. Ají, S.J., Kamen, M . D . , Ramson, S.L. and Wong, D.T.O. 1951. J. Biol. Chem. 189, 859. Swim, H.E. and Krampitz, L.O. 1954. J. Bacteriol. 67, 419. Lipmann, F. 1953. Bacteriol. Revs. 17, 1. Lynen, F. 1953. Federation Proc. 12, 683. Ochoa, S. 1954. Advances in Enzymol. 15, 183. Ají, S.J. and Wong, D.T.O. 1955. Arch. Biochem. Biophys. 54, 474. Krebs, H . A . 1937. Biochem. J. 31, 2005. Swim, H.E. and Krampitz, L.O. 1954. J. Bacteriol. 67, 426. Gilvarg, C. and Davis, B.D. 1956. J. Biol. Chem. 222, 307.

The Role of Vitamins and their Carrier Proteins in Citrate Fermentation Peter

Dimroth

Introduction Two vitamins, pantothenic acid and biotin, have been shown to be of central importance for the fermentation of citrate in Klebsiella pneumoniae and some other species of Enterobacteriaceae. By virtue of these vitamins, each enzyme of the citrate fermentation pathway, citrate lyase and oxaloacetate decarboxylase, performs an unique type of catalysis. As in other systems in which pantothenate and biotin participate, the vitamins gain biological activity by their covalent attachment to specific proteins. Such carrier proteins are the central parts of the enzyme complexes citrate lyase and oxaloacetate decarboxylase, specially designed to allow the prosthetic groups to get access to the different catalytic sites. Whereas the biological role of biotin exclusively is that of a prosthetic group, pantothenate in addition is a constituent of coenzyme A, as was detected by Fritz Lipmann.

Citrate Lyase Citrate lyase, in the presence of a divalent metal ion, e.g. Mg 2 + , catalyzes the reversible cleavage of citrate to acetate and oxaloacetate (Eq. 1). Mg 2

citrate

(1)

acetate + oxaloacetate

The catalytically active enzyme carries an acetyl-thioester residue at its active site which undergoes continuous turnover during the catalysis (1). The overall mechanism of action involves the consecutive reactions of acyl exchange (A) and acyl cleavage (B), as shown in Scheme 1 below:

VC / HOJCCHJ

Enzyme-S—COCH3

R \

ο II

/

/

-OH

,c; \

RCOCOjH

Enzyme-S-C-H2C

Scheme 1

Citrate: R = CH 2 C0 2 H; Citramalate: R = CH 3

C02H

CH,CO,H

\

•OH

C02H

192

Peter Dimroth

With this mechanism the enzyme elegantly manages the chemical problem of cleavage or formation of the carbon-carbon bond of citrate: the critical step is accomplished with thioester derivatives which are chemically much more reactive than free acids in performing aldol-type condensation reactions. The chemistry of the acyl cleavage reaction on citrate lyase is thus completely analogous to the reversible formation of citryl-CoA from acetyl-CoA and oxaloacetate on citrate synthase. Incubation of citrate lyase with hydroxylamine or mercaptans removed the essential acetyl residues with concomitant loss of catalytic activity (1). Partial hydrolysis of the acetyl-thioesters also occurs during the catalysis and leads thus to its own inactivation (2). The deacetylated enzyme species carrying a free SH-group at the active site recovers catalytic activity by chemical acetylation with e.g. acetic anhydride (1) or by enzymic acetylation with acetate, ATP and a specific ligase (3).

Subunit structure The K. pneumoniae citrate lyse complex, Mt 550,000, is composed of three different subunits with M r 55,000, 32,000, and 10,000 in a stoichiometric relationship of 1:1:1 (4-8). The composition of the lyase is therefore (α,β,γ) 6 . The smallest subunit was detected in experiments performed to determine the binding site of the acetyl residues (4). Dodecylsulfate gel-electrophoresis of citrate lyase labelled with 14 C-acetyl residues at its active site indicated that all of the radioactive label was located in the small subunit, subsequently named acyl carrier protein (ACP) (4). Citrate lyases from other bacteria are likewise composed from 6 α-, 6 β-, and 6 ychains with similar molecular weights as the corresponding subunits of K. pneumoniae (9-12). The only known exception is the lyase from E. coli, where the mass of the y-chain is 85 kDa and the structure of the enzyme is Acyl carrier protein N H

i

ft Figure 2

Structure of the prosthetic group of citrate lyase and citramalate lyase.

The ACP's of citrate lyase and fatty acid synthetase are not only distinct by their different prosthetic groups, but also by their amino acid sequences and by their biological activities. The ACP's of citrate lyase from K. pneumoniae and fatty acid synthetase from E. coli contain 78 and 77 amino acid residues, respectively, and show no obvious sequence homology (23). The 4'-phosphopantetheine is bound to serine-36, in the middle of the fatty acid synthetase ACP (18). In contrast, the binding site for the prosthetic group in the ACP of citrate lyase is serine-14, and thus in the N-terminal region of the protein (23).

The Role of Vitamins and Their Carrier Proteins in Citrate Fermentation

195

As expected from the structural differences, the ACP of citrate lyase could not substitute for the ACP of the E. coli fatty acid synthetase in the malonyl-CoA/C0 2 exchange reaction (4). K. pneumoniae contains a distinct ACP species for fatty acid biosynthesis that is identical or nearly identical to the ACP of fatty acid synthetase from E. coli (4,24). The biological activity of the ACP of citrate lyase was investigated by reconstitution with the catalytic subunits α and β. Citrate lyase activity was not only reconstituted from the homologous combination of the three subunits from the K. pneumoniae enzyme but also with ACP of citrate lyase from Streptococcus diacetilactis or Rhodopseudomonas gelatinosa, and with the ACP of citramalate lyase from Clostridium tetanomorphum (16). Therefore, the structural parts of these ACP's that are functionally important must have been conserved. The least relationship to the ACP of the K. pneumoniae citrate lyase was expected for the ACP of citramalate lyase from C. tetanomorphum·, the two enzymes catalyze different reactions and the two bacteria are phylogenetically very distinct. A comparison of the amino acid sequences of these acyl carrier proteins, however, clearly indicates that the two proteins are related (Figure 3) (unpublished results).

A r g - L e u - G l u - T h r - A l a - V e l - A s n - A r g - A l a - Gin 83 Leu-Arg-Trp-Gin-Gin 7478 Figure 3 Amino acid sequences of the ACP's of citramalate lyase (CML) and citrate lyase (CL) (23, and unpublished results).

196

Peter Dimroth

The two ACP's are of about the same size, have methionine at the ,/V-terminus, and contain no histidine and a single phenylalanine residue. The presence of a single cysteine and tryptophan residue in the citrate lyase ACP is not shared by the citramalate lyase carrier protein. The latter contains a single tyrosine which is missing in the ACP of citrate lyase. The sequence homology is most pronounced in a stretch of 6 identical residues TV-terminal of serine-14, the binding site for the prosthetic group in both proteins. Other homologous regions are distributed over about 3/4 of the N-terminal part of the proteins, whereas the C-terminal sequences are more distinct. It should also be noted that at positions where amino acids are not identical most substitutions are of a conservative character. In two areas deletions are apparent which have occured in each ACP two or three residues apart from each other. The two proline residues present in each ACP molecule were found in identical positions, 9 and 13 residues downstream the binding site of the prosthetic group. These proline residues could provide a hinge by which the prosthetic group acquires the necessary flexibility to move between the catalytic sites on the transferase and lyase subunits. It is quite interesting that two prolines in a region 2 7 - 3 3 residues upstream of the prosthetic group are highly conserved in biotin carrier proteins that also mediate between two different catalytic sites in the respective enzyme complexes. The sequence homology between the ACP's of citrate lyase and citramalate lyase firmly establishes the close relationship between these enzymes which may have evolved from a common ancestral gene. A high degree of structural conservation during evolution is expected in areas of the molecules that interact with other enzymes. These are the regions recognized by the catalytic subunits of the enzyme complexes and the recognition sites for the enzymes catalyzing the synthesis of the prosthetic group and its acetylation to the catalytically active carrier. Enzymes catalyzing deacetylation or degradation of the prosthetic group may also recognize specific sites on the ACP molecules. If multiple interactions would be connected to the same region of the ACP molecule, as can be anticipated for the region around the binding site for the prosthetic group, this would create an enormous evolutionary pressure against change in this region.

Transferase and lyase subunits The overall reaction catalyzed by the transferase subunit of citrate lyase is an ACP transfer from acetate to citrate and vice versa (Eq.2). The reaction probably proceeds via an acetyl-citryl anhydride intermediate by a single displacement mechanism (25). It is thus distinct from that of the related classical Co A transferases; these operate by a double displacement mechanism involving covalent enzyme-CoA derivatives in addition to anhydrides (26). Nonetheless, the transferase of citrate lyase is also a true CoA transferase because acetyl-CoA and citryl-CoA can replace the respective ACP thioesters as substrates (25). Remarkably, acetyldephospho-CoA was also a substrate of the enzyme, but acetyl-4'-phosphopantetheine was completely inactive (25). This specificity for the whole CoA molecule is shared by all known CoA transferases and demonstrates the requirements for the CoA-like prosthetic group in citrate lyase.

The Role of Vitamins and Their Carrier Proteins in Citrate Fermentation

197

The reaction catalyzed by the lyase subunit is the cleavage of citryl-ACP to acetylACP and oxaloacetate (Eq. 3). In the isolated state, the lyase is apparently highly specific for the ACP derivatives, but within the enzyme complex citryl-CoA is also recognized as a substrate (14). Presumably, the isolated lyase is in an inactive conformation and requires the interaction with the ACP subunit for generation of catalytic activity. This property of the lyase subunit of citrate lyase from K. pneumoniae is distinct from the isolated lyase subunit of ci tramalate lyase from C. tetanomorphum which catalyzes the cleavage of citramalyl-CoA to acetyl-CoA and pyruvate (16).

Oxaloacetate Decarboxylase The second step in citrate fermentation by Klebsiella pneumoniae is the decarboxylation of oxaloacetate to pyruvate and C 0 2 . Oxaloacetate decarboxylase is a membrane-bound, biotin-containing enzyme that is specifically activated by N a + ions (27). These properties would be quite unusual if the only function of the enzyme would be catalysis of a step in the pathway of citrate degradation. The membranelinked character of the decarboxylase and its activation by N a + were reminiscent, however, of transport enzymes such as the N a + / K + ATPase and suggested to us that oxaloacetate decarboxylase might act as a N a + pump (28,29). This concept was proven to be true in studies using inverted vesicles of K. pneumoniae cells. The vesicles accumulated N a + ions very rapidly in response to the decarboxylation of oxaloacetate, but virtually no N a + was taken up in the absence of the substrate or after inhibition of the decarboxylase with avidin (28,29). Thus, the free energy of oxaloacetate decarboxylation (Δ G°' « — 28 kJ/mol) is used to drive an active transport of N a + ions. This is a completely unexpected and new type of conversion of chemical energy which exists in addition to the well-known energy conservation mechanisms of electron transport and substrate-level phosphorylation. Under catalysis of the decarboxylase, a N a + concentration gradient and a membrane potential were generated which at the steady state corresponded to values of 49 and 65 mV, respectively (29). The total sodium motive force Δ (I (Na + ), which according to Eq. 4 is the sum of electrical and chemical potential, therefore amounts to 114mV. ®

=

(4)

After the function of oxaloacetate decarboxylase as a N a + pump was established, methylmalonyl-CoA decarboxylase of Veillonella alcalescens (30,31) and glutaconyl-CoA decarboxylase of Acidaminococcus fermentans (32) were found to be analogous N a + pumps. Until now the sodium ion transport decarboxylases have exclusively been found in anaerobic bacteria where they perform a dual function: each enzyme catalyzes an essential step of a specific fermentation pathway; the decarboxylation is coupled to N a + pumping thereby conserving the free energy of this reaction (for previous reviews see 33-35).

198

Peter Dimroth

Molecular and catalytic properties of oxaloacetate decarboxylase and related enzymes The biotin content of oxaloacetate decarboxylase as well as of other N a + transport decarboxylases has greatly facilitated their purification, which is achieved by affinity chromatography on monomeric avidin-Sepharose columns (36,37). A comparison of the subunit composition of three different decarboxylases (Table 1) indicates a remarkable degree of homology (31,32,38). Each enzyme contains a subunit in the range of M r 60,000-65,000 (α-chain) and a subunit in the range of Mt 33,000-35,000 (/5-chain). Another weight-homologous subunit of oxaloacetate decarboxylase and methylmalonyl-CoA decarboxylase is the y-chain (M r 12,000 and 14,000, respectively). A marked difference between the three enzymes exists with respect to the biotin carrier peptide: in oxaloacetate decarboxylase of K. pneumoniae this is integrated within the α-chain (38), whereas methylmalonyl-CoA decarboxylase of V. alcalescens (31) and glutaconyl-CoA decarboxylase oi A. fermentons (32) contain separate biotin carrying subunits with M r 18,500 and 120,000, respectively. It is remarkable that the sodium transport decarboxylases, which have three different substrate specificities and are found in bacteria that are not closely related phylogenetically, share a similar composition of a- and ß-subunits. These subunits are therefore expected to perform analogous functions in the enzyme complexes and may have highly conserved primary structures.

Figure 4 Hypothetical mechanism of oxaloacetate decarboxylase. The carboxyltransferase activity is represented by reaction A, and the carboxybiotin enzyme decarboxylase activity by reaction B. (Reproduced from ref. 34 with permission.)

The Role of Vitamins and Their Carrier Proteins in Citrate Fermentation Table 1

199

Subunit composition of the N a + transport decarboxylases

Enzyme

Molecular weight

Organism

Oxaloacetate decarboxylase K. pneumoniae

Methylmalonyl-CoA decarboxylase V. alcalescens

α-chain /¡-chain y-chain Biotin-carrier

65,000 34,000 12,000 α-chain

60,000 33,000 14,000 18,500

Glutaconyl-CoA decarboxylase A. fermentons 60,000 35,000 15,000 (?) 120,000

Studies on the decarboxylation mechanism The mechanism of oxaloacetate decarboxylation involves two consecutive steps (Figure 4) (38,39). The first step is a transfer of the carboxyl group from position 4 of oxaloacetate to the biotin prosthetic group. The thus generated jV-carboxybiotin enzyme intermediate is subsequently decarboxylated and thereby regenerates the free biotin enzyme. During the course of the reaction sequence, N a + ions are transported from the interior of the bacterial membrane to the exterior. The carboxyltransferase activity was demonstrated by the isotopie exchange between [l- 1 4 C]pyruvate and oxaloacetate and also by the transfer of 1 4 CO z from [4- 14 C]oxaloacetate to the biotin prosthetic group on the enzyme (38,39). Unlike the overall reaction, the carboxyltransferase activity is completely independent from the presence of N a + ions. The α-chain was shown to be the subunit responsible for the carboxyltransfer reaction (38). Upon freezing and thawing of the isolated membranes in presence of 1 M LiCl, the oxaloacetate decarboxylase complex dissociated releasing the α-chain into the supernatant. The solubilized a-subunit was subsequently purified to homogeneity by chromatography on a monomelic avidinSepharose column, and was shown to be a catalytically active carboxyltransferase (38). The α-chain was completely soluble in the absence of detergents and has thus properties of a peripheral membrane protein, comparable to the F t moiety of F j F 0 ATPases (40). In contrast, β- and y-subunits are highly hydrophobic proteins comparable to those forming the F 0 part of the ATPases. It is obvious that the membrane-bound β- and y-subunits are responsible for the transport of N a + ions across the membrane. A mixture of the two subunits catalyzed the Na + -dependent decarboxylation of the carboxybiotin enzyme. The decarboxylase activity is therefore probably coupled to the transport of N a + ions. A binding site for N a + is apparently located on the ß-subunit because this protein was specifically protected by N a + ions from tryptic hydrolysis (38). A decarboxylation mechanism involving two different catalytic sites, as indicated from the activities of the isolated subunits (38), is completely in accord with results from enzyme kinetic studies (41). A two-site mechanism is thus a common feature of all classes of biotin enzymes: carboxylases, transcarboxylase and decarboxylases (for reviews see 42-44). The carboxylases combine a biotin carboxykinase with a carboxyltransferase, transcarboxylase has two carboxyltransferases with different

200

Peter Dimroth

substrate specificity, and the decarboxylases consist of a carboxyltransferase and a carboxybiotin enzyme decarboxylase. Thus, all biotin enzymes contain a carboxyltransferase. Another common feature is the presence of a biotin carboxyl carrier peptide, which either can exist as a distinct subunit, or as a domain of a multifunctional polypeptide chain carrying one or all two catalytic activities in addition (44). The relationship among different biotin-dependent carboxylases and transcarboxylase is apparent from a remarkable degree of homology around the biotin binding site, even in enzymes from such distinct sources as bacteria and animals (45-49). Structural work was therefore performed on oxaloacetate decarboxylase to determine whether this relationship extends to this representative of the N a + transport decarboxylases. Limited proteolysis of oxaloacetate decarboxylase with trypsin, chymotrypsin or thermolysin always cleaved the α-chain at about the same site yielding a biotin-free fragment of about 51 kDa (38). The other biotin-containing fragment, obtained by tryptic hydrolysis, was isolated by avidin-Sepharose affinity chromatography (36). A single polypeptide of about 12 kDa was obtained. The sum of the molecular masses of the two fragments is about that of the native α-chain. These results suggest a construction of the α-chain from two different domains, a carboxyltransferase and a biotin carrier peptide. This construction is similar to that of other biotin enzymes. It is noticeable that concomitant with the fragmentation of the α-chain, the oxaloacetate decarboxylase activity disappeared. The covalent attachment of the two domains in the α-chain is therefore important for its function. Further homology of the decarboxylase α-chain with other biotin enzymes became apparent when the complete sequence of this subunit was determined by nucleotide sequencing of the cloned gene (50, and unpublished results). The biotin prosthetic group is bound to lysine, located 35 amino acid residues from the C-terminus, which is exactly the same position as in all other biotin enzymes investigated. The highly conserved region around the biotin binding site is also shared by the α-chain. Among the enzymes for which sequences are available, the greatest homology was found to the biotin peptide of transcarboxylase where 42 of the 86 C-terminal amino acid residues were identical. Interestingly, all biotin carrier proteins including that of the decarboxylase contain two proline residues in a region 27-33 amino acids upstream of the biocytin. Two conserved proline residues in proximity of the prosthetic group have also been found in the ACP's of citrate lyase and citramalate lyase. These residues may therefore be functionally important for carrier proteins that mediate between two different catalytic sites. The proline residues could provide a hinge by which the prosthetic groups obtain the necessary flexibility to flip-flop between these two sites.

Studies on the transport of Na + ions For proper investigations of the N a + transport activity of oxaloacetate decarboxylase, the purified enzyme was reconstituted into liposomes (51). These proteoliposomes were capable of accumulating N a + ions upon oxaloacetate decarboxylation, thereby creating a membrane potential of about 60 mV and a N a +

The Role of Vitamins and Their Carrier Proteins in Citrate Fermentation

201

Figure 5 Kinetics of oxaloacetate decarboxylation and N a + transport into reconstituted proteoliposomes. N a + accumulated (·); N a + accumulated in presence of 0.2 mM monensin ( · ) ; oxaloacetate decarboxylated (•); oxaloacetate decarboxylated in presence of 0.2 mM monensin (Δ). (Reproduced from ref. 34 with permission.)

concentration gradient equivalent to 50 mV (34). From the initial rates of N a + uptake and oxaloacetate decarboxylation, a stoichiometry of 2 N a + ions per decarboxylation step was determined (Figure 5) (34,52). It is noticeable that the stoichiometry changed when steeper gradients of N a + ions had developed, because decarboxylation continued even after a constant internal N a + concentration was reached. Even then partial coupling persisted, because the decarboxylation rate increased about twofold after monensin was added which abolished the N a + gradient (34,52). With the reconstituted proteoliposomal system but not with the soluble enzyme, the decarboxylation of oxaloacetate could be reversed (52). To make the carboxylation energetically feasible, an electrochemical N a + gradient of proper direction and magnitude was required. This could be generated by partially decarboxylating oxaloacetate, and the reversibility could therefore be measured by the isotopie exchange between 1 4 C 0 2 and oxaloacetate (Eq. 5). oxaloacetate + 2 Na + o u t + H + ^ 14

pyruvate + C 0 2 + 2 N a + i n

(Eq. 5)

The C0 2 -oxaloacetate exchange catalyzed by proteoliposomes containing oxaloacetate decarboxylase was completely dependent on the N a + ion gradient

202

Peter Dimroth

established during decarboxylation of part of the substrate because no exchange took place in the presence of monensin, which abolished the gradient (52). Another possibility of generating an electrochemical N a + gradient was by incorporating a second N a + pump into the same proteoliposomal membrane. Proteoliposomes containing oxaloacetate decarboxylase and methylmalonyl-CoA decarboxylase functioned as a transcarboxylase. The N a + gradient developed by one of the enzymes acting as decarboxylase was used by the other enzyme to drive the carboxylation of its decarboxylated substrate (Figure 6) (52). This system catalyzed the carboxylation of acetyl-CoA to malonyl-CoA by decarboxylation of oxaloacetate to pyruvate and vice versa. The N a + circuit provides the energetic coupling between the exergonic decarboxylation and the endergonic carboxylation reaction. No carboxylation, therefore, occured if the N a + gradient was dissipated by monensin. These are the first demonstrated examples of unfavorable carboxylation reactions being energized by an N a + gradient rather than by ATP hydrolysis. In summary, these experiments have shown that the two vitamins pantothenic acid and biotin function as prosthetic groups in the citrate fermentation enzymes citrate lyase and oxaloacetate decarboxylase. Each vitamin is attached to a specific carrier protein within the two enzyme complexes; the protein structures of the acyl carrier protein and the biotin carrier protein domain appear to contribute considerably to the biological function of these vitamins.

Figure 6 N a + circuit mediating the transcarboxylation from oxaloacetate and acetyl-CoA to pyruvate and malonyl-CoA and vice versa. (Reproduced from ref. 52 with permission.)

The Role of Vitamins and Their Carrier Proteins in Citrate Fermentation

203

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48.

Buckel, W., V. Buschmeier and H. Eggerer. 1971. Hoppe-Seyler's Z. Physiol. Chem. 352, 1195. Singh, M. and P.A. Srere. 1971. J. Biol. Chem. 246, 3847. Schmellenkamp, H. and H. Eggerer. 1974. Proc. Natl. Acad. Sci. USA 71, 1987. Dimroth, P., W. Dittmar, G. Walther and H. Eggerer. 1973. Eur. J. Biochem. 37, 305. Singh, M., D.E. Carpenter and P.A. Srere. 1974. Biochem. Biophys. Res. Commun. 59, 1211. Dimroth, P. and H. Eggerer. 1975. Eur. J. Biochem. 53, 227. Carpenter, O.E., M. Singh, E.G. Richards and P.A. Srere. 1975. J. Biol. Chem. 250, 3254. Singh, M., P.A. Srere, D.G. Klapper and J.D. Capra. 1976. J. Biol. Chem. 251, 2911. Singh, M. and P.A. Srere. 1975. J. Biol. Chem. 250, 5818. Hiremath, S.T., S. Paranjpe and C. SivaRaman. 1976. Biochem. Biophys. Res. Commun. 72, 1122. Giffhorn, F. and G. Gottschalk. 1978. FEBS Lett. 96, 175. Antranikian, G., C. Klinner, A. Kümmel, D. Schwanitz, T. Zimmermann, F. Mayer and G. Gottschalk. 1982. Eur. J. Biochem. 126, 35. Nilekani, S.P. and C. SivaRaman. 1983. Biochemistry 22, 4657. Dimroth, P. and H. Eggerer. 1975. Proc. Natl. Acad. Sci. USA 72, 3458. Buckel, W. and A. Bobi. 1976. Eur. J. Biochem. 64, 255. Dimroth, P., W. Buckel, R. Loyal and H. Eggerer. 1977. Eur. J. Biochem. 80, 469. Basu, Α., S. Subramanian, L.S., Hiremath and C. SivaRaman. 1983. Biochem. Biophys. Res. Commun. 114, 310. Prescott, D.J. and P.R. Vagelos. 1972. Adv. Enzymol. 36, 269. Dimroth, P. 1975. FEBS Lett. 51, 100. Dimroth, P. 1976. Eur. J. Biochem. 64, 269. Robinson, J.B., M. Singh and P.A. Srere. 1976. Proc. Natl. Acad. Sci. USA 73, 1872. Dimroth, P. and R. Loyal. 1977. FEBS Lett. 76, 280. Beyreuther, Κ., H. Böhmer and P. Dimroth. 1978. Eur. J. Biochem. 87, 101. Bayer, E. and H. Eggerer. 1978. Eur. J. Biochem. 86, 203. Dimroth, P., R. Loyal and H. Eggerer. 1977. Eur. J. Biochem. 80, 479. Jencks, W.P. 1973. In: The Enzymes (P. D. Boyer, ed.). 3rd edn., vol. 9, part B, p. 483. Academic Press, New York. Stern, J.R. 1967. Biochemistry 6, 3545. Dimroth, P. 1980, FEBS Lett. 122, 234. Dimroth, P. 1982. Eur. J. Biochem. 121, 443. Hilpert, W. and P. Dimroth. 1982. Nature (Lond.) 296, 584. Hilpert, W. and P. Dimroth. 1983. Eur. J. Biochem. 132, 579. Buckel, W. and R. Semmler. 1983. Eur. J. Biochem. 136, 427. Dimroth, P. 1982. Biosci. Rep. 2, 849. Dimroth, P. 1985. Ann. N.Y. Acad. Sci. 447, 72. Dimroth, P. 1987. Microbiol. Rev. 51, 320. Dimroth, P. 1982. FEBS Lett. 141, 59. Dimroth, P. 1986. Meth. Enzymol. 125, 530. Dimroth, P. and A. Thomer. 1983. Eur. J. Biochem. 137, 107. Dimroth, P. 1982. Eur. J. Biochem. 121, 435. Hatefi, Y. 1985. Ann. Rev. Biochem. 54, 1015. Dimroth, P. and A. Thomer. 1986. Eur. J. Biochem. 156, 157. Moss, J. and M.D. Lane. 1971. Adv. Enzymol. 35, 321. Wood, H.G. and R.E. Barden. 1977. Ann. Rev. Biochem. 46, 385. Obermayer, M. and F. Lynen. 1976. Trends Biochem. Sci. 1, 169. Maloy, W.L., B.U. Bowien, G.K. Zwolinski, K.G. Kumar, H.G. Wood, L.H. Ericsson and K. A. Walsh. 1979. J. Biol. Chem. 254, 11615. Sutton, M. R., R. R. Fall, A. M. Nervi, A. W. Alberts, P. R. Vagelos and R. A. Bradshaw. 1977. J. Biol. Chem. 252, 3934. Rylatt, D.B., D.B. Keech and J.C. Wallace. 1977. Arch. Biochem. Biophys. 183, 113. Takai, T., Κ. Wada and T. Tanabe. 1987. FEBS Lett. 212, 98.

204 49. 50. 51. 52.

Peter Dimroth Lamhonwah, A.M., F. Quan and R.A. Gravel. 1987. Arch. Biochem. Biophys. 254, 631. Schwarz, E. and D. Oesterhelt. 1985. EMBO J. 4, 1599. Dimroth, P. 1981. J. Biol. Chem. 256, 11974. Dimroth, P. and W. Hilpert. 1984. Biochemistry 23, 5360.

Lipmann's Influence on Firefly Luminescence W.D. McElroy

Introduction In 1940 when I was an undergraduate at Stanford University, I had the opportunity to participate in a microbiology course taught by Professor C.V. van Niel at the Hopkins Marine Station. This was where I first learned about the dynamic role played by phosphate in metabolic processes. Although I was a biology major at Stanford, this was my first introduction to the concept of phosphate bond energy and the importance of ATP. This was a new concept for most people in American biochemistry. In most universities phosphate was introduced as an essential macronutrient. The foundations of the "Dynamic Aspects of Biochemistry" which was being formulated in the English and European laboratories in the 1930-1940's had yet to migrate across the Atlantic. There were a few exceptions, and this changed rapidly during World War II when many outstanding biochemists from Europe came to the United States. In van Niel's microbiology reading list was a preprint by Fritz Lipmann on phosphate bond energy, which we all read several times during the summer of 1940. We later found out that this article would appear in Volume 1 of Advances in Enzymology in 1941. Not known to me at the time was a review by Herman Kalckar on the same subject, which was to appear in Chemical Reviews, also in 1941. I was convinced immediately that ATP must play a significant role in all processes requiring metabolic energy. It was to have a big influence on my thinking about the mechanism of bioluminescence. But it is necessary to review a little history before we consider ATP in light emission.

Discovery of Firefly Luciferin and Function of ATP It was in van Niel's course that I was first introduced to luminous bacteria, but it was not until I entered graduate school at Princeton University in 1941 that I became interested in bioluminescence. This was due entirely to Professor E. Newton Harvey, who was considered the father of this science in the United States. He introduced me to many different luminous forms, and I spent much of my spare time reading his books and papers on "Living Light". What became clear to me was that luminescence was an unusual process requiring a chemical reaction that could, in one step, generate 50-60 kilocalories in order to create an excited state that could emit visual light. This did not fit in with the concepts of Lipmann and Kalckar that most chemical reactions in biological systems took place with small energy changes

Abbreviations: LH 2 = luciferin; L = dehydroluciferin; L = 0 = oxyluciferin

206

W.D. McElroy

to insure reversibility. Thus the importance of phosphate bond energy became very clear as a mechanism for conserving energy and insuring reversibility. But this energy came in packages of 8 - 1 0 kcal, not 50-60 kcal. At the time, to my knowledge, there was only one other way to create these excited states in nature, and that was by using light itself, namely in photosynthesis and vision. It was this type of thinking that led one of my students, Dr. B. L. Strehler, and Dr. Bill Arnold to look for light emission in green plants, which they discovered while working together at Oak Ridge National Laboratories. These observations did not discourage me in my belief that in some way phosphate bond energy was involved in luminescence. In the years 1943-1946 I spent full time on war research, but my weekends and evenings were free to read and think about luminescence. This is when I learned the most about luminescence in casual conversations with Professor Harvey. He told me about the interesting experiments of R. DuBois in 1885-86 on the elaterid beetle, Pyrophorus, and the luminous clam, Pholas. He was able to obtain a cell-free extract that would emit light at room temperature. When he ground up fresh samples of these organisms, the crude extract would glow for a few minutes, but the light soon disappeared. He was able to restore light by adding a second extract that he obtained by heating a fresh sample of organisms to a temperature that prevented luminescence. By studying these extracts, DuBois concluded that there was present a heat stable substance that was used up in the light reaction. He suggested that this substance be called "luciferine". He further suggested that the substance that was destroyed by heat was a catalyst, and he proposed the term "luciferase". This concept of a luciferin-luciferase reaction (LH 2 + E + 0 2 -> light + L = 0 ) would dominate the field for the next 60-70 years. In 1946 I accepted a position in the Biology Department at Johns Hopkins University. That spring I realized that Baltimore was an excellent place for collecting large numbers of fireflies. With the aid of graduate students and local youngsters, we collected enough fireflies to enable us to repeat DuBois' experiment. It was easy to demonstrate that a hot water extract from lanterns would restore light in a cold water extract. The next logical step was to purify the presumed luciferin from the hot water extract and to study its properties. I soon learned that it could be precipitated with mercury and regenerated with H 2 S. The factor was destroyed at 100°C in 1 Ν H 2 S 0 4 . We also found that a crude muscle extract would restore light. Obviously this suggested that we should try ATP. Fortunately I had some partially purified ATP which John Gregg and I had isolated from frog eggs when we were studying the possible role of ATP in the gastrulation process (an evening project at Princeton). When I tried this ATP in a crude firefly extract, a brilliant light was observed (15). So in reality what DuBois had observed in 1887 was the effect of ATP on light emission, almost 40 years before the discovery of ATP. The luciferase could be precipitated from the crude extract with ammonium sulfate, but the light intensity was greatly decreased when the precipitate was dissolved in buffer and ATP added. Upon repeated addition of ATP it was clear that we could exhaust another factor that was necessary for light emission. Thus the search for a true luciferin was on, and, as it turned out, this factor that was used up was the true luciferin.

Lipmann's Influence on Firefly Luminescence

207

After many years of research and the use of millions of fireflies and the collaboration of graduate students, postdoctoral students and other colleagues at Hopkins we isolated and identified the luciferin, which ultimately led to its synthesis (3,23) (See Figure 1 for the structures of L, L = 0 and LH 2 ). Two key collaborators on this project were a post-doctoral student, Dr. Frank McCapra, and Dr. Emil White, a Professor in the Johns Hopkins Chemistry Department. Dr. Arda Green and I purified and crystallized the luciferase (13). Detailed studies on the mechanism (4-6,8-10,18,20) have led to the following formulation for the light reaction: LH2 + MgATP + E ?± E • LH2 - AMP + PP¡ 1 E • Lids - AMP + 0 2

E - L = 0 + C0 2 + H 2 0 AMP

E • L = 0 ->· light + E • L = 0 + AMP AMP E • L = 0 ç± E + L = 0

χαο OXYLUCIFERIN Η

D (-) LUCIFERIN 4'

3'

3

τ

r

1

DEHYDROLUCIFERIN Figure 1

1

The structure of luciferin (LH 2 ), dehydroluciferin (L) and oxyluciferin ( L = 0 ) .

· denotes tight binding - denotes covalent binding

208

W.D. McElroy

Dr. Seliger and I demonstrated that the quantum yield of the reaction was close to one (21). Thus Lipmann's article on phosphate bond energy greatly influenced my thinking on experiments that should be done with luminous organisms. We were lucky that we used fireflies first. Now I would like to discuss one more influence that Lipmann had on this system.

From Adenyl Luciferin to Luciferyl Adenylate In 1958 Rhodes and I published a brief note in Science concerning the first demonstration of the enzymatic synthesis of an acyl adenylate (19). The intense fluorescence of firefly dehydroluciferin made it possible to measure the disappearance of quantities which were less than the amount of enzyme added. We were able to measure the formation of L-AMP in the following reaction (L-AMP fluorescence is only about 5 percent of the fluorescence of free L): ATP + L + Ε τ± E · L - AMP + PP¡ The equilibrium constant for the association of "adenyl-oxyluciferin" 2 and the firefly luciferase (L-AMP + E E · L-AMP) was determined to be approximately 2 χ 109. We observed at the time that the tight binding of acyl adenylates to the activating enzymes probably was the reason it was difficult to obtain any net enzymatic synthesis of these reactive species. At that time, in personal correspondence with Dr. Meister, he indicated that he had been able to demonstrate the enzymatic synthesis of adenyl tryptophane. A note to this effect was added in proof (M. Karasek et al. 1958. J. Am. Chem. Soc. 80, 2335). During this time I had been discussing our results with Dr. Lipmann. Consequently when Dr. Rhodes and I had prepared a detailed manuscript to be submitted to the Journal of Biological Chemistry, we sent a copy to Dr. Lipmann for his review. He made very few suggestions, and at the end he indicated a "modest" proposal; i.e., change the name of adenyl-oxyluciferin to oxyluciferyl-adenylate. This we did, and I believe that this was the first time this nomenclature was used. Berg (1956) used adenyl acetate, and others used adenyl-amino acids, etc. For reasons that are not clear to me, we used this nomenclature in only two places - in the title "The synthesis and function of luciferyl-adenylate and oxyluciferyl-adenylate" and in the summary (20). Elsewhere in the text it was LH 2 -AMP and L-AMP. The fact that apparently the change did not impress us at the time is indicated in the acknowledgements, where we thanked Dr. Sidney Colowick but made no mention of Dr. Lipmann!! But to our knowledge he was the first to suggest this nomenclature for acyl adenylates.

2

In earlier publications no distinction was made between dehydroluciferin and oxyluciferin. Only after the isolation and identification of the latter were we certain that the enzymatic product of the light reaction contained oxygen. Dehydroluciferin is produced from luciferin by alkaline removal of two hydrogens or by a photochemical decomposition that also removes two hydrogens.

Lipmann's Influence on Firefly Luminescence

209

The Function of Coenzyme A in Firefly Luminescence The discovery of Coenzyme A by Lipmann and the events leading a number of workers to demonstrate its function in transfer reactions was an exciting period in biochemistry. It occurred to us to predict that if the reaction leading to the activation of luciferin (LH 2 ) and dehydroluciferin (L) by ATP was due to the formation of an acyl adenylate, the intermediate might react with CoA. We had shown previously that in the presence of excess L and inorganic pyrophosphatase, luciferase could be completely inhibited. One could obtain light from this reaction mixture by adding luciferin and pyrophosphate secondarily. This produced a flash of light which disappeared gradually with the hydrolysis of the pyrophosphate. This suggested that the following reaction was occurring: L + ATP + E ç± E • L - AMP + PP¡ This reaction leads to the depletion of free enzyme available for the concurrent reaction between LH 2 and ATP. PP¡ addition frees luciferase temporarily. I should point out that this inhibition of luciferase by dehydroluciferin (L) is quite different from the inhibition due to product formation in the light reaction. The product, oxyluciferin ( L = 0 ) , is bound tightly to the enzyme and slowly dissociates, allowing only a low level of luminescence after the initial flash (12). If CoA reacted with E · L-AMP in the presence of LH 2 and ATP, we should see a stimulation of light by CoA according to the following reactions: CoA + E • L - AMP Η,Ν - CHR - C?) 3 S J

The new class of amino acid derivatives consists of well crystallized individuals of zwitteronic character the negative charge most likely prevailing on the sulfur atom. As far as investigated they are more lipophilic than the respective parent compounds (larger ÄF-values), and - admixed with their S-free analogs - form homogenous crystals composed of both species up to 1 : 1 ratios (16). The wider crystal lattice, apparently, allows incorporation of the S-free analogs without disturbation of the lattice. We also investigated a possible acylating ability of the α-amino thioic acids, i. e. their usefulness for peptide synthesis. A preliminary test, reaction with NH 3 as an acyl acceptor, however, showed that no reaction occured with this nucleophile unless hydrogen carbonate ions were present. An enhancement of the reactivity also of aminoacyl thiols by HCO3 (see later) points to a reactive intermediate, likely an inner TV-carbonic carboxylic acid anhydride (Leuchs' anhydride). Accordingly, in the reverse reaction, a-aminothioic acids can be prepared by reaction of Leuchs' anhydrides with hydrogen sulfide (17). Η I RC I ΗΝ

Η I R-C+ NH,

C= 0 I 0 + SH"

'HCO,

Ö The reaction of acid chlorides with hydrogen sulfide successfully also led to the synthesis of a tripeptide with thioic acid terminal, glycylalanyl-thioisoleucine, and of β- and ω-aminothioic acids (18).

Th. Wieland

216

Biomimetic Peptide Syntheses α-Aminoacyl thiophenols in aqueous medium react with amino groups of α-amino acids or of a-aminoacyl thiophenols to yield di-, tri- and higher peptides. Thus, glycyl thiophenol in weakly alkaline water on standing at room temperature formed di-, tri- and higher polyglycines (19); in the presence of a second amino acid (valine) among other products diglycylvaline was detected (20). O O 11^ - iLH0NCH„C Η,Ν - CH.- C* Η„Ν - CH-CH(CH,)_—H-GlyGlyVal-OH L L I L L ι L I i L SC H C0 SC 6 H 5 6 5 2 In this paper in 1951, when the mechanism of ribosomal protein synthesis was yet far from being understood, we wrote: "Es erhebt sich nun die Frage, ob nicht schwefelgebundene Aminoacylreste bei der Biosynthese von Peptiden als energiereiche Aminoacyl-Reste fungieren können..." This question has been positively answered nearly 20 years later mainly in the laboratory of Fritz Lipmann (21). The activation of amino acids for the syntheses of gramicidin S, tyrocidine, bacitracin and many other microbial peptides occurs by their ATP-dependent linkage to SH groups of cysteine, and pantetheine (cysteamine) side chains in the respective synthesizing enzymes. One amino acid, energy-rich bound to the sulfur of a cysteine side chain, will aminolyze the acyl residue of a second amino acid, energyrich bound to the sulfur atom of the cysteamine part of the "transport arm" pantetheine. As a consequence a peptide is formed, energy-rich bound to the pantetheine sulfur atom. The next step consists of the aminolysis of the peptide activated on the transport arm by the following amino acid so generating the elongated peptide. This, now, is tranlocated back in an S—S transacylation reaction (thiolysis) to the sulfur atom of its S—CO-moiety by the amino group of the next cysteine-S-bound amino acid.

©

HS CH*

CH

CH,

CH¿

CH Z

NH

I"--"

s

©

I

s

CH;

CHj Cys

NH

Cys

CO

peptide

trans-

peptide

'

ffoorrm maa tti !o n

acylation

formation

O Hz I C Hi NH pantetheine

217

Sulfur in Biomimetic Peptide Syntheses

Both biological reactions, transamidation (peptide forming step) as well as transacylation have been mimicked in vitro with our S-aminoacyl thiols. Transacylations from 5-acylthiophenol to other thiols like GSH and CoA have been described previously (page 214), additional bioorganic thiol compounds just mentioned, cysteine and cysteamine, have been S-acylated in the same way.

Intramolecular Aminoacyl Migration The reaction of S-glycyl- (or other aminoacyl-)thiophenols with cysteine revealed an interesting consecutive step, an immediate migration of the initially transferred aminoacyl from cysteine sulfur to the ideally located amino group of the same molecule (22). C,H,S-COCHRNH, + HS-CH,

6 ?

L

ι L H-N-C-H 2 j CO Η

H0NCHRCO-S-CH0

L

ι L H0N-C-H 2 ι

^

CO Η

HS-CH,

ι L ι

H-NCHRCONH-CH

2

^

CO Η

A comparable easy peptide synthesis could not have been observed with other acceptor amino acids except histidine. Here, obviously, one imidazol nitrogen atom has a "receptor" function analog to the SH of cysteine; the /V-amino-acyl compound primarily formed at the imidazol ring, as an "energy rich" acyl derivative (23), transfers the aminoacyl group in a rapid intramolecular reaction to the amino group of the molecule -CH

//

aminoacylthiophenol

N—C—CHj. Η I H«N-C-CO,H Η

+ / HjN-CHIR C it 0

•CH -CH —CHi I HjN-C-COaH

H


5 to yield in the case of unsymmetrical diacyl amines a mixture of isomeric dipeptide amides. 1

COCHR N , HF{

3

COCHRZN3

D

Ä

.

1+

-COCHR N H , H < . + 3 COCHRZNH3

H2NC0CHR1NHC0CHR2NH2

+ H2NCOCHR2NHCOCHR1NH2

Pursuing this type of rearrangement with dipeptide cysteamides by loading their SH groups with a new a-amino-acyl residue and subjecting the compounds to the HCO3-catalyzed aminoacyl migration, indications were observed of a participation of the peptide nitrogen as an intermediate carrier of the energy-rich diaminoacyl system. If in a peptide chain an amide-nitrogen would be amino-acylated (e. g. by acyltransfer), by acyl migration of the peptide part an additional amino acid, H 2 N—CHR 2 —C0 2 H, would be incorporated into an intact peptide chain. A possible occurrence of this and of related reactions in nature can be discussed. -CO-HN-CH-CO-NH-CHR1-CO-NHI H-C-S 2 Ι CO I 2 H2N-CHR^

> ^

-CO-HN-CH-CO-N—CHR1-CO-NHI I HS-CH, CO 2 ( „ H„N-CHR 2

-CO-NH-CH-CO-NH-CHR2-CO-NH-CHR1-CO-NHI

HS-CH 2

Conclusion Speculations in the early 1950ies, then justified, as to a possible role of the thiol group, e.g. of glutathione, in the biological synthesis of peptides and proteins prompted us to synthesize S-aminoacyl derivatives, at first of thiophenol. Thiophenylesters of iV-protected amino acids, as high-energy compounds, have been found to be very potent reagents for in vitro peptide syntheses, i.e. transamidation. In addition, these substances were found to be suitable reagents for transfer of the acyl group to SH groups of other thiols. By transacylation several S-acyl - not only aminoacyl - compounds of biological interest like glutathione, coenzyme A, cysteine and cysteamine, respectively, were obtained. Since in natural peptide synthesis the amino acid building blocks are not protected at their amino groups, emphasis was laid on studies with free amino acyl residues

220

Th. Wieland

activated by binding to thiol sulfur. Aminoacyl derivatives of thiophenol. glutathione, and cysteamine were investigated in detail. In aqueous solution in the presence of second amino components they give rise to the formation of peptides - as in the non-ribosomal biological synthesis of microbial peptides - , rearrangements can occur by intramolecular migration of the acyl group from sulfur to nitrogen. Cysteine on reaction with an aminoacyl mercaptan instantaneously forms a dipeptide as a consequence of the activation of the amino acyl as an S'-compound formed in the first step. An analogous reaction sequence was observed with histidine, as indication of its imidazol moiety as carrier of an ^-activated aminoacyl rest. iV-acylimidazoles as energy-rich compounds! A further type of acyl activation was found in diacylamines. Valylglycyl-amine undergoes intramolecular rearrangement yielding two isomeric dipeptide amides. In contrast to the new activated amino acids discovered in the studies of amino-acid thioesters the prototypes, aminothioic acids, H 3 N—CHR—COS", normally do not exert any peptide-forming activities. The present review on an old field of research in the authors laboratory has become topical as amino acid activation by thioesterification has been revealed in nature mainly in F. Lipmann's laboratory. Whether the one or other reaction of amino acids or peptides observed in connection with laboratory work would occur naturally is still a question.

References 1. F. Lynen and E. Reichert. 1951. Angew. Chem. 63, 47. 2. F. Lipmann. 1945. J. Biol. Chem. 160, 173. 3. F. Lynen, E. Reichert and L. Rueff. 1951. Liebigs Ann. Chem. 574, 1. 4. Ch. J. Stewart and Th. Wieland. 1955. Nature 176, 316. 5. Th. Wieland, W. Schäfer und E. Bokelmann. 1951. Liebigs Ann. Chem. 573, 99. 6. Th. Wieland and H. Köppe. 1953. Liebigs Ann. Chem. 581, 1. 7. Th. Wieland, G. Pfleiderer and H.H. Lau. 1956. Biochem. Z. 327, 393. 8. Th. Wieland and L. Rueff. 1953. Angew. Chem. 65, 186. 9. Th. Wieland and E. Bokelmann. 1952. Angew. Chem. 64, 59. 10. Th. Wieland and F. Jaenicke. 1955. Chem. Ber. 88, 1967. 11. Th. Wieland, J. Franz and G. Pfleiderer. 1955. Chem. Ber. 88, 641. 12. E. Fischer. 1905. Ber. dtsch. chem. Ges. 38, 2914. 13. Th. Wieland and W. Schäfer. 1952. Liebigs Ann. Chem. 576, 104. 14. Th. Wieland and D. Sieber. 1953. Naturwissenschaften 40, 242. 15. Th. Wieland, D. Sieber and W. Bartmann. 1954. Chem. Ber. 87, 1093. 16. Th. Wieland and W. Bartmann. 1956. Chem. Ber. 89, 946. 17. Th. Wieland and K.E. Euler. 1958. Chem. Ber. 91, 2305. 18. Th. Wieland and K. Freter. 1954. Chem. Ber. 87, 1099. 19. Th. Wieland and W. Schäfer. 1951. Angew. Chem. 63, 146. 20. Th. Wieland and W. Schäfer. 1952. Liebigs Ann. Chem. 576, 101. 21. F. Lipmann. 1982. In: "Peptide Antibiotics, Biosynthesis and Function (H. Kleinkauf and Η. v. Döhren, eds.), de Gruyter, Berlin, New York, pp. 23 ff. 22. Th. Wieland, E. Bokelmann, L. Bauer, H.U. Lang and H. Lau. 1953. Liebigs Ann. Chem. 583, 129. 23. Th. Wieland and G. Schneider. 1953. Liebigs Ann. Chem. 580, 159. 24. H.A. Staab. 1956. Chem. Ber. 89, 1927. 25. H.A. Staab. 1957. Liebigs Ann. Chem. 609, 75. 26. H.A. Staab and K. Wendel. 1961. Angew. Chem. 73, 26.

Sulfur in Biomimetic Peptide Syntheses 27. 28. 29. 30. 31. 32. 33.

221

G.W. Andersen and R. Paul. 1958. J. Amer. Chem. Soc. 80, 4423. Th. Wieland and K. Vogeler. 1961. Angew. Chem. 73, 435. E.R. Stadtman and F.H. White jr. 1953. J. amer. chem. Soc. 75, 2022. M. Bergmann, V. du Vigneaud and L. Zervas. 1929. Ber. dtsch. chem. Ges. 62, 1909. Th. Wieland, R. Lambert, H.U. Lang and G. Schramm. 1956. Liebigs Ann. Chem. 597, 181. Th. Wieland and H. Möhr. 1956. Liebigs Ann. Chem. 599, 222. Th. Wieland and H. Urbach. 1958. Liebigs Ann. Chem. 613, 84.

The Function of Teichoic Acids in Walls and Membranes of Bacteria James Baddiley

Introduction Much of the research in my laboratory has been concerned with the structure, biosynthesis and function of polymers in bacterial cell walls and membranes. At first sight this might not appear to have a close connection with the several topics that Fritz Lipmann made so much his own. In fact our association on the structure of coenzyme A between 1950-1954 aroused my interest in the mechanism of biosynthesis of that molecule and it was during this work that my group discovered two nucleotides which were precursors of the hitherto unknown polymers, teichoic acids, in the cell wall. Moreover, at the end of our collaborative work I had the good fortune to spend a while as a guest in Fritz's laboratory where a lasting friendship developed and I found the frequent discussions with him a great stimulus in helping to formulate my ideas. Perhaps the most obvious function of the cell wall in bacteria is to protect the cell, including the surrounding cytoplasmic membrane, from the hazards of the external environment. In order to do this effectively and at the same time avoid undesirable restriction of growth, cell division and general metabolism it requires special properties. It must be strong in order to withstand considerable osmotic pressures, it must be able to prevent the passage of macromolecules and it must allow the passage of smaller metabolites and ions. Moreover, it must assist in creating the correct conditions for the optimum operation of the many enzymes and transport proteins in the membrane. This article is confined to a discussion of how some of these and other properties relate to molecular structure in the walls of Grampositive bacteria, with particular emphasis on the teichoic acids. In the Gram-positive bacteria the wall is a single undifferentiated layer representing about 20% of the dry weight of the cell. Variable amounts of protein are present but the main components are peptidoglycan, teichoic acid and occasionally an acidic polysaccharide, teichuronic acid. Peptidoglycan is a highly cross-linked polymer comprising glycan chains joined to each other by short peptide chains. The extent of cross-linking and the amino acid composition vary in different organisms but both free amino and carboxyl groups are present in all cases. The teichoic acids are polymers of 30 or more glycerol phosphate or ribitol phosphate units to which are attached glycosyl residues and D-alanine ester residues. There is a fairly wide range of structural variation in the teichoic acids in which hexoses and acetylaminohexoses and their phosphates occur either as appendages to the main chain or as a part of the chain; the number of alanine ester residues also varies in different organisms. The teichoic acids are attached to peptidoglycan in the wall by short linkage units containing glycerol phosphate and sugar residues. Teichuronic acids are polysac-

224

James Baddiley

charides of varying complexity containing uronic acid residues. Both teichoic and teichuronic acids are attached to the peptidoglycan through a terminal phosphodiester (1-3). As these wall polymers are attached covalently to the peptidoglycan, which is itself highly cross-linked, it is possible to regard the whole wall as a single macromolecular complex. However, since the relative amounts can be varied according to culture conditions in some organisms and as their biosynthetic routes are distinct from each other and can be controlled independently, it is convenient to regard the wall as comprising two or three distinct molecular species.

Importance of Cell Wall Components It is generally agreed that the widespread occurrence of peptidoglycan in the walls of bacteria indicates that it plays a vital role and its stable, highly cross-linked structure suggests that physical strength is its major function. The loss of physical strength of the wall brought about by interference with peptidoglycan synthesis through ßlactam antibiotics illustrates the important role played by this component. Similarly, the teichoic acids are very widely distributed amongst Gram-positive bacteria and in those organisms where they are absent it is usually found that a teichuronic acid is present. These acidic polymers are in fact major wall components and account for 30-50% of the dry weight of most walls, in a few cases representing as much as 60% of the wall. Such large amounts of material represent a considerable investment of biosynthetic effort and it is therefore likely that the presence of teichoic acid is advantageous to the organism. In addition to their abundance and widespread occurrence several other observations support the importance of teichoic acids. Amongst a large number of mutants of Bacillus subtilis selected for phage resistance, defects affecting teichoic acid were extremely limited (4) and none were described in which they were absent. In other mutant bacilli where the teichoic acid content is reduced there are gross deformations in cell shape (5,6). Moreover, genetic determinants for wall teichoic acid in B. subtilis strains can be transferred to other strains without impairment of cell function but in such studies no recombinants have been observed which are devoid of teichoic acid (7). These observations support the suggestion (2) that, although acidic polymers in the wall are essential, their main function is one in which structural detail is less important than general overall properties such as those arising through their possession of a large number of charged centres. Pneumococcus spp. illustrate their essential nature; in these organisms the wall and membrane teichoic acids (see later) contain choline (8,9) and the organisms are unable to synthesise this, which must be supplied as an essential component of the growth medium. As choline is not a component of any other cellular substance it must be concluded that the teichoic acids are indispensable. Two outstanding characteristics of teichoic acids were recognised at the time of their discovery, namely their highly charged nature and their pronounced serological properties. Both the wall and membrane polymers have been shown to be major antigens with either group or type specific reactions towards antisera (10,11).

The Function of Teichoic Acids in Walls and Membranes of Bacteria

225

However, this property can hardly confer advantage since serological activity of that kind is generally a stimulus to the immunological defence mechanisms of the host in its efforts to destroy invading bacteria. Thus it seems most unlikely that serological behaviour could be considered a function. On the other hand the possession of many anionic and cationic centres, which under normal conditions would be ionised, suggests that their charged state is related to function (12). The phosphodiester groups have a strong affinity for cations and, as expected, their affinity for divalent cations is considerably greater than for monovalent cations (13,14). Moreover, the presence of amino groups on the alanine ester residues results in a decrease in the cation affinity (15), suggesting that this might act as a control of cation binding. These properties of teichoic acids in solution are also observed in wall preparations (16). Although peptidoglycan contains carboxyl and amino groups, these appear to play a much less effective role in cation binding than do the phosphates in teichoic acids (13). The importance of these polyanionic polymers is strongly supported by growth studies under continuous culture conditions using Bacillus spp. When grown under phosphate limitation the cells no longer have teichoic acid in their walls but now possess teichuronic acid. If the phosphate concentration in the medium is increased the teichuronic acid is replaced by teichoic acid (17). Thus the phosphate concentration in the medium can profoundly alter the composition of the cell wall but nevertheless the presence of an acidic polymer is maintained throughout. It would also seem that, provided the level of inorganic phosphate in the medium is adequate, there is a preference in many organisms for teichoic rather than teichuronic acid in the wall. The teichoic acids associated with the membrane are polymers of glycerol phosphate attached to phosholipid and they too possess alanine ester residues. They are referred to as membrane teichoic acids and are probably even more widespread than the wall polymers (18). More recently they have been called lipoteichoic acids. They are not interchangeable with other acidic polymers under differing growth conditions. Thus it is possible that the structural requirements in close proximity to the membrane surface are more rigid than they are further out in the wall.

Cation Control From these observations it can be concluded that there would be a relatively high concentration of divalent cations in the wall and in the vicinity of the membrane. Such cations are essential for the integrity of the membrane itself, as it is known that association between divalent cations and the charged centres of phosholipids on the membrane surfaces plays an important part in the lipid bilayer structure. Moreover, the membrane is the location of a number of enzymes, including those involved in translocation of metabolities, proton and electron transfer and in the synthesis of phospholipids, glycolipids, peptidoglycan, polysaccharides, ATP apd teichoic and teichuronic acids themselves. In not all cases are the ion requirements known but in a number it is found that optimum activity requires a high concentration of divalent cations.

226

James Baddiley

The importance of teichoic acids in cation control in the vicinity of the membrane to ensure optimum activity of bound enzyme systems has been demonstrated experimentally (19). Washed membrane preparations from Bacillus licheniformis are able to synthesise wall teichoic acid from the nucleotide precursors CDP-glycerol and UDP-glucose. The membrane-bound enzyme system has a high requirement for magnesium ions with a maximum activity at about 15 mM Mg 2 + , the rate falling off sharply above or below this value. If bacteria are broken in a French pressure cell a preparation is obtained in which the wall and membrane are still in contact with each other and, unlike the washed membrane preparation, the wall and lipoteichoic acids remain in situ', in this preparation the teichoic acids are in salt association with Mg 2 + . It was found that these wall-membrane preparations catalyse the synthesis of wall teichoic acid at a maximum rate which is independent of the external concentration of Mg 2 + in the surrounding medium. It appears that the enzymes prefer to interact with the cations bound to wall and membrane teichoic acids which are able to maintain the Mg 2 + concentration at its optimum level, thus producing a powerful cation buffering in the vicinity of the enzyme complex. If the washed membrane preparation is pretreated with wall or membrane teichoic acid a cation buffering effect is observed, although not as powerful as with the wall-membrane preparation. Thus the teichoic acids scavenge divalent cations from the surrounding medium and make them available at optimum concentration to the membranebound enzymes. The amout of alanine ester in wall teichoic acid varies with the growth conditions and decreases with increasing amounts of sodium chloride in the medium (16). This is possibly connected with the cation binding properties since it can be argued that, since sodium ions would compete with magnesium ions, in order to maintain the optimum concentration of Mg 2 + for correct functioning of membrane bound enzymes the cation affinity of the teichoic acid must be increased by decreasing the alanine content. It is significant that the alanine of the lipoteichoic acid follows an identical course (20) and therefore is presumably under the same control. The volume occupied by the cell wall can be altered by varying the concentration of cations in the surrounding medium (36,37). This effect would presumably depend upon the number of anionic binding sites in the wall and thus be largely controlled by the teichoic or teichuronic acid. The significance of wall volume on bacterial metabolism and cell function and its possible control by teichoic acids has received relatively little attention but merits further investigation.

Importance in Biosynthesis In the biosynthesis of wall teichoic acid by membrane prepartions it can be shown that a carrier (LTC) in the membrane is able to accept units from nucleotide precursors to form the main polymer chain and perhaps to transfer this to linkage unit attached to undecaprenyl phosphate in the membrane (3,21). LTC is in fact the lipoteichoic acid (22) but its activity as an acceptor depends heavily on the amount of alanine in the preparation. Maximum activity is observed when no alanine ester residues are present and the activity decreases rapidly as the alanine ester content is

The Function of Teichoic Acids in Walls and Membranes of Bacteria

227

increased (23). As alanine-free lipoteichoic acid has not been detected in membranes it is not clear whether this mechanism of synthesis occurs in whole organisms. However, this function for lipoteichoic acids in whole cells can not be completely excluded as it is not known whether catalytic amounts of the alanine-free compound might occur in the immediate vicinity of the enzyme complex responsible for synthesis. Although the status of lipoteichoic acid in the synthesis of the main chain of wall teichoic acid is uncertain it does nevertheless appear to participate in the introduction of alanine ester residues into that polymer (24,25). The turnover of alanine ester in the wall is independent of the much slower turnover of teichoic acid itself. This is believed to arise through loss of alanine by presumably spontaneous hydrolysis and its replacement by alanine from the lipoteichoic acid, which in turn receives its alanine from intracellular alanine in reaction with ATP. The mechanism of transfer between polymer chains is not yet established and it is not known whether an enzyme is involved.

Effect on Autolysins The autolysins which participate in the processes of growth and cell separation can be influenced by teichoic acids. In pneunococci, which are highly autolytic, the wall teichoic acid contains choline residues. If choline is substituted by ethanolamine in the medium the resulting cells possess ethanolamine rather than choline in both wall and membrane teichoic acids (26). Such organisms are no longer autolytic and they grow as long filaments of unseparated cells. Thus a small structural change has a profound inhibitory effect on the autolysin. The choline residues are essential for attachment of the autolysin to the cell wall in its correct orientation for autolysis to occur (27). With other organisms lipoteichoic acid preparations are able to inhibit autolysins (28) but in these cases the alanine residues are absent and more recent experiments using preparations containing controlled amounts of alanine show that inhibitory activity decreases exponentially with increasing alanine content (29). Thus it would seem that much of the lipoteichoic acid in the membrane has little influence on autolysin activity. It is possible, nevertheless, that lysis might be influenced in localised areas by lipoteichoic acid.

Phage Receptor Sites Many bacteriophages show specificity towards teichoic acid in binding to the cell wall. The phage receptor sites are organised regions of the wall involving both teichoic acid and peptidoglycan (30,31) and the teichoic acid component is involved in the binding as well as in the specificity. Phage infection is obviously a disadvantage to bacteria as it results in their death. However, it can be a means of interchange of genetic material and consequently might have an important role in bacterial evolution. Nevertheless, although phage-cell interaction is a biologically important phenomenon, it is doubtful whether this represents a primary function of teichoic acid or of peptidoglycan.

228

James Baddiley

Other Possible Functions An interesting and probably important role for lipoteichoic acid has been suggested in the process of attachment of bacteria to the surface of the host cell during early stages of infection. It has been known for a considerable time that during cell growth lipoteichoic acid becomes detached from the membrane where its lipophilic acyl residues are normally intercalated with the lipid bilayer structure (11,18,32). The detached molecules pass through the wall into the surrounding medium where they can be detected either unchanged or in a deacylated form (28). Cell surface proteins are able to associate with the glycerol phosphate chains of lipoteichoic acids and it is proposed that complexes of this kind, involving several molecules of lipoteichoic acid, on the outer surface of the bacterial cell would possess exposed glycolipid which could interact with the lipid bilayer structure of the host cells (33). In this model the glycerol phosphate chains of some of the lipoteichoic acid molecules would associate with the bacterial wall teichoic acid. Further work is essential to establish this theory and it is not known whether cations play a part in the interactions. It would be possible to envisage binding between wall and membrane teichoic acids through divalent cations. Inter- and intrachain salt linkages involving phosphodiester residues and divalent cations can be demonstrated in bacterial cell walls (34) and their participation in the lipoteichoic acid-wall complex would seem likely. It has been suggested that teichoic acid might act as a phosphate reserve for the cell (35). However, it is doubtful that this could be a major function as the ability to achieve a substantial decrease in the teichoic acid content of the wall when organisms are grown under conditions of phosphate limitation is not a general property of bacteria which possess wall teichoic acids but is confined to a few varieties. Furthermore, although phosphatases that hydrolyse teichoic acids are known, they are uncommon and there is no convincing evidence that they are active during phosphate limitation of growth.

Conclusions In conclusion it seems likely that teichoic acids have several functions. The most obvious is to provide a polyionic component in the wall and especially in the vicinity of the membrane, in order to bind mainly divalent cations. Such cations are required for the integrity of the cell membrane, the optimum activity of membrane enzymes and perhaps also for the control of autolysins and cell-host interactions. In addition the lipoteichoic acids have a role in the transportation of alanine to the wall teichoic acid and possibly in other aspects of wall teichoic acid assembly. Other roles for lipoteichoic acids in control of autolysins and in the early stages of infection are supported experimentally but require further work for their substantiation.

The Function of Teichoic Acids in Walls and Membranes of Bacteria

229

References 1. Archibald, A. R. 1974. Adv. Microbial Physiol. 10, 53-95. 2. Baddiley, J. 1972. Essays Biochem. 8, 35-77. 3. Hancock, I.C. and Baddiley, J. 1985. In: The Enzymes of Biological Membranes (Martonosi, A.N., (ed.) Plenum Press, 2nd ed., pp. 279-307. 4. Young, F.E. 1967. Proc. Natn. Acad. Sci. USA 58, 2377-2384. 5. Boylan, R.J., Mendelson, N.H., Brooks, D. and Young, F.E. 1972. J. Bacterid. 110, 281-290. 6. Forsberg, C.W., Wyrick, P.B., Ward, J.B. and Rogers, H.J. 1973. J. Bacteriol. 113, 969-984. 7. Karamata, D., Pooley, H.M. and Monod, M. 1987. Mol. Gen. Genet. 207, 73-81. 8. Brundish, D.E. and Baddiley, J. 1968. Biochem. J. 110, 573-582. 9. Tomasz, A. 1967. Science 157, 694-697. 10. Baddiley, J. and Davison, A. L. 1961. J. Gen. Microbiol. 24, 295-299. 11. McCarty, M. 1959. J. Exp. Med. 109 361-378. 12. Archibald, A.R., Armstrong, J.J., Baddiley, J. and Hay, J.B. 1961. Nature 191, 570-572. 13. Heptinstall, S., Archibald, A.R. and Baddiley, J. 1970. Nature 225, 519-521. 14. Lambert, P.A., Hancock, I.C. and Baddiley, J. 1975. Biochem. J. 149, 519-524. 15. Lambert, P.A., Hancock, I.C. and Baddiley, J. 1975. Biochem. J. 151, 671-676. 16. Archibald, A.R., Baddiley, J. and Heptinstall, S. 1973. Biochim. Biophys. Acta 291, 629-634. 17. Ellwood, D.C. and Tempest, D.W. 1968. Biochem. J. 108, 40P. 18. Lambert, P.A., Hancock, I.C. and Baddiley, J. 1977. Biochim. Biophys. Acta, 472, 1-12. 19. Hughes, A.H., Hancock, I.C. and Baddiley, J. 1973. Biochem. J. 132, 83-93. 20. Fischer, W. and Rosei, P. 1980. FEBS Lett. 119, 224-226. 21. Fiedler, F. and Glaser, L. 1974. J. Biol. Chem. 249, 2684-2689. 22. Duckworth, M„ Archibald, A.R. and Baddiley, J. 1975. FEBS Lett. 53, 176-179. 23. Koch, H.U., Fischer, W. and Fiedler, F. 1982. J. Biol. Chem. 257, 9473-9479. 24. Fischer, W. and Koch, H.U. 1985. Biochem. Soc. Trans., 984-986. 25. Neuhaus, F.C. 1985. Biochem. Soc. Trans., 987-990. 26. Tomasz, Α., Westphal, M., Briles, E.B. and Fletcher, P. 1975. J. Supramol. Struct. 3, 1-16. 27. Garcia-Bustos, J.F. and Tomasz, A. 1987. J. Bacteriol. 169, 447-453. 28. Cleveland, R.F., Holtje, J.-V., Wicken, A. J., Tomasz, Α., Daneo-Moore, L. and Shockman, G.D. 1975. Biochem. Biophys. Res. Commun. 67, 1128-1135. 29. Fischer, W., Rosei, P. and Koch, H.U. 1981. J. Bacteriol. 146, 467-475. 30. Archibald, A.R. 1980. In: Virus Receptors and Recognition. Series B, 7. Chapman and Hall, London, pp. 7-26. 31. Glaser, L., Ionesco, H. and Schaeffer, P. 1966. Biochim. Biophys. Acta 124, 415-417. 32. Stewart, F.S. 1961. Nature 190, 464. 33. Ofek, I., Simpson, W.A. and Beachey, E.H. 1982. J. Bacteriol. 149, 426-433. 34. Baddiley, J., Hancock, I.C. and Sherwood, P.M.A. 1973. Nature 243, 43-45. 35. Grant, W.D. 1979. J. Bacteriol. 137, 35-43. 36. Marquis, R.E. 1968. J. Bacteriol. 95, 775-781. 37. Ou, L.T. and Marquis, R.E. 1970. J. Bacteriol. 101, 92-101.

The Amidotransferases: Origins of the Concept of Affinity Labeling of Enzymes John M. Buchanan

Fritz Lipmann's concept of adenosine triphosphate as a highly reactive source of energy in enzymatic reactions contributed enormously to the formulation of the several reactions involved in the pathways for the synthesis of the purine and pyrimidine ribonucleotides. By the 1950s his high energy "squiggle" phosphate bonds of ATP had explained the energetics of activation of substrates by transfer of phosphate with the formation of adenosine diphosphate. Primarily through the efforts of Arthur Kornberg the activation of compounds by the transfer of pyrophosphate groups was established. Subsequently, the transfer of the adenylate group in fatty acid and amino acid activation was discovered by Paul Berg and Mahlon Hoagland, respectively. All three of these mechanisms are represented in the enzymatic reactions of ribonucleotide biosynthesis. In the course of our enzymatic studies on purine nucleotide biosynthesis we focused our attention principally on two reactions, both of which contribute an amide nitrogen of glutamine to the purine ring (1). The importance of these reactions developed from our finding that two antibiotics, L-azaserine and 6-diazo-5-oxo-Lnorleucine, are analogs of glutamine, compete with the normal substrate at the active site and eventually cause an irreversible inactivation of the enzyme (Figure 1) (2-7). These findings set the stage for the discovery of affinity labeling, the subject of this review.

NHfCO-CH^CHjCH NHfCOOH L- GLUTAMINE N=N=CH-C0-0—CHjCHNHJ-COOH L- AZASERINE N=N=CH-C0-CH2 CHj CHNHj-COOH 6-DIAZ0-5-0X0 L-NORLEUCINE NH-CO-NH-CH-CHNH-COOH 2 2 2 L- ALBIZZI IN Figure 1

Antimetabolites of glutamine.

The energy required for the formation of the carbon-to-nitrogen bonds in both reactions mentioned above is derived from the cleavage of phosphodiester bonds of ATP. In the first of the two systems the activation of one substrate, ribose 5phosphate, to yield 5-phosphoribosyl 1-pyrophosphate (PRPP) (reaction 1) and the

232

John M. Buchanan

coupling of the latter with glutamine to yield phosphoribosylamine (PRA) (reaction 2) occur in two separate discrete steps involving individual enzymes. ATP + ribose 5-phosphate —• AMP + 5-phosphoribosyl 1-pyrophosphate (PRPP) (1) PRPP + glutamine + H 2 0 -» 5-phosphoribosylamine (PRA) + PP¡ + glutamate (2) ATP + 5'-phosphoribosylformylglycinamide (FGAR) + glutamine + H 2 0 —• ADP + Pj + 5'-phosphoribosylformylglycinamidine (FGAM) + glutamate (3) On the other hand, in a second glutamine-requiring system catalyzed by 5'phosphoribosylformylglycinamide (FGAR) amidotransferase (reaction 3) the activation and coupling steps occur on one enzyme, which is composed of two domains, one binding glutamine, and the second the remaining substrates, MgATP and FGAR (8,9). Any hypothetical intermediate such as a phosphorylated FGAR remains bound to the enzyme, the products not being released until the complete reaction has taken place. Ammonia may be utilized at a lower rate as a nitrogen source in the normal reaction and even when the glutamine site is blocked. Hence, if there is a separate site for NH 3 , it must be located within or near the domain of the enzyme that binds MgATP and FGAR.

Formation of enzyme-substrate complexes From initial rate studies of FGAR amidotransferase of chicken liver with competitive and product inhibitors H.-C. Li has been able to determine that the substrates add sequentially to form a quaternary complex with the enzyme before undergoing reaction with release of products, and that glutamine adds first to the enzyme with ATP and FGAR randomly thereafter (partially compulsory order mechanism) (10). As mentioned above, Mizobuchi (2) has reported the isolation of a glutamineenzyme complex from an incubation of FGAR amidotransferase with glutamine followed by filtration on Sephadex gel. The reaction leading to the formation of the complex was reversible (i 1/2 = 125 min at 2°C) and did not depend on the presence of the rest of the substrates. The isolation of the FGAR-MgATP enzyme complex by the same technique has also been reported (3). At 2°C the half life of his complex is approximately 62 min. The formation of this complex is not dependent on the presence of glutamine. All three components, FGAR, ATP and Mg 2 + , must be present for the formation of a complex that is isolatable by gel filtration. Thus, glutamine and the pair of substrates (MgATP, FGAR) can react randomly with the enzyme, whereas FGAR and ATP are mutually dependent on each other for the formation of a stable complex with the enzyme. However, the enzyme catalyzes an ATP. ADP exchange reaction, which is dependent on the presence of Mg 2 + , but is not affected by the separate addition of either FGAR or glutamine. These observations imply that glutamine and MgATP, at least, can react with the enzyme independently of the presence of the other substrates. Thus, even though there is as yet no experimental evidence for the participation of a phosphorylated enzyme intermediate in a catalytically active pathway (Figure 2), its role in the reaction

The Amidotransferases: Origins of the Concept of Affinity Labeling of Enzymes

233

h - s - c=o Mg

H

h2c h2n. ι

H?C C' r ν, H H qf N— Ribose- PO4

'C H s - t XlN— Ribose-PO4 /To

rN^ii

— O -

P-0 = i i

— Y:-*P-0—ADP :0

(H>

(I) H0S,0

ϊ

- S H

R

H

,0

/ H0HN\

\Γ

R I /"V s - c=0

R'

HO H ( o r

η

Η N— R ¡ bose- PO4

u γ. ''

(ΠΣ) Figure 2

2

c -

n

Ì.H

π V ' ° H

HoN ADP

NHoOH)

H

h

α '' HN 0

ADP

Ribose-P04

O -IP - O 0·

ADP

(IE)

Proposed mechanism for the reaction catalyzed by F G A R amidotransferase.

mechanism must be given serious consideration. The FGAR · MgATP-enzyme complex may in fact be better represented as an enzyme-P · ADP · FGAR complex since it has been shown that all elements of the reaction remain bound to the enzyme after complex formation. Thus, three states of the complex are theoretically possible. E-MgATP-FGAR

Ε - P · MgADP-FGAR

E· MgADP- FGARP

(4)

Attempts were made unsuccessfully to isolate an hypothetical phosphoryl FGAR intermediate (FGARP) by elution from a Dowex-l-Cl column at pH 2. It is appreciated that, if such a compound did exist, it might well have been unstable under these conditions of its isolation. Although to date the favored mechanism used to explain the reaction sequence shown in Figure 2 involves such a phosphorylated FGAR intermediate, we would like to propose as a possibility a revision of a previously published scheme (9) in

234

John M. Buchanan

which the amide of glutamine is transferred as a powerful nucleophilic "NH 2 ~ incipient anion" to the carbonyl carbon of FGAR forming a tetrahedral intermediate. The subsequent transfer of the carbonyl oxygen to the phosphorus of the enzyme-phosphate intermediate would result in the formation of 5'phosphoribosylformylglycinamidine (FGAM), ADP and inorganic phosphate, all of which would then be released from the enzyme site. Although a scheme of this kind might well be concerted, we would make no commitment at this point as to whether the reaction involves one or more than one step. However, the proposed mechanism does give equal weight to the nucleophilic "push" of an incipient NH 2 ~ anion and the electrophilic "pull" of a phosphate derived from the y phosphate of ATP. NH 3 would enter the scheme by an abstraction of a proton by a basic group on the enzyme. Thus, the NH 2 ~ anion becomes the common intermediate produced from ammonia or the carboxamide nitrogen of glutamine. The relatively greater reactivity of glutamine as compared to ammonia as a nitrogen source may thus be related to the rate of NH 2 " anion formation. In support of this mechanism one is reminded that glucosamine 6-phosphate synthetase utilizes glutamine but not ammonia as a nitrogen source, and that ATP is not required in the reaction (1). The carbinol oxygen of fructose 6-phosphate is presumably transferred to a proton to form water.

The glutamine active site It is now known from several studies on the glutamine-utilizing enzymes that the enzyme is composed of at least two domains, one of approximately 25,000 daltons that is concerned with the binding of glutamine, the transfer of an amido group and finally the release of an equivalent of product, glutamate. A second domain, which may be of variable size, depending on the enzyme source, contains a so-called amination site and a site for the remaining substrates, e.g. FGAR·MgATP in the case of FGAR aminotransferase. In some instances these two domains are fused to yield a larger enzyme. In the absence of the other substrates, glutamine undergoes a baseline reaction to yield glutamate and ammonia at about 0.5 percent the rate of reaction in a complete system. Similarly, in the absence of glutamine ammonia may substitute as a nitrogen source, but the reaction rate with ammonia is only 1/50 that with glutamine when both substrates are compared under optimal conditions. FGAR amidotransferase can also catalyze the hydrolysis of esters, thioesters and substituted amides of glutamic acid in addition to that of the amide bond of glutamine. The hydrolytic reactions are dependent on the presence of FGAR, ATP and Mg 2 + ; however in contrast to the reaction of glutamine they are not accompanied by the stoichiometric formation of ADP in the case of the amides and there is no ADP formation when esters and thioesters are substrates (11). The y-glutamyl-substituents may thus be wholely or partially transferred to a proton rather than to the keto function of FGAR. However, analogs of FGAM may be formed in proportion to the amount of ADP produced. Use of these derivatives thus permits a study of the catalytic properties of the glutamine site and its relationship to the ammonia site.

The Amidotransferases: Origins of the Concept of Affinity Labeling of Enzymes

235

Inhibition of F G A R amidotransferase by affinity reagents; identification of a sulfhydryl group at the active site Modification of the enzyme at any one of these sites (i.e. presence or absence of normal substrates or inhibitors) may affect reaction at the other sites. Let us consider first the characteristics of modification of the glutamine site by such reagents as L-azaserine, 6-diazo-5-oxo-L-norleucine, and a third antimetabolite, Lalbizziin (Figure 1) on the one hand, and cyanate, iodoacetate and iodoacetamide on the other (12). The first three compounds are competitive with glutamine in their reaction at the active site, but at different rates cause an irreversible inhibition of the enzyme activity. The methods and equations developed by Kitz and Wilson (13) and extended by Mares-Guia and Shaw (14) can be applied to these inhibitors if they participate in a reversibly formed enzyme-inhibitor complex prior to the irreversible reaction resulting in enzyme inactivation. Such a complex should show the saturation behavior expected from an enzyme provided the irreversible reaction rate is low. The anticipated reaction scheme forming inactive enzyme (E') is E + I è

[ E · I] ^

E'

(5)

Solution of the rate equations results in 1

^app

1 =

¥3

K

A;, 1 +

k3 1I

K

(6)

The experimentally determined reciprocal rate constant (1 /k a p p ) can then be plotted against 1/1 according to the above equation. If the intercept of this line on the ordinate is positive, this is taken to indicate the formation of a saturable complex of enzyme and inhibitor prior to the covalent reaction that results in the inactivation of the enzyme. If the line passes through the origin, the apparent rate constant is shown to be strictly proportional to inhibitor concentration. From the slope of this line and the positive ordinate intercept value, both the reaction rate constant (k 3 ) and the dissociation constant (k2/ki) for the enzyme inhibitor complex can be calculated. Albizziin satisfies completely these criteria of a competitive, active-site directed reagent. L-azaserine and DON likewise may be classified as such by the same analytical procedure. Being more closely related structurally to glutamine, D O N inhibits F G A R amidotransferase at 1 /40 the concentration needed by L-azaserine to effect an equivalent rate of inactivation. The stereospecificity and chain length of the inhibitors are important in the fitting of the antimetabolites to the active site since D-azaserine and the homolog of DON, 5-diazo-4-oxo-L-norvaline, are ineffective as inhibitors. All three inhibitors inactivate the enzyme by an enzyme catalyzed reaction with a crucial sulfhydryl group. The evidence for L-azaserine has been obtained by use of 14 C-labeled compound and isolation of the cysteinyl TV-acetyl serine adduct. The relative stability of the enzyme-albizziin adduct in alkaline solution also indicated the participation of a sulfhydryl group in its formation.

236

John M. Buchanan

It is important to point out that the mechanism of inhibition by azaserine and DON varies in an essential feature from that of albizziin. In the case of the former two inhibitors the diazo group is an excellent leaving group thus creating a carbonium ion that reacts chemically with the active site sulfhydryl group. The attack of the nucleophilic group on the carbonyl carbon of DON in a manner comparable to that with glutamine can occur in at least one instance resulting in a cleavage of a carbonto-carbon bond. For example, Hartmann (15) has shown that glutaminase reacts with L - D O N to yield glutamate and methanol (from breakdown of diazomethane). However, for each 70 cycles of the above reaction, the enzyme site eventually becomes alkylated and irreversibly inactivated. As yet no evidence is available for such a carbon-to-carbon cleavage of azaserine or D O N with F G A R amidotransferase. The reactions of azaserine and DON are thus considered to be out of register with the nucleophilic sulfur. However, the reaction of the active site sulfur with the

l / I (ΙΟ 3 )

M' 1

Figure 3 A comparison of the kinetics of inactivation of FGAR amidotransferase by potassium cyanate and albizziin.

The Amidotransferases: Origins of the Concept of Affinity Labeling of Enzymes

237

carbonyl of albizziin is in register. An TV-substituted ^-carbamoyl linkage ( - S - C O - N H - C H 2 - C H N H 2 - C O O H ) is probably formed with the elimination of ammonia. The rate of inactivation of the enzyme by azaserine or DON is not affected by the loading of the complementary site with FGAR-MgATP. However, contrary to expectation, the rate of inhibition of the enzyme by albizziin in the absence of the FGAR-MgATP pair was twice or three times greater than in its presence. This anomaly may be explained either by the effect of the FGAR · MgATP pair inhibiting the binding of albizziin to the enzyme or promoting its dissociation once bound. A second group of compounds including iodoacetate, iodoacetamide and cyanate reacts with the active site sulfhydryl at the glutamine site, but cannot be considered as active-site directed reagents in the same sense as are azaserine, DON and albizziin. None of the former group of inhibitors forms a saturable enzyme complex and therefore the plot of l/&app vs. 1/1 passes through the origin (Figure 3). Under forcing conditions all of the three reagents would react with more than the activesite sulfhydryl group. Although the rate of inactivation of the enzyme by cyanate is not affected by the FGAR · MgATP pair, the rate of inactivation by iodoacetate or iodoacetamide is accelerated 20 to 30 fold by the presence of the other substrates.

Comparison of the amino acid sequences around the active sites of the amidotransferases FGAR amidotransferase was the first of the amidotransferases for which an amino acid sequence around the active cysteine was obtained. The data were derived from proteolytic digestion of enzyme from Salmonella typhimurium treated with Lazaserine labeled with 14 C in the diazo carbon (7) and later with chicken liver enzyme labeled with 14 C-iodoacetate (16). In either case progress was frought with the difficulty of isolation and purification of minute amounts of proteolyticallyderived, sulfur-containing fragments, which were highly susceptible to autoxidation. Nevertheless, these studies constituted one of the first conclusive demonstrations along with that of Harris, Meriwether and Park (17) on 3-phosphoglyceraldehyde dehydrogenase that a sulfhydryl group of an amino residue of an enzyme could function as a nucleophilic agent in an enzymatic reaction. The amino acid residues at the glutamine active site for the Salmonella enzyme were ala-leu-gly-val-cys, and for the chicken liver enzyme, gly-val-cys-asp-asx-cys-glx. Because of the sparsity of sample a distinction of glu vs gin could not be made for the terminal residue, but as will be reviewed later this residue is in all probability a glutamine. The glutamyl active site of the chicken liver FGAR amidotransferase is apparently unique in that a 14C-labeled polypeptide of 21 to possibly 28 residues contained on acid hydrolysis only five amino acids, namely glycine, valine, glutamic acid, aspartic acid and cysteinyl residues either in disulfide linkage or as a carboxymethyl derivative. The total sequence of this polypeptide should, therefore, prove to be very interesting.

238

John M. Buchanan

Alignment

of

Conserved

Regions

of

E. c o l i G A P C C F S.

M S A P T G

P S II B S S P S A M S

subtil is

Consensus *

Amide T r a n s f e r

Domains

* NH„ HHÍ HHÍ mí HHÍ NHJ

-

-

P P P P P Ρ Q Τ L

V I I V Y A

F I L F L L

G V C V G M Q G I C L G H Q G V C L G H Q G I C L G H Q G I C L G M Q G V C Ν G C q

-

-

/ / / /

-

-

H Ρ E Η Ρ E Η Ρ E Η Ρ E - - / - - H P E - - / - H P E -

-

t y p h i m u r i um

F G A M S B,

Glutamine

A L G V C P V L G V C N G F Q - - / - - H P E Ρ - - G - C - G - Q - - / - - H Ρ E

-

-

R e a c t i v e c y s t e i n e . GMPS, g u a n y l a t e s y n t h e t a s e ; A S I I , anthranilate s y n t h e t a s e component I I ; PABS, p-aminobenzoate s y n t h e t a s e ; CPS, carbamoyl phosphate s y n t h e t a s e ; CTPS, c y t i d i n e t r i p h o s p h a t e s y n t h e t a s e ; F G A M S , FGAM s y n t h e t a s e o r F G A R a m i d o t r a n s f e r a s e

Figure 4 Conserved regions of several glutamine amide transfer domains (from Ebbole and Zalkin (24)).

As a result of the introduction of the technique of the sequencing of DNA nucleotides great strides have been made in the determination of the derived amino acid sequence of the several amidotransferases (Figure 4). This group of enzymes may be divided into two categories. In Group I are included GMP synthetase (GMPS) (18), anthranilate synthase component II (ASII) (19), p-aminobenzoate synthase subunit II (PABS) (20), carbamoyl phosphate synthetase (CPS) (21), CTP synthetase (CTPS) (22), FGAR amidotransferase (FGAMS) (23) from E. coli, and finally FGAR amidotransferase from B. subtilis (24). Taken as a group there are essentially two domains that contain either identical or similar residues. The first domain contains the active-site cysteine. The sequence gly-x-cys-x-gly-x-gln is strictly conserved in all types and species of enzymes so far studied. A possible exception to this statement is the finding of aspartate in the fifth residue for the chicken liver FGAR amidotransferase. In view of the fact that aspartate and glycine appeared as major and minor components, respectively, in this cycle of amino acid sequencing, and since aspartate may have been present as a carry over of the previous cycle, a reexamination of the analysis of the amino acid at this location should be made (16). The importance of the active-site cysteine in glutamine function has been confirmed by site-directed mutagenesis effecting replacement of the cysteine by a glycine in the case of carbamoyl phosphate synthetase from E. coli (25) and anthranilate synthetase from S. marcescens (26). Neither of the two mutant enzymes was enzymatically active.

The Amidotransferases: Origins of the Concept of Affinity Labeling of Enzymes

1. FORMATION

239

OF GLUTAMINYL ENZYME ADDUCT

n-SH + IB

> n - S " + HB

0

OH

D-S" + C-NH 2 R 2. TRANSFER

> n-S-C-NH 2 R

OF AMIDE OH

OH n-S-C-NHo + HB

Ι

* Ο- S - C - N HΛÍ

I

+ :B

R

R

r

,

• -S-C-NHÎ

h

s

> Π - S - C + "NHJ

h

3. HYDROLYSIS OF THIOESTER H 2 0 + :B

R

Π-S-C

+ OH"

>OH" + HB » n-S"

A n - S " + HB Figure 5

+ C-OH

h

> n-SH

+ :B

Hypothetical reaction sequence for glutamine amide transfer (from Amuro et al. (27)).

In the second domain there are three strictly conserved consecutive residues, histidine-proline-glutamic acid (ΗΡΕ). By site-specific substitution of histidine 170 by tyrosine in anthranilate synthase from S. marcescens Zalkin and his coworkers (27) have demonstrated that glutamine-dependent enzyme activity was undetectable, whereas the ammonia-dependent activity was unchanged. Affinity labeling of the active site Cys-84 in anthranilate synthetase II by DON was used to distinguish whether His-170 has a role in the formation or in the breakdown of the glutaminylCys-84 intermediate. The interpretation of the results of this experiment is that His170 functions as a general base to promote glutaminylation of Cys-84 (Figure 5). Reversion analysis indicated that His-170 is important catalytically rather than structurally. Zalkin has made the interesting proposal that Cys-84, His-170 and Glu-172 are aligned in a charge relay complex to increase the ionization of the sulfhydryl group of Cys 84, thus increasing its nucleophilicity. The histidine is thus involved in the proton exchange concerned with the formation of the y-glutaminyl complex and the subsequent transfer of the amide group. Further consideration is needed, however, about the mechanism of the hydrolysis of

John M. Buchanan

240

the y-glutamyl complex. Because of the relative stability of the thio ester, it seems likely that a further step involving another nucleophilic residue, e. g. glutamic acid, would be required to yield a product, such as an anhydride, that would easily undergo hydrolysis. The second group of amidotransferases are represented by PRPP-glutamine amidotransferase (28,29) and glucosamine-6-phosphate synthetase (30) from E. coli. In either case the active-site cysteine is located at the amino terminus of the enzyme, the initiating methionine or other residues having been lost by posttranslational processing. Again, both enzymes are inactivated by reaction with DON. PRPP-glutamine amidotransferase has also been isolated from B. subtilis and its deduced amino and sequence determined (31). Replacement of the TV-terminal cysteine of the B. subtilis enzyme with phenylalanine results in loss of enzymatic activity (32). Glucosamine-6-phosphate synthetase utilizes glutamine but not ammonia as a nitrogen source, whereas PRPP-glutamine amidotransferase utilizes both. As mentioned previously, there must be a residue on the enzyme (separate from the cysteine sulfhydryl) that can abstract a proton from N H 3 as a prelude to its participation as a nitrogen source in the reaction. Using the technique of intracistron complementation, Sampei and Mizobuchi (33) have demonstrated that PRPP-glutamine amidotransferase is comprised of three domains, one for binding of PRPP, one for glutamine (to yield a y-glutamyl-enzyme complex) and a short third domain that lies between the two. The PRPP-glutamine amidotransferases isolated from yeast (34), E. coli and B. subtilis contain a consensus sequence within this third domain that has a significance to be discussed in the next section. yeast E. coli B. subtilis

Y Y Y

M L M V V V

A A A

S S S

E E E

It should be noted that neither PRPP-glutamine amidotransferase nor glucosamine6-phosphate synthetase contains the consecutive ΗΡΕ tripeptide present as part of the active site of the enzymes of Group I. Possibly, another conserved histidine and glutamic acid are brought into proximity to Cys-1 by conformational adjustments to form a similar type of charge complex as described above.

Factors Involved in the Utilization of Ammonia The highly nucleophilic analogs of ammonia, e.g. hydroxylamine and hydrazine, have been useful in exploring the amination site and its relationship to the glutamine active site of the enzyme (8,12,35). The foregoing experiments with y-glutamylhydroxamide and •y-glutamylhydrazide were carried out in the possible anticipation that an enzyme-directed binding of a substrate-like reagent could be utilized to increase the effective concentration of reagent near the active site and thereby promote the reaction with the enzyme (11). The design of specific reagents of this type has been reviewed by Baker (36). For example, if both the amide

The Amidotransferases: Origins of the Concept of Affinity Labeling of Enzymes

241

nitrogen of glutamine, and ammonia (from NH 4 C1) enter the overall synthetic reaction forming F G A M from some similar intermediate, then it could be anticipated that a substituted nucleophile (e. g. y-glutamylhydrazide or y-glutamylhydroxamide) might react more selectively with the ammonia site than does the free nucleophile (e. g. hydrazine or hydroxylamine). Such, however, did not prove to be the case. The studies with these compounds were complicated by the fact that both nucleophiles and their y-glutamyl derivatives are inhibitors of 5'-phosphoribosyl 5aminoimidazole synthetase, a coupling enzyme required for the assay of F G A M . However, in later studies this technical difficulty was overcome sufficiently to permit the demonstration that hydroxylamine destroyed the ability of F G A R amidotransferase to utilize glutamine. This inactivation was a function of both inhibitor concentration and duration of incubation when the latter was carried out in the absence of glutamine prior to the assay reaction. However, the enzyme was fully reactive with ammonia as substrate (12). When the incubation with hydroxylamine was carried out in the presence of glutamine, the enzyme retained activity and y-glutamylhydroxamate was formed (35). Because the Km of glutamine is approximately 20 times lower than that of y-glutamylhydroxamate (11) the active site was occupied by glutamine, and hydroxylamine replaced water in the splitting of the yglutamyl thioester on the enzyme. The reactivity of the enzyme with ammonia is sometimes affected by the type of reagent used to block the glutamine site. Thus, with azaserine- and iodoacetamidetreated enzyme the ammonia activity is increased from 50 to 200 percent (9,12). With iodoacetate-inactivated enzyme the ammonia activity is decreased by 50 percent. With enzyme treated with either cyanate or albizziin there is no effect on the activity with ammonia. This uncoupled action between the agents reacting at the glutamine site and the extent of reaction with ammonia implies that even though the glutamyl site and the ammonia site act in concert on the normal substrate, glutamine, additional positional factors dependent on the structure of the inhibitor group introduced can also influence the enzyme catalysis. The above discussion illustrates the difficulty of establishing an area of the enzyme corresponding to an active site either for ammonia furnished as NH 4 C1 or derived from the y-carboxamide of glutamine. Following the finding of Zerner and his coworkers (37) that urease is a nickel-containing enzyme, we had hoped to identify a metal ligand that might be located at the ammonia site and be able to complex with ammonia. Such, however, proved not to be the case for highly purified F G A R amidotransferase. As mentioned above Sampei and Mizobuchi (33) have identified a third domain of PRPP-glutamine amidotransferase of E. coli containing a conserved sequence, Y - ASE, that until now has had no assigned function. They have recently obtained a mutant of the enzyme that contains a single mutation in which the glutamate (Glu209) has been changed to a lysine. This mutant cannot utilize either glutamine or ammonia for growth. It is propagated only by supplying adenine or by intracistronic complementation with a plasmid containing the normal sequence of nucleotides in the region where the mutation has occurred on the chromosomal gene. Although their data may have other interpretations, the fact that both glutamine and

242

John M. Buchanan

ammonia utilization is blocked by the mutation at Glu-209 argues for this area of the enzyme as the ammonia site. Further studies with mutations of the other amino acids in this conserved sequence are underway.

Origins of the Concept of Affinity Labeling of Enzymes Labeling of enzymes with active-site directed modifying reagents The history of the chemical modification of proteins to obtain information about their structural or functional properties is a long one. The chemicals that have been commonly used have been either acylating or alkylating reagents. Frequently these reagents have reacted with proteins with first order kinetics indicating that the rate or extent or reaction is primarily a function of the concentration of the reagent. When applied to the study of enzymatic reactions it is often difficult to determine whether such reagents are inhibiting the reactivity of an enzyme because of some nonspecific reaction with a group important to maintenance of the enzyme's structure or whether a functional group at the enzyme site has been covalently modified. In general it was this type of reagent that was initially used to study enzyme function. When additional supportive information was available about catalytic properties of a given enzyme, use of reagents of this kind tagged with an appropriate radioactive label was immensely successful. Otherwise, the interpretation of data from the modification of enzymes by these reagents could be inconclusive or at times misleading. The dilemma is best illustrated by a comparison of the use of modifying reagents in the exploration of the residues at the active sites of two enzymes, papain and fumarase. As early as 1939 Balls and Lineweaver (38) titrated crystalline papain with iodine and found approximately one sulfhydryl group per mole of protein. Treatment of papain with excess iodoacetic acid abolished the iodine reaction. In their comprehensive studies on the composition and structure of highly purified papain Finkle and Smith (39) reported in 1958 that slightly more than one equivalent of iodoacetic acid produces complete inhibition of the enzyme. By 1964 this single thiol group was located at Cys-25 near an aspartyl residue believed to contribute to its ionization and hence nucleophilicity (40). On the other hand the conditions required to modify the twelve thiol groups of fumarase by reagents such as />-mercuribenzoate, iodoacetate or iodoacetamide suggest that the thiols are not at the active site but are buried in the hydrophobic regions of the enzyme (41). Although the thiol groups do not appear to be in the active site, modification of the thiol groups leads to inactivation of the enzyme with loss of activity proportional to the number of thiol groups modified. Some evidence suggests that the thiol groups of fumarase may reside in or near the contact regions between subunits (42) and hence are important in maintaining the structural integrity of the enzyme. In an attempt to identify structures in the active site of

The Amidotransferases: Origins of the Concept of Affinity Labeling of Enzymes

243

fumarase by affinity labeling Bradshaw et al. (43) have encountered difficult problems associated with studies of this type because of the high reactivity with thiol groups of most compounds studied. "Thus, without careful consideration of the possibility of thiol group modification attempts at affinity labeling of fumarase may be misleading (44)". The successful use of modifying reagents has depended sometimes on unique properties of an enzyme such as: 1. the availability of an amino acid at the enzyme surface, 2. the unusual ionization of a special group such as the hydroxyl of a seryl residue in the case of the serine proteases, and 3. the good fortune of working with an enzyme that does not contain more than one available reactive side chain. With these precautions in mind there have been several successful uses of alkylating or acylating reagents to produce an active-site directed, irreversible inactivation of an enzyme. For example, the classical experiments of Moore and Stein with iodoacetate or bromoacetate identified two important histidyl residues at the active site of ribonuclease (45-48). The protection against enzyme modification aiforded by anion binding (phosphate or sulfate) indicated that these reagents were reacting at the active site of ribonuclease. Again, there was a supportive background of experiments that led up to the use of diisopropylfluorophosphate (DFP) for labeling the enzyme site of trypsin and chymotrypsin. As early as 1949 A.K. Balls and his coworkers (49-51) found that reaction with DFP produces a time-dependent, irreversible inactivation of chymotrypsin suggesting covalent modification. Protection was observed with substrates implying that the process may be active-site directed. Using [ 3 2 P]-DFP, Jansen et al. found stoichiometric incorporation of a radioactive phosphate label in which the kinetics of labeling matched enzyme inactivation. After hydrolysis in acid solution a radioactive derivative was isolated and identified as O-phosphoryl serine (52). The phosphorylated serine residue was later found to be serine-195 (53). Further characterization of the enzyme site was achieved by Hartley and Kilby (54). These workers tested the reactive ester, p-nitrophenylacetate, as a substrate for chymotrypsin. A burst of/7-nitrophenol was released in proportion to the amount of enzyme present. Thereafter, there was a slower but steady production of the product. Eventually with the 14C-labeled acetyl substrate, a radioactively labeled enzyme complex was isolated, which, although stable enough to permit isolation, could not be used for enzyme site characterization. Substituted acetyl derivatives proved to be more stable. The interaction of a-chymotrypsin with substrates and inhibitors results in conformational changes in the enzyme that can be measured by changes in the absorption spectrum at 290n. Moon. Sturtevant and Hess (53) have now reported that within the pH range 5.5 and 10 there is a direct relationship between the spectral changes at 290n and the phosphorylation of the enzyme by DFP, the latter being measured by the liberation of a proton. Their scheme places a reversibly formed EI complex, represented by the spectral change, in the sequence of reactions prior to the irreversible phosphorylation step. The Ks of the proposed enzyme inhibitor

244

John M. Buchanan

complex is too great to be measured by traditional kinetic procedures. In spite of these findings and even though it selectively reacts with only one serine among many on a given protease, DFP cannot operationally be placed in the category of an affinity reagent since it reacts indiscriminately with chymotrypsin and trypsin, two enzymes of such different substrate specificities.

Design and Characteristics of a Reagent for the Affinity Labeling of Enzymes These considerations lead to discussion of another type of active-site directed, irreversible inhibitor of enzyme activity mentioned above. This type of inhibitor contains within a given molecule both the potential of mimicking the structure of the normal substrate at the active site and also a reactive group that can irreversibly bind the inhibitor to the reacting group in the site. A reagent of this kind should show competitive kinetics with the normal substrate, yet cause an eventual irreversible inhibition of the enzyme according to reactions 5 and 6. If the irreversible step is relatively rapid compared to the rate of formation of the inhibitor-enzyme complex, the kinetics may appear to be a mixed inhibition, i. e. the lines plotting l/£ a p p vs. 1 /I at various inhibitor concentrations converge to the left of the ordinate rather than at the ordinate as in the case of a strictly competitive inhibition (equation 6). As far as I am aware, the naturally occurring antibiotics, L-azaserine and its synthetic analog, L - D O N , and later albizziin were the first reagents of this type to be recognized by the characteristics and kinetics of their inhibition of FGAR amidotransferase. In 1955 a short note described inhibition by L-azaserine of the enzymatic synthesis of inosinic acid in vitro and the protection against this inhibition afforded by glutamine (2). The accumulation of FGAR was noted with the statement that the site of inhibition would occur at a subsequent step in the series of reactions. In 1957 the complete kinetic analysis of the inhibition by both L-azaserine and L - D O N was published along with information about the importance of the stereospecificity and chain length of the latter (3). After a brief preliminary report in 1959 (4) the identification of the active site cysteine was established in 1963 as the residue reactive with 14 C-azaserine (6). In connection with their studies on the esters of methanosulfonic acid as irreversible inhibitors of acetylcholinesterase, Kitz and Wilson (13) had published in 1962 a kinetic analysis of the reaction of these inhibitors with the enzyme. In the same year S.J. Singer and his colleagues (56) had applied these concepts to the reactions of antigens with antibodies. The latter were responsible for coining the useful term "affinity-labeling" to reactions with this kind of inhibitor. By these criteria affinitylabeling by an active-site directed reagent could be distinguished from reactions by such compounds as DFP and iodoacetate among others, which may combine with a residue at the active site primarily because of the latter's unique reactivity. B.R. Baker recognized both the power and specificity of active-site directed, irreversible inhibitors and was responsible between 1961 and 1969 for devising many novel reagents capable of effecting the specific inhibition of several enzymes. In one

The Amidotransferases: Origins of the Concept of Affinity Labeling of Enzymes

245

of his earlier attempts to design a specific enzyme alkylator, Baker (57) synthesized 4-iodoacetamidosalicylic acid as a possible irreversible reagent for glutamic dehydrogenase. The salicyl moiety contributed the potential for association as a complex at the active site of the enzyme, whereas the iodoacetamido moiety contributed the capacity to react irreversibly with the enzyme. This paper was a landmark in the concept of affinity labling of enzymes because it represented one of the first efforts to design reagents of this type. The extensive research from Baker's laboratory is summarized in a review in 1964 (58) and in a book published in 1967 (36). The goals of Baker's laboratory were the obtaining of selective enzyme inhibitors frequently with the objective of therapeutic application. For the most part their studies were not concerned with establishing the details of selective protein modification. This latter approach was, however, a major objective of Elliot Shaw and his collaborators, who possibly more than anyone were responsible for designing and synthesizing selective enzyme inhibitors for the purpose of investigating the mechanism of proteolytic enzyme activity. Recognizing that the chloromethylketones are powerful alkylating reagents, Shaw synthesized a number of amino acid derivatives based on previous information about the properties of synthetic substrates for individual proteases. In a now classical study Schoellmann and Shaw (59,60) prepared tosyl L-phenylalanylchloromethylketone (TPCK) that is a specific affinity reagent for chymotrypsin. Later, tosyl L-leucylchloromethylketone was synthesized and found to be specific for trypsin (61,62). Both reagents exhibited a saturation effect in the kinetics of the enzyme inactivation process, indicating the intermediate formation of a reversible complex before the inactivation step. In either case a histidyl residue could be identified as an important component at the active site. As is now known from x-ray crystallographic studies, this histidine and an adjacent aspartic acid comprise a charge relay complex to effect the ionization of the hydroxyl group of a specific seryl residue. The more versatile aromatic sulfonyl derivatives may either sulfonylate or alkylate the active site residues depending on the structure of the particular compound. By preparing peptidyl derivatives of the sulfonyl or chloromethylketone inhibitors corresponding to peptide bonds a given enzyme is known to cleave, Shaw (63,64) has been able to develop compounds capable of the specific and selective inhibition of individual members of the serine proteases by one mechanism or the other. Two excellent reviews by Shaw on the research on the protease inhibitors up to 1970 are available (65,66).

Examination of proximal amino acid residues within the active site Compounds have been designed with reactive groups at either end of the molecule, which have the capacity of linking two amino acid residues at the active centers. Lawson and Schramm (67) have used the bromoacetamido analog of the pnitrophenyl ester of dimethylacetic acid ( B r - C H 2 - C O N H - C ( C H 3 ) 2 - C O O R ) to acylate first the reactive serine of chymotrypsin. Because of steric hindrance this intermediate has a moderate stability and hence a capacity for reaction with a

246

John M. Buchanan

proximal residue if positional interactions permit. By this method Lawson and Schramm were able to locate methionine 192 at the active center of chymotrypsin. In an even more novel approach Westheimer and his colleagues (68) recognized the utility of acyl enzymes for enzyme modification. Upon photolysis of diazoacetylchymotrypsin a carbene is generated as an acyl enzyme, which then irreversibly attacks some nearby side chain. In the particular instance cited above with 14 C-diazoacetylchymotrypsin, a covalent modification was achieved and O-carboxymethyl serine, 1-carboxymethylhistidine and O-carboxymethyltyrosine were present in the products (69). The attractive feature of this approach is its potential for labeling aliphatic side chains of all types by carbene insertion into C — H bonds (70). Chowdhry and Westheimer (71) have reviewed the extensive application of this method in the exploration of the active sites of several enzymes. This method is capable of describing best the chemical environment of the active center, short, of course, of an x-ray crystallographic analysis.

Suicide Inhibitors Bloch and his colleagues (72-74) are responsible for the discovery of the first example of a class of compounds known as "suicide" inhibitors of enzymatic action. These reagents have claimed the attention of biochemists and enzymologists by virtue of their novel characteristic of serving as pseudo substrates of the enzyme, participating in the normal reaction, and finally yielding a product that reacts rapidly and irreversibly with one of the amino acids at the active site to render the enzyme inactive. The discovery of "suicide" compounds was in fact part of a more general program originated by Bloch (72) in which he described for the first time the pathway for the synthesis of unsaturated fatty acids by microorganisms when operating under strictly anaerobic conditions. In the anaerobic synthesis of unsaturated fatty acids microorganisms utilize some aerobic reactions in common but differ decisively in the manner in which the unique double bond is introduced into the fatty acid chain, e. g. position "9" and " i l " for palmitoleic and oleic acids, respectively. Anaerobically Escherichia coli utilizes an enzyme, ß-hydroxydecanoyl thioester dehydrase, to catalyze the conversion of the /^-hydroxy decanoyl thioester, a normal intermediate in the synthesis of the saturated fatty acids, first to the trans α, β diene and then to the cis ß, y diene. By subsequent cycles of addition of two carbon units the longer naturally occurring unsaturated fatty acids are produced. The amino acids at the active site concerned with these hydrogen exchanges are now known to the histidyl and tyrosyl-residues. In the process of searching for competitive inhibitors of the reaction, Bloch (73) found that the triple bonded 3-decynoyl thio ester caused a striking irreversible inactivation of the dehydrase. Principally because of his previous research on the kinetic properties of the enzyme he deduced that one round of hydrogen abstraction and addition must have occurred to yield the 2,3-decadienoyl thio ester, which in fact is the true inhibitor that reacts chemically with an active-site histidyl residue. It

The Amidotransferases: Origins of the Concept of Affinity Labeling of Enzymes

247

was from this type of elegant enzymology and chemistry that the concept of a "suicide" inhibitor was born. The concept of "suicide" inhibition has been greatly extended to other systems over the last decade or more by Rando (75), Abeles (76) and Walsh (77). A comprehensive review of inhibitors of this type up to 1978 has been published by Seiler, Jung and Koch-Weser as a collection of reviews entitled "Enzyme-Activated Irreversible Inhibitors" (78).

Concluding Remarks A younger colleague of mine has characterized the last four or five decades as the "romantic" period of biochemistry and molecular biology. He explains that from his point of view this was the time when many of the great concepts of these two disciplines were developed as a necessary prelude to the almost unlimited explosion of information reaching our desks today. Fritz Lipmann was indeed one of biochemistry's great conceptualizers. In addition to his role in explaining the energetics of substrate activation by ATP, he isolated a new coenzyme, made important contributions to the enzymology of protein synthesis and discovered carbamoyl phosphate. The list of his accomplishments continues with his research on sulfate activation, gramicidin biosynthesis and later in his life the uptake of glucose by transformed cells. His modest and self effacing personality belied his agile and penetrating mind. I remember having lunch with Fritz at a not too recent meeting of the American Society of Biological Chemists. Fritz had just presented a poster on the effects of viral transformation on the uptake of glucose by cells in vitro. It was an enviable contribution but, in spite of this, he felt depressed with his own estimation of its relevance compared to the sweep of new information unfolding at these meetings. He was certainly not alone in his respect for the magnitude of some of the reports, particularly one by Sanger describing the complete nucleotide sequence of ΦΧ 174 D N A (79). Perhaps Fritz realized as did I that one era of biochemistry had come to an end and that the vision ahead was too great to be comprehended. But we are meeting now to celebrate the decades of scientific discoveries that Fritz knew best. Fortunately many of his scientific contemporaries are present to add to the mosaic of science that we all cherish in retrospect.

References 1. Buchanan, J. M. 1973. In: Advances in Enzymology and Related Areas of Molecular Biology (A. Meister, ed.) Wiley, New York. Vol.39, pp.91-183. 2. Hartman, S.C., Levenberg, Β. and Buchanan, J.M. 1955. J. Am. Chem. Soc. 77, 501-502. 3. Levenberg, Β., Melnick, I. and Buchanan, J.M. 1957. J. Biol. Chem. 225, 163-176. 4. Buchanan, J. M., Hartman, S.C., Herrmann, R. L. and Day, R.A. 1959. J. Cell, and Comp. Physiol. 54, Suppl. 1, 139-160. 5. French, T.C., Dawid, I.B., Day, R.A. and Buchanan, J.M. 1963. J. Biol. Chem. 238, 2171-2177.

248

John M. Buchanan

6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

Dawid, I. Β., French, T.C. and Buchanan, J.M. 1963. J. Biol. Chem. 238, 2178-2185. French, T.C., Dawid, I.B. and Buchanan, J.M. 1963. J. Biol. Chem. 238, 2186-2193 (1963). Mizobuchi, K. and Buchanan, J.M. 1968. J. Biol. Chem. 243, 4863-4877. Mizobuchi, K., Kenyon, G.L. and Buchanan, J.M. 1968. J. Biol. Chem. 243, 4863-4877. Li, H.-C. and Buchanan, J.M. 1971. J. Biol. Chem. 246, 4720-4726. Li, H.-C. and Buchanan, J.M. 1971. J. Biol. Chem. 246, 4713-4719. Schroeder, D.D., Allison, A.J. and Buchanan, J.M. 1969. J. Biol. Chem. 244, 5856-5865. Kitz, R. and Wilson, I.B. 1962. J. Biol. Chem., 237, 3245-3249. Mares-Guia, M. and Shaw, E. 1967. J. Biol. Chem. 242, 5782-5788. Hartman, S.C. 1971. In: The Enzymes 3rd ed. (P.D. Boyer, ed.) Academic Press, New York, Vol.4, pp.79-100. Ohnoki, S., Hong, B.-S. and Buchanan, J.M. 1977. Biochemistry 16, 1070-1076. Harris, I., Meriwether, B.P. and Park, J.H. 1963. Nature 198, 154-157. Zalkin, H., Argos, P., Narayana, S.V.L., Tiedeman, A.A. and Smith, J.M. 1985. J. Biol. Chem. 260, 3340-3354. Nichols, B.P., Miozzari, G.F., van Cleemput, M., Bennett, G.N. and Yanofsky, C. 1980. J. Mol. Biol. 142, 503-517. Kaplan, J.B., Merkel, W.K. and Nichols, B.P. 1985. J. Mol. Biol. 183, 327-340. Piette, J., Nyunoya, H., Lusty, C.J., Cunin, R., Weyens, G., Crabeel, M., Charleir, D., Glandorff, Ν. and Pierard, Α. 1984. Proc. Natl. Acad. Sci. U.S.A. 81, 4134-4138. Weng, M., Markaroff, C.A. and Zalkin, H. 1986. J. Biol. Chem. 261 5568-5574. Stubbe, J. and Smith, J.M. Personal Communication. Ebbole, D.J. and Zalkin, H. 1987. J. Biol. Chem. 262, 8274-8287. Rubino, S.D., Nyunoya, H. and Lusty, C.J. 1986. J. Biol. Chem. 261 11320-11327. Paluh, J.L., Zalkin, H., Betsch, D. and Weith, H.L. 1985. J. Biol. Chem. 260, 1889-1894. Amuro, N„ Paluh, J.L. and Zalkin, H. 1985. J. Biol. Chem. 260, 14844-14849. Tso, J.Y., Hermodson, M. A. and Zalkin, H. 1982. J. Biol. Chem. 257, 3532-3536. Tso, J.Y., Zalkin, H., van Cleemput, M., Yanofsky, C. and Smith, J.M. 1982. J. Biol. Chem. 257, 3525-3531. Walker, J.E., Gay, N.J., Saraste, M. and Eberle, A.N. 1984. Biochem. J. 224, 799-815. Markaroff, C.A., Zalkin, H., Switzer, R.L. and Vollmer, S.J. 1983. J. Biol. Chem. 258, 10586-10593. Mäntsälä, P. and Zalkin, H. 1984. J. Biol. Chem. 259, 14230-14326. Sampei, G. and Mizobuchi, K. (Personal Communication). Mäntsälä, P. and Zalkin, H. 1984. J. Biol. Chem. 259, 8478-8484. Mizobuchi, K. and Buchanan, J.M. 1968. J. Biol. Chem. 243, 4842-4852. Baker, B.R. 1967. Design of active-site directed irreversible enzyme inhibitors, Wiley, New York. Dixon, N.E., Gazzola, C., Blakeley, R.L. and Zerner, Β. 1976. Science 191, 1144-1150. Balls, A.K. and Lineweaver, H. 1939. Nature 144, 513. Finkle, B.J. and Smith, E.L. 1958. J. Biol. Chem. 230, 669-690. Light, Α., Frater, R., Kimmel, J.R. and Smith, E.L. 1964. Proc. Natl. Acad. Sci. U.S.A. 52, 1276-1289. Robinson, J.W., Bradshaw, R.A., Kanarek, L. and Hill, R.L. 1967. J. Biol. Chem. 242, 2709-2718. Teipel, J.W. and Hill, R.L. 1971. J. Biol. Chem. 246, 4859-4865. Bradshaw, R.A., Robinson, G.W., Hass, G.M. and Hill, R.L. 1969. J. Biol. Chem. 244, 1755-1763. Hill, R.L. and Teipel, J.W. 1971. In: The Enzymes, 3rd ed. (P.D. Boyer, ed.) Academic Press, New York, Vol. V, pp. 539-571. Barnard, E.A. and Stein, W.H. 1959. J. Mol. Biol. 1, 339-349. Gundlach, H.G., Stein, W.H. and Moore, S. 1959. J. Biol. Chem. 234, 1754-1760. Crestfield, A.M., Stein, W.H. and Moore, S. 1963. J. Biol. Chem. 238, 2413 -2420. Crestfield, A.M., Stein, W.H. and Moore, S. 1963. J. Biol. Chem. 238, 2421-2428. Jansen, E.F., Nutting, M.F., Jang, R. and Balls, A.K. 1949. J. Biol. Chem. 179, 189 -200. Jansen, E.F., Nutting, M.F. and Balls, A.K. 1949. J. Biol. Chem. 179, 201-204. Jansen, E.F., Nutting, M.F. and Balls, A.K. 1950. J. Biol. Chem. 185, 209-220.

16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51.

The Amidotransferases: Origins of the Concept of Affinity Labeling of Enzymes 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79.

249

Schafler, Ν.Κ., May, S.C. and Summerson, W.H. 1953. J. Biol. Chem. 202, 67-76. Hartley, B.S. 1964. Nature 201, 1284-1287. Hartley, B.S. and Kilby, B.A. 1954. Biochem. J. 56, 288-297. Moon, A.Y., Sturtevant, J.M. and Hess, G.P. 1965. J. Biol. Chem. 240, 4204-4212. Wofsy, L., Metzger, Η. and Singer, S.J. 1962. Biochemistry 1, 1030-1039. Baker, B.R., Lee, W.W., Tong, E. and Ross, L.O. 1961. J. Am. Chem. Soc. 83, 3713-3714. Baker, B.R. 1964. J. Pharm. Sci. 53, 347-364. Schoellmann, G. and Shaw, E. 1962. Biochem. Biophys. Res. Commun. 7, 36-40. Schoellmann, G. and Shaw, E. 1963. Biochemistry 2, 252-255. Shaw, E., Mares-Guia, M. and Cohen, W. 1965. Biochemistry 4, 2219-2224. Shaw, E. and Springhorn, S. 1967. Biochem. Biophys. Res. Commun. 27, 391-397. Kettner, C. and Shaw, E. 1978. Biochemistry 47, 4778-4784. Wong, S.C. and Shaw, E. 1976. Arch. Biochem. Biophys. 776, 113-118. Shaw, E. 1970. Physiological Reviews 50, 244-296. Shaw, E. 1970. In: The Enzymes, 3rd ed. (P.D. Boyer, ed.) Academic Press, New York, Vol. 1, pp. 91-146. Lawson, W.B. and Schramm, H.J. 1965. Biochemistry 4, 377-386. Singh, Α., Thornton, E.R. and Westheimer, F.H. 1962. J. Biol. Chem. 237, PC 3006-3008. Shafer, J., Baronowsky, P., Laursen, R., Finn, F. and Westheimer, F.H. 1966. J. Biol. Chem. 241, 421-427. Vaughan, R.J. and Westheimer, F.H. 1969. J. Am. Chem. Soc. 91, 217-218. Chowdhry, V. and Westheimer, F.H. 1979. In: Annu. Rev. Biochem. 48, 293-325. Brock, D„ Kass, L. and Bloch, Κ. 1967. J. Biol. Chem. 242, 4432-4440. Endo, K„ Helmkamp, G.M. and Bloch, Κ. 1970. J. Biol. Chem. 245, 4293-4296. Bloch, Κ. 1971. In: The Enzymes, 3rd ed. (P.D. Boyer, ed.) Academic Press, New York, Vol. V, pp. 441 -464. Rando, R. 1974. Science 185, 320-324. Abeles, R.H. and Maycock, A. L. 1976. Accts. Chem. Res. 9, 313-319. Walsh, C.T. 1977. In: Horizons in Biochemistry and Biophysics (E. Quaglieriello, ed.) AddisonWesley, Reading, Mass. Vol. 3, p. 36. Seiler, Ν., Jung, M. J., and Koch-Weser, J. eds. 1978. Enzyme-Activated Irreversible Inhibitors, Elsevier/North Holland Biomedical Press, Amsterdam. Sanger, F., Nicklen, S., and Coulson, A.R. 1977. Proc. Natl. Acad. Sci. U.S.A. 74, 5463 - 5467.

Intracellular Protein Degradation: Past, Present and Future S. Grisolia, E. Knecht and J. Hernández- Yago

Introduction It is widely accepted that intracellular proteins in eukaryotic and prokaryotic organisms are degraded and resynthesized continuously, under steady state conditions. However, this concept has emerged only after nearly two centuries of work. Excellent historical coverage of this subject has been given (1).

Past The french physiologist Francois Magendie (1783-1855) was probably the first to recognize the importance of nitrogenous food in nutrition. Other investigators, including John B. Lawes (1814-1900), Karl von Voit (1831-1908), Frederick G. Hopkins (1861-1947) and William C. Rose, extended these observations and identified protein as the principal source of nitrogen in the human diet. Under certain conditions, an organism losses more nitrogen than is assimilated in the diet, resulting in a negative nitrogen balance. This additional loss of nitrogen should come from intracellular sources. Rudolf Schoenheimer (1898-1941) was the first to use isotopie tracers in biochemical research. His book "The Dynamic State of Body Constituents" (2), published one year after his death, led to the new concept that although the intracellular amount of large molecules was fairly stable, they were "incessantly interchanging small pieces; in the case of proteins, aminoacids or short polypeptides". An interesting aspect of protein degradation is its apparent energy requirement, first briefly mentioned by Simpson (3) in rat liver slices but largely ignored; later it was suggested, as a chemotropic reagent, to be of vital importance for protein turnover (4). At present, although there is much evidence for its involvement in protein degradation, the mechanisms for the stimulation by ATP of protein degradation are still not clear. Summarizing, the main events of the past were: 1. the postulation by Magendie of the importance of nitrogenous foods; 2. the identification of proteins and of aminoacids as necessary in the diet, 3. the recognition of the dynamic state of body constituents, including proteins, 4. the recognition of the requirement for ATP in a large portion of intracellular protein degradation.

252

S. Grisolía, E. Knecht and J. Hernández-Yago

Present Metabolism involves many chemical reactions catalyzed by hundreds of enzymes. Several factors control intracellular enzyme levels and activity. Enzyme activity regulation in higher organisms is achieved by: 1. binding of metabolites to enzymes; 2. interconversion between covalently modified and unmodified forms or enzymes; 3. compartmentation; and 4. action of hormones. The intracellular level of an enzyme is determined by: 1. the kinetics of synthesis, usually controlled by the amount of mRNA available, and 2. the kinetics of its degradation. All intracellular proteins in eukaryotic and prokaryotic organisms are continuously being degraded, indeed, as predicted many years ago (4), it has now become apparent that many mechanisms of intracellular protein degradation are operating (5-16). In eukaryotic cells, a major pathway involves lysosomes (17-22), which sequester organelles by autophagy (a process, easily detected by electron microscopy, relatively well established (23-26)), and perhaps small bits of cytoplasm (23,27-32). In addition, other proteolytic systems have been decribed in mitochondrial (33-35) and cytosolic fractions from mammalian tissues (36-41). Protein degradation appears to be selective, since there are large differences in halflives among specific proteins. However, we do not know the mechanisms which accomplish this specificity. Regulation of protein degradation occurs at the proteolytic system level which may determine the overall rate, and at the protein substrate level which may determine the individual degradations (42,43). These may be affected by many factors as follows: structural features of proteins (44,45), including primary aminoacid sequence (46-48) and subcellular localization (49,50), several "marking" reactions, including covalent modification of proteins (e.g. phosphorylation (51), glycosylation (52), mixed function oxidation (53), formation of mixed disulfides (54), carbamylation (55), "labeling" with ubiquitin (7,39, but see 56 for the scope of this system and 57 for a new function of ubiquitin in protein

Table 1 Theories, approaches and main findings presently known on intracellular protein degradation. 1. The bulk of proteins are long-lived. Proteins of very short half-life possibly account for a very small portion of the total, ca. 2 % (58,59), including leading peptides proteolysis, but consume ca. 40% of the energy used for protein synthesis under steady state conditions. 2. Mechanisms to explain the vast differences in half-life of proteins are: a) participation of lysosomes (autophagy, microautophagy); b) correlation between molecular weight, hydrophobicity, isoelectric point, secondary structure, the terminal aminoacids, etc. and degradation rate. 3. The study of proteases and inhibitors in different tissues and cell organelles. 4. The increased and also decreased stability of enzymes to proteases induced by substrates and other environmental compounds. 5. Covalent modifications, including the ATP-induced increased proteolysis, an example of chemotropic agents (4), the probably quantitatively minor ubiquitin systems, etc.

Intracellular Protein Degradation: Past, Present and Future

253

degradation), etc.) and interaction of proteins with their cofactors and the substrates and products of the enzymatic reaction they catalyze (42,43,55). The theories, approaches and main findings presently known are resumed in Table 1. It is apparent that in spite of much work, we need more precise information regarding the extent, mechanism(s) and regulation of intracellular protein turnover. New experimental approaches are therefore required. Although many experiments have been carried out with specific proteases and other measurements, as already indicated in the references throughout, we will concentrate on experiments from our laboratory. We have employed several new approaches as well as continuing with the more standard ones included in Table 2. Table 2 tion.

Approaches used in the authors' laboratory to investigate intracellular protein degrada-

1. Model based on the entry and exit of proteins in mitochondria in order to detect signals which may regulate the close interrelationships between protein synthesis and degradation. 2. The study of protective effect of phosphorylated compounds such as 2,3-bisphosphoglycerate on protease substrates the stability of which is affected by ATP. 3. The importance of hepatocyte heterogeneity for protein degradation. 4. The influence of centrifugation on the disorganization of the Golgi apparatus, and on the concomitant inhibition of degradation of short-lived proteins, while the degradation of long-lived proteins remains unaffected. 5. The incorporation of lysosomal proteins into cultured cells, via endocytosis with liposomes.

1. Concerning interrelationships between protein synthesis and degradation, we have outlined a simple model. The procedure, illustrated in Figure 1, consists of incubating mitochondria, which had been pulse-labeled with a radioactive aminoacid, with cytosolic precursors of specific mitochondrial proteins and then measuring the release of proteins from the mitochondria (60). To refine this procedure we used mitochondrial protein precursors, such as apocytochrome c and cloned cDNAs of carbamoyl phosphate synthase and ornithine transcarbamoylase (61) in order to obtain specific m R N A s for in vitro translation. Results obtained using apocytochrome c, strongly suggest that mitochondrial protein turnover is regulated at the level of synthesis. 2. Low molecular weight phosphoryl compounds such as 2,3-bisphosphoglycerate and carbamoyl phosphate, present or formed in large amounts in some tissues, may protect such mitochondrial and cytosolic proteins as ornithine transcarbamoylase, carbamoylphosphate synthetase, glutamate dehydrogenase and glyceraldehyde 3phosphate dehydrogenase, from proteolytic inactivation by proteases (rat liver lysosomal extracts, pronase, elastase) (62-64). Since this effect is also observed in the accelerated proteolysis which occurs in the presence of ATP (65), this novel type of protection of the substrate protein may serve to clarify the regulation of protein turnover at the substrate level.

254

S. Grisolla, E. Knecht and J. Hernández-Yago

1. Synthesis of mitochondrial protein precursors

POLYRIBOSOME 2. Processing and import

4. Release of proteins and/or Polypeptides (?

MITOCHONDRIA 5. Lysosomal and non-lysosomal degradation

Figure 1 "Push-Pull" model for mitochondrial protein degradation. Most mitochondrial proteins are synthesized on cytoplasmic polyribosomes as larger precursors, which are imported (after cleavage of the leader peptide by a mitochondrial protease). The "push-pull" model postulates that the concentration of a protein in mitochondria may be regulated by the import of their protein precursors from the cytosol. Thus, under steady state conditions, whole or, perhaps, partially degraded (by hypothetical mitochondrial proteases) mature proteins are released into the cytosol as a function of the amount of imported proteins. For the sake of simplicity, autophagy, degradation by mitochondrial proteases or by proteases which may enter mitochondria (12) have not been included in the scheme.

3. Immunogold procedures at the electron microscopic level and immunohistochemical procedures, using polyclonal and monoclonal antibodies for carbamoyl phosphate synthase, ornithine transcarbamoylase and glutamate dehydrogenase, enzymes with vastly different half-lives (66-68), show intracellular homogeneity but intercellular heterogeneity in rat liver (69-74), suggesting a preferential degradation of these proteins by the autophagic-lysosomal system. 4. Degradation of short-lived proteins in cultured cells occurs by unknown mechanisms (75). Centrifugation results in disorganization of the Golgi apparatus together with inhibition of the degradation of short-lived proteins, but not of longlived proteins (76,77). This suggests that this organelle partakes in the degradation of these proteins by controlling the traffic of proteins or proteases to the degradation site(s). Since some precursors of mitochondrial proteins also have very short half-lives (78), it appears that processing and degradation of the signal sequences can contribute significantly to the degradation of short-lived proteins.

Intracellular Protein Degradation: Past, Present and Future

255

5. It has been suggested that the degradation of intracellular proteins in exponentially growing cultured cells occurs mainly by non-lysosomal mechanisms, because lysosomal inhibitors have no effect. To test the possible participation of lysosomes in this degradation, rat liver lysosomal proteins were introduced, via multilamellar liposomes, into L-132 cells which had been previously labeled with radioactive aminoacids. The liposomal content was released into the lysosomes of the cultured cells, as demonstrated by the uptake of ferritin on electron microscopy. Degradation rates of intracellular proteins increased with the uptake of lysosomal proteases, but were not affected by the lysosomal inhibitors chloroquine and leupeptin. Thus, it appears that lysosomal mechanisms insensitive to these inhibitors, participate in the degradation of proteins, in exponentially growing cultured cells.

Future The various approaches which we are currently exploring seem promising and may go a long way towards clarifying the complex mechanisms of intracellular protein degradation. However, as indicated above, even the mitochondrial system, which we selected as an initial model to follow entry and exit of proteins, is more complex than at first realized. Obviously, our task would be much simpler if we knew more about the characteristics of the binding of mature protein precursors and the mechanism for their entry. It appears that if the protein, once it is synthesized and leaves the ribosome, does not enter immediately into the mitochondria, it may be subject not only to degradation in the cytosol but to a number of post-transcriptional modifications. For example, SH-group modifications as postulated by Singer (79). Indeed, if one isolates the mitochondrial protein precursors and adds them to mitochondria, the entry is diminished. It will be important, therefore, to either develop effective ways to obtain mRNA in large quantities or to be able to control, more closely than is now possible, the treatment of reticulocytes and/or post-mitochondrial supernatants with RNAse. Obviously, if all the necessary factors for the synthesis were easily available it would be possible to improve the mitochondrial model. This, together with the increasing number of cDNAs which are becoming available, may permit exploring the mitochondrial model more fully. The highly radioactive 1 3 N (which requires a nearby cyclotron) may help to follow aminoacid incorporation better than, or in combination with, the more commonly available 3 H- or 14 C-aminoacids. But what is really needed are good and specific methods to rapidly identify and isolate individual proteins and their decay products. Predictably these should be available in the near future. There are a number of questions which we believe need clarifying before a better comprehension of protein turnover can be achieved. Although we know a great deal about the intermediate steps, factors and main components of protein synthesis, we still do not know how the cell and/or multicellular organisms regulate synthesis either qualitatively or quantitatively. For example, what is the proportion of mRNA per ribosome, which mRNA has priority in binding, if any, and how is such priority expressed in steady state and particularly

256

S. Grisolla, E. Knecht and J. Hernández-Yago

in extreme conditions (for example under starvation, where certain enzymes may even increase while others disappear). As far as protein degradation is concerned, we know next to nothing regarding selectivity, except that it occurs and that environmental factors as well as the intrinsic properties of the proteins themselves are responsible for their stability.

Conclusion One of the main unsolved problems in biology, intimately related to the concepts of Schoenheimer, is that in spite of the rapid and extensive protein turnover, the protein concentration per cell volume changes but little. As indicated above, the average cell contains many enzymes, and it is well known that in response to a number of signals (e.g. hormones, high protein diets, starvation, etc.) the concentration of one or various enzymes may increase or decrease dramatically. Granted that many of those enzymes form a very small percentage of the total and also that increases such as those due to inducers (e.g. hormones) are generally restablished rapidly. Even in extreme conditions, e.g. starvation, the percentage of variation of protein per cell volume is small. Indeed, an example of the critical maintenance under such conditions is the fact that in many animal species under starvation there is a predominant and selective cell loss but at the same time an increase of certain enzymes in organs such as liver, which may lose a great deal of its normal weight. Certainly, the maintenance of osmotic pressure is essential for the cell and in man and other animals, among other mechanisms, the production of small molecular weight proteins, such as albumin, predominates, to aid in this critical task. In a model cell (Figure 2) we represent the fact that the amount of protein synthesis normally exceeds the amount needed to replace that degraded, and thus to maintain the necessary equilibrium. It should be made quite clear that the cell, or the organ, must have mechanisms which permit the synthesis of many proteins in an exceedingly well organized manner, most likely under genetic control, so that no large excess of subunits or of complementary proteins are made. Nevertheless, a certain surplus must be produced to insure health and under these circumstances, the proteins which do not find either a morphological or a chemically stabilizing environment, will be rapidly destroyed, as first suggested by Wheatley et al. (80). Obviously, part of this necessary excess in synthesis is due to the existence of leading peptides and recycling of membrane proteins. At any rate, the bulk of proteins will be stabilized, either by "mophological" means (e.g. entry of proteins into the mitochondria) and/or by the metabolic environment. Nevertheless, and occasionally, there is destabilization and the protein/s will find themselves at the mercy of lysosomes and/or other proteolytic systems as mentioned above. In this context, it should be remembered that the life of a protein, usually expressed in terms of halflife, in the few cases which have been measured, follows a first order decay, indicating a stochastic process and, thus, that no molecule has a definite half-life, but that it is the result of the aggregate of many individual molecules. Granted that the chemical characteristics, including size and especially conformation, will determine stability, but even these will be influenced by the environment!

Intracellular Protein Degradation: Past, Present and Future

257

STABILIZED MATURE PROTEINS! stabilized χ—^-destabilized" ¡jChemical Stabilization: Destabilization: e.g. "wrong" ^e.g. substrates, otheiik conditions, covalent modifications by Fivironmental conditions)^ chemotropic agents, etc. Morphological Stabilization: te.g. entry into mitochondria, etc

PROTEIN SYNTHESIS

O ® O

Non-lysosomal proteases

O O o O

Non-stabilized newly synthesized proteins S"short"-ived proteins)

Figure 2

PROTEIN DEGRADAT

Autophagy

Microautophagy

Stabilization and destabilization of intracellular proteins.

At the present time, the main areas to clarify in this field are: 1. to establish the relative importance, under "in vivo" conditions, of the several structural features and "marking" reactions in determining the degradation rate of proteins, 2. to identify non-autophagic systems of intracellular protein degradation, and 3. to determine the relative importance of the various proteolytic systems in degrading different intracellular proteins under physiological and pathological conditions. Another important question is: what advantage, if any, does the cell or an animal derive from having such a variety of proteins, in terms of their half-,lives? Finally, although from the above it is evident that the extent and velocity of degradation of proteins depends mainly on their stabilization, it also appears evident that protein turnover is mainly regulated at the protein synthesis level.

258

S. Grisolia, E. Knecht and J. Hernández-Yago

Acknowledgement Supported by FISS, Comité Conjunto Hispano-Norteamericano, International Molecular Cytology Program IIC-KUMC and CAICYT (grants # 84-0574 and 832386).

References 1. Munro, H.N. 1964. Historical Introduction: the Origin and Growth of our Present Concepts of Protein Metabolism. In: Mammalian Protein Metabolism (H.N. Munro and J. B. Allison, eds.). Academic Press, New York. Vol.1, pp. 1-29. 2. Schoenheimer, R. 1942. The Dynamic State of Body Constituents. Harvard University Press, Cambridge, Mass. 3. Simpson, M.W. 1953. J. Biol. Chem. 201, 143-152. 4. Grisolla, S. and W. Hood. 1972. Chemotropic Basis for, and Regulation of, Protein Turnover. In: Biochemical Regulatory Mechanisms in Eukaryotic Cells (E. Kun and S. Grisolia, eds.). Wiley, New York, pp. 138-203. 5. Goldberg, A.L. and A.C. St. John. 1976. Annu. Rev. Biochem. 45, 747-803. 6. Holzer, H. and P.C. Heinrich. 1980. Annu. Rev. Biochem. 49, 63-91. 7. Hershko, A. and A. Ciechanover. 1982. Annu. Rev. Biochem. 51, 335-364. 8. Mortimore, G.E. 1982. Nutr. Rev. 40, 1-12. 9. Segal, H.L. and A.C. Parks. 1982. Crit. Rev. Food Sci. Nutr. 16, 141-147. 10. Duque-Magalhâes, M.C. 1984. Biochimie 66, 653 - 662. 11. Grisolia, S. and D. Wheatley. 1984. Life Chemistry Rep. 2, 257-297. 12. Grisolia, S., J. Hernández-Yago and E. Knecht. 1985. Curr. Top. Cell. Regul. 27, 387-396. 13. Mayer, R.J. and F. Doherty. 1986. FEBS Lett. 198, 181-193. 14. Beynon, R.J. and J.S. Bond. 1986. Amer. J. Physiol. 251, C141-C152. 15. Pontremoli, S. and E. Melloni. 1986. Annu. Rev. Biochem. 55, 455-481. 16. Bond, J.S. and P.E. Butler. 1987. Annu. Rev. Biochem. 56, 333-364. 17. Seglen, P.O., B. Grinde and A.E. Solheim. 1979. Eur. J. Biochem. 95, 215-225. 18. Mortimore, G.E. and W.F. Ward. 1981. J. Biol. Chem. 256, 7659-7665. 19. Pfeifer, U. 1981. Acta Biol. Med. Germ. 40, 1619-1624. 20. Kominami, E., S. Hashida, A. Khairallah and N. Katunuma. 1983. J. Biol. Chem. 258, 6093-6100. 21. Henell, F. and H. Glaumann, 1984. Lab. Invest. 51, 46-56. 22. Glaumann, H., J.L.E. Ericsson and L. Marzella. 1981. Int. Rev. Cytol. 73, 149-182. 23. De Duve, C. and R. Wattiaux. 1966 Annu. Rev. Physiol. 28, 435-492. 24. Ericsson, J.L.E. 1969. Autophagy. In: Lysosomes in Biology and Pathology (J.T. Dingle and H.D. Fell, eds.). North Holland, Amsterdam. Vol.2, pp. 345-394. 25. Novikoff, A.B. 1973. Lysosomes: a Personal Account. In: Lysosomes and Storage Diseases (H.G. Hers and F. van Hoof, eds.). Academic Press, New York. pp. 2-41. 26. Grinde, Β. 1985. Experientia 41, 1089-1230. 27. Saito, T. and Κ. Ogawa. 1974. Acta Histochem. Cytochem. 7, 1-18. 28. Dean, R.T. 1975. Biochem. Biophys. Res. Commun. 67, 604-609. 29. Mortimore, G.E. and C.M. Schworer. 1977. Nature 270, 174-176. 30. Knecht, E. and J. Hernández. 1978. Cell Tissue Res. 193, 473-489. 31. Hernández-Yago, J., E. Knecht, A. Martínez-Ramón and S. Grisolia. 1980. Cell Tissue Res. 205, 303-309. 32. Marzella, L., J. Ahlberg and H. Glaumann. 1980. Exp. Cell Res. 129, 460- 466. 33. Kalnov, S.L., L.A. Novikova, A.S. Zubatov and V.N. Luzikov. 1979. Biochem. J. 182, 195-202. 34. Desautels, M. and A.L. Goldberg. 1982. J. Biol. Chem. 257, 11673-11679. 35. Rapoport, S., W. Dubiel and M. Müller. 1982. FEBS Lett. 147, 93-96. 36. Etlinger, J.D. and A.L. Goldberg. 1977. Proc. Natl. Acad. Sci. USA 74, 54- 58.

Intracellular Protein Degradation: Past, Present and Future 37. 38. 39. 40. 41.

259

De Martino, G.N. and A.L. Goldberg. 1979. J. Biol. Chem. 254, 3712-3715. Müller, M., W. Dubiel, J. Rathmann and S. Rapoport. 1980. Eur. J. Biochem. 109, 405-410. Ciechanover, Α., D. Finley and A. Varshavsky. 1984. Cell 37, 57-66. Hough, R. and M. Rechsteiner. 1984. Proc. Natl. Acad. Sci. USA 81, 90-94. Waxman, L., J.M. Fagan, Κ. Tanaka and A.L. Goldberg. 1985. J. Biol. Chem. 260, 11994-12000. 42. Schimke, R.T. 1973. Adv. Enzymol. 37, 135-187. 43. Waterlow, J.C., P.J. Garlick and D.J. Millward. 1978. Protein Turnover in Mammalian Tissues and in the Whole Body. North Holland, Amsterdam, pp. 158-164. 44. Dice, J.F. and A.L. Goldberg. 1975. Proc. Natl. Acad. Sci. USA 72, 3893-3897. 45. Segal, H.L., D.M. Rothstein and J.R. Winkler. 1976. Biochem. Biophys. Res. Commun. 73, 79-84. 46. Bachmair, Α., D. Finley and A. Varshavsky. 1986. Science 234, 179-186. 47. Rogers, S., R. Welles and M. Rechsteiner. 1986. Science 234, 364-368. 48. Dean, R.T. 1987. FEBS Lett. 220, 278-282. 49. Bohley, P., H. Kirschke, J. Langner, S. Ansorge, B. Wiederanders and H. Hanson. 1971. In: Tissue Proteinases (A.J. Barrett and J.T. Dingle, eds). North-Holland, Amsterdam, pp. 187-219. 50. Russell, S.M., R.J. Burgers and R.J. Mayer. 1982. Biochim. Biophys. Acta 714, 34-45. 51. Pohling, G., W. Schäfer, M. von Herrath and H. Holzer. 1985. Curr. Top. Cell. Regul. 27, 317-334. 52. Beeley, J.G. 1977. Biochem. Biophys. Res. Commun. 76, 1051-1055. 53. Levine, R.L., C.N. Oliver, R.M. Fulks and E.R. Stadtman. 1981. Proc. Natl. Acad. Sci. USA 78, 2120-2124. 54. Francis, G.L. and F.J. Ballard. 1980. Biochem. J. 186, 581-590. 55. Grisolla, S. 1964. Physiol. Rev. 44, 657-712. 56. Saus, J., J. Timoneda, J. Hernández-Yago and S. Grisolía. 1982. FEBS Lett. 143, 225-227. 57. Fried, V.A., H.T. Smith, E. Hildebrandt and K. Weiner. 1987. Proc. Natl. Acad. Sci. USA 84, 3685-3689. 58. Knecht, E., J. Hernández-Yago, A. Martínez-Ramón and S. Grisolía. 1980. Exp. Cell Res. 125, 191-199. 59. Grisolía, S., J. Timoneda, J. Hernández-Yago, J. Soler, M.D. de Arriaga and R. Wallace. 1981. Acta Biol. Med. Germ. 40, 1407-1418. 60. Hernández-Yago, J., E. Knecht, V. Felipo, V. Miralles and S. Grisolía. 1983. Biochem. Biophys. Res. Commun. 113, 199-204. 61. González-Bosch, C., V. Miralles, J. Hernández-Yago and S. Grisolía. 1987. Biochem. Biophys. Res. Commun. 146, 1318-1323. 62. Navarro, A. and S. Grisolía. 1985. Eur. J. Biochem. 149, 175-180. 63. Roche, E., E. Knecht and S. Grisolía. 1987. Biochem. Biophys. Res. Commun. 142, 680-687. 64. Knecht, E. and E. Roche. 1986. FEBS Lett. 206, 339-342. 65. Roche, E., F. Aniento, E. Knecht and S. Grisolía. 1987. FEBS Lett. 221, 231-235. 66. Nicoletti, M., C. Guerri and S. Grisolía, S. 1977. Eur. J. Biochem. 75, 583-592. 67. Wallace, R., E. Knecht and S. Grisolía. 1986. FEBS Lett. 208, 427-430. 68. Grisolía, E., E. Knecht, J. Hernández-Yago and R. Wallace. 1980. Turnover and Degradation of Mitochondria and their Proteins. In: Protein Degradation in Health and Disease (Ciba Found. Symp. 75). Excerpta Medica, Amsterdam, pp. 167-188. 69. Knecht, E., J. Hernández, R. Wallace and S. Grisolía. 1979. J. Histochem. Cytochem. 27, 975-981. 70. Knecht, E., J. Hernández-Yago and S. Grisolía. 1984. Histochemistry 80, 359-362. 71. Knecht, Ε., A. Martínez-Ramón and S. Grisolía. 1986. J. Histochem. Cytochem. 34, 913-922. 72. Martínez-Ramón. Α., E. Knecht and S. Grisolía, 1987. J. Histochem. Cytochem. 35, 897-907. 73. Lamers, W., J.W. Gaasbeek-Janzen, Α.F.M. Moorman, R. Charles, E. Knecht, A. MartínezRamón and S. Grisolía. 1986. J. Histochem. Cytochem. 36, 41-48. 74. Vargas, J.L., E. O'Connor, E. Roche, E. Knecht and S. Grisolía. 1987. Biochem. Biophys. Res. Commun. 147, 535-541. 75. Knecht, E., J. Hernández-Yago and S. Grisolía. 1984. Exp. Cell Res. 154, 224-232. 76. Knecht, E., J. Hernández-Yago and S. Grisolía. 1982. FEBS Lett. 150, 473-476.

260

S. Grisolía, E. Knecht and J. Hernández-Yago

77. Knecht, E., E. Roche, J. Hernández-Yago, J.L. Vargas and S. Grisolía. 1986. Biomed. Biochim. Acta 45, 1575-1583. 78. Felipo, V. and S. Grisolla. 1987. FEBS Lett. 210, 173-176. 79. Mäher, P.A. and S.J. Singer. 1986. Proc. Natl. Acad. Sci. USA 83, 9001-9005. 80. Wheatley, D.N., M.R. Giddings and M.S. Inglis. 1980. Cell Biol. Int. Rep. 4, 1081-1090.

Lipmann's Squiggle and the Unification of Cellular Structure and Function Heinz Herrmann and Anne L. Hiskes

During the early part of Lipmann's career the molecular, functional, and structural properties of cells were investigated in separate disciplines (biochemistry, cell physiology, histology) as isolated compartments that did not and could not communicate with each other. The advances of molecular biology have transformed the three disciplines into one broad area within which we can move freely from the purely molecular to the structural and functional aspects of cells without awareness of any boundaries. Indeed, we are taking this apparent continuity for granted. At this particular occasion we might pause and ask how this decisive change has come about, what its sources are, what its deeper meaning may be and, last but not least, how Lipmann's work has contributed to this transition. As introduction we will describe the pivotal role of our understanding of cellular phosphorylations in this fundamental reorientation. This will provide, at the same time, an opportunity to view Lipmann's work in this context. The main forms of phosphorylation that can be regarded as mediators between cellular structure and function are summarized in Table 1.

The establishment of the phosphate bond as a main source of cellular energy It may be appropriate to begin by recalling once more the commonly recognized phases in the development of biochemical studies in modern times. It started as the organic chemistry of biological substances, that is the establishment of their chemical characteristics without regard for their involvement in cellular processes. This was followed by investigations of the modifications and conversions of biological materials by cellular enzymes. Test tube experiments with cell extracts led to the identification of the individual steps of metabolic transformations which, eventually, could be put together in a coherent scheme of the main metabolic pathways within the cell. A substantial part of the studies in this phase of explorations dealt with the anaerobic and oxidative conversion of glucose into phosphorylated metabolic intermediaris. It is noteworthy that this line of studies began with the question of the mechanism of muscle contraction, in particular, of how the energy produced by the catabolism of glucose was eventually utilized in the structural changes that accompany the contraction of muscle (1,23,31). This work led to the recognition that phosphorylation not only facilitates the conversion of metabolic intermediaries but that ATP, formed in the course of glucose catabolism, becomes a potential carrier of metabolic energy as summarized in Lipmann's classic paper in the Advances in Enzymology (47). In the excitement of

Heinz Herrmann and Anne L. Hiskes

262

Table 1 The main lines of the experimental unfolding of the concept of the high energy phosphate bond and its impact on the understanding and unification of cellular structure and function. Organic chemistry of biological substances about 1850-1920

Protein Kinases Phosphorylation of proteins with concomitant conformational and functional change Cytoskeletal phosphorylations Membrane phosphorylations Receptor phosphorylation and signal transduction Phosphorylation during endocytic clathrin assembly-disassembly Nuclear phosphorylations^ DNA replication and transcription Phosphorylations in translation

Characteristics of metabolic intermediaries and metabolic energetics 1920-1941 Lipmann's paper on high energy phosphate bonds, Lipmann's squiggle 1941 I Myosin ATPase activity Engelhardt-Ljubimova Needham - Szent-Györgyi 1939-1942 I Development of the sliding filament model Huxley, Hanson 1953-1969 i Elaboration of the sliding filament model 1969 to the present I Conceptual implications

ATPase dependent conformational and functional changes, Ca 2 + Mg 2 + ATPase and ion transport in general

The Aristotelian unification of biological structure and function Post-Renaissance approaches to the relationship between biological structure and function The contemporary unification of biological structure and function through molecular concepts

the rapidly progressing investigations of the intermediary metabolism of carbohydrates and, in particular, o f the various forms of phosporylation investigators lost sight of the structural aspects of muscle contraction. A l t h o u g h Lipmann's 1941 article describes m a n y metabolic uses o f phosphate b o n d s as energy carriers their use for the structural changes in muscle contraction is mentioned only in passing. This is documented by quoting the last lines of a half page summary about utilization of phosphate b o n d energy in muscle contraction (47). "It is an o b v i o u s deduction that the energy-rich phosphate should link up to the contracting protein. There is, however, n o experimental indication even to encourage such a view. A n opening for further experimental approaches might be found in Mirsky's work o n myosin coagulation in muscle rigor."

Lipmann's Squiggle and the Unification of Cellular Structure and Function

263

Myosin ATPase and muscle contraction One can see that at this point the concept of the conversion of metabolic energy into structural change, such as muscular contraction had'hardly changed since the lactic acid coagulation model. The minds of those who had so brilliantly carried the advance to this point did not emancipate themselves from the experimentation with systems that were devoid of cellular structure. It took a fresh approach to bridge the gap. Such a bridge was in the making while Lipmann wrote his famous 1941 paper. It started with the publications by Engelhardt and Ljubimova (19,20), SzentGyörgyi (82,83), and Needham (17,57,58) and their associates. Myosin was found to bind high energy phosphate in the form of ATP, to have the enzymatic properties of an ATPase and to represent at the same time the structural cell unit in which the liberated energy produces a conformational change that is essential for muscle contraction. Demonstrating the role of ATP in the structural modification of myosin this work ushered in the third phase of biochemical studies in which molecular events are taken out of the test tube and are placed into the structural context of the cell. This could be regarded as the origin of molecular biology 10-15 years before the advent of the DNA double helix. The eventual convergence of the investigations of the metabolic and structural aspects of phosphorylations was documented in a collection of papers assembled under the editorship of Kalckar (37). Parenthetically, the shift from the metabolic to the conformational, structural frame of reference could be seen as a typical case of a replacement of paradigms that is thought to characterize scientific revolutions (43). The significance of the ATPase activity of myosin for a unified view of muscle structure and function was not immediately recognized. Further progress depended on advances in the resolution of muscle fine structure that was attained only more than a decade after the initial observations of myosin ATPase activity. In this period the results of x-ray diffraction and electron microscopy were coordinated with measurements on contracting muscle using physiological methodology and muscle contraction could be interpreted as the result of interactions between actin and myosin in the form of the sliding filament model (34). A particular advantage of muscle for this type of analysis was the large quantity and the distinctiveness of the two major proteins in the muscle fiber and the amplification of the molecular organization that could be clearly recognized on the cytological and physiological levels. Therefore, muscle contraction was initially regarded as a special case of an involvement of phosphorylation in functional and structural aspects of cell maintenance. However, soon it was realized that the role of phosphorylations in linking cellular structures and functions is much more general, as indicated in the table. It is now well established that high phosphate bond energy can be released in two different reactions. In one reaction the conversion of ATP to ADP changes the conformation of the ATP binding ATPase protein. The conformation change can be transmitted from one domain in the same protein to another or even to another protein. In a different mechanism the inorganic phosphate released in the ATP-ADP conversion is covalently linked to a serine, threonine, or tyrosine residue in another protein with a concomitant change in the conformation of the

264

Heinz Herrmann and Anne L. Hiskes

receptor protein. This type of phosphorylation is catalyzed by protein kinases. A brief annotation summarizes some representative examples for such phosphorylations in diverse cell systems.

The ion transport related ATPases The ATP-ADP conversion is an energy source not only in muscle contraction but also in several systems for translocation of ions through membranes (Table 1). An example is the role of phosphorylation in the structural modification of the sarcoplasmic Ca 2 + ion transport system. It is a complex protein with three globular domains one of which has an ATPase activity and an adjoining domain with an aspartate residue that is phosphorylated by the ATPase portion of the complex. The binding of ATP to the ATPase subunit promotes the phosphorylation of the phosphate accepting domain. This induces a conformational change in the Ca binding domain that allows translocation of Ca 2 + through a helical transmembrane portion of the protein (35,49,67). The N a + K + dependent ATPase is another ion transferring system in which conformational change is the basis of ion transport that depends on ATP-ADP conversion (84). Similar processes are involved in proton translocation as part of the electron transport system. It should be pointed out that phosphorylation is not a general requirement for ion transport. For example, the anion transport through Band 3 protein does not seem to involve phosphorylation with a concomitant conformational change of an ATPase-type enzyme although binding of glycolytic enzymes to the cytoplasmic loops of the ATPase may be a phosphorylation-dependent interaction (36).

Modifications of the cytoskeletal elements by phosphorylation The alternative mechanism for the utilization of ATP by protein kinases became apparent as early as 1955 with the observation of a phosphorylation of glycogen Phosphorylase (22,42) leading to the recognition of phosphorylation of several enzymes of the intermediary carbohydrate metabolism by enzymes called protein kinases. It has been demonstrated only in the last decade that a wide range of structural proteins and of non-enzymatic cell components are phosphorylated by kinases and that protein kinases themselves are substrates for autophosphorylations. Proteins in the cytoplasm, the plasma membrane and the cell nucleus are phosphorylated in this way. Starting with phosphorylations of cytoskeletal components the ubiquity of these protein kinase dependent modifications is strikingly demonstrated. The changes in the state of microtubules and microfilaments, their increase and reduction in length and their interactions with each other and with other cytoskeletal elements and cell membranes are essential for many cell functions (54,64). The state of microtubules is regulated by a cascade of controls to achieve the required modifications.

Lipmann's Squiggle and the Unification of Cellular Structure and Function

265

To begin, ATP and GTP and their hydrolysis are involved in complex kinetics of the addition and the release of the protein subunits from actin filaments and microtubules during their differential assembly and of disassembly (for a recent reconsideration see Farrell (21)). The state of microtubules is modified by association with several groups of proteins (Microtubule associate proteins, MAPs). These proteins enhance the polymerization of the tubulin and the linking to other cell components such as microfilaments. The regulatory effectiveness of several of the MAPs is decreased by their phosphorylation (79). The levels and the sites of phosphorylation are determined by protein kinases and the removal of phosphates from the proteins by phosphatases (29,30,55,86,88). Some of these kinases are cAMP dependent, others are calcium and calmodulin dependent (78). The polymerization promoting effect of the MAPs is balanced by the direct depolymerizing effect of calcium and calmodulin on the tubulin molecules themselves. Recent investigations have demonstrated a remarkable role of microtubular organization and phosphorylation in membrane vesicle transport. The movement of microtubule associated vesicles can be observed within the cell and in in vitro preparations. Under in vitro conditions purified microtubules assembled from purified tubulin molecules do not show the transport phenomenon unless a soluble cell component and ATP are added. The soluble component has apparently ATPase activity that generates the energy for a conformational change that propels the particles along the microtubular tracks (1,25,46,65,74,75,76,87). Similar to the microtubules, the polymerization and the steady state length of the second main component of the cytoskeletal system, the microfilaments, are regulated by the kinetics of the ATP-ADP conversion that are associated with the actin subunits of these structures (9,10). In addition to the cross-linking with microtubules in maintaining the overall cytoskeletal organization, the microfilaments have a major role in the association of the cytoskeleton with the cell membrane (64). Several actin associated molecules may be involved in the actinmembrane association. One of these is vinculin, a protein kinase-phosphorylated protein that also is part of an elaborate system for the attachment of myofibrils to the sarcoplasmic membrane. One of the kinases that phosphorylate vinculin is activated in the course of the neoplastic cell transformation and may be part of the mechanism that produces the change in cell shape in this process. In the villi of the brushborder of intestinal cells a core of actin filaments connect with the cell plasma membrane by 15-30 nm side arms that protrude at 300 nm intervals from the actin filament in a helical array. These arms are removed by ATP probably due to the presence of a 110 kDa protein with ATPase activity (54). Similar to microtubules and microfilament systems, some of the proteins associated with the third class of cytoskeletal elements, the intermediate filaments are phosphorylated (62). One of them, plectin, is phosphorylated at two different sites by a cAMP dependent and independent kinase respectively. The phosphorylation of vimentin by a cAMP dependent protein kinase is hormonally regulated through activation by norepinephrine (7).

266

Heinz Herrmann and Anne L. Hiskes

A network of supporting proteins, a membrane associated cytoskeleton, is essential in maintaining the state of the erythrocyte plasma membrane (50). One of the main conditions for the structural changes and the coordinated functions of the system is the level of phosphorylation of spectrin, one of the main components of this system. Spectrin is phosphorylated by a cAMP dependent and by a cAMP independent kinases (59). The phosphorylation of spectrin modifies the interaction of spectrin and actin, another component of the cytoskeletal network of the erythrocyte. Spectrin has been considered as a cytoskeleton component that exerts a restraining effect on the molecules in the membrane bilayer and may play a role in the maintenance of lipid asymmetry that was mentioned earlier (80,89). Whether phosphorylation is involved in the shape change of the erythrocyte has remained controversial. So far, our survey has given examples of filamentous cytoplasmic structures, their state of polymerization, and their interactions controlled by phosphorylations. These molecular modifications lead to electron or light microscopically recognizable structural changes and their functional expression as in muscle contraction, in various forms of cell motility, or the translocations of membrane units in the cell interior. These examples of structure-function related phosphorylations give, to varying degrees, information about the mechanisms of these processes or suggest possibilities for establishing a conceptual continuity between structure and function as demonstrated more fully for muscle contraction and discussed later on in more general terms. Inclusion of these and further examples has the purpose of demonstrating the remarkable ubiquity and diversity of the systems in which phosphorylations are linked to structure-function modifications and to the progress towards structure-function unification. In this sense we will continue with the discussion of structure-function phosphorylations in plasma membranes and in the cell nucleus.

Phosphorylation and the dynamics of the plasma membrane The plasma membrane is a dynamic system with endocytic, pinocytotic or phagocytotic internalization and replacement taking place in a continuous process. Most of the endocytic internalization occurs in distinctive membrane areas, the coated pits, in which the cell surface layer is associated with a cage like assembly, comprising clathrin molecules of 180 kDa together with smaller molecules (30 and 32 kDa), the light chains (16). After internalization the clathrin cage is dissociated and the clathrin molecules are released. This disassembly requires energy that is generated by a 70 kDa uncoating ATPase (12). Binding of this factor, cleavage of ATP, and clathrin cage disassembly occur only if both the heavy and the light cage components are present. Cages depleted of the light chains do not induce ATP hydrolysis and do not dissociate (71,73). Apparently, two sites are involved in the binding of the ATPase (72). The binding to one of these occurs only in the presence of light cage protein fraction. These data demonstrate the involvement of ATP bond energy in one of the basic membrane mechanisms (see 69).

Lipmann's Squiggle and the Unification of Cellular Structure and Function

267

Protein kinase dependent changes of membrane receptors A second major task of the plasma membrane is the transduction of external signals into the interior of the cell. A number of transmembrane proteins are receptors that interact with the external molecular messengers such as hormones, neurotransmitters and growth factors. This elicits a conformational change in the exterior domain of the receptor protein that is transmitted to its cytoplasmic domain where it triggers intracellular processes. Several examples will indicate the involvement of phosphorylations in the transductions. One of the well analyzed signal transducing complexes is the receptor for acetylcholine that is released from the synaptic nerve endings of the cholinergic portion of the nervous system. The receptor is a large protein complex (270 kDa) that consists of five subunits (2a, 1/?, \y, 1 Lys)

spoOB, spoOD, spoOE, spoOF, and spoOH mutants

Spo" ->• Spo +

spoOB, spoOE, and spoOF mutants

P 2 8 transcripts:

wild-type cells

Spo~ in 0.7 M EtOH ->· Spo + in 0.7 M EtOH

• +

Spo" in high Glu -»• Spo + in high Glu spoOJ and spoOK mutants

no effect

Tabulated from Sharrock et al. (1984) Mol. Gen. Genet. 194, 260; Leung et al. (1985) Mol. Gen. Genet. 201, 96; Gilman and Chamberlin (1983) Cell 35, 285; and Yamashita et al. (1986) Mol. Gen. Genet. 205, 28.

In fact, the intergenic suppressor mutations, rvtA, studied by Leighton's group (32) and sof-l by Hoch, Saito and their associates (33,34), correct sporulation defects in the spoOB, spoOD, spoOE, spoOF and spoOH mutants, as shown in Table 2. The rvtA and sof-l mutations have been identified to be the same mutation in the spoOA structural gene. The twelfth amino acid residue, asparagine, was replaced by lysine due to the missense mutation resulting in the substitution of thymidylic acid by guanylic acid (34). It is tempting to think that this structurally modified spoOA protein can circumvent the functions of the spoOB, SpoOE and spoOF gene products for sporulation by performing not only the role of the spoOA protein, but also the functions of the spoOB, spoOE and spoOF proteins. However, alternatively the sof-l protein could have been modified in such a way that the spoOA protein would have been activated by the interaction with the spoOB, spoOE and spoOF gene products in the wild type cells for the initiation of sporulation.

From Phosphoenolpyruvate Carboxykinase to Sporulation

345

A further search by Hoch and his associates (28) for the homology of the spoOA protein with the known protein sequences revealed that it is highly homologous to the ompR protein of E. coli. This protein controls the gross permeability properties of the cell. It is also homologous to the E. coli dye or sfrA protein (35), of which mutation results in loss of the sex factor F expression, and an increase of sensitivity to dyes and antibiotics due to the aberration of envelope protein synthesis. Koshland and his associates (36) sequenced the cheY gene of Salmonella typhimurium, one of the Chemotaxis genes and found its product is highly homologous to the product of another Chemotaxis gene cheB and also to the ompR, dye and spoOA proteins. The sequence homology between the K. pneumoniae ntrC product, E. coli phoΒ product and spoOA product is also noted and discussed by Nixon et al. (37) and Drummond et al. (38). All the above proteins which are homologous to the spoOA product are membrane proteins and are involved in the transduction of environmental signals. In 1976, Schaeffer and his associates (39) proposed that an alteration of the cytoplasmic membrane in early-blocked spoO mutants was responsible for the pleiotropic effects of these mutants on the basis of their studies on nitrate reductase and inhibition of sporulation by ethanol. This sensitivity of sporulation to alcohol is also suppressed by the intergenic suppressor sof-l or rvtA, which restores sporulation of the spoOB, spoOE and spoOF mutants (32) (cf. Table 2). All these findings, together with the fact that these spoO gene products are present in vegetative cells, are in accordance with the notion that the spoOA protein is a protein which senses the changes of the environmental conditions and exerts its effect on expression of other sporulation-related genes. At present, however, it seems difficult to delineate the functions of the spoOA protein to account for all the current information. It was mentioned above that the spoOA product might function in the initiation of sporulation by interacting with RNA polymerase and allowing it to transcribe spore-specific genes analogous to sigma-factors. Ferrari et al. (27) also postulated that the spoOA protein interacts directly with RNA polymerase and even suggested that it can be σ 2 8 itself, which is a minor form of sigma-factors involved in highly specific transcription from a small number of promoter sites in the B. subtilis genome discovered by Chamberlin and his associates (40). Bacillus subtilis RNA polymerase exists in at least seven holoenzyme forms, as shown in Table 3, in addition to three phage-specific ones, each associated with a unique sigma-factor that confers a characteristic promoter specificity on the "core" RNA polymerase (12). The core enzyme consists of four subunits, a2ßß', just like the counterpart of the E. coli enzyme, σ 4 3 is the major sigma-factor present and active in vegetatively growing B. subtilis cells. Previously, it was called σ 55 , but Doi and his colleagues (41) have recently cloned and sequenced the gene, and the molecular weight of σ 4 3 was deduced to be 42,828. It has a strong sequence homology with the major sigmafactor of E. coli, σ 70 , and recognizes the same consensus DNA sequences, TTGACA at — 35 region and TATAAT at — 10 region (12). A part of the holoenzyme with σ 4 3 is considered to be active during sporulation, although Losick and his associates favor the idea that σ 4 3 is dissociated or inhibited in its activity in sporulating cells (cf. 11). σ 3 7 and σ 3 2 are minor sigma-factors discovered by Losick and his collaborators (42,43). Both of these sigma-factors are present in vegetative cells but are used for

346

Kiyoshi Kurahashi, Toshihiko Ikeuchi and Jun Kudoh

Table 3

Sporulation-related B. subtilis sigma factors and their promoter recognition sequence

Sigma Gene factor

Presence and function

Consensus D N A sequence -35

- 1 0

Promoters

rpoD

veg. and spo. cells (major sigma in cell)

TTGACA

TATAAT

σ43Ρ2, spoVG, aprA

sigB or rpoF

veg. cells; active in sporulation, but not required for sporulation *

AGNNTT

GGNATTNTT

aprA, spo VGP¡ ctc, σ 4 3 Ρ 3

veg. cells; active in sporulation

AAATC

TANTGNTTNTA

spoVGP2, ctc

spoOH

veg. cells; active in sporulation

spoIIG

induced during sporulation

TTNAAA

CATATT

veg. cells; active in veg. cells

none

CCGATAT

spoil AC

spoVG L promoter, ctc, aprA

T 3 _5 cells

From Chamberlin etal. (1986) in Welch Found. Conf. on Chem. Res. XXIX; and Doi and Wang (1986) Microbiol. Rev. 50, 227.

transcription of sporulation genes, as revealed by the extensive studies of Losick's group on the expression of cloned spoVG, spoVC and vegetative ctc genes (cf. 11). Apparently, all promoters of these genes are utilized by the holoenzymes with σ 3 7 or σ 32 . The gene for σ 3 7 , called s ig Β or rpoF, was mapped at 40° on the genetic map and its sequence was determined (44,45). The molecular weight of σ 3 7 was deduced to be 30,143 (45). Duncan et al. (45) reported that the null mutation of the rpoF gene had no effect on cell growth and sporulation in rich medium. σ 3 0 was discovered by Carter and Moran (46) very recently as a possible product of the spoOH gene. It has a role in spo VG transcription. It is present in the vegetative and sporulating phase of spo+ cells, but not in the spoOA and spoOB mutant cells. A mutation at abrB restores σ 3 0 to RNA polymerase in the spoOA strain. They presented two alternative models to explain the effects of spoOA and abrB mutations on the regulation of σ 3 0 activity: (1) spoOA positively regulates σ 3 0 synthesis and abrB may repress transcription of the σ 3 0 structural gene and spoOA may antagonize this repressor; or (2) spoOA may be required for the processing of σ 3 0 to an active form, and abrB and spoOA may affect expression of the processing enzyme. Dubnau et al. (47) constructed lacZ fusions in frame with the spoOH gene and studied its expression under various growth conditions and in various genetic backgrounds. They showed that expression of the spoOH gene was not repressed by glucose, but was induced by decoyinine, which induces sporulation in the presence of catabolites, as discovered by Freese and co-workers (48).

From Phosphoenolpyruvate Carboxykinase to Sporulation

347

σ 2 9 was discovered by Haldenwang, Lang and Losick (49). It is induced during sporulation, most abundant from t 2 to t 4 during differentiation and serves only transcription of sporulation genes (50). It is not essential, however, for the events before stage III (50). Szulmajster, Haldenwang and their colleagues identified it as a product of the spoIIG gene and found a high sequence homology with E. coli and B. subtilis major sigma-factors (51,52). It transcribes the spoIID gene. In this connection, it is interesting to know that Mandelstam, Piggot and their associates sequenced the spoil AC gene and found very high homologies of amino acid sequences between the predicted spoIIAC protein and σ29, E. coli σ 7 0 , and the heat shock regulatory gene product (54-56). Doi and Wang suggested that the product may be σ 2 2 which Fukuda et al. (57) isolated in 1975 (cf. 12). We may not be surprised if the products of some additional sporulation genes turn out to be minor sigma-factors. σ 2 8 was discovered by Chamberlin and his associates, present in vegetative cells and active in those cells, but there are only 20 to 30 promoters (P 28 ) regognized by σ 2 8 in the entire B. subtilis chromosomes (40,58). The transcription from promoters 28 decreases rapidly when cells enter sporulation. The level of σ 2 8 RNA polymerase in the spoOA mutant is the same as in the wild-type cells, but it was found that the transcription depends on the presence of the wild-type spoO genes, such as spoOA, spoOB, spoOE and spoOF. The second site suppressor mutation, rvtA, the mutation in the spoOA gene, as described above, restored the transcriptional function of the σ 2 8 RNA polymerase in the spoOB, spoOE and spoOF mutants (cf. Table 2). Since the rvtA mutation restored σ 2 8 function as well as the ability to sporulate of spoOB, spoOE and spoOF mutants, Chamberlin suggested that the σ 2 8 function might be involved in an early stage of the establishment or triggering of sporulation (59). The above results indicate that σ 2 8 is not the product of the spoOA gene, but may suggest that the product of the spoOA gene acts directly to modulate the activity of σ 2 8 . The simplest interpretation of these results seems to be as follows: The products of the spoOB, spoOE and spoOF genes are required to activate the product of the spoOA gene, but when a mutation occurs in the spoOA gene and the sof-l protein is produced, it is already an active form and can regulate the expression of the sporulation and sporulation-associated genes (including P 2 8 transcription) positively without further interaction with those spoO gene products. However, the analysis of the deduced primary structure of the spoOA protein by the method of Chou and Fasman did not predict the helix-turn-helix DNA binding motif seen in various positive or negative regulator proteins of transcription. An important problem of the initial event of sporulation that needs to be solved is how the asymmetric septation takes place in response to the deprivation of nutrients. Freese (60) states that whereas the asymmetric septation usually depends on a nutritional signal, forespore development depends only on the genetic constitution of the organism. He also assumes that there is ultimately one compound of low molecular weight which, in combination with a protein, suppresses sporulation. As long as there are enough nutrients, the cells continue to divide symmetrically. However, Freese and his associates discovered that decoyinine, a specific inhibitor of guanosine monophosphate synthesis, can induce sporulation dramatically in the presence of glucose or glycerol (48). They carefully

348 Table 4

Kiyoshi Kurahashi, Toshihiko Ikeuchi and Jun Kudoh Comparison of cellular responses under different sporulation conditions Sporulation initiated by

Condition End of growth in nutrient medium Decrease of GTP Sporulation of mutants blocked in citrate cycle or in gluconeogenesis Sporulation without Mn 2 + Induction of certain catabolite repressible enzyme Production of alkaline phosphatase »

+b — -

+ +

Direct deprivation of guanosine nucleotide synthesis"

+ + + -

+

a

Using decoyinine or a guanine auxotroph. + , the event occurs; —, the event does not occur. (From E. Freese. In: Sporulation and Germination (60)). b

analyzed the intracellular levels of various nucleotides (61). As shown in Table 4, when the cells sporulated at the end of vegetative growth with the exhaustion of nutrients or by deprivation of guanosine nucleotide synthesis, a decrease of intracellular GTP was observed (60). Manganese ion had long been considered essential for sporulation, but decoyinine induced sporulation in the absence of manganese ion. The work by Freese and his associates (62) indicates that the nutritional deprivation induces the decrease of intracellular GTP and GDP concentration that directs the cells towards asymmetric cell division, although the mechanism itself is unknown. Another way that the cells devised to overcome the suppressive effect of catabolites on sporulation is the crs mutation, a catabolite resistant mutation. Takahashi and Sun (63,64) isolated a number of such mutants. Recently, it was found by Doi, Leighton and their collaborators (65,66) that the crs A mutation among them is not only catabolite resistant for sporulation, but also extragenic suppressors of stage 0 sporulation genes, such as spoOB, spoOD, spoOE, spoOF, spoOJ and spoOK, as shown in Table 5. In addition, the mutation lies in the rpoD gene, the σ 4 3 coding sequence where the proline residue at 290 position was replaced by phenylalanine (67). These findings suggest that the major sigma-factor of B. subtilis, which has been considered to be involved only in transcription of vegetative functions, is also implicated in the initiation of sporulation. It can be conceived naively again that the altered σ 4 3 protein can bypass the functions of those spoO gene products and regulate positively the sporulation and sporulation-associated genes even in the presence of catabolites. Strains bearing the rvtA or sof-1 mutation, a spoOA allele, are also resistant to catabolite repression (cf. 34,68). All these findings suggest a close link between the spoOA gene product and σ 4 3 in their functions, although no amino acid sequence homologies were found between the two. Recently, Yamashita et al. (69) studied the effect of spoO mutations on the expression of spoOA- and spoOF-lacZ fusions in B. subtilis. They found that

From Phosphoenolpyruvate Carboxykinase to Sporulation Table 5

349

Properties of crsA intergenic suppressor mutation

Designation crsA (catabolite resistance for sporulation)

Mutation 43

in rpoD (σ ) gene CCT - TTT (Pro-290 ^ Phe)

Introduced into

Phenotypic change

wild-type cells

Spo~ in high Glu -> S p o + in high Glu Spo" in 0.7 M EtOH S p o + in 0.7 M EtOH

spoOB, spoOD, spoOE, spoOF, and spoOJ mutants

Spo" Spo + (some, partial)

spoOK mutant

Spo" Spo + and catabolite sensitive (Spo" in high Glu)

-

Tabulated from Kawamura et al. (1985) Proc. Natl. Acad. Sci. U S A 82, 8124; and Leung et al. (1985) Mol. Gen. Genet. 201, 96

expression of the spoOA-lacZ fusion was increased transiently at the end of the exponential phase of growth. The spoOF-driven /?-galactosidase synthesis was also induced at the late exponential phase of growth and continued to increase until the mid-sporulation phase. The expression of both spoOA- and spoOF-lacZ fusions was impaired in spoOB, spoOD, spoOE, spoOF and spoOH mutant cells, but not in spoOJ and spoOK mutants. The additional sof-l mutation restored the impaired expression in those mutants. Marahiel et al. (70) reported that the expression of a fusion gene of the Bacillus brevis tyrocidine structural gene (tycA) and the lacZ gene of E. coli in B. sub tills is under the control of the spoOA, spoOB and spoOE genes but not the spoOC, spoOF and spoOH genes and that its dependence of the expression on the spoOA gene product can be entirely bypassed by an abrB suppressor mutation. The introduction of this suppressor mutation caused tycA-lacZ to be transcribed constitutively even during the vegetative phase of growth. In a similar experiment by Zuber and Losick (71) for the expression of the spoVG-lacZ fusion gene, however, the synthesis of the fusion protein was found to be still under the control of catabolites. In order to account for all the known observations, we propose the following hypothetical schemes for the induction of sporulation and sporulation-associated events (see Figure 5). 1. The spoOH gene which is repressed by an active repressor complexed with GTP during the vegetative growth of the cells is transcribed by Εσ 43 upon dissociation of the repressor from the operator when the cellular GTP level is decreased. 2. The product of the spoOH gene, σ 30 , is first modified by a sporulation-associated protease to σ 30 '. We consider basal activity of the protease is enough to initiate the activation of σ 3 0 . σ 3 0 is in turn utilized to promote the transcription of spoO genes, for example spoOD and spoOE genes. We propose that these genes are

350

Kiyoshi Kurahashi, Toshihiko Ikeuchi and Jun Kudoh

From Phosphoenolpyruvate Carboxykinase to Sporulation

351

structural or regulatory genes responsible for the synthesis of an unknown catabolite factor equivalent to cAMP in E. coli. This factor is visualized as a factor, of which the level in the cell is raised upon depletion of catabolites. 3. Since the spoOB gene is expressed during vegetative growth of the cells, we are inclined to assign the role of a CAP (catabolite activator protein)-like protein to the spoOB product, SpoOB. 4. The spoOF gene is repressed during the vegetative growth, but upon exhaustion of nutrients and a decrease of the GTP level, the catabolite factor and SpoOB in conjunction may enhance the removal of the repressor from the spoOF promoter to let the spoOF gene be transcribed by Εσ 43 . 5. The product SpoOF now activates the product of the spoOA gene, SpoOA to SpoOA'. The spoOA gene is transcribed by Εσ 43 during the vegetative growth phase, but the transcription is enhanced at the end of the logarithmic phase of growth by SpoOB and the catabolite factor. 6. The SpoOA' then enhances the transcription of the septation genes by E • «o c HC-R I NH •* 0 C 1 HC -R M

To• c HC-R'" I CO NHe

STAGE 4

Figure 2 Fritz Lipmann's early model of a protein-template directing the formation of polypeptides. It does not connect however, the DNA-template with the polyenzyme structure (from his essay On the mechanism of some ATP-linked reactions and certain aspects of protein biosynthesis in The Mechanism of Enzyme Action, edited by W.D. McElroy and Bentley Glass in 1954).

mRNAs. The peptide antibiotics were in the right size to handle and sequence. A RNA-fraction of the gramicidin S-producer was characterized, directing the formation of the peptide with a ribosomal fraction of the Dubos-strain. That these data have been placed in JMB and Biochemistry illustrate quite well the readiness at that time to accept a mRNA template for D-residues, ornithine, and cyclic structures 5 .

The Competition The unravelling of the fundamental enzymic mechanism of nonribosomal peptide biosynthesis, a major late event in the Lipmann's lab history, has been speeded up an exciting competition mainly between Osaka, Oslo and New York, the results being rapidly published in PNAS, FEBS Letters, or the Journal of Biochemistry. While Trine-Lise Berg (later Zimmer) was the first to develop a cell-free system for gramicidin S (5), Kiyoshi Kurahashi had already detected the cyclodipeptide cDPhe-Pro (6), and in a key-paper had measured the ATP consumption for each s

John B. Hall and Theodore Winnick commented these data in a restudy (Biochemistry 5, 3844 (1966)) with the remarkable sentences "The mass of data accumulated in this laboratory over a period of several years in support of the ribosomal pathway seems too extensive and selfconsistent to be explained as the result of simple technical error or misinterpretation of experimental results. However, the fact remains that despite our best efforts, we have not been able to reproduce these experiments.

358

Hans von Döhren and Horst Kleinkauf

Figure 3

Birthday at Cold Spring Harbor 1969 (8).

peptide bond to be one for tyrocidine biosynthesis 6 . Another major achievement from his lab was the fractionation of the gramicidin S synthetase into two complementary fractions (7), which has not been a simple task in these early days of protein chromatography. When he met the Lipmann people at Cold Spring Harbor in 1969 (Figure 3), they had cought up to him within 2 years, combining his results and those of the Oslo group with the Rockefeller experiences on amino acid incorporation into proteins. Lipmann had speculated as early as 1941 on the squiggle phosphate in amino acid activation, and the participation of aminoacylthiolesters was proposed in 19527. Then starting in 1968, these ideas came active.

6

7

K. Fujikawa, T. Suzuki, and K. Kurahashi, BBA 161, 232 (1966). This measure of one squiggle per peptide bond represents the expected value for carboxyl activation as known from acetyl-CoA formation and charging of t R N A . So if no additional energy is required, the thiotemplate is as e.g. glutathione biosynthesis much more effective than protein biosynthesis. Recently Bose's group studying the mycobacillin multienzymes arrived at two moles ATP for each peptide bond in the phosphate template type of synthesis. One additional squiggle has been postulated for conformational changes facilitating elongation (S.K. Ghosh, S. Majumder, N . K . Mukhopadhyay, and S.K. Bose, Biochem. J. 240, 265 (1986). See Theodor Wieland's chapter in this volume.

Research on Nonribosomal Systems

359

A major step after separation of the gramicidin S multienzyme has been the activation assay measuring the exchange of 32 P-PP¡ in ATP by reversal of aminoacyladenylate formation. This assay dates back to the fifties, when Hoagland studied acetate activation with Lipmann, and later amino acid activation at Zamecnik's lab. Now the mechanism had to be studied with respect to chain elongation, peptidyl-transport, and the control of the sequence of events. Another key contribution then was the characterization of the intermediate amino acyl- and peptidyl-residues as thiolesters. This was accomplished by Wieland Gevers and Horst Kleinkauf (9) with the aid of a set of stability criteria summarized by Lothar Jaenicke and Feodor Lynen for acyl intermediates of fatty acid biosynthesis (10). This result led to the term thiotemplate mechanism coined by Trine-Lise Zimmer and Sören Laland in their essay The protein thiotemplate mechanism of synthesis for the peptide antibiotics produced by Bacillus brevis (11). One has to be

áMéaÉn /—

η e r

M i i c

c

-

k

.

^

v

,

.

^

^

^

^

^

»

Figure 4 Two ways of portraying a functional principle. Around 1970 the Norwegian artist Geir Helgen made a series of cartoons on the gramicidin S system that can be viewed at the Biokjemisk Institutt at Oslo. The poor pantetheine-fellow forced in the chain for his merry-go-round to collect one letter amino acids, but guess who collects the peptide? Right: The reductive approach from the New York lab.

360

Hans von Döhren and Horst Kleinkauf

Figure 5 The Lipmann's lab in 1970. From left, back row: B. Fedyniak, D. Richter, O. Griffith, C. Walsh, W. Gevers, I. Krisko, P. Herrlich, and R. Anthony. Middle: H. Kleinkauf, R. Roskoski, C. Gillespie, L. Lenard, F. Lipmann, W. Eisinger, M. Salas, L. Spector. Sitting: G. Ryan, J. Gordon, R Tao, M. Schweiger, Α. Joseph.

aware, that the aminoacyl-thioesters triggered many studies searching for similar intermediates during t R N A charging by ligases, without any similar clear evidence ever been obtained (12), similar to the search for a swinging arm in the ribosomal peptidyltransferase reaction. The transport of the growing intermediates between the enzyme sites, or between the enzymes, visualized by an acyl carrier protein analog swinging arm was an exceptionally hot topic at that time. Although the proposal of 4'-phosphopantetheine in peptide synthetases has later been credited by Fritz in some way to Feodor Lynen (13), Horst Kleinkauf recalls Fitzi's disbelief of his first evidence at a New York party in 1968. By that time Sören Laland was looking for lipoic acid as a possible key. Finally Lactobacillus pantothenic acid assays were run at both labs, and C.C. Gilhuus-Moe made the race for gramicidin S (14), while the Lipmann lab followed with the tyrocidine synthetases (15). The final proof for the involvement of the peripheral cofactor thiol in peptide transport was then provided by the Lipmann's with the isolation of the acid-stable intermediates attached to pantetheine fragments obtained from a pepsin digest of gramicidin S synthetase 2. Also the sequential events of initiation were established: Aminoacylation of the

Research on Nonribosomal Systems

361

first position (proline) of gramicidin S synthetase 2 was a prerequisite for the transfer of the starter aminoacyl-residue from synthetase 1 (D-phenylalanine) (16). Now that the principle of elongation had been caught in analogy to fatty acid synthesis (Figure 4), attention turned to the structure and organization of protein templates. The tyrocidine system was finally resolved into three components, the separation of the intermediate (230 kDa) and heavy enzyme (450 kDa) at first causing problems at both the Osaka Protein Research Institute and Rockefeller. Bob Roskoski had the significant share in verifying the reaction sequence in analogy to gramicidin S (17). These data then led to a postulated "subunit" size of 70 kDa for an amino acid activating function.

Hunting for Subunits During the seventies the proposed organization of the 70 kDa activating subunits grouped around a small peptidyl carrier with the swinging arm was under investigation from both genetic (Yoshitaka Saito, Hyogoken) and protein studies, now with Horst Kleinkaufs new founded lab in Berlin. Sung G. Lee with Lipmann in 1973 reported an unusually tight structure of the tyrocidine multienzyme complexes (18), with some evidence for dissociation into the proposed fragments. Hans Koischwitz (now von Döhren) claimed at the first Lipmann Meeting in Berlin in 1974 an integrated structure of the heavy enzyme of gramicidin S synthetase (GS 2) to be required for functioning (19). Actually evidence for such a single-chain structure had been described in an unpublished study by E. G. Astrup in the Oslo lab already in 1972 (20), but the result was too unexpected to be acceptable. Recollecting those years getting used to the big sizes, Lipmann himself was not easily convinced of the giant structures, but rather considered other unusual covalent connections between subunits, like lipids. Since the aminoterminal of GS 2 was blocked, a set of indirect evidences was collected for the integrity of the 280 kDa multienzyme (21). A most careful analysis has been conducted by Yoshitaka Saito's lab, isolating from nonproducer mutants multienzymes with defects in single reactions, and assuring their unchanged sizes (22). Now, at least from cloning work in Bacillus brevis evidence for chromosomal multigenes corresponding to the multienzymes of gramicidin S synthetase (GS 1 and GS 2) and tyrocidine synthetases (TY1 and TY2) has been collected (23). Sung Lee, with the tyrocidine system, remained the only one lucky enough to dissect the multienzymes into subunits, or better called active units, of the expected sizes (24). Despite many efforts, limited proteolysis in the gramicidin S and later enniatin system ended up with fragments of various sizes, but predominantly in the 100 kDa range, with at least one function impaired by the cleavage reaction (25)8. 8

We can recollect very similar problems and failures in the analysis of fatty acid multienzyme complexes. Now known to consist of two polypeptide chains (180 and 185 kDa) the yeast enzyme complex had first been assigned to contain 7 or 8 amino termini corresponding to the expected number of catalytic subunits. Then, in a study conducted at the late Porter's lab, acyl carrier proteins were dissociated from a number of fatty acid synthetases now known to be of multienzymic nature by structural gene analysis. These results are now considered artefacts derived by proteolytic nicking.

Hans von Döhren and Horst Kleinkauf

362 Table 1

Limiting capacity of multifunctional peptide synthetases

Multienzyme

Gramicidin S-l Tyrocidine-1 Gramicidin-1 Bacitracin-2 Tyrocidine-2 Enniatin Gramicidin S-2 Tyrocidine-3 Gramicidin-2 Bacitracin-1 Bacitracin-3 a b

Epimerizations

actual b

Amino acid activations

3 3 5 6 10 7 14 18 17 16 18

1 1 2 2 3 2 4 6 5 5 5

1 1

Size

Number of reactions catalyzed

(kDa)

minimal"

120 115 160 210 230 250 280 450 350 330 380

2 2 3 3 4 5 6 6 7 7 8

-

1 1 -

2 1 2

Peptide/ ester bonds

: _ 1 2 3 2 6 6 5 4 6

Modifications

_ -

1 -

1 -

2 -

minimal number: addition of an amino acid, epimerization, modification actual number: activation reactions (adenylate, thiolester), peptide/ester bond, epimerization, modification

Thus the peptide synthetases clearly represent multifunctional enzymes, a term introduced into biochemistry in the 1976 Annual Reviews of Biochemisty by Kaspar Kirschner and Hans Bisswanger (26). For additional confusion in terminology, however, these multifunctional enzymes have to interact for their task of peptide formation. Although this implies complexes of multifunctionals, no such stable associates have been detected so far in vitro. Table 1 shows a selection of multienzymes so far characterized (27).

The Mechanism That the thiotemplate multienzyme mechanism (K. Kurahashi) is indeed the mechanism of nonribosomal peptide synthesis has become evident from a collection of systems investigated so far (Table 2). Looking back at the projects trying to trace biosynthetic events, many aspects of some systems were uncovered, but many questions remained or emerged. The questions how big a peptide can be made by this mechanism was addressed by Harald Mohr with the alamethicin system of Trichoderma viride. This linear 19peptide with a terminal phenylalanine has been made by a partially purified fraction, and there is no doubt that its slightly larger congener, the 23-peptide suzukacillin, has a thiotemplate, too. The details of enzyme organization remain to be established (28). With the 1972 report of even larger ornithine containing peptides made by variants of the Gauze-Brazhnikova strain by the Russian Bacillus brevis specialists around G . G . Zharikova, we may speculate on a limit beyond 30 residues (29). By now we are meeting the field of heavily modified ribosomally made peptides containing D-amino acids, with nisin and subtilin as best known examples.

Research on Nonribosomal Systems Table 2

Current knowledge of enzymology of peptides

Peptide Linear Edeine Gramicidine Alamethicine

Structure 3

Organisms

Pénicillium cyclopium Escherichia coli Aspergillus quadricinctus Bacillus brevis A T C C 9999 Bacillus brevis A T C C 8185

Cyclosporin

Tolypocladium

Mycobacillin

Bacillus

Lactones Destruxin

Metarhizium

Actinomycin

Streptomyces

inflatum

subtilis

anisopliae clavuligerus

Branched cyclic Polymyxin Bacillus

polymyxa

Bacitracin

Bacillus

licheniformis

Depsipeptides Enniatin Beauvericin

Fusarium Beauveria

b

State b

Bacillus brevis Vm4 ß-Tyr-Ise-Dpr-Dahaa-Gly-spermidine pp, 3 me Bacillus brevis A T C C 8185 f-Val-Gly-Ala-D-Leu-Ala-D-Val-Val-D-Val- pp, 4 me? Trp-(D-Leu-Trp) 3 -ethanolamine Trichoderma viride AcAib-Pro-Aib-Ala-Aib-Ala-Gln-Aibpp, me? Val-Aib-Gly-Leu-Aib-Pro-Val-Aib-AibGlu-Gln-Pheol

Cyclic Cyclopeptin Enterochelin Ferrichrome Gramicidin S Tyrocidine

a

363

oxysporum bassiana

c (Anthranilate-Phe) c (Dhb-Ser)j c (Gly 3 -HyOrn 3 ) c (D-Phe-Pro-Val-Orn-Leu) 2 c (D-Phe-Pro-Phe-D-Phe-Asn-Gln-TyrVal-Orn-Leu) c (MeBmt-Abu-Sar-NMeLeu-ValNMeLeu-Ala-D-Ala-NMeLeuNMeLeu-NMeVal) c (Pro-D-Asp-D-Glu-(y)-Tyr-Asp-TyrSer-D-Asp-Leu-D-Glu-(y)-D-Asp-Ala-DAsp)

PP PP, 3 e ρ me ρ, 2 m e ρ, 3 m e

c (D-HMiv-Pro-Ile-NMeVal-NMeAla-/?Ala) Mha-c (Thr-D-Val-Pro-NMeGly-Val)

pp, me

Oct-Dab-Thr-Dab-c (Dab-Dab-LeuLeu-Dab-Dab-Thr) c (Ile-Cys)-Leu-D-Glu-Ile-c (LysD-Orn-Ile-Phe-His-D-Asp-Asn)

pp, me?

c (NMeVal-D-Hiv) 3 c (NMePhe-D-Hiv) 3

ρ me ρ me

ρ, me?

ρ, 3 m e

p, e-l-2me

p, 3 m e

A b u aminobutyrate, Ac acetyl, Aib aminoisobutyrate, c cyclo-, D a b diaminobutyrate, D a h a a 2,6diamino-7-hydroxyazaleic acid, D h b dihydroxybenzoate, D p r diaminopropionate, f formyl-, Hiv hydroxyisovalerate, HMiv hydroxymethylisovalerate, Ise isoserine, Oct octanoyl-, MeBmt (4R)-4((E)-2-butenyl)-4-7V-dimethyl-L-threonine, N M e ΛΓ-methyl-, pheol-phenylalaninol, Sar sarcosin ρ purified, p p partially purified, e enzyme, me multienzyme, ? actual number of multienzymes not known

Their structures were elucidated by the late Erhard Gross. Again Kiyoshi Kurahashi pioneered in this area establishing a 129-peptide precursor of the 32-peptide subtilin9. The proposed mechanism of a cysteine-thiol addition to dehydro residues leading to inversion results in an unusually tight structure. So far no studies have been carried out on single residue inversions within peptides. Such a structure has recently been identified in dermorphine, a skin peptide of the tree frog Phyllomedusa sauvagei by Kreil and coworkers (30). 9

K. Kurashashi, personal communication.

364

Hans von Döhren and Horst Kleinkauf

i o

| ^ ^

C CO Gly HS

SH

S•

,I o o

.

-SH

»SH

Ζ I Qly

A

"

."

ÇHJ /? V"

S

HS

I S

Figure 6 Adjacent thiol model for peptide synthesis. Such functionalized crown ethers have been used by Sasaki et al. as an enzyme model (32).

Summarizing the early results of the thioltemplate mechanism, Lipmann in his 1971 Science-paper (31) traces a process evolution from homomeric polymerization of fatty acid biosynthesis, over the limited heteromeric processes of the thiotemplate, to the unlimited ribosomal system. This view, relating very general aspects, led to several evolutionary speculations, looking at the thiotemplate as an early mechanism, a remnant of evolution. Indeed, models of adjacent active esters may be useful to trace such mechanisms (Figure 6). On the other hand, these systems may well be considered as highly evolved, most complex metabolic systems, especially looking at the requirement for multifunctional fused enzyme functions. If the sizes of protein templates with about IO6 Da for a 12 to 15 membered peptide are compared to a corresponding mRNA, one has to note, that the ribosomal machinery also involves the aminoacyl-tRNA synthetases and their cognate tRNAs, together with all factors. This permits an estimation of the biosynthetic potential required for peptide construction. Such an estimate could consist of size of the actual macromolecular machinery (e. g. in terms of NTPs and peptide bonds required), and efficiency of synthesis, one mole of ATP per peptide bond in the enzymic system versus additional GTP and proof reading energy in the ribosomal system. Indeed the size of the systems involved could be comparable for peptides of the size of 10 to 30 residues, depending on the sequence, then turning in favour of the RNA-directed system. Not accepting such a comparison, one could ask for important functions of such peptides for their producers, if these have maintained such complex metabolic systems. Disappointing however is to see nonproducer mutants to be largely unaffected. Peptide antibiotic production has at least historically proved to be of life saving importance in the Dubos selection experiment (1). The antibiotic point of

Research on Nonribosomal Systems

365

advantage is at least understandable from this point. The many speculative attempts to define an intrinsic function, like a transcriptional regulator, have not yet been convincing. Most recent hopes center around the identification of regulatory controls of these biosynthetic routes to trace possible functions in developmental processes (23,27). But aside from the unclear functions of these metabolites, more enzyme systems directing their formation are being studied. In a survey on peptide structures we have discussed that these generally can be reduced to linear precursors undergoing either cyclizations by peptide or ester bond formation, or end group modification (33). These few structural types of peptides could well be derived by such protein template elongation processes including template directed cyclizations. The activation of hydroxy acids has been shown to proceed in complete analogy to amino acid utilization including acyladenylates and thiolesters. In addition, a growing number of peptides have been identified containing not only amino and hydroxy acids, but also acetate elements. Thus enzyme systems mixing polypeptide and polyketide functional units can be assumed to exist. So new forms of enzyme organization have been discovered, leaving behind the concept of one gene - one enzyme rule. We have detected single multifunctional enzymes, interacting multifunctional enzymes, and enzymes interacting with multifunctional enzymes. Though this work has been confined to systems polymerizing carboxyl-activated compounds, such organizational forms have been found in other sequential pathways, too.

Now and Next At the 1985 conference on the Regulation of Secondary Metabolite Formation (34), it has been pointed out, that the genetics of antibiotic producers are in their very beginnings. The study of the complex regulatory mechanisms would be much easier performed with well studied systems like Escherichia coli, Bacillus subtilis, or Aspergillus nidulans, respectively. Within the past two years Mohamed Marahiel and coworkers have studied the regulatory region of the tycA - gene, the first multienzyme of the tyrocidine biosynthetic system (TY1) from Bacillus brevis ATCC 8185. TY1 was found to be expressed in E. coli constitutively, being under developmental regulation in B. subtilis as originally in B. brevis. Directly following the RNA polymerase binding site the tycA gene contains a transcribed region of about 300 base pairs, apparently accessible to a repressor type mechanism also present in B. subtilis. Making use of the still unclear link between antibiotic production and sporulation, Marahiel, in collaboration with Richard Losick's laboratory in a delightful set of experiments, has introduced the tycA regulatory site fused to a lacZ expression signal with the aid of an SPß-transducing phage system to the chromosome of several s/wO-mutants of B. subtilis. With these mutants being impaired at the very onset of sporulation at T0, it is clearly shown, that an SpoOA mutation not transcribing tycA, also not expressing a B. subtilis antibiotic, can still be led to expression by a second mutation in AbrB, an antibiotic repressor gene (35). So an antibiotic production regulating system has been found operative in both

366

Hans von Döhren and Horst Kleinkauf

strains of Bacillus, respectively, implicating homologies to be a sound basis in the analysis of secondary metabolite expression. At the same time structural homologies of enzymes prove to be of similar importance in the isolation of genes. Crossreactions of anti-gramicidin S-synthetase antibodies with various other peptide synthetases have been shown (36) and facilitated the isolation of tyrocidine synthetase structural genes (37,23). Sequence data now to be obtained will aid the understanding of the organization of the multifunctional enzyme systems. As a new organizational type of multienzyme recently the cyclosporine synthetase of Tolypocladium inflatum has been characterized (38). This enzyme system does not dissociate into a set of multienzymes, but can be purified as functional entity producing the cyclic undeca-peptide, a potent modulator of the immune system. In this case the enzyme system will be useful permitting the synthesis of peptide analogs for structure/function studies. Other systems at present under investigation in various laboratories focus on fungi interacting with insects or plants (destruxins - Metarhizium anisopliae, victorine - Cochliobolus victoriae, HC-toxin - Cochliobolus carbonum) or the improvement of productivity of cultures of Streptomycetes peptides or various jß-lactam producers. Wide uses of homologies should enhance those studies employing heterologous antibodies or gene probes. One of the key questions remaining has been the actual mechanism of protein templates. The rotating cofactor reaching individual enzyme sites, the Lynen-Lipmann model as Spirin has termed it 10 , does not account for all observations of the peptide forming thiol template. This template contains sites not to be distinguished by their enzymic mechanisms, as in fatty acid formation. It rather has to involve a kind of directed mechanism, as the mRNA movement, best envisioned by some conformational change upon ligand dissociation. This mechanism, together with the three dimensional set up will be under investigation in the near future.

References 1. Dubos, R.J. 1939. J. Exp. Med. 70, 1, 11. 2. Lipmann, F., O.K. Behrens and E.A. Kabat, D. Burk. 1940. Science 91, 21. 3. Lipmann, F., R . D . Hotchkiss and R.J. Dubos. 1941. J. Biol. Chem. 141,155; independently the structure had been investigated by Halvor N. Christensen and his colleagues from Lederle Laboratories. 4. Barry, J.M. and E. Ichihara. 1958. Nature 181, 1274. 5. Berg, T.L., L.O. Freholm and S.G. Laland. 1964. Abstr. 1st Meet. Fed. Eur. Biochem. Soc. London A9. 6. Kurahashi, K. 1961. Abstr. 5th Int. Congr. Biochem. Pergamon Press, Oxford, p. 37. 7. Tornino, S., M. Yamada, H. Itoh and Κ. Kurahashi. 1967. Biochemistry 6, 2552. 8. Kleinkauf, Η. and W. Gevers. 1969. Cold Spring Harb. Symp. Quant. Biol. 34, 805. 9. Gevers, W., H. Kleinkauf and F. Lipmann. 1969. Proc. Nat. Acad. Sci. 63, 1335. 10. Jaenicke, L. and F. Lynen. 1960. In: The Enzymes (P. D. Boyer, H. Lardy and K. Myrback, eds.) 2nd ed. Vol. 3, p. 3, New York. 11. Zimmer, T.-L. and S.G. Laland. 1973. Essays Biochem. 9, 31.

10

see Alexander Spirin's contribution in this volume.

Research on Nonribosomal Systems

367

12. Schimmel, P. 1987. Annu. Rev. Biochem. 56, 125. 13. Lipmann, F. 1984. Annu. Rev. Biochem. 53, 30. 14. Gilhuus-Moe, C.C., T. Kristensen, J.E. Bredesen, T.L. Zimmer and S.G. Laland. 1970. FEBS Lett. 7, 287. 15. Kleinkauf, H., W. Gevers, R. Roskoski and F. Lipmann. 1970. Biochem. Biophys. Res. Comm. 41, 1218. 16. Kleinkauf, H., R. Roskoski, Jr. and F. Lipmann. 1971. Proc. Nat. Acad. Sci. 68, 2069. 17. Roskoski, R., Jr., W. Gevers, H. Kleinkauf and F. Lipmann. 1970. Biochemistry 9, 4839, 4846. 18. Lee, S.G., R. Roskoski, Jr., K. Bauer and F. Lipmann. 1973. Biochemistry 12, 398. 19. Kleinkauf, H. and H. Koischwitz. 1974. In: Lipmann Symposium: Energy Biosynthesis and Regulation in Molecular Biology (D. Richter, ed.) de Gruyter Berlin, New York, pp. 336 - 344. 20. Astrup, E.G. 1979. Thesis University of Oslo 1972, cited from T.L. Zimmer, 0 . Froyshov, S.G. Laland in Economic Microbiology, Vol. 3, Secondary Products of Metabolism (A. H. Rose, ed.). Academic Press, p. 141. 21. Kleinkauf, H. 1979. Planta Med. 35, 1. 22. Hori, Κ., T. Kurotsu, M. Kanda, S. Miura, Y. Yamada and Y. Saito. 1982. J. Biochem. 91, 369, and references therein. 23. Kleinkauf, H., M. A. Marahiel, Η. ν. Döhren, G. Czekay, D. Frese, M. Krause and G. Mittenhuber. 1987. 18th FEBS Meet. Ljubljana, p. 10 and unpublished data. 24. Lee, S.G. and F. Lipmann. 1974. Proc. Nat. Acad. Sci. 71, 607; 1977. Proc. Nat. Acad. Sci. 74, 2343. 25. Skarpeid, H. J., T. L. Zimmer and H. v. Döhren. 1986. Biol. Chem. Hoppe-Seyler 367, 77; Suppl. 17th FEBS Meet. 26. Kirschner, Κ. and H. Bisswanger. 1976. Annu. Rev. Biochem. 45, 143. 27. Kleinkauf, Η. and Η. v. Döhren: Crit. Rev. Biotechnol. (in press). 28. Möhr, Η. 1977. Biosynthese von Alamethicin, Thesis T U Berlin Mohr, H. and H. Kleinkauf. 1978. Biochim. Biophys. Acta 526, 375. 29. Kleinkauf, H. and H. Koischwitz. 1978. Progr. Mol. Subcell. Biol. 6, 59. 30. Richter, K., R. Egger and G. Kreil. 2987. Science 238, 200. 31. Lipmann, F. 1971. Science 173, 875. 32. Sasaki, S., M. Shionoya and Κ. Koga. 1985. J. Am. Chem. Soc. 107, 3371. 33. Kleinkauf, Η. and Η. v. Döhren. 1981. Current Topics Microbiol. Immunol. 91, 129. 34. Kleinkauf, Η., Döhren, Η. von, Dornauer, Η. and Nesemann, G. (Editors): Regulation of Secondary Metabolite Formation, Verlag Chemie, Weinheim 1976. 35. Marahiel, Μ. Α., Zuber, P., Czekay, G. and Losick, R. 1987. J. Bacteriol. 169, 2215. 36. Bothe, D. 1986. Thesis. Technische Universität Berlin. 37. Marahiel, M. Α., Krause, M. and Skarpeid, H . J . 1985. Mol. Gen. genet. 201, 231. 38. Billich, A. and Zocher, R. 1987. J. Biol. Chem. 262, 17258.

Metabolism of Carnosine and Related Peptides Karl Bauer

Introduction Carnosine, the first peptide which had been isolated f r o m natural material, was already isolated in 1900 by Gulewitsch a n d Amiradzibi f r o m Liebigs m e a t extract (1). Subsequently, this substance was identified as ß-alanyl-histidine (2,3) a n d meanwhile various ω-aminoacyl amino acids such as /i-alanyl-l-methyl-histidine (anserine) ( 4 - 6 ) , ß-alanyl-ornithine (7), ß-alanyl-lysine (8) as well as the y-aminobutyric acid (GABA)-containing peptides, GABA-histidine (homocarnosine) (9), GABA-1 -methylhistidine (homoanserine) (10), G A B A - L y s (11) and G A B A - O r n (7) have been isolated f r o m excitable tissues, brain and muscle. Apparently due to the restricted availability of the precursors, carnosine a n d the ßalanyl-containing peptides are mainly f o u n d in skeletal muscles a n d m a m m a l i a n olfactory tissue (12). While small a m o u n t s of carnosine a n d anserine are also f o u n d in the central nervous system (CNS), homocarnosine and other y-amino-butyric acid-containing peptides are present uniquely in the CNS. In both cases, the ωaminoacyl amino acids represent m a j o r constituent of histidine metabolism; in some muscles carnosine and anserine are recognized as the m a j o r non-protein nitrogenous constituents which contribute as much as 0 . 5 - 1 % to the wet weight of the tissue.

Physiological Function of ω-Aminoacyl Amino Acids Surprisingly, n o information is yet available as to the biological function of these peptides. There have been m a n y suggestions concerning carnosine's role in glycolysis, oxidative phosphorylation muscle contraction (for review see 13) and wound healing (14). These suggestions, however, have not been verified a n d also not the earlier hypothesis that carnosine functions as a buffer in muscles t h a t maintain their energy supplies by anaerobic means. M o r e recently, it has been proposed that carnosine could serve as a neurotransmitter or n e u r o m o d u l a t o r in primary olfactory neurons (15). However, even this function remains subject to further investigations since electrophysiological d a t a supporting (16) and disputing (17) this view have been published. As to the physiological function of homocarnosine, the m a j o r ω-aminoacyl amino acid in the brain, even a speculative hypothesis does not exist. The clinical a p p r o a c h to find some clues as to the function of these peptides, namely to determine these peptides in biological fluids f r o m persons with a variety of pathological conditions, unfortunately, was also not successful either. Significant changes in urinary excretion of ω-aminoacyl amino acids a n d dipeptide concentrations in cerebrospinal fluid had been observed but it remained unclear

370

Karl Bauer

whether disturbed metabolism of these dipeptides and/or their precursors might be the cause or the result of the pathological disorders. Despite these negative results, there are good arguments to support the view that these substances are biologically important and do serve definite functions: they do not represent products of catabolic processes and they are not synthesized by general, constitutive enzymes but by a discrete enzyme, carnosine synthetase. Correspondingly, these peptides are hydrolyzed by discrete enzymes (e.g. carnosinase) but not by general peptidases such as aminopeptidases and carboxypeptidases. As another argument one might add the ω-aminoacyl amino acids are formed under energy consumption. This fact should be taken into account at least for some tissues which contain these peptides in very high concentrations. Therefore, it will be most important to study the biogenesis, degradation and distribution of coaminoacyl amino acids and especially to study the enzymes which catalyze their metabolism. It is hoped that studies along these lines may offer a clue as to the functions of ω-aminoacyl amino acids in excitable tissues.

Biosynthesis of ω-Aminoacyl Amino Acids Studies on intact animals (18,19) and tissue slices (20,21) provided first evidence that carnosine and anserine are synthesized directly from their component amino acids. Subsequent studies with cell free extracts then clearly demonstrated that carnosine and related peptides are synthesized enzymatically by an enzyme termed carnosine synthetase (22,23) and indicated that ß-alanyl adenylate might be formed as an intermediate since enzymatic synthesis of carnosine could be observed with synthetic ß-alanyl adenylate (22). Therefore, carnosine synthetase was catalogued as an AMP-forming L-histidine: /^-alanine ligase (EC 6.3.2.11). However, the exact mechanism of peptide bond formation could not be elucidated since, owing to the instability of the enzyme, only crude enzyme preparations could be used for the earlier studies. Recently, we succeeded in purifying carnosine synthetase from chick pectorial muscle and also from rat muscle. After purification by ion-exchange chromatography on DEAE-Cellulose, hydrophobic interaction chromatography, adsorption chromatography on hydroxyapatite and gel filtration, we obtained a highly active enzyme preparation devoid of contaminating ATP-hydrolyzing activities. With these enzyme preparations, we observed /^-alanine dependent A T P hydrolysis under formation of A D P and P¡. Interestingly, synthesis of carnosine was found to be strongly inhibited by ADP, P¡ and also by PP¡. These data indicate that the mechanism of carnosine synthesis resembles that of glutathion but clearly differs from the mechanisms of amino acid activation as suggested or observed in other cases (e.g. protein synthesis; synthesis of peptide antibiotics). The exact mechanism of ß-alanine activation (e. g. formation of an acyl phosphate as an intermediate, etc.) remains to be elucidated. In agreement with the findings by Kalyankar and Meister (22), we also observed that the highly purified enzyme from chicken muscle not only catalyzes the stereospecific synthesis of carnosine, but also the synthesis of anserine, /i-alanyl-

Metabolism of Carnosine and Related Peptides

371

lysine, β-alanyl-arginine as well as the y-aminobutyryl-containing peptides. However, the ability of the chicken muscle enzyme to utilize 1-methylhistidine may explain only to some extent the high concentrations of anserine found in chicken muscle. As pointed out by McManus (24) and others (25), anserine appears to be formed primarily by direct methylation of carnosine by an enzyme termed carnosine-A'-methyl-transferase which catalyzes the transfer of the methyl group of S-adenosylmethionine specifically to the nitrogen in the 1-position of the imidazole ring in carnosine. Furthermore, the broad substrate specificity of carnosine synthetase from chicken muscles not necessarily implies that all ω-aminoacyl amino acids are synthesized by only one enzyme and that the availability of the precursors especially β-alanine and GABA, exclusively determines the formation of the products. With carnosine synthetase from rat muscle, we observed a rather restricted substrate specificity. Surprisingly, this enzyme does not accept GABA as a substrate (E. Geßner and K. Bauer, unpublished) and, therefore, one might speculate that carnosine synthetase from rodent brain may also differ from the corresponding muscle enzyme. Although the carnosine synthetase from avian and rat muscles can be purified by the same fractionation procedures and apparently exhibit very similar chemical characteristics (e.g. inhibition by -SH modifying reagents), considerable differences between the avian and mammalian enzymes are also noticed. Using different monoclonal antibodies generated against rabbit muscle carnosine synthetase, Margolis and co-workers (26) recently demonstrated that these antibodies bind carnosine synthetase from mammalian tissues but not that from chicken muscle. Obviously, the avian enzyme lacks certain epitopes that are common to all mammalian enzymes. Furthermore, these enzymes also exhibit significant differences in their enzymological properties (e. g. inhibition by phosphate, the product of ATP hydrolysis) and especially in their substrate specificity towards acceptor amino acids. While 1-methyl-histidine and 3-methyl-histidine are only poor substrates for carnosine synthetase from mouse (27) and rat (E. Geßner, unpublished) tissues, they are readily accepted by the avian enzyme. Histamine is neither accepted by the mammalian nor by the avian enzymes. In this context it is interesting to point out that an enzyme from Carcinus means catalyzes the formation of ß-alanylhistamine from ß-alanine and histamine (28) but does not utilize histidine as an acceptor amino acid. Otherwise, however, this enzyme appears to be very similar to carnosine synthetase. Since similar peptides (e.g. y-glutamyl-histamine) are also synthesized by other invertebrates and insects, it will be interesting to study the phylogenic development of the dipeptide-synthesizing enzymes and the biological functions of ω-aminoacyl amino acids in different species.

Metabolism of ω-Aminoacyl Amino Acids by Brain and Muscle Cells in Culture When working with extremely heterogeneous tissues, it is very difficult to find some understanding on the yet unknown mechanisms regulating carnosine metabolism. This is especially true since the localization of the enzymes within these tissues is yet

372

Karl Bauer

unknown and therefore we were interested to take advantage of the cell culture systems meanwhile available. When neuronal cells from embryonic rat brain in primary culture were incubated with radiolabeled ß-alanine, we observed rapid uptake of the tracer but no incorporation into carnosine. The same results were obtained with various hypothalamic cell lines (in collaboration with F. de Vitry and A. Tixier-Vidal, Collège de France, Paris) or different neuroblastoma cells in culture (in collaboration with B. Hamprecht, Tübingen). In contrast, tracer incorporation studies with the rat C 6 rat glioma cell line (7) or with astroglia-rich primary cell cultures (29) clearly demonstrated that carnosine and related peptides are actively synthesized by these cells, but not degraded. When the astrogliarich cultures were treated with dibutyrylcyclic A M P or other agents that elevate intracellular levels of cyclic AMP, the cells assume astrocyte-like morphology, and carnosine synthesis was found to be strongly reduced (M. Schulz, unpublished). The parallelism between the effect of these substances on carnosine synthetase activity and on the observed morphological changes suggest that carnosine synthesis might be related to differentiation of glial cells. Further studies on the regulation of carnosine synthesis eventually may not only offer a clue as to the physiological function of ω-aminoacyl amino acids, but hopefully may also provide some insight into the differentiation and physiological function of the enigmatic astroglia cells. Our studies on the biosynthesis of carnosine by primary cultures of embryonic chicken breast muscle also indicate that carnosine synthesis may be related to processes of cell differentiation. Carnosine synthesis was found to by very low as long as the cultures consisted of mononucleated myoblasts and drastically increased when the cells fused to form polynucleated myotubes. After considerable amounts of carnosine already accumulated, anserine was formed very slowly with a lag phase of several hours. These results indicate that anserine is not formed as a by-product of carnosine synthesis but apparently via methylation of carnosine, as suggested already by others (23,24). Furthermore, these results also demonstrate that carnosine obviously is not a depositary product which is synthesized by other cells (e.g. motoneurons, Schwann cells etc.).

Uptake and Release of Carnosine In some muscles and especially in chicken pectorial muscle, carnosine is found in very high concentrations. With our muscle cell cultures, correspondingly we observed that carnosine rapidly accumulated in these cells and then released very slowly into the medium. In contrast, with the astroglial-rich cultures we found that the endogeneously synthesized peptides are released very rapidly. Inversely, with these cells we also observed that carnosine, when added to the culture medium, is rapidly taken up by an energy and sodium dependent process (30). Uptake and release of carnosine by astroglial cells was found to be mediated by a saturable, high-affinity transport system which is not specific for carnosine but apparently represents a general, dipeptide-specific transport system with broad substrate specificity. Tri- and oligopeptides are not transported and, surprisingly, significant

Metabolism of Carnosine and Related Peptides

373

uptake of homocarnosine was also not observed. With neuronal cell cultures, accumulation of carnosine was not observed but it remains to be investigated whether carnosine uptake is a specific feature of astroglia cells which could be used as marker for cell identification.

Degradation of ω-Aminoacyl Amino Acids Carnosine and other ω-aminoacyl amino acids are degraded by discrete enzymes distinctly different from general hydrolytic enzymes such as amino and carboxypeptidases or general dipeptidases. Hanson and Smith (31) were the first who described a carnosine degrading enzyme from swine kidneys. Since then, this metal dependent carnosinase from various tissues has been studied by several investigators (32-35). This enzyme exhibits high KM values with carnosine as substrate and highest activities are found in the kidneys of various animals. From rat brain, we recently identified also a carnosine degrading enzyme with very similar characteristics (36). This enzyme is inhibited by -SH reactive agents, stabilized by dithiothreitol and dependent on Μη 2 + ions for full biological activity. Interestingly, this enzyme hydrolyzes /J-Ala-Arg fifty times more effectively than carnosine but it does not degrade the corresponding y-aminobutyryl-dipeptides. While this enzyme exhibits high KM values with carnosine (Ä"M = 25 mM), another carnosinase, meanwhile identified and purified from mouse kidney (34) and rat brain (36) exhibits a low KM value (0.02 mM) towards this substrate. This carnosinase is not dependent on metal ions and is strongly inhibited by Μη 2 + as well as by heavy metal ions. Both carnosinases do not hydrolyze homocarnosine and it remains to be elucidated whether homocarnosine and other ω-aminoacyl amino acids are hydrolyzed by highly specific enzymes, not identified yet. The physiological functions of these enzymes also remains obscure, so far. In the kidneys, such enzymes may serve an important function in the hydrolysis of dietary or circulating dipeptides and thus for the resorption of the constituent amino acids. Whether they are important for the hydrolysis of ω-aminoacyl amino acids at their sites of synthesis or for the inactivation of the peptidergic signals remains to be investigated.

Perspectives Although the biological functions of carnosine and other ω-amino-acyl amino acids are yet unknown, their existence and the presence of discrete enzymes involved in the synthesis and degradation of these dipeptides strongly suggest a definite function. Since in some tissues these substances are present in high concentrations it is not surprising that pharmacological effects are not observed when these peptides are applied directly by various routes. Alternatively, one might exepct that their biological functions become apparent under experimentally induced conditions of imbalanced dipeptide metabolisms, e. g. by inhibiting carnosine synthetase in order

374

Karl Bauer

to generate a situation of carnosine or homocarnosine deficiency which, eventually, might even resemble certain pathological disorders. It is hoped that studies along these lines will finally provide the tools to attain an understanding o n the biological functions of these enigmatic substances in excitable tissues.

Acknowledgement I thank U . Heinrich, M . Schulz and E. Geßner for their help and invaluable contribution, B. Hamprecht, University of Tübingen, for obliging help and stimulating discussions, H. Kleinkauf for his interest and the Deutsche Forschungsgemeinschaft for financial support.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29.

Gulewitsch, W. and S. Amiradzibi. 1900. Ber. Dt. Chem. Ges. 33, 1902. Barger, G. and F. Tutin. 1918. Biochem. J. 12, 402. Baumann, L. and T. Ingwaldsen. 1918. J. Biol. Chem. 35, 263. Ackermann, D., O. Timpe and K. Poller. 1929. Z. Physiol. Chem. 183, 1. Du Vigneaud, V. and O.K. Behrens, 1939. Ergeb. Physiol. Biol. Chem. Exp. Pharmakol. 41, 917. Linneweh, W., A.W. Keil and F.A. Hoppe-Seyler. 1929. Z. Physiol. Chem. 183, 11. Bauer, K., J. Salnikow, F. de Vitry, A. Tixier-Vidal and H. Kleinkauf. 1979. J. Biol. Chem. 254, 6402. Matsuoka, M., T. Nakajima and I. Sano. 1969. Biochim. Biophys. Acta 177, 169. Pisano, J. J., J.D. Wilson, L. Cohen, D. Abraham and S. Udenfriend. 1961. J. Biol. Chem. 236, 499. Nakajima, T., F. Wolfgram and W.G. Clark. 1967. J. Neurochem. 14, 1107. Kumon, Α., Y. Matsuoka, T. Nakajima, Y. Kakimoto, Ν. Imaoka and I. Sano. 1970. Biochim. Biophys. Acta 200, 170. Margolis, F.L. 1974. Science 184, 909. Crush, K.G. 1970. Comp. Biochem. Physiol. 34, 3. Fitzpatrick, D.W. and H. Fisher. 1982. Surgery 91, 56. Margolis, F.L. 1978. Trends Neurosci. 1, 42. Gonzalez-Estrada, M.T. and W.J. Freeman. 1980. Brain Res. 202, 373. MacLeod, N.K. and D.W. Straughan. 1979. Exp. Brain Res. 34, 183. Harms, W.S. and T. Winnick. 1954. Biochim. Biophys. Acta 15, 480. Parshin, A.M. and T.A. Gordukhina. 1953. Doklady Akad. Nauk S.S.S.R. 88, 113. Williams, H.M. and W. A. Krehl. 1952. J. Biol. Chem. 196, 443. Razina, L.G. 1956. Doklady Acad. Nauk S.S.S.R. Ill, 161. Kalyankar, G. and A. Meister. 1959. J. Biol. Chem. 234, 3210. Winnick, R.E. and T. Winnick. 1959. Biochim. Biophys. Acta 31, 47. McManus, I.R. 1962. J. Biol. Chem. 237, 1207. Winnick, T. and R.E. Winnick. 1959. Nature 183, 1466. Margolis, F.L.M. Grillo, J. Hempstead and J.I. Morgon. 1987. J. Neurochem. 48, 593. Horinishi, H., M. Grillo and F.L. Margolis. 1978. J. Neurochem. 31, 909. Arnould, J.-M. 1987. J. Neurochem. 48, 1316. Bauer, K., H. Hallermayer, J. Salnikow, H. Kleinkauf and B. Hamprecht. 1982. J. Biol. Chem. 257, 3593.

Metabolism of Carnosine and Related Peptides 30. 31. 32. 33. 34.

375

Schulz, Μ., Β. Hamprecht, Η. Kleinkauf and Κ. Bauer. 1987. J. Neurochem. 49, 748. Hanson, H.T. and E.L. Smith. 1949. J. Biol. Chem. 179, 789. Rosenberg, A. 1960. Biochim. Biophys. Acta 45, 257. Lenney, J.F. 1976. Biochim. Biophys. Acta 429, 214. Margolis, F.L., M. Grillo, C.E. Brown, T.H. Williams, R . G . Pitcher and G . J . Elgar. 1979. Biochim. Biophys. Acta 744, Til. 35. Margolis, F.L., M. Grillo, Ν. Grannot-Reisfeld and A.I. Farbmann. 1983. Biochim. Biophys. Acta 744, 237. 36. Kunze, Ν., Η. Kleinkauf and Κ. Bauer. 1986. Eur. J. Biochem. 160, 605.

3 Molecular Biology Sharpens its Tools

When regarding recent advances in nucleic acid analysis the significance of the past decade strikes one immediately. As Wieland GEVERS conjectured in his Dahlem lecture, genetic engineering would be the field attracting Fritz LIPMANN today, a field that was established within the last years of his life. Analysis on the level of the genes certainly was the last barrier to be overcome for research in biology, the final tool for obtaining data at the molecular level, which could then be used as a basis for studies of macromolecular interactions still to be conducted. The steady refinement of protein analysis, with the breakthrough of Polyacrylamide gel electrophoresis in the sixties, followed by new sensitivities in sequencing and separation, were known in the laboratory run by the late Fritz LIPMANN. The cloning impact is particularly striking in gene structural studies and the processing of genetic information in transcription and translation, if we return to the papers presented at the first LIPMANN Symposium in 1974, looking at the tools used by the LIPMANN alumni then and now. The essays at the beginning of this chapter are concerned with the stability of DNA, with relevance to environmental hazards, D N A damage and repair, and drug resistance factors. Manfred SCHWEIGER and his colleagues have directed their work to D N A repair in human cells, focusing on the enzymes of repair reactions. Irving GOLDBERG discusses the mechanism of D N A damage by antitumor compounds tracing energy rich phosphate intermediates in sugar fragmentation. A. KAJI and coworkers describe studies of a bacterial drug resistance plasmid encoding an inherent temperature sensitive instability factor, a phenomenon whose biological significance has yet to be understood. Marion HIEROWSKI and his colleagues describe their recent work in the isolation of human prostatic growth factors, and have shown elevated levels in cancer tissues, which enhances D N A synthesis. F. J.S. LARA comments on the beginning studies on gene amplification in salivary gland chromosomes of Rhynchosciara larvae. A series of contributions follows that treats aspects of the regulation of expression of genes at various levels. Wilfried MOMMAERTS deals with possible switches of the myosin multigene family in connection with the innervation of cells. Peter HERRLICH and Michael KARIN describe the principle of transcription factor modification as a regulatory process, with T7 protein kinase and the mammalian AP I factor as examples. Werner MAAS summarizes his more than 30 years' involvement in the regulation of arginine biosynthesis, presenting here the arginine repressor as the culmination of his studies. Herbert WEISSBACH, Nathan BROT and their coworkers describe the complex regulation of methionine biosynthesis in E. coli, focusing on the terminal reactions effected by methionine and vitamin B 1 2 . B. G. LANE gives a thorough account of the development of the molecular biology of the wheat embryo germination process. M. SUIKO describes a thiol-directed purine analog inhibitor of tumor cell DNA-dependent R N A polymerase.

378

3 Molecular Biology Sharpens its Tools

The unravelling of the structural organization of the ribosome is clearly one of the most complex projects ever tackled in the life sciences. We have all followed the separation of the approximately 60 components by two-dimensional gel electrophoresis, sequencing of all the E. coli ribosomal proteins in WITTMANN'S laboratory, and now the efforts to deduce the three-dimensional arrangement of the components by cross-linking studies, neutron and X-ray scattering, and immune electron microscopy. H. G. WITTMANN and A. YONATH present an account of the most recent approaches. Of particular advantage has been the wide use of sources, especially Bacillus stearothermophilus and halobacteria, in the accomplishment of crystallographic steps by refined techniques. Structural features such as an intersubunit space to accomodate tRNA, elongation factors, and mRNA, and a tunnel for the possible "escape" of the nascent peptide chain are emerging. Masayasu NOMURA recollects his participation in the unravelling of the 30 S initiation model, then describes how through a long path of studying ribosomal genes - he again encountered initiation upon discovery of translational regulation, with ribosomal proteins acting as translational repressors. Still open questions are pointed out, such as the actual signal determining ribosome biosynthesis, possibly by modification of factors. The first eucaryotic genes cloned were rRNA genes, in Stanley COHEN'S laboratory in 1974. A. A. HADJIOLOV compares the organization of eucaryotic rRNA transcription units, and Alexander SPIRIN provides us with a detailed discussion of the energetics and dynamics of protein biosynthesis. Consequently, he retraces the many connections to the LIPMANN laboratory, including the nonribosomal system of enzymatic peptide biosynthesis. Lawrence Fox briefly reviews the highlights of protein biosynthetic research, finally commenting on the expression of recombinant material. Clelia GANOZA discusses in detail structural features of non-coding regions of mRNA, with the search for possible interacting sites of the initiator tRNA, in addition to codon-anticodon pairing. In a joint effort Clelia GANOZA, R. M . BAXTER, and L. Fox consider the peptidyl transferase center and the role of additional factors such as E F - P in the specific enhancement of translation rates. Finally Z . KUCAN reviews the role of the polyamine spermine on the various partial reactions involved in protein synthesis.

DNA Repair in Human Cells: Molecular Cloning of cDNAS Coding for Enzymes Related to Repair Manfred Schweiger, Rainer Schneider, Monica HirschKauffmann, Bernhard Auer, Helmut Klocker, Elisabeth Scherzer, Martin Thurnher, Herbert Herzog, Hans-Peter Vosberg and Erwin F. Wagner

Introduction DNA repair is of great relevance to life; indeed, its importance is increasing since the environmental load and agents injurious to DNA are permanently elevated due to tests of nuclear weapons, accidents of nuclear power stations, increasing levels of hazardous noxious chemicals and extensive UV exposure. Despite this great relevance of DNA repair, our knowledge on its molecular mechanisms are fragmentary. Most of the work performed on DNA repair has utilized lower organisms as Escherichia coli and yeast and only little with human cells. Human mutants of DNA repair, as in hereditary diseases, are not characterized on the molecular level. Very little is known about enzymes participating in DNA repair. Elucidation of the enzymatic defects of DNA repair diseases as well as of the enzymatic mechanisms of DNA repair is required. We have focused our attention on enzymes which play a role during repair of damaged cells.

ADP-Ribosyl Transferase (ADPRT) ADP-ribosyl transferase is involved in DNA repair (1): inhibition of this enzyme potentiates the lethality of chemicals noxious to DNA. DNA repair is accompanied by a strong stimulation of ADP-ribosylation (3) and concomitantly by consumption of N A D + (2). This is diminished in cells from patients with Fanconi's anemia, a hereditary disease associated with defective DNA repair (4). A deficient activity of ADP-ribosyl transferase might be the cause of this disease. Interestingly, ADP-ribosyl transferase is involved also in cellular differentiation (5) and tumorigenic cell transformation (6,7). Inhibition of the enzyme prevents tumorigenic cell transformation and induces differentiation of teratocarcinoma cells. Triggering of differentiation of these cells by retinole acid is accompanied by reduced activity of ADP-ribosyl transferase. The targets of ADP-ribosyl transferase of the nucleus are manifold: ADPRT itself (8), topoisomerases I (9) and II (10, 11), DNA ligase (12), DNA polymerase (13), histones (13), HMG proteins (13), endonuclease (13) and terminal deoxynucleotidyl transferase (13). The molecular mechanisms by which ADPRT controls central

Manfred Schweiger et al.

380

ANTIBODIES

OLIGONUCLEOTIDE

Q

Figure 1 Screening of lambda gtll-library with immuno- and hybridisation techniques: Paired filters ( Α - C , B - D ) , each containing about 200 plaques of recombinant lambda gtt 1, probed once (A, B) with antibodies against A D P R T (2. and 3. round of plaque purification respectively) and second (C, D) probed with an A D P R T specific oligonucleotide. Arrows in C indicate the signals above background.

processes are of special interest. Isolation and characterization of this enzyme and molecular cloning of its gene were required. We have facilitated the purification and characterization of the enzyme from human placenta by developing an efficient affinity chromatography procedure (14). With this technique, 500 μg of the enzyme can be purified from human placenta or from other sources within 12 hours. A D P R T from mammals, from birds and from lower animals such as snails show very similar properties (15,16). The molecular masses are around 116000 Da and the behaviour against inhibitors are comparable. These enzymes crossreact im-

D N A Repair in Human Cells

381

munologically indicating sequence identities. The molecular masses of ADPRT from normal fibroblasts and from patients with Fanconi's anemia or Cockayne syndrome are the same. The same holds true for the Michaelis-Menten constants. However, the maximal velocities of ADPRT in homogenates from Fanconi cells were significatly lower (17).

10 20 30 40 50 60 70 80 90 100 110 120 CAATTCCGAGAAATCTCTTACCTCAAGAAATTGAAGCTTAAAAAGCAGGACCCTATATTCCCCCCAGAAACCAGCCCCTCCCTGGCGGCCACGCCTCCGCCCTCCACAGCCTCCGCTCCT GluPheArgGluIleSerTy rLeuLysLysLeuLysValL/sLvsGlnAspArg I l e P h e P r o P r o G l u T h r S e r A l a S e r V a l A l a A l a T h r P r o P r o P r o S e r T h r A l a S e r A l a P r o 130 140 150 160 170 180 190 200 210 220 230 240 TTCAGCACATAAGCCATTATCCAACATGAAGATCCTGACTCTCGGGAAGCTGTCCCGCAAGAAGGArGAAGCTGCTGTCAACTCCTCTGCGTGAAGGCCATGATTGAGAAACTCGGGCGG Al·aAl·aVal·A5nSeΓSerAl·aSerAl·aAspLvsPΓQLeuSeΓAsnMetLysIleLeuThΓLeuCl·vLvsLeuSeΓAΓHAsnLysAspGl·uValLvsAl·aMet IleGluLvsLeuGlyGly 250 260 270 280 290 300 310 320 330 340 350 360 AAGTTGACGGGGACGGCCAACAAGGCTTCCCTGTGCATCAGCACCAAAAAGGAGGTGGAAAAGATCAATAAGAAGATGGACGAAGTAAAGGAAGCCAACATCCGAGITGTGTCTGAGGAC LysLeuThrGlyThrAlaAsnLysAlaSerLeuCvsIleSerThrLy sLysGluValGluLysMecAsnLysLysMecGluGluValLysGliiAlaAsnlleArgValValSerGluAsp 370 380 390 400 410 420 430 440 450 460 470 480 TTCCTCCAGGACGTCTCCGCCTCCACCAAGAGCCTTGAGGACTTGTTCTrAGCGCACATGTTGTCCCC'ITGGGGGGCAGAGGTGAAGGCAGAGCCTGTTGAAGTTGTGGCCCCAAGAGGG PheLeuGlnAsp^'alSerAlaSerThrLysSerLeuGlnGluLeuPheLeuAlaHisI leLeuSerProTrpGlyAlaGluValLvsAlaGluProValGluValValAIaProArgGly 490 500 510 520 530 540 550 560 570 580 590 600 AAGTCAGGGGCTGCGCTCTCCAAAAAAAGCAAGGGCCAGGTCAAGGAGGAAGCTATCAACAAATCTGAAAAGAGAATGAAATTAACTCITAAAGGAGCAGCAGCTGTGCATCCTGArrCT LysSerGlyAlaAlaLeuSerLysLysSerLysGlvGlnValLysGluGluGly IleAsnLysSerGluLysArgMetLysLeuThrLeiiLysGlyGlyAlaAlaValAspProAspSer 610 620 630 640 650 660 670 680 690 700 710 720 GGACTGGAACACTCTGCGCATGTCCTGGAGAAAGGTGGGAAGGTCTTCAATGCCACCCTTGGCCTGGTGGACATCGTTAAAGGAACCAACTCCTACTACAAGCTGCAGCTTCTGGAGCAC GXyLeuGluHisSerAlaHisValLeuGluLvsGlyGlyLysValPheAsnAlaThr LeuGlyLeuValAspIleVal Ly sGlyThrAsnSerTy rTyrLysLeuGlnLeuLeuGluAsp 730 740 750 760 770 780 790 800 810 820 830 840 GACAAGGAAAACAGGTATTGGATATTCAGGTCCTGGGGCCGTGTGGGTACGGTGTTCGGTAGCATCAAACTGGAACAGATGCCGTCCAAGGAGGATGCCAÎTGAGCACTTCATGAAATTA AspLysGluAsnArgTyrTrpIlePheArgSerTrpGlyArgValGlyThrValPheGLySerlleLysLeuGluGlnMecProSerLysGluAspAlalleGluHisPheMetLysLeu 850 860 870 880 890 900 910 920 930 940 950 960 TATGAAGAAAAAACCGGCAACGCTTGGCACTCCAAAAATTTCACGAAGTATCCCAAAAAGTTCTACCCCCTGGAGATTCACTATGGCCACGATGCCGAGGCAGTGAAGAAGCTGACAGTA TyrGluG LuLysThrGly AsnAlaTrpHi sSerl.y sAsnPheThr LysTyrProLy sLvsPheTvrPrnLeuGl u lleAspTyrGlyGlnAspAlaGluAlaValLysLysLeuThrVal 970 980 990 1000 1010 1020 1030 1040 1050 1060 1070 1080 AATCCTGGCACCAAGTCGAAGCTGCCCAAGCCAGTTCAGGACCTCArCAAGATGATCTTTGATGTGGAAAGTATGAAGAAAGCCATGGTCGAGTATGAGArGGACCTTCAGAAGATGGCC AsnProGlvThrLysSerLvsLeuProLysProValGlnAspLeuIleLvsMet IlePheAspValGluSerMetLysLvsAlaMetValGluTyrGluIleAspLeuGlnLysMetPro 1090 1100 1110 1120 1130 1140 1150 1160 1170 1180 1190 1200 TTGGGGAAGCTGAGCAAAAGGCAGATCCAGGCCGCATACTCCATCCTCAGTGAGGTCCAGCAGGGGGTGTCTCAGGGCAGCAGCGACTCTCAGATCCTGGATCTCTCAAATCGCTTTTAC LeuGlyLysLeuSerLysArgGlnlleGlnAlaAlaTyrSerlleLeuSerGluValGlnGlnAlaValSerGlrGlySerSerAspSerGlnlleLeuAspLeiiSerAsnArgPheTyr 1210 1220 1230 1240 1250 1260 1270 1280 1290 1300 1310 1320 ACCCTGATCCCCCACGAC'nTGGGATGAAGAAGCCTCCGCTCCTGAACAATGCAGACAGTGTGCAGGCCAAGGTGGAAATGCTrGACAACCTGCTGGACATCGAGGTGGCCTACAGTCTG ThrLeuIleProHisAspPheGlyMetLysLysProProLeuLeuAsnAsnAlaAspSerValGlnAlaLysValGluMetLeuAspAsnLeuLeuAspIleGluValAlaTyrSerLeu 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 CTCA(XKGAGCGTCTCATGATA(OCCAAGGATCCCAraATCTCAACTArGAGAAGCTCAAAACTC^ LeuArgGlyGlySerAspAspSerSerLysAspProIleAspValAsnTyrGluLysLeuLysThrAspIleLysValValAspArgAspSerGluGluAlaGluIlelleArgLysTyr 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 GTrAAGAACACTCATGCAAGCACACACAATGCGTATGACITGGAAGTf'ATCGATATGTTrAAGATAGAGCGTGAAGGCGAATGCCAGCGTTAGAAGCGGTTTAAlXAGCTTCArAACCGA ValLysAsnThrHisAlaThrThrHisAsnAlaTyrAspLeuGluV'allleAs^llePheLysI ÍeGluArgGluGlyGluCysGlnArgTyrLysProPheLysGlnLeuHisAsnArg 1570 1580 1590 1600 1610 1620 1630 1640 1650 1660 1670 1680 AGATTGCTGTGGCACGGGTCCAGGACCACCAACnTTGCTGGGATCCTGTCCCAGGGTCmGGATAGCCCCGCCTGAACCGCCCGTGACAGGCTACATGTrrœTAAAGGGATCTATTTC ArgLeuLeuTrpHisGlySerArgThrThr AsnPheAlaGly IleLeuSerGlnClyLeuArglleAlaProProGluAlaProValThrGlyTyrMetPheGly LysGly I l e T y r P h e

1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 GCTGACATGGTCTCCAAGAGTGCCAACTACTGCCATACGTCTCAGGGAGACCCAATAGGCTTAATCCTCTTGGGAGAAGTTCCCCTTGGAAACATGTATGAACTGAAGCACGCTTCACAT AlaAspMetValSerLysSerAlaAsnTyrCysHisThrSerGlnGlyAspProIleGlyLeuIleLeuLeuGlyGluValAlaLeuGlyAsnMetTyrGluLeiiLysHisAlaSerHis 1810 1820 1830 1840 1850 1860 1870 1880 1890 1900 1910 1920 ATCAGCAAGTTACCCAAGGGCAAGCACAGTGTCAAAGGTrrGGGCAAAACTACCCCTCATCCTTCACCTAACATrAGTCTGGATGGTGTAGACGTTCCTCTTGGGACCGGGATTTCATCT IleSerLysLeuProLysGlyLysHisSerValLysGlyLeuGlyLysThrThrProAspProSerAlaAsrlleSerLeuAspGlyValAspValProLeuGlyThrGlylleSerSer 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020 2030 2040 GGTGTGAATGACACCTCTCTACTATATAAGCAGTACATTGTCTATGATATOCTCAGGTAAATCTGAAGTATCTGCTGAAACTGAAATTCAATnTAAGACCTCCCTCTGGTAATTCGGA GlyValAsnAspThrSerLeuLeuTyrAsnGluTy r lleValTy rAspIleAlaGlnValAsrLeuLysTyrLeuLeuLysLeuLysPheAsnPheLysThrSerLeuTrpEnd 2050 2060 2070 2080 GAGCTAGCCGAGTCACACCCGGTGGCTCTGGTATGAATTC

Figure 2 Nucleotide sequence of insert of ADPRT-G8 encoding ADPRT: The underlined amino acid sequences are identical to the sequenced CNBr peptides of ADPRT.

382

Manfred Schweiger et al.

With milligram quantities of enzyme available we raised specific monoclonal as well as polyclonal antibodies and partially sequenced the enzyme. From the amino acid sequences we deduced oligonucleotides and synthesized them. Specific antibodies as well as specific oligonucleotides were employed for cloning of an A D P R T c D N A (18). A D P R T represents about 0.01 percent or less of the soluble proteins in term placentas. This very low abundancy might be the cause for the lack of success in our screenings of several million plaques of placental, liver or heart muscle lambda gtl 1 c D N A libraries. A c D N A library was constructed from m R N A prepared from Hela cells, since the A D P R T content of fast proliferating cells is eleve sd. Additonally, we enriched the c D N A for sequences of more than about 2500 bp. From this c D N A we have constructed a lambda gtl 1 library, which enabled us to detect several positive clones by immunoscreening (Figure 1). Four of these clones were plaque purified through 3 cycles. The strongest signals derived from a clone (ADPRT-G8) which also hybridized with a specific A D P R T oligonucleotide.

O

3

1 2 24 48 72 Ko ÍSS.:

Figure 3 Northern blot of total R N A after serum-stimulation hybridized with A D P R T cDNA: Each lane containes 7 μg total cellular RNA from human fibroblasts harvested at the indicated times (hours) after serum stimulation.

D N A Repair in Human Cells

383

By several lines of evidence we have proven that this clone indeed carries information for ADPRT: This clone reacts specifically with antibodies against A D P R T in the screening procedure. It hybridized to two specific oligonucleotides which were deduced from amino acid sequences from ADPRT. The c D N A clone directed the synthesis of a fusion protein, which reacted with affinity purified antibodies to human A D P R T in Western blot analysis. In the c D N A clone, nucleotide sequences were found that encode peptides which are found in two CNBr peptides from A D P R T (Figure 2). The cDNA detected a m R N A species in Northern blots which is sufficient in length to direct the synthesis of 116,000 Da A D P R T (Figure 3). Finally, the restriction pattern and the sequence are in agreement with independently identified A D P R T clones (19,20). The availability of the A D P R T clones provides the means to study the control of the synthesis of this enzyme. Formation of ADP-ribosyl transferase is induced at the level of transcription when cell proliferation is stimulated e.g. by release from serum starvation. The A D P R T clones will elucidate the mechanisms leading to modulated levels of activity of A D P R T during D N A repair, cell proliferation, differentiation and transformation. It should also enable us to analyse the molecular defect of Fanconi's anemia.

Topoisomerase I Topoisomerase I is one of the targets of ADP-ribosyl transfer by A D P R T and thereby a candidate for a mediator of its function (21). Additionally there are indications that topoisomerase I might participate in D N A repair (21). Therefore we attempted the molecular cloning of cDNA encoding human topoisomerase I. Topoisomerase I was purified from bovine tissue and antibodies were raised in rabbits. With the anti-bovine-topoisomerase I IgG a human c D N A library in lambda gtl 1 was screened. Several positively reacting clones were isolated by plaque purification. These clones had a size of 2 kb. The restriction pattern was determined and by protein blot technique a fusion peptide was detected, which specifically reacted with bovine topoisomerase I antibodies. The c D N A fragment was excised from lambda g t l l and subcloned in Bluescript M l 3 + vector for sequencing. Sequences of the 2.0 kb D N A segment were determined. Experiments providing further confirmation for the identity of this clone are in progress. The cloning of the cDNA encoding topoisomerase I provides us with a tool to elucidate the regulation of this enzyme in the course of D N A repair and should supply us with informations in its role in D N A repair.

RNAse Inhibitor from Human Cells During repair of D N A damage the cell relies preferentially on the existing messenger RNAs, although R N A is affected by nucleic acid damaging agents as well. But R N A is present in multiple copies therefore a portion of each species remains unaffected. Since one pyrimidine dimer blocks the transcription of a gene, as we

384

Manfred Schweiger et al.

1

2

3

Figure 4 One step purification of RNAse inhibitor: Polyacrylamide-gelelectrophoresis (silver stained): 1. Human placenta homogenate, 2. Flow trough of affinity column (RN Ase-Sepharose), 3. Eluate of affinity column.

could show by a newly developed in vivo D N A repair assay (22), it should be advantageous for the cell, if R N A is stabilized as long as transcription is reduced during D N A repair. After completion of repair the damaged R N A will be degraded quickly and substituted by intact m R N A molecules. Such a transient differential preservation and exchange of m R N A molecules could be achieved by regulating RNAse activities and specificities after D N A damage. Human cells indeed possess a protein that could mediate such a function, namely the RNAse inhibitor, which we discovered 20 years ago (23). This RNAse inhibitor is now commonly used in molecular biology to prevent degradation of R N A during in vitro protein synthesis, in vitro transcription, c D N A synthesis and isolation of polysomes, it forms a 1 : 1 complex with RNAse A (/£, = 0.3 nM) and contains highly labile free thiol groups which are essential of its function (24). Recent reports (25,26) suggest that the ribonuclease inhibitor participates in the in vivo regulation of angiogenin, a blood vessel inducing protein from human carcinoma cell, the primary structure of angiogenin is highly homologous to that of the pancreatic ribonuclease. The finding

D N A Repair in Human Cells

1

2

3 4

5

6

385

7

1 2 3 4 5 6 7

1 ν Figure 5 Southern blotting of RNAse inhibitor cDNA clones with a specific oligonucleotide: Left: Ethidiumbromide stain, Right: Hybridization with oligonucleotide. Lanes: (1) Digestion of RNAse inhibitor c D N A clone with EcoRI/Hindlll, (2) KpnI/SstI, (3) EcoRI/PvuII, (4) Eco RI/BamHI, (5) EcoRI/Kpnl, (6) EcoRI/SstI, (7) EcoRI.

that ribonuclease inhibitor abolishes both the biological and the ribonucleolytic activities of angiogenin might have pharmacologic and/or therapeutic implications. Although many properties and functions of the inhibitor have been reported, the structure of the protein and the exact nature of its interaction with RNAse and homologous enzymes or its regulation are still unknown. This and its possible role during D N A repair prompted us to isolate c D N A clones of the RNAse inhibitor to provide a solid basis for studying its structure, function and regulation at a molecular level. A large scale and high yield, one step, purification procedure was developed for ribonuclease inhibitor from human placenta, which led to homogeneous preparations in SDS Polyacrylamide gels (Figure 4). The N-terminus of the protein is blocked; therefore, to obtain an aminoacid sequence, the purified protein was CNBr digested, peptides were separated on a H P L C column and subjected to gasphase sequencing. A corresponding oligonucleotide was synthesized, specific antibodies against the protein were raised in rabbits and used to screen a HeLa lambda gtl 1 expression library. Several identical clones with an insert of about 900 bp were found. Restriction and southern blot analysis of these clones showed that the insert hybridized specifically with the synthetic oligonucleotide (Figure 5). In Western blot analysis of lambda lysogens of these clones a fusion protein of M r 140,000 reacted with antibodies against the ribonuclease inhibitor. These results confirm that we have successfully isolated a c D N A clone for human RNAse inhibitor.

386

Manfred Schweiger et al.

Conclusion The now available c D N A clones for ADPRT, topoisomerase I and RNAse inhibitor provide means to analyse the regulation of synthesis of these enzymes during D N A repair and might supply us with some insights in their roles during D N A repair. It will be most interesting to search for deficiencies of their genes in one of the hereditary diseases. The cause for the reduced consumption of N A D + after D N A damage in cells from patients with Fanconi's anemia will be studied with the cDNA clone for ADPRT.

Acknowledgements We are very grateful to many colleagues who helped us or gave us advice at various stages of this work: Drs. Arnold, Beyreuther, de Groot, Heinze, Hermans, Herrmann, Jungblut, Kühne, Mertz, Reinke, Shall, Shekel, Stöffler-Meilicke, Traub, Winnacker, Wintersberger, Wittmann-Liebold. We thank Drs. Peter Nielson and J. Skowronski for supplying us with additional human c D N A libraries. This work was generously supported by the Deutsche Forschungsgemeinschaft, by the Bundesministerium für Wissenschaft und Forschung der Republik Österreich, by the Thyssen-Stiftung and by EMBL.

References: 1. Shall, S. 1985. In: ADP-Ribosylation of Proteins (F.R. Althaus, H. Hilz and S. Shall, eds.). Springer, Berlin, Heidelberg, New York, Tokyo, pp. 9-29. 2. Skidmore, C. J., Davies, M.I., Goodwin, P.M., Halldorson, H., Lewis, P. J., Shall, S. and Zia'ee, A.A. 1979. Eur. J. Biochem. 101, 135-142. 3. Juarez-Salinas, H., Sims, H.L. and Jacobson, M.K. 1979. Nature 282, 740-741. 4. Klocker, H., Auer, Β., Hirsch-Kauffmann, M., Altmann, H., Burtscher, H. J. and Schweiger, M. 1983. EMBO J. 2, 303-307. 5. Ohashi, Y., Ueda, K., Hayaishi, O., Ikai, K. and Niwa, O. 1984. Proc. Natl. Acad. Sci. U.S.A. 81, 7132-7136. 6. Kun, E., Kristen, E., Milo, G.E., Kurian, P. and Kumari, H.L. 1983. Proc. Natl. Acad. Sci, U.S.A. 80, 7219-7223. 7. Borek, C., Morgan, W.F., Ong, A. and Cleaver, J.E. 1984. Proc. Natl. Acad. Sci. U.S.A. 81, 243-247. 8. Ogata, N„ Ueda, K„ Kawaichi, M. and Hayaishi, O. 1981. J. Biol. Chem. 256, 4135-4137. 9. Jongstra-Bilen, J., Ittel, Μ. E., Niedergang, C., Vosberg, H.P. and Mandel, P. 1983. Eur. J. Biochem. 136, 391-396. 10. Ferro, A.M., Higgins, N.P. and Olivera, Β.M. 1983. J. Biol. Chem. 258, 6000-6003. 11. Darby, M.K., Schmitt, Β., Jongstra-Bilen, J. and Vosberg, H.P. 1985, EMBO J. 4, 2129-2134. 12. Creissen, D. and Shall, S. 1982. Nature 296, 271-272. 13. Yoshihara, K., Itaya, Α., Tanaka, Y., Ohashi, Y., Ito, K., Teraoka, H., Tsukada, K., Matsukage, A. and Kamiya, T. 1985. In: ADP-Ribosylation of Proteins (F.R. Althaus, H. Hilz, S. Shall, eds.). Springer, Berlin, Heidelberg, New York, Tokyo, pp. 82-92. 14. Burtscher, H.J., Auer, Β., Klocker, Η., Schweiger, M. and Hirsch-Kauflfmann, M. 1986. Anal. Biochem. 152, 285-290. 15. Burtscher, H. J., Schneider, R., Klocker, Hi, Auer, Β., Hirsch-Kauflfmann, M. and Schweiger, M. 1987. J. Comp. Physiol. Β.157, 567-572.

D N A Repair in Human Cells

387

16. Burtscher, H.J., Schneider, R., Klocker, H., Auer, B., Hirsch-Kauffmann, M. and Schweiger, M., 1987. Biochem. J. 248, 859-864. 17. Schweiger, M., Auer, B., Burtscher, H.J., Hirsch-Kauffmann, M., Klocker, H. and Schneider, R. 1986. Eur. J. Biochem. 165, 235-242. 18. Schneider, R., Auer, B., Kühne, C., Herzog, H., Klocker, H., Burtscher, H.J., HirschKauffmann, M., Wintersberger, U. and Schweiger, M. 1987. Eur. J. Cell Biol. 44, 302-307. 19. Alkhatib, H. M., Chen, D., Cherney, B., Bhatia, K., Notario, V., Giri, C., Stein, G., Slattery, E., Roeder, R . G . and Smulson, M . E . 1987. Proc. Natl. Acad. Sci. USA 84, 1224-1228. 20. Suzuki, H., Uchida, Κ., Shima, Η., Sato, T. and Miwa, M. 1987. Biochem. Biophys. Res. Comm. 146, 403-409. 21. Vosberg, H.P. 1985. Curr. Topics Microbiol, and Imm. 114, 19-102. 22. Klocker, H., Schneider, R., Burtscher, H. J., Auer, B., Hirsch-Kauffmann, M. and Schweiger M. 1985. Eur. J. Cell Biol. 39, 346-351. 23. Traub, P., Zillig, W., Milette, R.L. and Schweiger, M. 1966. Z. f. Physiol. Chem. 343, 261 - 2 7 5 . 24. Blackburn, P., Glynn, W. and Moore, S. 1977. J. Biol. Chem. 252, 5904-5910. 25. Shapiro, R. and Vallee, B.L. 1987. Proc. Natl. Acad. Sci. U.S.A. 84, 2238-2241. 26. Weiner, H.L., Weiner, L.H. and Swain, J.L. 1987. Science 237, 280-282.

Acyl ~ Phosphate Intermediates in Oxidative DNA Sugar Damage by Antibiotics Irving H. Goldberg

We are indebted to the genius of Fritz Lipmann for our current understanding of the role of energy-rich intermediates in biological processes, in particular those involving biosynthetic reactions (1). I believe that he would have found it "amusing" to learn that ~ P containing compounds also serve as energy-rich intermediates in DNA damage reactions. Recently, in studying the mechanisms of oxidative DNA sugar damage (and their biological consequences) by certain antitumor antibiotics, namely neocarzinostatin, we found evidence for the formation of 3'-formyl~ phosphate-ended DNA and possibly 5'-acyl ~ phosphatecontaining DNA as high-energy DNA sugar damage intermediates. In this report I shall briefly review our current understanding of the molecular mechanisms responsible for the formation of these and other novel oxidative DNA sugar damage products. Whereas many antitumor agents, such as ionizing radiation and bleomycin, damage DNA sugar indirectly via generation of "reactive" oxygen, generally as the hydroxyl radical, a thiol-activated form of neocarzinostatin (NCS) itself acts as a targeted radical that directly attacks DNA deoxyribose of mainly thymidylate residues to produce specific, novel damage products (see refs. 2 and 3 and refs. therein). Recently, antitumor antibiotics (calichemicins (4) and esperamicins (5)) have been described that may be viewed as second generation NCS-like agents, because of structural and mechanistic similarities. These agents are the most potent antitumor drugs presently known. NCS consists of a labile, non-protein chromophore (NCS-Chrom) that is tightly and specifically bound to its apoprotein. All the biological activity resides in the nonprotein chromophore; the apoprotein acts to protect NCS-Chrom and to serve as its carrier. NCS-Chrom consists of three subunits: a 5-methyl-7-methoxynaphthoate, a 2,6-dideoxy-2-methylamino-galactose moiety, and an interconnecting highly unsaturated C 12 -subunit containing a novel, highly strained bicylo(7,3,0)dodecadiyne system, bearing a cyclic carbonate moiety and an epoxide (Figure 1). Based on hydrodynamic, spectroscopic, and electric dichroism measurements, a model for the reversible binding of NCS-Chrom to DNA has been proposed in which the naphthoate moiety intercalates between adjacent DNA base pairs and the positively-charged aminosugar is attracted by negative electrostatic forces to the minor groove of the DNA. These types of interactions serve to position the reactive diyne-containing ring system close to the C-5' of deoxyribose of mainly thymidylate residues. Thiols active NCS-Chrom, in vitro and in vivo, by addition to the C 12 -subunit of the chromophore, to convert the DNA-bound drug into a presumably radical species

390

Irving H.Goldberg

Acyl ~ Phosphate Intermediates in Oxidative D N A Sugar Damage by Antibiotics

391

that abstracts a hydrogen atom from C-5' of deoxyribose, generating a carboncentered radical (Scheme 1, pathway A). Under aerobic conditions, dioxygen adds to the 5'-carbon-centered radical (1) (pathway B), as shown by 1 8 0 studies, to form a peroxyl radical (2) that degrades to form D N A strand breaks with 5'-terminal nucleoside aldehyde (3). These lesions, breaks with a phosphate at the 3'-end and nucleoside (thymidine) 5'-aldehyde at the 5'-end, constitute over 80% of the D N A strand breakage under aerobic conditions. In the absence of 0 2 , the carbon-centered radical on the deoxyribose may instead react with the NCS-Chrom containing the abstracted hydrogen atom (pathway C) to form a novel, stable covalent chromophore-DNA sugar adduct (4). This analysis, involving a common D N A damage intermediate, is consistent with studies showing a competition between stable chromophore-DNA adduct formation under anoxia and D N A strand breakage and base release in the presence of dioxygen. Evidence for the existence of a labile chromophore-DNA adduct (5), with properties suggesting that it functions as an intermediate in D N A breakage, suggests that it may be formed by an alternative pathway (D, E, or F) in which a radical form of the drug adds to the peroxyl radical on the D N A sugar (pathway F), or the drug itself forms a peroxyl radical species that adds to the carbon-centered radical at C-5' of deoxyribose (pathways D and E). In Scheme 1 the 5'-carbon of deoxyribose is the predominant site of attack by the activated drug. It is consistent with the D N A sugar damage products formed and with the results of 1 8 0 and hydrogen atom abstraction studies. It should be noted, however, that C-1' appears to be the primary site of hydrogen atom abstraction when radiolytically-activated NCS reacts with D N A to form alkali-labile, abasic lesions. A similar mechanism may be involved in the formation of the mutagenic, abasic sites found at the C-residue in the sequence A G C (see ref. 3). It appears that the A G C sequence generates a structure that modifies the interaction with NCSChrom so as to alter the attack site specificity, resulting in hydrogen atom abstraction from a different carbon in deoxyribose and in a different D N A sugar lesion. Nucleoside 5'-aldehyde is the predominant product at the 5'-end of a D N A strand break under the conditions ordinarily used to study NCS-induced D N A damage in vitro (see ref. 6 and references therein). On D N A sequencing gels, however, a band of 3 2 P radioactivity (from a D N A restriction fragment labeled with 3 2 P 0 4 at its 3'end) can also be seen, with a mobility somewhat faster than that of the nucleoside 5'aldehyde ended D N A fragment (but not as fast as the Maxam-Gilbert procedure produced marker that has a phosphate at its end and is also one nucleotide shorter). The intensity of this band varies with the exact incubation conditions. Under ordinary conditions (10 mM thiol, I m M EDTA), this band is minimal. When EDTA is omitted entirely or added immediately after the cleavage reaction, however, the band becomes as intense as that for the nucleoside 5'-aldehyde ended fragment, unless the thiol is increased greatly (to 150 mM). The mobility of this band is identical to that of a D N A fragment bearing nucleoside 5'-aldehyde that has been mildly oxidized by sodium hypoiodite to its 5'-carboxylic acid derivative, suggesting that the unknown band might also have a terminal 5'-carboxylic acid (alkaline phosphatase studies show it does not have a phosphate at the end). Other

Irving H.Goldberg

392

H0-^=0 9 HO-CH

H(H¡=0 Β

9 G" +

0 ho-N)

o HC-P=0

Oo

H0-P=0 8 (2)

(3)

I Η0-^=0 o H0-f>=0

0 H0-P=0

10 Scheme 2

transition metal chelators, such as deferoxamine or DETAPAC are similar to EDTA in their effects. If the unknown band is a 5'-carboxylic acid ended DNA fragment, this may have important implications for the mechanism of the oxidative strand cleavage, suggesting the involvement of a disproportionation reaction in its formation and that of the 5'-aldehyde (Scheme 2). In fact, there is precedence in the literature for the bimolecular decay of secondary peroxyl radicals via a short-lived tetroxide intermediate (6) by the Russell mechanism (Scheme 2). Reaction 1 in Scheme 2 would give rise to equal amounts of the hydroxyl (7) and the carbonyl (8) derivatives. The former (as the hemiacetal) would break down to give a strand break with nucleoside 5'-aldehyde (?) at the 5'-end; the latter (a 5'-acyl phosphate) would breakdown to form a strand break with a nucleoside 5'-carboxylic acid (10) at the 5'end. It is also possible, as shown below for a formyl phosphate derivative, that the energy-rich acyl moiety can be transferred to neighboring nucleophilic molecules. It must be pointed out, however, that it has not yet been conclusively shown that the nucleoside 5'-carboxylic formation does not result from oxidation of the 5'-aldehyde after the cleavage reaction, without the involvement of an acyl intermediate, although the result mentioned above with the delayed addition of EDTA is against

Acyl~ Phosphate Intermediates in Oxidative D N A Sugar Damage by Antibiotics

393

this possibility. If the unknown band on the gel in a 5'-carboxylic acid ended DNA fragment, compound (5) in Scheme 1 may have to be changed to a tetroxide intermediate (6). It is important to point out that if a transition metal is involved in this reaction, as is implied by the results with chelators, it appears not to be required in the initial DNA damaging reaction, since strand breaks and total damage products are unaltered by metal removal. Rather, a metal, possibly iron, may determine the ultimate distribution of the products, as indicated above. Until recently, the molecular mechanism responsible for the 10-15% of the strand breaks that lacked nucleoside 5'-aldehyde at the 5'-end and, instead, possessed a phosphate moiety at each end of the break, i.e., a gap from which both base and sugar were eliminated, was obscure, although there was evidence that formate formation from C-5' was involved (7). The finding that nitroaromatic radiation sensitizers, such as misonidazole [l-(2-hydroxy-3-methoxypropyl)-2-nitroimidazole], act as dioxygen substitutes in the NCS-induced DNA damage reaction and generate predominantly breaks with phosphate at each end (8) enabled us to characterize the chemistry of the lesion at C-5' and to propose a mechanism for the formation of this type of strand break in both types of reactions (9). When misonidazole substitutes for dioxygen in the NCS-DNA reaction, the attack site specificity, both in terms of base residue attacked (mainly thymidylate) and deoxyribose carbon (C-5') from which a hydrogen atom is abstracted, remain the same as in the aerobic reaction. The amount of DNA damage produced varies with the one-electron reduction potential of the nitroaromatic compound, such that more electron-affinic compounds are more effective. Under anaerobic conditions with misonidazole, NCS causes in addition to the release of thymine, the release of substantial amounts of a 5'[ 3 H]-labeled sugar fragment (from 5'[ 3 H]thymidinelabeled DNA) that elutes from HPLC at 4 min (10). This substance is uncharged but when treated with either acid or alkali is converted into formic acid. We have identified the formic acid precursor as formyl-Tris (Tris being an amino group containing molecule used as the buffer in the reaction) (9). Similarly, when hydroxylamine ("the Lipmann reagent") substitutes for Tris, formyl hydroxamate is formed. The amount of the latter formed is comparable to the spontaneously released thymine. During the reaction, 14C-labeled misonidazole is reduced to form a glutathione adduct. Under anaerobic conditions in the presence of misonidazole a reactive form of formate from C-5' of deoxyribose of thymidylate appears to be generated, and the formate can be transferred to available nucleophiles. These results, plus the evidence that misonidazole reduction is DNA-dependent, suggest a mechanism (Scheme 3) in which the carbon-centered radical formed at C-5' (1) by hydrogen atom abstraction by thiol-activated NCS reacts anaerobically with misonidazole to form a nitroxide radical intermediate (2) (reaction 1), which fragments (reaction 2) to produce an oxyradical at C-5' (3). ^-scission (reaction 3) results in cleavage between C-5' and C-4' with the generation of 3'-formyl phosphate-ended DNA, a high-energy form of formate. The latter either spontaneously hydrolyzes, releasing formate and creating a 3'-phosphate-end, or transfers the formyl moiety to available nucleophiles. A similar mechanism, involving dioxygen addition, is probably responsible for the 10-15% DNA gap formation in the aerobic reaction.

394

Irving H.Goldberg

φ

Φ.

?

Η-C—θ4-Ν—R

0

Φ

.0

ο

H-C-0* — — >

φ 1

φ

2

ir

φ 3

OH

1(3)

ÇH2-CHCH20CH3 Rno2=LJ Β

N

°

2

Ν

, ® Υ

I

+

ρ H-C-0H 6


• > · f f f t ' » ' y f i

4

6

8

10

12

14

INCUBATION TIME IN HOUR Figure 4 Growth of cAMP-deficient mutants CA7902, CA7902 (Rtsl), and CA7902 (R100) in the presence or absence of cAMR CA7902 with or without R factor was inoculated into TSB (Trypticase Soy Broth) and incubated at 32 °C or 42 °C in the presence or absence of exogenous cAMP. Cell growth was monitored periodically by measuring the turbidity in a Klett-Summerson colorimeter with a red filter. Symbols: O E. coli CA7902; · E. coli CA7902 harboring Rtsl; Δ E. coli CA7902 (R100). (A) 42°C with OraM cAMP; (B) 42°C with 1 mM cAMP.

order of molecular weights. This fragment mixture was allowed to randomly relígate and the resulting mixture was used to transform E. coli. The bacteria harboring Rtsl "mini" plasmid was then selected by Kanamycin resistance phenotype. We have identified a number of plasmids which retained the tsg characteristics. The smallest of them was pFY556 which contained E- (21.5 kb), G- (11.3 kb) and D- (25.7 kb) fragments (12). Survey of tsg miniplasmids suggested that G-fragment is essential for the tsg phenotype [Table 2, (12)]. We first thought that the G-fragment contained the tsg gene. An unexpected observation, however, was made recently which suggests that the role of G-fragment in the expression of tsg is that of a "gene activator" rather than itself containing tsg. This finding came from studies of an Rtsl D N A fragment created by Sail digestion. Sail digestion of Rtsl DNA produced a 3.65 kb fragment which contained the Kanamycin resistance gene. A surprising finding was that this fragment contained the tsg gene also (13). Thus, pUC19 (a vector) containing this 3.65 kb fragment conferred tsg phenotype to host bacteria. This pUC19 derivative with tsg was named p F A N l O l . It should be pointed out that the tsg phenotype of pFANlOl is not caused by a simple insertion of any D N A of this size into pUC19. Thus, myc gene placed in pUC19 did not exhibit tsg effect. These findings were somewhat surprising

402 Table 2

A. Kaji et al. Fragment composition of the BamHI derived Rtsl miniplasmids BamHI Fragments, Mdal

Plasmid pAK8 pFY505 pFY551 pFY553 pFY556 pFY545 pFY560 pFY601 pFY603

tsg + + + + +

1.1

3.7

4.3

6.0

+ ?

+ +

+

+

+ +

(G) 8.0

(E) 14.1

(D) 18.6

+

+ + + +

—

—

—

—

—

—

+

—

—

—

-

—

+ + + + +

—

—

—

-

—

+

—

—

—

—

—

—

+

—

—

-

-

-

-

-

+ + +

+ +

+ + +

+ + + + +

22.9

+ —

+ —

— — —

Intal Μτ

78.7 54.7 63.6 45.0 40.7 32.7 33.8 37.0 32.7

Figure 5 Tsg phenotype of various plasmids in E. coli 20SO incubated at 32 °C or 42 °C. Optical density was measured at 540nm. (Α) Δ ; pFANlOl. (B) • ; pTK301 (pUC19 + BamHI Km' fragment of Rtsl (E-fragment)). *; pTK401 (pUC19 + 2.8 kb myc gene). (C) V ; pFY556. (D) · ; pUC19, χ ; E. coli 20SO having no plasmids. — Growth at 32°C, - -- growth at 42°C.

403

Rtsl: A Multiphenotypic, Unusual Temperature

Bom

HI

Sail

Nrul 1 I

5

Pvul Soli Sali Nrul \ / ^ if I

IO κ * kmr (3.65 Kb)

NruI

Eco RI Sail Eco RI 1 EcoRI (?) EcuRK?)^ /.Bom HI Τ

15

-1 τ

20 21.5

Ikb

Figure 6 A restriction enzyme map of the BamHI Km' fragment (Ε-fragment) of Rtsl. EcoRI site with the question mark indicates that there is one EcoRI site at either one of these positions.

since the Bam HI fragment of Rtsl DNA encoding Kanamycin resistance (Efragment) does not give tsg phenotype in this assay (Figure 5). As can be seen from the restriction map of Ε-fragment (Figure 6), the E-fragment (not active as tsg) contains within its DNA the 3.65 kb Sail fragment which is active as tsg. On the basis of these observations, we postulate that the tsg gene resides in the 3.65 kb Sail fragment. The expression of this gene is regulated by a negative element which resides in the Bam HI 21.5 kb Ε-fragment. Therefore, the Efragment, when placed in a vector, does not express tsg. To make the situation further complicated, the Ε-fragment starts expressing its tsg gene when connected with the G-fragment (11.3 kb Bam HI fragment of Rtsl DNA). As can be seen from Figure 5, pFY556 which consists of G-, E- and D-fragments, gives the tsg phenotype. We therefore postulate that the G-fragment activates the dormant tsg in the E-fragment. In other words, the tsg gene is controlled by a negative element as well as by a positive element. If this is indeed the case, elucidation of the molecular mechanism underlying tsg gene expression may give us understanding of a new regulatory mechanism of gene expression which has hitherto not been known. It should be pointed out that this is only a preliminary hypothesis and future experiments will prove or disprove these predictions.

Temperature Dependent Instability of Rtsl In addition to the three temperature dependent phenotypes discussed above, Rtsl exhibits yet another intriguing phenotype, i.e. temperature dependent instability (tdi). As shown in Table 3, during prolonged incubation at 42 °C of low density cultures of cell harboring Rtsl, large number of R-segregants accumulate (14). Our original belief that this was mostly due to overgrowth of R-segregants at 42 °C was disproven by the finding that a 5.6 MDa Bam HI fragment of Rtsl (Hfragment) confers the tdi characteristic to vector plasmids such as pBR322 (7) without tsg. pFK896 is such a plasmid and is unstable at 42 °C (7) (Figure 7). Since pBR322 is a multicopy plasmid (15), it is conceivable that the loss of this plasmid may be gradual. We thought that we would observe cells harboring lower copy numbers of pFK896 as an intermediate state. Contrary to this expectation, it was found that when cells lose pFK896, they lose it very abruptly and no "intermediate" state exists (data not shown).

A. Kaji et al.

404 Table 3 Culture

Loss of R factors from host cells Incubation temperature

Incubation period

Viable count

(°C)

(h)

(cells/ml)

Percent colonies sensitive to kanamycin or ampicillin (%)

27 42

0 24 24

0.42 x 103 2.2 χ 109 5.8 x 108

0 0 0

27 42

0 24 24

0.98 x 103 1.3 x 109 2.5 x 10 4

0 0 77

R28K

Rtsl

Cultures of E. coli 2 0 S 0 / R t s l and R28K were diluted to 0.5 to 1 x IO3 cells per ml and incubated in trypticase soy broth at 27 °C or 42 °C for 24 h without shaking. Before and after incubation samples were plated and the resulting clones were tested for resistance to 25 μg/ml kanamycin or 20 μg/ml of ampicillin (Bristol Labs).

To elucidate the mechanism of tdi, we first attempted to localize the tdi gene in the Rtsl D N A . After series of restriction enzyme digestions and functional tests, we finally localized tdi to a 786 bp DNA fragment with Ava I and PvuII ends. We now sequenced this 786 bp fragment (data not shown) and called this fragment as b'. The sequence of b' showed that there are several open reading frames and one χ sequence. Deletion studies of b' fragment together with functional tests revealed that three regions of b' are essential for the expression of tdi. They are the χ sequence, open reading frame I (ORFI) coding for a protein consisting of 140 amino acids, and the tail region consisting of 156 bp nucleotides. Furthermore, we identified the protein apparently coded by O R F I in E. coli mini cells harboring pFK896 or other vector plasmids containing b' fragment. With the use of site-directed mutagenesis, we have inserted a C residue in the middle of ORFI. This frame shift mutation abolished the tdi activity of b'. Despite these detailed sequence information regarding tdi, the exact mechanism of tdi expression remains unknown. Survey of the literature (16-23) indicates that a specific region called par is essential for proper partitioning of stringent type plasmids and deletion of par causes plasmid instability. For multicopy plasmids such as pFK896, formation, of multimers has been suggested as a mechanism for plasmid loss. It is conceivable that if plasmids form one large multimer, one of the daughter cells would not receive the plasmid during the cell division, leading to loss of plasmid in that daughter cell. It should be pointed out that the 786 bp D N A of tdi was derived from Rtsl which is a stringent type plasmid. Furthermore, the 786 pb D N A , when inserted into pSClOl, a stringent type plasmid, causes instability at 42 °C but not at 32°C. In addition, expression of tdi is dependent only on RecA. On the other hand, this phenotype is not dependent on the RecBCD pathway (data not shown). If the χ region is involved in the expression of tdi as suggested by our deletion studies of b' fragment, it would be through a hitherto unknown new mechanism which involves only RecA rather than the generally accepted mechanism dependent on Ree A Β C D (24,25). To the best of our knowledge, although proteins necessary for plasmid

Rtsl: A Multiphenotypic, Unusual Temperature

405

Figure 7 Time course of loss of the idi plasmids. Escherichia coli 20SO harboring the plasmids were inoculted at a density of 2 x l 0 3 cells/ml in Trypticase Soy Broth. Samples were taken at various times, and antibiotic (ampicillin or kanamycin)-resistance colonies and total number of colonies were counted. pFK896 (tdi-pBR322, Δ); pYH156 (tdi-pSC105, Á); pBR322 ( O ) ; pSC105 ( · ) .

stability have been postulated, no plasmid-encoded functions mediating instability have been reported. Therefore, if ORFI protein is required for instability as our data suggest, it will represent the first case where a plasmid coded enzyme participates in plasmid loss. Since instability due to tdi is expressed only at 42 °C, an interesting possibility exists that this enzyme is heat activated.

Epilogue We described four temperature dependent phenotypes exhibited by R t s l . Another phenotype, temperature sensitive conjugative transfer of Rtsl (6) has not been studied in detail. This phenotype is probably caused by slow Rtsl D N A synthesis at 42 °C since conjugative transfer of plasmid requires simultaneous plasmid D N A synthesis. Our original impression that all the thermosensitive phenotypes of Rtsl are regulated by one thermosensitive gene was quickly discarded because at least three phenotypes, tdi, tsg and T4-phage restriction are expressed by three completely independent thermosensitive genes situated in different portions of Rtsl D N A . A possibility exists that tsg and tsc (temperature sensitive CCC formation) may be related to each other because all tsg Rtsl mini-plasmid so far tested posessed tsc phenotype (12). In addition to elucidation of each phenotype at the molecular level, consideration of how such unusual temperature dependent phenotypes ended up in

406

A. Kaji et al.

one plasmid w o u l d be o f interest. H o w d o such multiphenotypic plasmids like R t s l confer u p o n host bacteria selective advantages other than kanamycin resistance? We feel that continued questioning of such biological problems, regardless of h o w trivial they m a y appear o n the surface, should contribute toward advancement of science in general, and this is the lesson that our beloved Dr. Lipmann taught us all.

References 1. Kaji, Α., A. Tanaka, K. Hsia, J. Sabran, T. Furuse, Y. Iwasaki and H. Park. 1985. In: Cellular Regulation and Malignant Growth, (S. Ebashi, ed.). Japan Sei. Soc. Press, Tokyo/Springer, Berlin, p. 427. 2. DiJoseph, C.G. and A. Kaji. 1974. J. Bacteriol. 120, 1364-1369. 3. Terawaki, Y., H. Takayasu and T. Akiba. 1967. J. Bacteriol. 94, 687-690. 4. DiJoseph, C.G., Μ. E. Bayer and A. Kaji. 1973. J. Bacteriol. 115, 399-410. 5. Ishaq, M. and A. Kaji. 1980. J. Biol. Chem. 255, 4040-4047. 6. Terawaki, Y., Y. Kakizawa, H. Takayasu and Y. Yoshikawa. 1968. Nature 219, 284-285. 7. Okawa, N., H. Yoshimoto and A. Kaji. 1985. Plasmid 13, 88-98. 8. Horii, Z.I. and A.J. Clark. 1973. J. Mol. Biol. 80, 327-344. 9. Dharmalingam, K. and E.B. Goldberg. 1976. Nature 260, 406-410. 10. Yamamoto, T. and A. Kaji. 1977. J. Bacteriol. 132, 90-99. 11. Yamamoto, T., T. Yokota and A. Kaji. 1977. J. Bacteriol. 132, 80-89. 12. Finver, S. and A. Kaji. unpublished data. 13. Okawa, N., M. Tanaka, S. Finver and A. Kaji. 1987. Biochem. Biophys. Res. Commun. 142, 1084-1088. 14. Dijoseph, C.G. and A. Kaji. 1974. Proc. Nat. Acad. Sci. U.S.A. 71, 2515-2519. 15. Clewell, D.B. 1972. J. Bacteriol. 110, 667-676. 16. Austin, S. and A. Abeles. 1983. J. Mol. Biol. 169, 373-387. 17. Gerdes, K., P.B. Rasmussen and S. Molin. 1986. Proc. Nat. Acad. Sci. U.S.A. 83, 3116-3120. 18. Miki, T., A.M. Easton and R.H. Rownd. 1980. J. Bacteriol. 141, 87-99. 19. Miller, C.A., W.T. Tucker and S.N. Cohen. 1983. Gene 24, 309-315. 20. Mori, H., A. Kondo, A. Ohshima, T. Ogura and S. Hiraga. 1986. J. Mol. Biol. 192, 1-15. 21. Cohen, A. and Clark, A.J. 1986. J. Bacteriol. 167, 327-335. 22. Hakkaart, M.J.J., P.J.M. van den Elzen, E. Veltkamp and H.J.J. Nijkamp. 1984. Cell. 36, 203-209. 23. Summers, D.K. and D.J. Sherratt. 1984. Cell. 36, 1097-1103. 24. Smith, G.R. 1983. Cell. 34, 709-710. 25. Biek, D.P. and S.N. Cohen. 1986. J. Bacteriol. 167, 594-603.

The Partial Dependency of Human Prostatic Growth Factor on Steroid Hormones in Stimulating Thymidine Incorporation into DNA Marion T. Hierowskif*, Jerry W. Sullivan

Michael W. McDonald,

Lily Dunn and

Introduction The androgen dependency of prostate cancer has been recognized since the pioneering work of Huggins almost five decades ago (1), and although exact mechanisms are uncertain, androgen appears to play a role in the etiology of benign prostatic hyperplasia. It appears that all prostate cells whether normal, malignant or undergoing benign prostatic hyperplasia, require androgen to maintain cellular content and functional activity. Recognition of the roles of 5a-reductase, dihydrotestosterone and androgen receptors has improved our understanding of the cellular mechanisms of androgen action, but the exact pathways through which androgen stimulates cell growth or division remain unknown. Work over the past 2 decades, much of it in tissue culture, has demonstrated a new class of trophic substances, distinct from steroid hormones, known as growth factors. Growth factors are naturally occurring proteins capable of stimulating cell growth and replication in tissue culture and, presumably, in vivo. Investigation of the possible role of growth factors in normal and abnormal prostate growth has been limited so far. Matuo et al. used heparin-Sepharose chromatography to partially purify growth factors from normal and cancerous rat prostate cell lines, and demonstrated a positive correlation between the metastatic potential of different clonal sublines and an increased ratio of growth factor with high affinity to heparin relative to growth factor with low affinity to heparin (2). Maehama et al. used ion exchange, reverse-phase and gel permeation chromatography to isolate a protein from rat prostate capable of stimulating incorporation of 3 H-thymidine into rat kidney cells grown in tissue culture (3). Two groups have isolated human prostatic growth factors from benign prostatic hyperplasia tissue (4,5) but their existence in prostate cancer remains unclear. Investigations involving another hormone-dependent neoplasm, breast cancer, have recently shown that 17/?-estradiol induces formation of growth factors in tissue cultures of MCF-7 breast cancer cells, and has suggested that these estrogeninduced growth factors may enhance the metastatic potential of MCF-7 cells in nude mice (6). * deceased during printing of this colume

408

Marion T. Hierowski et al.

We have used heparin-Sepharose chromatography to study normal prostate, benign prostatic hyperplasia and prostate cancer tissues in an attempt to further define the role of growth factors in the human prostate. To assess the possible interaction of androgen and growth factors, we have studied the ability of growth factors obtained from human prostate tissues to stimulate, in the presence and absence of steroid hormones, incorporation of 3 H-thymidine into DNA of two human prostate cancer cell lines, androgen unresponsive DU145 and androgen dependent LNCaP.

Materials and Methods Prostate tissue Four specimens of normal prostate were obtained at autopsy from young men who died of trauma, 8 specimens of benign prostatic hyperplasia and 5 specimens of prostate cancer were obtained at open surgical procedures or autopsy. All specimens were immediately placed in 0.9% NaCl in ice, weighed, examined histologically and stored in liquid nitrogen.

Preparation of cytosol Prostate tissues were homogenized in 10 volumes (v/w) of 10 mM Tris-HCl (pH 7.5) containing 0.25 M sucrose, 3 mM CaCl 2 and 2 mM phenylmethyl sulfonylfluoride. The homogenates were centrifuged at 1,500 χ g for 10 min, and the supernatants further centrifuged at 105,000 χ g, for 1 h. The resulting supernatants represented cytosols and were used as starting material for measurement of growth factor activity, relative content of low affinity growth factor activity and partial purification of growth factor.

Growth factor assay Growth factor activity was determined by an assay to measure incorporation of 3 Hthymidine into DNA of Swiss 3T3. 2 χ 104 cells were plated in 24 multiwell plates with 1 ml of RPMI 1640 containing 3% calf serum; 5 h later the medium was replaced by 1 ml of RPMI 1640 containing 1 % calf serum. After 48 h samples of 20 μΐ of cytosol or purified hPGF (ca. 12 ng of protein) were added. 24 h after the addition of the sample, 10μΐ o f 2 x l O ~ 5 M 3 H thymidine (Amersham solution 40 μΟ/ιηΙ) was added. After 3 h of incubation with 3 H thymidine the cells were washed once with phosphate-buffered saline and 3 times with 1 ml of cold 10% (w/v) trichloroacetic acid. DNA was solubilized with 1 ml of 0.5 Ν NaOH and 0.5 ml of the solution neutralized with 5 Ν HCl and mixed with 6 ml of ACS in scintillation fluid (Amersham). Radioactivity was counted with a Packard Tri-CarbR530 scintillation counter and all measurements done in duplicate. One unit of growth factor activity is the amount of growth factor equivalent to the activity of calf serum that can induce half-maximal incorporation of labelled thymidine into the DNA of Swiss 3T3 cells.

Human Prostatic Growth Factor

409

Measurement of total growth factor activity and relative content of low affinity prostatic growth factor Cytosol from each of the normal prostate, benign prostatic hyperplasia and prostate cancer tissues was adjusted with 4 M NaCl to a final concentration of 0.5 M and divided in 2 equal portions. One portion was mixed with an equal volume of heparin-Sepharose gel equilibrated with 10 mM Tris-HCl (pH 7.5) containing 0.5 M NaCl and the other portion with Sepharose C1-6B gel without heparin. Both portions were then incubated at 25 °C for 3 h with constant shaking and then centrifuged. The supernatant of each portion was subjected to growth factor activity assay. Growth factor activity in the supernatant obtained from cytosol mixed with plain Sepharose represented total growth factor activity, both high affinity to heparin and low affinity. Growth factor activity in the supernatant obtained from cytosol mixed with heparin-Sepharose represented low affinity growth factor, since high affinity to heparin growth factor adhered to the heparin-Sepharose gel and was removed at centrifugation. Relative content by the ratio of low affinity growth factor to total growth factor activity.

Partial purification of prostatic growth factor Heparin-Sepharose column chromatography was used to examine 1 specimen of prostate cancer, 1 specimen of benign prostatic hyperplasia and 3 specimens of normal prostate according to the methods of Matuo et al. and Nishi et al. (2,4). 250 ml of cytosol (protein content 0.6 mg/ml) were supplemented with 36 ml of 4 M NaCl to a final concentration of 0.5 M and applied to a 1.0 χ 11.0 cm column of heparin-Sepharose CL-6B (Pharmacia Fine Chemicals, AB, Uppsala, Sweden). After washing, a gradient of 250 ml of 0.5-3.0 M NaCl was applied. Fraction numbers 1 - 1 0 were 23 ml in volume and fractions 11 - 78; 5 ml in volume. Fractions 43-60 were combined and subjected to a second heparin-Sepharose chromatography. High affinity fractions from the second chromatography represented partially purified high affinity prostatic growth factor.

Combined action of high affinity growth factor and steroid hormones 2 x 104 cells from Swiss/3T3, DU145 and LNCaP cell lines were plated in RPMI containing 1 % calf serum for the first 5 h followed by 1 % charcoal-treated calf serum (7). After 48 h 20 μΐ samples of partially purified growth factor from either prostate cancer or benign prostatic hyperplasia, representing 12 ng of protein, were added, together with either 10" 8 M 5-a-dihydrotestosterone or 10 ~9 M 17/?estradiol. 2 χ 10" 5 M 3 H thymidine was added 24 h later and after 3 h of incubation DNA was solubilized and radioactivity counted as described in the description of the growth factor assay.

410

Marion T. Hierowski et al.

Determination of physicochemical properties of partially purified growth factor Isoelectric points were determined by subjecting 25 μΐ containing 2.5 μg protein samples of partially purified growth factor to thin layer analytical isoelectric focusing with Ampholine carrier ampholytes in BioRad equipment. Molecular weight of partially purified growth factor was estimated using a SDS-PAGE.

Cell lines and culture Albino mouse embryonic fibroblasts (Swiss/3T3) and human prostatic adenocarcinoma DU145 8 cell lines were purchased from American Type Culture Collection (Rockville, MD). The LNCaP 9 , 1 0 cell line was donated by Dr. Horoszewicz of Roswell Park Memorial Institute, Buffalo, NY. The three cell lines were grown in RPMI 1640 (Flow Lab, Scotland) supplemented with 10% (v/v) inactivated calf serum (Gibco, Scotland), 50 μg/ml penicillin, 50 μg/ml streptomycin, and 2.2 μιηοΐ/ml L-glutamine. Costar plastic culture flasks (surface area 75 cm 2 , No. 3275) were used to maintain the cells. Cultures were incubated in 5% carbon dioxide, 95% air atmosphere at 37°C.

Statistics Statistical analysis was done using the unpaired Student t-test planned to accept groups of unequal size.

Results Total growth factor activity and relative content of low affinity prostatic growth factor Total growth factor activity and relative content of low affinity prostatic growth factor in cytosols of normal prostate, benign prostatic hyperplasia and prostate cancer are shown in Table 1. Both prostate cancer and benign prostatic hyperplasia had significantly greater amounts of growth factor activity than normal prostate (P i í

* i

H ^ T y f f f ^ f I r H f i l l i " 4 ^ f ^ í H ^ f f ï ^ ' H i

Figure 4 Electron micrograph of a two-dimensional array of 70S ribosomes from Bacillus stearothermophilus. An optical diffraction pattern is inserted.

Figure 5 Computer graphic display of the reconstructed model of 70S ribosomes stained with gold-thioglucose. The small and large ribosomal subunits are marked by S and L, respectively.

Architecture of Ribosomal Particles

487

Figure 6 Outline of a 20 Â thick section in the middle of the reconstructed 70S model (Figure 5). T: part of the tunnel.

only partially resolved, and evidence for its existence is obtained by investigating sections through reconstructed particles (Figure 6). The two ribosomal subunits are fairly well separated. Only the two ends of the small subunit are in contact with the large one. The overall shapes of both subunits have been compared with models which have previously been suggested for these particles (4,5). There is a good agreement between the 50S particle as seen in Figure 1 and the one within the 70S ribosome (Figure 5). In general, there is also a similarity between the model of the small subunit obtained by visualization of single particles and that revealed by our studies. However, isolated 30S particles seem to be shorter than the reconstructed ones within the 70S particles. As discussed above for the 50S ribosomal subunit, this may be a consequence of the contact of the isolated particles with the flat electron microscope grids, whereas particles within the crystalline arrays are held together by their interactions with the neighboring particles as well as by interparticle crystalline forces. In this way the conformation of the particles is conserved, and the influence of the flat grid is decreased. We have also reconstructed models from ribosomal arrays stained with uranyl acetate instead of with gold-thioglucose (13). The arrays used for these studies were large and well ordered. As mentioned above, uranyl acetate, in contrast to goldthioglucose, functions also as a positive stain, since it may interact with selected portions of the object. In the case of ribosomes the extent of interaction of uranyl acetate depends on the accessibility of R N A . As seen in Figure 7, regions which were stained by uranyl acetate are located on the large subunit in its surface area which faces the internal empty space. Furthermore, penetration of uranyl acetate to the region assigned as "collar's ridge" on the small subunit was also observed. In both cases the staining of these areas with uranyl acetate may stem either from the existence of exposed ribosomal r R N A regions and/or from the presence of m R N A and t R N A in these locations.

488

H. G. Wittmann and A.Yonath

Figure 7 Computer graphie display of the reconstructed model of 70S ribosomes stained with uranyl acetate. Areas of substantial interaction with the stain are marked by UA.

X-Ray Crystallographic Studies The best method for the determination of the spatial structure of biological macromolecules is X-ray crystallography. This technique has advanced rapidly, especially in collecting, processing and analyzing crystallographic data. Synchrotron radiation provides the most intense, intrinsically parallel X-ray beam, and it is essential for crystallographic data collection from single crystals of ribosomal particles because of their large unit cell dimensions. Three-dimensional crystals have been obtained from 70S ribosomes as well as from 30S and 50S ribosomal subunits of various bacteria, and in all cases only biologically active ribosomal particles could be crystallized. The best crystals for Xray structure analysis obtained so far are those of the 50S subunits from the thermophilic Bacillus stearothermophilus (14) and from the extremely halophilic Halobacterium marismortui (15).

50S Ribosomal Particles from Bacillus stearothermophilus Because ribosomes from eubacteria, to which Bacillus stearothermophilus belongs, disintegrate at high salt concentrations, volatile organic solvents had to be used as précipitants. A modified version of the standard hanging drop technique, using glass plates, was developed (16). The crystallization droplets, which contained either no precipitant or an extremely small quantity of it, were equilibrated with a reservoir containing the precipitant. Attempts to increase the size of these crystals by a drastic slow-down of the crystallization process failed, since ribosomal particles from this

Architecture of Ribosomal Particles

489

bacterium are unstable and may deteriorate before they are able to aggregate and form proper nucleation centers. Growing crystals from volatile organic solvents imposes many technical difficulties in manipulating, data collection and heavy-atom derivatization. In fact, any handling of the crystals, such as removing or reorienting the crystals, or replacing the growth medium by a different solution, is extremely difficult. Thus, for this system seeding was virtually impossible. However, reducing the size of the exposed

Figure 8 (a) Crystals of ribosomal 50S particles f r o m Bacillus stearothermophilus. (b) X-ray diffraction patterns of the m a j o r zones obtained with s y n c h r o t r o n radiation, (c) Electron micrographs of thin sections through the crystals. The section on the right shows the open packing of this crystal form, and is approximately perpendicular to that on the left. Optical diffraction patterns are inserted. For details see ref. (17).

490

H.G. Wittmann and A. Yonath

surface of the crystallizing droplet led to the production of large crystals (16). This was achieved by growing crystals directly in X-ray capillaries. Crystals from the 50S subunits of Bacillus stearothermophilus, which may reach a length of 2.0 mm and a cross-section of 0.4 mm, are obtained at 4°C from mixtures of methanol and ethylene glycol. Since most of the crystals grow with one of their faces adhering to the walls of the capillaries, it is possible to irradiate them without removing the original growth solution. Although many crystals grow with their long axes parallel to the capillary axis, a fair number grow in different directions. Thus, we could determine the unit cell constants and obtain diffraction patterns from all of the zones without manipulating the crystals. The crystals are loosely packed (Figure 8) in a unit cell of 360 χ 680 χ 920 Â. They diffract to 13-18 À resolution. They often last a few hours in the synchrotron beam, but the higher resolution terms are lost within 5 - 1 0 minutes. Although valuable crystallographic data have been obtained (14) from the crystals grown with mixtures of alcohols, there were technical difficulties in manipulating, data collection and heavy-atom derivatization. Therefore, we searched for an alternative, and were able to grow crystals of 50S subunits from Bacillus stearothermophilus ribosomes using polyethylene glycol. The particles in these crystals, in contrast to those grown from alcohols, are tightly packed. Crystallographic measurements show that the new crystals diffract to better than 14 Â, and periodic spacings of 260 Â, 320 À and 700 Â were calculated from their diffraction patterns. Since these crystals can be much better handled than those grown from alcohols, they are considerably more suitable for crystallographic studies. The process of crystal growth is initiated by nucleation. Although many biological molecules and complexes have been crystallized, currently little is known about the mechanism of nucleation. Most of the data available concerning the process of nucleation of crystals of biological systems are based on rather indirect evidence, such as monitoring aggregation under crystallization conditions by scattering techniques (18). Because ribosomal particles are large enough to be detected by electron microscopy, crystals of ribosomal particles provide an excellent system for direct investigation of nucleation. The crystallization process was examined by electron microscopy. It was found that the first step in crystal growth is unspecific aggregation, and that nucleation starts by a rearrangement within the aggregates (19).

50S Ribosomal Particles from Halobacterium marismortui In contrast to ribosomal particles from eubacteria, those from halobacteria are structurally stable and functionally active at high salt concentrations. Crystals from these bacteria grow under conditions which mimic, to some extent, the natural environment within the bacteria. For obtaining large and ordered crystals, advantage has been taken of the delicate equilibrium of mono- and divalent ions needed for the growth of halobacteria as well as of the major role played by the Mg 2 +

Architecture of Ribosomal Particles

Figure 9 (a) Crystals of ribosomal 50S particles from Halobacterium marismortui. (b) X-ray diffraction pattern obtained at — 180°C with synchrotron radiation.

491

Bar: 0.2 mm.

concentration in crystallization of ribosomal particles. It was found earlier that three-dimensional crystals of 50S ribosomal subunits from Bacillus stearothermophilus grow in relatively low Mg 2 + concentration, whereas the production of twodimensional arrays requires a high Mg 2 + concentration, at which growth of threedimensional crystals is prohibited. Similarly, for spontaneous crystal growth of 50S subunits from Halobacterium marismortui, the lower the Mg 2 + concentration is, the thicker the crystals are. With these points in mind, a variation of the standard seeding procedure has been developed. Thin crystals of the 50S subunits from Halobacterium marismortui, grown spontaneously under the lowest possible Mg 2 + concentration, are transferred to mixtures in which the M g 2 + concentration is so low that the transferred crystals dissolve, but after several days new crystals can be observed. These are well ordered and 10-20 fold thicker than the original seeds. Orthorhombic crystals of the 50S subunits from Halobacterium marismortui grow as fragile thin plates with a size of 0.6 χ 0.6 χ 0.2 mm (Figure 9). They diffract to a resolution of 5.5 Â (15), and have relatively small and compact unit cells of a = 214 Â, b = 300 Â, c = 590 Á, in contrast to the open structure of the large crystals of Bacillus stearothermophilus. Although between — 2 °C and 4 °C up to 15 photographs can be taken from an individual crystal, the high resolution terms appear only on the first 2 - 3 X-ray diffraction patterns. Hence, under these conditions over 260 crystals were irradiated in order to obtain a complete data set. However, at cryo-temperature, i.e. — 180°C, irradiated crystals show hardly any radiation damage after days. Thus, for the first time, a full data set could be collected from a single crystal (unpublished results).

492

H. G. Wittmann and A. Yonath

Concluding Remarks As objects for crystallographic studies, ribosomal particles are of enormous size and have no internal symmetry. Furthermore, they are flexible and have a tendency to disintegrate. Therefore, crystallization of these particles appeared to be extremely difficult. However, as a result of a systematic exploration of numerous crystallization conditions and by development of innovative experimental techniques, crystals of ribosomes and their subunits could be obtained. They have been subjected to Xray structure analysis and diffract to 5.5 Â in the case of the 50S ribosomal particles from halobacteria. Furthermore, three-dimensional image reconstruction of crystalline arrays of ribosomal particles have revealed interesting features, such as a tunnel for the nascent polypeptide chain within the 50S ribosomal subunit and a wide space between the 30S and 50S subunits which is large enough to accommodate tRNAs, elongation factors and mRNA. It can be expected that our crystallographic studies, combined with the results from other structural and functional approaches, will eventually yield a model of the ribosome in sufficient detail to allow an understanding of the process of protein biosynthesis at a truly molecular level.

References 1. Wittmann, H . G . 1982. Ann. Rev. Biochem. 51, 155. 2. Noller, H. F. 1984. Ann. Rev. Biochem. 53, 119. 3. Wittmann-Liebold, B. 1986. In: Structure, Function and Genetics of Ribosomes (B. Hardesty and G. Kramer, eds). Springer, Berlin, Heidelberg, New York, p. 326. 4. Wittmann, H . G . 1983. Ann. Rev. Biochem. 52, 35. 5. Hardesty, B. and Kramer, G. (eds.). 1986. Structure, Function and Genetics of Ribosomes. Springer, Berlin, Heidelberg, New York. 6. Yonath, Α., Leonard, K . R . and Wittmann, H . G . 1987. Science 236, 813. 7. Mathews, B.W. 1968. J. Mol. Biol. 33, 491. 8. Malkin, L.I. and Rich, A. 1967. J. Mol. Biol. 26, 329. 9. Blobel, B. and Sabatini, D . D . 1970. J. Cell. Biol. 45, 130. 10. Smith, W.P., Tai, P.C. and Davis, B.D. 1978. Proc. Natl. Acad. Sci. USA 75, 5922. 11. Barnabeau, C. and Lake, J. A. 1982. Proc. Natl. Acad. Sci. USA 79, 3111. 12. Milligan, R. A. and Unwin, P.N.T. 1986. Nature 319, 693. 13. Arad, T., Piefke, J., Weinstein, S., Gewitz, H.S., Yonath, A. and Wittmann, H . G . 1987. Biochemie, 69, 1001. 14. Yonath, Α., Saper, M. Α., Makowski, I., Müssig, J., Piefke, J., Bartunik, H.D., Bartels, Κ. S. and Wittmann, H . G . 1986. J. Mol. Biol. 187, 633. 15. Makowski, I., Frolow, F., Saper, M. Α., Shoham, M., Wittmann, H . G . and Yonath, A. 1987. J. Mol. Biol. 193, 819. 16. Yonath, Α., Müssig, J. and Wittmann, H . G . 1982. J. Cell Biochem. 19, 145. 17. Yonath, Α., Leonard, K.R., Weinstein, S. and Wittmann, H . G . 1988. Cold Spring Harbor Symp., in press. 18. Kam, Z„ Shore, H.B. and Fehder, G. 1978. J. Mol. Biol. 123, 539. 19. Yonath, Α., Khavitch, G., Tesche, B., Müssig, J., Lorenz, S., Erdmann, V.A. and Wittmann, H . G . 1982. Biochem. Internat. 5, 629.

Initiation of Protein Synthesis: Early Participation and Recent Revisit Masayasu

Nomura

The study of protein synthesis was at its peak in the 1960's. Following the discovery of poly U-dependent polyphenylalanine synthesis by Nirenberg and Matthei (1), many biochemists were engaged in the elucidation of the genetic code and this historic task was essentially complete by 1966. At the same time, the biochemical mechanisms involved in protein synthesis were also studied extensively in many laboratories. Fritz Lipmann was one of the leading biochemists in this field. As early as 1941, Lipmann speculated on and discussed the mechanism of amino acid polymerization in his famous review (2). His fundamental contributions to the subject starting from the late 1950's through the 1960's, such as the identification of polypeptide chain elongation factors, are well known and will undoubtedly be recounted in this volume by many contributors. I would just like to note here that I, too, was one of many students who were inspired by reading his 1941 review. I read it in the early 1950's when I was just starting my research career as a student in Tokyo, working on the metabolic pathway involved in the degradation of tartaric acid by microorganisms. I still remember the excitement I felt in reading the review, and would like, with many other scientists, to share my appreciation of his great contributions to biochemistry. In this article, I would first like to recount the work we carried out on the initiation of protein synthesis some 20 years ago, when I had a close professional relationship with Fritz Lipmann in the field of protein synthesis. I shall then describe briefly some aspects of our current research which have a bearing on some of the problems remaining in connection with the initiation of protein synthesis.

Initiation of Protein Synthesis; 20 Years Ago Bacterial ribosomes consist of two non-identical 30S and 50S subunits, which form 70S ribosomes. Such a bipatite ribosome structure is found in all known living cells. In the late 1950's I was working in Jim Watson's laboratory, then at Harvard, as well as in the late S. Spiegelman's laboratory at the University of Illinois, when this bipartite subunit structure and the Mg 2 + -dependent dissociation and reassociation reaction (70S = 3 0 S ^ 5 0 S ) were just being established for the E. coli ribosome (3). I remember that we were curious about its meaning. The significance of the double helical structure of D N A was, of course, clarified by Watson and Crick several years earlier. Thus we wondered whether the unequal bipartite structure was necessary for ribosome function, or perhaps for the mechanism of ribosome synthesis. For example, it was formally possible to think about a mechanism in which the 30S subunits are assembled on the 50S subunits and vice versa. These kinds of questions

494

Masayasu Nomura

stayed with me and influenced my thinking on the initiation of protein synthesis when we (and other investigators) discovered in 1967 that initiation starts on the small subunits in vitro. There must have been similar influences on other investigators who considered and demonstrated cyclic dissociation and reassociation of subunits during ribosome function. In fact, at the time of these discoveries we (4,5), as well as others (e.g., ref. 6), thought we had found the reason for the existence of this bipartite structure, although it is still not entirely clear why the bipartite structure evolved as the only ribosome structure in all living cells (see below). Of course, people had been asking how ribosome functions are divided between the two subunits. Thus, it had been shown that some synthetic mRNAs, e.g., poly U, can bind to 30S subunits (and sometimes stimulate binding of the corresponding tRNA, e.g., phe-tRNA), but do not bind to 50S subunits (7-9). Similarly, at about the same time as the 30S subunit initiation model was proposed, Monro demonstrated that the 50S subunit alone can carry out the peptidyl transferase reaction (10). Nevertheless, these studies were carried out in a way analogous to the studies of active centers of enzymes without considering the significance of the dissociation and reassociation of ribosomes. Most people did not think about the possibility that initiation starts on the small subunits. Another factor which helped our discovery of the initiation on 30S subunits was simply an accidental one. In the 1960's, I was working on the reconstitution of ribosomes (as well as on the mode of action of colicins) at the University of Wisconsin. Having succeeded in "partial" reconstitution of 30S and 50S ribosomal subunits individually (11) and having begun to analyze the functional roles of protein and R N A components, as well as trying to achieve complete reconstitution, we were carrying out in vitro protein synthesis reactions using various proteindeficient ribosomal particles as well as reconstituted ribosomal subunits. For reconstitution studies, we used purified 30S and 50S subunits as controls; that is, our standard in vitro reaction mixtures contained 30S and 50S subunits rather than 70S ribosomes. In contrast, most workers in the field used 70S ribosomes. For example, two types of in vitro experiments used extensively for the determination of codon assignment at that time were: (a) the incorporation of radioactive amino acid into protein directed by synthetic polynucleotides with random or defined order, and (b) trinucleotide-dependent binding of aminoacyl-tRNA to ribosomes. These in vitro reactions were conveniently carried out using 70S ribosomes under optimized reaction conditions (usually in the presence of 10 to 20 mM M g 2 + which prevents 70S ribosome dissociation). Similarly, to fractionate chain elongation factors required for protein synthesis, people, such as those in Lipmann's laboratory, used a poly U-dependent polyphenylalanine synthesizing system, and there was no necessity to use purified 30S and 50S subunits. I suspect that this is probably part of the reason why people studying in vitro protein synthesis assumed that protein synthesis starts with 70S ribosomes; with synthetic polynucleotides, 70S ribosomes could attach to m R N A and initiate protein synthesis at random sites in vitro. In our case, we were eager to demonstrate some partial reactions carried out by the 30S subunits alone or by the 50S subunits alone that could be used to assess the functional capacity of reconstituted subunits. Thus, in early 1967, I was simply trying to devise specific initiation reactions using 30S subunits alone and synthetic

Initiation of Protein Synthesis

495

polynucleotides or A U G trinucleotide as m R N A . I knew that people had difficulty in demonstrating fMet-tRNA binding to 70S ribosomes with natural mRNAs, such as R N A phage R N A , as templates under conditions where synthetic polynucleotides or triplet A U G was shown to stimulate fMet-tRNA binding. Nevertheless, Charles Lowry, then one of my students, and I tried the binding reaction with 30S subunits alone and found that phage f2 R N A gave good stimulation of binding of fMet-tRNA to 30S subunits (4). We also found that, under the same experimental conditions, the efficient binding of fMet-tRNA directed by f2 R N A or random A U G polymer was observed only with 30S subunits, but not with a mixture of 30S plus 50S subunits or 70S ribosomes, although the apparent inhibitory action of 50S subunits was mostly eliminated by the addition of excess initiation factors (4,12). By contrast, a 30S-50S mixture or 70S ribosomes were more efficient than 30S subunits in the polynucleotide-directed binding of aminoacyl-tRNAs other than fMet-tRNA. From these experimental observations, we proposed that the initiation of protein synthesis takes place on the 30S subunit and that the 50S subunit joins only after the formation of a complex consisting of the 30S subunit, m R N A , and f M e t - t R N A (4,12). In the nucleic acid Gordon Conference of the same year, I gave a short talk on this subject. Fritz Lipmann was the chairman of the session. I remember that he did not seem to really be convinced of this new proposal, and I felt the necessity of doing more rigorous tests, just to be able to convince him. However, several other people thought that the new model was reasonable. For example, Robert Thach, who was then at Harvard University, had observed that although the AUG-dependent fMett R N A binding reaction (in the presence of initiation factor IF1 and IF2) itself had an optimum M g 2 + concentration of about l O m M , preincubation of his 70S ribosome preparations at a lower M g 2 + concentration (0.5 mM) caused a large stimulation of the subsequent binding reaction (13). Such observations could be best explained on the basis of dissociation of 70S ribosomes during preincubation caused by the lowered M g 2 + concentration (13). Binding of fMet-tRNA to 30S subunits directed by synthetic or natural m R N A s followed by joining of 50S subunits was quickly confirmed in several laboratories (14,15). Nevertheless, more experiments were needed to really establish that protein synthesis starts with 30S subunits rather than 70S ribosomes. Of course, as already mentioned, my laboratory was not the only laboratory which seriously considered the importance of ribosomal subunits in protein synthesis at that time. David Schlessinger and his coworkers at Washington University were studying polysomes in E. coli extracts prepared by supposedly gentle methods. They observed that the extracts contained only polysomes and free subunits, but not single 70S ribosomes, and suggested that single ribosomes do not exist in vivo (16), implying that protein synthesis starts with free ribosomal subunits. Although other laboratories observed the presence of single ribosomes in similar studies and disagreed with Schlessinger's conclusion (see e.g., ref. 17; for reviews, see 18,19), Schlessinger and his coworkers showed that the "native" 30S and 50S subunits found in crude extracts can form 70S ribosomes only in the presence of m R N A and t R N A and that m R N A containing the A U G sequence is more effective than poly U in promoting association (20).

496

Masayasu Nomura

Another laboratory which made an important contribution to this subject was that of Matthew Meselson at Harvard. Using heavy ribosomes labeled with heavy isotopes, Raymond Kaempher, Meselson and Herschel Raskas demonstrated that during growth E. coli ribosomes undergo subunit exchange (6), and that subunit exchange takes place in crude bacterial extracts only under conditions of protein synthesis (21). These experiments were entirely consistent with the 30S subunit initiation model, and gave strong support to the model. In addition, Kaempfer also demonstrated a similar subunit exchange in an eukaryotic system (22), suggesting that the basic concept of translation initiation on small ribosomal subunits is correct in eukaryotic systems too. Nevertheless, the possibility existed that the binding of fMet-tRNA takes place on 70S ribosomes rather than the 30S subunits and that the observed subunit exchange is coupled with some other step of protein synthesis. Therefore, Christine Guthrie, then a graduate student in my laboratory, and I carried out further experiments to test the 30S subunit initiation model. We used 70S ribosomes labelled with heavy isotopes and carried out tRNA binding experiments in the presence of excess light 50S subunits. A polynucleotide preparation containing A, U, G in random order was used as mRNA to direct binding of 14 C-Val-tRNA and 3 H-fMet-tRNA in the same reaction mixture. We found that after equilibrium density gradient centrifugation in CsCl, 14 C-Val-tRNA was at the density of the original heavy 70S ribosomes, whereas 3 H-fMet-tRNA was mostly at the density of hybrid ribosomes consisting of heavy 30S and light 50S subunits (23). From these and other results, we were able to conclude that the binding of fMet-tRNA to 70S ribosomes does not take place directly, but through the formation of 30S-mRNA-fMet-tRNA as an intermediate. In the original formulation of the model entertained by us, Meselson's group and David Schlessinger's group, it was assumed that only 30S and 50S subunits exist in the cellular pool, and that 70S ribosomes can be formed only after the formation of the 30S initiation complex on mRNA. In addition, as described above, it was shown that 30S subunits have an ability to form the initiation complex at the initiator AUG codon, but do not respond to other codons efficiently, whereas 70S ribosomes have the ability to respond to all other codons and bind corresponding aminoacyl tRNAs. Thus we argued that the bipartite structure is fundamentally important to prevent non-specific initiation of protein synthesis on mRNA (5,6). However, after considerable debate, it was subsequently concluded that free 70S ribosomes do exist in bacterial cells in vivo (17-19). In eukaryotic cells, the presence of large amounts of non-functioning 80S ribosomes is often observed unequivocally, and these 80S ribosomes are prevented from entering into translation by other mechanisms (e.g., in unfertilized eggs or cultured mammalian cells growing under suboptimal conditions). In fact, we could conceive (hypothetical) mechanisms allowing free 70S ribosomes to start only at correct initiation sites without dissociation into subunits. Therefore, it is still not entirely clear why Nature has selected the bipartite ribosome structure and mode of translation involving cyclic dissociation and reassociation. I have described above only one aspect of research on the initiation of protein synthesis in which I myself participated in the late 1960's. Of course, there were other important discoveries in the history of research related to this subject. The

Initiation of Protein Synthesis

497

most important was proving that protein synthesis in E. coli starts with fromylmethionine at the unique codon, A U G (or G U G ) , on m R N A . Starting from the discovery of the existence of fMet-tRNA by Marcker and Sanger in 1964 (24), several early workers made important contributions to establishing this conclusion (e.g. 25-27). The discovery of initiation factors essential for the initiation of translation of natural m R N A (28-30) was also important in revealing the complexity of the biochemistry of initiation of protein synthesis, and suggesting the possible presence of regulatory mechanisms acting on the initiation step in gene expression. I should also add that one of the initiation factors, IF3, was later demonstrated to be active in dissociating 70S ribosomes into subunits (31,32), thus explaining the earlier observations in our laboratory mentioned above; in the absence of initiation factors, 30S subunits were more active than 70S ribosomes and the addition of excess initiation factors abolished the apparent inhibition of the binding of fMet-tRNA by 50S subunits. Although our major effort at that time was on the mechanism of ribosome assembly both in vitro and in vivo, we did some more studies on the initiation of protein synthesis in connection with ribosome reconstitution, asking which ribosomal components were important in determining the specificity of initiation. By carrying out heterologous reconstitution using ribosomal components from E. coli and Bacillus stearothermophilus we obtained evidence indicating that the 16S r R N A is important in determining the specificity of initiation characteristic of the species. Hence, the suggestion was made that direct interaction of some parts of 16S r R N A with m R N A is involved in recognition of initiation signals on natural m R N A (33). Of course, we now know that the 3'-end of 16S r R N A interacts with a sequence ("Shine-Dalgarno sequence") near the initiator A U G codon on prokaryotic m R N A and the importance of 16S r R N A in initiation has been well established (34,35). However, since the 3'-end of 16S r R N A complementary to the Shine-Dalgarno sequence was found to be identical between E. coli and B. stearothermophilus (36), the role of 16S r R N A demonstrated in our original heterologous reconstitution experiments cannot be explained on the basis of the interaction involving the ShineDalgarno sequence and the 3'-end of 16S r R N A only (see 36,37). Therefore, the question of how 30S ribosomal subunits recognize certain initiation sites preferentially, and not other potentially active initiation sites, has still not been clarified.

Initiation of Protein Synthesis Revisited In the early 1970's, the direction of research in my laboratory changed. During my sabbatical stay at the University of Aarhus, Denmark in 1971, I thought about future research projects and decided to study the regulation of ribosome synthesis, and upon returning to Madison from Denmark, we gradually started to jdo genetics and regulation studies. As I have written elsewhere (38, 39) about how we struggled to isolate r-protein genes (this was before gene cloning techniques were developed), and how we eventually succeeded in isolating transducing phages carrying many r-protein genes (and other essential genes, such as genes for protein elongation factors, EF-G and EF-Tu, and R N A polymerase subunit α), I will not recount this

498

Masayasu Nomura

here. In any event, after the isolation of these transducing phages, we became busy identifying genes, mapping them physically, and defining transcription units. In addition to r-protein genes, we also started to work on rRNA genes. Thus, by the mid 1970's, we had stopped working on ribosome structure and function, including the initiation of protein synthesis, and had started to concentrate on the characterization of ribosomal genes and studies of the regulation of their expression. One of the first major questions we wanted to study was how the expression of genes for 52 r-proteins and 3 rRNAs is regulated to ensure balanced and coordinated synthesis of all of these ribosomal components. At that time, the known mechanisms for gene expression were almost exclusively based on transcriptional control, basically similar to the original operon theory developed by Jacob and Monod (40). In fact, people (including us) working on the regulation of r-protein synthesis anticipated the regulation at the level of transcription, and some supporting evidence and arguments for this thesis had actually been published by other workers (e.g., 41,42). In addition, the regulation of rRNA genes is certainly transcriptional regulation. Thus, for quite a few years our concern was with genes and transcription, and not with translation. I did not expect that I would come back to study the initiation of protein synthesis once more in connection with the regulation of r-protein gene expression. The first experimental evidence to indicate the presence of translational regulation came from our gene dosage experiments using a lysogenic strain carrying a transducing phage which has a set of r-protein genes in the spc and alpha r-protein operon. Based on these experiments we proposed a model of feedback regulation at the level of translation (43; see Figure 1). The postulated feedback regulation at the level of translation was quickly confirmed by in vitro protein synthesis experiments by ourselves (44-47) as well as other workers (48,59), and several r-proteins were identified as translational repressors. The repressor activity of these r-proteins was also demonstrated in vivo by a variety of methods (see 38, 39 for historical accounts and 50, 51 for reviews). One remarkable feature of this regulation was that not all the r-proteins were repressors, but usually one r-protein repressor inhibited the translation of itself as well as all other r-proteins encoded in the same polycistronic mRNA. For example, the L l l operon consits of the gene for L l l and LI (Figure 1), and LI is the translational repressor. In the initial in vitro experiments, we studied the regulation using coupled transcription-translation systems using a DNA template carrying both genes. The addition of purified LI inhibited the synthesis of both L l l and LI (but not other proteins from the template), and it was demonstrated that the inhibition was at the level of translation and not at the level of transcription (44). The question was then whether LI inhibits the synthesis of each of the two proteins individually, or the inhibition of synthesis of both proteins is a result of an interaction of the LI repressor with a single target site on the mRNA. Using various experimental approaches, both in vitro and in vivo, we were able to demonstrate that the second possibility is correct; that is, the repressor LI inhibits translation of L l l directly by interacting with a target site near the translation initiation site of the LI 1 cistron, and inhibition of LI synthesis is indirectly caused because the translation of LI does not take place in the absence of translation of the preceding LI 1 cistron

499

Initiation of Protein Synthesis

L 11 OPERON

Ρ — AUG

m RNA

rpjA

rp^K

UAA

\

AUG

UAA

ΝΛΛΛΛΛΛΛΛΛΛΛΛΛΛΛΛΛΛΛΛΛΛΛΛΛΛΛΛΛΛ/*-

I I ι

A

\

\

LI \

RIBOSOME ENTRY SITE

Elj

I L RIBOSOME

ASSEMBLY

Figure 1 Sequential translation of E. coli ribosomal protein L l l and LI and its feedback repression by LI. The LI 1 r-protein operon in E. coli consists of the genes for LI 1 (rplK) and LI (rplA). After syntheis of the bicistronic mRNA, ribosomes can engage at the entry site on the LI 1 mRNA. The L l l m R N A is then translated, followed by sequential translation of the LI coding sequences. No independent translation initiation takes place at the initiation region for LI. indicated by a crossed arrow below the initiation codon for LI. This mechanism ensures the equimolar synthesis of both L l l and LI. Once the L l l and LI proteins are synthesized, two different paths can be taken that affect the continued translation of these proteins from their mRNA. In one path (solid lines) L l l and LI participate in the assembly of ribosomes and the further translation of LI 1 and LI from the mRNA continues. In the absence of ribosome assembly (dashed lines) "free" L l l and LI accumulate. The prepressor r-protein LI (indicated by a box) can then interact with the bicistronic mRNA near the ribosome entry site and block the further, unnecessary, translation of both L l l and LI. Thus the synthesis of L l l and LI is identical and coordinated to ribosome assembly.

(52,53). We called this phenomenon, the dependence of the translation of LI on the translation of the preceding L l l , sequential translation. The presence of sequential translation was also demonstrated for several other r-protein opérons we were studying (50). Although our experiments demonstrating sequential translation were straightforward and convincing, and a similar conclusion had just been made by Oppenheim and Yanofsky for the trp operon from analyses of polar mutants (54; called translational coupling), many people were surprised at this conclusion. I suspect that people had been so accustomed to cotranscription of genes as the basis of coregulation of linked genes that they had forgotten the possibility of "cotranslation" of genes on polycistronic m R N A i.e., sequential translation or translational coupling as we now call it. The question of translation of several proteins from a single polycistronic m R N A was not a new question. In the 1960's, soon after the physical demonstration of the

500

Masayasu Nomura

presence of polycistronic m R N A in E. coli, people, including myself, wondered whether translation of these proteins from the m R N A takes place independently, or takes place sequentially starting from the first cistron of the m R N A (e.g., see 5,55,56). Actually, there was an example of sequential translation (translational coupling); translation of the replicase cistrons of R N A phage f2 is dependent on the translation of the preceding coat cistron, as originally suggested by the discovery of a polar effect caused by an amber mutation in the coat cistron on the translation of the replicase cistron (57) and by the kinetics of appearance of these proteins in in vitro systems (58,59). However, it was always thought that R N A phages could be exceptional, and the question regarding "natural" polycistronic m R N A remained unanswered for many years. I should also note that in the 1960's some people even pointed out that possible sequential translation of polycistronic m R N A might play a role in connection with coregulation of inked genes by a repressor (55). In fact, when Jacob and Monod proposed the operon theory from their studies on the lactose and phage λ systems, they considered two possibilities, one with a repressor acting on a target site on D N A and another with a repressor acting on a target site on m R N A , and favored the former (40). On rereading their classical paper published in 1961 we realize that the arguments they used for favoring the former are not compelling. Of course, they were correct with respect to their systems (and many other opérons studied subsequently), and their work exerted a strong influence on people working on the regulation of gene expression in prokaryotic systems. As I have already explained above, coregulation of the expression of many r-protein genes uses r-protein repressors, but the repressors act on target sites on m R N A with sequential translation as the basis of coregulation of these linked genes; that is, the second alternative discarded by Jacob and Monod has turned out to be the case in our r-protein systems. As mentioned above, in addition to r-protein opérons, translational coupling was first observed for the trp operon (54) and subsequently for several other bacterial opérons. Undoubtedly, the presence of translational coupling must be quite common and important for the regulation of gene expression. Several obvious questions then arise, What is the significance of coregulation of linked genes at the level of translation rather than transcription (at least in the case of many r-protein genes)? What is the mechanism which prevents independent translation initiation of distal cistrons? The "masked" distal cistron, have, in most instances, good ShineDalgarno sequences and yet fail to allow independent translation initiation. Is this simply due to the presence of a secondary structure masking the Shine-Dalgarno sequences and/or the A U G initiator codons? Our recent work on the translational coupling of LI 1 and LI suggests that masking by secondary structures involving simple base pairing is probably not sufficient for explaining the absence of independent initiation of translation at the distal cistron (60). How does the ribosome translating an upstream cistron behave at the intercistronic junction before the translation of the next cistron starts? Is it the same ribosomes that translate the next cistron or is there exchange (both subunits or possibly only the 50S subunit) with free ribosomal subunits in the pool? I now realize most of these questions were asked and discussed some 20 years ago by some investigators

Initiation of Protein Synthesis

501

working on the initiation of protein synthesis, including myself (e.g., ref. 5), but the questions were not solved at that time. First of all, the question of sequential vs. independent translation of polycistronic m R N A was not solved at that time (except for R N A phages) as mentioned above. Technologies were not sufficiently advaced at that time to give definitive answers. Thanks to advances in various technologies, especially recombinant D N A technologies, we have established sequential translation (or translational coupling, as it is generally called now) in a convincing way, and questions regarding the mechanisms responsible for masking and unmasking of distal cistrons should now be amenable to experimental analysis. The outcome of these studies should also have some relevance to the regulation of translation initiation on eukaryotic m R N A . It was originally recognized that in eukaryotic systems ribosomes rarely initiate at internal initiation codons and that the presence of the 5'-end of the m R N A is required for the initiation of protein synthesis. This led to the proposal of the scanning model by Kozak (61,62); the 40S ribosomal subunit binds initially at or near the 5'-end of m R N A and then migrates down to the first A U G codon, at which site a 60S subunit joins and the first peptide bond formation takes place. Although much emphasis was given to the differences in the initiation between eukaryotes and prokaryotes in the past, it should be noted that in the E. coli r-protein opérons discussed above, independent initiation does not take place at the initiation codon of distal cistrons. Only ribosomes migrating from the upstream (translating the upstream cistrons) can cause translation of these distal cistrons as a result of chain termination followed by the act of reinitiation. The existence of a similar situation, translation of distal open reading frames as a result of chain termination followed by reinitiation, has also been recognized recently in eukaryotic systems (for a review, see 63). In fact, the presence of open reading frames in the upstream leader region appears to be used for regulation in some systems as revealed by recent studies on the regulation of translation of yeast G C N 4 m R N A by Hinnenbush and other workers (see e. g. 64). Although a general rule for the initiation of protein synthesis in eukaryotes has been formulated as the modified scanning model by Kozak (63), precise information on the signals which specify the initiation sites is still not clearly established, and biochemical mechanisms responsible for the selection of the initiation site are unknown. Undoubtedly, the mechanisms must be complex so that sophisticated regulation can be exerted on the initiation step, as exemplified by the yeast G C N 4 system mentioned above. Another question regarding the initiation of protein synthesis that we are currently interested in concerns the factors which determine the rate of total protein synthesis in growing E. coli cells. Under a set of nutritional conditions, bacterial cells attain a certain growth rate, that is, a certain total protein synthesis rate. What determines the total protein synthesis rate? The work by Arthur Koch and his coworkers in the early 1970's clearly demonstrated that the rate of total protein synthesis in bacteria is not limited by the amount of ribosomes in vivo (65,66). Our recent experiments using a conditional r R N A expression system has also given the same conclusion (67). In addition, we have recently demonstrated that the rate of total protein synthesis is not determined either by the supply of m R N A (68), or by the amounts of initiation factor IF2 (69). These latter conclusions were obtained using genetically

502

Masayasu Nomura

engineered bacterial systems where the chromosomal genes for R N A polymerase subunits β and β' or for IF2 are replaced by the β, β' genes or the IF2 gene under lac promoter/operator control, respectively, and R N A polymerase concentration or IF2 concentration can be decreased by decreasing the concentration of lac operon inducers. We suspect that all the components for protein synthesis are probably present in excess and that the initiation reaction itself (possibly together with the elongation reaction), is directly regulated by some system which senses the cell's "growth capacity" under given nutritional conditions. One specific model we would like to test is that the act of translation itself generates some feedback signals to inhibit the initiation reaction, when the total protein synthesis rate exceeds the cell's growth capacity, e. g., the supply of energy and substrates. It is even possible that the same feedback signal(s) is used for the regulation of total protein synthesis (the activity of the translational apparatus) and for the regulation of ribosome biosynthesis (the amount of translational apparatus). The latter subject, the pathway used to generate feedback signals for the regulation of ribosome synthesis, has been discussed recently (69,70). It is known that in eukaryotes phosphorylation of the α subunit of eIF2 appears to be used to regulate the initiation of protein synthesis in several instances, e.g. in heme-deficient reticulocytes and interferontreated or heat-shocked mammalian cells (for a review, see 71). However, it is not known whether such phosphorylation determines the rate of total protein synthesis under most "normal" conditions. Other factors, such as the concentrations of K + and Ca 2 + , and pH, are also known to affect the rate of the initiation of protein synthesis and might represent elements involved in the global regulation of protein synthesis. Protein synthesis was an exciting subject for study in the 1960's. Twenty years later, the subject, the regulation of protein synthesis in particular, is still fascinating and important questions remain to be elucidated.

Acknowledgements I thank my present, as well as my previous, coworkers, who have participated in the work described in this article. I also thank Drs. S. M. Arfin and J. R. Cole for their reading of this manuscript. Our research described here has been continuously supported by grants from the National Institute of Health (currently GM35949) and from the National Science Foundation (currently PCM79-10616).

References 1. 2. 3. 4. 5.

Nirenberg, M.W. and J.H. Matthei. 1961. Proc. Natl. Acad. Sci. USA. 47, 1588. Lipmann, F. 1941, Adv. Enzymol. I 99. Tissières, A. and J.D. Watson. 1958. Nature 182, 778. Nomura, M. and C.V. Lowry. 1967. Proc. Natl. Acad. Sci. USA 58, 946. Nomura, M. 1968. In: 4. Wissenschaftliche Konferenz der Gesellschaft Deutscher Naturforscher und Ärzte. Springer, Berlin, p. 50. 6. Kaempfer, R.O.R., M. Meselson and H.J. Raskas. 1968. J. Mol. Biol. 31, 277. 7. Takanami, M. and T. Okamoto. 1963. J. Mol. Biol. 7, 323.

Initiation of Protein Synthesis 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58.

503

Pestka, S. and M. Nirenberg. 1966. J. Mol. Biol. 28, 145. Suzuka, I., H. Kaji and A. Kaji. 1966. Proc. Natl. Acad. Sci. USA 55, 1483. Monroe, R.E. 1967. J. Mol. Biol. 26, 147. Hosokawa, K., R . K . Fujimura and M. Nomura. 1966. Proc. Natl. Acad. Sci. USA 55, 198. Nomura, M., C.V. Lowry and C. Guthrie. 1967. Proc. Natl. Acad. Sci. USA 58, 1493. Ohta, T., S. Sarker and R.E. Thach. 1967. Proc. Natl. Acad. Sci. USA 58, 1638. Ghosh, H.P. and H . G . Khorana. 1967. Proc. Natl. Acad. Sci. USA 58, 2455. Hille, Μ. Β., M.J. Miller, Κ. Iwasaki and A.J. Wahba. 1967. Proc. Natl. Acad. Sci. USA 58, 1652. Mangiarotti, G. and D. Schlessinger. 1966. J. Mol. Biol. 20, 123. Kohler, R.E., E. Ron and B.D. Davis. 1968. J. Mol. Biol. 36, 71. Kaempfer, R. 1974. In: Ribosomes (M. Nomura, A. Tissières and P. Lengyel, eds.). Cold Spring Harbor Laboratory, p. 679). Davis, B.D. 1974. In: Ribosomes (M. Nomura, A. Tissières and P. Lengyel, eds.). Cold Spring Harbor Laboratory, p. 705. Schlessinger, D., G. Mangiarotti and D. Apirion. 1967. Proc. Natl. Acad. Sci. USA 58, 1782. Kaempfer, R. 1968. Proc. Natl. Acad. Sci. USA 61, 106. Kaempfer, R. 1969. Nature 222, 950. Guthrie, C. and M. Nomura. 1968. Nature 219, 232. Marcker, K. A. and F. Sanger. 1964. J. Mol. Biol. 8, 835. Webster, R.E., D . K . Engelhardt and N. Zinder. 1966. Proc. Natl. Acad. Sci. USA 55, 155. Adams, J.M. and M . R . Capecchi. 1966. Proc. Natl. Acad. Sci. USA 55, 147. Clark, B.F.C, and K.A. Marcker, 1966, J. Mol. Biol. 17, 394. Stanley, jun., W.M., M. Salas, A.J. Wahba and S. Ochoa. 1966. Proc. Natl. Acad. Sci. USA 56, 290. Revel, M. and F. Gros. 1966. Biochem. Biophys. Res. Commun. 25, 124. Eisenstadt, J. and G. Brawerman. 1966. Biochemistry 5, 2777. Subramanian, A.R. and B.D. Davis. 1970. Nature 228, 1273. Sabol, S., M. Sillero, Κ. Iwasaki and S. Ochoa. 1970. Nature 228, 1269. Held, W.A., W.R. Gette and M. Nomura. 1974. Biochemistry 13, 2115. Shine, J. and L. Dalgarno. 1974. Proc. Natl. Acad. Sci. USA 71, 1342. Steitz, J. A. and K. Jakes. 1975. Proc. Natl. Acad. Sci. USA 72, 4734. Sprague, K.U., J. A. Steitz, R.M. Grenley and C.E. Stocking. 1977. Nature 267, 462. Nomura, M. Cold Spring Harbor Symp. Quant. Biol 52, (in press) Nomura, M. 1981. In: Science and Scientists: Essays by Biochemists, Biologists and Chemists (M. Kageyama, ed.). Japan Scientific Societies Press, p. 133. Nomura, M. 1984. Scientific American 250, 102. Jacob, F. and M. Monod. 1961. J. Mol. Biol. 3, 318. Maale, O. 1969. Dev. Biol. (Suppl.) 3, 33. Gausing, K. 1977. J. Mol. Biol. 115, 335. Fallon, A.M., C.S. Jinks, G . D . Strycharz and M. Nomura. 1979. Proc. Natl. Acad. Sci. USA 76, 3411. Yates, J.L., A.E. Arfsten and M. Nomura. 1980. Proc. Natl. Acad. Sci. USA 77, 1837. Yates, J.L. and M. Nomura. 1980. Cell 21, 517. Dean, D., J.L. Yates and M. Nomura. 1981. Cell 24, 413. Yates, J.L., D. Dean, W.A. Strycharz and M. Nomura. 1981. Nature 294, 190. Fukuda, R. 1980. Mol. Gen. Genet. 178, 483. Brot, N„ R. Caldwell and H. Weissbach. 1980. Proc. Natl. Acad. Sci. USA 77, 2592. Nomura, M., R. Gourse and G. Baughman. 1984. Ann. Rev. Biochem. 53, 75. Lindahl, L. and J. Zengel. 1986. Ann. Rev. Genet. 20, 297. Yates, J.L. and M. Nomura. 1981. Cell 24, 243. Baughman, G. and M. Nomura. 1983. Cell 34, 979. Oppenheim, D.S. and C. Yanofsky. 1980. Genetics 95, 785. Ames, B.N. and P.E. Hartman. 1963. Cold Spring Harbor Symp. Quant. Biol. 28, 349. Berberich, Μ. Α., J. S. Kovachand R.F. Goldberger. 1967. Proc. Natl. Acad. Sci. USA 57,1857. Engelhardt, D.L., R . D . Webster and N.D. Zinder. 1967. J. Mol. Biol. 29, 45. Ohtaka, Y. and S. Spiegelman. 1963. Science 142, 493.

504

Masayasu Nomura

59. 60. 61. 62. 63. 64. 65. 66. 67.

Lodish, H.F. 1968. Nature 220, 345. Sor, F., M. Bolotin-Fukuhara and M. Nomura. 1987. J. Bact. 169, 3495. Kozak, M. 1978. Cell. 15, 1109. Kozak, M. 1983. Microbiol. Rev. 47, 1. Kozak, M. 1986. Cell 47, 481. Mueller, P.P. and A. Hinnenbusch. 1986. Cell 45, 201. Koch, A.L. 1970. J. Theoret. Biol. 28, 203. Koch, A.L. 1971. Adv. Microbiol. Physiol. 6, 147. Gourse, R.L., Y. Takebe, R. A. Sharrock and M. Nomura. 1985. Proc. Natl. Acad. Sci. USA 82. 1069. Nomura, M., D.M. Bedwell, M. Yamagishi, J.R. Cole and J.M. Kolb. 1987. In: RNA Polymerase and Regulation of Transcription (W. S Reznikoff, R. R. Burgess, J. E. Dahlberg, C.A. Gross, M.R. Record, Jr. and M.R. Wickens, eds.). Elsevier, New York, p. 137. Cole, J. R., C. L. Olsson, J. W. B. Hershey, M. Grunberg-Manago and M. Nomura. J. Mol. Biol. 198, 383. Yamagishi, M., H.A. de Boer and M. Nomura. J. Mol. Biol. 198, 547. Pain, V.M. 1986. Biochem. J. 235, 625.

68. 69. 70. 71.

Structure of Ribosomal RNA Genes of Eukaryotes: Some Solved and Unsolved Questions A.A.

Hadjiolov

Introduction Mammalian cells produce 0.2 to 4.5 χ IO3 ribosomes/min/nucleus (1). Since the initial observation that ribosomes turn over even in quiescent cells (2,3) it is clear that ribosome biogenesis is a continuous process taking place in both growing and non-growing cells. This implies the incessant and presumably coordinate expression of about 100 ribosomal genes (coding for rRNA, r-proteins, specific enzymes and factors) involved in ribosome biogenesis. Since pre-rRNA and rRNA serve as backbones in the assembly of preribosomes and ribosomes, elucidation of the organization and structure of rRNA genes and the formation of their product mature rRNAs - is of prime importance in understanding ribosome structure, function and biogenesis. Ribosomal RNA genes were the first eukaryotic genes to be cloned in 1974 (4) and since then numerous powerful molecular genetics techniques resulted in the rapid accumulation of important information on their structure and expression. Our laboratory has been involved in the cloning and sequence analysis of some eukaryotic rRNA genes (Saccharomyces cerevisiae and Rattus norvegicus). Here I shall briefly outline and discuss some questions related to the organization and structure of the multiple rRNA genes in eukaryotes and possible mechanisms of their evolution.

Basic Organization of Eukaryotic rRNA Genes Analysis of the rDNA repeating units from several eukaryotes revealed the following basic features in their organization (review in ref. 1): 1. The multiple rRNA genes are tandemly arrayed in the nucleolus organizer loci of one or more (up to six) specific "nucleolar chromosomes". 2. The rRNA transcription unit (transcribed by RNA polymerase I) follows invariably the arrangement: (promoter) - ETS - S-rDNA - ITS 1 - 5.8 S rDNA ITS 2 - L-rDNA - ETS (ETS = external transcribed spacer; ITS = internal transcribed spacer). 3. Transcription units are always separated by NTS sequences (NTS = nontranscribed spacer) showing considerable variations among species both in size and in primary structure.

506

A.A. Hadjiolov

While the above basic features in the organization of eukaryotic r R N A genes are now firmly established, the genetic factors directing their preservation remain largely unknown. In trying to understand the molecular mechanisms ensuring the rather complex organization of the multiple r R N A genes in eukaryotes the following observations should be taken into consideration. 1. The different segments constituting the r D N A repeating units seem to evolve at different rates. The sequences of the S-rRNA, 5.8 r R N A and L-rRNA are certainly the most strongly conserved. For example, the sequence of the about 3.4 kb L-rRNA gene of S. cerevisiae (5) and S. carlsbergensis (6) is practically identical, while only limited differences in the L-rRNA gene sequence are observed when closely related species like Mus musculus (7) and Rattus norvegicus (8) or human and chimpanzee (9) are compared. It is noteworthy that the S-rRNA gene appears to be more strongly conserved as compared to the L-rRNA gene. The size and largely the sequence of the S-rRNA gene show only little changes in the 15 eukaryotic genes with presently known primary structure (10), while the L-rRNA gene varies from 3.4 kb in Saccharomyces to about 5.0 kb in Homo (see below). In contrast, the transcribed spacer sequences (both ETS and ITS) are rapidly evolving from species to species, a fact which is strikingly evident when r R N A gene and flanking sequences from different species are aligned (11). The changes are so abrupt that the transition gene-spacer may be outlined within a few nucleotides upstream or downstream from the gene sequence. Finally, the NTS sequences display considerable heterogeneity in size and structure among the few eukaryotic species studied till now (see réf. 1). 2. The strong conservation of r R N A gene sequences within the r D N A repeating unit seems to be related to their existence in tandem arrays clustered in the nucleolus organizer(s). In a few cases isolated r R N A genes ("orphons") have been identified, namely in Drosophila (12) and the human (13, 14) genome. It is significant that these single r R N A genes are clearly inactive pseudogenes, which have largely diverged in sequence from the bulk of the clustred r R N A genes (14). 3. Although the r D N A repeating units show considerable variations among even closely related species, the sequence of the r D N A is basically the same in apparently all of the multiple repeating units present in the single or numerous clusters of tandemly arrayed r R N A genes. The structural homogeneity of the r D N A repeating units within the species was first observed when Xenopus laevis and Xenopus borealis rDNAs were compared (15) and has since been confirmed in numerous studies (see ref. 16). Our own studies support this conclusion at the sequence level when different and presumably independent rDNA clones of the same organism have been investigated as in the case of the NTS from Saccharomyces (17) or Rattus norvegicus (18,19) or the ITS of the rat (20,21). Only minor differences are recorded even in the highly divergent NTS r D N A sequences. More detailed sequence analyses of different Xenopus laevis r D N A clones revealed only minor differences in the ITS and none in the S-rDNA (22), while intraspecies heterogeneity is most likely due to the presence of a varying number of repeated elements within the NTS segment (23). The intraspecies homogeneity of the multiple r D N A repeating units implies the operation of a rapid and specific mechanism designated by different authors as horizontal or concerted evolution. The occurence of the horizontal evolution phenomenon is still a major enigma in modern biology and elucidation of its

Structure of Ribosomal RNA Genes of Eukaryotes

507

mechanisms will certainly unravel important facets in the comportment of eukaryotic genomes. Before discussing some aspects of this phenomenon, I shall briefly outline the outcome of studies on the

GC-Rich Segments in the L-rRNA Gene of Eukaryotes As mentioned above the L-rRNA in eukaryotes is characterized by marked increases in size, vertebrate L-rRNA exceeding by about 1500 nucleotides their counterpart in yeast. Earlier studies revealed that the size difference is largely accounted for by the presence of rather large (G + C)-rich segments (GCS) in mammalian L-rRNA. The presence of such GCS in L-rRNA was first observed (quite unexpectedly) in studies on the limited enzymatic digestion of isolated mammalian r R N A (24-26). Their location along the polymer chain was nicely visualized upon electron microscopy of partially denatured L-rRNA molecules (27,28). At present, the complete sequence of several vertebrate L-rRNAs is known, namely that of Xenopus laevis (29), Mus musculus (7), Rattus norvegicus (8,30) and Homo sapiens (9). Thus it is now possible to compare at the sequence level the vertebrate L-rRNAs with their S. cerevisiae (5) counterpart. The results of such a comparison are shown in Figure 1. The results shown in Figure 1 allow the following conclusions: a) severalregions of high or moderate homology exist in eukaryotic L-rRNA, distributed all along the polymer chain; b) expansion segments are located at specific and identical sites along the L-rRNA chain. Analysis of the nucleotide composition of these expansion segments reveals that, at least in the case of the large ones (1,4,6 and 8), they are characterized by a G + C

Figure 1 Distribution of conserved, non-conserved and expansion segments along the L-rRNA chain of vertebrates as compared to the L-rRNA sequence of S. cerevisiae. X. 1. = Xenopus laevis (29), M.m. = Mus musculus (7), R.n. = Rattus norvegicus (8), H.s. = Homo sapiens (9). Regions of more than 80% sequence homology are designated by black areas aligned along the yeast sequence. The position and the size of identified GCS (numbered 1 to 8) are denoted by full horizontal lines.

508

Α. Α. Hadjiolov

content in the range of 80 to 85% (about 65% for total L-rRNA) and therefore may be considered as true GCS. A more detailed sequence comparison reveals that the GCS of the L-rRNA from different vertebrate species are related in structure and further expansion seems to take place by "growth from the tip" of putative doublehelical loops (31,32). The sequence analysis of GCS demonstrates that most likely they do not originate as insertion sequences transposed from other genomic sites, but are formed as a result of unequal crossing-over and further fixation in the L-rRNA gene. The sequence analyses of GCS elements in eukaryotic L-rRNA genes have to be considered in relation to the basic characteristics of eukaryotic r R N A genes outlined above and in the context of current views on the still enigmatic horizontal evolution phenomenon. At present considerable evidence has been accumulated to explain satisfactorily the rapid spreading of an r D N A sequence among the multiple repeating units. It is plausible that the horizontal evolution phenomenon of r D N A repeating units in eukaryotes is due to rapid "molecular drive" changes including mechanisms of unequal crossing-over, gene conversion and (in some cases) transposition (16). The "molecular drive" mechanisms could thus serve to homogenize continuously the sequences of the r D N A repeating units within a species. However, the "molecular drive" mechanisms may not be the only ones to explain the comportment of r D N A repeating units. As stressed above the different constituents of the r D N A repeating unit evolve clearly at different rates. Moreover, the rapid expansion of the GCS in the L-rRNA gene is in contrast to the high degree of conservation of homologous sequences within this gene. These facts strongly suggest that selection pressures are most probably acting on the changes in r D N A repeating units taking place under the action of "molecular drive" phenomena. How are these selection pressures operating? It is still premature to visualize mechanisms acting on NTS, ETS or ITS r D N A sequences. However, the changes taking place in L-rRNA GCS deserve further emphasis. The pattern outlined in Figure 1 clearly shows that at least in the case of GCS 1 and 6 the observed changes are clearly unidirectional in evolution. Therefore, we should envisage the existence of some "sensor mechanism" superimposed over the rapidly ocurring "molecular drive" changes. In other words I consider useful to propose that eukaryote (or at least vertebrate) germ cells possess "sensor mechanisms" discriminating between large ribosomal subunits with expanded vs. shortened GCS segments obtained by rapidly acting "molecular drive" mechanisms. Hopefully, further studies on the structure and comportment of r D N A repeating units in eukaryotes will help to unravel important facets in the biology of their genome.

References 1. Hadjiolov, A.A. 1985. The Nucleolus and Ribosome Biogenesis, pp.267, Springer, Wien/New York. 2. Loeb, J.N., R.R. Howell and G.M. Tomkins. 1965. Science 149, 1093. 3. Hadjiolov, A.A. 1966. Biochim. Biophys. Acta 119, 547. 4. Morrow, J.F., S.N. Cohen, A. Chang, M.W. Boyer, H.M. Goodman and R.B. Helling. 1974. Proc. Natnl. Acad. Sci. USA 71, 1743.

Structure of Ribosomal R N A Genes of Eukaryotes

509

5. Georgiev, O.I., Ν. Nikolaev, A.A. Hadjiolov, K . G . Skryabin, V.M. Zakharyev and A . A . Bayev. 1981. Nucleic Acids Res. 9, 6853. 6. Veldman, G . M . , J. Klootwijk, V.C.H.F. deRegt, R.J. Planta, C. Branlant, A. Krol and J.-P. Ebel. 1981. Nucleic Acids Res. 9, 6935. 7. Hassouna, N., B. Michot and J.-P. Bachellerie. 1984. Nucleic Acids Res. 12, 3563. 8. Hadjiolov, Α. Α., O.I. Georgiev, V.V. Nosikov and L.P. Yavachev. 1984. Nucleic Acids Res. 12, 3677. 9. Gonzales, I.L., J.L. Gorski, T.J. Campen, D.J. Dorney, J . M . Erickson, J.E. Sylvester and R . D . Schmickel. 1985. Proc. Natnl. Acad. Sci. USA, 82, 7666. 10. Huysmans, E. and R. DeWachter. 1986. Nucleic Acids Res. 14, r73. 11. Hadjiolova, K.V., O.I. Georgiev, V.V. Nosikov and A.A. Hadjiolov. 1984. Biochem. J. 220, 105. 12. Childs, G., R. Maxson, R.H. Cohn and L. Kedes. 1981. Cell 23, 651. 13. Brownell, E., M. Krystal and N. Arnheim. 1983. Mol. Biol. Evol. /, 29. 14. Munro, J., R . H . Burdon and D.P. Leader. 1986. Gene 48, 65. 15. Brown, D.D., P.C. Wensink and E. Jordan. 1972. J. Mol. Biol. 63, 57. 16. Dover, G.A., R.B. Flavell. 1984. Cell 38, 622. 17. Skryabin, K.G., M. A. Eladarov, V.L. Larionov, A.A. Bayev, J. Klootwijk, V.C.H.F. deRegt, G . M . Veldman, R.J. Planta, O.I. Georgiev and A.A. Hadjiolov. 1984. Nucleic Acids Res. 12, 2955. 18. Yavachev, L.P., O.I. Georgiev, E.A. Braga, T. A. Avdonina, Α.Ε. Bogomolova, V.B. Zhurkin, V.V. Nosikov and A.A. Hadjiolov. 1986. Nucleic Acids Res. 14, 2799. 19. Mroczka, D.N., Β. Cassidy, Η. Busch and L.I. Rothblum. 1984. J. Mol. Biol. 174, 141. 20. Subrahmanyam, C.S., Β. Cassidy, Η. Busch and L.I. Rothblum. 1982. Nucleic Acids Res. 10, 3667. 21. Georgiev, O. I., V.V. Nosikov, E. A. Braga and A.A. Hadjiolov. 1984. Biochem. Internat.«, 225. 22. Stewart, M.A., L . M . C . Hall and B.E.H. Maden. 1983. Nucleic Acids Res. 11, 629. 23. Wellauer, P.Κ., R . H . Reeder, I.B. Dawid and D . D . Brown. 1976. J. Mol. Biol. 105, 487. 24. Hadjiolov, A.A., P.V. Venkov, L.B. Dolapchiev and D . D . Genchev. 1967. Biochim. Biophys. Acta 142, 111. 25. Hadjiolov, A.A. and G . I . Milchev. 1967. Compt. Rend. Acad. Sci. Bulgarie 20, 1333. 26. Delihas, Ν. 1967. Biochemistry 6, 3356. 27. Wellauer, P.K. and I.B. Dawid. 1973. Proc. Natnl. Acad. Sci. USA 70, 2827. 28. Schibler, U „ T. Wyler and O. Hagenbuchle. 1975. J. Mol. Biol. 94, 503. 29. Ware, V.C., B.W. Tague, C.G. Clark, R.L. Gourse, R.C. Brand and S.A. Gerbi. 1983 Nucleic Acids Res. 11, 7795. 30. Chan, Y.-L., J. Olvera and I.G. Wool. 1983. Nucleic Acids Res. 11, 7819. 31. Michot, B., Hassouna, N. and J.-P. Bachellerie. 1984. Nucleic Acids Res. 12, 4259. 32. Michot, B. and J.-P. Bachellerie. 1987. Biochimie 69, 11.

Energetics and Dynamics of the ProteinSynthesizing Machinery Alexander S. Spirin

Carboxyl Activation and Spontaneous Headward Elongation The classical studies of Fritz Lipmann on carboxyl activation introduced comprehension and language into biochemical energetics (1,2). In addition, these studies were a prerequisite and a starting point for the discovery and understanding of the mechanism of amino acid activation and polymerization. Similar to the two-step acetyl activation, the activation of an amino acid was found to proceed in two steps (3). The primary activation catalyzed by aminoacyl-tRNA synthetase involves the amino acid carboxyl group which accepts the adenylic acid residue from ATP, thus resulting in the formation of a mixed carboxylphosphoanhydride, aminoacyl adenylate:

Ï

NH, - CH - C 3 \

?"

0

L. - Ρ δ + - O - CH,

OH OH The aminoacyl adenylate reacts then on the same activating enzyme with tRNA which replaces the adenylic acid residue, thus producing an ester, aminoacyl-tRNA: + ι NH3 - CH - C ^

^-CH2

H

- (NMP)

Va tRNA

In Lipmann's terms, at first the high pyrophosphoryl group potential of ATP converts into the phosphoryl group potential of the phosphorylacylate, and then the latter converts into the acyl group potential of the ester (2,4). The eventual sense of the activation is the acquisition of a high acyl group potential expressed as a partial positive charge δ + at the carbonyl carbon which will be attacked by a nucleophilic

Alexander S. Spinn

512

acceptor molecule. Thus, the transformation of aminoacyl adenylate into aminoacyl-tRNA is very important not only for the subsequent decoding of a messenger polynucleotide, but also from the energetics point of view, since the transfer of δ + from Ρ of the anhydride to C of the ester is realized in this reaction. The latter is a crucial moment for the subsequent polymerization providing the attraction of the unshared pair of electrons on the amino group of the acceptor molecule to the ester carbon and hence the formation of the tetrahedral intermediate in the transpeptidation reaction (5,6): o R1 II I tRNA' - O - C - C H - N H 6+ +

tRNA" - O - C - CH - NH2

tRNA' tRNA"

OH R' I I - O - C - C H - N H -

I

- O - C - C H - N H

II

O

I

R" Ο

R·

tRNA" - O - C - CH - NH - C - CH - NH

II

O

I

R"

tRNA'-OH In considering the condensation or polymerization of activated molecules, a special nomenclature was proposed by Lipmann for definition of the mechanism of molecule participation in the reaction (4). According to the nomenclature, the activated part of the polymerizing unit is the "head", whereas the passive part of the other unit is the "tail". In the case of the polypeptide synthesis, carboxyl is the head and amino group is the tail. The elongation of a polypeptide from aminoacyl-tRNA substrates is considered as a headword chain growth: the activated head of the polypeptide chain reacts with the tail of the next amino acid residue while the head of this adding residue also carries on itself the activating annex (tRNA): Aa n ~ tRNA' + A a ~ t R N A "

Aa n + 1 ~ t R N A " + tRNA'

Thus, the chain is always headed by the activating annex (tRNA) at the growing end. As a result of the reaction the annex is replaced by the tail of the adding residue,

Energetics and Dynamics of the Protein-Synthesizing Machinery

513

but this residue itself has an activated head which will be ready to accept the next residue, and so on (4). The principle of the headward chain growth was discovered by Lipmann's group as early as at the the beginning of 50's in studies on the synthesis of acetoacetate from acetyl-CoA (see review (4)) and later was found to be realized both in protein biosynthesis (7) and m the synthesis of polypeptide antibiotics (8,9). It is important that the carboxyl activation in the polypeptide syntheses provides both the kinetic possibility for the reaction of peptide bond formation (transpeptidation) and the thermodynamic spontaneity of the amino acid polymerization as a downhill process. Indeed, if a peptide is polymerized from activated amino acids (aminoacyl-tRNA's in protein synthesis or aminoacyl thioesters in the synthesis of polypeptide antibiotics), the acyl group potential (in Lipmann's terms) decreases strongly upon the formation of the amide bond. The change of the standard free energy is not less than — 5 kcal/mol (4) (more likely, even — 7.5kcal/mol (10)). Hence, the process of the headward elongation on the ribosome η Aa ~ tRNA

Aan ~ tRNA + (n - 1 ) tRNA,

or on the corresponding multisubunit enzyme η Aa ~ S—R

Aan ~ S—R + (n - 1 ) HS—R

(where R is a thiol-carrying group of an enzyme subunit), is thermodynamically ensured and must not require any additional energy sources. This was always evident in the case of the non-ribosomal synthesis of peptides. The ribosome-driven elongation of a polypeptide, however, was known to require GTP (11), which is cleaved during polymerization into GDP and orthophosphate (12,13). The cleavage (hydrolysis) of GTP was found to be directly connected with the cyclic work of two soluble proteins, the so-called elongation factors T u (EF-TU) and G (EF-G) (14-17), which will be discussed below. From the experimental data available and the theoretical considerations on the sufficiently high acyl group potential in the reacting amino acid residues, Nishizuka and Lipmann arrived at the conclusion that neither GTP, nor elongation factors are required for the ribosomecatalyzed formation of peptide bonds, but that they are rather needed for some mechanical displacements called translocations on, or within, the ribosome (18). Eventually the experimental results of several laboratories have led to the conception of the elongation cycle for the translating ribosome as represented schematically in Figure 1. As a result of the cycle, one codon of a template polynucleotide is read out and one peptide bond is formed. The cycle consists of three successive steps: aminoacyl-tRNA binding, transpeptidation, and translocation. The process of tRNA aminoacylation feeds the whole elongation cycle both with building materials (amino acids) and with energy (at the expense of energy of ATP). The codon-dependent binding of aminoacyl-tRNA takes place at the socalled A site of the ribosome (the upper "shelf' in the schematically drawn ribosome), while the other, Ρ site (the lower "shelf'), is occupied by peptidyl-tRNA (i.e., tRNA which has brought over a preceding amino acid residue and so carries a nascent peptide). The binding is catalyzed by an elongation factor (EF-TJ with GTP. During the next

514

Alexander S.Spirin Αο + ΔΤΡ

tRNA

AMP + P~P¡

Αα-tRNA

Figure 1 Elongation cycle of the ribosome (lower) fed with amino acids and ATP energy through the reaction of tRNA aminoacylation (upper).

step the aminoacyl-tRNA in the A site reacts with the peptidyl-tRNA in the Ρ site resulting in transpeptidation. The reaction is catalyzed by the peptidyl transferase center of the ribosome itself. Now the elongated peptidyl-tRNA occupies the A site, and the deacylated tRNA produced in the reaction is in the Ρ site. Further follows translocation which is the displacement of the peptidyl-tRNA together with the bound template codon from the A site into the Ρ site and the release of the deacylated tRNA from the Ρ site. The translocation is catalyzed by another elongation factor (EF-G) with GTP. As a result, the original post-translocation state of the cycle is again attained where the peptidyl-tRNA is in the Ρ site, and the A site with a new vacant codon is ready to accept the next aminoacyl-tRNA; the ribosome can start the next cycle. Thus, the process of elongation (translation) consists in the repetition of the cycles. At the same time, Gordon in Lipmann's laboratory seems to be the first who demonstrated the possibility of ribosomal synthesis of a polypeptide in the absence of EF-G, considered as a "translocase" (19). Soon after, Pestka (20,21) and then Gavrilova et al. (22-25) reported that the template-directed elongation of a polypeptide on the ribosome can proceed from aminoacyl-tRNA's in the absence of all the elongation factors and GTP, though in this case it was significantly slower than the factor-promoted (GTP-fed) elongation. The factor-free elongation could be stimulated by a modification or a removal of one of the ribosomal proteins, SI2 (27); this fact suggested that the factor-independent elongation is specially blocked in the intact ribosome until the elongation factor with GTP intervenes, removes the block, and additionally accelerates the elongation cycle. In any way, ribosomal synthesis of polypeptides from aminoacyl-tRNA proved in fact to be a thermodynamically spontaneous (downhill) process, and the additional energy of GTP seemed to be spent for the acceleration (catalysis) of non-chemical functions of the ribosome, such as aminoacyl-tRNA binding and translocation (see below).

Energetics and Dynamics of the Protein-Synthesizing Machinery

515

Elongation Factors and GTP As in the studies on carboxyl activation, Fritz Lipmann was among pioneers in the discovery of the elongation factors and the factor-mediated flow of additional energy in the form of GTP for the ribosome-driven protein synthesis (12-18). At first the role of EF-G was understood as a ribosome-dependent GTPase catalyzing the post-transpeptidation non-chemical step in the elongation cycle called translocation (15,18). Later EF-TU was also acknowledged to be a ribosome-dependent GTPase catalyzing the other non-chemical step of the elongation cycle, the codondirected aminoacyl-tRNA binding (28-31). (As a matter of fact, the first indication on the GTPase function of the protein, TF-1, enhancing the binding of aminoacyltRNA to the ribosome was obtained earlier with the rabbit reticulocyte system in the group of Schweet (32)). Both EF-G and EF-TU were found to be true hydrolytic GTPases cleaving GTP into GDP and orthophosphate by means of water, without any intermediate chemical transfer of groups to substrates of the synthesis, to themselves or to ribosomal sites (33,34). Thus, in the given cases the energetic role of nucleoside triphosphates was not in the activation of a chemical group. As already mentioned, the transpeptidation between peptidyl-tRNA and aminoacyl-tRNA is an exergonic reaction. It has been experimentally shown that this reaction can be the only energy source for the elongation cycle (factor-free GTP-independent elongation) (26). Moreover, the codon-directed binding of aminoacyl-tRNA to the peptidyl-tRNA-carrying ribosome (evidently, to the A site of the ribosome) is also spontaneous and can proceed without the elongation factors and GTP ("non-enzymic" binding) (35). If the pre-translocation state ribosomes are prepared and used in experiment, the translocation can also be demonstrated to proceed spontaneously, without the elongation factors and GTP ("non-enzymic" translocation) (36,37); the fact that the affinity of an N-blocked amino acid residuebonded tRNA to the Ρ site of the ribosome is significantly higher than to the A site (38,39) seems to underlie the spontaneity of translocation. Nevertheless, participation of the elongation factors with GTP in the codondependent aminoacyl-tRNA binding and translocation steps does impart a number of new properties to these processes and to all the elongation cycle: it makes them significantly faster, more accurate, and more resistant (26,40). The catalytic role of the elongation factors is the most evident, has been repeatedly discussed, and does not need special comments. Indeed, EF-TU with GTP makes the binding of amino-acyl-tRNA to the ribosome fast, in contrast to the relatively slow "non-enzymic" binding (see, e.g. (17,30)). EF-G with GTP strongly accelerates the translocation (41). The ribosome-catalyzed transpeptidation seems to be always the fastest step of the elongation cycle. As a result, due to the participation of the elongation factors with GTP, the elongation becomes at least two orders of magnitude faster than that without the factors (24,26,40). The peculiarity is that the catalytical action of these proteins is coupled with the cleavage (hydrolysis) of GTP. This is considered unusual for enzymic catalysis. However, if enzymic catalysis means a decrease of activation barriers of a reaction owing to the affinity of the enzyme for the transition state of a substrate, the

516

Alexander S. Spirin

completion of the reaction and the vacation of the enzyme must be paid from the free energy change of the chemical reaction catalyzed. In other words, the exergonic chemical reaction is necessary in order to compensate the free energy gain which has been attained during the formation of the enzyme-substrate complex. In the case of the elongation factors, the catalyzed process is not a chemical reaction. Therefore, it is likely that if an elongation factor with GTP also decreases the kinetic barrier of a certain step due to the affinity for an intermediate (conformational transition) state

EF-VGDP

Figure 2 A plausible sequence of events during EF-T U · GTP-catalyzed ("enzymic") binding of aminoacyl-tRNA with the A site of a translating ribosome. The ribosome is drawn by a contour resembling that of its non-overlap projection, the small subunit up and the large one below. The residues of t R N A are given as wedgy L-shaped filled figures, amino acid residues as small open circles, and EF-T U as large circles, either hatched with G T P or open with GDP. (1) The post-translocation state ribosome with a vacant codon in the A site scans t R N A anticodons among Aa-tRNA · EF-T U · G T P complexes. (2) In the case of cognate anticodon recognition the ribosome firmly binds the corresponding Aat R N A - EF-T U · G T P complex at the expense of both the codon-anticodon interaction and the affinity of EF-T U · G T P for the ribosome. As a result, Aa-tRNA is in a "transition" (barrier) binding state (its anticodon is already in the A site). (3) Hydrolysis of G T P on EF-T U is induced by the ribosome resulting in a change of the EF-T U conformation and, hence, the reduction of its affinity for Aa-tRNA and the ribosome. Aa-tRNA is in an intermediate metastable state from which it can be released (fall out) into solution in the case of incomplete fit to the codon (correction stage). (4) EF-T U · G D P is released from the ribosome, and Aa-tRNA finally settles into the A site. A productive state ready for the next step of the elongation cycle (transpeptidation) is attained.

Energetics and Dynamics of the Protein-Synthesizing Machinery

517

of the tRNA-ribosome complex, the completion of the step and the desorption of the factor will require a significant energy compensation at the expense of an exergonic chemical reaction. It is the hydrolysis of GTP that can be such a reaction. Thus, catalysis in the given case is GTP-dependent because it is not a chemical reaction but a conformational change that is subjected to the catalysis. I believe (and predict) that "enzymic" catalyses of conformational changes of macromolecules and their complexes must be accompanied by nucleoside triphosphate hydrolysis in other cases as well. The sequence of events during catalysis of a conformational change of the tRNAribosome complex appears as follows (Figures 2 and 3). The protein (elongation

Figure 3 A plausible sequence of events in EF-G · GTP-catalyzed ("enzymic") translocation. Symbols are the same as in Figure 2, but large double-circles are EF-G, either hatched with G T P or open with GDP. (1) The pre-translocation state ribosome, with peptidyl-tRNA in the A site and deacylated t R N A in the Ρ site, interacts with EF-G · GTP. (2) A firm complex of EF-G · G T P seems to be formed with a "transition" (barrier) state of the translocating ribosome when the peptidyl-tRNA and the deacylated t R N A are pulled out from their A and Ρ sites, respectively, and the peptidyl-tRNA anticodon is already in the Ρ site. (3) Hydrolysis of G T P on EF-G is induced by the ribosome resulting in a change of the EF-G conformation, so that its affinity for the ribosome strongly decreases. It seems that the affinity of the deacylated t R N A for the ribosome complex is also strongly reduced. (4) EF-G · G D P is released from the ribosome, the peptidyl-tRNA finally settles in the Ρ site, and the A site is fully vacant and ready to accept the next aminoacyl-tRNA (post-translocation state).

518

Alexander S. Spirin

factor) itself does not posses an essential affinity for the ribosome, tRNA or tRNAribosome complex. The binding of GTP to the protein imparts the affinity for them, and specifically for an intermediate high-barrier sub-state in the process of either aminoacyl-tRNA binding or translocation (in the case of either EF-TU · GTP or EFG · GTP, respectively). At the same time, the ribosome induces a GTPase activity of the factor resulting in the cleavage of GTP on the protein and, hence, in the loss of its affinity for the ribosome. Thus, the free energy of GTP hydrolysis is expended for desorption of the catalyst protein from the ribosome, and so opens the way for the next stage of the elongation cycle. Another aspect of the GTP-dependent catalysis of the corresponding steps of the elongation cycle is the increase of accuracy of the process. This is especially evident in the case of the codon-dependent binding of aminoacyl-tRNA. The point is that the participation of the elongation factor with GTP in the aminoacyl-tRNA binding gives the possibility to introduce an additional correction mechanism into the binding process (42). Errors in the binding of aminoacyl-tRNA to the ribosome arise because of a certain probability of non-cognate codon-anticodon interactions, especially as concerns the aminoacyl-tRNA's close to the cognate aminoacyl-tRNA in their codon specificity ("near-cognate" interactions). Naturally, the affinity of an aminoacyl-tRNA for the cognate codon and, respectively, the life-time of a cognate codon-anticodon complex is greater than those in the cases of non-cognate interactions. It is evident that the binding errors will be determined by the difference in the affinities (or the difference in the life-times) between the cognate and noncognate complexes: the greater the difference, the lower is the error level. The main point in the error correction mechanism is that the codon-anticodon complex on the ribosome receives two independent chances to dissociate. The primary codondirected binding of aminoacyl-tRNA is realized as an interaction of the ternary aminoacyl-tRNA · EF-TU · GTP complex with the codon-programmed ribosome; thus, the first selection is accomplished here, on the basis of the interaction difference between the cognate and non-cognate ternary complexes. Then, after the GTP cleavage (which is practically irreversible and so removes the subsequent phase of the process from the equilibrium with the preceding phase), the bound aminoacyltRNA may again dissociate, and the probability of its dissociation will be the higher the worse is its interaction with a codon; hence, if a non-cognate tRNA is bound, it will be released with a higher probability (correction, or proofreading). Therefore, when EF-TU binding and GTP cleavage are involved, the probability of the final binding of a non-cognate aminoacyl-tRNA with the codon-programmed ribosome is equal to the product of the probabilities of its binding at two separate phases of the process. Experimentally, the contribution of EF-TU to the reduction of the translation errors was directly demonstrated in our laboratory (43); it also follows from the increased expenditure of GTP per one peptide bond formed (44-46) and from the augmentation of the number of EF-TU cycles per peptide bond (47) during miscoding. The third aspect of the GTP-dependent catalysis involving the elongation factors is the increase of the resistance of the corresponding steps of the elongation cycle, including drug resistance, resistance to unfavorable shifts of medium conditions, and the capacity to overcome barriers such as a developed secondary structure of a

Energetics and Dynamics of the Protein-Synthesizing Machinery

519

template polynucleotide, etc. (48) (see also (40)). For example, the antibiotic tetracycline, which is an inhibitor of the aminoacyl-tRNA binding step, exerts a halfinhibition of poly(U) translation at a ratio of 50 to 100 molecules per ribosome in the absence of EF-TU (i.e. in the factor-free and the EF-G-promoted translation systems) and at a ratio of about 2000 molecules per ribosome in the presence of EFT u with GTP (i. e. in the EF-T u -promoted and the two-factor-promoted complete translation systems). On the other hand, in order to half-inhibit the poly(U) translation by erythromycin, considered as an inhibitor of translocation, about 100 molecules per ribosome are required in the absence of EF-G (in the factor-free and the EF-T u -promoted translation systems) and at least, 10,000 in the presence of EFG with GTP (in the EF-G-promoted and the complete translation systems). The poly(U) translation is half-suppressed at 0.15 M ethanol in the factor-free system, and only at 1.5 M ethanol in the complete factor-promoted system. EF-TU counteracts the inhibiting effect of a lowered Mg 2 + concentration on the aminoacyltRNA binding step; similarly, EF-G counteracts the suppressing action of increased Mg 2 + concentrations on translocation (40). Translation of template polynucleotides with developed secondary structures more strictly requires the presence of the elongation factors with GTP, as compared with translation of poly(U), poly(U,C), poly(U,I) and poly(A) which are easier translatable in the factor-free mode (24,25,48). All the three apparent manifestations of the involvement of the elongation factors with GTP in elongation seem to be a result of imparting a higher potential to the translating ribosome. Thus, the elongation factors supply the working ribosome with an excess power which is realized in the high rates of the individual steps and the whole process, as well as in their higher reliability.

Conformational Changes of the Translating Ribosome The realization of a mechanical function during translocation (the transfer of rather significant molecular masses from one site to another) and the dependence of the process on the hydrolysis of nucleoside triphosphate into nucleoside diphosphate and orthophosphate suggested some analogy between the ribosome and the muscle. The possibility of "an alternate contraction and expansion of the ribosome" (13) or "a pulsating ribosome contraction" (18) in the elongation cycle was first mentioned by F. Lipmann with associates. Later, continuing this idea, I proposed the hypothesis on a periodical mutual mobility (of the unlocking and locking type) of the two coupled ribosomal subunits as a possible driving mechanism for translocation (49,50). The experimental solution of the question on mechanical changes of the translating ribosome, however, turned out to be not simple, and the first unambiguous results appeared only recently. Firstly, physical measurements required all (or the majority of) the ribosomal particles under study to be functionally active and to represent the same functional state. This is not the case for routine ribosome preparations and routine cell-free translation systems, where only a fraction of ribosomes is active in elongation and

520

Alexander S. Spirin

the ribosomes work asynchronously. The difficulty was overcome only by the end of the last decade owing to the development of a special column technique for preparing translationally active one-functional-state ribosomes (51,52). Secondly, ribosomes in different functional states, e.g. in the pre-translocation and post-translocation states, were found to contain different numbers of tRNA residues per ribosome, to say nothing of the possibility of different positions of the tRNA residues on the ribosome. Hence, small differences in physical parameters, if recorded by hydrodynamic, light or X-ray scattering and other relevant methods, cannot be interpreted unambiguously. For instance, the pre-translocation state ribosomes were shown to have a slightly higher sedimentation coefficient than the post-translocation state ribosomes, but this could be explained either by their higher compactness, or by their higher mass due to the presence of an additional tRNA residue per particle (53). The latter difficulty was overcome by using the technique of neutron scattering with contrast variation (54). The point is that this technique provides a unique possibility to make RNA selectively "invisible" for neutrons (contrast-matched) in the proper H 2 0 — D 2 0 mixture (ca. 70% D 2 0), and thus to watch the behaviour of just the protein moiety. The result was that the post-translocation state ribosomes had a somewhat higher radius of gyration than the pre-translocation state ribosomes, as judged from the difference of their total protein moieties (without elongation factors). In other words, translocation was found to make ribosomes slightly less compact.

Figure 4 Depiction of 70S ribosome model in the overlap projection. The 30S subunit faces the viewer with the head up; the movable L7/L12 stalk of the 50S subunit is directed to the right. The sites of the ribosome interacting with the elongation factors (EF-T„ and EF-G) are shaded. It is likely that the interaction of EF with the distal part of the stalk immobilizes it in a trypsin-sensitive conformation (see the text).

During the last years the indications on more local changes in the ribosome structure during functioning appeared. The most evident among them is the change of the lateral rod-like protuberance (L7/L12 stalk) of the large (50S) subunit of the ribosome (Figure 4) upon interaction with EF-G · GTR Proton magnetic resonance studies have revealed that, while the majority of the ribosomal proteins are firmly fixed within the particle, the protein L7/L12 or its part in situ possesses an

Energetics and Dynamics of the Protein-Synthesizing Machinery

521

independent mobility displayed by the corresponding narrow signals in the PMR spectrum of the intact ribosomes (55,56). In this mobile state, as well as in the isolated state, the protein L7/L12 is relatively resistant against the treatment by trypsin (57). The interaction of the ribosome with EF-G • GTP (a non-cleavable analog of GTP was used in the experiment), however, immobilizes the protein L7/L12 (58) and simultaneously makes it very sensitive to trypsinolysis (57). After the GTP cleavage, even if EF-G continues to be bound with the ribosome (GTP with fusidic acid, which does not hinder the GTP hydrolysis, but retains EF-G · GDP on the ribosome, were used in the experiment), the protein L7/L12 again becomes mobile (58) and trypsin-resistant (57). Hence, EF-G with GTP somehow changes and fixes the conformation of the L7/L12 stalk in a trypsin-sensitive state, and the hydrolysis of GTP by itself, despite the continuing interaction of the ribosome with EF-G, release the stalk conformation. It can be supposed (59) that EF-G · GTP interacts both at the base of the L7/L12 stalk (60) and with its distal part (see Figure 4), and the hydrolysis of GTP releases the distal part and thus allows its mobility. It is not clear yet how this is coupled with the catalysis of translocation. The study of partial contributions of the small (30S) and the large (50S) subunits of the ribosome to the total change of its compactness during translocation (61) has shown that the radii of gyration of the protein moiety (without elongation factors) of the 50S subunit are the same in the pre-translocation and the post-translocation states, i.e. the compactness of the 50S subunit seems to be unchanged. The situation with the 30S subunit, however, proves to be different: the radius of gyration of its protein component somewhat increases, i.e. the compactness slightly decreases, as a result of translocation (61). According to the preliminary data of Serdyuk et al., the contribution of this decrease of the compactness is not sufficient to explain the total change of the radius of gyration observed for the whole ribosome (see above), i.e. it does not abolish the idea (49,50) on some drawing apart (unlocking) of the ribosomal subunits during translocation. Nevertheless, the result indicates that the 30S subunit itself also somehow changes. An independent indication of a change in the 30S subunit is the observation that its two trypsin-resistant proteins, SI5 and SI8, become trypsin-sensitive as a result of the attachment of EF-G · GTP to the ribosome and continue to be trypsin-sensitive after GTP cleavage, until EFG · GDP is released from the ribosome (57). Thus, there is evidence that the translating ribosome is a dynamic structure indeed. The experiments available suggest the existence of at least three types of large-block mobilities: an intersubunit displacement (unlocking) during translocation, a mobility of the L7/L12 stalk governed by elongation factors, and a less identified change in the small (30S) ribosomal subunit.

Ligand Translocations in the Ribosome, in a Multienzyme Complex, and between Soluble Proteins The basic phenomenon of intraribosomal translocation (see Figure 3) seems to be the transfer of the tRNA residue from the A site to the Ρ site (62). There are all grounds to believe that the coupled shift of mRNA by one triplet of nucleotides is

522

Alexander S. Spirin

driven by this tRNA displacement. It is likely that the deacylated tRNA leaves the Ρ site simultaneously, being shifted by the displacement of its bound codon and/or pushed by the incoming peptidyl-tRNA; there is experimental evidence that the acceptor end of the deacylated tRNA may be somewhat moved from the Ρ site even earlier, as an immediate result of transpeptidation (63). Energetic and kinematic aspects of translocation and the corresponding models have been discussed in detail in a special review (62). The question of a molecular pathway of the tRNA transfer from the A site to the Ρ site, however, remains open. It is evident that a large ligand characterized by a multicenter binding with its sites must be passed from site to site through intermediate states, in order to avoid 1. high kinetic barriers at the exit from one site and when entering the other site, and 2. falling out into solution during the transfer. The consideration of examples of a sequential transfer of a metabolite between subunits in a multienzyme complex and a direct "hand-to-hand" passage of a metabolite between soluble enzymes may be helpful for an understanding of the general principles of the translocation pathway. The enzymic machinery for the synthesis of polypeptide antibiotics in Bacillus brevis (8,9,64,65-67) is an example which seems to be the closest to the proteinsynthesizing system. Similarly to the aminoacyl-tRNA synthetases, the enzymes of the antibiotic-synthesizing complexes carry out a two-step activation of the amino acid carboxyls: firstly an amino acid reacts with ATP and is transformed into an enzyme-bound aminoacyl adenylate, and then the anhydride bond is replaced, but by a thioester bond with the protein itself, instead of the ester bond as in aminoacyltRNA in the case of protein biosynthesis. Further, the elongation of a polypeptide on the enzyme, just as on the ribosome, proceeds by the mechanism of the headward chain growth, where, after the formation of a next peptide bond, the last added amino acid ever remains activated. What is a pathway of the growing peptide in this case? The enzyme with a molecular mass of about 100,000 Da binds phenylalanine, transforms it into phenylalanyl adenylate (at the expense of ATP) and then transfers the phenylalanyl residue on a SH group of the enzyme itself (into the thioester bond). The racemization of the L-phenylalanyl residue into D-phenylalanyl is then accomplished on the enzyme. An analogous activation of proline occurs on the other (larger) enzyme (but without racemization). These two enzymes are separate, and their stable complex has never been detected. It is evident, however, that for the subsequent formation of the peptide bond between the two enzyme-bonded aminoacyl residues the two enzymes must directly interact with each other. The formation of a transient pair complex is assumed to occur. The transient interaction must be strictly oriented in order to provide the approaching of the amino group of the prolyl residue to the carbonyl carbon (C á + ) of the phenylalanyl residue. Transpeptidation takes place in this transient couple; the phenylalanyl residue is transferred from the thioester bond with the first enzyme onto the amino group of proline which is in the thioester bond with the second enzyme:

Energetics and Dynamics of the Protein-Synthesizing Machinery

0 ΪΊ^Ι II I E_S C CH - NH_ F ..2 δ+

523

E 0 - S - C - CH - NH p IL \ / C

.E

F

0 ?7H7 Il I - S - C - CH - NH„ „6 + " 2

HO ? 7 H 7 . EF - S - C , - CH - NH_ "2 6+ E ' - S - CC- C H - N II \ / 0"

t 6+ EP' - S - C - CH - NH ft \ / C

E„- SH F

3H6

+

3H6

C

3H6

0 ?7H7 5+ E - S - C - C H - N - C - C H - NH_ P

H

^ y

» 2

0 C

3H6

Thus, the first peptide bond is formed, the enzyme pair dissociates, and the dipeptide remains on the second enzyme. Besides the chemical reaction of transpeptidation, this first cycle in the synthesis of a polypeptide includes the translocation of the phenylalanyl residue from one enzyme to the other. The translocation seems to proceed as a direct, "hand-to-hand" transfer of the residue between the two soluble enzymes. To achieve this, the functional sites of the enzymes must specifically approach each other; it is likely that the residue in an intermediate state of the transfer is shared between the two sites. When the enzymes move apart, the residue remains on the second site. The direct transfer of metabolites between soluble enzymes via transient enzymeenzyme complexes seems to be a more general phenomenon than it used to be thought among enzymologists. During the last years Bernhard et al. have demonstrated and studied the direct transfer of nicotinamide adenine dinucleotide (NAD) between dehydrogenases (68-70), as well as 1,3-diphosphoglycerate between glyceraldehyde-3-phosphate dehydrogenase and 3-phosphoglycerate kinase (71) via an enzyme-substrate-enzyme complex (Figure 5). In addition to the specific pair association of the enzymes and the approaching of their functional sites ("docking"), the mechanism of direct transfer of a ligand has been found to involve a conformational change of the ligand. Thus, if an original substrate conformation fits the binding site of a donor protein, the alternative conformation proves adjusted to the binding site of an acceptor protein. For example, the ^«-conformation of NADH on glyceraldehyde-3-phosphate dehydrogenase is transformed into the anticonformation as a result of the transfer of the co-enzyme to alcohol dehydrogenase or lactate dehydrogenase (69,70). In the case under consideration this is due to the fact that the donor and the acceptor dehydrogenases bind the different (opposite) faces of the nicotinamide ring of NADH, while the binding pattern of the adenylic part of NADH is similar in both.

Alexander S. Spirin

524

E,· M

E2

Ε,·Μ·Εϋ

E,

Ε2·Μ

Figure 5 Schematic representation of the direct ("hand-to-hand") transfer of a ligand (metabolite) between two proteins (enzymes) (after Srivastava and Bernhard (70)). E, and E 2 are two successive proteins in a chain of metabolite transfer; M is a ligand. The proteins "dock" to each other in a specifically oriented manner ("mouth-to-mouth") due to the complementarity of their surfaces around the active centers. The ligand is passed from one active center to the other within the transient E¡ • Μ · E 2 complex, presumably changing its conformation during the transfer.

It is not known how common for other systems are the conclusions drawn from the studies of the direct transfer of NADH between dehydrogenases. Nevertheless, the interaction of two successively occupied binding sites with opposite faces of a ligand is noteworthy. It can be assumed that the intermediate state in the ligand transfer between approached sites, especially in the case of large ligands, is a simultaneous interaction of the ligand or of a part of it with both sites, when one site binds, say, the "right" face of the ligand and the other site the "left". In more detail this possibility will be discussed below in considering the transfer of tRNA; the necessity of a conformational change of a ligand (tRNA) during its transfer will be also discussed there. Returning to the synthesis of polypeptide antibiotics by the enzymic systems of Bacillus brevis I shall dwell on the most studied case, the synthesis of gramicidin S, considered in detail in many reviews; the recent surveys of Kleinkauf et al. should be especially mentioned (66,67). In this case, after the formation of the first peptide bond, D-phenylalanyl-prolyl is found on a cystein residue of the so-called large enzyme (molecular mass of about 300,000 Da) which carries also pre-activated cysteine-bonded residues of valine, ornithine and leucine on its other domains. In addition to these four domains, the large multienzyme includes a domain having a covalently bonded phosphopantetheine residue with its free SH group on the distal end. The phosphopantetheine SH group attacks the thioester bond holding the D-phenylalanyl-prolyl residue, so that the transthiolation occurs and the peptide is found on the phosphopantetheine residue. It is believed that the peptide on the long (2.0 nm) phosphopantetheine arm moves to the next activated cysteine-bonded aminoacyl residue (valine) and offers the newly-formed thioester group to the attack by the amino group. Transpeptidation proceeds, and the tripeptide formed (Dphenylalanyl-prolyl-valyl) is found on the cysteine of the second domain of the multienzyme. Now the freed SH group of the phosphopantetheine residue picks up the tripeptide and transports it to the next amino acid residue (ornithine). And so on. If the large enzyme is represented as a disk with the activated cysteine-bonded amino acids along the periphery and the phosphopantetheine-carrying domain in the center, a "merry-go-round" mechanism for translocation of the growing peptide can be imagined (Figure 6): the thioester bond at the end of the phosphopantetheine

Energetics and Dynamics of the Protein-Synthesizing Machinery D-Phe

525

D-Phe Pro

Leu~S·

-S¡jVal-Pro-D-Phe-» /

Orn

Órn

Orn

3 H

transt h i o l a t i o n , etc

Orn

Figure 6 Schematic representation of the successive transfer of a growing peptide between domains in the heavy multienzyme of gramicidin S synthesis with the use of the movable phosphopantetheine arm (after Lipmann-Kleinkauf (8,72,74)). The scheme shows one cycle including transthiolation (the dipeptide is passed from the proline-activating domain onto the free SH group of the phosphopantetheine arm), translocation (the peptide on the phosphopantetheine arm is moved to the valine-activating domain), and transpeptidation (the free amino group of the valine residue attacks the thioester carboxyl of the peptidyl residue resulting in the formation of the tripeptide bonded to the SH group of the domain and the release of the SH group of the phosphopantetheine arm). Further the transthiolation will occur again, but on the valine-activating domain, and so on. The completion of the third cycle will result in the formation of pentapeptide DPhe-Pro-Val-Orn-Leu on the leucine-activating domain of the multienzyme.

arm (about 2.0 nm away from the center) goes round the activated thiol-bonded amino acid residues, with alternating transthiolations and transpeptidations (8,72). A mechanism of this type was first proposed by Lynen for the multienzyme fatty acid synthase of yeast (73) where the phosphopantetheine residue is also suggested to play the role of a long moving arm. The "protein template mechanism" was further elaborated and spread to other peptide-synthesizing systems (see reviews by Kleinkauf and associates (74-77)). It is apparent that the mechanism of translocation by means of a movable arm in a multienzyme proposed by Lynen and Lipmann is the opposite, in a certain sense, to the mechanism of the direct "hand-to-hand" transfer via the formation of a transient couple of two successive soluble enzymes as suggested by Bernhard. Indeed, in a multienzyme the transfer is virtually accomplished through a medium: a ligand leaves one active site, moves a relatively long distance (by the scale of the protein globule size) in a perienzyme space, and enters another active site. In other words, in the case of soluble enzymes a free diffusional mobility of the proteins themselves is allowed but then the pathway of ligand transfer can be strictly determined and very short, the medium being by-passed; on the contrary, in a multienzyme, protein domains or subunits are more or less fixed, and hence the transfer requires more freedom for ligand movements and a longer translocational pathway (including the overcoming of barriers at the exits and entrances of active sites), than in the former case. Turning to the protein-synthesizing machinery, it can be expected that both cases are realized here. A free aminoacyl-tRNA synthetase activates an amino acid and produces an aminoacyl-tRNA, but the product is rather slowly released into solution, as known for at least some of the synthetases. Therefore the direct transfer of the aminoacyl-tRNA from the synthetase to another soluble protein, EF-T U · GTP, may be a more natural way for the product than its release into a

Alexander S. Spirin

526

Aminoacyl - t R N A synthetase

EF-Ti» (EF-1 ]

Figure 7 A scheme showing the interactions of aminoacyl-tRNA synthetase and EF-T U (EF-1) with two different sides of a tRNA molecule. The tRNA is represented as a ribbon model, the anticodon up and the corner facing the viewer. A large surface of the aminoacyl-tRNA synthetase contacts with the side where the dihydrouridylic loop and helix are located. EF-T U interacts on the other side, mainly covering the "left" side of the Τψ-acceptor limb of the L-shaped tRNA molecule.

Figure 8 Schematic representation of the direct ("hand-to-hand") transfer of aminoacyl-tRNA from the aminoacyl-tRNA synthetase to EF-TU through the formation of the hypothetical transient ARSase • Aa-tRNA · EF-T U · GTP complex. The mutual arrangement of the tRNA and the proteins is as in Figure 7.

medium. Analysis of tRNA interactions with aminoacyl-tRNA synthetases and EF-TU suggests that the tRNA contacts with these two proteins by its different sides. Direct X-ray crystallographic data of the Strasbourg group on the aspartyltRNA synthetase • tRNA A s p complex demonstrate that the synthetase covers predominantly the side of the L-shaped tRNA molecule where the dihydrouridylic loop is localized (the right side if the tRNA is observed from the corner, anticodon up; see Figure 7) (78). At the same time, EF-TU has been reported to protect from chemical and enzymatic modifications the side of the tRNA where Tip and variable loops are situated (the left side in Figure 7), being in contact mainly with the acceptor-Ttp-helix limb of the molecule (79). Consequently, the simultaneous interaction of an aminoacyl-tRNA with both the synthetase and EF-TU is possible (though experimentally not shown). This is just a prerequisite for the direct transfer mechanism. In Figure 8 a hypothetical scheme of the "hand-to-hand" transfer of

Energetics and Dynamics of the Protein-Synthesizing Machinery

527

aminoacyl-tRNA from the synthetase to EF-TU via the formation of the transient ARSase · Aa-tRNA · EF-TU · GTP complex is presented. From general considerations it seems likely that a conformational change of tRNA must occur in the transient two-protein complex in order to prevent a mutual "sticking" of the proteins and to make the complex metastable. The molecule of tRNA can be principally in two alternative conformations, the so-called "closed" conformation which has been demonstrated for crystals of tRNA P h e (80,81) and tRNAp" (82), and the "open" one shown in crystals of tRNA A s p (83). Perhaps, if the "closed" conformation of tRNA is fixed on EF-TU, the synthetase holds tRNA in its "open" conformation.

Figure 9 Schematic representation of the direct transfer of aminoacyl-tRNA from EF-TU to the A site of the ribosome (RS) through the formation of the intermediate RS · Aa-tRNA · EF-TU · GTP complex. The complex dissociates as a result of GTP hydrolysis and loss of the affinity of EF-TU for Aa-tRNA and the ribosome; Aa-tRNA remains on the A site of the ribosome.

The fact of the direct transfer of aminoacyl-tRNA from EF-TU to the A site of the ribosome is well known (Figure 9). In this case, however, the intermediate complex, A-site · Aa-tRNA · EF-TU · GTP, is stable, until GTP cleavage takes place (correspondingly, it is quite stable with a non-cleavable analog of GTP). The cleavage of GTP results in destabilization of the complex and hence the dissociation of EF-TU while the aminoacyl-tRNA remains in the A site. Evidently the GTP cleavage provides a one-way transfer of the ligand from EF-TU to the A site (in contrast to the previous case where there must be an equilibrium distribution of the same ligand between the synthetase and EF-TJ. There are some indications that the transfer of aminoacyl-tRNA from EF-TU to the A site of the ribosome is accompanied by a change of the "closed" tRNA conformation into the "open" one (83,84). The ribosome can be considered as a "multi-site" complex. The mechanism of intraribosomal translocations, particularly the translocation of tRNA from the A site to the Ρ site, is one of the most intriguing problems of ribosomology. If this proceeds according to the type of translocations in multienzyme complexes, the residue of tRNA, being flexibly coupled with the ribosome (through the messenger

528

Alexander S. Spirin

Figure 10 A drawing representing a hypothetical intermediate state in the transfer of peptidyltRN A from the A site to the Ρ site of the ribosome during translocation with the use of the EF-Ggoverned movable L7/L12 stalk. The ribosome is depicted in the nonoverlap projection, the 30S subunit up and the 50S subunit down. The L7/L12 stalk in the initial and final states is directed towards the viewer. In the intermediate state EF-G (with GTP) is depicted as a two-domain protein at the base of the stalk; the stalk is bent to the 30S subunit head; EF-G and the bent stalk together hold the peptidyl-tRNA in a "transition state" of translocation; the deacylated tRNA may be still retained on the ribosome in a rapidly exchangeable state outside the Ρ site.

codon and, perhaps, the peptidyl residue), must first leave the A site and then enter the Ρ site; the distance of the transfer of the central part (corner) of the tRNA will be about 3.0-3.5 nm (85,86). By analogy with the proposed role of the phosphopantetheine arm in the peptidyl translocations from one site to another in the case of the gramicidin S synthetase, it can be speculated that the movable L7/L12 stalk is a transferring arm for tRNA in the ribosome. The flexibility and position of the stalk is governed by the elongation factors (59). It may be that the L7/L12 arm with the help of EF-G · GTP draws out the tRNA from the A site and then, upon GTP cleavage, releases it for binding to the Ρ site (Figure 10). An alternative mechanism of the type of direct ligand transfer between free proteins cannot be excluded either. The question is the possibility of a "hand-to-hand" relaying of tRNA from the A site to the Ρ site. This seems to require such a largeblock mobility of the ribosome that at a certain moment the A and Ρ sites could closely approach each other. Then, the tRNA residue of the A site could directly switch over to the Ρ site, similar to the molecule of NADH when it passes from one dehydrogenase to another (see (69)). Correspondingly, if the A site interacts with the dihydrouridylic ("right") side of the tRNA molecule (see Figure 9), the Ρ site should accept the Τψ ("left") side of the molecule (as in Figure 10). It is noteworthy that the tRNA residue seems to change its conformation from the "open" one in the A site to the "closed" in the Ρ site (87) (see also review in (59)). This is consistent with the mechanism of direct ligand transfer between transiently approached binding sites.

Compartmentation of Eukaryotic Translation Proteins on Polyribosomes Prokaryotic systems are usually implied when aminoacyl-tRNA synthetases, the elongation factors and other translation-serving enzymes are spoken of as free soluble proteins. These proteins, however, are not distributed uniformly in the eukaryotic cytoplasm and extracts of eukaryotic cells, but due to their non-specific

Energetics and Dynamics of the Protein-Synthesizing Machinery

529

affinity for RNA have a tendency to be concentrated (compartmentized) on polyribosomes (88,89). Hence, in the eukaryotic systems, in contrast to the prokaryotic ones, the organization of the protein-synthesizing machinery approaches that of a multienzyme complex but without a strict stoichiometry in the ratios of the synthetases, the factors and other proteins within the polyribosome fraction. Among other proteins associated with eukaryotic polyribosomes the glyceraldehyde-3-phosphate dehydrogenase has been found to be one of the major components; this key glycolytic enzyme, together with the polyribosome-bound phosphoglycerate kinase, is supposed to supply translation with GTP in situ (90). The question arises how the transfer of metabolites and other ligands is accomplished between the proteins adsorbed on RNA. (An analogous problem exists in all cases when successive enzymes of a metabolic chain, e.g. enzymes of glycolysis, are adsorbed on surfaces, such as a membrane or a cytoskeleton.) The point is that the fixation of a protein on some surface seems to be incompatible with a direct transfer mechanism (70) which requires a diffusional mobility of proteins for "docking". On the other hand, the mechanism of a movable arm in a multienzyme complex requires a rigid spatial organization of proteins; this seems to be unlikely in the case of surface-adsorbed enzymes. A solution of this dilemma has been recently proposed in the hypothesis of my collaborator A. G. Ryazanov who assumed that an enzyme (a protein) possesses an affinity for a surface and is capable of adsorbing on it unless the protein is bound with a metabolite (a ligand) (91). Indeed, the eukaryotic threonyl-tRNA synthetase has been shown to lose its affinity for a high-polymer RNA when it carries a bound threonine tRNA (92). The glyceraldehyde-3-phosphate dehydrogenase loses the affinity for RNA and leaves polyribosomes upon NADH binding (93). Many examples of desorbing effects of metabolites on the enzymes labilely associated with membranes and the cytoskeleton can be also given. Thus, it is postulated that a metabolite (a substrate or a product) mobilizes a surface-adsorbed enzyme resulting in its desorption and free diffusion. Now the freely diffusing enzyme can "dock" to the next enzyme which is vacant and so sits on the surface; the product of the first enzyme can be directly handed over to the second one. Then the second enzyme, with its bound substrate, or after transforming it into the product, dissociates from

-è.

\

A site of the ribosome

Figure 12 Schematic representation of the "relay at surface" for aminoacyl-tRNA passing through the polyribosome bound proteins, aminoacyl-tRNA synthetase and EF-1, to the A site of the ribosome. The explanation is in the text.

the surface, while the first one immediately re-adsorbs. This mechanism has been called "relay at a surface" and is schematically represented in Figure 11. As seen, it can provide both for the compartmentation of successively working proteins on and near a surface and for the direct ("hand-to-hand") transfer of metabolites along the enzyme chain. The scheme of the "relay at a surface" for the case of transfer of aminoacyl-tRNA through polyribosome-bound aminoacyl-tRNA synthetase and EF-1 to the A site of the eukaryotic ribosome is given in Figure 12. Since aminoacyl-tRNA synthetases and elongation factors in eukaryotes are found to be associated with polyribosomes, one may consider all this as a united, though labile and mobile, protein-synthesizing complex. Undoubtedly, a mobilization of certain proteins of the complex occurs periodically in the process of functioning to allow their free diffusion and specific "docking" to other proteins or sites of the polyribosome. According to the "relay-at-surface" hypothesis, such a mobilization of polyribosome-bound aminoacyl-tRNA synthetases and elongation factors (i.e., their desorption from polyribosomal RNA surfaces) must be induced by aminoacyltRNA or, in the case of EF-2, by GTR Thus, the eukaryotic protein-synthesizing complex could represent a compromise between two apparently incompatible mechanisms of ligand transfer, namely the Bernhard mechanism of transfer via the formation of transient protein couples (70) and the Lynen-Lipmann mechanism involving an organized multi-domain or multi-subunit particle with fixed remote sites and a movable transferring arm (8,73). It is not excluded that the principle of temporary mobilization of certain subunits by ligands (metabolites) may be realized also in some of well organized multienzyme complexes.

Acknowledgements I am very grateful to my colleagues Alexey Ryazanov and Eugeny Shakhnovitch for fruitful discussions.

Energetics and Dynamics of the Protein-Synthesizing Machinery

531

References 1. Lipmann, F. 1941. Advan. Enzymol. 1, 99-162. 2. Lipmann, F. 1960. In: Molecular Biology (D. Nachmansohn, ed.). Academic Press, New York, London, pp. 37-47. 3. Hoagland, M.B. 1960. In: Nucleic Acids (E. Chargaff and J . N . Davidson, eds.). Academic Press, New York. Vol. 3, pp. 349-408. 4. Lipmann, F. 1968. In: Essays in Biochemistry (P.N. Campbell and G . D . Greville, eds.). Academic Press, London. Vol.4, pp. 1 - 2 3 . 5. Lim, V.l. and A.S. Spirin. 1986. J. Mol. Biol. 188, 565-574. 6. Spirin, A.S. 1986. Ribosome Structure and Protein Biosynthesis. Benjamin/Cummings, Menlo Park, California. 7. Lipmann, F. 1969. Science 164, 1024-1031. 8. Lipmann, F. 1971. Science 173, 875 - 884. 9. Kleinkauf, H. and W. Gevers. 1969. Cold Spring Harbor Symp. Quant. Biol. 34, 805-813. 10. Watson, J . D . 1976. Molecular Biology of the Gene. Benjamin, Menlo Park, California. 11. Keller, E.B. and P.C. Zamecnik. 1956. J. Biol. Chem. 221, 4 5 - 5 9 . 12. Nathans, D., G. von Ehrenstein, R. Monro and F. Lipmann. 1962. Fed. Proc. 21, 127-133. 13. Conway, J.W. and F. Lipmann. 1964. Proc. Nat. Acad. Sci. USA 52, 1462-1469. 14. Allende, J. E., R. Monro and F. Lipmann. 1964. Proc. Nat. Acad. Sci. USA 51, 1211-1216. 15. Nishizuka, Y. and F. Lipmann. 1966. Proc. Nat. Acad. Sci. USA 55, 212-219. 16. Lucas-Lenard, J. and F. Lipmann. 1966. Proc. Nat. Acad. Sci. USA 55, 1562-1566. 17. Lucas-Lenard, J. and A.-L. Haenni. 1968. Proc. Nat. Acad. Sci. USA 59, 554-560. 18. Nishizuka, Y. and F. Lipmann. 1966. Arch. Biochem. Biophys. 116, 344-351. 19. Gordon, J. and F. Lipmann. 1967. J. Mol. Biol. 23, 23-33. 20. Pestka, S. 1968. J. Biol. Chem. 243, 2810-2820. 21. Pestka, S. 1969. J. Biol. Chem. 244, 1533-1539. 22. Gavrilova, L.P. and V.V. Smolyaninov. 1971. Molekül. Biol. (USSR) 5, 883-891. 23. Gavrilova, L.P. and A.S. Spirin. 1971. FEBS Lett. 17, 324-326. 24. Gavrilova, L. P., O. E. Kostiashkina, V. E. Koteliansky, N. M. Rutkevitch and A. S. Spirin. 1976. J. Mol. Biol. 101, 537-552. 25. Rutkevitch, N . M . and L.P. Gavrilova. 1982. FEBS Lett. 143, 115-118. 26. Spirin, A.S. 1978. In: Progress in Nucleic Acid Research and Molecular Biology (W.E. Cohn, ed.). Academic Press, New York. Vol.21, pp. 39-62. 27. Gavrilova, L.P., V.E. Koteliansky and A.S. Spirin. 1974. FEBS Lett. 45, 324-328. 28. Ravel, J . M . 1967. Proc. Nat. Acad. Sci. USA 57, 1811-1816. 29. Ravel, J.M., R . L . Shorey, C.W. Garner, R . C . Dawkins and W. Shive. 1969. Cold Spring Harbor Symp. Quant. Biol. 34, 321-330. 30. Skoultchi, Α., Y. Ono, J. Waterson and P. Lengyel. 1969. Cold Spring Harbor Symp. Quant. Biol. 34, 437-454. 31. Lucas-Lenard, J., P. Tao and A.-L. Haenni. 1969. Cold Spring Harbor Symp. Quant. Biol. 34, 455-462. 32. Arlinghaus, R., J. Schaelfer and R. Schweet. 1964. Proc. Nat. Acad. Sci. USA 51, 1291-1299. 33. Webb, M . R . and J.F. Eccleston. 1981. J. Biol. Chem. 256, 7734-7737. 34. Eccleston, J.F. and M . R . Webb. 1982. J. Biol. Chem. 257, 5046-5049. 35. Kuriki, Y. and A. Kaji. 1967. J. Mol. Biol. 25, 407-423. 36. Hamel, E., M. Koka and T. Nakamoto. 1972. J. Biol. Chem. 247, 805-814. 37. Belitsina, N.V., L.P. Gavrilova and A.S. Spirin. 1975. Dokl. Akad. Nauk SSSR 224, 1205-1208. 38. Odinzov, V.B. and S.V. Kirillov. 1978. Nucl. Acids Res. 5, 3871-3879. 39. Lili, R., J . M . Robertson and W. Wintermeyer. 1986. Biochemistry 25, 3245-3255. 40. Chetverin, A.B. and A.S. Spirin. 1982. Biochim. Biophys. Acta 683, 153-179. 41. Belitsina, N.V. and A.S. Spirin. 1979. Eur. J. Biochem. 94, 315-320. 42. Hopfield, J.J. 1974. Proc. Nat. Acad. Sci. USA 71, 4135-4139. 43. Gavrilova, L.P., I.N. Perminova and A.S. Spirin. 1981. J. Mol. Biol. 149, 6 9 - 7 8 . 44. Thompson, R.C. and P.J. Stone. 1977. Proc. Nat. Acad. Sci. USA 74, 198-202.

532

Alexander S. Spirin

45. 46. 47. 48. 49. 50. 51.

Yates, J.L. 1979. J. Biol. Chem. 254, 11550-11554. Kakhniashvili, D.G., S.K. Smailov and L.P. Gavrilova. 1986. FEBS Lett. 196, 103-107. Ruusala, T., M. Ehrenberg and C.G. Kurland. 1982. The EMBO J. 1, 741-745. Spirin, A.S., O.E. Kostiashkina and J. Jonák. 1976. J. Mol. Biol. 101, 553-562. Spirin, A.S. 1968. Currents in Modern Biol. 2, 115-127. Spirin, A.S. 1969. Cold Spring Harbor Symp. Quant. Biol. 34, 197-207. Baranov, V.l., N.V. Belitsina and A.S. Spirin. 1979. In: Methods in Enzymology (K. Moldave and L. Grossman, eds.) Academic Press, New York. Vol. 59, Part G, pp. 382-398. Baranov, V.l. 1983. Bioorgan. Khimiya (USSR) 9, 1650-1657. Baranov, V.l., L. A. Ryabova and A.S. Spirin. 1984. Molekül. Biol. (USSR) 18, 350-357. Spirin, A.S., V.l. Baranov, G.S. Polubesovand I.N. Serdyuk. 1987. J. Mol. Biol. 194,119-126. Gudkov, A.T., G . M . Gongadze, V.N. Bushuev and M.S. Okon. 1982. FEBS Lett. 138, 229-232. Cowgill, C. Α., Β.G. Nickols, J.W. Kenny, P. Butler, E.M. Bradbury and R.R. Traut. 1984. J. Biol. Chem. 259, 15257-15263. Gudkov, A.T. and G . M . Gongadze. 1984. FEBS Lett. 176, 32-36. Gongadze, G.M., A.T. Gudkov, V.N. Bushuev and N.F. Sepetov. 1984. Dokl. Akad. Nauk SSSR 279, 230-232. Spirin, A.S. 1987. Biochimie 69, 949-956. Girshovich, A.S., T.V. Kurtskhalia, Yu. A. Ovchinnikov and V.D. Vasiliev. 1981. FEBS Lett. 130, 54-59. Serdyuk, I.N. and A.S. Spirin. 1986. In: Structure, Function, and Genetics of Ribosomes (B. Hardesty and G. Kramer, eds.) Springer, Berlin, Heidelberg, New York, pp. 425-437. Spirin, A.S. 1985. In: Progress in Nucleic Acid Research and Molecular Biology (W. Cohn, ed.) Academic Press, New York. Vol.32, pp.75-114. Hardesty, B., O.W. Odom and H.-Y. Deng. 1986. In: Structure, Function, and Genetics of Ribosomes (B. Hardesty and G. Kramer, eds.) Springer, Berlin, Heidelberg, New York, pp. 495-508. Kurahashi, K., M. Yamada, K. Mori, K. Fujikawa, M. Kambe, Y. Imae, E. Sato, H. Takahashi and Y. Sakamoto. 1969. Cold Spring Harbor Symp. Quant. Biol. 34, 815-826. Laland, S.G. and T.-L. Zimmer. 1973. In: Essays in Biochemistry (P.N. Campbell and F. Dickens, eds.) Academic Press, London. Vol.9, p p . 3 1 - 5 7 . Kleinkauf, H. and H. Koischwitz. 1980. In: Multifunctional Proteins (H. Bisswanger and E. Schmincke-Ott, eds.) Wiley, New York, pp. 217-233. Kleinkauf, H. and H. Koischwitz. 1980. In: Chemical Recognition in Biology (F. Chapeville and A.-L. Haenni, eds.). Springer, Berlin, Heidelberg, New York, pp. 427-438. Srivastava, D . K . and S.A. Bernhard. 1984. Biochemistry 23, 4538-4545. Srivastava, D.K., S.A. Bernhard, R. Langridge and J.A. McClarin. 1985. Biochemistry 24, 629-635. Srivastava, D . K . and S.A. Bernhard. 1986. Science 234, 1081-1086. Weber, J.P. and S.A. Bernhard. 1982. Biochemistry 21, 4189-4194. Kleinkauf, H., R. Roskoski and F. Lipmann. 1971. Proc. Nat. Acad. Sci. USA 68, 2069-2072. Lynen, F. 1969. Methods Enzymol. 14, 17-33. Kleinkauf, H. 1979. Planta Medica 35, 1 - 1 8 . Kleinkauf, H. and H. von Döhren. 1981. Curr. Topics in Microbiol. Immunol. 91, 129-177. Kleinkauf, H. and H. von Döhren. 1982. In: Peptide Antibiotics-Biosynthesis and Function (H. Kleinkauf and H. von Döhren, eds.) de Gruyter, Berlin, pp. 3 - 2 1 . Kleinkauf, H. and H. von Döhren. 1983. In: Biochemistry and Genetic Regulation of Commercially Important Antibiotics (L.C. Vining, ed.). Addison-Wesley, USA, pp.95-145. Podjarny, Α., Β. Rees, J.C. Thierry, J. Cavarelli, J.C. Jesior, M. Roth, A. Lewitt-Bentley, R. Kahn, Β. Lorber, J.-P. Ebel, R. Giege and D. Moras. 1987. J. Biomol. Struct. Dyn. 5, 187-198. Wikman, F.P., G.E. Siboska, H.U. Petersen and B.F.C. Clark. 1982. EMBO J. 1, 1095-1100. Kim, S.H., F.L. Suddath, G.J. Quigley, A. McPherson, J.L. Sussman, A . H . J . Wang, N.C. Seeman and A. Rich. 1974. Science 185, 435-440. Robertus, J.D., J.E. Ladner, J.T. Finch, D. Rhodes, R.S. Brown, B.F.C. Clark and A. Klug. 1974. Nature 250, 546-551. Woo, N.H., Β. A. Roe and A. Rich. 1980. Nature 286, 346-351. Moras, D., Μ. B. Comarmond, J. Fischer, R. Weiss, J. C. Thierry, J.-P. Ebel and R. Giegé. 1980. Nature 288, 669-674.

52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63.

64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83.

Energetics and Dynamics of the Protein-Synthesizing Machinery

533

84. Spirin, A.S. and V.l. Lim. 1986. In: Structure, Function, and Genetics of Ribosomes (B. Hardesty and G. Kramer, eds.) Springer, Berlin, Heidelberg, New York, pp. 556-572. 85. Johnson, A.E., H.J. Adkins, E.A. Matthews and C.R. Cantor. 1982. J. Mol. Biol. 156, 113-140. 86. Paulsen, H., J . M . Robertson and W. Wintermeyer. 1983. J. Mol. Biol. 167, 411-426. 87. Jorgensen, J., G . E . Siboska, F.P. Wikman and B.F.C. Clark. 1985. Eur. J. Biochem. 153, 203-209. 88. Spirin, A.S. and M . A . Ajtkhozhin. 1985. Trends in Biochem. Sci. 10, 162-165. 89. Ryazanov, A.G., L.P. Ovchinnikov and A.S. Spirin. 1987. BioSystems. 20, 275-288. 90. Ryazanov, A . G . 1985. FEBS Lett. 192, 131-134. 91. Ryazanov, A . G . and A.S. Spirin. 1988. Biokhimiya (USSR) 53 (in press). 92. Fedorov, A.N., A . G . Ryazanov and L.P. Ovchinnikov. 1986. Biokhimiya (USSR) 51, 1541-1547. 93. Ryazanov, A.G., L.I. Ashmarina and V.l. Muronetz. 1987. Eur. J. Biochem. 171, 301-305.

Expression J. Lawrence Fox

It is a pleasure to be able to contribute to this Festschrift in honor of Dr. Lipmann. When I came to the Lipmann laboratory in 1966 I was interested in learning enough about protein synthesis to begin to understand how the expression of proteins was controlled by cells. While Dr. Lipmann was primarily interested in the contributions to free energy (or whatever else is the appropriate concept for open biological systems) from what I would consider to be the enthalpic side of the equation, my interests were primarily from the entropie side. In particular, I was interested in how information flowed and what the transduction mechanisms were. After more than twenty years, I am still searching for an understanding of these questions. The pioneering experiments of Zamecnik (1) in the 50's demonstrated the requirements for ribonuclear particles in protein synthesis. In 1961 Nirenberg (2) showed that a poly-U message directed the incorporation of a specific amino acid, phenylalanine, into a polypeptide. The experiment of Chapeville et al. (3) proved that the amino acid for polypeptide synthesis was donated from an activated form on a tRNA which was directed during protein synthesis by the tRNA, no longer the amino acid itself. The Lipmann laboratory (4) went on to elucidate the Τ and G factors and to show that Τ factor consisted of a heat stable and a heat unstable component. Marcker and Sanger (5) demonstrated the prevalence of N-terminal methionine. The elucidation of formyl-methionine and the work of the Ochoa laboratory (6) in defining the initiation factors provided an understanding of the initiation process. Cappechi (7) and we (Clelia Ganoza and myself; Ref. 8) demonstrated the existence of specific termination factors for chain termination. The extrapolation of these components to the eukaryotic world was provided by the laboratories of Moldave (9) and Hardesty (10) during the 1970's. During the 70's and 80's the Wittmanns (11), in particular, have focused on the issues of the roles of the ribosomal proteins. By a like token, Schimmel's laboratory (12) has explored the function of amino acyl-tRNA synthetases. Many other laboratories, too numerous to mention in such a succinct review, have examined the differences (and similarities) of protein synthesis up and down the phylogenetic tree. The development of recombinant DNA technology during the 70's has opened the door to permit one to answer many new questions. Since Berg (13) expressed the SV40 Τ antigen in 1974, the power of recombinant DNA technology has been clear and has been applied to both scientific and commercial purposes. While the prokaryotic expression of proteins appeared to follow the Jacob and Monod (14) concepts of induction and repression rather religiously (a number of positive and negative effectors have now been isolated and characterized), the eukaryotic rules proved to be much more complex.

536

J. Lawrence Fox

The most exciting (and far from understood) element to be found in eukaryotic genomes is the intron (15). Why the majority of a structural gene in a eukaryotic system should be devoted to intervening sequences which are spliced out of the messenger R N A utilized in protein synthesis is not clear. The role of R N A itself in catalyzing the R N A splicing process (16) for excision of introns has provided the first evidence of catalytic function residing in molecules other than proteins. Obviously an evolutionary advantage must exist for the role of introns, but exactly what this is far from clear. In addition to the introns, one next finds that the promoter is not sufficient to obtain normal levels of expression. "Enhancer" regions (17) have been defined which may alter expression by up to three orders of magnitude. These enhancers may react to proteins from other DNAs, so the concept of trans-acting factors (18) has evolved as well. Finally, in some systems transcription terminator regions (19) have been identified, so the 5' end, the middle, and the 3' end of eukaryotic messengers play specific roles in protein expression. While these elements largely define the roles of transcription, transcriptional control and translation, the majority of human (and other higher eukaryotic) proteins possess post-translational modifications of their proteins (20). While there are at least 50 different kinds of modifications, glycosylation is probably the most universal of these post-translational modifications. The role of many of these modifications is not obvious. For some proteins, such as erythropoietin (21), the presence of the correct glycosylation is essential for activity. For others, the sugars may be important for secretion, stability or receptor recognition, but may not adversely affect the behavior of exogenous proteins used therapeutically. The development of recombinant D N A technology over the past decade has made it possible to express essentially any protein or mutant protein of interest. The choice of a suitable host cell for this expression is often the most delicate part of the process and entails the greatest effort. The first choice is invariably a bacterium. E. coli has been the preferred bacterium; it can be grown cheaply to high cell density with resultant high protein yields. For very large proteins or proteins with extensive disulfide bonding, the low redox potential of these cells may lead to insoluble inclusion bodies. Recovery of active protein from such inclusion bodies is often poor. E. coli are incapable of most post-translational modifications, so if these are found to be essential to activity, this is not a suitable host cell. Other bacteria are not so widely utilized. Bacillus strains drew considerable attention a few years ago (22) for their potential secreting abilities. Unfortunately, most bacilli produce prodigious quantities of proteases, so to date this has not been a truly attractive alternative. Moving up the evolutionary scale are the yeasts (23). Yeasts can be grown to very high cell densities using well known fermentation technologies. However, the plasmid systems commonly used are not receptive to very large proteins. While yeasts can glycosylate proteins, they apparently are restricted to a rather simple set of reactions and can not produce the more complex glycosylations often encountered in human proteins.

Expression

537

The baculovirus/insect cell systems (24) also have received great press in recent years. Unfortunately, the most interesting work conducted with these expression systems has been done in industry and is not yet available in the literature. The initial promises of high protein yield, protein secretion and complex glycosylation are also largely unmet. Thus, the cells of choice for human protein expression are largely mammalian cell systems. Great strides in the past couple of years have permitted air lift (25) and hollow fiber (26) growth conditions that yield over 100 mg protein per liter of media in, typically, 2 4 - 4 8 hours. Since these will be secreted proteins, the media can be slowly harvested and the cells remain in production for long periods of time. We have some cell lines in continuous production for a year now.

Acknowledgement The knowledge base from which these views arose resulted from support by the National Science Foundation, the National Institutes of Health, the Robert A. Welch Foundation, and Abbott Laboratories.

References 1. Zamecnik, P.C. and E.B. Keller. 1954. J. Biol. Chem. 209, 337. 2. Matthaei, J.H. and M.W. Nirenberg. 1961. Proc. Natl. Acad. Sci. U.S.A. 47, 1580. 3. Chapeville, F., F. Lipmann, G. von Ehrenstein, Β. Weisblum, W.J. Roy, Jr., and S. Benzer. 1962. Proc. Natl. Acad. Sci. U.S.A. 48, 1086. 4. Lucas-Lenard, J. and F. Lipmann. 1971. Ann. Rev. Biochem. 40, 409. 5. Marcker, K. and F. Sanger. 1964. J. Mol. Biol. 8, 835. 6. Salas, Μ., Μ. Β. Hille, J. A. Last, A.J. Wahba, and S. Ochoa. 1967. Proc. Natl. Acad. Sci. U.S.A. 57, 387. 7. Cappechi, M . R . and H . A . Klein. 1970. Nature 226, 1029. 8. Fox, J.L. and M . C . Ganoza. 1968. Biochem. Biophys. Res. Commun. 32, 1064. 9. Rao, P. and K. Moldave. 1969. J. Mol. Biol. 46, 447. 10. Ravel, J.M., R.C. Dawkins, Jr., S. Lax, O.W. Odom, and B. Hardesty. 1973. Arch. Biochem. Biophys. 155, 332. 11. Wittmann, H . G . 1983. Ann. Rev. Biochem. 52, 35. 12. Schimmel, P.R. and D. Soil. 1979. Ann. Rev. Biochem. 48, 601. 13. Cole, C.N., T. Landers, S.P. Goff, S. Manteiul-Brutlag and P. Berg. 1977. J. Virol. 24, 277. 14. Jacob, F. and J. Monod. 1961. J. Mol. Biol. 3, 318. 15. Nevins, J . R . 1983. Ann. Rev. Biochem. 52, 441. 16. Cech, T.R. 1983. Cell 34, 713. 17. Gluzman, V. and T. Shenk. 1983. Enhancers and Eukaryotic Gene Expression. Cold Spring Harbor Laboratory. 18. Tijan, R. 1981. Cell 26, 1. 19. Platt, T. 1986. Ann. Rev. Biochem. 55, 339. 20. Wold, F. 1984. Trends Biochem. Sci. 9, 256. 21. Lin, F.-K., S. Suggs, C.-H. Lin, J. K. Browne, R. Smalling, J. C. Egrie, K. K. Chen, G. M. Fox, F. Martin, Z. Stabinsky, S.M. Badrawi, P.-H. Lai and E. Goldwasser. 1985. Proc. Natl. Acad. Sci. U.S.A. 82, 7580. 22. Doi, R . H . 1984. Biotechnology and Genetic Engineering Reviews 2, 121. 23. K. Struhl. 1986. In: Maximizing Gene Expression (W. Reznikoff and L. Gold, eds). Butterworths, Boston, p. 35.

538

J. Lawrence Fox

24. Summers, M.D. and G.E. Smith. 1987. A Manual of Methods for Baculovirus Vectors and Insect Cell Culture Procedures. Texas Agricultural Experiment Stations Bulletin No. 1555. 25. Arathoon, W.R. and J.R. Birch. 1986. Science 232, 1390. 26. Ku, K., M.J. Kuo, J. Delente, B.S. Wildi, and J. Feder. 1981. Biotechnol. Bioeng. 23, 79.

Punctuation in the Genetic Code: A Plausible Basis for the Degeneracy of the Code to Initiate Translation M. Clelia Ganoza

Introduction It is perhaps not always realized that the punctuation of the genetic code (the sequences that begin and end translation) is as important as the code itself. Proteins cannot be made until a specific sequence of bases in mRNA is placed in the correct reading frame and until the correct non-coding sequences are reached in-phase for signalling the end of protein synthesis. The features in mRNA that signal the correct starting point are not well understood; particularly unclear still is the solution to the complex code which begins the translation process. Similarly, part of the code that punctuates the end of translation has been defined, as have aspects of the mechanism of this reaction, but genetic and biochemical work indicate that some termination triplets are ambiguous in that they specify insertion of an amino acid as well. In some cases this device is used to make two proteins from one mRNA. A recent model of the reaction predicts the features of mRNA that may govern such regulation. 1 Genetic, biochemical data and theoretical considerations briefly discussed here lead us to propose that the code to initiate protein synthesis is flexible and degenerate. Certain persistant features of this code permit it to be categorized into classes. The obvious corollary is that the recognition system for start is also flexible in a way that can accommodate the different classes of start signals. We have searched for the one reactant, the "reader" for initiation which has the properties of constancy through evolution and the flexibility to sense or recognize the various classes of start signals that we observe. We propose it to be tRNAi. We have deduced a model for the initiation reaction constistent with the following: 1. certain features of the primary sequence of tRNAi distinguish it from elongator tRNAs; 2. unlike elongator tRNAs the coding properties of tRNAi are unique, 3. mutations that disrupt tRNAi · mRNA interactions affect translation, 4. a statistically significant number of signals complementary to tRNAi occur in prestart sequences and 5. signals complementary to the tRNAi loops enhance prediction of translational reading frames. Below are summarized observations that lead to these conclusions.

1

Ganoza, M.C. et al. (1984) J. Biol. Chem. 259, 14101-14104.

540

M. Clelia Ganoza

Variant and Invariant Features of the Genetic Code and Its Recognition It is generally accepted that the genetic code is, with but minor exceptions2 (1,2) a universal code (3). Sequence analysis of DNA and RNA genomes and their corresponding proteins lends strong support to this general idea. The paradox is that the apparatus that deciphers the genetic code has been considerably altered by evolution, resulting in ribosomal particles of totally different sedimentation value, immunological cross-reactivity, as well as different sequence in regions of the RNA that form about two thirds of these particles (4). Strong functional evidence for the dramatic change in the decoding apparatus comes from the observation that not all mRNAs can be dedoced by all ribosomes. Thus, although small amounts of proteins of the correct size can be made by hepatic extracts or cultured animal cells programmed with RNA bacteriophages, aberrant levels of their synthesis and relative expression are the rule rather than the exception (5 a 5g). Furthermore, most bacterial extracts, specifically those of E. coli, are unable to initiate the synthesis of globin or many other eukaryotic proteins (6). Although translation efficiency is enhanced by addition of m 7 G to the 5' side of some mRNAs (7,8) the general statement holds that one cannot predict whether a given mRNA will be faithfully expressed in one species or another. Hence, although the code is universal, the apparatus that deciphers this code has changed in its recognition properties. Perhaps the most important single reaction to examine in this regard is the initiation of the translation process.

Common and Diverse Features of the Signals Specifying Initiation of Translation. In a functional sense the translational start codon defines the genes of an organism because it specifies the start of its proteins. Intuitively then, one expects the start domain of all proteins to share common features. Conversely, the flexibility of such features might be used to advantage by increasing the genetic coding potential. Sequence studies indicate that the most conserved single feature of the translational start domain is the AUG triplet itself which codes for tRNAi. Of importance is the observation that certain unusual triplets UUG, GUG, AUU, U U G and UAA(A) (the latter only in mitochondria) can also, less commonly, precede the N-terminus of proteins. One of the problems which we view as a universal is that the initiation codons are ambiguous, since they specify initiation as well as insertion of amino acids into internal portions of proteins. These codons also occur randomly at out-of-phase sites in most mRNAs studied. It has been argued (9) that eukaryotic cells have a 2

It has been suggested that human mtDNA differs from the universal code in that UGA codes for methionine not isoleucine, and AGA as well as AGG are terminators rather than arginine codons (1,2).

Punctuation in the Genetic Code

541

specialized mechanism for initiating translation such that the 5' proximal end of the m R N A [usually a 7-methyl-guanine (CAP)] is recognized followed by migration of the 40 S particle to the "correct" A U G (9). Some experimental evidence support this idea (9,10,10a). However, m R N A s of eukaryotes sometimes contain a fairly long region of non-coding bases with more than one A U G . Hence eukaryotic ribosomes may also have to discriminate between the A U G s that specify initiation and those that do not. Furthermore, capping of m R N A does not resolve this problem as many m R N A s do not contain caps and are expressed as efficiently, often more efficiently than those that do have the capped structure (9). Finally, two other examples are noted, the first being the case of bicistronic m R N A s (9) and the second is the existence of overlapping genes in viruses that normally grow on eukaryotic hosts (11,12). The point clearly established by the latter observation is that the coding issue is fundamentally the same in both eukaryotic and prokaryotic cells, since ribosomes must recognize an internal A U G as a starter. This view of the mechanism of initiation differs from ones where the monocistric character of most eukaryotic m R N A s or the position of the starting triplet near the 5' end is emphasized over the inherent problem of recognizing an ambiguous start codon (9,10). Another difference between eukaryotic and prokaryotic ribosomes that has frequently been noted is that the latter often contain a purine-rich sequence at the 5' side of the A U G which is complementary to the 16S r R N A 3' end ( the "Shine and Dalgarno" region) (13). In prokaryotes, most current work on the recognition of start signals has focussed on examining the untranslated clusters of nucleotides which contain the "ShineDalgarno" region at the 5' termini of the codons specifying the N-terminus of proteins (14-16). Several hundred ribosome binding sites have been sequenced from prokaryotes in an attempt to determine the characteristics essential to an initiation region. We have examined several hundred of these intercistronic regions by computer-aided techniques (17). From this collection, I selected a few sequences and placed them into arbitrary classes to emphasize a few salient features. Of the sequences examined, relative to A U G , G U G and U U G are used rarely to initiate synthesis. The nucleotide sequences of the intercistronic regions examined are all different. They vary in lenght from overlap to a few hundred bases (14-16). Overlapping genes can occur in or out of the aligned reading frame (24-26). Introduction by mutation of a nonsense codon can sensitize ribosomes to re-initiate translation with either AUG, U U G or G U G (27). Very few messengers of prokaryotes can translate in the apparent absence of a leader sequence (28,28 a). Many intercistronic regions contain an in-or out-of-phase termination codon usually, but not always, U A A or U G A (18,19). A striking common feature of the start sites is that many of them can pair with the 3' OH end of the 16 S rRNA; however, this region of complementarity is variable in length. The suggestion of Shine and Dalgarno that this pairing is essential has received much experimental support from both genetic and biochemical experiments (14-17).

542

M. Clelia Ganoza

Examination of the intercistronic sequences as well as internal sequences shows that complementarity with the 16S rRNA cannot account for specificity. First, a large number of regions in mRNAs, most of which do not code for protein starts, complement the 16 S rRNA and code for internal methionines, valines or leucines (14-23). Secondly, some of the regions of complementarity are not sufficiently stable to account for the fact that the rate of protein synthesis increases at temperatures that melt weak RNA-RNA interactions (16). Furthermore, several normal E. coli proteins lack this region of complementarity in the leader sequence preceding the start AUG (29-32). We observe no correlation between the degree of complementarity with the 16 S rRNA on one hand and the efficiency of translation on the other. For example, about twenty times more coat protein (where 3 or 4 bases complement) (33) is made by all RNA bacteriophages than maturation protein (where 8 bases can pair the 16S rRNA). Finally, many eukaryotic mRNAs do not contain regions of complementarity with the exposed 3' end of 18 S rRNA (9,10). It is possible that either a larger sequence or spatially restricted signal specifies initiation. To date no unique secondary structure has been observed in a large number of mRNAs that have been examined (15-17). Interestingly, the potential local secondary structure of start regions is markedly different from that of the mRNA coding regions (17). A good correlation has been found between the efficiently translated messengers and the tendency of their start codons to be unpaired with other bases of the intercistronic region. On the other hand, little correlation, either positive or negative, relate poorly expressed mRNAs and the position of AUG in the stem and loop structures of their start regions (16,17). We conclude from this survey that secondary structural features can be important determinants of efficient initiation regions, but - as with the Shine and Dalgarno region - are not always sufficient and are sometimes not necessary. The relative position of AUG and the Shine-Dalgarno domain allow arbitrary classes of secondary structures to be drawn (34) but their interpretation remains obscured by the fact that many alternate structures can be drawn from each sample. It is possible that ribosomes sense a common higher-order structure in mRNA but there is no evidence, so far, on this issue (14-17). It is then quite logical to examine whether other components of the ribosomal apparatus, particularly those involved in the process of codon recognition, can impart further specificity to initiation. A special tRNA is required for initiation of translation in all organisms so far investigated (35). Phylogenetic relationships between phe-tRNA and methionine (tRNA¡) have been reported (36,37). Inspection of the initiator (tRNAi) has revealed that this molecule exhibits a remarkable degree of homology in a number of species studied (38). This high degree of sequence conservation suggested to us that tRNAi may contain essential information that need not reside in the anticodon loop itself. Closer inspection of tRNAi reveals other features that may be important in discriminating initiation signals. We note first that although the tertiary structure of E. coli tRNAi is not significantly different from that of yeast phenylalanyl-tRNA, there are

Punctuation in the Genetic Code

543

features in the linear structure that tend to characterize it and distinguish it from elongator-tRNAs (39-41). The anticodon, D and Τ loops are remarkable in this respect. We call attention to the following points: 1. The absence of modified bases in the anticodon loop of E. coli t R N A i could permit the pairing of portions larger than the anticodon with m R N A s ( 4 2 - 4 6 ) . In some messengers, rather extensive signals for the anticodon loop are found. They could compensate for the lack of an extended ribosome binding sites in the short m R N A s of Acl and 434 cl phage repressor (28,28 a) (See Figure 1 for examples). 2. The D loop of tRNAi contains only one dihydrouridine (which base pairs as U), leaving seven nucleotides around the triplet C C U available for pairing. Indeed, it has been shown that the region, itself can bind complementary oligonucleotides

A.

E.

COLI

GAL E

E.

COLI

TN903

PHAGE

434

LAMBDA

B.

C.

EL

GAL

UGUGAAGAUUGGGGGUAAAUAACAGAGGUGCGUUAUGAGUAUUUCUUCCAGGGUAAAAAGCAAAAGAAUU.

A

AUAACAAAAAUAGCCUUCCUCUAAAGGUGGCAUCAUGACUGUUCAAGCUGAAAAAAAGCACUCUGCAUUUU CGUUCCUCGUGCUUAGUAACUAAGGAUGAAAUGCAUGUCUAAGACAGCAUCUUCGCGUAACUCUCUCAGCG

Τ

CCAUCCACAGGGAUAUCCCGAUUAAGAAACGACCAUGACGCAAUUUAAUCCCGUUGAUCAUCCACAUCGCC

BETA A

ACACAGCAGUACUUCACUGAGUAUAAGAGGACAUAUGCCUAAAUUACCGCGUGGUCUGCGUUUCGGAGCCG

L A M B D A Ν 111 E.COLI

RPL

UACGGGGCGGCGACCUCG£SS£I111UUCGCUAUUUAU6AAAAUUUUCCGGUUUAAGGCGUUCCGUCUCUCGC Ν (LIHÎ

SPC)

LAMBDA 0 E.COLI

D.

E.

GUCUGCUUACAUAAACAGUAAUACAAGGGGUGUUAUGA6CCAUAUUCAACGGGAAACGUCUUGCUCG

AGAUAUUUAUCCCUUGCGGUGAUAÊAUIIUAACGU^AUGAGCACAAAAAAGAAACCAUUAACACAAGAGCAGC

BETA R E P L I C A S E

E.COLI 0

R.

CL

E . C O L I TNA 0

AUACCAUAAGCCUAAUGGAGCGAAUUAUGAGAGUUCUGGUUACCGGUGGUAGCGGUUACAUUG KANAMYCIN

GUCUCAGUAGUAGUUGACAUUÌÈCGGAGCCUAAAAUGAUCCAAGAACAGACUAUGCUGAACGUCGCCGACA ACCUAACAUUGAUU£4BS11ACAGGGAGAAGGCGCAUGAGACUCGAAAGCGUAGCUAAAUUUCAUUCGCCAA

ST

TN1681

E . C O L I RPL E

I

CAAAUAUCCGUGAAACAACAUGACGGGAGGUAACAUGAAAAAGCUAAUÊIIUËÊCAAUUUUUAUUUCUGUAU

(L5)

ACAGCGAAACUAUCAAGUAAUUUGGAGUAGIIACGAIIGGCGAAACIIGCAIIGAIIIIAr.llACAAAGArGAAGIlAG

E.COLI

TUF

A

(EF

Tu)

E.COLI

TUF Β

(EF

Tu)

UGAAGGGGAGAGCACAAUAGUAAGGAAUAUAGCCGUGUCUAAAGAAAAAMIIMACGUACAAAACCGCACG UCACCGAUUUAUCCGUGUCUUAGAGGGACAAUCGAUGUCUAAAGAAAAGUUUGAACGUACAAAACr.Gr.ACG

LAMBDA G I T

CUAUUACAACCCCUACA£U11U£AUGAGUAUAGAAAUGGAUCCACUCGUUAUUCUCGGACGAGUGUUCAGUA

MS2

UCUCUAGAUAGAGCCCUCAACCGSAFILLUUSIIGCAUGGCUUCUAACUUUACUCAGUUCGUUCUCGUCGACA

COAT

FD V I I I E.COLI

UUACGUAUUUUACCCGUUUAAUGGAAACUUCCUCAUGAAAAAGUCUUUAGUCCUCAAAGCCUCCGUAGCCG ARO Η

CUGCCGUAGAAGCAACAAAUUUCUGAGACUUGUAAUGAACAGAACUGACGAACUCCGUACUGCGCGUAUUG

E.COLI

1 6 S RNA

E.COLI

TRNAI

3'

END(3'-5')

(3"-5' )

AUUCCUCCACJAGGUUGGCGUCC... ACC 3'-OH kinase reaction which is irreversible (13). Those containing no common nucleic acid bases or substituted at the vicinal 2'-OH, for example FMN, NADP, 2-O-methylAMP or ribose-5-phosphate, are inert (3). These properties of Streptomyces enzymes distinctly contrast with those of above-mentioned two classes of enzymes and seemed puzzling regarding their intracellular functions. Molecular sizes also differ characteristically among these three classes as 75, 55 and 25 kDa, respectively. In spite of its puzzling nature and in view of the then emerging stimulating results on ppGpp and related subjects, and also being circumstantially prompted, it was readily conjectured that Streptomyces enzyme and its variety of 3'pyrophosphorylated reaction products, especially those base-analogous to ppGpp, might have some specific biological functions in such as cellular development or differentiation, intercellular communication or transmembrane control. (p)ppApp had been implicated in morphological and physiological changes of bacterial cells such as cellular elongation and sporulation (14,15). So we prepared these nucleotides using purified enzyme and carried out some preliminary experiments to see their possible effects on various kinds of gene expression machinery and subcellular activities using mostly cell-free systems. Structural identification of 3'-pyrophosphates was a simple task: 3'-phosphate linkage specific Pénicillium nuclease P¡ liberated inorganic pyrophosphate stoichiometrically. Several typical results will be presented here. To begin with, Streptomyces pyrophosphokinase was found not to respond to stringent control in vitro·, the enzyme was neither activated nor inhibited in either of the two reactions tested (ATP + GTP -> AMP + pppGpp) or (ATP + ATP-» AMP + pppApp ->· pApp + pApp) either in the presence or absence of homologous unloaded mixed tRNAs and ribosomes complexed with poly-A,U,G. We then tested effects of eight kinds of (p)ppNpp (N = A,G,C,U) on E. coli-DNA directed stable RNAs and protein syntheses by coupled transcription-translation system from S 30 lysate of E. coli CP 78 relA + (16). Only (p)ppGpp strongly inhibited both the syntheses while the others all were without appreciable effects. None of these eight nucleotides affected poly-U dependent phenylalanine incorporation by the same lysate. These results may indicate that only (p)ppGpp are the sole natural effectors

Enzymology of 3'-Squiggled Nucleotides

605

among all the tested nucleotides though not necessarily excluding the possibility that we have observed only the overall results. Further detailed analysis using respective markers is needed. Streptomyces lysate systems under development may shed some light on this problem. Our interest in natural occurrence and physiological activity of 3'-pyrophospho nucleotides in general further prompted us to examine their effects on eucaryote systems also. Although no significant effect was noted with SV40 DNA fragments-directed transcription by Heia cell lysate, unexpected and surprising was the finding that ppCpp specifically stimulated silkworm chorion mRNA-coded chorion synthesis by wheat germ lysate as much as 2.5 fold (17). We do not know about its natural occurrence in insects presently, but we nevertheless consider recently reported occurrence of ppGpp, ppApp (18) and adenosine 3'phospho «-butyl diester (19) as growth regulators in a lower eucaryote yeast cells as well as in procaryotes and of oligo-3'-phospho glyceroyl-ATP in rat (20) are rather pleasing signs on our side. More of this kind of nucleotides containing or derived from 3'-pyrophosphate may continue appearing. Our studies at the enzyme activity level were more clear-cut, resulting in conclusive data. Thus, when NAD and FAD were reacted with dATP by pyrophosphokinase, they yielded the respective adenosine 3'-pyrophosphorylated products. They were isolated on DEAE-cellulose column and liberated pApp on venom pyrophosphatase digestion to structural identification. Spectrophotometic cofactor activity assay using yeast alcohol dehydrogenase and hog kidney D-amino acid oxidase apoenzyme respectively showed total inactivity of both the 3'-pyrophosphates either as the coenzyme or as inhibitor (21). Our interest was then focused on CoA when I first became associated with Lipmann in 1978. To prepare 3'-pyrophospho CoA (CoApp), dephospho-CoA was employed as the acceptor. The resultant CoApp was isolated on DEAE-Sephadex column and precipitated as lithium salt with alcohol. The coenzyme activity determination using phosphotransacetylase from CI. kluyveri, E. coli and Leuconostoc mesenteroides invariably gave identical, equally surprising results: CoApp was found about twice more active than normal (3'-monophospho) CoA (11). According to the kinetic analysis of the reaction CoApp very likely interacts with the same site on the enzyme molecule as does normal CoA with the affinity about twice increased (22). As adenosine 3'-monophospho-5'-diphosphate moiety of CoA was recently implicated in providing energy for binding of acyl CoA substrates to the CoA transferase protein (23), a similar mechanism arising from the more highly energized adenylic acid moiety might explain the enhanced activity of CoApp. It is also well known that 3'-dephospho CoA generally has very low catalytic activity. Since the reaction product acetyl phosphate is an essential factor in osmotic-shock sensitive glutamine transport mechanism of some bacterial strains (24), CoApp may have a role in some phases of bacterial nitrogen economy as ppGpp in stringent control. When bacteria are starved for nitrogen, the transport machinery must work more efficiently. This might be aided by CoApp, enabling increased synthesis of acetyl phosphate. We then chose E. coli phosphoenolpyruvate carboxylase as the next target of CoApp vs. CoA. The enzyme is an allosteric enzyme playing anaplerotic role by replenishment of oxaloacetic acid to tricarboxylic acid cycle. Acetyl CoA, fructose

606

Jun-Ichiro Mukai

1,6-bisphosphate and some fatty acids are known positive effectors. The enzyme is also subject to stringent control, being activated by ppGpp. Our experimental results with a purified enzyme were far more convincing of the enhanced activity of acetyl CoApp relative to the normal counterpart, acetyl CoA or ppGpp as an effector (22). That is, the extent of activation was not only much higher than that by ppGpp on molar concentration basis but also about six times higher than that by acetyl CoA in terms of the affinity to enzyme molecule. This in turn led to about 5 fold increased affinity of the enzyme to phosphoenolpyruvate. This must be a considerable expedience for bacteria facing amino acid deficiency in view of the physiological role of the enzyme. We add that no enzyme tested has ever refused CoApp; it was more or less (up to two-fold) preferred to normal CoA, or accepted at least at the same rate, never at a slower rate. The enzymes tested are acetyl CoA synthetase from yeast, succinic thiokinase from porcine heart, and citrate synthases from the former and pigeon breast muscle. Various attempts, however, have been unsuccessful thus far in identifying CoApp in Streptomyces cells grown under different conditions. In parallel with CoApp experiments, we conducted an analogous test with pAp as the cofactor of sulfate transfer (11). Superphosphorylation of pAp to pApp completely abolished the cofactor activity for a sulfotransferase purified from dog liver. Now it seems that 3'-pyrophospho nucleotides are mainly concerned with cellular signaling and regulatory roles while 5'-counterparts energetic processes. Finally we must ask ourselves. What kinds of roles does Streptomyces pyrophosphokinase play? What are the natural substrates or products of the enzyme? Do these 3'-pyrophosphoryl nucleotides, especially CoApp and ppCpp, really exist in nature? We have ever so many things to do to answer these questions.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

Lipmann, F. 1971. Adv. Enz. Reg. 9, 5. Sy, J. 1974. Proc. Natl. Acad. Sci. U S A 71, 3470. Mukai, J. I., Kukita, T., Murao and S. Nishino, T. 1978. J. Biochem., Tokyo. 83, 1209. Oki, T., Yoshimoto, Α., Ogasawara, T., Sato, S. and Takamatsu, A. 1976. Arch. Microbiol. 107, 183. Lipmann, F. and Sy, J. 1976. Progr. Nucl. Acids Res. Mol. Biol. 17, 1. Cashel, M. 1975. Ann. Rev. Microbiol. 29, 301. Sy, J. 1976. Biochemistry 15, 606. Richter, D., Fehr, S. and Harder, R. 1979. Eur. J. Biochem. 99, 57. Murao, S. and Nishino, T. 1974. Agr. Biol. Chem. 38, 2483. Mukai, J. I., Razzaque, Α., Hanada, Y., Murao, S. and Nishino, T. 1980. Anal. Biochem. 104, 136. Mukai, J.I., Sy, J. and Lipmann, F. 1983. Proc. Natl. Acad. Sci. U S A 80, 2899. Simoncsits, András. 1980. Nucl. Acids Res. 8, 4111. Mukai, J.I., Razzaque, Α., Kukita, T., Murao, S. and Nishino, T. 1978. Nucl. Acids. Res. sp. pubi. 5, s451. Hamagishi, Y., Tone, H., Oki, T. and Inui, T. 1980. Arch. Microbiol. 125, 285. Murao, S., Kameda, M., Nishino, T. and Hamagishi, Y. 1980. Agrie. Biol. Chem. 44, 2773. Mukai, J.I. and Koguchi, S. 1982. FEBS Lett. 141, 251. Mukai, J.I., Furuichi, M., Murao, S. and Nishino, T. 1981. Agrie. Biol. Chem. 45, 1775.

Enzymology of 3'-Squiggled Nucleotides 18. 19. 20. 21. 22. 23. 24.

607

Silverman, R.H. and Atherly, A.G. 1979. Microbiol. Rev. 43, 27. Kawaguchi, T., Asahi, T., Satoh, T., Uozumi, T. and Beppu, T. 1984. J. Antibiotics. 37, 1587. Hutchinson, W.L., Ratclifle, P.J. and Mowbray, J. 1986. Biochem. J. 240, 997. Razzaque, Α., Hanada, Y., Mukai, J. I., Murao, S. and Nishino, T. 1978. FEBS Lett. 93, 335. Mukai, J.I. 1984. Nucl. Acids Res. sp. pubi. 15, s77. Fierke, C. and Jencks, W.P. 1986. J. Biol. Chem. 261, 7603. Hunt, A.G. and Hong, J.S. 1983. Biochemistry 22, 844.

Pyridine Nucleotides as Group Transfering Coenzymes H.

Hilz

Introduction The pioneering work of Otto Warburg and Hans von Euler in the thirties established two 'cozymases', D P N (NAD) and TPN (NADP), as essential components of tissue respiration and led to their isolation and identification (1-3). Subsequently, the fundamental role of these pyridine nucleotides as hydrogen transfering coenzymes of many oxido reductases was discovered. In the last two decades, it became apparent that the chemical potential of these coenzymes comprised more than a reactive pyridine for hydrogen transfer. According to Fritz Lipmann's concept of energy-rich bonds (4), N A D as well as N A D P posess two such bonds: a glycosidic linkage of ADP-ribose to the quarternary nitrogen of the nicotinamide moiety which exhibits a potential of 8.2 kcal/mol (5), and a pyrophosphate group with an energy store of about 7 - 8 kcal/mol. Thus, the pyridine nucleotides are potential group transfering coenzymes and when we look at these molecules with "squigglesensitized" eyes (Figure 1), activated forms of at least four different groups are seen. Here, the role of N A D as group transfering cosubstrate will be reviewed. Attention will also be focused on two new types of covalent modification that originate f r o m the activated components in NADP.

NAD-derived Group Transfer In 1966 and 1967, reports from three laboratories described a novel type of a polymer, poly(ADP-ribose), that was formed in nuclei at the expense of N A D (6-8). Two years later, transfer of a single ADP-ribosyl residue from N A D to elongation factor II by diphtheria toxin was demonstrated (9). Mammalian tissues also contain various classes of such mono(ADP-ribosyl) protein conjugates that exhibit different chemical sensitivities and independent changes (10), thus demonstrating that cellular mono(ADP-ribosyl)ation reactions comprise more than the initiation sites for poly(ADPR) formation (11). An additional reaction with the intriguing feature of involving N A D in a non-oxidative manner was also found in the sixties: Ligation of D N A strands in E. coli by a utilization of the second energyrich bond in N A D (12-14). The three types of NAD-derived group transfer will be discussed in separate sections.

Abbreviations: ADPR = ADP-ribose, phospho ADPR = 2'-phospo ADP-ribose

H. Hilz

610

I 1

I 1

active AO PR

I I — Figure 1

active A M P

' — a c t i v e 2 p h o s p h o AMP active 2 ' - phospho ADPR

Activated groups in pyridine nucleotides.

Mono(ADP-ribosyl)ation reactions Transfer of ADP-ribosyl group in N A D to protein acceptors represents a classical type of covalent modification. The attachment of a charged group to a specific acceptor amino acid residue is thought to impose marked changes in structure and function of the modified polypeptide (15-17). Dramatic functional changes associated with ADP-ribosylation are seen especially in the case of bacterial toxins that acquire ADP-ribosyl transferase activity by a host-mediated activation step

Pyridine Nucleotides as Group Transfering Coenzymes Table 1

611

Mono(ADP-ribosyl)ation reactions

ADP-ribosyl transferase

Acceptor

Target amino acids

Affected process

Referenees

A. Prokaryotic Enzymes Diphtheria toxin Pseudomonas aeruginosa toxin P. aeruginosa exoenzyme S

elongation factor II elongation factor II elongation factor I (?)

diphthamide diphthamide ?

protein synthesis protein synthesis protein synthesis

9 36 37

Cholera toxin E. coli heat labile enterotoxin

G , proteins; transducin G s proteins; transducin

arginine arginine

adenylate cyclase; photoreception

38-40 41,42

Pertussis toxin

G¡ proteins; transducin

cysteine

43,44

Botulinum toxin C2 Botulinum toxin Botulinum neurotoxins Clostridium difficile toxin Β

actin protein P20 membraneous P21 protein P90

7 ? ? ?

adenylate cyclase; photoreception cytoskeleton ? ? microfilament ?

Bacteriophage T4

R N A polymerase and other E. coli proteins dinitrogenase reductase

arginine

host RNA-synthesis

36, 52

? arginine

? nitrogen fixation

53 35

arginine arginine

?

Bacteriophage N 4 Rhodospirillum r. transferase Eukaryotic Enzymes avian erythrocyte transferase avian nuclear transferase

mammalian cytoplasmic transferase mammalian cytosolic transferase mammalian cytosolic transferase

45-47 48 49, 50 51

7

22 54

arginine

?

55

?

cysteine

7

30

elongation factor II

diphthamide

protein synthesis

21

multiple proteins histone H I , glycogen phosph. kinase, h. c-Ha-ras protein multiple proteins

(18). Specific proteins and distinct amino acid residues serve as acceptors of ADPR groups under the catalytic action of the various toxins. The consequences are in almost all cases irreversible interference with catalytic and regulatory functions of the host cell. Table 1 summarizes important examples of toxin-mediated and bacteriophage-derived ADP-ribosylation reactions and their effects. Target specific ADP-ribosyl transferases like Cholera and Pertussis toxins also provided powerful tools for the elucidation of transmembrane signaling and the function of G proteins (cf. 19). The irreversible nature of toxin-catalyzed ADP-ribosyl transfer appears to preclude analogous reactions in eukaryotic cells as regulatory functions usually require reversibility of the modulating events. Indeed, cellular ADP-ribosylation not only differs from toxin-mediated modifications with respect to the acceptors or acceptor sites; it also seems to be reversible as indicated by the existence of degrading enzymes (20). There appears to be one exception so far: Iglewski et al. described a liver ADP-ribosyl transferase that modifies elongation factor II at the same unique diphthamide residue as does diphtheria toxin (21). However, since a single molecule of diphtheria toxin is sufficient to kill a cell by irreversible inhibition of protein synthesis, the biological significance of the analogous cellular enzyme remains unclear.

612

H. Hilz

There are two subclasses of eukaryotic transferases which involve arginine residues in the acceptor molecules, one modifying a wide variety of proteins as well as low molecular weight guanidino compounds (22-25). Its biological role is not known. The second type is represented by a transferase isolated from chicken liver nuclei. This enzyme recognizes a specific arginine residue in histone HI and in certain cytoplasmic proteins like glycogen Phosphorylase kinase that is part of a specific sequence Arg—Arg—Ala —Ser—Leu also serving cAMP-dependent protein kinase A as an acceptor site. ADP-ribosylation of an arginine prevents subsequent phosphorylation of serine by protein kinase A (26). The same ADP-ribosyl transferase from chicken nuclei is also capable of modifying human c—H —ras protein (27). Based on the specific cleavage by mercuri ions of thioglycosides as present in ADPribosyl G¡ protein (28), significant amounts of S-linked ADPR protein conjugates was recently found in rat liver (29). It may well be that the cysteine ADP-ribosyl transferase described by Tanuma et al. (30) is responsible for the formation of these conjugates. In mitochondria apparent modification of 55kDa and 31 kDa polypeptides with labeled NAD (31-38) may not represent a mono-(ADP-ribosyl)ation reaction: In mitochondrial homogenates and in submitochondrial particles NAD is rapidly converted to free ADPR which proved as good a 'substrate' for the formation of acidinsoluble conjugates as ( 3 H)NAD. Since formation of a glycosidic linkage requires activated ADPR, not free ADPR, and since heat treatment of the mitochondrial preparations does not prevent subsequent incorporation of label from free ADPR, most of this mitochondrial reaction appears to be non-enzymic (34). So far, a regulatory function of mono(ADP-ribosyl)ation under in vivo conditions has only been shown for the nitrogenase reaction in Rhodospirillum rubrum (35): An ADP-ribosyl transferase was purified from the bacterium and shown to inactivate dinitrogenase reductase. Furthermore, an ADP-ribosyl reductase glycohydrolase was isolated thus providing a reversible system. It has also been demonstrated that the ADP-ribosylation cascade is reversible and dynamic in vivo. Effectors that initiate the modification and thereby slow down or block nitrogen fixation include ammonium ions, glutamine, darkness and uncouplers (35).

Poly(ADP-ribosyl)ation of nuclear protein The nuclear system was the first ADP-ribosylation reaction to be detected (6-8). The structure of the polymer (Figure 2) indicates O-glycosidic linkages which are present in the bond between the polymer and the acceptor proteins (ester glycoside) as well as in the ribose (1" -» 2') ribose phosphate - phosphate backbone of the polymer and in the branch points (ribose (1"' -> 2") ribose (1" ->· 2') ribose; (56,57). A single enzyme of 116 kDa, poly(ADPR) polymerase, is responsible for the entire structure of the modifying group (58): This unusual catalyst transfers the first ADPR residue to the acceptor protein, it elongates the chain, and it introduces branches. Its activity depends on nicked dsDNA (59), which is bound to a defined domain of the enzyme (60). Activation by DNA fragmentation occurs also in

Pyridine Nucleotides as Group Transfering Coenzymes

613

nh2

u> Protein HO

OH

mono ( A D P - r i b o s y l ) protein

p-N

p-

Protein

poly ( A D P - ribosyl ) protein Figre 2

ADP-ribosyl proteins.

permeabilized cells (61) as well as in vivo (62). Inactivation proceeds, at least in vitro, by auto poly(ADP-ribosyl)ation at another specific domain, presumably by interfering with DNA binding (63). The 'concentration' of poly(ADPR) polymerase in HeLa cells is about 240,000 molecules per nucleus (64). Other mammalian cells have very similar amounts, independent of their origin, their growth characteristics and their state of transformation. The basic structural domains of the enzyme appear to be highly conserved as the existence of a 116 kDa polypeptide cross-reacting with anti pig polymerase antibodies could be detected even in slime molds like Physarum polycephalum (64). A great number of poly(ADPR) polymerase inhibitors have been described ranging from nicotinamide (65) and benzamide (66) to thymidine (65) and theophylline (58). The most frequently used compounds are nicotinamide analogs like 3-aminobenzamide, benzamide and methoxybenzamide (66). None of the inhibitors available,

614

H. Hilz

however, is truely specific for the polymerase (16). Although these compounds interfer with the polymerase at rather low concentrations, side effects on various reactions (nicotinamide N-methyltransferase, PRPP formation, cAMP phosphodiesterase, purine nucleotide synthesis (16), may lead to misinterpretations especially when the inhibitors are applied in vivo. Acceptor polypeptides: A large number of polypeptides has been found to be modified in vitro. They encompass nearly all chromosomal proteins known to interact with nucleic acids like histones and HMG proteins, or to use nucleic acids as substrates (DNA polymerases, endonucleases, topoisomerases). An (incomplete) list of such ,in vitro substrates' of poly(ADPR) polymerase is given in Table 2. The plentitudo of acceptors found in experiments with isolated nuclei or purified poly(ADPR) polymerase sharply contrasts with the paucity of ADP-ribosylated polypeptides isolated from intact tissues. So far, only histone H2B 1 (67,68), traces of histone HI (69) and auto modified poly(ADPR) polymerase (70) have been definitely identified as acceptors after treatment of cells with an alkylating agent. However, it should be mentioned that the concentration of modified acceptors even in maximally stimulated cells remains low (10-70 pmol/mg DNA, assuming an average chain length of five ADPR units and two chains per polypeptide (11)). It

Table 2

ADP-ribosylation of nuclear proteins

Acceptor protein A. In vitro Histones A24 protein H M G proteins D N A polymerase α D N A pol α-primase D N A polymerase β terminal deoxy nucleotidyl transferase D N A ligase II Ca 2 + , Mg 2 + -dep. endonuclease topoisomerase I and II R N A polymerase RNases hnRNA associated protein nuclear lamina proteins poly (ADPR) polymerase B. In vivo (in response to alkylation) Histones: H2B H4 HI topoisomerase I H M G proteins poly(ADPR)polymerase

1

Modifying A D P R group

References

polymer polymer polymer polymer polymer polymer polymer polymer polymer polymer polymer polymer polymer polymer

78-81 82 83,84 85 85 85 86 85 85 86, 86 a 87 88 89 89 a 90

monomer and polymer monomer monomer polymer oligomer polymer

67, 68 69,75 69,91 70 71 70

Histone H2B is also the major acceptor of mono(ADP-ribosyl) groups in alkylated 3T3 cells (75).

Pyridine Nucleotides as Group Transfering Coenzymes

615

appears that the extremely high turnover of poly(ADPR) residues under these conditions (r 1 / 2 900 s _ 1

Stahl & Jencks e

150s" 1

70s" 1 > 1000 s" 1

" Experiments with rabbit skeletal muscle FSR were carried out at room temperature (mostly at 25 °C), whereas those with frog skeletal muscle FSR were done at 15°C. b Stahl and Jencks (personal communication) assumed that a conformational change following ATP binding was rate-limiting. Other investigators, however, thought the conformational change was too fast to be rate-limiting. c Ogawa and Harafuji (40) assumed that a conformational change subsequent to ATP binding was rate-limiting, whereas other investigators assumed that ATP binding itself was rate-limiting. d Ogawa and Harafuji (40) assumed that the conformational change by Ca-binding was too fast to be rate-limiting, while others assumed that it was rate-limiting. e personal communication. Compare also Ref. 41.

756

Yasuo Ogawa

reflected in a larger ACp of the Ca transport reaction by rabbit FSR than by frog FSR (38), which is revealed to be an entropy-driven reaction (39). We examined the transient kinetics of EP formation of frog FSR in a rapid-quench apparatus which had 4 syringes to carry out experiments where Ca 2 + or ATP was added in varied sequence to Ca 2 + free FSR after a specified preincubation period with a ligand (40). After a short preincubation such as 10-20 ms with Ca 2 + , ATP started the reaction of EP formation in an exponential time course after a lag time of 2 3 ms; this lag time was independent of ATP concentration. With prolongation of the preincubation period with Ca 2 + , both the level of EP maximally attained and the apparent rate constants for its formation increased under the same final concentrations of Ca 2 + and ATP, the rate constant per unit level of EP remaining constant. In these cases, no lag time was observed (Figure 8). The increase in EP level and apparent rate constant occurred in an exponential time course with the preincubation period. These results indicate that E may undergo a slow conformational change to Ca • *E after Ca binding which is probably rapid, and that Ca · *E can be reactive with ATP to transform into EP · Ca in second order kinetics. The rate-limiting conformational change, E to Ca • *E, may proceed in first-order kinetics, the rate constant being estimated as 11 s _ 1 at 15°C. On the other hand, when the reaction was started by addition of Ca 2 + after a short preincubation of 10 20 ms with ATP, EP formation began instantaneously without any delay and followed an exponential time course. The rate constant was larger than the counterpart of the reversed addition sequence. When the preincubation period with ATP was prolonged to about 84 ms, a burst in EP formation was observed as Stahl and Jencks (41) reported with rabbit FSR (Fig. 8). The time course of EP formation can be described as two components with large and small rate constants (Figure 8). The amplitude of the component with a large rate constant increased with the increase in ATP concentration, which was consistent with AMPOPCP binding to FSR under a Ca 2 + free condition. These observations, with other lines of evidence already published (42,43), indicate that E undergoes another conformational change to E§ · ATP after ATP binding and that this E§ · ATP can react with Ca 2 + to form EP. ATP binding may be too fast to be rate-limiting, but the subsequent conformational change, E to E§ · ATP, is rate-limiting, although it is much faster than the conformational change, E to Ca · *E. Similar experiments with varied concentrations of ATP gave estimates of rate constants for reaction steps to form EP as shown in Table 2. This finding indicates that an ATP activation route exists in addition to the well-known Ca 2+ -activation route as in the E r E 2 model, and that the former route is much faster. A similar conclusion was obtained with rabbit FSR by Stahl and Jencks (41). Interestingly, recent detailed analysis of their own results (personal communication) strongly suggests that there may be negative cooperative interaction between Ca 2 + and ATP as shown in Figures 4B and 5B. The myoplasmic Ca 2 + is estimated to be around 0.1 μΜ at rest and to reach transiently 3 - 5 μΜ during twitch. Then, E would be at the E 2 state at rest according to the Ei-E-2 model. Frog skeletal muscle recovers from twitch in about 100 ms at 15°C. Since there is abundant ATP in the myoplasm, the ATP-activation route is more important for muscle relaxation (45).

Comparative Aspects of the SR

757

The probable occurence of the two alternative routes for the activation and mutual interaction of ligand binding would explain some previously unresolved points. First, kinetically at least two sites for ATP were assumed as mentioned above: catalytic and regulatory sites. However, the primary amino acid sequence for C a 2 + ATPase protein of rabbit FSR deduced from its c D N A revealed only one ATP site (46,47). The downward deflection in the Lineweaver-Burk plot would easily be explained now by the mutual interaction between C a 2 + and ATP and the two alternative activation routes with different rate constants. Second, transient overshoot in the EP-level can similarly be explained by the two alternative pathways of the reaction with the knowledge of where the first reaction turn began.

Concluding Remarks Discussion here is limited to some interesting points on the mechanism of Ca uptake and the accompanying Ca 2 + -ATPase reaction, although other minor differences remain to be discussed. Different characteristics are also observed about Ca release, another important function of SR. Briefly, rabbit FSR is more permeable to Ca 2 + at rest than bullfrog FSR. This is also observed in in situ SR of skinned fiber. Therefore, passive loading with bullfrog FSR is usually unsuccessful. Frog FSR, in turn, retains some amount of Ca which is releasable by caffeine and A23187 in contrast to rabbit FSR which contains no or very little Ca a couple of days after the preparation (48). Sensitivity to some Ca-releasing drugs is also different (7,49-51). Halothane causes more Ca release at higher temperature from rabbit FSR, while it causes greater release at lower temperature from bullfrog FSR (51). Many investigators consider that the species difference might be trivial, if it exists at all. As shown here, however, some of these differences may be critical. Comparison may cast light on some aspects which would not be recognized in other approaches and contribute toward elucidating the mechanism functioning in the sarcoplasmic reticulum.

Acknowledgements Without the days spent in Dr. Fritz Lipmann's laboratory from September, 1972 to February, 1975, the author would likely not have even considered a scientific career. His thanks are also due to Miss Kazuyo Nakajima for her expert secretarial assistance.

References 1. Ebashi, S. and Y. Ogawa. 1988. In: Handbook of Experimental Pharmacology, Vol. 83: Calcium in Drug Actions (P. F. Baker, ed.). Springer, Heidelberg, pp. 31-56. 2. Ebashi, S..1961. J. Biochem. 50, 236-244. 3. Ebashi, S. and F. Lipmann. 1962. J. Cell Biol. 14, 389-400. 4. Weber, A. and S. Winicur. 1961. J. Biol. Chem. 236, 3198-3202.

758

Yasuo Ogawa

5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

Ebashi, S., A. Kodama and F. Ebashi. 1968. J. Biochem. 64, 465-477. Hasselbach, W. and M. Makinose. 1961. Biochem. Z. 333, 518-528. Weber, Α., R. Herz, and I. Reiss. 1966. Biochem. Z. 345, 329-369. Weber, A. and R. Herz, 1968. J. Gen. Physiol. 52, 750-759. Weber, Α.. 1968. J. Gen. Physiol. 52, 760-772. Tada, M., T. Yamamoto and Y. Tonomura. 1978. Physiol. Rev. 58, 1-79. de Meis, L.. 1981. The Sarcoplasmic Reticulum. Wiley, New York. Ikemoto, N.. 1982. Ann. Rev. Physiol. 44, 297-317. Hasselbach, W. and H. Oetliker. 1983. Ann. Rev. Physiol. 45, 325-339. Martonosi, A. and T.J. Beeler. 1983. In: Handbook of Physiology, Sect. 10: Skeletal Muscle, (L.D. Peachey, R. Adrian and S.R. Geiger, eds.). Am. Physiol. Soc., Bethesda, Md. pp. 417-485. Inesi, G.. 1985. Ann. Rev. Physiol. 47, 573-601. Fleischer, S. and Y. Tonomura (eds.). 1985. Structure and Function of Sarcoplasmic Reticulum. Academic Press, Orlando Kurebayashi, N. and Y. Ogawa. 1982. J. Biochem. 92, 907-913. Kurebayashi, Ν., Y. Ogawa and H. Harafuji. 1982. J. Biochem. 92, 915-920. Ogawa, Y.. 1972. J. Biochem. 71, 571-573. Ogawa, Y.. 1972. In: Role of Membranes in Secretory Processes (L. Bolis, R.D. Keynes and W. Wilbrandt, eds.). North-Holland, Amsterdam, pp. 170-174. Ogawa, Y. and S. Ebashi. 1973. In: Organization of Energy-Transducing Membranes (M. Nakao and L. Packer, eds.). University of Tokyo Press, Tokyo, pp. 127-140. Inesi, G.. 1971. Science 171, 901-903. Pucell, A. and A. Martonosi. 1971. J. Biol. Chem. 246, 3389-3397. de Meis, L. and W. Hasselbach. 1971. J. Biol. Chem. 246, 4759-4763. de Meis, L. and M.C. Fialho de Mello. 1973. J. Biol. Chem. 248, 3691-3703. Segel, I.Η.. 1975. Enzyme Kinetics. Wiley, New York. pp. 847-883. Kurebayashi, Ν., T. Kodama and Y. Ogawa. 1980. J. Biochem. 88, 871-876. Ogawa, Y. and N. Kurebayashi. 1982. J. Muscle Res. Cell Motil. 3, 39-56. Hasselbach, W. and M. Makinose. 1962. Biochem. Biophys. Res. Commun. 7, 132-136. Makinose, M. and W. Boll. 1979. In: Cation Flux Across Biomembranes (Y. Mukohata and L. Packer, eds.). Academic Press, New York pp. 89-100. Yamada, S. and N. Ikemoto. 1980. J. Biol. Chem. 255, 3108-3119. Makinose, M.. 1969. Eur. J. Biochem. 10, 74-82. Ogawa, Y., N. Kurebayashi and H. Harafuji. 1987. J. Biochem. 100, 1305-1318. Pang, D.C. and F.N. Briggs. 1977. J. Biol. Chem. 252, 3252-3266. Cable, M.B., J.J. Feher and F.N. Briggs. 1985. Biochemistry 24, 5612-5619. Meissner, G.. 1973. Biochim. Biophys. Acta 298, 906-926. Ebashi, F. and I. Yamanouchi. 1964. J. Biochem. 55, 504-509. Kodama, T., Ν. Kurebayashi and Y. Ogawa. 1980. J. Biochem. 88, 1259-1265. Kodama, T., Ν. Kurebayashi, Η. Harafuji and Y. Ogawa. 1982. J. Biol. Chem. 257, 4238-4241. Ogawa, Y. and H. Harafuji. 1987. J. Biochem. 100, 1319-1328. Stahl, Ν. and W.P. Jencks. 1984. Biochemistry 23, 5389-5392. Fernandez-Belda, F., M. Kurzmack and G. Inesi. 1984. J. Biol. Chem. 259, 9687-9698. Dupont, Υ., Ν. Bennett and J.-J. Lacapere. 1982. Ann. Ν. Y. Acad. Sci. 402, 569-572. Froehlich, J.P. and E.W. Taylor. 1975. J. Biol. Chem. 250, 2013-2021. Ogawa, Y., N. Kurebayashi, A. Irimajiri and T. Hanai. 1981. In: Adv. Physiol. Sci. Vol.5. Molecular and Cellular Aspects of Muscle Function (E. Varga, A. Kôvér, T. Kovács, L. Kovács, eds.). Pergamon Press, London/Akadémiai Kiadó, Budapest, pp. 417-435. MacLennan, D.H., C.J. Brandi, B. Korczak and N.M. Green. 1985. Nature 316, 696-700. Brandl, C.J., N.M. Green, Β. Korczak and D.H. MacLennan. 1986. Cell 44, 597-607. Ogawa, Y., H. Harafuji and N. Kurebayashi. 1985. In: Structure and Function of Sarcoplasmic Reticulum (S. Fleischer and Y. Tonomura, eds.). Academic Press, Orlando, pp. 411-427. Ogawa, Y.. 1970. J. Biochem. 67, 667-683. Ogawa, Y. and S. Ebashi. 1976. J. Biochem. 80, 1149-1157. Ogawa, Y. and N. Kurebayashi. 1982. J. Biochem. 92, 899-905.

15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51.

Caltrin: A Versatile Regulator of Calcium Transport in Spermatozoa Henry Lardy, Jovenal San Agustín and Carlos

Coronel

In this male-dominated audience it is well to reflect that not only are we males not capable of reproducing our own species, as is the female; we are not even able to contribute spermatozoa competent for fertilizing the eggs produced by the female. Mammalian sperm are not capable of fertilizing eggs at the time of mating. They must spend some time in the female reproductive tract before they acquire that ability. In the late 1940s and early 1950s many investigators attempted in vitro fertilization of mammalian eggs but the success rate was negligible. In 1951, Austin (1) working in Australia and Chang (2) at the Worcester Foundation in Massachusetts found that sperm recovered from the reproductive tract of the rabbit some hours after mating, were fully capable of fertilizing rabbit eggs in vitro. The physiological changes that the sperm undergo in the female reproductive tract to become competent are called "capacitation". When mammalian eggs are ovulated they are covered with a gelatinous protein layer termed the zona pellucida and surrounded by layers of cells - the cumulus oophorus. In natural mating the sperm must bore a hole through these layers before it can reach the vitellin membrane which it penetrates readily. Eggs from which these layers have been removed by mild chemical treatments can be fertilized in vitro by freshly collected spermatozoa. A period of capaci tation in the female is not required. Part of the capacitation process is the "acrosome reaction" (3). The acrosome is an organelle at the head of each sperm cell that is enclosed by its own membrane and under the sperm plasma membrane. It contains hyaluronidase, a proteolytic enzyme, acrosin, and other hydrolytic enzymes that aid sperm penetration of the egg by digesting a pathway through the cumulus oophorus and the zona pellucida. During capacitation the outer membrane of the acrosome undergoes fusion with the plasma membrane that covers it, vesiculation and discomposition of the membranes occur and the stored hydrolytic enzymes are released. It is important that the acrosome reaction not occur at the time of mating for, if it did, the hydrolytic enzymes would be left behind as the spermatozoa swim up the female reproductive tract. Ideally, the acrosome reaction should occur when the sperm encounter the egg(s) in the oviduct and that is how Nature has timed the event. The time required for capacitation in various species is proportional to the distance the sperm must travel before meeting the female gamete. As you will see, caltrin, the subject of this paper honoring the memory of Fritz Lipmann, plays a key role in timing the capacitation process. Following the initial discoveries of Chang and Austin, biologists began studing the process of capacitation and in the laboratories of Dan (4) and of Yanagimachi (5) it

760

Henry Lardy

was found that in vitro capacitation could be achieved if sperm were incubated for some hours in a medium containing buffer salts, lactate and/or pyruvate and calcium. The function of calcium was not understood until it was demonstrated that in the presence of medium calcium the calcium ionophore A23187 (6) could induce the acrosome reaction without the prolonged incubation (7-9). Using 45 CaCl 2 we demonstrated that the acrosome reaction occurs within minutes of calcium uptake induced by the ionophore (9). Detailed studies in several laboratories indicate that calcium uptake by the spermatozoa results in activation of phospholipase A 2 in the acrosome. The resulting hydrolyses of phospholipids in the acrosomal membrane produces lysophospholipids which facilitate fusion of the acrosomal and plasma membranes and eventual discomposition of both. The hydrolytic enzymes are thus released. Babcock et al. (10,11) had earlier investigated the transport of calcium and phosphate into spermatozoa and observed striking differences between epididymal and ejaculated cells. Detailed studies revealed that bovine epididymal spermatozoa contain only 6.4 + 0.8 nmol of mobilizable (with A23187) calcium per 108 cells and their total calcium is 9.9 + 0 . 6 ng atoms/10 8 cells. When incubated in medium containing 0.2 to 0.5 mM Ca 2 + these cells take up from 20 to 30 nmol of Ca (12). Bovine seminal plasma contains ca 9 mM calcium and we were therefore astonished to find that ejaculated sperm contain no more mobilizable calcium (6.7 + 0.9 nmol/10 8 cells) than do epididymal cells (12). Calcium could be brought into ejaculated cells with the aid of A23187 so it appeared that the permeability of sperm to calcium was altered on ejaculation. Sperm-free seminal fluid was then found to inhibit uptake of Ca 2 + into epididymal cells (12). The inhibitory factor was isolated from seminal fluid and found to be a small protein with an isoelectric point in the range of pH 9 (13). Because it is a ca/cium transport /whibitor we named it caltrin. Sequencing the protein (14) revealed it to contain 47 amino acid residues and a molecular weight of 5411; it also revealed that the sequence of the first 25 amino acids was identical with that reported (15) for a protein isolated from seminal plasma by Reddy and Bhargava (16) and shown to possess antibacterial activity. The two proteins are identical and the erroneous structure resulted from a transposition of two peptides in deducing the sequence; there was also one lysine de trop. The structure found by Lewis et al. (14) has been confirmed in Bhargava's laboratory (17). Caltrin does not bind calcium (13) but does bind tenaciously to spermatozoa for repeated washing of ejaculated sperm does not relieve the inhibition of calcium transport. Plasma membrane vesicles prepared from ejaculated cells take up much less calcium than similar vesicles from epididymal cells (18). Using rabbit polyclonal antibodies to bovine caltrin and goat anti-rabbit IgG labeled with fluorescein isothiocyanate, caltrin binding to spermatozoa was found to be specifically located over the acrosome, the tail, and at a spot on the neck that joins the head and midpiece (19). No fluorescence occurred from the lower half of the sperm head nor from the midpiece where the mitochondria, end-to-end in a helix, surround the axoneme. It is easy to see how caltrin can prevent premature development of the acrosome reaction by blocking calcium channels where it binds

Caltrin

761

over the acrosomal membranes. There is also a logical explanation for the function of caltrin on the tail. When ejaculated, sperm swim with moderate speed on a linear course. This facilitates their movement up the female tract to the oviduct where they encounter the egg(s). Calcium causes a hyperactivation of motility characterized by more rapid beat and wider excursion of the tail. Sperm so activated swim in tight circles and it has been postulated that this hyperactivated movement promotes their penetration of the debris that surrounds the egg. Ejaculated bovine sperm do not exhibit hyperactivity until capacitated, presumably because caltrin blocks calcium penetration. If caltrin prevents premature occurrence of both the acrosome reaction and hyperactivation it is reasonable to ask what eventually permits calcium to enter and to stimulate these events. The answer is that caltrin itself becomes a promoter of calcium uptake (19). When caltrin is separated from the constituents of seminal fluid and stored for some weeks as a homogeneous protein in 35% glycerol at — 20 °C it loses its ability to

time (min) Figure 1 Calcium uptake in bovine epididymal spermatozoa. Epididymal spermatozoa (8 χ IO8 cells/ml) previously incubated for 15 min in N K M plus 10 mM DL-/?-hydroxybutyrate were transferred to the assay medium without CaCl 2 and containing the indicated additions: ( · ) none; ( T ) 30 M A23187; ( A ) 5 M C1CCP; ( • ) 0.400 mg/ml caltrin (old preparation): (X) 0.400 mg/ml fresh caltrin; ( • ) 0.400 mg/ml fresh caltrin passed through CM Sephadex G-25 cation exchanger. The final cell count was 4 χ 107 cells/ml and the final volume was 1 ml. Ten minutes after addition of cells to the assay medium, CaCl 2 containing 4 5 Ca was added to a final concentration of 0.2 mM (t = 0). At indicated times during the assay, 0.200-ml aliquots were taken out for determination of calcium uptake. A representative result of several experiments is shown. From Ref. 19.

762

Henry Lardy

inhibit and becomes an enhancer of calcium transport. This change in physiological function coincides with a change in conformation of the protein as is indicated by enhancement of tetracycline fluorescence. Caltrin prepared by the method of Reddy and Bhargava (16) for the preparation of bovine seminal plasmin was found not to inhibit calcium transport but to stimulate in the same manner as our samples of caltrin that had been stored for a month or more. Careful study disclosed that the change in behavior of the protein occurred when this basic protein was adsorbed on and eluted from CM-Sephadex cation exchanger (19). This prompted an examination of the anionic constituents of seminal plasma to determine whether they might influence the properties of caltrin. Fractionation revealed the presence of two anions that restored inhibitory behavior to caltrin that had been treated with CM-Sephadex. One of these effective anions was characterized as citrate, an abundant component of seminal fluid that is contributed by the prostate gland and the seminal vesicles (20). Incubating old (enhancer form) caltrin with citrate at low pH followed by gel filtration to remove excess citrate restores the ability of the protein to inhibit calcium uptake into epididymal bull sperm. A second anion fraction that restores inhibiting activity to deionized caltrin is now under investigation. Figure 1 depicts some of these relevant date. Our tentative interpretation of these findings is that, at the time of ejaculation, caltrin binds to spermatozoa so as to prevent calcium uptake and, consequently, a premature acrosome reaction. As the sperm move up the female tract, away from the anionic components of the seminal fluid, the regulatory anions slowly diffuse away. The resulting conformational change in the caltrin molecule converts it to an enhancer of calcium uptake. The acrosome reaction is thus induced at about the time sperm reach the egg(s), the hydrolytic enzymes facilitate penetration of the layers surrounding the egg(s), and the union of male and female gametes then occurs. Much more work will be required to establish or disprove this hypothetical sequence. Calcium transport is not regulated in the same manner in all species. In contrast to the epididymal spermatozoa of the bull, those of the guinea pig, rabbit, rat and dog do not take up external calcium. Guinea pig epididymal sperm do take up calcium after they have been held at 37 °C for 20 to 60 min in a medium at pH 7.8 containing NaCl and N a H C 0 3 to isotonicity plus pyruvate or lactate as energy sources (9,21). After the preincubation period of guinea pig epididymal sperm, calcium uptake induces the acrosome reaction (7-9). The uptake of calcium is depressed by fluid from the reproductive tract of the male guinea pig and this fluid contains a protein that cross reacts with rabbit anti-bovine caltrin globulin (22). The caltrin-like protein has a molecular weight of 6200 and is produced in the seminal vesicles as is bovine caltrin. Whether the guinea pig caltrin becomes a stimulator of calcium transport in the manner of the bovine protein is being investigated. Finally, we may ask whether caltrin plays a role in regulating calcium transport in somatic cells. We have used the rabbit antibodies to bovine caltrin to search for the presence of this protein in several bovine organs. The Ouchterlony procedure detects caltrin in bull seminal vesicles but in no other tissue tested.

Caltrin

763

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.

Austin, C.R. 1951. Aust. J. Sei. Res. B. 4, 581. Chang, M.C. 1951. Nature 168, 697. Chang, M.C. 1984. J. Androl. 5, 45. Dan, J.C. 1954. Biol. Bull. 107, 335. Yanagimachi, R. and N. Usui. 1974. Exp. Cell Res. 89, 161. Reed, P.W. and H.A. Lardy. 1972. J. Biol. Chem. 247, 6970. Summers, R. G., P. Talbot, E. M. Keough, B. L. Hylander and L. E. Franklin. 1976. J. Exp. Zool. 196, 381. Green, D.P.L. 1978. J. Cell Sci. 32, 137. Singh, J.P., D.F. Babcock and H.A. Lardy. 1978. Biochem. J. 172, 549. Babcock, D.F., N.L. First and H.A. Lardy. 1975. J. Biol. Chem. 250, 6488. Babcock, D.F., N.L. First and H.A. Lardy. 1976. J. Biol. Chem. 251, 3881. Babcock, D.F., J.P. Singh and H.A. Lardy. 1979. Dev. Biol. 69, 85. Rufo, G.Α., J.P. Singh, D.F. Babcock and H.A. Lardy. 1982. J. Biol. Chem. 257, 4627. Lewis, R.V., J. San Agustín, W. Kruggel and H.A. Lardy. 1985. Proc. Natl. Acad. Sci. USA 82, 6490. Theil, R. and K.H. Scheit. 1983. EMBO J. 2, 1159. Reddy, E.S.P. and P.M. Bhargava. 1979. Nature (London) 279, 725. Sitaram, N., V.K. Kumari and P.M. Bhargava. 1986. FEBS Lett. 201, 233. Rufo, G.Α., Jr., P.K. Schoffand H.A. Lardy. 1984. J. Biol. Chem. 259, 2547. San Agustín, J.T., P. Hughes and H.A. Lardy. 1987. FASEB J. 1, 60. Mann, T. 1964. In: The Biochemistry of Semen and of the Male Reproductive Tract. Wiley, New York. Coronel, C.E., H.A. Lardy. 1987. Biol. Reprod. (In press). Coronel, C.E., J. San Agustín, H.A. Lardy. 1987. Submitted to Biol. Reprod.

Protein Kinase C, the Structural Heterogeneity and Differential Expression in Rat Brain K. Ogita, U. Kikkawa, K. Ase, M. S. Shearman, Y. Nishizuka, Y. Ono, T. Fujii, T. Kurokawa, Κ. Igarashi, Ν. Saito and C. Tanaka

Introduction Receptor-mediated hydrolysis of inositol phospholipids is now accepted to be a common mechanism for transducing various extracellular signals into the cell (1,2). At an early phase of cellular responses, inositol-l,4,5-trisphosphate (IP 3 ) mobilizes Ca 2 + , whereas diacylglycerol activates protein kinase C. These two intracellular mediators are generated from the hydrolysis of a single membrane component, phosphatidylinositol-4,5-bisphosphate (PIP 2 )· Although once considered as a single entity, enzymological and molecular cloning analysis has revealed that protein kinase C exists as a family of multiple subspecies with subtle individual characteristics. Biochemical and immunocytochemical studies indicate their differential regional expression and distinct cellular localization. Presumably, each subspecies relays information from different signals across the membrane. This article will briefly summarize some aspects of the structure and expression of the protein kinase C family.

Structural Heterogeneity The cDNA clones α, ßl, ßll, and γ were isolated from bovine, rat, rabbit and human brain libraries in several laboratories (3-7). This integrated nomenclature and its correspondence to those of other workers (3-6) are as given in our preceding papers (8,9). The complete amino acid sequences of four subspecies of protein kinase C are shown in Figure 1 (see also Ref. 10). The sequences of each subspecies in different animal species are remarkably similar, and the structures of the four subspecies are highly homologous and closely related to one another. Their common structure is composed of a single polypeptide chain with four conserved (C t — C 4 ) and five variable (V¡ — V 5 ) regions as schematically given in Figure 2. ßl- and ßll-subspecies are derived from a single RNA transcript by alternative splicing, and differ from each other only in a short range of about 50 amino acid residues in their carboxyl-terminal end regions (5-8). The first conserved region C, contains a tandem repeat of cysteine-rich sequence, CX2CX13CX2CX7CX7C, where C is cysteine and X represents any amino acid. This sequence agrees with the consensus of "Zn 2 + -finger", which is found in many DNA-binding proteins that are related to transcriptional regulation (11). This implies a potential role of protein kinase C to control some nuclear functions, but no indication for this role is available at resent. The C¡ and C 2 regions compose the

766

Κ. Ogita et al.

M M '

« w u> < o o o σ b< h h k >J J U tJ u u u u a ο αw u u u u ο α α α ΗM u u y o u o δ ww o b b h b ft ft ft Pi ft < < ft α σ a a oo ou < Λ -

m σι V so m m m η γο r> >-

CO rt Hιη r» oo oo 0) m m mm so Oí Γ> Ν «0 Ν00σι * » ν τ Μ Μ Μ Ο CQ CÛ> -

M W O >C •J > M >U ft Ζ ft XM X E è κ σ υw " ft w e 5 Xh www ÍH PH Ix υ υ u W < M (h h h ~ W W X w σ σ o> o o o •J M 2 ξ S t-t M f·* > •J M σ « σ ο α o ft e ft ft ft ft b> t* t* "W tJ OI > > > ft ft ft I

Ν Γ» rtΓ» ηf»ίο so so so so σ»