Snake and Spider Toxins: Methods and Protocols [1st ed. 2020] 978-1-4939-9844-9, 978-1-4939-9845-6

Table of contents :
Front Matter ....Pages i-x
Front Matter ....Pages 1-1
Snake- and Spider-Venom-Derived Toxins as Lead Compounds for Drug Development (Philip Lazarovici)....Pages 3-26
Analytics for Bioactivity Profiling of Complex Mixtures with a Focus on Venoms (Marija Mladic, Wilfried M. A. Niessen, Govert W. Somsen, Jeroen Kool)....Pages 27-49
Front Matter ....Pages 51-51
Venom Collection from Spiders and Snakes: Voluntary and Involuntary Extractions (“Milking”) and Venom Gland Extractions (William K. Hayes, Gerad A. Fox, David R. Nelsen)....Pages 53-71
Production and Purification of Recombinant Toxins (Matan Geron)....Pages 73-84
Front Matter ....Pages 85-85
RNA-Sequencing of Snake Venom Glands (Khin Than Yee, Olga Vasieva, Ponlapat Rojnuckarin)....Pages 87-96
Exploring Toxin Evolution: Venom Protein Transcript Sequencing and Transcriptome-Guided High-Throughput Proteomics (Cassandra M. Modahl, Jordi Durban, Stephen P. Mackessy)....Pages 97-127
Three-Dimensional Structure Determination of Peptides Using Solution Nuclear Magnetic Resonance Spectroscopy (Christina I. Schroeder, K. Johan Rosengren)....Pages 129-162
Determining the Structures of the Snake and Spider Toxins by X-Rays (Anwar Ullah, Rehana Masood, Zafar Hayat, Ahmed Hafeez)....Pages 163-172
MALDI-TOF Mass Spectrometric Profiling of Spider Venoms (Ondrej Šedo, Stano Pekár, Zbyněk Zdráhal)....Pages 173-181
Front Matter ....Pages 183-183
Methods for Evaluation of a Snake Venom-Derived Disintegrin in Animal Models of Human Cancer (Stephen D. Swenson, Catalina Silva-Hirschberg, Francis S. Markland)....Pages 185-204
Cell-Based Adhesion Assays for Isolation of Snake Venom’s Integrin Antagonists (Philip Lazarovici, Cezary Marcinkiewicz, Peter I. Lelkes)....Pages 205-223
Using C. elegans to Study the Effects of Toxins in Sensory Ion Channels In Vivo (Valeria Vásquez)....Pages 225-238
Measurements of Cell Death Induced by Snake and Spider’s Venoms and Derived Toxins (Yossi Maatuf, Avi Priel, Philip Lazarovici)....Pages 239-268
Using Toxins in Brain Slice Recordings (Alexey Bingor, Rami Yaka)....Pages 269-274
High-Throughput Calcium Imaging Screen of Toxins’ Function in Dissociated Sensory Neurons (Yossi Maatuf, Avi Priel)....Pages 275-282
Synthesizing and Expressing Native Ion Channels (Shana L. Geffeney, Charles T. Hanifin)....Pages 283-290
Back Matter ....Pages 291-293

Citation preview

Methods in Molecular Biology 2068

Avi Priel Editor

Snake and Spider Toxins Methods and Protocols

METHODS

IN

MOLECULAR BIOLOGY

Series Editor John M. Walker School of Life and Medical Sciences University of Hertfordshire Hatfield, Hertfordshire, UK

For further volumes: http://www.springer.com/series/7651

For over 35 years, biological scientists have come to rely on the research protocols and methodologies in the critically acclaimed Methods in Molecular Biology series. The series was the first to introduce the step-by-step protocols approach that has become the standard in all biomedical protocol publishing. Each protocol is provided in readily-reproducible step-bystep fashion, opening with an introductory overview, a list of the materials and reagents needed to complete the experiment, and followed by a detailed procedure that is supported with a helpful notes section offering tips and tricks of the trade as well as troubleshooting advice. These hallmark features were introduced by series editor Dr. John Walker and constitute the key ingredient in each and every volume of the Methods in Molecular Biology series. Tested and trusted, comprehensive and reliable, all protocols from the series are indexed in PubMed.

Snake and Spider Toxins Methods and Protocols

Edited by

Avi Priel Faculty of Medicine, School of Pharmacy, Institute for Drug Research, The Hebrew University of Jerusalem, Jerusalem, Israel

Editor Avi Priel Faculty of Medicine School of Pharmacy Institute for Drug Research The Hebrew University of Jerusalem Jerusalem, Israel

ISSN 1064-3745 ISSN 1940-6029 (electronic) Methods in Molecular Biology ISBN 978-1-4939-9844-9 ISBN 978-1-4939-9845-6 (eBook) https://doi.org/10.1007/978-1-4939-9845-6 © Springer Science+Business Media, LLC, part of Springer Nature 2020 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Humana imprint is published by the registered company Springer Science+Business Media, LLC, part of Springer Nature. The registered company address is: 233 Spring Street, New York, NY 10013, U.S.A.

Preface In addition to being scientific discipline in toxicology, the research of animal toxins provides powerful tools for the deciphering of various biological mechanisms; both their toxic effects and extensive use as experimental tools consequent from their coevolution with their targets. Specifically, snake and spider toxins serve as excellent sources for our understanding of the development of peptide toxins, and different labs throughout the world routinely employ a wide array of these toxins for studying complex biological processes. As such, various methodologies have been developed for both handling and investigating peptide toxins and their use as molecular tools. The goal of this book is to share common techniques and protocols across different biological disciplines for the study and application of snake and spider peptide toxins. This book comprises of 16 chapters, including 2 overviews (Part I) describing the use of toxins in the process of drug development and the analysis of bioactivity of complex mixtures like venoms. The next 14 chapters provide detailed protocols describing the extraction of venom glands and the recombinant production of toxins (Part II), the characterization of toxins from the RNA level to peptide structure determination (Part III), and the determination of toxins’ biological function (Part IV). These protocols employ different cellular and animal models and various techniques involving toxins. Snake and Spider Toxins: Methods and Protocols is the outcome of the joint effort of 36 contributors from 10 countries. I wish to thank each of the authors for their contribution and willingness to share their valuable knowledge. Finally, I wish to thank Prof. John Walker for his guidance and help throughout the process of editing this book. Jerusalem, Israel

Avi Priel

v

Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

PART I

OVERVIEW

1 Snake- and Spider-Venom-Derived Toxins as Lead Compounds for Drug Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Philip Lazarovici 2 Analytics for Bioactivity Profiling of Complex Mixtures with a Focus on Venoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Marija Mladic, Wilfried M. A. Niessen, Govert W. Somsen, and Jeroen Kool

PART II

3

27

TOXINS PURIFICATION AND PRODUCTION

3 Venom Collection from Spiders and Snakes: Voluntary and Involuntary Extractions (“Milking”) and Venom Gland Extractions . . . . . . . William K. Hayes, Gerad A. Fox, and David R. Nelsen 4 Production and Purification of Recombinant Toxins . . . . . . . . . . . . . . . . . . . . . . . . Matan Geron

PART III

v ix

53 73

TOXINS CHARACTERIZATION

5 RNA-Sequencing of Snake Venom Glands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 Khin Than Yee, Olga Vasieva, and Ponlapat Rojnuckarin 6 Exploring Toxin Evolution: Venom Protein Transcript Sequencing and Transcriptome-Guided High-Throughput Proteomics . . . . . . . . . . . . . . . . . . . 97 Cassandra M. Modahl, Jordi Durban, and Stephen P. Mackessy 7 Three-Dimensional Structure Determination of Peptides Using Solution Nuclear Magnetic Resonance Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . 129 Christina I. Schroeder and K. Johan Rosengren 8 Determining the Structures of the Snake and Spider Toxins by X-Rays . . . . . . . . 163 Anwar Ullah, Rehana Masood, Zafar Hayat, and Ahmed Hafeez 9 MALDI-TOF Mass Spectrometric Profiling of Spider Venoms. . . . . . . . . . . . . . . . 173 Ondrej Sˇedo, Stano Peka´r, and Zbyneˇk Zdra´hal

PART IV 10

CHARACTERIZATION OF TOXINS FUNCTION

Methods for Evaluation of a Snake Venom-Derived Disintegrin in Animal Models of Human Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 Stephen D. Swenson, Catalina Silva-Hirschberg, and Francis S. Markland

vii

viii

Contents

11

Cell-Based Adhesion Assays for Isolation of Snake Venom’s Integrin Antagonists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Philip Lazarovici, Cezary Marcinkiewicz, and Peter I. Lelkes 12 Using C. elegans to Study the Effects of Toxins in Sensory Ion Channels In Vivo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Valeria Va´squez 13 Measurements of Cell Death Induced by Snake and Spider’s Venoms and Derived Toxins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yossi Maatuf, Avi Priel, and Philip Lazarovici 14 Using Toxins in Brain Slice Recordings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Alexey Bingor and Rami Yaka 15 High-Throughput Calcium Imaging Screen of Toxins’ Function in Dissociated Sensory Neurons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yossi Maatuf and Avi Priel 16 Synthesizing and Expressing Native Ion Channels . . . . . . . . . . . . . . . . . . . . . . . . . . Shana L. Geffeney and Charles T. Hanifin Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

205

225

239 269

275 283 291

Contributors ALEXEY BINGOR  Faculty of Medicine, School of Pharmacy, Institute for Drug Research, The Hebrew University of Jerusalem, Jerusalem, Israel JORDI DURBAN  Unidad de Veno´mica Evolutiva y Traslacional, Instituto de Biomedicina de Vale`ncia, Consejo Superior de Investigaciones Cientı´ficas, Vale`ncia, Spain GERAD A. FOX  Department of Earth and Biological Sciences, School of Medicine, Loma Linda University, Loma Linda, CA, USA SHANA L. GEFFENEY  Department of Biology, Utah State University-Uintah Basin, Vernal, UT, USA MATAN GERON  Faculty of Medicine, School of Pharmacy, The Institute for Drug Research, The Hebrew University of Jerusalem, Jerusalem, Israel AHMED HAFEEZ  Department of Biosciences, COMSATS University, Islamabad, Pakistan CHARLES T. HANIFIN  Department of Biology, Utah State University-Uintah Basin, Vernal, UT, USA ZAFAR HAYAT  Department of Biosciences, COMSATS University, Islamabad, Pakistan WILLIAM K. HAYES  Department of Earth and Biological Sciences, School of Medicine, Loma Linda University, Loma Linda, CA, USA JEROEN KOOL  AIMMS Division of BioMolecular Analysis, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands PHILIP LAZAROVICI  Faculty of Medicine, School of Pharmacy, Institute for Drug Research, The Hebrew University of Jerusalem, Jerusalem, Israel PETER I. LELKES  Department of Bioengineering, College of Engineering, Temple University, Philadelphia, PA, USA YOSSI MAATUF  Faculty of Medicine, School of Pharmacy, Institute for Drug Research, The Hebrew University of Jerusalem, Jerusalem, Israel STEPHEN P. MACKESSY  School of Biological Sciences, University of Northern Colorado, Greeley, CO, USA CEZARY MARCINKIEWICZ  Department of Bioengineering, College of Engineering, Temple University, Philadelphia, PA, USA FRANCIS S. MARKLAND  Department of Biochemistry and Molecular Medicine, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA; Norris Comprehensive Cancer Center, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA REHANA MASOOD  Department of Biochemistry, Shaheed Benazir Bhutto Women University, Peshawar, Pakistan MARIJA MLADIC  AIMMS Division of BioMolecular Analysis, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands CASSANDRA M. MODAHL  School of Biological Sciences, University of Northern Colorado, Greeley, CO, USA; Department of Biological Sciences, National University of Singapore, Singapore, Singapore DAVID R. NELSEN  Department of Biology, Southern Adventist University, Collegedale, TN, USA WILFRIED M. A. NIESSEN  AIMMS Division of BioMolecular Analysis, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands; Hyphen MassSpec, Voorhout, The Netherlands

ix

x

Contributors

STANO PEKA´R  Faculty of Science, Department of Botany and Zoology, Masaryk University, Brno, Czech Republic AVI PRIEL  Faculty of Medicine, School of Pharmacy, Institute for Drug Research, The Hebrew University of Jerusalem, Jerusalem, Israel PONLAPAT ROJNUCKARIN  Faculty of Medicine, Department of Medicine, Chulalongkorn University, Bangkok, Thailand K. JOHAN ROSENGREN  School of Biomedical Sciences, The University of Queensland, Brisbane, QLD, Australia CHRISTINA I. SCHROEDER  Institute for Molecular Bioscience, The University of Queensland, Brisbane, QLD, Australia ONDREJ SˇEDO  Central European Institute of Technology, Masaryk University, Brno, Czech Republic CATALINA SILVA-HIRSCHBERG  Department of Molecular Microbiology and Immunology, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA; Primary Speciality Center San Lazaro, Santiago, Chile, USA; Norris Comprehensive Cancer Center, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA GOVERT W. SOMSEN  AIMMS Division of BioMolecular Analysis, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands STEPHEN D. SWENSON  Department of Biochemistry and Molecular Medicine, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA; Department of Neurological Surgery, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA ANWAR ULLAH  Department of Biosciences, COMSATS University, Islamabad, Pakistan OLGA VASIEVA  Institute of Integrative Biology, University of Liverpool, Liverpool, UK; Ingenet Ltd, London, UK VALERIA VA´SQUEZ  Department of Physiology, College of Medicine, University of Tennessee Health Science Center, Memphis, TN, USA RAMI YAKA  Faculty of Medicine, School of Pharmacy, Institute for Drug Research, The Hebrew University of Jerusalem, Jerusalem, Israel KHIN THAN YEE  Biochemistry Research Division, Department of Medical Research, Yangon, Myanmar ZBYNEˇK ZDRA´HAL  Central European Institute of Technology, Masaryk University, Brno, Czech Republic; Faculty of Science, National Centre for Biomolecular Research, Masaryk University, Brno, Czech Republic

Part I Overview

Chapter 1 Snake- and Spider-Venom-Derived Toxins as Lead Compounds for Drug Development Philip Lazarovici Abstract Snake and spider venoms have been developed by nature as a defense mechanism against predators or to immobilize their prey by blocking the cardiovascular, respiratory, and/or nervous systems. Consequently, predators are deterred from approaching their prey by painful sensations. At a molecular level, the targeted physiological systems are blocked or stimulated by peptide toxins which, once injected into the body, modulate, though not exclusively, important cell membrane ion channels and receptors. Millions of years of constant evolution have led to the evolvement of complex venom libraries of optimized protein toxins, making them more potent, more selective, resistant to proteases, less immunogenic, and improved in terms of pharmacokinetic (PK) properties. The resulting advantage is that they induce long-term and potent pharmacodynamic (PD) effects toward unique molecular targets of therapeutic importance such as coagulation cascade proteins, receptors, and ionic channels. This optimization process has been enabled by the diversification of peptide sequences (mainly by gene duplication) and an upscaling of the complexity of toxin peptide scaffold structures, through implementation of multiple disulfide bridges and sequence-active motif diversification, leading to a wide diversity of chemical structures. This combination of pharmaceutical properties has made venom toxins valuable both as pharmacological tools and as leads for drug development. These highly tunable molecules can be tailored to achieve desirable biocompatibility and biodegradability with simultaneously selective and potent therapeutic effects. This brief overview provides basic definitions, rules, and methodologies and describes successful examples of a few drugs developed from snake toxins that are currently used in the clinic for therapy of several diseases as well as new molecular entities in clinical development based on spider-venom-derived peptide toxins. Key words Snake, Spider, Venom, Toxin, Peptide, Protein, Lead compound, Drug, Therapeutic applications

1

Introduction The key motivation behind this overview was the observation that during the last decades there has been a surge in pharmaceutical projects exploiting the extraordinary biological potency and target selectivity of snakes and spider venom toxins to develop novel drugs and diagnostics for human diseases [1], or as tools to study mammalian cell physiology and pharmacology [2, 3]. The pioneering

Avi Priel (ed.), Snake and Spider Toxins: Methods and Protocols, Methods in Molecular Biology, vol. 2068, https://doi.org/10.1007/978-1-4939-9845-6_1, © Springer Science+Business Media, LLC, part of Springer Nature 2020

3

4

Philip Lazarovici

studies of Bristol-Myers Squibb Co. in developing the angiotensinconverting enzyme inhibitor, captopril [4], based on identification and characterization of Brazilian pit viper (Bothrops jararaca) venom peptides was a turning point which stimulated many pharmaceutical companies to invest in venom-based drug discovery programs. To date, several of the currently FDA-approved drug products are developed based on some snake venom toxins and are characterized by distinct cardiovascular therapeutic effects by targeting thrombin, fibrinogen, and integrin receptors [5]. Rapid advances in proteomics, genomics, and transcriptomics of spider venoms have since resulted in affordable new technology platforms [6, 7] that also stimulate the use of spider venom peptides for drug discovery and clinical development. Therefore, this overview is presenting pharmaceutical definitions, principles, and methodologies which may further contribute to the use of snake and spider venom toxins as lead compounds in drug discovery. We anticipate that combining the huge amount of information on this topic in one brief overview will facilitate and promote future research on venomics as a platform for drug discovery and development. This primer is organized as answers to major, basic essential questions: What is a classical synthetic lead compound in drug discovery? Can peptides be used as lead compounds for drug development? What do the “most desired” peptide drugs look like? Why do snake- and spider-venom-derived toxins represent good lead compounds for peptide drug development? Which therapeutic applications are known for drugs developed from snake-venom-derived protein toxins and spider-venom-derived toxin peptides? What is a common preclinical pipeline for drug development based on snake- and spider-venom-derived protein/peptide toxins?

2

What Is a Classical Synthetic Lead Compound in Drug Discovery? Drug discovery encompasses processes ranging from target selection and validation to the selection of a drug candidate. While comprehensive drug discovery workflows are implemented predominantly in the big pharma, early discovery focuses on academia serves to identify probe molecules that can serve as tools to study targets or pathways. A successful early discovery program is built on strong established target definition and validation using a diverse set of pharmacological and cell-based assays with functional relevance to the biological system being studied. The drug discovery process involves the use of high-throughput screening (HTS) techniques to screen thousands of compounds from large to small organic synthetic molecule libraries to identify potential drug candidates for a given target (receptor, ionic channel, enzyme, etc.). The molecules which show good activity in the HTS assay are called “actives.” These are compounds that show good pharmacological

Venom Derived Toxins in Drug Development

5

activity for the specific molecular target investigated based on potent inhibitory concentration 50% (IC50) or effective concentration 50% (EC50) values. Structurally similar “actives” are clustered into groups called “clusters” and rescreened. From the top “cluster,” the most active compounds are resynthesized with high purity and rescreened for potency and physical properties upon checking for compliance with Lipinski’s rule of five. This pharmaceutical rule claims that drug molecules must have a molecular weight less than 500 g/mol, a partition coefficient (logP—a measure of hydrophobicity) less than 5, no more than 5 hydrogen bond donors, and no more than 10 hydrogen bond acceptors [8]. The best compounds from these screens are selected for further screening and are called “hits.” From these “hits” several compounds are selected for further optimization, based on good structure-activity relationship (SAR) and pharmacokinetic ADME (adsorption, distribution, metabolism, elimination) as well as toxicity (safety) profiles. These are known as the “lead compounds.” These compounds are used as templates for designing other compounds through chemical modifications and/or protein engineering and are screened using in vitro and in vivo assays. The compounds with the best pharmacological and pharmacokinetic profiles are finally selected as “drug candidates” [9].

3

Can Peptides Be Used as Lead Compounds for Drug Development? Peptides have the potential to provide drugs with the specificity and efficacy of large protein molecules combined with the smaller dose size, simpler routes of delivery, and lower cost of manufacturing as found with small, non-peptide molecules. They are, therefore, uniquely suited for the treatment of many chronic diseases which impact large patient populations. Peptides are a unique group of molecules. The 20 natural amino acids, along with many more non-natural amino acids and a series of posttranslational modifications, such as sugar and lipid incorporation, enable them to generate incredibly diverse molecules which the body sees as natural. This means that our immune system is generally tolerant of them and can pass them safely through the body without accumulation. They have the exquisite selectivity of other biologics such as antibodies but, are much smaller—making them both easier to manufacture and able to reach places that antibodies can’t. In addition, they break down to amino acids which are easily eliminated from the human body. Therapeutic peptides have taken a long time to come of age. Many of the early peptide-based therapeutics were obtained from animal tissue. The first chemical synthesis of a therapeutic peptide was that of oxytocin hormone in 1953. Recombinant synthesis of proteins was introduced in 1974, and recombinant human insulin,

6

Philip Lazarovici

the first approved peptide therapeutic to be manufactured by recombinant fermentation, was introduced in 1982. All in all, about 65 peptide-based drug products have reached FDA approval, most of them in the last three decades. Many of these peptides are hormones containing up to 41 amino acids which can be prepared by common solid-phase synthetic approaches [10]. In spite of the increasing rate of approval and the large number of candidates in the pharmaceutical pipe, therapeutic peptides as a drug class still face significant challenges. They are generally perceived as being (1) rapidly eliminated in vivo, unless chemical modifications are made; (2) labile during storage at ambient temperatures; and (3) not orally available, requiring injection and being associated with self-administration compliance issues. These challenges might be considered as a list of independent hurdles, but if a peptidebased drug is to be designed rationally for use in the clinic, all these challenges should be addressed early in a more integrative approach for each drug candidate and development process. Unfortunately, this is often not the case. Most candidates take “injectables” as default and, having started down that path, reach later stage clinical development with a number of built-in formulation issues, which remain unresolved until post-approval modifications can be made. Changes in formulation or delivery modality in late clinical trials can put the approval process at risk, so many decisions that will profoundly affect the market success of the drug product need to be considered and fixed early in the drug research and development process.

4

What Do the “Most Desired” Peptide Drugs Look Like? Although there are exceptions, such as human parathyroid hormones, luteinizing hormone-releasing hormones, and other peptides that have to be administered in a pulsatile manner to have therapeutic efficacy, there is little doubt that the preferred therapeutic form would be an orally delivered tablet or capsule containing a long-acting peptide drug that is stable at ambient temperatures and costs no more than the injectable equivalent. Although an oral formulation may not always be possible, other non-parenteral or alternative delivery platforms may be more than adequate to achieve a suitable drug form of sufficient efficacy and satisfactory compliance. There is never going to be a single set of guidelines for reaching that goal, but a rational evaluation of the intended drug at the start of the development process can help achieve this goal. The first question has to be whether, the chosen peptide in its present form is a suitable candidate for treating its intended clinical therapeutic indication. If the peptide is eliminated rapidly in vivo, as most peptide sequences are, chemical modifications to the peptide, either within the sequence or by conjugation

Venom Derived Toxins in Drug Development

7

to a polymer or a lipid, need to be considered to obtain a therapeutic “prodrug” candidate that is less easily degraded or less amenable to renal clearance. There is now a wide range of natural and synthetic polymeric conjugates available, including polyethylene glycol (PEG), hydroxyethyl starch (HES), human serum albumin (HSA), XTEN (a recombinant polypeptide), PAS (a recombinant polypeptide containing only proline, serine, and alanine), polyglutamic acid, etc. By making a covalent modification, a new molecular entity (NME) is created that may have markedly different pharmacokinetic, pharmacodynamic, and immunogenic characteristics. If the conjugation is reversible, the active pharmaceutical ingredient (API) will have the nature of a prodrug, which does not need to interact directly with the drug target; if it is non-reversible, then interaction with the target is obligatory. Covalent conjugation typically reduces potency, but this should be more than compensated for by the extended half-life. An alternative approach is to incorporate the peptide into a biodegradable long-acting-release (LAR) matrix, such as a poly D, L-lactide-co-glycolide (PLGA) polymer, or a hydrogel. Because the release profile of an LAR matrix can be designed, it is possible to program biphasic or multiphasic release of the peptide drug [11, 12].

5 Why Snake- and Spider-Venom-Derived Toxins Represent Good Lead Compounds for Peptide Drug Development? The pharmaceutical industry is currently facing unparalleled challenges to develop innovative new drugs. Although the low annual number of new drugs approved by the Food and Drug Administration (FDA) has not changed much or even decreased, research and development (R&D) investment per drug is escalating at a rapid rate. Both safety and efficacy hurdles are responsible for the rising cost in drug discovery and development. The purpose of drug design is to find the optimal structure that possesses high specificity for the biological target while decreasing the likelihood of side effects. HTS of combinatorial chemical libraries is very expensive, thus contributing heavily to drug development cost, forcing the pharmaceutical companies to rely more on academic and in house innovation, and rekindling their interest in natural products as a source of NMEs. The complex molecular scaffolds found in venom toxins represent a wide chemical diversity that is unmatched by synthetic molecules [13, 14]. Indeed, the opportunity of using venom-derived toxins for drug development can be considered as the optimal phase of “the nature drug discovery program” operating over hundreds of millions of years of evolution resulting with very potent and selective hits. With scientific advancements in modern molecular and cellular biology, analytical chemistry, and pharmacology, the unique properties of natural products including

8

Philip Lazarovici

venoms and toxins are being harnessed in order to exploit the chemical and structural biodiversity of these types of products in relation to their therapeutic effect. Often, new toxins of interest in drug design units are used for the rearrangement of chemical entities in order to generate new molecules which can be formulated into clinically useful drugs [15]. Snakes and spider venoms are abundant natural sources of NMEs, which have been relentlessly enhanced by evolution through natural selection [16]. Since most venomous animals rely heavily on their venom for survival, a constant selection pressure exists on toxin potency and efficacy. As a result, toxins typically possess good pharmaceutical characteristics such as low molecular mass, stability, folding by disulfide bridges apart from high selectivity, and affinity for a wide variety of molecular targets in mammalian physiological systems. This activity is more difficult to replicate by the synthesis of small molecules, making therefore snake and spider toxins extremely valuable as lead compounds in drug development. However, there are many technological issues to overcome, as toxin peptides are hard to obtain and characterize and the amount obtained, in particular from spider venom, is insufficient to perform all the necessary animal experiments and in vitro assays. Fortunately, novel methodologies of heterologous protein expression in bacteria, yeast, insect, and mammalian cells, as well as up scaling peptide chemical synthesis, help to provide enough quantities and allow chemical and pharmacological improvement of these toxins. 5.1 Which Therapeutic Applications Are Known for Drugs Developed from Snake-Venom-Derived Protein Toxins?

Many snakes specifically evolved to immobilize vertebrate prey through the use of large amounts of venom injected into their prey by fangs or teeth. The complex synergistic action of venom on prey tissue is mediated by its composition including several dozen of different families of toxic proteins and peptides with or without enzymatic activity [17]. Phospholipases A2, metalloproteases, serine proteases, and three-finger toxins are the most abundant protein constituents of snake venom. Some of these are neurotoxins targeting and modifying the neuromuscular junction synaptic activity, using different types of pre- and postsynaptic nicotinic and muscarinic receptors, or serve as acetylcholine esterase inhibitors. Other toxins modulate the cardiovascular system, represented by a wide range of pro- and anticoagulant proteins, disintegrins which block integrin receptors with different relative selectivity, fibrinolytic agents, natriuretic peptides, etc. Some of these toxins have been developed into research and diagnostic tools or drugs used to treat diseases of the cardiovascular system, such as blood coagulopathies and high blood pressure. To understand this approach we should exemplify two antiplatelet (anticoagulant) drugs in the clinic, tirofiban and integrilin; a compound in clinical trials for heart failure therapy, named Cenderitide; and an antiplatelet (anticoagulant) compound in preclinical research, named Vipegitide (Table 1).

b

Bothrops jararaca Angiotensin-converting enzyme c Echis carinatus d Acute coronary syndrome e Sistrurus miliarius barbouri f Dendroaspis angusticeps g Chronic heart failure h Vipera palaestinae i Pseudonaja textilis iv intravenous, aa amino acid

a

Haempatch CoVase

Thrombin Thrombin

Fibrinolysis, hemorrhage Hemorrhage Hemorrhage

Australian common brown snakei

Plasmin

Textilinin

Thrombosis

α2β1 Integrin

13 aa - Peptidomimetic antagonist Recombinant protein 7000 Da inhibitor Factor Xa-like protein Factor Va-like protein

Israeli viperh

Vipegitide

Preclinical research

CHFg

Guanylyl cyclase receptors A and B

Eastern green mambaf 37 aa - Peptide agonist

Cenderitide Clinical trials

ACSd

αIIb β3 Integrin

Pygmy rattlesnakee

ACSd

αIIb β3 Integrin

Integrilin

Hypertension

ACEb

Therapeutic application

Saw-scaled viperc

Non-peptide mimetic inhibitor Non-peptide mimetic antagonist 6 aa - Cyclic peptide antagonist

Drug in clinic Pit vipera

Captopril

Target

Tirofiban

Pharmacological definition

Development phase Snake name

Name

iv iv

iv

iv

iv

iv

iv

Oral, iv

[47] [47]

[46]

[24]

[45]

[19]

[44]

[43]

Administration routes Reference

Table 1 Selected, FDA-approved drugs and new molecular entities in chemical trials and preclinical research, developed from snake-venom-derived protein toxin

Venom Derived Toxins in Drug Development 9

10

Philip Lazarovici

Members of the snake-venom-derived “disintegrin” peptide family containing the Arg-Gly-Asp (RGD) amino acid sequence are among the most potent inhibitors of the binding of adhesive proteins to platelet glycoprotein (GP) IIb-IIIa. The first antiplatelet drug derived from a snake venom disintegrin was Tirofiban (Aggrastat). Tirofiban is a non-peptide molecule that was developed based on the structure of the RGD tripeptide motif present in the parent disintegrin molecule, echistatin isolated from the venom of the saw-scaled viper [18]. Through the synthesis and optimization of the RGD-motif lead peptide, a peptidomimetic was developed and optimized which ultimately led to the development of this selective inhibitor of platelet aggregation (Table 1). However, GPIIb-IIIa antagonists containing the RGD sequence are not integrin specific and inhibit the adhesive functions of many other RGD-dependent integrins resulting in many side effects. The disintegrin peptide, barbourin, isolated from the southeastern pigmy rattlesnake, contains a conservative amino acid substitution of Lys (K) for Arg (R) in the RGD sequence resulting with a KGD motif highly specific for GPIIb-IIIa. Using this information, a series of conformational constrained, disulfide-bridged peptides, containing the KGD amino acid sequence, were synthesized. Incorporation of the KGD sequence into a cyclic peptide template, followed by systematic optimization of the cyclic ring size, optimization of secondary hydrophobic binding-site interactions, and derivatization of the lysyl side-chain functionality of the KGD sequence, has resulted in peptide analogs which display inhibitory potency and GPIIb-IIIa selectivity comparable to that of barbourin [19]. Eptifibatide (Integrilin), the drug derivative of barbourin, is a cyclic heptapeptide that mimics the tertiary structure found in the parent toxin which allows it to bind receptors with the KGD integrin recognition sequence (Table 1). Eptifibatide is a competitive antagonist for the activated platelet glycoprotein IIb/IIIa receptor. Its mechanism of action involves preventing the binding and cross-linking of fibrinogen to the platelet surface. Arterial injury induced by percutaneous coronary interventions (PCI) such as balloon angioplasty and stenting and the spontaneously occurring disease process known as the acute coronary syndrome (ACS) share a common underlying pathophysiology. In both situations, disruption of integrity of the arterial wall initiates a cascade of platelet activation, adhesion, and aggregation. Ultimately, this pathological process may proceed to arterial thrombosis unless controlled or modified. Advances in understanding how the platelet plays a pivotal role in this process have significantly enhanced therapy for patients with ACS and have resulted in important reductions in thrombotic complications from PCI procedures. Central to these advances has been the understanding of eptifibatide interaction with platelets. The binding of eptifibatide to the integrin receptor involves displacement of receptor-associated Ca2+

Venom Derived Toxins in Drug Development

11

from the activated binding site. Through a series of small clinical trials and laboratory studies an effective dose schedule was determined. Modeling of the drug based on its two-compartment pharmacokinetics defined the role of a double-bolus initiation of therapy. In a large-scale clinical trial using this double bolus followed by infusion regimen in PCI procedures, clinical efficacy was shown to be significantly improved [20]. To date, in addition to the dual-antiplatelet therapy of aspirin and clopidogrel and systematic stent implantation, the use of the eptifibatide proved beneficial in improving early outcome of percutaneous coronary intervention (PCI), especially in higher risk clinical and/or anatomical subsets (Table 1). Another snake venom, toxin-based compound, in clinical trials for heart failure, is Cenderitide (Table 1). Cenderitide consists of the 5-amino acid amino-terminus and the 17-amino acid disulfide ring of C-type natriuretic peptide (CNP) to which the 15-amino acid carboxy-terminal of Dendroaspis natriuretic peptide (DNP) was fused. DNP is a natriuretic peptide (NP) isolated from the venom of the eastern green mamba snake (Dendroaspis angusticeps). CNP is considered to be the oldest member of the family of natriuretic peptides, which, contrary to its name, does not exhibit kidney effects. It lacks a carboxyl-terminus and is highly specific for binding to the particulate guanylyl cyclase receptor B (pGC-B). Furthermore, the lack of a carboxyl-terminus renders CNP highly susceptible to cleavage by neprilysin (NEP). Physiologically, CNP via activation of pGC-B exhibits anti-fibrotic properties and is also involved in bone metabolism and repair of the vascular endothelium. Atrial NP (ANP), brain NP (BNP), and snake DNP, however, bind to pGC-A and regulate body fluid homeostasis and induce natriuresis and vasodilation. Thus, the chimeric peptide Cenderitide, also termed CD-NP, was designed to combine the beneficial effects of CNP, which inhibits fibrosis, and DNP which exerts natriuretic activities by simultaneously activating both pCG-A and pCG-B receptors. Cenderitide was evaluated in 12 clinical trials for therapeutic efficacy in stable, chronic heart failure, preservation of left ventricular function, and moderate kidney dysfunction (ClinicalTrials.Gov). Four-hour infusion of Cenderitide was found to be safe and well tolerated, and significantly increased plasma cGMP levels and urinary cGMP excretion without adverse effects, with no change in blood pressure, indicating a favorable safety profile and expected pharmacological effects in human heart failure [21]. Another, typical basic methodology to select a lead compound from a snake venom is exemplified by our preclinical studies using disintegrins from the venom of Vipera palestinae, to develop Vipegitide, an antiplatelet (anticoagulant) lead compound [22–25]. In the first step, the snakes were maintained under good laboratory practice (GLP) conditions and were manually milked over 1 year, and the venom from several snakes was pooled to several gram quantities (alternatively the venom can be purchased from several

12

Philip Lazarovici

reliable commercial sources). In the second step, the venom was extracted in cold water and separated by high-pressure liquid chromatography (HPLC) on C18 column, using a linear gradient of increasing acetonitrile concentration. HTS of the fractions, used in cell adhesion assays of cells overexpressing α2β1 integrin, identified fraction 12 as active (selective antagonist of α2β1). The third step of purification used ion-exchange chromatography that was performed with a Mono Q column resulting with a very pure protein named VP 12 [26]. Purity of VP 12 was tested by SDS-PAGE under reduced and non-reduced conditions indicating a dimeric structure of this molecule. In the next step, the amino acid sequences of the two subunits of VP 12 were established by a combination of mass spectrometry and N-terminal sequencing of their proteolytic fragments. For this purpose, the VP 12 molecule was reduced and cysteines were ethylpyridylated (EP); the ethylpyridylated subunits were separated on reverse-phase HPLC and named VP 12A and VP 12B. Thereafter, MALDI-TOF mass spectroscopic analysis of unmodified VP 12 yielded a single molecular ion of 30,387 Da, whereas EP-VP 12A and EP-VP 12B subunits showed 15,981 Da and 15,893 Da, respectively. Based on these values, we calculated the number of cysteines in the VP 12 molecule using the equation: NCys ¼ (MPE  MNat)/106.3, where NCys is a number of cysteine residues per molecule, MPE is the mass of the reduced and ethylpyridylated protein, MNat is the mass of the native protein, and 106.3 is the mass increment due to the ethylpyridylation of one thiol group. NCys ¼ [(15,981 + 15,893)  30,387]/106.3 ¼ 13.99; therefore, the calculated number of cysteines in the whole molecule was 14. Based on the structure of other C-lectin-type protein isoforms, which has similar activity as VP 12, we predicted that each subunit contained an equal number of seven cysteines. Complete amino acid sequences of both subunits of VP 12 were established using a standard procedure including N-terminal sequencing of separated EP subunits and their tryptic, overlapping fragments, indicating that VP 12 belongs to the C-type lectin-related protein family [26]. In another approach, to specifically isolate additional antagonists targeting the α2β1 integrin, we developed a protocol based on affinity chromatography using extracellular recombinant α2β1 integrin-A domain, immobilized to a resin, resulting with the isolation of the toxin VP-i. We found that VP-i binds to the α2 integrin A domain and that it significantly inhibited adhesion of various cells to type I collagen as well as cell migration. Moreover, we found that VP-i differed structurally from the previously purified VP 12, although not functionally, and therefore represented another venom lead compound which can be utilized for further drug development [25]. In the next steps, linear and folded peptides containing the integrin-binding motif W1KTSRTSHY9 were used as lead compounds to synthesize a novel peptidomimetic antagonist of α2β1

Venom Derived Toxins in Drug Development

13

integrin, with platelet aggregation-inhibiting activity, named Vipegitide. Vipegitide is a 13-amino acid, folded peptidomimetic molecule, containing two α-aminoisobutyric acid residues at positions 6 and 8, but found not stable in human serum. Substitution of glycine and tryptophan residues at positions 1 and 2, respectively, with a unit of two polyethylene glycol (PEG) molecules yielded the peptidomimetic Vipegitide-PEG2, which was stable in human serum for over 3 h. Vipegitide and Vipegitide-PEG2 were characterized by high potency (7  1010 M and 1.5  1010 M, respectively) and intermediate efficacy (40% and 35%, respectively) as well as selectivity toward α2 integrin in inhibition of adhesion of α1/ α2 integrin-overexpressing cells toward respective collagen ligands. Interaction of both peptidomimetics with an extracellular active domain of α2 integrin was confirmed in a cell-free receptor-binding assay using recombinant α2 integrin extracellular A-domain. Integrin α2β1 receptor is found on the platelet membrane and triggers collagen-induced platelet aggregation. Vipegitide and VipegitidePEG2 inhibited α2β1 integrin-mediated adhesion of human and murine platelets under static and flow conditions by 50%. They efficiently blocked collagen I-induced platelet aggregation in platelet-rich plasma and whole-human blood. Higher potency of Vipegitide than Vipegitide-PEG2 was consistent with results of computer modeling of these molecules in water. These peptidomimetic NMEs were acutely tolerated in mice upon intravenous bolus injection of 50 mg/kg. These results underline the use of Vipegitide and Vipegitide-PEG2 molecules as platelet aggregationinhibiting lead drug compounds and pharmacological tools in antithrombotic therapy [24]. Table 1 also presents additional snake venoms derived NMEs in preclinical research: (1) textilinin-1, a selective plasmin inhibitor whose application as an antifibrinolytic agent can reduce blood loss associated with surgeries; (2) haempatch, a factor Xa-like protein characterized by potent procoagulant effects developed as a hemostatic agent to reduce blood loss resulting from surgery or trauma; and (3) CoVase, a procoagulant cofactor that may have applications as a systemic antibleeding agent in the treatment of hemorrhage. Finally, it is important to note that a wide range of other protein toxins found in snake venoms have been used in the development of laboratory diagnostic kits available in hematology laboratories of hospitals, to measure levels of fibrinogen, prothrombin, blood-clotting factors, and protein C associated with hemostatic disorders. 5.2 Which Therapeutic Applications Are Known for SpiderVenom-Derived Toxin Peptides?

At present, about 39,000 species of spiders exist worldwide. With the exception of several hundred species, all of these are poisonous using their venom’ toxins for prey catching. Some spiders beeing small produce small amounts of venom. However, they contain very potent toxins to quickly paralyze their prey therefore, explaining the minute amount of venom produced. Once the pray is paralyzed, the spiders ingest the liquefied tissues due to hydrolytic activity

14

Philip Lazarovici

of venom enzymes. The main components of spider venoms are large-molecular-weight enzymes and peptides of molecular weight between 1 and 10 kDa. The high number of molecular targets known to be modulated by spider toxins, with high selectivity and potency, makes the spider venom-isolated peptides attractive lead compounds. These peptides’ primary sequence is typically less than 45-amino acid residues in length, forming a three-dimensional fold structure, known as the “inhibitor cystine knot (ICK) motif,” which confers to these toxins’ chemical, thermal, and biological stability. ICK is defined as an antiparallel β-sheet stabilized by a cysteine knot comprising a ring formed by two disulfide bridges and the intervening sections of polypeptide backbone, with a third disulfide bridge piercing the ring to create a pseudo-knot. In addition, spider venom contains small peptides without disulfide bonds, which permeate cells’ plasma membrane by pore formation. Since spiders employ their venom to paralyze the prey, it is not surprising that these venoms mainly contain peptides that fast paralyze the nervous system by targeting and modulating the activity of neuronal ion channels and ligand-gated ion channel receptors such as sodium channels (NaV), potassium channels (KV), calcium channels (CaV), acidsensing ion channels (ASIC), mechanosensitive channels (MS), transient receptor potential channels (TRP), and certain purinergic receptors (P2X). Not only do most of these peptides have selectivity for a given category of ion channel, but also they can have anything from relative selectivity to exquisite selectivity for a given channel subtype [27]. This potential for high target affinity in the range of picomolar to nanomolar and selectivity for ionic channels and receptor subtypes makes spider-venom peptides an ideal natural source of lead compounds for therapy of pain and channelopathies (such as epilepsy and arrhythmias), diseases caused by disturbed function of ion channel subunits. Although a wide range of generic names have been reported for spider peptide toxins, in the last decade a more rational nomenclature for naming peptide toxins has been adopted by ArachnoServer and UniProtKB [28], helping in pharmaceutical applications. Although none of the spider-venom-derived peptide toxins and related analogs have emerged yet as drugs in the clinic, many of them are in preclinical development and/or clinical trials predominantly targeting NaVs and CaVs. Voltage-gated NaVs and voltage-gated CaVs play key roles in action potential generation in sensory pain-mediating neurons and consequently these channels have become important targets for the development of analgesic drugs. In Table 2 we present selected spider-venom-derived NMEs targeting NaV and CaV, investigated by academic laboratories and pharma companies, to envision their potential for pain therapeutic applications. The NaV channel family is comprised of nine subtypes, NaV1.1–NaV1.9, and spider-venom peptides have been identified that target each of these subtypes. NaV1.7, NaV1.8, and NaV1.9 are expressed predominantly in nociceptive neurons and are

Venom Derived Toxins in Drug Development

15

Table 2 Selected new molecular peptide entities in preclinical research for analgesia (iv injection), developed from spider-venom-derived peptide toxins Name

Spider name

Pharmacological definition

Target

Reference

CcoTx-1 GpTx-1 ProTx-II JNJ 63955918 HwTx-IV

Ceratogyrus cornuatus Grammostola porteri Thrixopelma pruriens

NaV1.7

Ornithoctonus huwena

33 aa Peptide inhibitor 34 aa Peptide inhibitor 30 aa Peptide inhibitor ProTx-II peptide analog 35 aa Peptide inhibitor

[29] [30, 48] [49, 50] [33] [51]

ω-Aga-IVA ω-Aga-IVB

Agelenopsis aperta

48 aa Peptide inhibitor

CaV2.1

[52, 53] [54]

HwTx-X SNX-325

Ornithoctonus huwena Segestria florentina

28 aa Peptide inhibitor 49 aa Peptide inhibitor

CaV2.2

[55] [56]

CSTX-1

Cupiennius salei

74 aa Peptide inhibitor

CaV1.1–1.4

[57]

ω-TRTX-Cc1a

Pelinobius muticus

39 aa Peptide inhibitor

CaV1.2/1.3

[58]

thought to play important roles in pain signaling. The 1.7 subtype of the voltage-gated sodium channels (NaV) has emerged as one of the hottest drug targets for channel modulators over the past decade. Thus, selective inhibitors of NaV1.7 based mainly on tarantula spider venom toxins, are potential leads for development of future analgesic drugs towards different types of pain. During this process of drug development caution must be taken since nonspecific blockades of other NaV subtypes (lack of selectivity) may lead to substantial side effects, including seizures, arrhythmias, and impaired motor function. Pfizer Co. developed potent and selective blockers of NaV1.7 with improved therapeutic properties using ceratotoxin-1 (CcoTx1) isolated from the venom of the tarantula Ceratogyrus cornuatus [29]. Employing structure-activity relationship and chemical modification, the research and development group of the company created potent (IC50 in nanomolar range) and selective (80-fold and 20-fold over the closely related NaV1.2 and NaV1.6 channels, respectively, and 1000-fold over skeletal NaV1.4 and cardiac NaV1.5 channels) peptide drug inhibitors of NaV1.7. Amgen Co. identified and characterized GpTx-1, a known antagonist of TTX-sensitive sodium channels, from the venom of another tarantula spider, Grammostola porteri [30]. GpTx-1 was first reported as a CaV channel blocker after isolation from venom of the closely related Chilean tarantula, Grammostola rosea, and named GTx1 as also identified in the venom of the Chilean copper tarantula, Paraphysa scrofa. On the basis of its potency and selectivity, GpTx-1 was used as a lead compound to target Nav1.7. The synthetic, folded-analog (Ala5, Phe6, Leu26, Arg28) GpTx-1 was found to be very potent (IC50 of 1.6 nM against NaV1.7) and selective (about thousand fold less selective

16

Philip Lazarovici

for NaV1.4, and NaV1.5) peptide drug inhibitors of NaV1.7 [31]. Merck Co. focused their development program on optimizing ProTX-II as a lead compound for a painkiller drug selectively targeting NaV1.7 [32]. Protoxin-II (ProTX-II) was isolated from the Peruvian green velvet tarantula, Thrixopelma pruriens, and is composed of 30 amino acids and 3 disulfide bonds which adopt an ICK motif structure. ProTx-II inhibits both tetrodotoxin-sensitive and tetrodotoxin-resistant voltage-gated sodium channels. ProTxII inhibits activation by shifting the voltage dependence of channel activation to more positive potentials. ProTx-II blocks NaV1.7 with an IC50 value of 300 pM, and NaV1.2, NaV 1.5, and NaV 1.6 with IC50 of 41 nM, 79 nM, and 26 nM, respectively. Using ProTX-II as a scaffold, Janssen Co. screened a library of over 1500 venom-derived peptides and identified JNJ63955918 as a potent, highly selective, closed-state Nav1.7 blocking peptide. They show that JNJ63955918 induced a pharmacological insensitivity to pain that closely recapitulates key features of the Nav1.7 knockout phenotype seen in mice. Moreover they demonstrated that this NME has a high degree of selectivity, coupled with a closed-state-dependent mechanism of action which is required for strong efficacy. Their research indicated that peptides such as JNJ63955918 may represent a viable, non-opioid alternative, for the pharmacological treatment of severe pain [33]. Medimmune Co. developed a triple mutant of HwTx-IV (E1G, E4G, Y33W) which is one of the most potent blockers of NaV1.7 reported to date (IC50 of 500 pM). Huwentoxin IV (HwTx-IV) lead compound is a 35-amino acid peptide toxin which also adopts an ICK fold, inhibiting NaV1.7 with an IC50 of 26 nM, isolated from the venom of the Chinese bird-eating spider Selenocosmia huwena. Characterization of the toxin-channel interaction revealed that the peptide binds to one of the four voltage sensor domains (VSDs) of the NaV1.7 channel, in contrast to small molecules-local anesthetic drugs that modulate the pore region. The channel pore is more highly conserved among the different NaV1 subtypes compared to the VSDs, making VSDs attractive targets for the development of selective drugs [31, 34]. There are many reports indicating the existence of drug development programs at academic institutions and companies, as described above, which have further optimized the potencies and selectivity of other spider-venom-derived peptide lead compounds as inhibitors of NaV1.7. The huge pharmaceutical advantage of spider-venom-derived toxin peptides, as emphasized in Table 2 with tarantula-derived lead compounds, demonstrates the strong opportunity and feasibility for development of more selective analgesic drugs targeting NaV1.7. The CaV family of channels is composed of about ten subtypes denoted CaV1.1–1.4 (L-type), CaV2.1 (P/Q-type), CaV2.2 (N-type), CaV2.3 (R-type), and CaV3.1–3.3 (T-type) based on their biophysical, electrophysiological, and pharmacological

Venom Derived Toxins in Drug Development

17

properties. CaV2.1, CaV2.2, and CaV3 are the best validated targets for the development of analgesic drugs [35]. This was supported by the pharmaceutical development of the nonnarcotic pain reliever Ziconotide (Prialt), an ICK peptide (derived from the venom of the cone snail Conus magus) selectively targeting CaV2.2 which is an FDA-approved analgesic drug available in the clinic. Many spider-venom-derived toxins target and inhibit CaV channels, but very few of these peptides display selectivity for a particular CaV channel subtype. For example, PnTx3-6, a neurotoxin purified from the venom of the spider Phoneutria nigriventer is a nonselective blocker of CaV1, CaV2.1, CaV2.2, and CaV2.3 by physically occluding the channels that was shown to be analgesic in mice models of neuropathic pain [36, 37]. This lack of complete selectivity does not necessarily exclude this peptide toxin from the drug development process, since different chemical modifications may eventually lead to a more selective lead compound. Table 2 presents a few selected examples of spider-venom-derived NMEs, developed by pharma companies or reported in academic research, targeting CaVs, to envision their potential for pain therapeutic applications. Pfizer Co. was among the first to use, as lead compounds, spider-venom-derived ω-agatoxin- IVA and IVB, 48-amino acid peptides isolated from the venom of the American funnel-web spider Agelenopsis aperta that target CaV2.1 channels with nanomolar high affinity. These peptides are the most selective CaV2.1 blockers described to date, which upon binding to the channel shift the activation voltage of the channel to nonphysiological positive potentials and confer relief from different types of pain. Other spider-venom-derived toxin peptides that display selectivity for other calcium channel subtypes, the CaV2.2 channels, are HWTX-X, a 28-amino acid peptide with three disulfide bridges ICK toxin peptide, purified from the venom of the spider Ornithoctonus huwena and SNX-325, a 49-amino acid peptide with four disulfide bonds from the venom of the spider Segestria florentina. However, the characterization of the analgesic effect of these toxins has not yet been reported. Other spider-venom-derived toxin peptides target with relative selectivity neuronal CaV1.1-1.4. Typical examples are represented by CSTX-1, a 74-amino acid peptide with four disulfide bridge toxin isolated from the venom of the hunting spider Cupiennius salei and ω-TRTX-Cc1a, a 39-amino acid peptide with three disulfide bridges from the venom of the tarantula Citharischius crawshayi (now Pelinobius muticus). Table 2 indicates that the spider-venom-derived peptide toxin NMEs in preclinical research for analgesia, using the intravenous delivery route, are ICK peptides in the range of 30–74 amino acids folded by three or four cysteine bridges. Drug development of these peptides may result with smaller size lead compounds. Definitely, further screening of spider venoms might provide a source of NMEs for drug development to cope with the unmet clinical need for selective inhibitors of

18

Philip Lazarovici

Fig. 1 A flowchart of the screening, identification, lead discovery, and in vitro and in vivo preclinical development of drug candidates of new molecular entities, based on snake- and spider-derived protein/peptide toxins. Bioassay-guided

Venom Derived Toxins in Drug Development

19

CaV2.1 in pain relief. Continued methodologies and technical advances in the spider-venom-based drug discovery process are likely to uncover many new leads for other therapeutic applications besides analgesia.

6 What Is a Common Preclinical Pipeline for Drug Development Based on Snake- and Spider-Venom-Derived Protein/Peptide Toxins? As a consequence of their high potency and selectivity, snake- and spider-venom-derived proteins and peptide toxins have proved particularly useful for in vitro and in vivo proof-of-concept pharmacological studies. However, to use them for drug development, a number of important issues associated with the development of the suitable lead compound, its safety, pharmacokinetics and pharmacodynamics, and delivery route need to be investigated. Optimization of toxin-derived lead compound structure-activity relationships, efficient delivery to peripheral and central targets, and its therapeutic index will help to determine whether or not these peptide toxins can be considered candidates for drug development. In the preceding sections we surveyed a number of snakeand spider-venom peptides with successful potential therapeutic applications (Tables 1 and 2). The snake and spider “venomics to drug” approach, as presented in Fig. 1, starts with collection of venom and its fractionation by gel permeation, ion exchange, and reverse-phase HPLC, followed by bioassay-guided screening to detect active toxins (Fig. 1, steps 1, 2). They are characterized by a combination of N-terminal sequencing, SDS-PAGE, and mass spectrometric determination of the molecular masses and the cysteine content (Fig. 1, steps 2–5). Protein fractions indicating a single electrophoretic band, molecular mass, and N-terminal ä Fig. 1 (continued) isolation and purification approaches are based on pharmacological screening of crude venoms fractions obtained by gel permeation, ion exchange, and/or reverse-phase high-performance liquid chromatography (HPLC). Purified venom peptides are sequenced using mass spectroscopy and/or Edman degradation. Alternatively, sequence-based venom peptide discovery incorporating transcriptomics, proteomic, and bioinformatic approaches can contribute to the identification of bioactive venom protein/peptide toxin hit, making it a complementary and additive strategy to the pharmacological screening-based toxin drug discovery. Synthetic or recombinant toxins can be prepared and evaluated by the same pharmacological assays. Structure-activity relationship (SAR) studies will further generate lead compounds which are refined based on optimal potency, selectivity, efficacy, and safety. Characterization of these compounds’ pharmacokinetics (ADME properties) and safety profiles will allow the selection of drug candidates to be evaluated in animal disease models and the clinical trials

20

Philip Lazarovici

sequence can be straightforwardly assigned by BLAST analysis to a known protein family (Fig. 1, step 6). Large-molecular-weight toxins are reduced and cysteines are ethylpyridylated (EP); the ethylpyridylated subunits are again separated on reverse-phase HPLC and processed for MALDI-TOF mass spectroscopic analysis. Toxins indicating heterogeneous or blocked C- or N-termini are analyzed by SDS-PAGE and the bands of interest are subjected to reduction, carbamidomethylation, and in-gel tryptic digestion. The resulting tryptic peptides are then analyzed by MALDI-TOF mass fingerprinting followed by amino acid sequence determination of selected doubly and triply charged peptide ions by collisioninduced dissociation tandem mass spectrometry (Fig. 1, steps 4, 5). In general, snake and spider venoms were screened in mediumto high-throughput assays against targets of therapeutic interest (receptors, ionic channels), and then “hit isolated fractions” were chromatographically fractionated and the individual fractions were rescreened in order to isolate proteins/peptides responsible for bioactivity. In some cases, incomplete sequence information acquired via MS/MS and/or Edman degradation was used for designing primers to amplify transcripts encoding the toxin of interest from a venom gland cDNA library and used for production of recombinant toxins and purified spider peptides were submitted for solid-phase peptide synthesis. Both approaches were carried out in milligram scale (Fig. 1, step 7). Technologies such as nanoscale nuclear magnetic resonance (NMR) for structure determination, full-length combinatorial chemical synthesis, transcriptomics, and proteomics are all crucial and complementary technologies for the successful development of snake- and spider-venom-derived toxins as lead drug compounds in steps 1–7 described in Fig. 1. Once the protein or peptide sequence is identified, the minimal binding motif (sequence) to the receptor or ionic channel is further determined (Fig. 1, step 7a) and used for the synthesis of cyclic, smallmolecule mimetics of venom peptide toxins (Fig. 1, step 8). The peptide motif that mediates the interaction of these peptides with their cognate integrin receptors or ion channels can be remarkably small (3–10 amino acids). As long as high-quality amounts of protein/peptide toxins are available, this enables their crystallography and X-ray analysis, computer modeling, design of non-peptide mimetic, identification of small-molecule mimetics via in silico screening of chemical libraries, or a combination of these approaches. In general, active, cyclic, and small peptides are obtained following the study of structure-function relationships (SAR) (Fig. 1, step 8). Achieving water solubility and adequate affinity to the receptor/ionic channel site is the main goal of the “venomics to drug” pipeline research and is critical in this step (Fig. 1, step 9). To reach sufficient selectivity and potency in vivo it is necessary to find NMEs with binding affinities in the low nanomolar or even picomolar range. Binding assays are relatively

Venom Derived Toxins in Drug Development

21

simple and highly suitable for calculating the affinity for the molecular target. Molecules with high affinity move on to further molecular and cellular assays that determine the functional effects on the target molecule in order to distinguish agonistic from antagonistic receptor effects or openers or closers of ionic channels. Binding and functional assays define the in vitro segment of this selection process. Refined lead candidate molecules with sufficient binding affinity, selectivity, and functional efficacy and a good therapeutic window (Fig. 1, steps 9, 10), estimated by cell death in vitro assays (see Chap. 11), are taken forward into in vivo safety and pharmacokinetic evaluation and selection processes (Fig. 1, step 11). Based on the data achieved, the optimal molecules from the ADME and toxicological evaluations are selected as drug candidates (Fig. 1, step 12) and in general licensed by academic laboratories or further developed by pharma companies toward clinical trials (Fig. 1, step 13). From the ADME point of view, we would like to stress that in the spider-venom-derived peptide toxins, the structure is stabilized by an ICK motif. When rapid hydrolysis proves to be disadvantageous for a spider-venom peptide of therapeutic interest, strategies such as D-amino acid substitution of susceptible residues, cyclization to reduce conformational flexibility, protection of the termini via C-terminal amidation, or use of N-terminal alkylation could be employed to improve metabolic stability (proteolytic resistance). Because of their inherent proteolytic resistance, the plasma half-life of ICK peptides is likely to be determined by the rate at which they are metabolized and eliminated. There are different strategies that can be employed to reduce their degradation and clearance rates, such as increasing the peptide mass by PEGylation, conjugation to carrier proteins, or making peptides more hydrophobic in order to enhance their nonspecific binding to serum albumin. As emphasized in Tables 1 and 2, for cardiovascular diseases and chronic conditions such as persistent pain, injection of snake- and spidervenom-derived proteins or peptide drugs is likely to be an acceptable route of administration. In certain pathologies, alternative local routes of administration such as intranasal, dermal, pulmonary, and/or intrathecal may be a viable option if the difficulties associated with peptide delivery are outweighed by the benefits of treatment. Generally, however, oral delivery is likely to be desirable but not yet considered for snake- and spider-venom-derived protein/peptide drugs. A major barrier to successful oral delivery of peptide and protein molecules is their inherent instability in the lumen of the gastrointestinal tract. In human gastric fluid, the larger peptide drugs including somatostatin, calcitonin, secretin, glucagon, and insulin are metabolized rapidly, while the smaller peptides like oxytocin, vasopressin, and desmopressin, containing disulfide bridges between their cysteine amino acids, partially cyclize their structure and result in high resistance to pepsin cleavage and show good stability in gastric fluids. In human small intestinal fluid, however, both small and large peptides degraded

22

Philip Lazarovici

rapidly with the exception of the cyclic peptide cyclosporine and the disulfide bridge-containing peptides octreotide and desmopressin, which showed good stability [38]. We would like to propose that the intrinsic stability of ICK peptides is likely to facilitate the development of oral delivery strategies since they will presumably have much higher stability in the gastrointestinal tract and as a result increased bioavailability upon oral delivery. In case of unsatisfactory results, the snake- and spider-derived drug candidates may be improved by substituting the disulfide bond cyclization to backbone cyclization to increase stability and methylation to increase intestinal permeability. The approach of backbone cyclization is a general method well accepted in medicinal chemistry, by which conformational constraint is imposed on peptides. In backbone cyclization, a peptidomimetic is formed by covalently interconnecting atoms in the backbone (N and/or C) of a target linear peptide to form a ring. The advantage of backbone cyclization over other modes of peptide cyclization is that cyclization is achieved mainly using backbone atoms and not side chains that are essential for biological activity. This method has been shown to dramatically enhance the metabolic stability of substance P in serum without affecting its selectivity for neurokinin receptor [39] and similarly for somatostatin analogs [40]. Recent structural studies on libraries of cyclic hexapeptides led to the identification of common backbone conformations that are very important to the oral availability of peptides [41]. With this background we predict that the presence of N-methylated cis-peptide bonds at certain locations of snake- and spider-derived-peptide drug candidates may increase their metabolic stability and intestinal permeability through a suitable conformational preorganization [42]. The high degree of method innovation in the field of snake and spider toxins will generate a new wave of drug research and development. Interdisciplinary research using new technologies, as described in this book, will be essential for the future success of using snake and spider toxins as novel NMEs that can make significant contributions to the cure of human diseases. Through the combination of venomics (a combination of MS and molecular biology methodology), transcriptomics, proteomics, recombinant proteins molecular pharming, and combinatorial peptide synthesis, the contribution of snake and spider toxins to the future drug development seems to be very promising.

Acknowledgments Philip Lazarovici holds the Jacob Gitlin Chair in Physiology and is affiliated and supported by the Adolph and Klara Brettler Medical Research Center, the David R Bloom Center for Pharmacy, and the Grass Center for Drug Design and Synthesis of Novel Therapeutics at the Hebrew University of Jerusalem, Israel.

Venom Derived Toxins in Drug Development

23

References 1. King GF (2011) Venoms as a platform for human drugs: translating toxins into therapeutics. Expert Opin Biol Ther 11:1469–1484. https://doi.org/10.1517/14712598.2011. 621940 2. de Souza JM, Goncalves BDC, Gomez MV, Vieira LB, Ribeiro FM (2018) Animal toxins as therapeutic tools to treat neurodegenerative diseases. Front Pharmacol 9:145. https://doi. org/10.3389/fphar.2018.00145 3. Morales-Lazaro SL, Hernandez-Garcia E, Serrano-Flores B, Rosenbaum T (2015) Organic toxins as tools to understand ion channel mechanisms and structure. Curr Top Med Chem 15:581–603. https://doi.org/10. 2174/1568026615666150217110710 4. Cushman DW, Ondetti MA (1991) History of the design of captopril and related inhibitors of angiotensin converting enzyme. Hypertension 17:589 LP–589592. http://hyper.ahajournals. org/content/17/4/589.abstract 5. Koh CY, Kini RM (2012) From snake venom toxins to therapeutics – cardiovascular examples. Toxicon 59:497–506. https://doi.org/ 10.1016/j.toxicon.2011.03.017 6. Liang S (2008) Proteome and peptidome profiling of spider venoms. Expert Rev Proteomics 5:731–746. https://doi.org/10.1586/ 14789450.5.5.731 7. Sanggaard KW, Bechsgaard JS, Fang X, Duan J, Dyrlund TF, Gupta V, Jiang X, Cheng L, Fan D, Feng Y, Han L, Huang Z, Wu Z, Liao L, Settepani V, Thogersen IB, Vanthournout B, Wang T, Zhu Y, Funch P, Enghild JJ, Schauser L, Andersen SU, Villesen P, Schierup MH, Bilde T, Wang J (2014) Spider genomes provide insight into composition and evolution of venom and silk. Nat Commun 5:3765. https://doi.org/10. 1038/ncomms4765 8. Lipinski CA, Lombardo F, Dominy BW, Feeney PJ (2001) Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv Drug Deliv Rev 46:3–26 9. Roy A (2018) Early probe and drug discovery in academia: a minireview. High Throughput 7:E4. https://doi.org/10.3390/ht7010004 10. Bray BL (2003) Large-scale manufacture of peptide therapeutics by chemical synthesis. Nat Rev Drug Discov 2:587–593. https:// doi.org/10.1038/nrd1133 11. Kaur P, Garg V, Gulati M, Singh SK (2016) Oral delivery of antidiabetic polypeptide-k:

journey so far and the road ahead. Curr Drug Deliv 13:236–244 12. Bruno BJ, Miller GD, Lim CS (2013) Basics and recent advances in peptide and protein drug delivery. Ther Deliv 4:1443–1467. https://doi.org/10.4155/tde.13.104 13. Peigneur S, Tytgat J (2018) Toxins in drug discovery and pharmacology. Toxins (Basel) 10:E126. https://doi.org/10.3390/ toxins10030126 14. Harvey AL (2014) Toxins and drug discovery. Toxicon 92:193–200. https://doi.org/10. 1016/j.toxicon.2014.10.020 15. Mathur S, Hoskins C (2017) Drug development: lessons from nature. Biomed Rep 6:612–614. https://doi.org/10.3892/br. 2017.909 16. Casewell NR, Wuster W, Vonk FJ, Harrison RA, Fry BG (2013) Complex cocktails: the evolutionary novelty of venoms. Trends Ecol Evol 28:219–229. https://doi.org/10.1016/ j.tree.2012.10.020 17. Xiong S, Huang C (2018) Synergistic strategies of predominant toxins in snake venoms. Toxicol Lett 287:142–154. https://doi.org/10. 1016/j.toxlet.2018.02.004 18. Starnes HB, Patel AA, Stouffer GA (2011) Optimal use of platelet glycoprotein IIb/IIIa receptor antagonists in patients undergoing percutaneous coronary interventions. Drugs 71:2009–2030. https://doi.org/10.2165/ 11595010-000000000-00000 19. Scarborough RM, Naughton MA, Teng W, Rose JW, Phillips DR, Nannizzi L, Arfsten A, Campbell AM, Charo IF (1993) Design of potent and specific integrin antagonists. Peptide antagonists with high specificity for glycoprotein IIb-IIIa. J Biol Chem 268:1066–1073 20. Gilchrist IC (2003) Platelet glycoprotein IIb/IIIa inhibitors in percutaneous coronary intervention: focus on the pharmacokineticpharmacodynamic relationships of eptifibatide. Clin Pharmacokinet 42:703–720. https://doi.org/10.2165/00003088200342080-00001 21. Kawakami R, Lee CYW, Scott C, Bailey KR, Schirger JA, Chen HH, Benike SL, Cannone V, Martin FL, Sangaralingham SJ, Ichiki T, Burnett JCJ (2017) A human study to evaluate safety, tolerability, and cyclic GMP activating properties of Cenderitide in subjects with stable chronic heart failure. Clin Pharmacol Ther 104:546. https://doi.org/10.1002/ cpt.974

24

Philip Lazarovici

22. Momic T, Arlinghaus FT, Arien-Zakay H, Katzhendler J, Eble JA, Marcinkiewicz C, Lazarovici P (2011) Pharmacological aspects of Vipera xantina palestinae venom. Toxins (Basel) 3:1420–1432. https://doi.org/10. 3390/toxins3111420 23. Momic T, Cohen G, Reich R, Arlinghaus FT, Eble JA, Marcinkiewicz C, Lazarovici P (2012) Vixapatin (VP12), a c-type lectin-protein from Vipera xantina palestinae venom: characterization as a novel anti-angiogenic compound. Toxins (Basel) 4:862–877. https://doi.org/ 10.3390/toxins4100862 24. Momic T, Katzhendler J, Shai E, Noy E, Senderowitz H, Eble JA, Marcinkiewicz C, Varon D, Lazarovici P (2015) Vipegitide: a folded peptidomimetic partial antagonist of alpha2beta1 integrin with antiplatelet aggregation activity. Drug Des Devel Ther 9:291–304. https://doi.org/10.2147/DDDT.S72844 25. Arlinghaus FT, Momic T, Ammar NA, Shai E, Spectre G, Varon D, Marcinkiewicz C, Heide H, Lazarovici P, Eble JA (2013) Identification of alpha2beta1 integrin inhibitor VP-i with anti-platelet properties in the venom of Vipera palaestinae. Toxicon 64:96–105. https://doi.org/10.1016/j.toxicon.2013.01. 001 26. Staniszewska I, Walsh EM, Rothman VL, Gaathon A, Tuszynski GP, Calvete JJ, Lazarovici P, Marcinkiewicz C (2009) Effect of VP12 and viperistatin on inhibition of collagen-receptor-dependent melanoma metastasis. Cancer Biol Ther 8:1507–1516 27. Lahiani A, Yavin E, Lazarovici P (2017) The molecular basis of toxins’ interactions with intracellular signaling via discrete portals. Toxins (Basel) 9:1–58. https://doi.org/10.3390/ toxins9030107 28. King GF, Gentz MC, Escoubas P, Nicholson GM (2008) A rational nomenclature for naming peptide toxins from spiders and other venomous animals. Toxicon 52:264–276. https:// doi.org/10.1016/j.toxicon.2008.05.020 29. Shcherbatko A, Rossi A, Foletti D, Zhu G, Bogin O, Galindo Casas M, Rickert M, HasaMoreno A, Bartsevich V, Crameri A, Steiner AR, Henningsen R, Gill A, Pons J, Shelton DL, Rajpal A, Strop P (2016) Engineering highly potent and selective microproteins against Nav1.7 Sodium Channel for treatment of pain. J Biol Chem 291:13974–13986. https://doi.org/10.1074/jbc.M116.725978 30. Murray JK, Ligutti J, Liu D, Zou A, Poppe L, Li H, Andrews KL, Moyer BD, McDonough SI, Favreau P, Stocklin R, Miranda LP (2015) Engineering potent and selective analogues of GpTx-1, a tarantula venom peptide antagonist

of the Na(V)1.7 sodium channel. J Med Chem 58:2299–2314. https://doi.org/10.1021/ jm501765v 31. Pennington MW, Czerwinski A, Norton RS (2017) Peptide therapeutics from venom: current status and potential. Bioorg Med Chem 26:2738. https://doi.org/10.1016/j.bmc. 2017.09.029 32. Priest BT, Blumenthal KM, Smith JJ, Warren VA, Smith MM (2007) ProTx-I and ProTx-II: gating modifiers of voltage-gated sodium channels. Toxicon 49:194–201. https://doi.org/ 10.1016/j.toxicon.2006.09.014 33. Flinspach M, Xu Q, Piekarz AD, Fellows R, Hagan R, Gibbs A, Liu Y, Neff RA, Freedman J, Eckert WA, Zhou M, Bonesteel R, Pennington MW, Eddinger KA, Yaksh TL, Hunter M, Swanson RV, Wickenden AD (2017) Insensitivity to pain induced by a potent selective closed-state Nav1.7 inhibitor. Sci Rep 7:39662. https://doi.org/10.1038/ srep39662 34. Revell JD, Lund P-E, Linley JE, Metcalfe J, Burmeister N, Sridharan S, Jones C, Jermutus L, Bednarek MA (2013) Potency optimization of Huwentoxin-IV on hNav1.7: a neurotoxin TTX-S sodium-channel antagonist from the venom of the Chinese bird-eating spider Selenocosmia huwena. Peptides 44:40–46. https://doi.org/10.1016/j. peptides.2013.03.011 35. Lee S (2013) Pharmacological inhibition of voltage-gated ca(2+) channels for chronic pain relief. Curr Neuropharmacol 11:606–620. https://doi.org/10.2174/ 1570159X11311060005 36. Vieira LB, Kushmerick C, Hildebrand ME, Garcia E, Stea A, Cordeiro MN, Richardson M, Gomez MV, Snutch TP (2005) Inhibition of high voltage-activated calcium channels by spider toxin PnTx3-6. J Pharmacol Exp Ther 314:1370–1377. https://doi. org/10.1124/jpet.105.087023 37. Dalmolin GD, Silva CR, Rigo FK, Gomes GM, Cordeiro M d N, Richardson M, Silva MAR, Prado MAM, Gomez MV, Ferreira J (2011) Antinociceptive effect of Brazilian armed spider venom toxin Tx3-3 in animal models of neuropathic pain. Pain 152:2224–2232. https://doi. org/10.1016/j.pain.2011.04.015 38. Wang J, Yadav V, Smart AL, Tajiri S, Basit AW (2015) Toward oral delivery of biopharmaceuticals: an assessment of the gastrointestinal stability of 17 peptide drugs. Mol Pharm 12:966–973. https://doi.org/10.1021/ mp500809f 39. Byk G, Halle D, Zeltser I, Bitan G, Selinger Z, Gilon C (1996) Synthesis and biological

Venom Derived Toxins in Drug Development activity of NK-1 selective, N-backbone cyclic analogs of the C-terminal hexapeptide of substance P. J Med Chem 39:3174–3178. https:// doi.org/10.1021/jm960154i 40. Afargan M, Janson ET, Gelerman G, Rosenfeld R, Ziv O, Karpov O, Wolf A, Bracha M, Shohat D, Liapakis G, Gilon C, Hoffman A, Stephensky D, Oberg K (2001) Novel long-acting somatostatin analog with endocrine selectivity: potent suppression of growth hormone but not of insulin. Endocrinology 142:477–486. https://doi.org/10. 1210/endo.142.1.7880 41. Marelli UK, Ovadia O, Frank AO, Chatterjee J, Gilon C, Hoffman A, Kessler H (2015) Cis-peptide bonds: a key for intestinal permeability of peptides? Chemistry 21:15148–15152. https://doi.org/10.1002/ chem.201501600 42. Weinmuller M, Rechenmacher F, Kiran Marelli U, Reichart F, Kapp TG, Rader AFB, Di Leva FS, Marinelli L, Novellino E, MunozFelix JM, Hodivala-Dilke K, Schumacher A, Fanous J, Gilon C, Hoffman A, Kessler H (2017) Overcoming the lack of oral availability of cyclic hexapeptides: design of a selective and orally available ligand for the integrin alphavbeta3. Angew Chem Int Ed Engl 56:16405–16409. https://doi.org/10.1002/ anie.201709709 43. Waheed H, Moin SF, Choudhary MI (2017) Snake venom: from deadly toxins to life-saving therapeutics. Curr Med Chem 24:1874–1891. https://doi.org/10.2174/ 0929867324666170605091546 44. Barrett JS, Murphy G, Peerlinck K, De Lepeleire I, Gould RJ, Panebianco D, Hand E, Deckmyn H, Vermylen J, Arnout J (1994) Pharmacokinetics and pharmacodynamics of MK-383, a selective non-peptide platelet glycoprotein-IIb/IIIa receptor antagonist, in healthy men. Clin Pharmacol Ther 56:377–388 45. Wojta J (2016) Cenderitide: a multivalent designer-peptide-agonist of particulate guanylyl cyclase receptors with considerable therapeutic potential in cardiorenal disease states. Eur Hear J Cardiovasc Pharmacother 2:106–107. https://doi.org/10.1093/ ehjcvp/pvv043 46. Flight SM, Johnson LA, Du QS, Warner RL, Trabi M, Gaffney PJ, Lavin MF, de Jersey J, Masci PP (2009) Textilinin-1, an alternative anti-bleeding agent to aprotinin: importance of plasmin inhibition in controlling blood loss. Br J Haematol 145:207–211. https://doi.org/10. 1111/j.1365-2141.2009.07605.x

25

47. Earl STH, Masci PP, de Jersey J, Lavin MF, Dixon J (2012) Drug development from Australian elapid snake venoms and the venomics pipeline of candidates for haemostasis: Textilinin-1 (Q8008), Haempatch (Q8009) and CoVase (V0801). Toxicon 59:456–463. https://doi.org/10.1016/j. toxicon.2010.12.010 48. Deuis JR, Wingerd JS, Winter Z, Durek T, Dekan Z, Sousa SR, Zimmermann K, Hoffmann T, Weidner C, Nassar MA, Alewood PF, Lewis RJ, Vetter I (2016) Analgesic effects of GpTx-1, PF-04856264 and CNV1014802 in a mouse model of NaV1.7-mediated pain. Toxins (Basel) 8:E78. https://doi.org/10. 3390/toxins8030078 49. Wright ZVF, McCarthy S, Dickman R, Reyes FE, Sanchez-Martinez S, Cryar A, Kilford I, Hall A, Takle AK, Topf M, Gonen T, Thalassinos K, Tabor AB (2017) The role of disulfide bond replacements in analogues of the tarantula toxin ProTx-II and their effects on inhibition of the voltage-gated sodium Ion Channel Nav1.7. J Am Chem Soc 139:13063–13075. https://doi.org/10. 1021/jacs.7b06506 50. Park JH, Carlin KP, Wu G, Ilyin VI, Musza LL, Blake PR, Kyle DJ (2014) Studies examining the relationship between the chemical structure of protoxin II and its activity on voltage gated sodium channels. J Med Chem 57:6623–6631. https://doi.org/10.1021/jm500687u 51. Liu Y, Wu Z, Tang D, Xun X, Liu L, Li X, Nie D, Xiang Y, Yi J, Yi J (2014) Analgesic effects of Huwentoxin-IV on animal models of inflammatory and neuropathic pain. Protein Pept Lett 21:153–158 52. Sluka KA (1998) Blockade of N- and P/Q-type calcium channels reduces the secondary heat hyperalgesia induced by acute inflammation. J Pharmacol Exp Ther 287:232–237 53. Nebe J, Ebersberger A, Vanegas H, Schaible HG (1999) Effects of omega-agatoxin IVA, a P-type calcium channel antagonist, on the development of spinal neuronal hyperexcitability caused by knee inflammation in rats. J Neurophysiol 81:2620–2626. https://doi.org/10. 1152/jn.1999.81.6.2620 54. Adams ME (2004) Agatoxins: ion channel specific toxins from the American funnel web spider, Agelenopsis aperta. Toxicon 43:509–525. https://doi.org/10.1016/j.toxicon.2004.02. 004 55. Liu Z, Dai J, Dai L, Deng M, Hu Z, Hu W, Liang S (2006) Function and solution structure of Huwentoxin-X, a specific blocker of N-type calcium channels, from the Chinese

26

Philip Lazarovici

bird spider Ornithoctonus huwena. J Biol Chem 281:8628–8635. https://doi.org/10. 1074/jbc.M513542200 56. Newcomb R, Palma A, Fox J, Gaur S, Lau K, Chung D, Cong R, Bell JR, Horne B, Nadasdi L (1995) SNX-325, a novel calcium antagonist from the spider Segestria florentina. Biochemistry 34:8341–8347 57. Kubista H, Mafra RA, Chong Y, Nicholson GM, Beirao PSL, Cruz JS, Boehm S, Nentwig W, Kuhn-Nentwig L (2007) CSTX1, a toxin from the venom of the hunting spider

Cupiennius salei, is a selective blocker of L-type calcium channels in mammalian neurons. Neuropharmacology 52:1650–1662. https://doi. org/10.1016/j.neuropharm.2007.03.012 58. Klint JK, Berecki G, Durek T, Mobli M, Knapp O, King GF, Adams DJ, Alewood PF, Rash LD (2014) Isolation, synthesis and characterization of omega-TRTX-Cc1a, a novel tarantula venom peptide that selectively targets L-type Cav channels. Biochem Pharmacol 89:276–286. https://doi.org/10.1016/j.bcp. 2014.02.008

Chapter 2 Analytics for Bioactivity Profiling of Complex Mixtures with a Focus on Venoms Marija Mladic, Wilfried M. A. Niessen, Govert W. Somsen, and Jeroen Kool Abstract This chapter introduces bioactivity and bioaffinity terms in relation to mixture profiling and gives the significance of bioactivity and/or bioaffinity profiling of biologically active mixtures in general, and for bioactive mixtures in drug discovery research in particular. Further, the chapter gives an overview of the common and less common analytical approaches for bioactivity profiling of bioactive mixtures. Special focus is put on bioassay-guided fractionation as the standard technique employed (in identification and purification of bioactive molecules from a bioactive mixture), and on state-of-the-art post-column bioactivity profiling approaches, also providing examples and limitations of these analytical methods. On-column and pre-column bioactivity profiling analytics is also discussed. Examples of bioactive molecules identified and purified from different natural products are given with emphasis on molecules isolated from animal venoms. Finally, this chapter briefly discusses the importance of bioactivity profiling of metabolic mixtures in drug discovery. Key words Analytical approaches, Bioactivity profiling, Natural products, Venoms, Drug discovery, Metabolic mixtures

1 Bioactivity/Bioaffinity Profiling of Mixtures for Toxin Identification, Drug Discovery, and Natural Product Profiling Bioactivity describes the property of a compound or a set of compounds to modulate biochemical and physiological functions of living organisms. Identification of bioactive compounds in mixtures, such as venoms, plant extracts, and metabolic mixtures, is an important challenge in different scientific and application fields that are often highly intertwined. Most of the times, bioactive compounds are present in very complex mixtures. These mixtures can at the same time contain multiple bioactives with different mechanisms of action or with synergistic and/or antagonistic effects. Bioactivity profiling of a mixture of compounds means bioactivity assessment of mixture constituents and identification of

Avi Priel (ed.), Snake and Spider Toxins: Methods and Protocols, Methods in Molecular Biology, vol. 2068, https://doi.org/10.1007/978-1-4939-9845-6_2, © Springer Science+Business Media, LLC, part of Springer Nature 2020

27

28

Marija Mladic et al.

possibly present bioactives. Bioactivity profiling is generally directed at finding bioactives for a particular target, most often a drug target. Commonly, it is used in different screening programs that aim at the identification of bioactive compounds from different sample sources. Identifying toxins in venoms for better understanding of venom functioning, for example, is an important research field. In close relation to this, discovery of new biopharmaceutical lead compounds from venoms is an important topic within pharmaceutical sciences. The analysis of metabolic mixtures of drugs generated either in vivo or in vitro for potential bioactive metabolites represents yet another application of bioactivity profiling. Furthermore, bioactivity profiling of mixtures is important in toxicology, forensics, natural product sciences, and environmental and food analysis. In all these fields, a combination of similar analytical and biological tools is used in order to identify biologically active compounds from the large variety of complex mixtures. In this introductory chapter, analytical approaches for bioactivity profiling of mixtures of interest are discussed, providing a complete picture of the state of the art in the analysis of biologically active samples. Special attention will be devoted to venom analysis and bioactivity-guided identification of toxins. An important tool in identification and characterization of venom toxins, and in discovery of new lead compounds, is highthroughput screening (HTS). HTS is the process of testing large libraries of diverse chemical structures against disease targets to identify “hits,” i.e., bioactive compounds, that modulate a particular biochemical pathway. HTS assays are performed in 96-, 384-, or 1536-well microtiter plates. Robotics, liquid-handling devices, and sensitive detectors are brought together with advanced data processing and control software to optimize sample throughput. Based on ligand-target interactions, HTS is characterized by its simplicity, rapidness, low cost, and high efficiency, as well as a high information outcome. HTS has been developed to rapidly conduct millions of chemical, genetic, and/or pharmacological tests in order to identify potential drug candidates [1]. HTS is an excellent tool for testing the bioactivity of large libraries of pure compounds. However, the identification of one or more bioactive components in a mixture requires a combination of analytical strategies for separation and chemical detection, often involving mass spectrometry (MS) and/or nuclear magnetic resonance spectroscopy (NMR), with biological assays for the bioactivity detection. Very often, bioactive mixtures are analyzed by their bioaffinity rather than their bioactivity. Bioaffinity represents the property of a compound to bind to the target protein, which is generally considered as the precondition for the compound to cause a biological effect. In case of venoms, bioactivity is directly probed by measuring, e.g., coagulopathic properties, phospholipid hydrolysis, and/or cytolytic activities. Prior to bioactivity assessment, extraction from the

Analytics for Bioactivity Profiling of Complex Mixtures with a Focus on Venoms

29

starting material (for example a venom or plant extract) and pre-concentration of the compounds of interest may be needed. Finally, the correct attribution of bioactivity to a certain chemical structure has to be confirmed by additional tests and full characterization of the bioactive compound is required, chemically, pharmacologically, as well as toxicologically. 1.1 Bioassay-Guided Fractionation

Over the years, different strategies have been developed for bioactivity analysis of complex mixtures, such as venoms, and identification of the bioactive components (e.g., venom-derived toxins). The first and still most frequently used approach is (bio)assay-guided fractionation (BGF), which was initially introduced for profiling natural extracts of plants. Similar approaches (often under different names) were developed and used in other fields where bioactivity profiling of complex mixtures is of interest [2]. A well-known example in this regard is effect-directed analysis (EDA) used for bioactivity assessment and identification of toxic pollutants present in environmental samples. As the main focus of this book is toward analytical strategies for toxin identification, bioactivity profiling of environmental extracts will not be discussed further. Strategies used in this particular research field were recently reviewed by Jonker et al. [3]. The BGF approach has been successfully used for identification of many lead compounds from natural products, such as animal venoms and extracts from plants. The traditional BGF is a multistep approach which most often consists of four steps after optional sample preparation: (1) initial separation of the mixture of interest followed by low-resolution fractionation (microfractionation); (2) bioassaying of collected fractions and identification of the bioactive fraction(s); (3) an additional (preferably orthogonal) separation step, followed by fractionation and bioactivity assessment performed on the bioactive fraction to isolate the pure compound for chemical analysis; and (4) (chemical analysis and) structure elucidation of the bioactive compound. In many cases, after two rounds of fractionation, the fractions are still too complex for bioactive compound identification and further rounds of fractionations are needed. The sample preparation that precedes BGF can be simple dilution of a lyophilized sample with optional centrifugation, as is often done in case of animal venoms. In case of plant material, the sample preparation step usually includes grinding and sawing of starting material followed by subsequent extraction of the bioactive sample using different organic solvents. The choice of the initial separation highly depends on the nature of the sample analyzed, but most often it is performed using liquid chromatography (LC) based on reversed-phase (RPLC), size-exclusion (SEC), or ion-exchange (IEX) separation mechanisms. SEC and IEX are especially useful for the initial separation of mixtures mainly containing peptides and proteins such as animal venoms [4, 5]. Other

30

Marija Mladic et al.

approaches include gel chromatography followed by LC [6] or preparative TLC [7]. Subsequent fractionation of the separated mixture, also known as microfractionation, is performed manually in low resolution (long time intervals for fraction collection), generally resulting in 1 mL fractions. The bioactivity testing of the fractions collected is usually performed after freeze-drying and reconstitution in a bioassay-compatible solvent, and it may include multiple up-concentration or dilution pipetting steps. From our point of view (see below), this approach is to be considered as off-line bioactivity testing since the bioassays are performed separately from the separation and detection. A variety of in vivo and in vitro bioassays are employed for the bioactivity assessment. Traditionally, bioassays included testing of the fractions on isolated animal organs or tissues or testing on whole animals. With the development of cell culturing techniques and the possibility of isolation and/or overexpression of venom toxins (e.g., peptides and enzymes), the choice and the possibilities of bioassays have drastically increased. Many successful applications of various bioassays using the BGF approach have been reported since for snake and insect venoms and for plant extracts [4, 7–12]. When one or more bioactive fractions are found, efforts are put in chemical analysis of the bioactive compound. Depending on the complexity of the bioactive fractions, an orthogonal separation can be performed with another round of microfractionation and bioassaying. In some cases, only one step of separation, fractionation, and bioactivity testing is needed. After sufficiently pure samples have been obtained, the chemical analysis of the bioactive compound is performed using high-resolution MS and NMR. MS is widely used because of its high sensitivity and the amount of information it provides. In the case of small molecules, UV-VIS data (absorbance of ultraviolet or visible light spectrometry) obtained with photodiode array (PDA) detectors can be useful in dereplication (identification of already known compounds) and assigning of an unknown bioactive compound to a certain chemical class [13]. However, NMR is most often needed for the final confirmation of the structure. This requires relatively high amounts of pure compounds. In case of bioactive peptides and proteins, MS-based proteomics approaches are widely used for amino acid sequencing, and for the characterization of posttranslational modifications and the positions of cysteine bridges. However, MS-based proteomics approaches do not provide full structure characterization. There may be issues with distinguishing between the isomeric amino acids leucine and isoleucine. For characterization of the 3D structures of proteins, other techniques like X-ray crystallography, electron microscopy, and NMR are needed. Huge efforts put in isolation of sufficient material for NMR prolong the time needed for identification of the bioactive. Finally, full chemical and pharmacological characterization of the compound is performed. Compounds

Analytics for Bioactivity Profiling of Complex Mixtures with a Focus on Venoms

31

obtained from natural products, such as venoms, plants, and microorganisms, using the BGF approach have often reached the market as drugs in their original form. However, next to that, semisynthetic approaches, pharmacophore identification, and structureactivity relationship [14] studies on these compounds resulted in many new drugs with improved properties. This is for instance the case for various betalactam [15, 16] and aminoglycoside [17] antibiotics and for statins [18]. 1.2 Pre-column and On-Column Approaches for Bioaffinity Profiling of Mixtures

Even though highly successful and still widely used, the BGF approach is very laborious and time consuming. Therefore, many academic groups have put significant effort in the development of alternative approaches for analysis of bioactive mixtures [19]. Part of these efforts were due to advances in separation sciences, in hyphenation of separation technologies to MS, and especially in hyphenation of bioassays to separation techniques and MS detection [20]. Approaches based on protein-ligand affinity separation will be briefly discussed here first (see Jonker et al. [19] for an elaborate discussion). Different techniques based on ultrafiltration, size-exclusion separation, and centrifugation can be used to analyze affinity of compounds binding to a solubilized target. Continuous ultrafiltration [21, 22] is a technique for bioaffinity screening of mixtures by injecting the target protein into an ultrafiltration chamber, featuring a molecular weight cutoff membrane, after which a bioactive mixture can be pumped through the chamber. After washing, a disruption step or sufficient elution time for the bound ligands to dissociate from the target protein allows processing of the eluted ligands toward MS for identification. This technique has been applied in a HTS format for screening of a compound library for affinity toward a Streptococcal enzyme [23]. Furthermore, online coupling of ultrafiltration with LC–MS was developed to screen natural extracts and other mixtures for bioaffinity toward the relevant targets [24–26]. Pulsed ultrafiltration is a further development and improvement of this approach [27]. In pulsed ultrafiltration, continuous infusion is replaced by injection of a small amount of sample into the ultrafiltration unit. Molecules that do not bind to the target protein are flushed away through the molecular weight cutoff membrane using a continuous buffer flow. Again, sufficient elution time or use of a dissociation buffer allows migration of the ligands toward the MS for their subsequent identification. Pulsed ultrafiltration has been applied in a metabolic stability study [28] and in the screening for inhibitors of the retinoid X receptor [29]. In the field of natural extract screening, the use of pulsed ultrafiltration has been described for affinity selection of cyclooxygenase inhibitors from medicinal plants [30]. SEC coupled to MS can be applied as an affinity selection method to screen binding of multiple ligands to a receptor

32

Marija Mladic et al.

[31, 32]. This technique is based on the incubation of mixtures of ligands with a target protein followed by a rapid SEC purification of the ligand-target complexes. A subsequent disruption step, commonly in continuous-flow format, precedes either direct analysis of the released ligands by MS or a final SPE–(LC)–MS analysis. The SEC-based affinity trapping approach was developed in an automated HTS format by Neogenesis and commercialized as the Automated Ligand Identification System (ALIS) [33]. In ALIS, a small online SEC affinity purification coupled to MS is used to screen the ligands binding to the relevant target protein. The method has been successfully used as an automated SEC–LC–MS setup to screen for ligands of lipid phosphatase [34] and the muscarinic M2 acetylcholine receptor [35]. Similar methodologies were reported to study ligand binding to G protein-coupled receptors (GPCRs), an important class of drug targets [36]. Another similar approach utilizing in-solution ligand binding to a target protein is based on His-tagged target proteins with high affinity for immobilized metal affinity columns (IMAC). Protein-ligand trapping on an IMAC column is followed by column washing with a disruption eluent allowing for release of the ligands from their target protein and subsequent identification [37]. Alternatively, magnetic beads with IMAC properties have been used for ligand fishing in an off-line format [38, 39]. An online format of this technique has also been developed. Here, the process of washing and target protein-ligand disruption occurs online in an analytical chromatography setup by switching the eluents applied over the beads. Beads are trapped and finally released from the online system by use of a retractable magnet [40, 41]. The MS binding assay approach [42] is a technique based on a conventional radioligand binding assay in which the radioactivity detection of bound ligand is replaced by non-radiolabeled ligand detection using MS. In short, ligands are incubated with the target protein in the presence of ligand and the bound fraction is separated from the unbound one using vacuum filtration. The fraction with bound ligand is subjected to a disruption buffer, to release the ligand, which is subsequently detected by MS. MS binding assays were developed for analysis of inhibitors of γ-aminobutyric acid (GABA) transporters [43]. Other approaches include frontal and zonal affinity chromatography [44], which are based on chromatographic affinity columns with an immobilized target protein. In frontal affinity analysis, a mixture of interest is continuously infused onto the column, while in the zonal affinity analysis the mixture of interest is injected as a plug onto the column. In both cases, an MS is used as detector. The affinity of the ligands for the immobilized drug target on the column determines the elution order: a longer elution time implies a higher affinity. Affinity chromatography techniques are applied to study binding affinities of the compounds [45–47] as well as in the

Analytics for Bioactivity Profiling of Complex Mixtures with a Focus on Venoms

33

screening of mixtures of compounds for bioactives [48]. Alternatively, capillary electrophoresis (CE) can be applied as a separation technique to determine affinity interactions of proteins and ligands. Affinity capillary electrophoresis (ACE) [49] has been used to study affinity and purity of llama-derived antibodies, also referred to as nanobodies [50]. Immunoaffinity capillary electrophoresis (IA-CE) has been applied to study the affinity of peptides in complex matrixes [51]. Applications of ACE and IA-CE can be found in a review by Haselberg et al. [52]. 1.3 Post-column Bioactivity Profiling of Mixtures

As discussed above, in the originally developed BGF approach, the bioactivity profiling of the mixture is performed in an off-line postcolumn fashion; that is, the bioactivity/bioaffinity of the collected fractions is assessed after a separation step that is decoupled from the chemical analysis. On the other hand, the described alternative approaches for bioaffinity analysis are based on either pre-column bioassays followed by separation or on-column bioaffinity analysis and separation. In these cases, the bioassays are coupled to the chemical detection, e.g., MS identification, either in an off-line or in an online mode. In the last two decades, our group has made considerable efforts in the development of advanced strategies for post-column bioactivity profiling of mixtures [20]. These strategies are based on online (and at-line) post-column screening of complex mixtures and comprise parallel bioactivity/bioaffinity assessment and chemical analysis, which in turn allows for more efficient bioactivity profiling of mixtures. An important advantage of postcolumn analytics over the pre-column and on-column bioassay strategies is the possibility to also implement bioactivity analysis rather than only bioaffinity assessment. This implies that bioactivity analysis using functional cellular bioassays [53] and bioassays based on enzyme activity [54, 55] is within the scope. The concepts of online post-column screening [56, 57] and advances in this field [20] have been reviewed in detail. Here, the discussion is focused at the online bioaffinity/bioactivity analysis with parallel MS detection. The basic principle of current online screening approaches lies in coupling of RPLC separation to continuous-flow bioassays (mostly fluorescence-based assays) with parallel MS detection. Parallel bioassaying and MS detection are enabled by a post-column split of the LC effluent in an approximately 1:9 ratio in most of the methods developed. In this case, the smaller fraction (10%) is subjected to the bioassay and larger fraction is directed to the MS. Typically, two types of bioassays are used: bioassays based on the enzymatic conversion of a fluorogenic substrate into a fluorescence product [58, 59] and bioassays based on fluorescence enhancement upon binding of a tracer ligand to its target receptor or binding protein [60, 61]. The mixing of the LC effluent and the bioassay reagents is achieved by using two superloops followed by a continuous-flow incubation in a reaction coil

34

Marija Mladic et al.

after each superloop. A superloop represents a syringe through which the corresponding solution is delivered using an external LC pump. The first superloop and reaction coil allow interaction of the separated compounds with the target molecule (an enzyme or a receptor), while the second superloop introduces the fluorescent ligand that interacts with the ligand-target mixture. The flow rates and tubing dimensions determine the incubation times in the reaction coils and are limited to a few minutes in order to avoid severe post-column band broadening. In some cases, additional off-line NMR analysis was incorporated for full chemical characterization of the bioactive compounds [62]. The online screening approach was successfully demonstrated in proof-of-principle studies as well as in screening of natural products [63] and metabolic mixtures [62]. However, there are some important drawbacks and limitations of this approach. The main limitation of online screening is its restriction to bioassays with short incubation times. Thus, for example enzymatic assays with long incubation times and also elaborate radioligand binding affinity assays (that also comprise many pipetting steps including a filtration step) cannot be performed in an online approach. Obviously, cell-based assays such as gene-reporter assays requiring long incubation times cannot be used in this setup either. Other disadvantages include the necessity for modification and adaptation of the traditional microtiter plate-based assays for online screening approaches; compatibility issues between LC mobile-phase composition and solvent composition permitted in the bioassay, especially with respect to organic solvent content; and need of (relatively) high concentrations of target receptor or enzyme to provide a robust system with an adequate dynamic range. Finally, suitable fluorescence enhancement ligands or substrates that are enzymatically converted into a fluorescent product are needed for the bioassay. To address some of the drawbacks of the online screening approach, various modifications and alternative approaches have been developed. For example, the possibility of a direct MS bioassay readout has been explored for the analysis of mixtures toward targets for which fluorescence tracers or ligands are not available [64], but this approach did not find broad application due to issues with sensitivity, matrix effects, and ion-source contamination [20]. For samples only available in small amounts, such as animal venoms, the online approach was miniaturized toward using a nanoLC separation with parallel MS detection and the bioassay being performed on a microfluidic chip [65]. This allowed sample injections of 5 μg via a 10–500 nL injection loop in comparison to the 250 μg injections in the conventional online setup using a 10–50 μL injection loop. The miniaturized online approach was applied to the screening of snake and cone snail venoms, and toad skin extracts for bioaffinity assessment toward nicotinic

Analytics for Bioactivity Profiling of Complex Mixtures with a Focus on Venoms

35

acetylcholine receptors (nAChR) by using its binding protein analog [66–68]. Finally, the limitation regarding the choice of bioassays was tackled by introducing a post-column high-resolution fractionation onto microtiter plates prior to bioassaying. This approach was named at-line post-column bioaffinity profiling and, to some extent, was highly similar to the online approach. In this approach, the LC effluent was split post-column: the larger fraction was guided to a modified autoinjector that served as the fractionation system, whereas the smaller fraction is directed to the MS. Prior to fractionation, the bioassay reagents were mixed with the LC effluent using two syringe pumps to deliver the reagents in a continuous-flow format. This approach allowed for coupling of the LC separation to more types of bioassays since there are no limitations in incubation time [69]. Further development of the at-line fractionation approach was advanced by introducing the original BGF concept; that is, the fractionation step was introduced directly after the flow split followed by a solvent evaporation step (freeze-drying) and subsequent bioassaying. Compared to the original BGF approach, advancements are direct fractionation and solvent evaporation on microtiter well plates and a parallel postcolumn split to MS. In this way, potential bioassay interferences related to the presence of organic solvents used in the preceding RPLC separation are eliminated. These modifications resulted in a technique known as at-line nanofractionation and was for example applied in 96- and in 384-well plate formats to study metabolic mixtures of ligands toward the histamine receptor H4 [53, 70]. In these studies, both bioaffinity and bioactivity of metabolic mixtures were assessed using radioligand binding assays and a β-galactosidase reporter gene assay. Initially, at-line nanofractionation was developed as a decoupled approach. Namely, two separate chromatographic runs were performed for bioaffinity/bioactivity profiling and for MS detection. One LC-UV run was hyphenated to MS detection and the other was used in nanofractionation coupled to the bioassay. Later on, the technique was applied to venom analysis with focus on bioactive peptides, e.g., for screening of snake venoms for bioactive substances toward thrombin, factor Xa, and angiotensinconverting enzyme [49], which are drug targets for pathologies of the cardiovascular system. Both thrombin and factor Xa are enzymes involved in the coagulation cascade and valid drug targets for the treatment of abnormal coagulation. In case of venom profiling for thrombin and factor Xa, in total 39 venoms were screened employing RPLC coupled to fluorescence-based enzyme activity assays and parallel MS detection [71]. Bioactivity profiles of the venoms screened were studied for the presence of inhibitors, but in most cases only proteases were detected. In case inhibitors were detected (i.e., negative peaks were observed) the bioactivity

36

Marija Mladic et al.

was assigned to components with m/z values obtained from the parallel MS measurements. Furthermore, components with protease activity (detected as positive peaks in the bioactivity chromatogram) of the different venoms screened were compared in an attempt to understand the functional phylogeny. Venom profiling directed toward ACE inhibitors is another interesting example of using nanofractionation for venom profiling considering the inhibiting ACE activity of venoms. As an advancement to the earlier developed nanofractionation approaches, the use of hydrophilic interaction liquid chromatography (HILIC) was investigated as a complementary tool to RPLC separation [72]. All venoms were initially screened in RPLC mode and the most interesting candidates were subjected to rescreening with RPLC and an additional screening in HILIC mode. Since venoms comprise very complex samples, several constituents in a fraction may be responsible for the bioactivity. The additional use of the HILIC mode allowed narrowing down of the number of candidates responsible for the bioactivity per venom analyzed. Additionally, the bioactive fractions were, directly from the wells, subjected to nanoLC–MS/ MS analysis, and amino acid sequences of the relevant peptides were determined from the fragmentation patterns observed in the MS/MS spectra. Yet another study using similar analytics describes the assessment of coagulation activity of individual venom toxins in crude venoms. The assay is a time-course spectrophotometric measurement which kinetically measures the clotting profile of plasma. A major effect following envenomation by many venomous animals is perturbation of blood coagulation caused by hemotoxic venom toxins. The approach rapidly assessed which procoagulant and anticoagulant activities were present in hemotoxic venoms, and these activities were directly correlated with their accurate masses from the parallel obtained MS data [73].

2

Profiling Bioactive Mixture Classes in Life Sciences with Emphasis on Venoms

2.1 Venoms for Toxin Identification and as Source of Bioactive Compounds

Venoms are complex mixtures of peptides and proteins affecting various physiological processes upon delivery in the soft tissue of a prey organism in a process called envenomation. For identification of individual venom toxins in a venom, and for studying their individual toxic effects, toxins in venoms have to be purified first and during purification their native functional structures must be preserved. Examples of traditional BGF procedures for identification and biological assessment of different venom toxin classes are given in the following paragraph. Hyaluronidases are enzymes known as spreading factors, cleaving hyaluronan which is a non-proteoglycan polysaccharide. During envenoming, the victim may have a systemic collapse if circulating hyaluronan is degraded in the bloodstream. In one

Analytics for Bioactivity Profiling of Complex Mixtures with a Focus on Venoms

37

study, this enzyme was purified through three chromatographic steps, cation exchange (CEX), gel filtration, and hydrophobic chromatography, and then identified using fast protein liquid chromatography with Edman degradation sequencing [74]. In another study, RPLC was used for protein purification from crude venom, followed by hyaluronidase identification using MS approaches [75]. When looking at the isolation of coagulopathic toxins from venoms, in one study Echis carinatus venom was subjected to chromatographic purification using a two-step approach. First, a gel filtration separation of bioactive fractions by Sephadex G-75 was performed and subsequently anion-exchange chromatography (AEX) by DEAE-Sepharose. Collected fractions were monitored for their ability to modulate coagulation of plasma from which anticoagulant fractions were assessed for purity and mass estimation using SDS/PAGE [76]. In a study directed at profiling Bothrops atrox venom, P-I class metalloproteases and acidic phospholipase A2s (PLA2) were purified for identification using several chromatographic separations, initiated by Sephacryl S-200 [77]. Fractions having the monitored enzymatic activity of interest were further separated using AEX followed by an ultrafiltration step. RPLC was used to measure the purity of the enzymes fractionated, while enzymes were characterized by determining their molecular masses, isoelectric points, specific functional activity, and partial amino acid sequencing. Three-finger toxins (3FTx) are polypeptides that consist of 57–82-amino-acid residues and have characteristic three loops that protrude from the central hydrophobic globule. For purification of a 3FTx from Naja kaouthia venom, firstly gel filtration was applied from which the 3FTx of interest was further purified by IEX [78]. From there, the fraction containing the 3FTx under study was subjected to RPLC to obtain the pure toxin for biological studies. L-amino acid oxidase (LAAO) was purified from Crotalus durissus terrificus venom using two chromatographic steps: SEC followed by Phenyl Sepharose FF chromatography, where the bioactivity was traced using an enzymatic assay [79]. Identity was confirmed using MS. Disintegrins affect cell-cell and cell-matrix interaction by binding to membrane proteins called integrins. The isolation of disintegrins from Agkistrodon contortrix laticinctus venom was performed by a two-step procedure in which applying CEX was the first step [80]. This was followed by RPLC of bioactive fractions to locate nonenzymatic venom peptides such as disintegrins, next to enzymatically active zones assessed using bioassays relevant to these toxicities. The isolation of two heterodimeric disintegrins was accomplished and confirmed by MS approaches. Bradykinin (BK) is a peptide associated with several physiological processes such as inflammatory responses and induction of nociception and hyperalgesia. Bradykinin-like peptides in venoms

38

Marija Mladic et al.

comprise a class of angiotensin-I-converting enzyme [49] inhibitors and as a consequence modulate blood pressure regulation. In one study, peptide purification started fractionating crude venom from Bitis gabonica Rhinoceros using SEC, after which the lowmolecular-mass fraction was further purified using RPLC [81]. A bioassay assessing ACE inhibition and in vivo edema potentiation showed functional activity of the purified peptide, while identity was confirmed using MS/MS. Bioactive components found in venoms are characterized by high potency and high selectivity for their target making venoms also a valuable source of lead compounds in drug discovery. The use of venoms in treatment of some medical conditions dates back a long time [82]. Snake venoms have been used to treat rheumatism and gastrointestinal and other disorders, while spider venoms have been known to possess anti-asthmatic and anticancer activity. Moreover, animal venoms represent a source of important pharmacological tools in receptor and disease studies. Venom-based drug discovery started with the isolation of peptide from the venom of Brazilian viper, Bothrops jararaca, with bradykinin-potentiating properties [83, 84]. This ultimately led to the development of captopril [85], the first marketed ACE inhibitor. ACE is an enzyme involved in the regulation of blood pressure and its inhibition has been successfully used in the treatment of hypertension and heart failure. Since the discovery of captopril, the interest in bioactive compounds from venoms has increased resulting in six new drugs and many potential candidates [82]. For example, two peptides found in snake venoms were developed into eptifibatide and tirofiban, antiplatelet agents acting as glycoprotein IIb/IIIa inhibitors. Eptifibatide (Integrilin) [86] is a cyclic heptapeptide derived from a snake disintegrin isolated from the venom of Sistrurus miliarius barbouri, the southeastern pygmy rattlesnake, while tirofiban (Aggrastat) [87] is a small molecule that was based on a peptide found in the venom of Echis carinatus, the saw-scaled viper. Moreover, snake venoms are constantly being explored as sources of leads in drug discovery aimed at a wide range of syndromes and diseases [82, 88–92]. Peptides from snake venoms are also used as research tools [93]. Alpha-bungarotoxin, a neurotoxin isolated from the venom of Taiwanese banded krait, Bungarus multicinctus, is a competitive and irreversible antagonist of the nAChR that is used in studies of neuromuscular junctions, nAChRs, and it contributed to the understanding of multiple sclerosis [94]. One particular source of venoms relevant in both medicine and as source of pharmacological tools is cone snails [95]. Marine cone snails are predatory sea snails that immobilize their pray via a venomous harpoon sting. Venom of cone snails provides an excellent source of bioactive peptides for discovery of new lead compounds [96–98]. The biggest success story is related to the development of ziconotide (Prialt) [99], an exact synthetic replicate of a

Analytics for Bioactivity Profiling of Complex Mixtures with a Focus on Venoms

39

ω-conotoxin originally found in the cone snail Conus magus. This highly selective blocker of voltage-sensitive calcium channels was approved for the treatment of chronic pain. Vetter and Lewis [98] highlighted other conopeptides acting on sodium ion channels and various receptors (i.e., acetylcholine, neurotensin, noradrenaline, and 5-hydroxytryptamine receptors). However, little is known about their progress toward clinical development. A μ-conotoxin (XEP-018) from Conus consors venom was reported to be moving into preclinical development on account of its particularly long duration of action [100] but its current status is unknown. Another example of successful drug development from animal venoms is exenatide, which was isolated from the venom of Gila monster, Heloderma suspectum [101]. Exenatide is approved for treatment of type II diabetes based on its similarity to the glucagon-like peptide-1, which plays a role in the control of appetite and glucose levels in blood [102, 103]. Many other venomous species, such as spiders and sea anemones, are currently being studied for the potential therapeutic use [104–106]. The possibilities include bioactives that act through binding to calcium, potassium, or chlorine channels; inhibition of nAChRs; inhibition of thrombin or factor Xa; inhibition of noradrenaline transporter; blocking of N-methyl-D-aspartate (NMDA) receptors; and many others [107]. Further exploration of the therapeutic potential of venoms will open new opportunities for drug discovery. Evidently, the same analytical procedures are in place when purifying venom toxins for their structural characterization and assessing their toxicological effect in in vitro bioassays and in vivo experiments. The huge potential for finding new leads in venoms and developing them into drugs lies in the diversity of venomous animals and the compounds they contain. Moreover, the venom-based peptides have the favorable characteristics of being highly selective and potent toward respective drug targets, which makes them very interesting for discovery of new leads. Importantly, the high content of disulfide bridges of some venomous peptides, securing the three-dimensional structure needed for very potent and specific protein interactions, provides stability against proteolytic degradation and adds to their chemical and thermal stability. Furthermore, the small size of venom peptidederived drugs contributes to their low immunogenicity and opens the possibility for different administration routes [108]. 2.2 Natural Products in Drug Discovery

The use of natural products in the treatment of a large number of diseases goes deep in the past, probably as far as humans exist. The medicinal and/or toxic properties of primarily plants, but also animal extracts, were known to old civilizations who documented their use in these purposes. The earliest records on medical use of plants date from around 2600 BCE and were made in ancient Mesopotamia. Other old documentation on natural medication encompasses Egyptian “Ebers Papyrus” from around 1500 BCE, Chinese “Materia Medica” with first written records dating from

40

Marija Mladic et al.

around 1100 BCE, and Indian “Ayurveda” dating from around 1000 BCE. The grounds of the medical and pharmaceutical knowledge in the Western world were set by ancient Greeks and Romans, followed by the middle age Arab influences who brought knowledge of the West and East together (69) [109]. However, the isolation of active substances from natural products and understanding of their pharmacological activity have a more recent history, beginning in 1804 with the isolation of morphine from Opium crudum produced by cut seed pods of the poppy, Papaver somniferum [110]. The isolation, purification, pharmacological characterization, and precise dosing of many pharmaceutically active compounds from natural origin have since then followed, thereby setting up the grounds of modern medicine. Development of analytical methodologies played an important role here as the success rate of new chemical entities (NCE) discovered increased with the advances of analytical chemistry in separation and in identification techniques as well as advances in the equipment used. As mentioned in the section dedicated to the bioactivity profiling of mixtures, the BGF approach played a key role in the identification of NCEs from natural products. By 1990, more than 80% of all drugs on the market were natural products, semisynthetic analogs of natural products, or compounds synthesized based on the pharmacophores of natural products [111]. Furthermore, many naturally occurring compounds served as pharmacological tools to study different receptors and diseases. For example, five of the seven different sodium channels were characterized using venom-based compounds [104]. Besides, the names for the nicotinic and muscarinic acetylcholine receptors were derived from the plant alkaloids selectively binding to these receptors. Even though the development of combinatorial chemistry and HTS caused the profit-driven pharmaceutical industry to decrease its interest in natural products a few decades ago, they remain an important source of lead compounds. The interest in natural products as sources for NCEs is gaining interest again since the output from traditional pipelines is decreasing, especially in the last decade. Moreover, the development of analytical techniques and screening methods together with the use of new genome mining approaches enabled faster dereplication and identification of new bioactive compounds from natural products and consequently contributed to the revival of the interest in natural products as potential drugs. In the period between 1981 and 2010, 50% of NCEs approved in the category of small molecules were natural product related [112]. In 2010 only, 10 out of 20 new small molecular drugs were of natural origin [112]. The role of plants, microorganisms, and venoms as sources of active substances in drug discovery will be briefly discussed in the next paragraphs. For more extensive information on this topic, a number of reviews can be consulted [113–117].

Analytics for Bioactivity Profiling of Complex Mixtures with a Focus on Venoms

2.3 Plants as Source of Active Substances

41

Traditional medicine based on the use of plants has had a significant impact on modern medicine and drug discovery. In fact, the isolation of morphine at the very beginning of nineteenth century represents the first successful drug discovery effort in the history of medicine and pharmacy. This was followed by the isolation of a large number of other active substances from plants that were further studied and applied in the treatment of the entire diapason of diseases. The original chemical structures of many of these compounds still represent important drugs such as morphine, digoxin, digitoxin, and theophylline [118]. Morphine is used for management of severe pain [119, 120], while the cardiac glycosides digoxin and digitoxin, isolated from Digitalis species, are used as antiarrhythmic agents [121]. Theophylline, isolated from Theobroma cacao plant, is a bronchodilator used to treat patients with asthma and chronical obstructive pulmonary disease (COPD) [122]. There are also many examples of drugs and whole chemical classes of drugs that are based on the compounds found in plants. For example, another bronchodilator, tiotropium bromide that acts as the antimuscarinic agent, was developed based on atropine, an alkaloid found in Atropa belladonna [123, 124]. Papaverine, another active substance isolated from the poppy plant, served as a starting point for the development of verapamil, an L-type calcium channel-blocking agent used in the treatment of different pathologies related to the heart [118, 125]. Metformin and other bisguanidines that are used in the management of diabetes type II and metabolic syndrome were synthesized based on galegine, an active compound found in Galega officinalis [126]. Local anesthetics were developed based on the pharmacophore of cocaine, which was isolated from the leaves of coca, Erythroxylum coca [127]. Atracurium, a neuromuscular blocker used to control the muscle tonus of patients under general anesthesia, has a base in d-tubocurarine, an alkaloid found in various plants of Menispermaceae family [128]. Besides, the role of ethnomedicine is based on the use of whole plants or their parts as herbal remedies and stays important in many cultures. They are used in primary health protection in less developed countries or as adjuvant or alternative therapy in developed countries. Moreover, plant-based drug discovery may benefit from current efforts put in the analysis and characterization of bioactive compounds from plants used in traditional Chinese medicine (TCM). Studies conducted so far on screening of TCM components for new leads show promising results and compounds with anticancer, anti-inflammatory, antiviral, antimalarial, and stroke prevention activity have been reported [129–131]. Furthermore, clinical studies on the influence of TCM on the overall condition of patients with cancer show that TCM as adjuvant to chemotherapy seems to improve the quality of life of these patients [132].

42

Marija Mladic et al.

2.4 Microorganisms as Source of Active Substances

The serendipitous discovery of antibacterial properties of penicillin from a Penicillium fungi extract in 1928 by Alexander Fleming and its purification in 1940s initiated a revolution in the treatment of bacterial infections and opened an era of drugs isolated from microorganisms [133]. In fact, microorganisms are the second most important source of active substances. Penicillin, which is still used as a drug, showed a relatively wide spectrum of antibacterial action, but poor bioavailability upon oral administration due to the sensitivity to stomach acid. The introduction of different functional groups using semisynthetic approaches led to the development of other penicillins that are stable for oral administration and have an even wider spectrum of action or are targeting specific types of bacteria [134]. Interestingly, microorganisms yielded most of all known antibiotic chemical classes. Throughout the years, many more antibacterial agents from fungi and bacteria, such as tetracyclines [135], cephalosporins [134], aminoglycosides [136], and macrolides [137], were discovered and developed into drugs followed by their semisynthetic analogues. Microorganisms also served as source for the development of some hypolipidemics, anticancer drugs, immunosuppressives, anthelmintics, and antiparasitics [138]. Many of these drugs made a significant impact on the treatment of the corresponding diseases, such as statins in the treatment of dyslipidemia or cyclosporine A (CSA) in the prevention of allograft rejection. CSA was isolated in 1970 from the Tolypocladium inflatum fungi and made a turn in the therapy of allograft rejection due to the highly specific mechanism of action, serving as a base for a new generation of immunosuppressives [139]. CSA is a cyclic peptide that consists of 11 amino acids and probably the best known small peptide from natural products that made it to the clinic. The cyclic form and modifications of the side chains contribute to its relative stability and good bioavailability upon oral administration. This is an important property considering the fact that many drugs based on linear peptides require intravenous application due to the sensitivity to stomach acid. Therefore, the study of the influence of peptide cyclization and/or side-chain modifications on their specificity, bioactivity, and oral bioavailability may lead to the development of peptide drugs with improved properties allowing further development and wider application of peptidergic drugs [140]. Furthermore, microorganisms in general continue to be an important source of active substances for drug discovery considering the large number of unexplored species as well as the improvement in analytical techniques and increasing amount of genomic information. Finally, the growing concern worldwide regarding the increases in resistance of bacteria and multiresistant bacteria to existing antibiotics gave a great impulse in the last decade on both academic and pharma research in the discovery of new classes of antibacterial agents.

Analytics for Bioactivity Profiling of Complex Mixtures with a Focus on Venoms

2.5 Bioactivity/ Bioaffinity Profiling of Metabolic Mixtures in Drug Discovery

43

One of the important reasons for drug discontinuation is related to its metabolism [141]. During the selection of candidates for the preclinical studies, all potential lead compounds are screened for their metabolic stability. This is mostly conducted using LC–MS analysis of the metabolic mixture generated using various methods [142]. The metabolic stability of a drug is important because it highly influences its bioavailability and consequently determines the formulation and route of administration. For example, highly lipophilic molecules (e.g., human estrogens and androgens) are characterized by the first-pass effect, i.e., extensive liver metabolism resulting in a poor oral bioavailability. For example, before more stable molecules than estradiol (e.g., ethinylestradiol) were synthesized to allow oral administration of estrogens and improve bioavailability [143], estrogens were applied as intramuscular injections. Besides, drug metabolism may result in the formation of one or more bioactive metabolites with different selectivity, bioaffinity, and/or bioactivity. Metabolic profiling including bioactivity/bioaffinity assessment of drug metabolites and of lead compounds is therefore also very important for full pharmacological and toxicological characterization. In this scenario, bioactivity profiling of the individual metabolites of a lead compound represents an important step in drug discovery. Moreover, the bioactivity profile of metabolites of a particular drug gives valuable information about the liable “hot spots” (important functional groups in biotransformation) in the parent compound, which can in turn be used for structural improvements in the design of new drugs. In fact, there are examples where the bioactive metabolite of a known drug reached the market [144] (e.g., benzodiazepine oxazepam, being a metabolite of diazepam). Full chemical and pharmacological characterization of the metabolites is usually only performed at a later stage in drug discovery/development, since the purification of all metabolites for MS- and NMR-based structure elucidation as well as full pharmacological characterization is an elaborate and costly process. Besides, the rationale behind late full characterization of drug metabolites is that in general more relevant information is obtained in the preclinical studies after exposing an experimental animal to a drug. That way, all metabolic pathways can be assessed and generally good prediction models of drug metabolism in humans can be made. However, the discovery of an unfavorable metabolic profile connected to the formation of bioactive metabolites at this stage will most probably result in discontinuation of development of the lead compound. This further means a significant loss of time and money invested in the particular lead compound. Therefore, the pharmaceutical industry could benefit from implementation of early-stage bioactivity/bioaffinity profiling of metabolites in drug discovery to lower the risk that a lead candidate will fail to reach clinical studies and consequently cut the attrition costs. The benefit of early-stage metabolic

44

Marija Mladic et al.

and bioactivity/bioaffinity profiling is also reflected in valuable information provided to medicinal chemists, allowing them to design structural improvements of promising lead compounds. The development of advanced analytical methodologies that integrate bioassays and chemical analysis as discussed in this chapter (e.g., online and at-line approaches) is necessary to accomplish this. In this regard, the at-line nanofractionation methodology has an important advantage compared to most other approaches as it offers the possibility of direct parallel MS analysis next to most types of bioassays available [20, 53, 145]. All in all, this chapter discussed the different analytical approaches for profiling biologically active mixtures to identify their bioactive components, with the emphasis on venom toxins. References 1. Liu B, Li S, Hu J (2004) Technological advances in high-throughput screening. Am J Pharmacogenomics 4(4):263–276 2. Weller MG (2012) A unifying review of bioassay-guided fractionation, effect-directed analysis and related techniques. Sensors (Basel) 12(7):9181–9209 3. Jonker W et al (2015) Methodologies for effect-directed analysis: environmental applications, food analysis, and drug discovery, in analyzing biomolecular interactions by mass spectrometry. Wiley-VCH, Weinheim, pp 109–163 4. Chen W et al (2015) Fasxiator, a novel factor XIa inhibitor from snake venom, and its sitespecific mutagenesis to improve potency and selectivity. J Thromb Haemost 13 (2):248–261 5. Graudins A et al (2012) Cloning and activity of a novel alpha-latrotoxin from red-back spider venom. Biochem Pharmacol 83 (1):170–183 6. Feng J, Yang XW, Wang RF (2011) Bio-assay guided isolation and identification of alphaglucosidase inhibitors from the leaves of Aquilaria sinensis. Phytochemistry 72 (2–3):242–247 7. Crawford AD et al (2011) Zebrafish bioassayguided natural product discovery: isolation of angiogenesis inhibitors from East African medicinal plants. PLoS One 6(2):e14694 8. Eng Kiat Loo A, Huang D (2007) Assayguided fractionation study of α-amylase inhibitors from Garcinia mangostana pericarp. J Agri Food Chem 55(24):9805–9810 9. Su B-N et al (2002) Activity-guided fractionation of the seeds of Ziziphus jujuba using a

cyclooxygenase-2 inhibitory assay. Planta Medica 68(12):1125–1128 10. Scher JM et al (2004) Bioactivity guided isolation of antifungal compounds from the liverwort Bazzania trilobata (L.) SF Gray. Phytochemistry 65(18):2583–2588 11. Ho CC, Kumaran A, Hwang LS (2009) Bio-assay guided isolation and identification of anti-Alzheimer active compounds from the root of Angelica sinensis. Food Chem 114 (1):246–252 12. Awad R et al (2009) Bioassay-guided fractionation of lemon balm (Melissa officinalis L.) using an in vitro measure of GABA transaminase activity. Phytotherapy Res 23 (8):1075–1081 13. Wu H et al (2013) Recent developments in qualitative and quantitative analysis of phytochemical constituents and their metabolites using liquid chromatography-mass spectrometry. J Pharm Biomed Anal 72:267–291 14. Bakker RA et al (2004) Constitutively active Gq/11-coupled receptors enable signaling by co-expressed G(i/o)-coupled receptors. J Biol Chem 279(7):5152–5161 15. Ohi N et al (1986) Semisynthetic BetaLactam antibiotics. 1. Synthesis and antibacterial activity of new ureidopenicillin derivatives having catechol moieties. J Antibiotics 39(2):230–241 16. Elander RP (2003) Industrial production of beta-lactam antibiotics. Appl Microbiol Biotechnol 61(5–6):385–392 17. Kondo S, Hotta K (1999) Semisynthetic aminoglycoside antibiotics: development and enzymatic modifications. J Infect Chemother 5(1):1–9

Analytics for Bioactivity Profiling of Complex Mixtures with a Focus on Venoms 18. Manzoni M, Rollini N (2002) Biosynthesis and biotechnological production of statins by filamentous fungi and application of these cholesterol-lowering drugs. Appl Microbiol Biotechnol 58(5):555–564 19. Jonker N et al (2011) Recent developments in protein-ligand affinity mass spectrometry. Anal Bioanal Chem 399(8):2669–2681 20. Kool J et al (2011) Advances in mass spectrometry-based post-column bioaffinity profiling of mixtures. Anal Bioanal Chem 399(8):2655–2668 21. Kinawi A, Teller C (1979) Determination of drug-albumin binding in buffered bovine serum-albumin solutions applying a modified ultrafiltration process. Arzneimittel-Forschung/Drug Res 29–2(10):1495–1500 22. Zlotos G et al (1998) Determination of protein binding of gyrase inhibitors by means of continuous ultrafiltration. Pergamon-Elsevier Science, Amsterdam 23. Comess KM et al (2006) An ultraefficient affinity-based high-throughout screening process: application to bacterial cell wall biosynthesis enzyme MurF. J Biomol Screen 11 (7):743–754 24. Li HL et al (2009) Screening and structural characterization of alpha-Glucosidase inhibitors from hawthorn leaf flavonoids extract by ultrafiltration LC-DAD-MSn and SORI-CID FTICR MS. J Am Soc Mass Spectrom 20 (8):1496–1503 25. Ionita IA, Akhlaghi F (2010) Quantification of unbound prednisolone, prednisone, cortisol and cortisone in human plasma by ultrafiltration and direct injection into liquid chromatography tandem mass spectrometry. Ann Clin Biochem 47(Pt 4):350–357 26. Mulabagal V, Calderon AI (2010) Development of an ultrafiltration-liquid chromatography/mass spectrometry (UF-LC/MS) based ligand-binding assay and an LC/MS based functional assay for Mycobacterium tuberculosis shikimate kinase. Anal Chem 82 (9):3616–3621 27. vanBreemen RB et al (1997) Pulsed ultrafiltration mass spectrometry: a new method for screening combinatorial libraries. Anal Chem 69(11):2159–2164 28. Geun Shin Y, Bolton JL, van Breemen RB (2002) Screening drugs for metabolic stability using pulsed ultrafiltration mass spectrometry. Comb Chem High Throughput Screen 5 (1):59–64 29. Liu DT et al (2007) Screening for ligands of human retinoid X receptor-alpha. Using

45

ultrafiltration mass spectrometry. Anal Chem 79(24):9398–9402 30. Cao H et al (2010) Discovery of cyclooxygenase inhibitors from medicinal plants used to treat inflammation. Pharmacol Res 61 (6):519–524 31. Annis DA et al (2004) A general technique to rank protein-ligand binding affinities and determine allosteric versus direct binding site competition in compound mixtures. J Am Chem Soc 126(47):15495–15503 32. Derks RJE et al (2006) SEC-MS as an approach to isolate and directly identifying small molecular GPCR-ligands from complex mixtures without labeling. Chromatographia 64(7–8):379–385 33. Annis DA et al (2004) An affinity selectionmass spectrometry method for the identification of small molecule ligands from selfencoded combinatorial libraries—discovery of a novel antagonist of E-coli dihydrofolate reductase. Int J Mass Spectrom 238(2):77–83 34. Annis DA et al (2009) Inhibitors of the lipid phosphatase SHIP2 discovered by highthroughput affinity selection-mass spectrometry screening of combinatorial libraries. Comb Chem High Throughput Screen 12 (8):760–771 35. Whitehurst CE et al (2006) Discovery and characterization of orthosteric and allosteric muscarinic M-2 acetylcholine receptor ligands by affinity selection-mass spectrometry. J Biomol Screen 11(2):194–207 36. Whitehurst CE, Annis DA (2008) Affinity selection-mass spectrometry and its emerging application to the high throughput screening of G protein-coupled receptors. Comb Chem High Throughput Screen 11(6):427–438 37. Jonker N et al (2008) Screening of proteinligand interactions using dynamic proteinaffinity chromatography solid-phase extraction-liquid chromatography-mass spectrometry. J Chromatogr A 1205(1–2):71–77 38. Hu F, Deng C, Zhang X (2008) Development of high performance liquid chromatography with immobilized enzyme onto magnetic nanospheres for screening enzyme inhibitor. J Chromatogr B 871(1):67–71 39. Marsza MP et al (2008) Ligand and protein fishing with heat shock protein 90 coated magnetic beads. Anal Chem 80 (19):7571–7575 40. Jonker N et al (2009) Online magnetic bead dynamic protein-affinity selection coupled to LC-MS for the screening of pharmacologically active compounds. Anal Chem 81 (11):4263–4270

46

Marija Mladic et al.

41. Pochet L et al (2011) Online magnetic bead based dynamic protein affinity selection coupled to LC-MS for the screening of acetylcholine binding protein ligands. J Chromatogr B Analyt Technol Biomed Life Sci 879 (20):1781–1788 42. Ho¨fner G, Wanner KT (2015) MS binding assays. In: Analyzing biomolecular interactions by mass spectrometry. Wiley-VCH, Weinheim, pp 165–198 43. Zepperitz C, Hofner G, Wanner KT (2006) MS-binding assays: kinetic, saturation, and competitive experiments based on quantitation of bound marker as exemplified by the GABA transporter mGAT1. ChemMedChem 1(2):208–217 44. Singh NS, Jiang Z, Moaddel R (2015) Frontal and zonal affinity chromatography coupled to mass spectrometry. In: Analyzing biomolecular interactions by mass spectrometry. WileyVCH, Weinheim, pp 241–270 45. Loun B, Hage DS (1992) Characterization of thyroxine-albumin binding using highperformance affinity chromatography. I. Interactions at the warfarin and indole sites of albumin. J Chromatogr 579 (2):225–235 46. Chan NW et al (2003) Frontal affinity chromatography-mass spectrometry assay technology for multiple stages of drug discovery: applications of a chromatographic biosensor. Anal Biochem 319(1):1–12 47. Moaddel R et al (2005) Enantioselective binding to the human organic cation transporter-1 (hOCT1) determined using an immobilized hOCT1 liquid chromatographic stationary phase. Chirality 17(8):501–506 48. Calleri E et al (2010) Frontal affinity chromatography-mass spectrometry useful for characterization of new ligands for GPR17 receptor. J Med Chem 53 (9):3489–3501 49. Pharmaceuticals N. Tegretol (carbamazepine) extended-release tablets prescribing information. In N. Pharmaceuticals (ed). 2003 50. Haselberg R et al (2014) Capillary electrophoresis-based assessment of nanobody affinity and purity. Anal Chim Acta 818:1–6 51. Medina-Casanellas S et al (2012) Preparation and evaluation of an immunoaffinity sorbent for the analysis of opioid peptides by on-line immunoaffinity solid-phase extraction capillary electrophoresis-mass spectrometry. Anal Chim Acta 717:134–142 52. Haselberg R, Somsen GW (2015) Online affinity assessment and immunoaffinity sample

pretreatment in capillary electrophoresis–mass spectrometry, in analyzing biomolecular interactions by mass spectrometry. WileyVCH, Weinheim, pp 271–298 53. Nijmeijer S et al (2012) Development of a profiling strategy for metabolic mixtures by combining chromatography and mass spectrometry with cell-based GPCR signaling. J Biomol Screen 17(10):1329–1338 54. Giera M et al (2009) Microfractionation revisited: a 1536 well high resolution screening assay. Anal Chem 81(13):5460–5466 55. Falck D et al (2013) Development of on-line liquid chromatography-biochemical detection for soluble epoxide hydrolase inhibitors in mixtures. Chromatographia 76 (1–2):13–21 56. Oosterkamp AJ et al (1997) Theoretical concepts of on-line liquid chromatographic- biochemical detection systems II. Detection systems based on labelled affinity proteins. J Chromatogr A 787(1–2):37–46 57. Oosterkamp AJ et al (1997) Theoretical concepts of on-line liquid chromatographicbiochemical detection systems. I. Detection systems based on labelled ligands. J Chromatogr A 787(1–2):27–35 58. Marques LA et al (2010) Production and on-line acetylcholinesterase bioactivity profiling of chemical and biological degradation products of tacrine. J Pharm Biomed Anal 53(3):609–616 59. Kool J et al (2007) Cytochrome P450 bio-affinity detection coupled to gradient HPLC: on-line screening of affinities to cytochrome P4501A2 and 2D6. J Chromatogr B 858(1):49–58 60. Falck D et al (2010) Development of an online p38alpha mitogen-activated protein kinase binding assay and integration of LC-HR-MS. Anal Bioanal Chem 398 (4):1771–1780 61. Kool J et al (2010) Online fluorescence enhancement assay for the acetylcholine binding protein with parallel mass spectrometric identification. J Med Chem 53 (12):4720–4730 62. de Vlieger JS et al (2010) Determination and identification of estrogenic compounds generated with biosynthetic enzymes using hyphenated screening assays, high resolution mass spectrometry and off-line NMR. J Chromatogr B Analyt Technol Biomed Life Sci 878 (7–8):667–674 63. Schenk T et al (2003) A generic assay for phosphate-consuming or-releasing enzymes coupled on-line to liquid chromatography

Analytics for Bioactivity Profiling of Complex Mixtures with a Focus on Venoms for lead finding in natural products. Anal Biochem 316(1):118–126 64. Hogenboom A et al (2001) Continuous-flow, on-line monitoring of biospecific interactions using electrospray mass spectrometry. Anal Chem 73(16):3816–3823 65. Heus F et al (2010) Development of a microfluidic confocal fluorescence detection system for the hyphenation of nano-LC to on-line biochemical assays. Anal Bioanal Chem 398 (7–8):3023–3032 66. Otvos RA et al (2013) Analytical workflow for rapid screening and purification of bioactives from venom proteomes. Toxicon 76:270–281 67. Heus F et al (2014) Miniaturized bioaffinity assessment coupled to mass spectrometry for guided purification of bioactives from toad and cone snail. Biology 3(1):139–156 68. Heus F et al (2013) An efficient analytical platform for on-line microfluidic profiling of neuroactive snake venoms towards nicotinic receptor affinity. Toxicon 61:112–124 69. Giera M et al (2010) Structural elucidation of biologically active neomycin N-octyl derivatives in a regioisomeric mixture by means of liquid chromatography/ion trap time-offlight mass spectrometry. Rapid Commun Mass Spectrom 24(10):1439–1446 70. Kool J et al (2012) High-resolution metabolic profiling towards G protein-coupled receptors: rapid and comprehensive screening of histamine H(4) receptor ligands. J Chromatogr A 1259:213–220 71. Mladic M et al (2016) At-line nanofractionation with parallel mass spectrometry and bioactivity assessment for the rapid screening of thrombin and factor Xa inhibitors in snake venoms. Toxicon 110:79–89 72. Mladic M et al (2017) Rapid screening and identification of ACE inhibitors in snake venoms using at-line nanofractionation LC-MS. Anal Bioanal Chem 409 (25):5987–5997 73. Still KS et al (2017) Multipurpose HTS coagulation analysis: assay development and assessment of coagulopathic snake venoms. Toxins (Basel) 9(12):382 74. Bordon KC et al (2012) Isolation, enzymatic characterization and antiedematogenic activity of the first reported rattlesnake hyaluronidase from Crotalus durissus terrificus venom. Biochimie 94(12):2740–2748 75. Wiezel GA et al (2015) Identification of hyaluronidase and phospholipase B in Lachesis muta rhombeata venom. Toxicon 107 (Pt B):359–368

47

76. Babaie M et al (2013) Isolation and partial purification of anticoagulant fractions from the venom of the Iranian snake Echis carinatus. Acta Biochim Pol 60(1):17–20 77. Menaldo DL et al (2015) Purification procedure for the isolation of a P-I metalloprotease and an acidic phospholipase A2 from Bothrops atrox snake venom. J Venom Anim Toxins Incl Trop Dis 21:28 78. Osipov AV et al (2017) New paradoxical three-finger toxin from the cobra Naja kaouthia venom: Isolation and characterization. Dokl Biochem Biophys 475(1):264–266 79. Teixeira TL et al (2016) Isolation, characterization and screening of the in vitro cytotoxic activity of a novel L-amino acid oxidase (LAAOcdt) from Crotalus durissus terrificus venom on human cancer cell lines. Toxicon 119:203–217 80. Rodriguez-Acosta A et al (2016) Biological and biochemical characterization of venom from the broad-banded copperhead (Agkistrodon contortrix laticinctus): isolation of two new dimeric disintegrins. Anim Biol Leiden Neth 66(2):173–187 81. Fucase TM et al (2017) Isolation and biochemical characterization of bradykininpotentiating peptides from Bitis gabonica rhinoceros. J Venom Anim Toxins Incl Trop Dis 23:33 82. King GF (2011) Venoms as a platform for human drugs: translating toxins into therapeutics. Expert Opin Biol Ther 11 (11):1469–1484 83. Ferreira SH, Bartelt DC, Greene LJ (1970) Isolation of bradykinin-potentiating peptides from Bothrops jararaca venom. Biochemistry 9(13):2583–2593 84. Ondetti MA et al (1971) Angiotensinconverting enzyme inhibitors from the venom of Bothrops jararaca. Isolation, elucidation of structure, and synthesis. Biochemistry 10(22):4033–4039 85. Smith CG, Vane JR (2003) The discovery of captopril. FASEB J 17(8):788–789 86. Scarborough RM (1999) Development of eptifibatide. Am Heart J 138(6 Pt 1):1093–1104 87. Cook JJ et al (1999) Tirofiban (Aggrastat (R)). Cardiovasc Drug Rev 17(3):199–224 88. Earl ST et al (2012) Drug development from Australian elapid snake venoms and the Venomics pipeline of candidates for haemostasis: Textilinin-1 (Q8008), Haempatch (Q8009) and CoVase (V0801). Toxicon 59 (4):456–463

48

Marija Mladic et al.

89. Koh CY, Kini RM (2012) From snake venom toxins to therapeutics–cardiovascular examples. Toxicon 59(4):497–506 90. Vink S et al (2012) Natriuretic peptide drug leads from snake venom. Toxicon 59 (4):434–445 91. Diochot S et al (2012) Black mamba venom peptides target acid-sensing ion channels to abolish pain. Nature 490(7421):552–555 92. Pu XC, Wong PT, Gopalakrishnakone P (1995) A novel analgesic toxin (hannalgesin) from the venom of king cobra (Ophiophagus hannah). Toxicon 33(11):1425–1431 93. McCleary RJ, Kini RM (2013) Non-enzymatic proteins from snake venoms: a gold mine of pharmacological tools and drug leads. Toxicon 62:56–74 94. Kini RM, Doley R (2010) Structure, function and evolution of three-finger toxins: mini proteins with multiple targets. Toxicon 56 (6):855–867 95. Lewis RJ et al (2012) Conus venom peptide pharmacology. Pharmacol Rev 64 (2):259–298 96. Twede VD et al (2009) Neuroprotective and cardioprotective conopeptides: an emerging class of drug leads. Curr Opin Drug Discov Devel 12(2):231–239 97. Essack M, Bajic VB, Archer JA (2012) Conotoxins that confer therapeutic possibilities. Mar Drugs 10(6):1244–1265 98. Vetter I, Lewis RJ (2012) Therapeutic potential of cone snail venom peptides (conopeptides). Curr Top Med Chem 12 (14):1546–1552 99. Miljanich GP (2004) Ziconotide: neuronal calcium channel blocker for treating severe chronic pain. Curr Med Chem 11 (23):3029–3040 100. Sto¨cklin R. XEP-018: a new myorelaxant peptide lead compound from the venom of the cone snail Conus consors. Proceedings of 7th Annual Peptide Therapeutics Symposium, 2012. p. 32 101. Furman BL (2012) The development of Byetta (exenatide) from the venom of the Gila monster as an anti-diabetic agent. Toxicon 59(4):464–471 102. Turton M et al (1996) A role for glucagonlike peptide-1 in the central regulation of feeding. Nature 379(6560):69–72 103. Doyle ME, Egan JM (2000) Glucagon-like peptide-1. Recent Prog Horm Res 56:377–399 104. Klint JK et al (2012) Spider-venom peptides that target voltage-gated sodium channels:

pharmacological tools and potential therapeutic leads. Toxicon 60(4):478–491 105. Mortari MR, Cunha AOS (2013) New perspectives in drug discovery using neuroactive molecules from the venom of arthropods. IntechOpen, London 106. Yang S et al (2013) Discovery of a selective NaV1. 7 inhibitor from centipede venom with analgesic efficacy exceeding morphine in rodent pain models. Proc Natl Acad Sci 110(43):17534–17539 107. Lewis RJ, Garcia ML (2003) Therapeutic potential of venom peptides. Nat Rev Drug Discov 2(10):790–802 108. Mathias NR, Hussain MA (2010) Non-invasive systemic drug delivery: developability considerations for alternate routes of administration. J Pharm Sci 99(1):1–20 109. Anderson S (2005) Making medicines: a brief history of pharmacy and pharmaceuticals. Pharmaceutical Press, London 110. Hamilton GR, Baskett TF (2000) In the arms of Morpheus the development of morphine for postoperative pain relief. Can J Anaesth 47 (4):367–374 111. Sneader W (1996) Drug prototypes and their exploitation. Wiley, Chichester 112. Newman DJ, Cragg GM (2012) Natural products as sources of new drugs over the 30 years from 1981 to 2010. J Nat Prod 75 (3):311–335 113. Koehn FE, Carter GT (2005) The evolving role of natural products in drug discovery. Nat Rev Drug Discov 4(3):206–220 114. Harvey AL (2008) Natural products in drug discovery. Drug Discov Today 13 (19–20):894–901 115. Harvey AL (2014) Toxins and drug discovery. Toxicon 92:193–200 116. Cragg GM, Newman DJ (2013) Natural products: a continuing source of novel drug leads. Biochim Biophys Acta 1830 (6):3670–3695 117. Chin YW et al (2006) Drug discovery from natural sources. AAPS J 8(2):E239–E253 118. Fabricant DS, Farnsworth NR (2001) The value of plants used in traditional medicine for drug discovery. Environ Health Perspect 109(Suppl 1):69–75 119. Hanks GW et al (2001) Morphine and alternative opioids in cancer pain: the EAPC recommendations. Br J Cancer 84 (5):587–593 120. Furlan AD et al (2006) Opioids for chronic noncancer pain: a meta-analysis of

Analytics for Bioactivity Profiling of Complex Mixtures with a Focus on Venoms effectiveness and side effects. CMAJ 174 (11):1589–1594 121. Campbell TJ, Williams KM (2001) Therapeutic drug monitoring: antiarrhythmic drugs. Br J Clin Pharmacol 52:21s–34s 122. Barnes PJ (2006) Theophylline for COPD. Thorax 61(9):742–744 123. Barnes PJ (2001) Tiotropium bromide. Expert Opin Investig Drugs 10(4):733–740 124. Hansel TT, Barnes PJ (2002) Tiotropium bromide: a novel once-daily anticholinergic bronchodilator for the treatment of COPD. Drugs Today (Barc) 38(9):585–600 125. Elliott WJ, Ram CV (2011) Calcium channel blockers. J Clin Hypertens (Greenwich) 13 (9):687–689 126. Bailey CJ, Day C (2004) Metformin: its botanical background. Pract Diab Int 21 (3):115–117 127. Ruetsch YA, Boni T, Borgeat A (2001) From cocaine to ropivacaine: the history of local anesthetic drugs. Curr Top Med Chem 1 (3):175–182 128. Raghavendra T (2002) Neuromuscular blocking drugs: discovery and development. J R Soc Med 95(7):363–367 129. Graziose R, Lila MA, Raskin I (2010) Merging traditional Chinese medicine with modern drug discovery technologies to find novel drugs and functional foods. Curr Drug Discov Technol 7(1):2 130. Gong X, Sucher NJ (1999) Stroke therapy in traditional Chinese medicine (TCM): prospects for drug discovery and development. Trends Pharmacol Sci 20(5):191–196 131. Hsiao W, Liu L (2010) The role of traditional Chinese herbal medicines in cancer therapy— from TCM theory to mechanistic insights. Planta Medica 76(11):1118 132. Lao L, Xu L, Xu S (2012) Traditional Chinese medicine. In: Integrative pediatric oncology. Springer, Beijing, pp 125–135 133. Scriabine A (1999) Discovery and development of major drugs currently in use. In:

49

Pharmaceutical innovation: revolutionizing human health. Chemical Heritage Press, Philadelphia, pp 148–270 134. Page MG (2012) Beta-lactam antibiotics. In: Antibiotic discovery and development. Springer, New York, pp 79–117 135. Chopra I, Roberts M (2001) Tetracycline antibiotics: mode of action, applications, molecular biology, and epidemiology of bacterial resistance. Microbiol Mol Biol Rev 65 (2):232–260; second page, table of contents 136. Forge A, Schacht J (2000) Aminoglycoside antibiotics. Audiol Neurootol 5(1):3–22 137. Omura S (2002) Macrolide antibiotics: chemistry, biology, and practice. Academic Press, Orlando 138. Berdy J (2005) Bioactive microbial metabolites. J Antibiot (Tokyo) 58(1):1–26 139. Colombo D, Ammirati E (2011) Cyclosporine in transplantation—a history of converging timelines. J Biol Regul Homeost Agents 25(4):493–504 140. Craik DJ et al (2013) The future of peptidebased drugs. Chem Biol Drug Des 81 (1):136–147 141. Kola I, Landis J (2004) Can the pharmaceutical industry reduce attrition rates? Nat Rev Drug Discov 3(8):711–716 142. Cusack KP et al (2013) Emerging technologies for metabolite generation and structural diversification. Bioorg Med Chem Lett 23 (20):5471–5483 143. Dhont M (2010) History of oral contraception. Eur J Contracept Reprod Health Care 15(S2):S12–S18 144. Lin JH, Lu AY (1997) Role of pharmacokinetics and metabolism in drug discovery and development. Pharmacol Rev 49(4):403–449 145. Mladic M et al (2015) At-line coupling of LC-MS to bioaffinity and selectivity assessment for metabolic profiling of ligands towards chemokine receptors CXCR1 and CXCR2. J Chromatogr B Analyt Technol Biomed Life Sci 1002:42–53

Part II Toxins Purification and Production

Chapter 3 Venom Collection from Spiders and Snakes: Voluntary and Involuntary Extractions (“Milking”) and Venom Gland Extractions William K. Hayes, Gerad A. Fox, and David R. Nelsen Abstract Venom collection (often called “milking”) provides the toxic secretions essential for studying animal venoms and/or generating venom products. Methods of venom collection vary widely, falling into three broad categories: voluntary venom extraction (inducing the animal to willingly release its venom), involuntary venom extraction (glandular massage, electrical stimulation, or administration of induction chemicals to promote venom expulsion), and venom gland extraction (surgical aspiration or trituration of homogenized gland tissue). Choice of method requires consideration of animal species, animal welfare, human safety (avoiding envenomation), venom yield and composition desired, and level of toxin purity required. Here, we summarize the materials and methods used to obtain venom by each of these approaches from spiders and snakes. Key words Arachnida, Serpentes, Venom composition, Venom expression, Venom milking, Venom yield, Fangs, Anesthesia

1

Introduction Venom extraction (or “milking”) usually comprises the initial task in studying animal venoms and/or generating venom products. Although venom composition and evolution can be inferred from transcriptomes derived from venom gland tissue, researchers often require the venom itself for quantifying yields, assessing composition, evaluating function, fractionating components, and supplying other needs. Spiders and snakes synthesize and store their venom within paired glands. In spiders, these glands reside in the basal segment (paturon) of the chelicerae, or may extend more posteriorly into the carapace. In snakes, they lie in the upper jaw of the head, but can extend posteriorly beyond the head along the sides of the body of some atractaspidid snakes [1]. Spiders expel their venom gland contents through a pair of hollow cheliceral fangs [1]. Front-fanged

Avi Priel (ed.), Snake and Spider Toxins: Methods and Protocols, Methods in Molecular Biology, vol. 2068, https://doi.org/10.1007/978-1-4939-9845-6_3, © Springer Science+Business Media, LLC, part of Springer Nature 2020

53

54

William K. Hayes et al.

snakes similarly extrude venom through a pair of hollow maxillary fangs, whereas colubrid snakes use normal or enlarged (sometimes numerous) rear maxillary teeth that may be grooved or ungrooved, but never hollow [2]. Two important properties of venom synthesis and storage can influence the composition of venom. First, asynchronous venom component regeneration can lead to varying venom composition in the days or weeks following venom usage [3–5]. Second, venom component heterogeneity can result in venom composition changes during venom emergence from the gland [6–8], presumably because different regions of the venom gland secrete and store different components. Venom component heterogeneity appears to be trivial in at least the front-fanged snakes [1], which possess a large lumen within their venom glands that minimizes regional secretion storage, but it may be important in spiders [9–11]. Methods of venom collection vary widely, falling into three broad categories: voluntary expression, involuntary expression, and gland extraction [9, 12, 13]. Choice of method requires consideration of animal species, animal welfare, human safety (avoiding envenomation), venom yield and composition desired, and level of toxin purity required. Different collection methods can provide varying quantities and composition of secretion, which may be important depending on the purpose of venom collection. Most investigators desire venom extracted from replete venom glands, which requires a period of venom disuse (generally 1 week or more) to ensure maximal yields and adequate representation of all venom components. Investigators may also attempt to extract the full venom gland contents (to the extent possible), which ensures presence of all venom components but may be injurious to the animal and requires a lengthier regeneration period between successive venom collections. Voluntary methods of venom extraction depend on the animal’s willingness to expulse venom. As such, they comprise the best means of quantifying the amount of venom potentially expended in “normal” defensive contexts [14, 15]. The investigator provokes the animal to express venom into a venom collection device (e.g., beaker, tube, or vial), usually by inducing it to bite through a membrane covering. In some snakes and spiders, the venom can be collected directly from the fangs. These methods are unlikely to fully empty the glands. Involuntary methods of venom extraction take partial or complete control of venom expulsion away from the animal. In doing so, the venom glands can be more completely emptied. Venom expression can be manipulated via electrical stimulation of the muscles surrounding the glands (more often in spiders), glandular massage (snakes), or administration of induction chemicals (colubrid and small elapid snakes) that enhance glandular secretion [16]. Anesthesia is often applied to minimize trauma to the animal.

Venom Collection from Spiders and Snakes

55

Some evidence suggests that gland massage and electrical stimulation could injure the glands or even cause death (e.g. [9, 17, 18]; nevertheless, these procedures are widely employed. Gland extractions may be necessary for spiders and small snakes, but these are invasive and require anesthesia and (almost always) euthanasia. Gland extractions may also be useful when it is easier to collect and store frozen animals until dissection is convenient or when the skills and equipment required for milking are not available [9]. Venom can be extracted from a surgically aspirated or excised gland by either directly capturing the glandular products or trituration of the homogenized tissue to extract the toxic components [19, 20]. Contamination of venom by digestive fluids and/or hemolymph (spiders), or by salivary secretions and/or blood (snakes [21]), can be problematic, especially for involuntary methods and gland extractions. Precautions should be undertaken to minimize contamination, venom samples are often centrifuged to remove cellular debris, and the samples can be evaluated for purity. Venom samples may also be subject to autodegradation. Researchers very seldom do so, but an enzyme inhibitor could be added to the venom sample to minimize proteolytic degradation. This might be helpful for study of venom composition, but is contraindicated for study of enzymatic activity. Some venoms (e.g., those from viperid snakes) are remarkably stable due to endogenous mechanisms associated with storage in the gland, and can withstand days of wet storage at moderate environmental temperatures (20–37  C), which minimizes concerns about degradation and sample treatment [22]. In this chapter, we summarize the materials and methods typically used for procuring venom from spiders and snakes. Because of profound variation in size and temperament of different species and individual animals, the best approach to venom extraction will vary substantially, and different investigators will often use different approaches even with the same species. The best results are usually achieved through trial and error. The protocols presented here are based on those of prior authors and generally adapted to our own experience. We summarize methods for voluntary, involuntary, and gland extractions for spiders first, followed by snakes.

2

Materials

2.1 Voluntary Venom Extraction of Spiders: Spontaneous Expulsion

1. Venom collection device: Small beaker, tube, or vial (e.g., 1.5 μL microfuge tube). The device can be held by hand or secured to a clamp stand. A micropipette (e.g., 5 μL disposable PCR micropipette) or capillary tube can be used for direct collection from the fangs. The capillary tube can be modified to make it smaller if necessary (see Note 1). A Kimwipe® can

56

William K. Hayes et al.

even be used to wipe venom from fangs, with the venom subsequently rinsed from the Kimwipe®. 2. Membrane: Optional, to cover venom collection device. Parafilm® is often used, which can be secured by stretch across opening of device; if fang penetration is desired, membrane must be thinner than that typically used for snake venom extractions due to the comparatively small, fragile fangs of spiders. 3. Manipulation and restraint device(s): Tweezers, forceps, clamps, and/or gloves. Tweezers and/or forceps are usually used for safe handling, though some spiders can be restrained by gloved hand. Clamps can be secured to a stand. 4. Venom collection equipment: Pipette with tips, micropipette, modified capillary tube, and/or microdispenser; these are essential for collecting and transferring small venom quantities. Scissors may be necessary to collect small portions of venom from the membrane, especially when the venom accumulates on the underside of the membrane. 5. Buffer solution or ultrapure water: Optional, to dilute and retrieve small quantities of venom and/or rapidly desiccating venom, which can occlude the micropipette or capillary tube tip (see Note 2). 6. Venom storage tubes: Usually microfuge tubes in which samples can be stored indefinitely in a freezer. Can be pre-labeled for efficiency. 7. Ice: To keep venom samples chilled during collection and transfer to a freezer. 8. Dissection microscope: Optional, necessary for smaller spiders. 9. Calipers: Optional, to record the spider’s length and/or other measurements if morphological data are required. Venom yields increase, generally exponentially, with growth in body size, and venom composition often changes as well. 10. Linen test magnifier: Optional, to better view and record length of venom column in micropipette or capillary tube, which can be used to measure venom yield (see Note 3). 2.2 Involuntary Venom Extraction of Spiders: Electrical Stimulation

1. Assemble the same items listed in Subheading 2.1. 2. Source of electric pulse: Stimulator or circuit board to generate electrical impulse, electrodes and/or modified tweezers or forceps, and optional accessories such as foot switch (see Note 4). A foot-operated switch may be the best means of controlling delivery of the electrical impulse. 3. Vacuum aspiration assembly: Optional, for removal of regurgitate to avoid venom contamination. The assembly should

Venom Collection from Spiders and Snakes

57

include a small vacuum pump (aquarium pump works well), 22-gauge needle (or other gauge tailored to the size of spider), two lengths of surgical tubing, Erlenmeyer flask of sufficient size to remain stable on table, two-hole rubber stopper, two glass tubes (or pipette tips with ends cut off), and a clamp stand. Connect each length of surgical tubing to a glass tube, and then insert glass tubes into each opening of the two-hole rubber stopper. Place stopper into Erlenmeyer flask. Connect one length of surgical tubing to the aquarium vacuum, and the other length of surgical tubing to the 22-gauge needle. The needle should be secured to the clamp stand. 4. Anesthesia assembly: Source of CO2, Erlenmeyer flask of sufficient size to accommodate spider, two-hole rubber stopper, solid rubber stopper, one length of surgical tubing, and 1 mL pipette tip. Connect length of surgical tubing to 1 mL pipette tip that had its tip cut off. Push the tapered end of the pipette tip into one of the holes in the rubber stopper. Connect the other end of the surgical tubing to the CO2 source. The second hole of the rubber stopper is left open. 2.3 Venom Gland Extraction of Spiders

1. Dissection arena: Glass petri dish works well; bottom can be coated in paraffin wax (see Note 5). A second petri dish or the cutoff cap of a 1.5 mL microfuge tube can be used for dissecting isolated glands. 2. Surgical implements: Extra fine-tipped tweezers or forceps, 00 insect needles, microsurgery dissection knife. 3. Venom storage tubes. 4. Chilled saline solution (e.g., 223 mM NaCl, 6.8 mM KCl, 8 mM CaCl2, 5.1 mM MgCl2, 10 mM 4-(2-hydroxyethyl)-1piperazineethanesulfonic acid (HEPES), at pH 7.8 [23]). 5. Ice and/or ice pack. 6. Dissecting scope.

2.4 Voluntary Venom Extraction of Snakes: Spontaneous Expulsion

1. Venom collection device: Beaker, tube, or vial. A pipette tip or capillary tube can be used for direct collection from the fangs, especially for small and short-fanged snakes. Some investigators secure the collection device to a clamp stand, but others hold the device by hand. 2. Membrane: Optional, to cover the venom collection device. Parafilm® is often used, which can be secured by stretch, but various plastic or rubber membranes can also suffice, which are generally secured by rubber band or tape. 3. Snake handling and immobilization tools: Can include snake stick, snake tongs, snake pinning tool, transparent cylinder (roughly one-half to two-thirds of snake length, and diameter

58

William K. Hayes et al.

slightly larger than head width), very thick gloves (e.g., Venom Defender by HexArmor®, Grand Rapids, Michigan, USA, though no glove should be considered fully fang-proof), thick folded towel, and/or large rubber-tipped forceps to manipulate the snake and its head while providing protection from envenomation. 4. Pipettor and tips: Often essential for transferring small venom quantities. 5. Storage tubes: Usually microfuge tubes in which samples can be stored indefinitely in a freezer. 6. Ice: Optional, to keep venom samples chilled between collection and transfer to a freezer. 7. Tape measure and/or calipers: Optional, to record the snake’s length and/or other measurements if morphological data are required. Venom yields increase, generally exponentially, with growth in body size, and venom composition often changes as well. 8. Aspiration device: Optional, for venom collection from small fangs or teeth, such as those of colubrids and small elapids; consider a capillary tube attached to a micropipette (see also Ref. 24). 9. Purified water, physiological saline, or carbonate water: Optional to more efficiently recover small venom quantities from collection device. 2.5 Involuntary Venom Extraction of Snakes: Manual Expulsion

1. Assemble the same items listed in Subheading 2.4.

2.6 Involuntary Venom Extraction of Snakes: Electrical Stimulation

1. Assemble the same items listed in Subheadings 2.4 and 2.5 (anesthetic is essential).

2.7 Involuntary Venom Extraction of Snakes: Pilocarpine Stimulation

1. Assemble the same items listed in Subheadings 2.4 and 2.5 (anesthetic is generally necessary).

2. Anesthetic: Optional, to reduce the risks associated with manual or electrical expression of the venom glands; typically, an inhalant such as isoflurane, sevoflurane, halothane, or methoxyflurane, but injectables such as ketamine or propofol may also be used.

2. Source of electric pulse: Stimulator (e.g., peripheral nerve stimulator) and electrodes (e.g., alligator clip).

2. Syringe and needles: To inject induction chemical. 3. Pilocarpine hydrochloride: A parasympathomimetic agent that stimulates venom gland secretion.

Venom Collection from Spiders and Snakes

2.8 Venom Gland Extraction of Snakes

59

1. Assemble items 4–9 of Subheading 2.4. 2. Dissection arena (a table top will suffice): Isolated glands can be dissected in one or several petri dishes. 3. Surgical implements: Surgical knives, scissors, tweezers, and/or forceps. 4. Suture material: To ligate the venom glands. 5. Absorbent material: Cotton, gauze, and/or paper toweling. 6. Chilled saline solution. 7. Ice and/or ice pack. 8. Dissecting scope (more essential for small snakes).

3

Methods

3.1 Voluntary Venom Extraction of Spiders: Spontaneous Expulsion

This method depends on the spider’s natural tendency to bite and expulse venom defensively. Some spiders refuse to cooperate. 1. Set up venom collection device (small beaker, tube, vial (e.g., 1.5 μL microfuge tube)), covering it with a membrane and/or securing it to a clamp stand if desired. The device can also be held by hand. 2. Manipulate spider to induce venom expulsion. This usually involves positioning the spider so that its chelicerae come into contact with the membrane, but some spiders will expulse venom with physical prodding of the fangs. A dissecting microscope will be essential to guide the process for very small spiders (see Note 6). Biting, resultant puncture marks in the membrane, and venom droplets are generally easy to see, especially with the aid of a microscope (see Note 7). 3. Place spider back into its container. 4. Examine outer surface of membrane, if used, and remove any debris (see Note 8). 5. Retrieve venom using micropipette or capillary tube and place into storage tube on ice. For very small samples, the portion of membrane containing venom may be cut and placed into preferred buffer solution or ultrapure water, and allowed to soak and/or be rinsed several times before removing membrane from solution. 6. Place venom sample into freezer for short-term storage. Venom sample may be centrifuged and supernatant removed to further reduce potential contamination, though this is impractical for very small samples. Samples should be lyophilized before freezing for long-term storage.

60

William K. Hayes et al.

Fig. 1 Involuntary venom extraction (milking) of spiders via electrical stimulation. (a) Setup, including light source (LS); dissecting microscope (DM); stimulator (S); vacuum pump (VP); Erlenmeyer flask (EF); two-hole rubber stopper (RS); electrified aspiration needle (EN) secured by clamp to a clamp stand (CS) and attached to stimulator via wire lead and to vacuum system to aspirate oral secretions; two vacuum tubes (VT), connecting VP to EF, and EF to EN; electrified tweezers (ET) attached to stimulator via wire lead; and capillary tube (CT) to collect venom from spider fangs. (b) Venom collection from small spider (Latrodectus hesperus), with fangs raised against electrified aspiration needle, and opening of needle pressed against spider’s mouth; electrified tweezers, with insulated side across ventral surface of cephalothorax, and saline-soaked electrified side across dorsal surface, close to chelicerae; and micropipette to collect venom from fangs. (c) Venom collection from large spider (Aphonopelma eutylenum), with electrified spatula #1 beneath spider fangs; electrified spatula #2 pressed against mouth; and capillary tube to collect venom from fangs. Electrical milking is the most frequently used approach to venom collection from spiders. Photos: David R. Nelsen and William K. Hayes 3.2 Involuntary Venom Extraction of Spiders: Electrical Stimulation

Although the setup can be rather complex (Fig. 1a), researchers rely mainly on this method for venom collection from spiders. Because spider body size and morphology span a wide range, exact method used will have to be modified to accommodate the variation. Spiders typically survive milking and can be milked again in the future, but long-term effects of repeated milking remain largely unstudied

Venom Collection from Spiders and Snakes

61

[9, 17, 18]. Some gland types may be impacted more than others, affecting both the volume and composition of regenerated venom [9]. Special attention must be given to avoid venom sample contamination from stomach regurgitation. Protocols have been described by others (e.g., [9, 25–28]). 1. Assemble the electrical stimulator. Attach electrodes, modified tweezers or forceps, and/or other peripherals (e.g., foot switch, power supply) to stimulator or circuit board. 2. Adjust current phase, current pulse, and/or voltage to desired settings. Optimal settings will need adjustment depending on spider species and size, electrode placement, and conduction. Investigators report a wide range of stimulation parameters, including either alternating or direct current, and voltages from 3 to 50 V (generally corresponding to the size of spider). Many authors fail to indicate amperage, frequency if alternating current, and duration and number of stimulations delivered. Excessive electrical charge can injure or kill the spider, so begin with lower levels of stimulation and increase until optimal delivery is achieved. 3. Prepare saturated saline solution (e.g., 15% solution [27]). The solution is used to wet the absorbent side of the tweezers or forceps to improve conductance of the electric pulse. 4. Secure tweezers or forceps with an insulated two-pronged clamp attached to base. Position tweezers or forceps under a dissection microscope and adjust the height of clamp-tweezer/ forceps assembly as needed on base. 5. Anesthetize spider. Place spider into Erlenmeyer flask and use a two-hole rubber stopper to close the opening. Connect a length of surgical tubing to a 1 mL pipette tip that has had its tip cut off. Push the tapered end of the pipette tip into one of the holes in the rubber stopper. Connect the other end of the surgical tubing to the CO2 source. Allow CO2 to flow slowly into flask until air inside flask has been replaced with CO2 gas. Replace two-hole rubber stopper with solid rubber stopper and allow spider to remain in flask for 5–10 min or more, depending on the species and spider size. If spider is left in CO2 gas for an insufficient time, it may resume activity during the milking. 6. While spider is being anesthetized, saturate absorbent side of tweezers or forceps with saline solution and turn on aquarium pump. 7. Remove spider from anesthesia flask and place spider ventral side up between clamped tweezers or forceps. The insulated side of tweezers or forceps should lie across the ventral surface of the spider and the saline-soaked side across the dorsal surface of the cephalothorax, close to the chelicera (Fig. 1b; see Note 9).

62

William K. Hayes et al.

8. Wash spider chelicerae and mouth with ultrapure water, and use suction needle to remove this water. Next, place suction needle across oral opening, between the oral opening and chelicera, to impede venom contamination from regurgitation triggered by electric shock. Then position pipette tips or modified capillary tubes near or onto the fang tips (see Note 10). 9. Apply the electric shock; this can be most readily controlled with a foot switch. Collect venom into pipette tip or modified capillary tube. Several rounds of shock delivery may be necessary. For larger spiders, two spatulas can be used to deliver the shock in place of tweezers or forceps, and without the aspiration needle (Fig. 1c). 10. Use pipette or microdispenser to expel collected venom into storage tube. Treat and store the venom sample as described in Subheading 3.1. 3.3 Venom Gland Extraction of Spiders

This approach, which kills the spider, removes the full venom glands to procure the venom itself. Equipment and technique will depend on size and morphology of the spider. Smaller spiders may require pinning to a wax-bottomed dissection plate (see Note 5), whereas larger spiders can be manipulated more readily by fingers and implements. Dissection should be conducted with ample saline so that osmotic shock does not rupture the venom glands or induce diffusion of venom out of the gland. Small spiders should be dissected while immersed entirely in chilled saline. 1. Anesthetize spider, as described in Subheading 3.2. 2. Secure spider by gloved hand, tweezers, and/or forceps, or by inserting one or two 00 insect pins through the abdomen and into the wax-bottomed dissection plate (see Note 11). 3. If the spider is small, completely immerse it in chilled saline. 4. Excise the glands. Orthognath spiders (e.g., tarantulas) possess glands that lie entirely within the paturons (bases) of the chelicerae. Grasp the paturons with pliers or strong forceps and tear them free, cut away the medial face with scissors, and then remove the gland from the anterior-lateral angle of the paturon [9]. Labidognath spiders have glands that extend posteriorly beyond the paturons and into the carapace. Grasp laterally across their bases with fine-tipped tweezers and gently rotate forward until the pleural membrane along the posterior edge of the chelicerae ruptures, pulling the glands free from the carapace. A knife can be used if necessary to cut the connective tissue between the chelicerae and carapace to help free the glands (see Note 12). 5. Place the isolated glands in a small glass dish on ice, and add chilled saline. Small glands can be placed in a drop of saline

Venom Collection from Spiders and Snakes

63

within the well of a cutoff cap of a 1.5 mL microfuge tube resting on an ice pack (see Note 13). Use tweezers, forceps, or dissecting needles (insect pins) to separate venom glands from the chelicerae, and then remove the chelicerae and any other extraneous tissue. 6. Rupture venom glands within the saline drop using the tweezers, forceps, or dissecting needles. Use of dissecting needles can minimize fluid loss due to capillary fluid movement up the tweezers or forceps tips. Gently agitate the glands, and then remove the emptied sacs. Glands from multiple individuals can be dissected within the same drop of saline. 7. Transfer venom-saline mixture to storage tube; this can be accomplished simply by placing the tube cap back onto the tube. 8. Centrifuge the sample (typically comprising glands from multiple individuals) for 10 min at 10,000  g and 6  C; use a 500 μL pipette to remove supernatant and place in a new microfuge tube. 9. Treat and store the venom-saline sample as described in Subheading 3.1. 3.4 Voluntary Venom Extraction of Snakes: Spontaneous Expulsion

This method requires restraint of the snake by any of numerous methods, and depends on the snake’s natural tendency to bite and expulse venom defensively. Some snakes refuse to cooperate. Exact approach will vary depending on snake handler’s preference, species and size of snake, and individual snake’s disposition. 1. Alternative 1, via bare hands (Fig. 2a): Place snake on a solid object (table or ground); firmly press a snake hook or pinning tool onto the snake’s head to immobilize it (without pressing hard enough to cause injury); and then grasp the back of the head (and the neck) manually while avoiding the fangs. Large snakes often require simultaneous control of the body to overcome twisting or thrashing that could injure the snake’s neck or disrupt the grasp on the snake’s head; this can be accomplished by using a second hand, a hook, or other restraining device (see Note 14). Alternative 2, via gloves or folded towel (Fig. 2b, e): Hold the tail end of the snake with one hand (if safe to do so), and then grab the snake’s head with the gloved hand (see Note 14) or a hand holding a folded towel (folded thick enough that the snake cannot bite through). Alternatively, place the snake into a bucket or trash can, allow snake to crawl upward to the top with head reaching the edge—at which point the body is extended and the snake cannot readily lunge—and then grasp the head.

64

William K. Hayes et al.

Fig. 2 Venom extraction (milking) of snakes via spontaneous expulsion (voluntary) or manual expulsion (involuntary). (a) Extraction via bare hands, with fingers and thumb firmly securing the snake’s head and neck, and snake’s mouth pressed against membrane-covered flask to encourage bite. (b) Extraction aided by the use of heavy-duty (but not necessarily fang-proof) gloves. (c) Extraction via fang inserted into capillary tube. (d) Extraction of snake immobilized within transparent cylinder and grasped with rubber-tipped forceps. (e) Extraction aided by the use of folded cloth towel (snake handler’s hand not visible). The snake often expulses venom spontaneously, but if it refuses to do so the thumb and index finger (of free hand) can be used to manually compress the venom fangs (involuntary manual expulsion), which forces venom out of the fangs and into the venom collection device. Spontaneous expulsion and manual expulsion comprise the most frequently used approaches to venom collection from snakes. Photos: William K. Hayes and Gerad A. Fox

Alternative 3, via cylinder (Fig. 2d): Induce the snake to crawl part way (preferably half the body length of more) into a transparent (plastic or glass) cylinder of appropriate diameter (slightly larger than head width), and then secure with one hand both the cylinder and the emerging portion of the snake’s posterior. Maneuver the snake forward until the snake’s head emerges at the far end, which can then be grasped (though not always necessary) by hand using the pinning tool, or by using rubber-tipped forceps [29]. 2. Maneuver the snake’s head to the venom collection device, and encourage the snake to bite or otherwise engage the device sufficiently to prompt venom discharge. Some snakes refuse to bite or expel venom using these procedures, though additional agitation, such as tapping the snake’s body or rubbing the ventral scales anteriorly, can help. Some snakes, such as

Venom Collection from Spiders and Snakes

65

atractaspidids and small elapids and viperids, can be stimulated to deliver venom by sliding pipette tips or capillary tubes over the fang (Fig. 2c) and wiggling the tube if necessary to provoke venom flow. 3. After venom discharge, return snake to its container (see Note 15). 4. Transfer venom to storage container (usually a small tube) and keep on ice until placed into a freezer for longer term storage. Sample may also be centrifuged to remove cellular debris, and lyophilized. 3.5 Involuntary Venom Extraction of Snakes: Manual Expulsion

This method is similar to Subheading 3.4, except that the investigator manually compresses the venom glands to expel the venom. These two approaches (spontaneous and manual expulsion) comprise the most frequently used methods of venom extraction for snakes. 1. Pin and capture the snake as described in Subheading 3.4. Alternatively, the snake can be anesthetized for safer handling. Inhalants can be delivered by inserting the anterior portion of the snake into a transparent cylinder (as described in Subheading 3.4), closing off the distal end of the tube with a membrane, and then injecting a small quantity of anesthetic through the membrane and onto a piece of cotton or paper towel just inside of the membrane [30]. Injectable anesthetics can be delivered by syringe and needle. See Note 16 for additional details on anesthesia. 2. Place snake’s head against the venom collection device, forcing the snake to open its mouth with the fangs extended over the device’s edge. 3. Place thumb and finger against the venom glands and squeeze firmly to force venom out of fangs and into venom collection device. 4. Return snake to container (see Note 15) and treat and store the venom sample as described in Subheading 3.4.

3.6 Involuntary Venom Extraction of Snakes: Electrical Stimulation

This seldom used method is similar to Subheadings 3.4 and 3.5, except that the investigator applies electrical stimulation to the venom glands to induce venom gland compression of an anesthetized snake. 1. Assemble the electrical stimulator. Attach the two electrodes to stimulator. 2. Adjust current phase, current pulse, and/or voltage to desired settings. Optimal settings will depend on snake species and size; similar to papers on electrical stimulation of spiders, investigators report a wide range of stimulation parameters,

66

William K. Hayes et al.

including either alternating or direct current, voltages from 1.5 to 10 V, and amperage of 10–30 mA, with stimulation generally corresponding to the size of snake (see Note 17). Begin with lower levels of stimulation and increase until optimal delivery is achieved. 3. Anesthetize the snake, as described in Subheading 3.5. 4. Place snake’s head against the venom collection device, forcing the snake to open its mouth with the fangs extended over the device’s edge. 5. Stimulate one venom gland, followed by the other. To do so, place one electrode against the posterior of the gland, and the other electrode against the anterior; alligator-clip electrodes can be clamped onto the skin if it is loose enough. Moistening the contacts with a small amount of saline or alcohol will increase conduction. Stimulation should cause venom expulsion from one fang—and sometimes both fangs—into the venom collection device. 6. Treat and store the venom sample as described in Subheading 3.4. 7. Return snake to its container after it becomes clear that recovery from anesthetic is well underway. 3.7 Involuntary Venom Extraction of Snakes: Pilocarpine Stimulation

This method is applied to most colubrid (rear-fanged) and some small elapid snakes. Injection of pilocarpine into an anesthetized snake stimulates glandular secretion, which substantially increases venom yield without appreciably affecting venom composition [31]. Additional but less often used methods of venom extraction have also been described for colubrids [24]. 1. Anesthetize the snake, as described in Subheading 3.5. 2. Inject pilocarpine hydrochloride, 5–10 mg/kg I.P. 3. Place snake on its back, and gently prop its mouth open (by hand, tape, or a small ball). 4. Position capillary tube or pipette tip around the largest rear maxillary teeth to collect the venom as it is secreted. 5. Treat and store the venom sample as described in Subheading 3.4. 6. Return snake to its container after it becomes clear that recovery from anesthetic is well underway.

3.8 Venom Gland Extraction of Snakes

This method is very rarely used for venom collection from snakes, as it is lethal and generally unnecessary for anything but the smallest snakes. However, gland excision is now used routinely to obtain tissue for transcriptome analysis, so venom could be obtained at this time, though the glands are milked a few days ahead of the

Venom Collection from Spiders and Snakes

67

procedure and therefore contain a limited amount of regenerated venom. As an alternative to gland excision, the venom could be aspirated from the lumen of intact glands via syringe. The steps below are more general than those provided for spiders in Subheading 3.3. 1. Set up dissection arena appropriate for the size of snake. Place the petri dish(es) on an ice pack. 2. Euthanize snake via overdose of anesthetic (see Subheading 3.5); consider exsanguinating the heart to ensure non-revival. 3. Cut and peel back the skin of the cheeks that covers the venom glands and the posterior portion of the venom duct. 4. Cut and peel back the musculature and connective tissue that surrounds the venom glands and the posterior portion of the venom ducts, being careful not to puncture the glands or ducts. 5. Ligate the exposed venom ducts to avoid venom drainage as the glands are removed. 6. Remove the venom glands; place them into a petri dish with chilled saline, and continue removing extraneous tissue adhering to the glands. 7. Macerate the venom glands into a clean saline solution (consider another clean petri dish). 8. Aspirate the venom-saline fluid and centrifuge it to remove the solid debris. 9. Treat and store the venom-saline sample as described in Subheading 3.4.

4

Notes 1. To modify a capillary tube to create a finer tip, heat the tube with a Bunsen burner or other flame sources (even a candle flame will work). Gently pull the capillary tube to elongate the tip to produce a fine point. Under a dissection microscope, carefully break the tip to create a beveled end; inspect the broken tip to ensure that glass fragments have been cleared away. Water may be necessary to clear glass shards from the capillary tube using a microdispenser (e.g., Drummond Captrol III) to expel the ultrapure water from the capillary tube. Modified capillary tubes can be cleaned and reused; a hot plate can help remove water after thorough washing. 2. The best buffer would be that used for subsequent analysis, such as for RP-HPLC. Alternatively, use an organic solvent like ethanol or acetonitrile and dry material immediately after collection to stabilize. An enzyme inhibitor could also be added to

68

William K. Hayes et al.

minimize proteolytic degradation, though few investigators do this. 3. For smaller volumes, the length of venom column can be photographed in a microscope and quantified using imaging software. 4. Forceps can be modified to improve conductance and better fit the unique morphology of the species. Consider tapering the forceps to a fine tip, coating one side with tape or a rubber or Plasti Dip spray (Plasti Dip International, Blaine, MN, USA) for insulation, and wrapping the other side in a sponge or other absorbent materials to hold saline or buffer solution. 5. Melt paraffin wax onto the bottom of glass petri dish sitting on a hot plate; the coating should be thick enough to securely hold insect pins but thin enough for spider to be submerged in saline. Allow paraffin wax to cool slowly and completely. 6. Spiders are more prone to bite with gentle compression of the abdomen and/or pinching of leg(s). Be careful not to press spider so vigorously into the parafilm that chelicerae and fang movement become impeded. 7. Different species and individuals exhibit varying temperaments and willingness to bite. Some can be reasonably handled by gloved hands, but the deadliest spiders must be treated with extreme caution. Gloves stretched tight can be pierced more readily by fangs, so a loose fit may be desirable. Legs are often very delicate, especially those of web-dwelling species, and leg autotomy (voluntary loss) is a common defensive strategy. Vigorous handling may result in serious injury or death to the spider. Handlers can take advantage of the natural behaviors of the spider (e.g., proclivity to flee and thanatosis) and gentle, well-placed prods. 8. Spider hairs and other materials may accumulate on the surface of the parafilm after venom collection. Such debris may be inconsequential if antibodies are used to capture the venom in an ELISA assay, but debris may contaminate the sample for other analyses. The venom dries and usually appears as a small disk on either or both sides of the parafilm. Consider wiping the outer surface of the parafilm with a dry Kimwipe®, and using fine-tipped forceps to remove large pieces of debris like hair. 9. The forceps can serve as a barrier against leg contraction. The electrical pulse may cause leg contraction that may obstruct one’s view of the fangs. The fangs will also likely contract against the chelicerae during the shock. 10. By placing each fang tip within the bevel of the pipette tip or capillary tube, the fangs can contract across the glass during the

Venom Collection from Spiders and Snakes

69

electrical stimulation, and the expelled venom goes directly into the tube. One can add a small amount of preferred buffer or ultrapure water to pipette tip or capillary tube to prevent venom from drying too quickly and gumming up it up. 11. When dissecting very small spiders, one can pin 5–8 spiders on the same petri dish to increase efficiency, dissecting all of them within a few minutes. 12. The venom glands sometimes remain in the cephalothorax, in which case the venom duct should be visible and can be used to remove the gland. 13. Coloring the outside of the cap with black ink can increase visibility of the translucent venom glands. 14. Snakes can be grasped by thumb and forefinger (especially if small), by thumb and middle finger with the forefinger placed on top of the head (often for medium snakes), or by the full hand and thumb (especially if large). The thumb and at least one finger should be placed securely behind the jaws; a digit placed beneath the chin of the snake could sustain a bite if the snake’s fang stabs through its lower jaw. For the first two grips, grasp the neck firmly with the remaining fingers to minimize twisting or thrashing that could injure the snake or pull it from the handler’s grasp. Small snakes can be held entirely by the grasp of the head, but larger snakes require additional support of the body. We have been told of severe envenomation resulting from the use of HexArmor gloves, with the rigid fangs of elapid snakes more likely to penetrate, but at least one case was described for a viperid. Clearly, caution is warranted when using these gloves. 15. Snakes are especially prone to retaliate upon release from a manual grasp. By releasing the snake upside down, the snake’s initial reaction is to right itself before attempting to strike or bite, which gives the handler more time to move hands safely beyond striking range. 16. Snakes are exposed to inhalants as needed. Adequate anesthesia is achieved when body becomes unresponsive to tapping and muscle tension of tail disappears, at which time the snake can be removed from the tube and should remain unresponsive for the brief duration of venom collection. Ketamine can be delivered at 20–80 mg/kg S.C. or I.M., and propofol at 5–10 mg/ kg I.V., with dose depending on the depth of anesthesia desired [32]. Righting reflex, tail muscle tension, and general body movements are good indicators of anesthesia depth. Heart and respiratory activity are also good indicators of anesthesia depth. Heart activity can be monitored by placing a finger on the heart, or by placing the snake upside down and watching for the rhythm. Breathing can become shallow and difficult to

70

William K. Hayes et al.

discern visually when the snake is under deep anesthesia. If oversedation is suspected, and especially if heart activity ceases, artificial respiration may be given by placing a large pipette tube into the trachea (aperture in floor of mouth) and gently blowing into it. Always maintain cautious handling of snake during anesthesia, as snakes can sometimes awaken with little to no warning. 17. One recent study delivered 5-s tetanic trains of 10–30 mA (amperage increasing with snake size) to cottonmouths (Agkistrodon piscivorus) [33]. The reader can consult prior studies of electrical stimulation cited by these authors. References 1. Weinstein SA, Smith T, Kardong KV (2009) Reptile venom glands: form, function and future. In: Mackessy SP (ed) Handbook of reptile venoms and toxins. CRC Press, Boca Raton, pp 65–94 2. Vonk FJ, Admiraal JF, Jackson K, Reshef R, de Bakker MA, Vanderschoot K et al (2008) Evolutionary origin and development of snake fangs. Nature 454:630–633 3. Perret BA (1977) Venom regeneration in tarantula spiders. I. Analysis of venom produced at different time intervals. Comp Biochem Physiol A Mol Integr Physiol 56:607–613 4. Luna MS, Valente RH, Perales J, Vieira ML, Yamanouye N (2013) Activation of Bothrops jararaca snake venom gland and venom production: a proteome approach. J Proteome 94:460–472 5. Cooper AM, Kelln WJ, Hayes WK (2014) Venom regeneration in the centipede Scolopendra polymorpha: evidence for asynchronous venom component synthesis. Zoology 117:398–414 6. Inceoglu B, Lango J, Jing J, Chen L, Doymaz F, Pessah IN et al (2003) One scorpion, two venoms: prevenom of Parabuthus transvaalicus acts as an alternative type of venom with distinct mechanism of action. Proc Natl Acad Sci U S A 100:922–927 7. Nisani Z, Hayes WK (2011) Defensive stinging by Parabuthus transvaalicus scorpions: risk assessment and venom metering. Anim Behav 81:627–633 8. Nisani Z, Boskovic DS, Dunbar SG, Kelln W, Hayes WK (2012) Investigating the chemical profile of regenerated scorpion (Parabuthus transvaalicus) venom in relation to metabolic cost and toxicity. Toxicon 60:315–323

9. Kristensen C (2005) Comments on the natural expression and artificial extraction of venom gland components from spiders. Toxin Rev 24:257–270 10. Morgenstern D, Hamilton B, Sher D, Jones A, Mattius G, Zlotkin E et al (2012) The bio-logic of venom complexity. Toxicon 60:241–242 11. Cooper AM, Nelsen DR, Hayes WK (2017) The strategic use of venom by spiders. In: Gopalakrishnakone P, Malhotra A (eds) Evolution of venomous animals and their toxins. Springer Science+Business Media, Dordrecht, pp 145–166 12. Meadows PE Russell FE (1970) Milking of arthropods. Toxicon 8:311–312 13. Glenn JL, Straight RC (1982) The rattlesnakes and their venom yield and lethal toxicity. In: Tu AT (ed) Rattlesnake venoms: their action and treatment. Marcel Dekker, New York, pp 3–119 14. Hayes WK (2008) The snake venom-metering controversy: levels of analysis, assumptions, and evidence. In: Hayes WK, Beaman KR, Cardwell MD, Bush SP (eds) The biology of rattlesnakes. Loma Linda University Press, Loma Linda, pp 191–220 15. Nelsen DR, Hayes WK (2014) Poke but don’t pinch: risk assessment and venom metering in the western black widow spider (Latrodectus hesperus). Anim Behav 89:107–114 16. Rosenberg HI (1992) An improved method for collecting secretion from Duvernoy’s gland of colubrid snakes. Copeia 1992:244–246 17. di Tada IE, Martori RA, Doucet ME, Abalos JW (1976) Venom yield with different milking procedures. In: Rosenberg P (ed) Toxins: animal, plant and microbial. Pergamon Press, Oxford, pp 3–7

Venom Collection from Spiders and Snakes 18. Cooper AM, Fox GA, Nelsen DR, Hayes WK (2014) Variation in venom yield and protein concentration of the centipedes Scolopendra polymorpha and S. subspinipes. Toxicon 82:30–51 19. Bettini S (1978) Arthropod venoms. Springer, Berlin 20. Bu¨cherl W, Buckley EE (1971) Venomous animals and their venom. In: Venomous invertebrates, vol 3. Academic Press, Orlando 21. Hayes WK, Lavı´n-Murcio P, Kardong KV (1993) Delivery of Duvernoy’s secretion into prey by the brown tree snake, Boiga irregularis (Serpentes: Colubridae). Toxicon 31:881–887 22. Munekiyo SM, Mackessy SP (1998) Effects of temperature and storage conditions on the electrophoretic, toxic and enzymatic stability of venom components. Comp Biochem Physiol B Biochem Mol Biol 119:119–127 23. Maier L, Root TM, Seyfarth E-A (1987) Heterogeneity of spider leg muscle: histochemistry and electrophysiology of identified fibers in the claw levator. J Comp Physiol B 157:285–294 24. Weinstein SA, Kardong KV (1994) Properties of Duvernoy’s secretions from opisthoglyphous and aglyphous colubrid snakes. Toxicon 32:1161–1185 25. Grothaus RH, Howell DE (1967) A new technique for the recovery of spider venom. J Kansas Entomol Soc 40:37–41 26. Rocha-e-Silva TA, Sutti R, Hyslop S (2009) Milking and partial characterization of venom

71

from the Brazilian spider Vitalius dubius (Theraphosidae). Toxicon 53:153–161 27. Garb JE (2014) Extraction of venom and venom gland microdissections from spiders for proteomic and transcriptomic analyses. J Vis Exp 93:e51618. https://doi.org/10. 3791/51618 28. Besson T, Debayle D, Diochot S, Salinas M, Lingueglia E (2016) Low cost venom extractor based on arduino® board for electrical venom extraction from arthropods and other small animals. Toxicon 118:156–161 29. Hogan MP (2015) Field venom extractions: saving fingers with tubes, forceps, and nerf bullets. Herpetol Rev 46:339–342 30. Hardy DL, Greene HW (1999) Surgery on rattlesnakes in the field for implantation of transmitters. Sonoran Herpetol 12:25–27 31. de Morais-Zani K, Serino-Silva C, da Costa Galizio N, Tasima LJ, Pagotto JF, da Rocha MMT et al (2018) Does the administration of pilocarpine prior to venom milking influence the composition of Micrurus corallinus venom? J Proteome 174:17–27 32. Mader DR (2005) Reptile medicine and surgery. Elsevier, Amsterdam 33. McCleary RJ, Heard DJ (2010) Venom extraction from anesthetized Florida cottonmouths, Agkistrodon piscivorus conanti, using a portable nerve stimulator. Toxicon 55:250–255

Chapter 4 Production and Purification of Recombinant Toxins Matan Geron Abstract Recombinant expression of toxins enables us to produce adequate quantities of these proteins which can be used to perform experiments at molecular, cellular, and behavioral levels. Furthermore, toxins can be edited by using simple molecular biology methods when producing them recombinantly. Thus, in many cases establishing a protocol for the recombinant expression of a toxin of interest is crucial in exploring the structure and function of the toxin and its effectors. To date, Escherichia coli (E. coli) represents the most widely used heterologous expression system in which recombinant proteins are usually accumulated in the bacterium cytoplasm. However, as many animal toxins contain disulfide bonds they tend to be misfolded and aggregate when found in the reducing E. coli cytoplasm. In contrast, conditions in the bacterium periplasm allow disulfide bond formation and correct folding of such toxins. Here, we describe a protocol for the production and purification of bioactive recombinant disulfide-rich toxins via periplasmic expression. Key words Recombinant toxins, Escherichia coli (E. coli) bacteria, Periplasm, Disulfide bonds, HPLC, TEV protease, Affinity chromatography

1

Introduction Structural and functional characterization of a toxin requires a considerable amount of that protein [1, 2]. However, in most cases efforts to collect toxins from their native sources involve tedious and/or expensive procedures which result in low yields rendering the whole process inapplicable [1]. Therefore, achieving a sufficient amount of material is a major challenge in studying a desired toxin [1, 2]. To overcome this obstacle, a protein of choice can be synthetically produced by solid-phase peptide synthesis (SPPS) which requires a subsequent time- and effort-consuming step of folding [1]. This step can be further complicated for toxins as many of them possess several disulfide bonds which are crucial for their folding and stability [1]. Another approach which is widely used is to recombinantly produce the venom peptides in a suitable host from which high yield of correctly folded proteins could be purified

Avi Priel (ed.), Snake and Spider Toxins: Methods and Protocols, Methods in Molecular Biology, vol. 2068, https://doi.org/10.1007/978-1-4939-9845-6_4, © Springer Science+Business Media, LLC, part of Springer Nature 2020

73

74

Matan Geron

in a cost-effective manner [3]. Furthermore, performing genome modifications is simple and quick in recombinant production. Heterologous expression systems being successfully used in research laboratories include different bacteria, yeasts, insect cells, and mammalian cells [4]. However, the bacterium Escherichia coli (E. coli) is the most commonly used host for the recombinant production of proteins as it presents several attractive advantages [5]. First, as for other bacteria, it is quick, easy, and cheap to culture E. coli due to its fast growth capability in simple media [5]. Second, E. coli is readily transformed by well-characterized vectors [5]. Third, E. coli can accumulate very large quantities of heterologous protein [4]. Finally, E. coli genetics are thoroughly studied and as a result many strains of E. coli were developed to further facilitate heterologous expression [6]. Nevertheless, heterologous expression of toxins in E. coli also presents a few issues. Limiting factors include the absence of chaperones, presence of proteases, and aggregation of overexpressed proteins [7, 8]. However, the most prominent obstacle is that most proteins are expressed in the bacteria’s cytoplasm where sophisticated machinery to perform posttranslational modifications and correct S-S bonding are lacking [1, 3, 7, 9]. In addition, the reducing cytoplasmic environment further undermines the probability of correct folding in disulfiderich toxins [10]. Thus, these toxins are usually poorly soluble and aggregate as inactive inclusion bodies in the cytoplasm [3, 4, 7]. Therefore, these toxins’ purification requires further denaturation and refolding steps [1, 4, 7, 11]. In contrast to the cytoplasmic environment, conditions in the E. coli’s periplasm favor the formation of S-S bonds as the periplasmic environment is oxidizing [7, 8, 12]. Moreover, the periplasm contains disulfide bond-forming (Dsb) proteins and isomerases which catalyze the formation and rearrangement of disulfide bonds to produce a stable and correctly folded toxin while the proteolytic activity is lower in this compartment [11, 12]. Thus, in order to obtain an active disulfide-rich toxin, it can be directed to the E. coli periplasm. Here, we present an optimized protocol for the production and purification of disulfide-rich toxins based on exporting the toxin to the bacterial periplasm which contains seven steps. Initially, one has to select the specific E. coli strain to be used as a host. BL21 is a popular strain in use as it lacks some proteases which can reduce protein yield [3]. Next, a proper vector containing the gene of the toxin of interest should be constructed. To allow the export of the protein to the periplasm and the success of the following purification steps, several sequences are inserted to the vector to produce a fusion protein with the toxin. These insertions include (listed from N- to C-terminal of the fusion protein) the following: (1) MalE signal sequence which targets the fusion protein to the periplasm. (2) His6 tag, a sequence of 6 histidines which binds nickel and

Disulfide-Rich Recombinant Toxins

75

Fig. 1 Production of a recombinant toxin via periplasmic expression. (a) Schematic illustration of a pLic-MBPToxin construct including T7 promoter, lac operator, MalE signal sequence for periplasmic export, His6 affinity tag, maltose-binding protein (MBP) fusion tag, TEV protease recognition site, and toxin. (b) SDS-PAGE analysis of fusion protein expression. Lane 1, molecular weight markers. Lane 2, protein expression before induction

76

Matan Geron

facilitates the fusion protein purification by affinity chromatography on a Ni-NTA column. (3) Maltose-binding protein (MBP), a commonly used 43 kDa fusion partner protein that enhances the solubility of the fusion protein and aids in toxin folding [3]. (4) TEV recognition site, a seven-residue (ENLYFQG) sequence which is recognized by the TEV protease [13]. Cleavage at this site allows the release of the toxin from its fusion partners before the final purification step. (5) Desired toxin to be produced (Fig. 1). Control over production is achieved by a T7 promoter and a lac operator, upstream to the fusion protein gene, which enable strong and inducible protein expression [11]. Once produced, the fusion protein is extracted by lysing the bacteria using high-shear mechanical pressure. Cell content is then subjected to affinity chromatography. The lysate is passed through a column containing immobilized nickel to which the His-tagged fusion proteins bind whereas other proteins are eluted in the flow-through. Next, the fusion protein is eluted separately by washing the column with a solution containing a high concentration of imidazole which competes with the histidines over binding to the nickel. The purified fusion protein is then cleaved by TEV protease to yield a free toxin. TEV protease activity requires reducing environment which could reshuffle the toxin’s disulfide bonds. Thus, the cleavage reaction is performed in mild reducing conditions that preserve both TEV protease activity and correct fold of the toxin. The liberated toxin is further purified from the other components of the cleavage mixture using reversed-phase HPLC. The collected HPLC fraction containing the toxin is then lyophilized and dissolved. Finally, toxin concentration is measured via UV absorption of the toxin’s aromatic residues at 280 nm (A280). In summary, this optimized protocol represents a facile method to produce a high yield of active disulfide-rich toxins using cheap materials (albeit expensive instrumentation) over a time course of just 4–5 days.

ä Fig. 1 (continued) with IPTG (“uninduced”). Lane 3, protein expression after induction with IPTG (“induced”). (c) SDS-PAGE analysis of fusion protein extraction via nickel affinity chromatography. Lane 1, molecular weight markers. Lane 2, lysate of fusion protein-expressing BL21(DE3) bacteria. Lane 3, flow-through of lysate subjected to Ni-NTA column. Lane 4, purified protein eluted from Ni-NTA column. (d) SDS-PAGE analysis of fusion protein cleavage by TEV protease. Lane 1, molecular weight markers. Lane 2, extracted fusion protein. Lanes 3–7, TEV protease cleavage of proteins shown in lane 2 at the indicated time. (e) Left: Reversed-phase HPLC (C18) chromatogram of recombinant MBP-DkTx fusion protein after TEV protease cleavage. A red asterisk denotes the retention of the active recombinant toxin. Right: Lane 1, molecular weight markers. Lane 2, extracted fusion protein. Lane 3, TEV protease cleavage of proteins shown in lane 2. Lanes 4–8, HPLC fractions indicated in the left panel

Disulfide-Rich Recombinant Toxins

2

77

Materials Prepare all solutions with double-distilled water (DDW; 18 MΩ-cm at 25
C) and analytical grade reagents.

2.1 Cloning and Transformation

1. pLic-MalE-His6-MBP vector, 2. Calf intestine phosphatase (CIP). 3. Electrophoresis gel 1%: Add ethidium bromide (EtBr) to a bottle containing 2% TAE electrophoresis buffer (50) in DDW to make a final concentration of 1 μg/mL. Dissolve 0.5 g agarose in 50 mL EtBr-TAE solution by heating the mixture. Mix well, pour into a gel-casting frame, and wait for the gel to solidify (~30 min). 4. PureLink gel extraction kit. 5. T4 ligase. 6. XL1–Blue competent cells. 7. LB medium and plates: For 1 L medium, dissolve 10 g tryptone, 5 g yeast extract, and 10 g NaCl in DDW. Adjust pH with 200 μL NaOH 5 N. Sterilize medium using an autoclave. To produce LB plates, add 15 g agar per 1 L LB medium and a stirrer before sterilization. Mix well the sterilized LB + agar solution and then coat 10 mm round dishes. Allow plate content to solidify for at least 4 h. Store at 4
C. 8. LB + ampicillin medium and plates: For both medium and plates, add ampicillin solution to sterilized media after they cool down to 50
C to make a final concentration of 100 μg/ mL and mix. 9. NucleoSpin mini prep kit. 10. BL21(DE3) competent E. coli. 11. Glycerol 50% solution: Add DDW to a cylinder containing glycerol to produce a 50% glycerol solution and mix. Sterilize using an autoclave. 12. 1.5 mL Cryo-vials.

2.2 Fusion Protein Induction and Purification

1. Isopropyl β-D-1-thiogalactopyranoside (IPTG): Make stock solutions by dissolving IPTG in DDW. Store aliquots at 20
C. 2. DPBS w/o Ca2+. 3. Extraction buffer: 50 mM NaH2PO4, 500 mM NaCl, 20 mM imidazole, 2 mM MgCl2, DNaseI, pH ¼ 8.0. Prepare fresh extraction buffer before use by adding MgCl2 and DNaseI to an already prepared binding buffer. The amount of DNaseI added is dependent on the specific unit strength of the enzyme.

78

Matan Geron

4. Phenylmethylsulfonyl fluoride (PMSF): Make a stock solution by dissolving PMSF in DMSO. Store aliquots at 20
C protected from light. 2.3 Toxin Purification

1. Binding buffer: 50 mM NaH2PO4, 500 mM NaCl, 20 mM imidazole, pH ¼ 8.0. Store at 4
C. 2. Elution buffer: 50 mM NaH2PO4, 500 mM NaCl, 500 mM imidazole, pH ¼ 7.4. Store at 4
C. 3. Amicon ultracentrifugal filter unit. 4. Storage buffer: Tris–HCl, NaCl, pH ¼ 7.4. Store at 4
C. 5. TEV protease. 6. Cleavage buffer: 25 mM Tris–HCl, 250 mM NaCl, 0.6 mM GSH, 0.4 mM GSSG, pH ¼ 7.0. Prepare fresh cleavage buffer before each use. 7. Syringe-driven filter (0.22 μm). 8. Buffer A: 0.1% TFA in DDW. Store at room temperature. 9. Buffer B: Acetonitrile, HPLC grade. 10. Liquid nitrogen.

3

Methods Perform all procedures in room temperature unless otherwise specified.

3.1 Subcloning the Desired Toxin Gene into an Appropriate Vector

1. Design a toxin gene preceded by a TEV recognition site. Flank this sequence with the appropriate restriction sites so it could be inserted to the vector after the MBP gene (see Notes 1 and 2). 2. Digest the vector, insert (2 μg each) using the chosen restriction enzymes, and add calf intestine phosphatase (CIP) to the vector restriction reaction only, according to the manufacturer’s instructions. 3. Run the digested vector and insert using gel electrophoresis. Cut the desired vector and insert fragments, and purify them using a gel cleanup kit according to the manufacturer’s instructions. Determine the concentration of fragments by UV absorbance at 260 nm using a NanoDrop device. 4. Ligate purified fragments (1:3 vector-to-insert molar ratio) using T4 ligase according to the manufacturer’s instructions. 5. Transform ligation product into chemically competent bacteria (use 1:9 DNA-to-bacteria volume ratio). Inoculate a 1 mL LB solution with the transformed bacteria and incubate at 37
C

Disulfide-Rich Recombinant Toxins

79

and 250 rpm for 1 h. Streak 100 μL of the inoculated LB on a plate containing LB + ampicillin. Incubate overnight at 37
C. 6. Transfer single colonies into separate tubes containing LB + ampicillin. Incubate the tubes in 37
C and 250 rpm overnight. Extract plasmids using a mini-prep kit and measure their concentration by spectrophotometry using NanoDrop. Keep plasmids at 20
C. Sequence plasmids to confirm successful subcloning. 3.2 Preparing Glycerol Stocks of BL21(DE3) Bacteria

1. Transform purified plasmids into BL21(DE3) bacteria as was described in Subheading 3.1, step 5. Inoculate a 1 mL LB solution with the transformed bacteria and incubate at 37
C and 250 rpm for 1 h. In order to isolate a single colony, streak the inoculated LB on a plate containing LB + ampicillin 100 μg/mL. Incubate overnight at 37
C (see Note 3). 2. Transfer a single colony into a tube containing 2.5 mL LB + ampicillin 100 μg/mL. Incubate the tube at 37
C and 250 rpm for 4 h. Transfer 20 μL of the inoculated solution into a tube containing 10 mL fresh LB + ampicillin 100 μg/mL. Incubate the tube at 37
C and 250 rpm overnight. 3. Mix the bacteria with an equal volume of a sterilized 50% glycerol solution. Quickly prepare aliquots in pre-labeled cryo-vials. Store aliquots at 80
C.

3.3 Inducing Fusion Protein Expression

1. Thaw a glycerol stock vial and immediately inoculate a freshly sterilized LB solution supplemented with ampicillin (final concentration ¼ 100 μg/mL). Incubate the flask at 37
C and 180 rpm. Monitor bacterial growth by cuvette-based spectrophotometry using NanoDrop until OD600 ¼ 0.8–1.0 is reached (see Notes 4–6). 2. Add IPTG (final concentration ¼ 250 μM) in order to induce fusion protein expression. Incubate bacteria at 16
C and 180 rpm overnight. 3. Centrifuge bacteria at 4
C and 6000  g for 15 min. Discard the supernatant and wash precipitated bacteria by gently resuspending them in DPBS w/o Ca2+ using a pipette. Transfer bacteria to tubes and centrifuge at 4
C and 3000  g for 30 min. Discard supernatant and store the bacterial pellet at 80
C.

3.4 Extracting Fusion Protein from Bacteria

1. Let bacterial pellet thaw on ice. Gently resuspend bacteria in extraction buffer by pipetting (use approximately 10 mL extraction buffer per 1 g pellet). Pass the entire volume of bacterial suspension through a homogenizer until the bacteria are evenly dispersed (~6 times). In order to block the digestion of the fusion protein by liberated proteases, add PMSF (final concentration ¼ 0.3–0.5 mM) and mix well (see Notes 7–9).

80

Matan Geron

2. In the Microfluidizer, set the pressure to 80 kPa. Fill the cooling bath with ice to prevent denaturation of proteins due to the heat produced by high shear rates. Pass 20–40 mL buffer A to wash the instrument tubing and discard flow-through. Pass the entire volume of bacteria followed by another 20–30 mL of buffer A through the microfluidizer and collect the lysate. Add more PMSF to retain the desired final concentration and mix (see Note 10). 3. Quickly distribute the lysate to centrifuge tubes and centrifuge at 4
C and 40,000  g for 40 min. Discard the pellet and keep the supernatant on ice. 3.5 Affinity Chromatography Purification

1. Prime the chromatography system and Ni-NTA column by washing them with 5 column volumes of DDW followed by 10 column volumes of binding buffer. Monitor the UV absorbance of the flow-through throughout the chromatography process (see Notes 11 and 12). 2. Immerse the reservoir containing the lysate supernatant in ice. Pass the lysate through the Ni-NTA column to capture His-tagged proteins. Discard the flow-through. 3. Wash the column with 12 column volumes of binding buffer. Perform further wash with 3 column volumes of 5% elution buffer and 95% binding buffer to eliminate impurities nonspecifically bound to the column. Discard the flow-through (see Note 13). 4. Elute the His-tagged fusion protein with 48% elution buffer and 52% binding buffer and collect the eluted fractions. Perfuse the column until all the protein was eluted. 5. In order to concentrate the eluted protein and remove imidazole from the solution, transfer the collected fractions to an ultracentrifugal filter unit with an appropriate molecular weight cutoff. Add TN buffer until the filter unit is full. Centrifuge at 4
C and 3000  g for 30 min. Discard the flowthrough and add TN buffer to the concentrate until the unit is full. Centrifuge as before. Repeat TN buffer addition and centrifugation two more times (see Notes 14 and 15). 6. Measure the concentration of fusion protein using the expected extinction coefficient and molecular weight calculated from the amino acid sequence. The concentrated fusion protein can be stored at 20
C for a few days.

3.6 Fusion Protein Cleavage

1. To cleave the toxin from the fusion protein, add 2 mg fusion protein to 1 mL cleavage buffer and gently mix. Next, add the TEV protease to the cleavage buffer (25 μg protease per 1 mg protein to be cleaved) and gently mix (see Note 16). 2. Allow the reaction to proceed at 23
C for 24 h.

Disulfide-Rich Recombinant Toxins

3.7 Toxin Purification by RP-HPLC

81

1. To concentrate the proteins, transfer the cleavage reaction solution to an ultracentrifugal filter unit with an appropriate molecular weight cutoff. Centrifuge at 4
C and 3000  g for 75 min. Filter the concentrate through a syringe-driven filter (0.22 μm) in order to avoid the injection of sediments to the HPLC column. Use a minimal diameter filter unit to minimize sample loss or dilution (see Note 17). 2. Set the flow to 3 mL/min and equilibrate a C18 HPLC column with 5% acetonitrile and 95% buffer A (see Notes 18 and 19). 3. Load column with the products of the cleavage reaction at an amount appropriate with the column capacity. Isolate the different components of the cleavage reaction using a gradient (20–50% acetonitrile in 50 min). Monitor UV absorbance at 280 nm throughout the chromatography process. Collect fractions containing the toxin of interest. 4. Pierce the cap of centrifuge tubes and transfer the collected fractions to them. Flash freeze the fractions using liquid nitrogen and lyophilize them. 5. After the solvent has evaporated, dissolve the toxin in DDW (or another solvent of your choice). Measure the concentration of fusion protein using the expected extinction coefficient and molecular weight calculated from the amino acid sequence. Prepare aliquots and store them at 80
C (see Notes 20 and 21).

4

Notes 1. To improve fusion protein yield the codons of the passenger toxin could be optimized for expression in E. coli as this bacterium may use different tRNAs than the toxin’s native source [6, 14]. Alternatively, E. coli strains which contain additional tRNAs could be used [3, 11]. 2. TEV protease cleaves one peptide bond before the C-terminal of its recognition site [15]. Thus, TEV cleavage leaves one residue attached to the toxin’s N-terminal which is unlikely to affect the protein’s structure or function [1]. However, the residue in the C-terminal of the TEV recognition site is not canonical and can be exchanged with several other residues to coincide with the toxin N-terminal. While Gly, Ser, and Ala are preferable in this recognition site position, other amino acids could be used at the expense of cleavage efficiency [1]. Nonetheless, many substitutions produce a reasonably cleavable recognition site [16]. 3. Specialized E. coli strains can be used to achieve correctly folded disulfide-rich toxins which are expressed in the cytoplasm

82

Matan Geron

[17]. These strains’ cytoplasm contains an oxidative environment and/or Dsb proteins such as in the Origami strains (Novagen) or the SHuffle strain (New England Biolabs) [3, 17]. 4. Keep 1–5 μL samples from each step of the protocol to be analyzed by SDS-PAGE (12% gel) and Coomassie staining. This is especially recommended when standardizing your protocol and allows troubleshooting. Dilute samples using DDW and add reducing and denaturing agents according to the manufacturer’s instructions to produce linear and negatively charged proteins. Run proteins using electrophoresis according to the instructions of the manufacturer. Rinse the gel with DDW and stain proteins with a Coomassie dye for 90 min while shaking. Visualize proteins and record staining using an imager. 5. In order to preserve the disulfide bonds in the toxin of interest, do not use strong reducing agents throughout this protocol. 6. To accelerate bacterial growth and maximize fermentation volume, baffled bottom flasks with vented closures can be used. These flasks allow increased mixing and gas exchange. 7. Selective extraction of periplasmic proteins and subsequently less exposure to proteases and impurities can be obtained by osmotic shock. However, this method of destabilizing the outer envelope produces incomplete extraction and typically lower yields than cell disruption lysis [1]. 8. If access to a microfluidizer to perform cell disruption is not available, other whole-cell lysis methods can be used to recover the fusion protein. These methods include sonication, freeze/ thaw cycles, chemical lysis, and French press. However, each method has its drawbacks and thus can be time-consuming, expensive, laborious, and not suited with processing large volumes or can yield incomplete lysis of bacteria [1]. 9. Solutions and liquids used in the processes of cell disruption, affinity chromatography, and RP-HPLC should be filtered before use. 10. Add PMSF right before bacterial lysis as this substance is highly unstable in aqueous solutions. 11. MBP can also be used to purify the fusion protein as MBP-binding amylose columns are commercially available. Thus, it is possible to use this affinity chromatography method instead of Ni-NTA columns or in adjunct. However, in many cases, the amylose-binding site of MBP is occupied and blocked by the passenger toxin rendering the amylose affinity purification ineffective [1]. 12. Do not use buffers containing chelators (e.g., EDTA, EGTA) in immobilized metal affinity chromatography (IMAC)

Disulfide-Rich Recombinant Toxins

83

columns such as Ni-NTA. Nickel ions in the affinity matrix required for His6 tag binding are chelated and stripped from the column by chelators [3]. 13. Washing the Ni-NTA column with 5% elution buffer results in a minor loss of fusion protein along with impurities nonspecifically bound to the resin. Skipping this wash step is possible if maximal yield is preferential over sample purity. 14. Use an ultracentrifugal filter unit with the highest molecular weight cutoff able to retain the protein of interest. This enables the removal of smaller proteins to the flow-through and enhances sample purity. 15. Perform up-and-down pipetting every 10–15 min of centrifugation to avoid protein aggregation on the filter. 16. TEV protease can be produced in-house or bought. TEV protease from different sources could produce different efficiency and kinetics in different settings. Thus, it is important to screen and optimize cleavage conditions (e.g., amount of protease, buffer, temperature, reaction time) in small-scale reactions before starting this protocol [3]. 17. If required, an additional purification step to remove the fusion tag can be performed prior to HPLC. Possible purification methods employed include affinity, size exclusion, and ion-exchange chromatographies. 18. To produce better yields, the length of the alkyl chain in the RP-HPLC column resin should be inversely proportional to the hydrophobicity of the toxin. Large highly hydrophobic toxins should be purified with either a C4 or a C8 column while less hydrophobic toxins can be purified using C18 columns [1]. 19. RP-HPLC conditions should be optimized for each toxin purification procedure. We start with the conditions mentioned above. 20. If crucial, very accurate measurements of toxin concentration can be performed by quantitative amino acid analysis. However, for mundane use, concentration can be measured by A280 using a spectrophotometer. Cuvette-based measurements require larger volume yet they are usually more accurate than micro-volume measurements. Thus, when using NanoDrop for measuring micro-volumes perform triplicate measurements of three protein dilutions [1]. 21. The identity of the dissolved protein can be verified by SDS-PAGE electrophoresis or more accurately by subjecting a sample to MALDI/TOF mass spectrometry. Toxin activity should also be tested after first using this protocol for a toxin of interest.

84

Matan Geron

References 1. Klint JK, Senff S, Saez NJ et al (2013) Production of recombinant disulfide-rich venom peptides for structural and functional analysis via expression in the periplasm of E. coli. PLoS One 8:e63865. https://doi.org/10.1371/ journal.pone.0063865 2. Dilworth MV, Piel MS, Bettaney KE et al (2018) Microbial expression systems for membrane proteins. Methods 147:3. https://doi. org/10.1016/J.YMETH.2018.04.009 3. Costa S, Almeida A, Castro A, Domingues L (2014) Fusion tags for protein solubility, purification, and immunogenicity in Escherichia coli: the novel Fh8 system. Front Microbiol 5:63. https://doi.org/10.3389/fmicb.2014. 00063 4. Demain AL, Vaishnav P (2009) Production of recombinant proteins by microbes and higher organisms. Biotechnol Adv 27:297–306. https://doi.org/10.1016/J.BIOTECHADV. 2009.01.008 ¨ ztu¨rk S, Ergu¨n BG, C 5. O ¸ alık P (2017) Double promoter expression systems for recombinant protein production by industrial microorganisms. Appl Microbiol Biotechnol 101:7459–7475. https://doi.org/10.1007/ s00253-017-8487-y 6. Makino T, Skretas G, Georgiou G (2011) Strain engineering for improved expression of recombinant proteins in bacteria. Microb Cell Factories 10:32. https://doi.org/10.1186/ 1475-2859-10-32 7. Berlec A, Sˇtrukelj B (2013) Current state and recent advances in biopharmaceutical production in Escherichia coli, yeasts and mammalian cells. J Ind Microbiol Biotechnol 40:257–274. https://doi.org/10.1007/ s10295-013-1235-0 8. Jalalirad R (2013) Selective and efficient extraction of recombinant proteins from the periplasm of Escherichia coli using low concentrations of chemicals. J Ind Microbiol Biotechnol 40:1117–1129. https://doi.org/ 10.1007/s10295-013-1307-1 9. Low KO, Muhammad Mahadi N, Illias R (2013) Optimisation of signal peptide for recombinant protein secretion in bacterial

hosts. Appl Microbiol Biotechnol 97:3811–3826. https://doi.org/10.1007/ s00253-013-4831-z 10. Freudl R (2018) Signal peptides for recombinant protein secretion in bacterial expression systems. Microb Cell Factories 17:52. https:// doi.org/10.1186/s12934-018-0901-3 11. Terpe K (2006) Overview of bacterial expression systems for heterologous protein production: from molecular and biochemical fundamentals to commercial systems. Appl Microbiol Biotechnol 72:211–222. https:// doi.org/10.1007/s00253-006-0465-8 12. Devi VS, Mittl PRE (2011) Monitoring the disulfide bond formation of a cysteine-rich repeat protein from helicobacter pylori in the periplasm of Escherichia coli. Curr Microbiol 62:903–907. https://doi.org/10.1007/ s00284-010-9803-2 13. Raran-Kurussi S, Waugh DS (2012) The ability to enhance the solubility of its fusion partners is an intrinsic property of maltose-binding protein but their folding is either spontaneous or chaperone-mediated. PLoS One 7:e49589. https://doi.org/10.1371/journal.pone. 0049589 14. Kudla G, Murray AW, Tollervey D, Plotkin JB (2009) Coding-sequence determinants of expression in Escherichia coli. Science 324:255–258. https://doi.org/10.1126/sci ence.1170160 15. Arnau J, Lauritzen C, Petersen GE, Pedersen J (2006) Current strategies for the use of affinity tags and tag removal for the purification of recombinant proteins. Protein Expr Purif 48:1–13. https://doi.org/10.1016/j.pep. 2005.12.002 16. Kapust RB, Too¨zseo´r J, Copeland TD, Waugh DS (2002) The P10 specificity of tobacco etch virus protease. Biochem Biophys Res Commun 294:949–955. https://doi.org/10.1016/ S0006-291X(02)00574-0 17. Becker S, Terlau H (2008) Toxins from cone snails: properties, applications and biotechnological production. Appl Microbiol Biotechnol 79:1–9. https://doi.org/10.1007/s00253008-1385-6

Part III Toxins Characterization

Chapter 5 RNA-Sequencing of Snake Venom Glands Khin Than Yee, Olga Vasieva, and Ponlapat Rojnuckarin Abstract Next-generation sequencing (NGS), particularly RNA-sequencing (RNA-Seq) technique, allows detection and quantification of different RNA transcripts in a tissue sample, and in our case toxin transcripts from snake venom glands. Using this approach, novel toxin transcripts can be detected and abundancies of different isoforms of each toxin measured. The analytical pipeline can be briefly outlined as follows. Isolation of mRNA from tissue under RNase-free condition is essential to keep mRNA intact before sequencing. After mRNA fragmentation, the adapters are added to both ends of the fragments to synthesize complementary cDNAs. The obtained cDNA library is then sequenced on Illumina HiSeq 2000 platform. Quality of millions of reads produced from the NGS is checked and the sequences corresponding to the adapters and low-quality reads are removed. Subsequently, the NGS data are subjected to the workflow of de novo assembly, quantification of expression levels, annotation of transcripts, and identification of ORFs, signal peptides, structurally conserved domains, and functional motifs. In this report we describe the listed methodological steps and techniques in details and refer to the platforms and software that may be adopted for similar studies. Key words Next-generation sequencing, RNA-Sequencing, Snake venom glands, Toxin transcripts, De novo assembly

1

Introduction Next-generation sequencing (NGS) of Myanmar Russell’s viper (MRV) venom gland was performed to investigate composition of the transcriptome of the venom gland. RNA-sequencing (RNA-Seq) allows us to determine the primary sequence and a relative abundance of each transcript [1]. In terms of quality, NGS provides a better sequencing depth and coverage than those of expressed sequence tag (EST) sequencing by Sanger technology [2]. It allows discovery of the low-abundance toxin transcripts in snake venom gland. Noteworthy, the de novo assembly of fulllength transcripts from RNA-Seq data can be performed without a reference genome [3] that is of importance in cases of non-model organisms, such as Russell’s viper. The usual procedure of a transcriptome analysis starts from messenger RNA extracted from the

Avi Priel (ed.), Snake and Spider Toxins: Methods and Protocols, Methods in Molecular Biology, vol. 2068, https://doi.org/10.1007/978-1-4939-9845-6_5, © Springer Science+Business Media, LLC, part of Springer Nature 2020

87

88

Khin Than Yee et al.

total RNA of the snake venom glands. The following procedures can vary as there are different developed platforms and approaches that can be adapted to the requirements of a specific experiment. In our study, the messenger RNA-sequencing libraries were prepared using the TruSeq RNA sample preparation kit (Illumina) before sequencing on the Illumina HiSeq2000 platform. The raw data from the platform was trimmed by Trimmomatic application and the quality was checked by FastQC before and after trimming process. Subsequently, de novo reconstruction of transcriptomes was done with the Trinity software. The quantification of transcripts from RNA-Seq data was estimated with the application of RSEM software. Further bioinformatics analysis of toxin transcripts has been performed via bioinformatics pipeline—comprising Blastn, Blastx, ORF, and SignalP 4.1 Server.

2

Materials For venom gland dissection, all utensils should be autoclaved and the working bench cleaned with soap and water. Glands are to be dissected under aseptic condition. The tissue should be kept in RNAlater solution at 4
C overnight and to 80
C until the RNA isolation procedure (see Note 1). For RNA extraction, autoclaved mortar, pestle, forceps, and scissors are to be used. All solutions are to be prepared in RNase-free plasticware or RNasefree glassware. The tissue sample is to be kept on ice during the preparation and stored at 80
C.

2.1 Venom Gland Extraction

1. Chloroform. 2. RNAlater solution (Ambion Inc., Canada, USA). 3. Ethanol: 75% Ethanol in distilled water. 4. Soap and water.

2.2 Total RNA Extraction

1. RNase® AWAY (Carl Roth GmbH + Co.KG, Karlsruhe, Germany). 2. TRIzol® LS Reagent (Ambion, Life technologies, Carlsbad, California). 3. Chloroform. 4. Isopropanol. 5. Ethanol: 75% Ethanol in DEPC-treated nuclease-free water.

2.3 mRNA Extraction (PolyATtract® mRNA Isolation System II, Promega)

1. Biotinylated Oligo(dT) Probe (50 pmol/μL). 2. 20 SSC solution (3 M NaCl, 0.3 M Na3 citrate·2H2O, pH 7.0): 0.5 SSC solution and 0.1 SSC solution in RNase-free water.

RNA-Sequencing of Snake Venom Glands

89

3. Streptavidin MagneSphere® Paramagnetic Particles. 4. Nuclease-free water. 5. mRNA user tubes. 6. MagneSphere® Magnetic Separation Stand for 12  75 mm tubes. 2.4 Precipitation and Concentration of mRNA

1. 3 M Sodium acetate, pH 5.2.

2.5 Quality Check of NGS Data

1. FastQC (http://www.bioinformatics.babraham.ac.uk/pro jects/fastqc).

2. Isopropanol. 3. 75% Ethanol.

2. Trimmomatic (0.32) page¼trimmomatic). 2.6 De Novo Assembly and Quantification of Transcripts 2.7 Annotation of Transcripts

(http://www.usadellab.org/cms/?

1. Trinity (r20140717) (http://trinityrnaseq.sourceforge.net/). 2. RSEM (1.2.15) (http://deweylab.biostat.wisc.edu/rsem/).

1. Blastn (https://blast.ncbi.nlm.nih.gov/blast/Blast.cgi?PRO GRAM¼blastn&PAGE_TYPE¼BlastSearch&LINK_ LOC¼blasthome). 2. Blastx (https://blast.ncbi.nlm.nih.gov/blast/Blast.cgi? PROGRAM¼blastx&PAGE_TYPE¼BlastSearch&BLAST_ SPEC¼&LINK_LOC¼blasttab&LAST_PAGE¼blastn).

2.8 Analysis of Toxin Transcripts

1. ORFfinder (https://www.ncbi.nlm.nih.gov/orffinder/). 2. Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/). 3. SignalP 4.1 Server (http://www.cbs.dtu.dk/services/SignalP/).

3

Methods Carry out gland dissection in a sterile environment and as fast as possible without affecting the quality of dissection. Carry out all the procedure in RNase-free working place (cleaned by RNase AWAY solution and 100% ethanol) at room temperature unless otherwise specified.

3.1 Venom Gland Extraction

1. Handle the snake gently and anesthetize with chloroform. 2. Monitor the snake for at least 20 min to the point of loss of reflex. 3. Place the anesthetized snake in the clean dissection tray.

90

Khin Than Yee et al.

4. Disinfect the surface of the head of the snake quickly with alcohol spray and wipe dry (see Note 2). 5. Identify the venom gland which lies in the temporal region behind the eye. 6. Make a skin-deep incision at the temporal region and remove the skin carefully to expose the underlying venom gland (see Note 3). 7. Remove the gland from the head (see Note 4). 8. Trim the gland swiftly into tiny pieces (dimension ”. 2. Submit the job (see Note 17).

4

Notes 1. The experiment plan must be approved by the local Animal Care and Use Committee. 2. Be careful of reflex bite from the severed head. Use a metallic clip, piece of cloth, or rubber band to constraint the mouth part if necessary. 3. A venom gland is usually round or spindle shaped with a grayish, slightly glistening capsule. It may be possible to observe the surrounding musculature and to track the duct. 4. Muscle or connective fibers surrounding the gland should not be included. Blood stains, if present, can be rinsed off briefly with autoclaved pure water or RNAlater solution. 5. It is advisable to put 3–5 cubes of tissue (150 bp) since this additional information can be useful for assembly, and paired-end reads can be merged by such programs such as PEAR (Paired-End reAd mergeR) [20] or FLASH (Fast Length Adjustment of SHort reads) [21] to create overall longer reads that also improve assembly. 3. Strand information: strand origin of a read. In order to quantify gene expression accurately, it is important to retain the strand specificity of origin for each transcript. This will allow one to identify from which overlapping gene the RNA transcript has originated. There are many steps to produce a high-quality assembly, but the assembly has many downstream applications (refer to https:// omicstools.com/rna-seq-categogy), such as evaluating toxin gene expression, selection, or use of the predicted translated products as custom databases for protein identifications (Fig. 2), so accuracy should be a major goal. Command examples for some of the programs discussed in the preceding text are given (Box 1), but individual documentation for each program should be referenced.

Fig. 2 Protocol overview for venom gland transcriptomics. Protocol overview shows each step to be performed for venom gland transcriptomic work, including the processing of next-generation sequencing reads, de novo transcriptome assembly, gene expression determination, toxin transcript identification, positive selection analysis, and integration of high-throughput proteomics with transcriptomics. Procedures discussed in the text are indicated by section numbers (red boxes)

108

Cassandra M. Modahl et al.

Box 1 Abridged Pipeline Example Commands. A few command examples are given; documentation for each program should be referenced for all command arguments and parameters, and only examples are provided. All CPU/thread arguments should be modified based on computing resources: ############################## ### FASTQC example command ### ############################## SYNOPSIS Usage: fastqc seqfile1 seqfile2 .. seqfileN fastqc [-o output dir] [--(no)extract] [-f fastq|bam|sam] [-c contaminant file] seqfile1 .. seqfileN fastqc

RAWDATA_PAIR_1.fastq.gz

RAWDATA_PAIR_2.fastq.gz

-o

OUTPUT_DIRECTORY ################################### ### TRIMMOMATIC example command ### ################################### SYNOPSIS Usage: PE

[-threads

]

[-phred33|-phred64]

[-trimlog

] [-quiet] [-validatePairs] [-basein | ] [-baseout | ] ... or: SE

[-threads

]

[-phred33|-phred64]

[-trimlog

] [-quiet] ... java -jar trimmomatic-0.35.jar PE -threads 4 -phred33 RAWDATA_PAIR_1.fastq.gz

RAWDATA_PAIR_2.fastq.gz

OUTPUT_R1-unpaired.fastq

OUTPUT_R1-paired.fastq

OUTPUT_R2-paired.fastq

OUTPUT_R2-

unpaired.fastq ILLUMINACLIP:TruSeq3-PE-2.fa:2:40:15 SLIDINGWINDOW:4:15 LEADING:20 TRAILING:20 MINLEN:50 HEADCROP:9 ############################ ### PEAR example command ### ############################ SYNOPSIS Usage: pear

(continued)

Exploring Toxin Evolution

109

Standard (mandatory): -f, --forward-fastq Forward paired-end FASTQ file. -r, --reverse-fastq Reverse paired-end FASTQ file. -o, --output Output filename. pear

-f

INPUT_R1-paired.fastq

-r

INPUT_R2-paired.fastq

-o

OUTPUT_NAME ############################# ### FLASH example command ### ############################# SYNOPSIS Usage: flash [OPTIONS] MATES_1.FASTQ MATES_2.FASTQ flash [OPTIONS] --interleaved-input (MATES.FASTQ | -) flash [OPTIONS] --tab-delimited-input (MATES.TAB | -) flash -o OUTPUT_PREFIX -t 5 INPUT_R1-paired.fastq INPUT_R2-paired. fastq -r 140 -f 350 -s 50 -d OUTPUT_DIRECTORY ############################### ### TRINITY example command ### ############################### SYNOPSIS #Usage: # --seqType :type of reads: (’fa’ or ’fq’) # # --max_memory :suggested max memory to use by #Trinity where limiting can be enabled. (jellyfish, sorting, etc) #provided in Gb of RAM, ie. ’--max_memory 10G’ # # If paired reads: # --left :left reads, one or more file names #(separated by commas, no spaces) # --right :right reads, one or more file names #(separated by commas, no spaces) # # Or, if unpaired reads: # --single :single reads, one or more file names, #commadelimited (note, if single file contains pairs, can use #flag: -run_as_paired ) # # Or,

(continued)

110

#

Cassandra M. Modahl et al.

--samples_file

tab-delimited

text

file

#indicating

biological replicate relationships. #ex. #cond_A cond_A_rep1 A_rep1_left.fq A_rep1_right.fq #cond_A cond_A_rep2 A_rep2_left.fq A_rep2_right.fq #cond_B cond_B_rep1 B_rep1_left.fq B_rep1_right.fq #cond_B cond_B_rep2 B_rep2_left.fq B_rep2_right.fq # Trinity --seqType fq --max_memory 50G --left INPUT_R1-paired.fastq. gz --right INPUT_R2-paired.fastq.gz --CPU 6 --full_cleanup --min_contig_length 100 --verbose ############################## ### CD-HIT example command ### ############################## SYNOPSIS Usage: cd-hit-est [Options] cd-hit-est -i INPUT_SEQUENCE -o OUTPUT_SEQUENCE -c 1 -n 8 ############################## ### BLAST+ example command ### ############################## SYNOPSIS Usage: blastx [-h] [-help] [-import_search_strategy filename] [-export_search_strategy

filename]

[-task

task_name]

[-db

database_name] [-dbsize num_letters] [-gilist filename] [-seqidlist filename] [-negative_gilist filename] [-entrez_query entrez_query] [-db_soft_mask

filtering_algorithm]

[-db_hard_mask

filtering_algorithm] [-subject

subject_input_file]

[-subject_loc

range]

[-query

input_file] [-out output_file] [-evalue evalue] [-word_size int_value] [-gapopen open_penalty] [-gapextend extend_penalty] [-qcov_hsp_perc float_value] [-max_hsps int_value] [-xdrop_ungap float_value] [-xdrop_gap float_value] [-xdrop_gap_final float_value] [-searchsp int_value] [-sum_stats

bool_value]

[-max_intron_length

length]

[-seg

SEG_options]

(continued)

Exploring Toxin Evolution

111

[-soft_masking soft_masking] [-matrix matrix_name] [-threshold float_value] [-culling_limit int_value] [-best_hit_overhang

float_value]

[-best_hit_score_edge

float_value] [-window_size int_value] [-ungapped] [-lcase_masking] [-query_loc range] [-strand strand] [-parse_deflines] [-query_gencode int_value] [-outfmt format] [-show_gis] [-num_descriptions int_value] [-num_alignments int_value] [-line_length line_length] [-html] [-max_target_seqs

num_sequences]

[-num_threads

int_value]

[-remote] [-comp_based_stats compo] [-use_sw_tback] [-version] blastx -query INPUT_SEQUENCE -db nr -max_target_seqs 3 -num_threads 8 -outfmt ’6 std stitle’ -out Blastx_nr_outfmt6 ############################## ### RSEM example command ### ############################## SYNOPSIS Usage: rsem-prepare-reference

[options]

reference_fasta_file

(s) reference_name rsem-calculate-expression

[options]

upstream_read_file

(s) reference_name sample_name rsem-calculate-expression [options] --paired-end upstream_read_file(s) downstream_read_file(s) reference_name sample_name rsem-calculate-expression [options] --alignments [--paired-end] input reference_name sample_name rsem-prepare-reference [options] INPUT_SEQUENCE INPUT_SEQUENCE. rsem.ref rsem-calculate-expression --paired-end -p 5 --bowtie2 INPUT_R1paired.fastq

INPUT_R2-paired.fastq

INPUT_SEQUENCE.rsem.ref

INPUT_SEQUENCE.rsem.results

3.3.2 NGS Data Quality Checks

The first step upon receiving sequencing reads is to conduct initial quality checks (QC). These QC results can be obtained by loading the read fastq files into the Java program FastQC [18]. This widely used quality control tool for high-throughput sequence data provides a modular set of analyses that can give an impression of potential problems during the library construction and the sequencing run. The following parameters need to be evaluated,

112

Cassandra M. Modahl et al.

and reads filtered to match these criteria, to be used reliably in the assembly: 1. Overall read quality should be greater than a quality score of 20. 2. Adapter contamination should be absent. 3. Proper read length should be at least 36 bp. There are several available open-source tools that can be used to remove low-quality reads and adapter contamination, but Trimmomatic [19] is a commonly used software for this purpose and performs well in that a sliding window is used to evaluate base quality instead of just read quality averaged. Base quality is reported in a Phred-like score, which is the log value of the error probability (probability of incorrect base calling ¼ 10Q/10; Q ¼ Phred score). A quality score (Q) of 20 indicates that there is a 1 in 100 chance that the base call is incorrect. Because low-quality bases are observed on read ends, when these are removed, a minimum length is also set to keep reads long enough to be informative for the assembly. The Trimmomatic package also contains common adaptor sequences that can be selected for removal. These qualitycontrolled and adaptor-removed filtered fastq files should then be checked again by FastQC before they are used as input for transcriptome assembly. Paired-end reads can also be merged with programs such as PEAR [20] or FLASH [21] and then used as input into assemblers such as Extender, leveraging longer sequence lengths. However, some assemblers do require the paired-end read information for contig construction. Paired-end read merging can be used for assembling small transcripts, as some animal toxins can be quite small, such as those from arthropod venoms. 3.3.3 De Novo Transcriptome Assembly

Venom gland transcriptomes are notoriously difficult to assemble because of the abundance of transcript isoforms and the high levels of expression of these isoforms. However, it is important that toxin transcripts are properly assembled because there is exceptional functional diversity in many toxin families, and minor differences in sequence can greatly alter binding and overall activity. Trinity [17] is currently one of the most popular de novo RNA-seq assemblers, with over 2500 citations. Trinity partitions RNA-seq reads into many independent de Bruijn graphs and with parallel computing reconstructs transcripts from these graphs. Three different software modules are used in Trinity contig construction: Inchworm, Chrysalis, and Butterfly. Inchworm assembles reads into unique sequences using a k-mer-base approach, where each read is partitioned into smaller nucleotide strings of k length. Next, Chrysalis clusters related reads and constructs a de Bruijn graph for each cluster of related sequences. Finally, Butterfly

Exploring Toxin Evolution

113

analyzes the de Bruijn graphs and read pairings to report all plausible transcript sequences. Assembly run times are quite quick, usually completed within 24 h (approximately one-half to 1 h per million reads). The Trinity software package contains many useful Perl scripts, such as those for transcript quantification, differential expression, coding region identification, translation (Transdecoder; https://github.com/TransDecoder/TransDecoder/wiki), and annotation (Trinotate pipeline; https://trinotate.github.io/). However, it has also been noted that Trinity does not perform well in distinguishing between highly similar paralogous or homologous transcripts [51], and because this is often the case with toxins Trinity has been reported to miss toxin transcripts during assembly, or to assemble only partial sequences [52]. Trinity has also been reported to struggle with assembling highly expressed transcripts [53], which is also often the case for toxin genes expressed in the venom gland [54]. These limitations are likely due to the smaller k-mer size (a fixed k-mer of 25) used for Trinity assemblies, because small k-mers are better for assembling minimally expressed genes while larger k-mers perform better for abundantly expressed genes [55]. Extender [22], a Java program, was designed to improve upon the issues observed using Trinity, and other de Bruijn graph assemblers such as ABySS [56] and Velvet [57], by utilizing a hashtag table and extending contigs based upon long overlaps. Extender also has faster run times, comparable to Trinity, but has smaller RAM requirements. A larger k-mer size can be used for Extender assemblies and because of an overlap versus a de Bruijn graph algorithm, there are fewer alternative paths and therefore less assembly errors are introduced. Extender has been used for multiple venom gland assemblies and performs well when assembling highly expressed transcripts within a venom gland [58]. Reads are first merged with PEAR [20] or FLASH [21] and then used as input into Extender. Extender also performs best when a large number of reads are used, >30 million, but it does produce fewer overall contigs in comparison to Trinity, likely excluding complete transcript diversity. Another assembler, VTBuilder [23], was also designed to address the issues observed with assembling multi-isoform transcriptomes, making it ideal for venom gland transcriptomes. The VTBuilder assembly algorithm is more similar to reference-guided genome assemblies. Reads are partitioned and a guide sequence is generated from these reads. Reads are then mapped as scaffold-like alignments and reconstructed as contigs representing the transcript isoform diversity present. Unfortunately, the current VTBuilder version only allows up to 5 M reads to be used for assemblies and only works effectively with read lengths equal to or greater than 250 bp. With shorter reads, it has been noted as having

114

Cassandra M. Modahl et al.

performance equal to if not lower in comparison to Trinity when assembling snake venom gland transcriptomes and an RNA spike-in (RNA transcripts of known sequence and quantity used as a control) [13]. Overall, given that each assembler has its own advantages and disadvantages, using multiple assemblers might be the best approach to achieve total and accurate transcript diversity. This is quickly becoming the preferred method of transcriptome assembly, considering that a transcriptome is a heterogeneous mixture of transcripts of different sizes, GC content, complexity regions, expression levels, etc., and one assembler algorithm is likely not best for every transcript. There is also an advantage to generating multiple assemblies with different parameters, such as k-mer values, because the optimal k-mer value for an assembly will depend on the read length, sequencing depth, and read error rate [55], especially in cases where transcript abundances differ tremendously, as mentioned above. A disadvantage to using many different k-mer values is that this has been found to increase the number of fusion/ chimeric transcripts when compared to single k-mer methods [59]. Therefore, multiple assemblers and multiple parameters should be explored in addition to quality control checks. There are some pipelines that include such programs as CAP3 [60] that have been used to merge assembled contigs from multiple assemblers into a final transcriptome set. This DNA sequence assembly program constructs multiple sequence alignments between contigs and then generates a consensus sequence. It can end up merging contigs from separate isoforms, emphasizing again the importance of proper assemble quality control checks. Programs such as TransRate (http://hibberdlab.com/transrate/) can assess the quality of a transcriptome assembly [61]. In order to evaluate assembly performance, several metrics such as N50, average contig length, total assembled nucleotides, maximum contig length, total number of contigs, and number of singletons have largely been taken into consideration [62]. However, which metrics actually reveal assembly quality is unclear, and standard quality metrics commonly used are repurposed from genome assembly. Further, because of the redundancy of using multiple assemblers, both redundancy removal and selection of the truest set of transcripts will be required. There are several redundancy removal software available, such as the CD-HIT suite software [26] or Exonerate [25], and script pipelines like those provided by EvidentialGene [24] identify high-quality transcripts. The EvidentialGene script pipeline has been shown to perform optimally when dealing with multiple transcriptome assemblies that include duplicated gene copies, and this is a feature of venom gland tissue transcriptomes. Moreover, the EvidentialGene pipeline has been found to be ideal for working with multiple transcript isoforms because transcripts are pooled into one super-set of sequences and then the “best” set of transcripts from this set is selected based on the coding

Exploring Toxin Evolution

115

sequence and protein length, emphasizing transcript coding potential. 3.3.4 Toxin Gene Identification and Expression Quantification

The best way to evaluate the quality of the final overall assembly is by the identification of full-length transcripts for toxins known to be present within the venom. In this sense, as mentioned above, the Trinity software package provides a Perl script based on sequence homology that could be used in order to decipher which toxinidentified transcripts expand throughout the entire length of protein sequence. Hence, evaluating the quality of coding sequences, such as if a full-length transcript starts with a methionine and ends with a stop codon, is better than relying on a value like “N50,” which is not very relevant to transcriptome assemblies, because a higher N50 value and the presence of many long contigs can be the result of misassemblies. However, it should also be noted that excluding partially assembled transcripts can lead to underestimation of venom complexity, as partial transcripts can contain valid variants. BLAST+ (Basic Local Alignment Search Tool) [27], which is run from the command line of a computer (accessed through the terminal for Unix-like operating systems), is commonly used for toxin annotation and is based upon database searches. The databases used include the nonredundant protein database available on NCBI (National Center for Biotechnology Information) or the UniProt database. Custom databases, such as a collection of venom protein sequences, can be created and have been found to be equally successful at the identification of toxin sequences, as long as there exists homology to known toxins. There are also a few specific toxin databases that have been assembled (reviewed in [16]). It should be noted that it is possible to find toxin identities using BLASTn that might be missed using BLASTx or BLASTp. This was observed in the case of the Boiga irregularis venom gland transcriptome, where Trinity assembled many partial transcripts that showed untranslated region transcript bias and were unable to be identified with BLASTx, but were identified as toxin transcripts with BLASTn [14]. In this sense, given the fact that mobile elements such as saurian SINEs and LINEs have been largely characterized in all major lineages of squamate reptiles, it is best to mask repeat nucleotide sequences with Repeat Masker (addressing http://www. repeatmasker.org). The program makes use of Repbase (http:// www.girinst.org/repbase/), a comprehensive database of repetitive element consensus sequences, reducing running times of the BLAST annotation process. BLAST+ can have very long run times, and with a full transcriptome (20,000 plus contigs) and using a single workstation it can easily run for a month (if not longer) to generate results. A way to speed this up, besides assigning more processing cores to allow

116

Cassandra M. Modahl et al.

for parallel computing, is to split up files and run them separately, and this is recommended. The program Diamond [28] has a much faster algorithm, faster than the stand-alone BLAST+ by about 20,000 times, and is highly recommended for BLASTx or BLASTp searches. Regarding the annotation process, within databases submissions are sometimes given the identification of “hypothetical protein,” “transcribed mRNA,” or even a mis-annotated description; some may not have complete identities when they are submitted, and some might even be partial sequences. Therefore, it is best to report at least the top three BLAST hits in case the top hit given has one of these non-informative labels or is incomplete. A filtering round using a list of keywords (including the acronyms of all known toxin protein families described so far) to distinguish putative snake venom toxins from non-toxin (ribosomal, mitochondrial, nuclear, etc.) proteins should be carried out over the BLAST hit results. The main issue with using previous toxin datasets on an identity search is that only toxin sequences similar to known toxins are identified. Other programs, such as HMMER [63] with the Pfam database [64] or InterPro [65], are sequence analysis programs that use hidden Markov models to identify domains for unknown proteins, and these can be useful to find unknown or novel toxins. Venom components are secreted cell products and therefore a signal peptide sequence should be present. This is a common criterion used to identify potential toxins and is accomplished by evaluating translated transcripts for signal peptides with SignalP [29]. SignalP can be downloaded and run from a command line for large FASTA files with many sequences. Protein sequences can also be evaluated for transmembrane domains, which are suggestive of non-secreted cell products, and this is done through the use of the program tmHMM [30] that employs hidden Markov models to identify membrane-bound protein regions. It is likely that if a protein has membrane-bound regions, it is not a venom component; however, there are no unequivocal certainties, because these proteins could be posttranslationally processed or in the case of a signal peptide there are other mechanisms of cellular export observed as well [66]. To identify a venom protein transcript confidently, venom gland transcriptomics must be combined with venom proteomics (though posttranscriptional regulation may result in no translated product). Transcript abundances are usually determined based on reads mapping to the de novo-assembled transcriptome and provide within-sample normalization for feature-length and library-size effects. They are reported as RPKM or FPKM (reads/fragments per kilobase of exon model per million mapped reads) [67] and TPM (transcripts per million), which is currently the most accepted quantification method. In order to estimate transcript abundances

Exploring Toxin Evolution

117

from full-length transcripts, several software packages have been developed. One of the most commonly used software packages for this is RSEM (RNA-Seq by expectation-maximization) [31]. This software package uses Bowtie/Bowtie2 [32] as the read aligner, utilizing a Burrows-Wheeler index to keep its memory requirements small. Because multiple transcript isoforms are present for many toxin genes, multi-mapping reads are frequently observed. One should note that for mapping programs like Bowtie2 (the read alignment program used for RSEM quantification), the search for alignments for a given read is randomized. This means that if Bowtie2 encounters a set of equally parsimonious alignments during mapping, one of these alignments is randomly picked. This allows for quick transcript quantification (RSEM run times are usual less than 48 h on a single workstation, depending on read numbers), but any transcript isoform quantification should be seen as a measure of relative abundance only. 3.4 Toxin-Positive Selection

Two primary modes of toxin evolution have been proposed: purifying and positive selection [68, 69]. It has been suggested that positive selection is the dominant driver of snake venom evolution [70], especially for highly expressed venom protein transcripts [13]. Additionally, it has been observed that abundant venom protein superfamilies experience weaker selective constraints because of multiple gene copies, allowing for the accumulation of deleterious mutations, and therefore also neutral evolution [71]. Even though there are multiple models that can be used to examine selection pressures, it must be noted that for large venom protein families that exhibit structural and functional diversity, toxin evolution can be complex. The most common method of selection evaluation, and one of the easiest to perform, is analyzing toxin transcripts for positive selection. This method examines single-nucleotide polymorphisms (SNPs) within codons, identifying if nonsynonymous or synonymous substitutions are occurring more frequently between homologs. SNPs have been well documented in venom protein transcripts and linked to toxin functional diversification [72]. The ratio of nonsynonymous to synonymous substitutions, ω, can be used to determine if selection is acting on the overall protein and/or specific regions. Values of ω < 1 are suggestive of negative purifying selection, ω ¼ 1 is suggestive of neutral evolution, and values ω > 1 indicate positive selection. There are several positive selection models that can be used. Branch models allow the ω ratio to vary among branches in a phylogeny to detect positive selection acting on particular lineages [73], and site models allow ω ratios to vary for sites (codons) [74]. There are also models that incorporate both branch and site evaluations, allowing ω to vary for both sites within the protein and across branches on the tree to detect positive selection affecting a

118

Cassandra M. Modahl et al.

few sites along particular lineages (foreground branches) [75, 76]. The most frequently used software for positive selection analysis is PAML (phylogenetic analysis by maximum likelihood) [37], specifically the codeml module. Usually a series of models within PAML are run, and model likelihood values are compared. Toxin evolution evaluation has been incorporated into venom gland transcriptome assembly publications because of the need of transcript sequences to determine selection occurring for a toxin family. It is of interest to identify which codons experience increased mutation rates since positive selection has indeed been linked to toxin-active sites and molecular surface residues [6, 72]. To set up sequences for a codeml analysis, it is ideal if orthologous toxin sequences are used to compare sequence variation across species and identify which coding regions are more variable. However, identifying orthologous sequences can be particularly challenging with venom toxins. Large multi-isoform toxin families exist because gene duplications result in multiple paralogs, and different paralogs can be evolving under different selection pressures. Correct orthologous sequences between species must be identified from these gene families. Using BLAST identities, especially reciprocal BLAST outputs, potential orthologous toxin genes might be able to be identified. Once a set of toxin sequences are chosen, PAML will need a nucleotide alignment file and tree file as input for codeml. The alignment will need to be in in PHYLIP format with sequence names identical to those present in the tree file. Each sequence also needs to have the same number of characters. The tree file will need to be in Newick format. Nucleotide models used for tree construction will not matter for PAML, but users should make sure that it is appropriate for their data set. PartitionFinder, Jmodeltest, or MEGA [36] can be used for model selection. Tree construction can be completed using either a maximum likelihood or a Bayesian approach. A suggested open-source pipeline to use is either Aliview [33] or Jalview [35] for the generation of a multiple sequence alignment with either a Clustal or a MUSCLE alignment algorithm, and SeaView [34] to construct a maximum likelihood tree once nucleotide model selection has been performed. The alignment and tree files will need to be designated in the codeml control file, as well as the resulting output file name and all models to be run for comparisons. Some commonly used PAML models include M0 (one ratio), M1a (neutral), M2a (selection), M3 (discrete), M7 (beta), and M8 (beta&ω). Model M0 estimates a constant ω rate and is compared to model M3, which allows ω to vary across sites. M1a is a model of neutral evolution, where all sites are assumed to be under either negative or neutral selection and is compared to M2a, a model of positive selection. A Bayes empirical Bayes (BEB) approach is useful for identifying specific amino acids under positive selection by

Exploring Toxin Evolution

119

calculating the posterior probabilities of a particular amino acid belonging to a given selection class (neutral, conserved, or highly variable). These BEB calculations are performed with the M8 model, run in comparison to the M7 model. Once likelihood values are generated for each model, comparisons can be made with negative twice the difference in log likelihoods between each model compared to a χ2 distribution. The length of time it takes to run PAML codeml is dependent on sequence number and models used, but it is usually completed within 24 h and can be executed easily on a desktop or laptop computer. Another software that has been successfully used for toxin selection analysis is HyPhy [77]. HyPhy hypothesis testing using phylogenies is similar to PAML in that it carries out likelihoodbased analyses on multiple alignments to find rates and patterns of sequence evolution. HyPhy can be executed from the DataMonkey server [38]. Tests for positive, negative, and episodic selection can all be performed on the DataMonkey server [78]. 3.5 High-Throughput Proteomics Integration

Venom gland transcriptomes will then be used as databases for locus-specific matching of proteomic data. Although some top-down proteomics strategies are being developed for proteome profiling, characterization of venoms is usually completed with a bottom-up tandem mass spectrometry (MS/MS) approach, where proteins are first digested with proteases such as trypsin (most commonly used), chymotrypsin, or Glu-C, and then MS/MS produces spectra of fragmented singly charged peptide ions that can be matched to databases for protein identification (peptide mass fingerprinting) or can be used for de novo sequence determination [79]. Collision-induced dissociation (CID) is the most popular MS/MS technique for this type of analysis. This technique creates a series of backbone fragmentations at the peptide bond, resulting in b- and y-fragment ions, and using Mascot, SEQUEST, or other search engines, databases are searched to identify unknown proteins based on their peptide fragment spectra. However, MS/MS peptide identification relying on available online protein sequence databases can overlook unique protein isoforms and/or be unsuccessful at recognizing novel toxins. Animal venoms can contain many different peptide and protein isoforms, and given that venoms experience high levels of variation even within species, such as ontogenetic [80–83] and regional venom variation [84–86], the use of public databases can be disadvantageous when attempting to characterize unexplored venoms. Venom compositional variation has direct implications for antiserum development and efficacy, and proper identification of toxin diversity is critical. Therefore, the use of an individual or speciesspecific transcriptome can greatly improve venom proteomic profiling.

120

Cassandra M. Modahl et al.

There are several programs that allow for the input of a custom protein database, such as a translated venom gland transcriptome, as a FASTA file. Some of the more popular software that have this capability are listed in Subheading 2. Another important consideration when using custom databases, such as a species-specific transcriptome, is that there could be mis-assemblies or missing transcripts within these databases, and therefore searches against publicly available databases also are still advisable. Peptide to translated transcriptome matches assigned by these tools can also have false positives, and therefore a false discovery rate (FDR) metric is often used for confidence assessment [87]. False-positive screening is performed with the inclusion of a decoy database, where incorrect “decoy” sequences are added to the search space. This decoy database can be useful for the design of FDR filtering criteria [88]. An integrated transcriptomics and proteomics (venomics) approach is ideal for not only more accurate and complete identification of venom proteins, but also for better protein quantification [89]. There are several label-free methods of MS/MS quantification of venom components, such as normalized spectral abundance factors (NSAF) [89–91], which normalizes for protein length, or the use of an internal standard of known concentration that is then used to determine unknown concentrations of proteins based upon peptide intensities [92], similarly used for iBAQ [93]. The use of a species-specific or even individual-specific translated transcriptome database can aid in the quantification of venom components, such as providing exact protein sizes for NSAF calculations. Some proteomic programs can also generate their own quantification numbers, such as the emPAI (Exponentially Modified Protein Abundance Index) number [94] from ProteinPilot and Mascot. Additionally, other researchers have relied on the use of chromatogram peak areas for venom component quantification and perform a reversed-phase high-performance liquid chromatography (RP-HPLC) separation before the digestion and identification of proteins [95]. In cases where peaks consist of multiple proteins, gel densitometry is used to determine the abundance of different proteins within a single peak. It is also important to note that although the translated transcriptome is ideal as a species-specific database for MS/MS peptide identifications, there is not always a quantitative correspondence between the transcriptome and proteome. Transcripts from an assembled transcriptome can be used to obtain the full amino acid sequence of a protein. Using proteomic methodologies (such as N-terminal sequencing and MS/MS de novo sequence determinations from many peptide fragments) to acquire full amino acid sequences of proteins can be labor intensive and expensive. Additionally, with these approaches, complete protein sequences are not guaranteed, as some proteins are N-terminally blocked, do not exhibit sequence for protease digestion, or do not ionize well for MS/MS.

Exploring Toxin Evolution

121

A combined transcriptomic and proteomic approach is often necessary to identify toxins, but the presence of a transcript alone does not mean that it is a translated and secreted venom component [96]. Because the basic definition of a venom is as a secretion, it is therefore of great importance that venom proteomes are characterized, in addition to venom gland transcriptomes, to determine which transcripts belong to secreted venom components. Venom proteins originated from homologs that performed non-venomrelated, physiological functions within tissues [97], and misidentification of these physiological proteins and peptides as toxins could distort our view of toxin evolution, especially when they are included in cladistics and selection analyses. This integration of transcriptomics and proteomics improves the accuracy of either approach used alone.

4

Notes 1. Venom protein-specific primer needs to be designed from venom protein transcripts. The best way to accomplish this, if the target sequence is unknown, is by performing a multiple sequence alignment with a collection of similar transcript sequences. Venom protein superfamilies tend to have conserved signal peptide regions and this region is ideal to design primers to target multiple venom protein transcripts within a single superfamily. It is best to incorporate some degenerate nucleotide bases, such as Y (designated for C or T nucleotides) and W (for A and T nucleotides), to improve amplification of all transcripts within a superfamily. Refer to specific instructions that companies have designated for ordering degenerate bases. Usually, 1–4 degenerate bases should be used; more degenerate bases will result in nonspecific binding and amplification. It is also best to run PCR products using agarose gel electrophoresis and excise the band belonging to the estimated transcript size, as this will also help avoid nonspecific transcripts. Modahl and Mackessy (2016) list several primers that have been successfully used to amplify multiple transcript isoforms within a single snake venom protein superfamily; this publication also has details regarding primer design and PCR for 30 RACE. 2. Make sure that X-gal, ampicillin, and IPTG are added after autoclaving agar, and when agar has cooled to approximately 50  C. 3. RNA is degraded by RNases that occur in the environment, on skin, and in bacteria or mold that may be present on airborne dust particles. RNase contamination is prevented by always wearing gloves, only using plasticware that is labeled “RNasefree” (treat any glassware with RNase inhibitors), using filtered

122

Cassandra M. Modahl et al.

pipette tips and micropipettes that are designated only for RNA work, and cleaning the work area with RNase inhibitors, such as RNase Away (ThermoFisher Scientific). Also, make sure that all reagents used are molecular grade and are only used for RNA work (this includes water, which must be treated beforehand with DEPC). It is better to be overly cautious when working to avoid environmental RNases than to be neglectful and end up with degraded RNA. Next-generation sequencing technology in particular requires high-quality RNA for library input, and some sequencing centers will even refuse to sequence RNA that falls below a RNA quality threshold. RNA is also unstable, and experiments should be planned to avoid multiple freeze-thaw cycles. RNA should be reversetranscribed as quickly as possible to avoid degradation. Any long-term storage of RNA should be done at 80  C and any tissue that will be used later for RNA isolation should also be stored at 80  C but within a RNAlater stabilizing buffer for best preservation. If tissue samples will be used within 1–2 months and are small, such as venom glands from arthropods, they can be directly collected and stored in TRIzol. This is actually recommended considering that it can be hard to remove small samples from RNAlater. Isolated RNA should be kept on dry ice during any transport. 4. In the case of total RNA isolated from rear-fanged snake venom, a DNase I digestion (amplification grade; Invitrogen) must be performed to remove all traces of DNA before beginning the 30 RACE procedure. Venoms collected from rearfanged snakes tend to have more DNA contamination that will interfere with later steps. 5. Touch-down PCR is used for this procedure. This means that the first set of repeated cycles has a higher annealing temperature to encourage specific primer binding, and the remaining repeated cycles have a lower annealing temperature to increase overall copy number. This is different than the nested PCR that is described in the manual for the 30 RACE system for rapid amplification of cDNA ends (ThermoFisher Scientific). The PCR method detailed in this chapter and modified from the ThermoFisher Scientific kit protocol has been shown to be successful [42]. 6. Ways to troubleshoot PCR to improve amplification: (1) If you had a total reaction volume of 25 μL, sometimes doubling reagent volumes and increasing the total volume to 50 μL can improve amplification. (2) Lower the annealing temperature. However, a lower annealing temperature can result in an increase in nonspecific PCR products. (3) Increase the number of cycles. However, too many (>40) cycles increase the chance of polymerase errors. (4) Increase the time associated with the

Exploring Toxin Evolution

123

68  C extension, sometimes necessary with longer transcripts. (5) Too much cDNA template can inhibit PCR. Try 1:2 or 1:10 dilutions of cDNA template before it is added to the PCR. 7. Make sure that when bacterial work is completed, precautions are taken for all work to be conducted under sterile conditions. All microcentrifuge tubes and pipette tips should be autoclaved, as well as all prepared LB broth and agar. Any items that come in contact with the bacteria must be discarded as biohazard waste. 8. RNA quality can be determined using a Bioanalyzer. The RIN (RNA Integrity Number) is calculated on a Bioanalyzer by evaluating the ratio between the ribosomal RNA (rRNA) subunits 28S and 18S [98]; this is used to establish the extent of RNase sample degradation. A RIN of at least 7 or 8 is considered acceptable. Spectrophotometry ratios measured on a Nanodrop are also good evaluations of protein or chemical contamination of isolated RNA. The 260/280 absorbance ratio of RNA should be approximately 2.0 to be lacking significant protein contamination, and the 260/230 ratio should also approximate 2.0–2.2 to demonstrate the absence of residual phenol or guanidine that can be carried over from the RNA isolation protocol. References 1. Mackessy SP (2010) The field of reptile toxinology: snakes, lizards and their venoms. In: Mackessy SP (ed) Handbook of venoms and toxins of reptiles. CRC Press/Taylor & Francis Group, Boca Raton, FL, pp 2–23 2. Gibbs HL, Rossiter W (2008) Rapid evolution by positive selection and gene gain and loss: PLA2 venom genes in closely related Sistrurus rattlesnakes with divergent diets. J Mol Evol 66 (2):151–166 3. Dowell NL, Giorgianni MW, Kassner VA, Selegue JE, Sanchez EE, Carroll SB (2016) The deep origin and recent loss of venom toxin genes in rattlesnakes. Curr Biol 26 (18):2434–2445 4. Safavi-Hemami H, Lu A, Li Q, Fedosov AE, Biggs J, Corneli PS, Seger J, Yandell M, Olivera BM (2016) Venom insulins of cone snails diversify rapidly and track prey taxa. Mol Biol Evol 33(11):2924–2934 5. Gendreau KL, Haney RA, Schwager EE, Wierschin T, Stanke M, Richards S, Garb JE (2017) House spider genome uncovers evolutionary shifts in the diversity and expression of black widow venom proteins associated with extreme toxicity. BMC Genomics 18:14 6. Doley R, Mackessy SP, Kini RM (2009) Role of accelerated segment switch in exons to alter

targeting (ASSET) in the molecular evolution of snake venom proteins. BMC Evol Biol 9:146 7. Li M, Fry BG, Kini RM (2005) Putting the brakes on snake venom evolution: the unique molecular evolutionary patterns of Aipysurus eydouxii (Marbled Sea snake) phospholipase A2 toxins. Mol Biol Evol 22(4):934–941 8. Whittington AC, Mason AJ, Rokyta DR (2018) A single mutation unlocks cascading exaptations in the origin of a potent pitviper neurotoxin. Mol Biol Evol 35(4):887–898 9. Nei M, Gu X, Sitnikova T (1997) Evolution by the birth-and-death process in multigene families of the vertebrate immune system. Proc Natl Acad Sci U S A 94(15):7799–7806 10. Margres MJ, Bigelow AT, Lemmon EM, Lemmon AR, Rokyta DR (2017) Selection to increase expression, not sequence diversity, precedes gene family origin and expansion in rattlesnake venom. Genetics 206 (3):1569–1580 11. Calvete JJ, Sanz L, Angulo Y, Lomonte B, Gutierrez JM (2009) Venoms, venomics, antivenomics. FEBS Lett 583(11):1736–1743 12. Calvete JJ (2014) Next-generation snake venomics: protein-locus resolution through

124

Cassandra M. Modahl et al.

venom proteome decomplexation. Expert Rev Proteomics 11(3):315–329 13. Aird SD, Aggarwal S, Villar-Briones A, Tin MM, Terada K, Mikheyev AS (2015) Snake venoms are integrated systems, but abundant venom proteins evolve more rapidly. BMC Genomics 16:647 14. Pla D, Petras D, Saviola AJ, Modahl CM, Sanz L, Perez A, Juarez E, Frietze S, Dorrestein PC, Mackessy SP, Calvete JJ (2017) Transcriptomics-guided bottom-up and top-down venomics of neonate and adult specimens of the arboreal rear-fanged Brown Treesnake, Boiga irregularis, from Guam. J Proteome 174:71–84 15. Modahl CM, Frietze S, Mackessy SP (2018) Transcriptome-facilitated proteomic characterization of rear-fanged snake venoms reveal abundant metalloproteinases with enhanced activity. J Proteome 187:223–234 16. Kaas Q, Craik DJ (2015) Bioinformatics-aided venomics. Toxins 7(6):2159–2187 17. Grabherr MG, Haas BJ, Yassour M, Levin JZ, Thompson DA, Amit I, Adiconis X, Fan L, Raychowdhury R, Zeng Q, Chen Z, Mauceli E, Hacohen N, Gnirke A, Rhind N, di Palma F, Birren BW, Nusbaum C, LindbladToh K, Friedman N, Regev A (2011) Fulllength transcriptome assembly from RNA-Seq data without a reference genome. Nat Biotechnol 29(7):644–652 18. Andrews S, FastQC. A quality control tool for high throughput sequence data. http://www. bioinformaticsbabrahamacuk/projects/ fastqc/ 19. Bolger AM, Lohse M, Usadel B (2014) Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30 (15):2114–2120 20. Zhang J, Kobert K, Flouri T, Stamatakis A (2014) PEAR: a fast and accurate Illumina paired-end reAd mergeR. Bioinformatics 30 (5):614–620 21. Magoc T, Salzberg SL (2011) FLASH: fast length adjustment of short reads to improve genome assemblies. Bioinformatics 27 (21):2957–2963 22. Rokyta DR, Lemmon AR, Margres MJ, Aronow K (2012) The venom-gland transcriptome of the eastern diamondback rattlesnake (Crotalus adamanteus). BMC Genomics 13:312 23. Archer J, Whiteley G, Casewell NR, Harrison RA, Wagstaff SC (2014) VTBuilder: a tool for the assembly of multi isoform transcriptomes. BMC Bioinformatics 15:389

24. Gilbert D. Gene-omes built from mRNA seq not genome DNA [version 1; not peer reviewed]. F1000 Research. 5:1695 (poster) (https://doi.org/10.7490/f1000research. 1112594.1) 25. Slater GS, Birney E (2005) Automated generation of heuristics for biological sequence comparison. BMC Bioinformatics 6:31 26. Fu L, Niu B, Zhu Z, Wu S, Li W (2012) CD-HIT: accelerated for clustering the nextgeneration sequencing data. Bioinformatics 28 (23):3150–3152 27. Camacho C, Coulouris G, Avagyan V, Ma N, Papadopoulos J, Bealer K, Madden TL (2009) BLAST+: architecture and applications. BMC Bioinformatics 10:421–421 28. Buchfink B, Xie C, Huson DH (2015) Fast and sensitive protein alignment using DIAMOND. Nat Methods 12:59–60 29. Petersen TN, Brunak S, von Heijne G, Nielsen H (2011) SignalP 4.0: discriminating signal peptides from transmembrane regions. Nat Methods 8(10):785–786 30. Krogh A, Larsson B, von Heijne G, Sonnhammer EL (2001) Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. J Mol Biol 305(3):567–580 31. Li B, Dewey CN (2011) RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics 12:323 32. Langmead B, Salzberg SL (2012) Fast gappedread alignment with bowtie 2. Nat Methods 9 (4):357–359 33. Larsson A (2014) AliView: a fast and lightweight alignment viewer and editor for large datasets. Bioinformatics 30(22):3276–3278 34. Gouy M, Guindon S, Gascuel O (2010) SeaView version 4: a multiplatform graphical user interface for sequence alignment and phylogenetic tree building. Mol Biol Evol 27 (2):221–224 35. Waterhouse AM, Procter JB, Martin DMA, Clamp M, Barton GJ (2009) Jalview version 2—a multiple sequence alignment editor and analysis workbench. Bioinformatics 25 (9):1189–1191 36. Kumar S, Stecher G, Tamura K (2016) MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol Biol Evol 33(7):1870–1874 37. Yang Z (2007) PAML 4: phylogenetic analysis by maximum likelihood. Mol Biol Evol 24 (8):1586–1591

Exploring Toxin Evolution 38. Delport W, Poon AFY, Frost SDW, Kosakovsky Pond SL (2010) Datamonkey 2010: a suite of phylogenetic analysis tools for evolutionary biology. Bioinformatics 26(19):2455–2457 39. Searle BC (2010) Scaffold: a bioinformatic tool for validating MS/MS-based proteomic studies. Proteomics 10(6):1265–1269 40. Zhang J, Xin L, Shan B, Chen W, Xie M, Yuen D, Zhang W, Zhang Z, Lajoie GA, Ma B (2012) PEAKS DB: de novo sequencing assisted database search for sensitive and accurate peptide identification. Mol Cell Proteomics 11(4):M111.010587 41. Vaudel M, Barsnes H, Berven FS, Sickmann A, Martens L (2011) SearchGUI: an open-source graphical user interface for simultaneous OMSSA and X!Tandem searches. Proteomics 11(5):996–999 42. Modahl CM, Mackessy SP (2016) Full-length venom protein cDNA sequences from venomderived mRNA: exploring compositional variation and adaptive multigene evolution. PLoS Negl Trop Dis 10(6):e0004587 43. Shendure J, Ji H (2008) Next-generation DNA sequencing. Nat Biotech 26(10):1135–1145 44. Goodwin S, McPherson JD, McCombie WR (2016) Coming of age: ten years of nextgeneration sequencing technologies. Nat Rev Genet 17(6):333–351 45. Parkinson J, Blaxter M (2009) Expressed sequence tags: an overview. Methods Mol Biol 533:1–12 46. Rotenberg D, Bamberger ES, Kochva E (1971) Studies on ribonucleic acid synthesis in the venom glands of Vipera palaestinae (Ophidia, Reptilia). J Biochem 121:609–612 47. Hargreaves AD, Swain MT, Hegarty MJ, Logan DW, Mulley JF (2014) Restriction and recruitment – gene duplication and the origin and evolution of snake venom toxins. Genome Biol Evol 8:2088–2095 48. Reyes-Velasco J, Card DC, Andrew AL, Shaney KJ, Adams RH, Schield DR, Casewell NR, Mackessy SP, Castoe TA (2015) Expression of venom gene homologs in diverse python tissues suggests a new model for the evolution of snake venom. Mol Biol Evol 32(1):173–183 49. Junqueira-de-Azevedo IL, Bastos CM, Ho PL, Luna MS, Yamanouye N, Casewell NR (2015) Venom-related transcripts from Bothrops jararaca tissues provide novel molecular insights into the production and evolution of snake venom. Mol Biol Evol 32(3):754–766 50. Hargreaves AD, Mulley JF (2015) Assessing the utility of the Oxford Nanopore MinION

125

for snake venom gland cDNA sequencing. PeerJ 3:e1441 51. Nakasugi K, Crowhurst R, Bally J, Waterhouse P (2014) Combining transcriptome assemblies from multiple de novo assemblers in the Allotetraploid plant Nicotiana benthamiana. PLoS One 9(3):e91776 52. Macrander J, Broe M, Daly M (2015) Multicopy venom genes hidden in de novo transcriptome assemblies, a cautionary tale with the snakelocks sea anemone Anemonia sulcata (pennant, 1977). Toxicon 108:184–188 53. Honaas LA, Wafula EK, Wickett NJ, Der JP, Zhang Y, Edger PP, Altman NS, Pires JC, Leebens-Mack JH, dePamphilis CW (2016) Selecting superior de novo transcriptome assemblies: lessons learned by leveraging the best plant genome. PLoS One 11(1): e0146062 54. Aird SD, Watanabe Y, Villar-Briones A, Roy MC, Terada K, Mikheyev AS (2013) Quantitative high-throughput profiling of snake venom gland transcriptomes and proteomes (Ovophis okinavensis and Protobothrops flavoviridis). BMC Genomics 14:790 55. Gruenheit N, Deusch O, Esser C, Becker M, Voelckel C, Lockhart P (2012) Cutoffs and k-mers: implications from a transcriptome study in allopolyploid plants. BMC Genomics 13:92 56. Simpson JT, Wong K, Jackman SD, Schein JE, Jones SJM, Birol I˙ (2009) ABySS: a parallel assembler for short read sequence data. Genome Res 19(6):1117–1123 57. Zerbino DR (2010) Using the velvet de novo assembler for short-read sequencing technologies. Curr Protoc Bioinformatics Chapter 11: Unit 11.5 58. McGivern JJ, Wray KP, Margres MJ, Couch ME, Mackessy SP, Rokyta DR (2014) RNA-seq and high-definition mass spectrometry reveal the complex and divergent venoms of two rear-fanged colubrid snakes. BMC Genomics 15:1061 59. Zhao QY, Wang Y, Kong YM, Luo D, Li X, Hao P (2011) Optimizing de novo transcriptome assembly from short-read RNA-Seq data: a comparative study. BMC Bioinformatics 12 (14):S2 60. Huang X, Madan A (1999) CAP3: a DNA sequence assembly program. Genome Res 9 (9):868–877 61. Smith-Unna R, Boursnell C, Patro R, Hibberd JM, Kelly S (2016) TransRate: reference-free quality assessment of de novo transcriptome assemblies. Genome Res 26(8):1134–1144

126

Cassandra M. Modahl et al.

62. O’Neil ST, Emrich SJ (2013) Assessing de novo transcriptome assembly metrics for consistency and utility. BMC Genomics 14:465 63. Finn RD, Clements J, Eddy SR (2011) HMMER web server: interactive sequence similarity searching. Nucleic Acids Res 39: W29–W37 64. Finn RD, Coggill P, Eberhardt RY, Eddy SR, Mistry J, Mitchell AL, Potter SC, Punta M, Qureshi M, Sangrador-Vegas A, Salazar GA, Tate J, Bateman A (2016) The Pfam protein families database: towards a more sustainable future. Nucleic Acids Res 44(D1):D279–D285 65. Hunter S, Jones P, Mitchell A, Apweiler R, Attwood TK, Bateman A, Bernard T, Binns D, Bork P, Burge S, de Castro E, Coggill P, Corbett M, Das U, Daugherty L, Duquenne L, Finn RD, Fraser M, Gough J, Haft D, Hulo N, Kahn D, Kelly E, Letunic I, Lonsdale D, Lopez R, Madera M, Maslen J, McAnulla C, McDowall J, McMenamin C, Mi H, Mutowo-Muellenet P, Mulder N, Natale D, Orengo C, Pesseat S, Punta M, Quinn AF, Rivoire C, Sangrador-Vegas A, Selengut JD, Sigrist CJ, Scheremetjew M, Tate J, Thimmajanarthanan M, Thomas PD, Wu CH, Yeats C, Yong SY (2012) InterPro in 2011: new developments in the family and domain prediction database. Nucleic Acids Res 40:D306–D312 66. Rabouille C (2017) Pathways of unconventional protein secretion. Trends Cell Biol 27 (3):230–240 67. Mortazavi A, Williams BA, McCue K, Schaeffer L, Wold B (2008) Mapping and quantifying mammalian transcriptomes by RNA-Seq. Nat Methods 5(7):621–628 68. Sunagar K, Moran Y (2015) The rise and fall of an evolutionary innovation: contrasting strategies of venom evolution in ancient and young animals. PLoS Genet 11(10):e1005596 69. Sunagar K, Undheim EA, Scheib H, Gren EC, Cochran C, Person CE, Koludarov I, Kelln W, Hayes WK, King GF, Antunes A, Fry BG (2014) Intraspecific venom variation in the medically significant southern Pacific rattlesnake (Crotalus oreganus helleri): biodiscovery, clinical and evolutionary implications. J Proteome 99:68–83 70. Rokyta DR, Wray KP, Lemmon AR, Lemmon EM, Caudle BS (2011) A high-throughput venom-gland transcriptome for the eastern diamondback rattlesnake (Crotalus adamanteus) and evidence for pervasive positive selection across toxin classes. Toxicon 57(5):657–671 71. Aird SD, Arora J, Barua A, Qiu L, Terada K, Mikheyev AS (2017) Population genomic analysis of a pitviper reveals microevolutionary

forces underlying venom chemistry. Genome Biol Evol 9(10):2640–2649 72. Sunagar K, Jackson T, Undheim E, Ali S, Antunes A, Fry BG (2013) Three-fingered RAVERs: rapid accumulation of variations in exposed residues of snake venom toxins. Toxins 5(11):2172–2208 73. Nielsen R, Yang Z (1998) Likelihood models for detecting positively selected amino acid sites and applications to the HIV-1 envelope gene. Genetics 148(3):929–936 74. Yang Z (2000) Maximum likelihood estimation on large phylogenies and analysis of adaptive evolution in human influenza virus a. J Mol Evol 51:423–432 75. Yang Z, Wong WS, Nielsen R (2005) Bayes empirical Bayes inference of amino acid sites under positive selection. Mol Biol Evol 22 (4):1107–1118 76. Zhang J, Nielsen R, Yang Z (2005) Evaluation of an improved branch-site likelihood method for detecting positive selection at the molecular level. Mol Biol Evol 22(12):2472–2479 77. Pond SL, Frost SD, Muse SV (2005) HyPhy: hypothesis testing using phylogenies. Bioinformatics 21(5):676–679 78. Pond SL, Frost SD (2005) Datamonkey: rapid detection of selective pressure on individual sites of codon alignments. Bioinformatics 21 (10):2531–2533 79. Chapeaurouge A, Silva A, Carvalho P, McCleary RJR, Modahl CM, Perales J, Kini RM, Mackessy SP (2018) Proteomic deep mining the venom of the red-headed krait, Bungarus flaviceps. Toxins 10(9):E373 80. Mackessy SP (1988) Venom ontogeny in the pacific rattlesnakes Crotalus viridis helleri and C. v. oreganus. Copeia 1:92–101 ˜ ez V, Toro MF, 81. Saldarriaga MM, Otero R, Nu´n Dı´az A, Gutie´rrez JM (2003) Ontogenetic variability of Bothrops atrox and Bothrops asper snake venoms from Colombia. Toxicon 42 (4):405–411 82. Saviola AJ, Pla D, Sanz L, Castoe TA, Calvete JJ, Mackessy SP (2015) Comparative venomics of the prairie rattlesnake (Crotalus viridis viridis) from Colorado: identification of a novel pattern of ontogenetic changes in venom composition and assessment of the immunoreactivity of the commercial antivenom CroFab(R). J Proteome 121:28–43 83. Rokyta DR, Margres MJ, Ward MJ, Sanchez EE (2017) The genetics of venom ontogeny in the eastern diamondback rattlesnake (Crotalus adamanteus). PeerJ 5:e3249 84. Massey DJ, Calvete JJ, Sa´nchez EE, Sanz L, Richards K, Curtis R, Boesen K (2012)

Exploring Toxin Evolution Venom variability and envenoming severity outcomes of the Crotalus scutulatus scutulatus (Mojave rattlesnake) from Southern Arizona. J Proteome 75(9):2576–2587 85. Rokyta DR, Wray KP, Margres MJ (2013) The genesis of an exceptionally lethal venom in the timber rattlesnake (Crotalus horridus) revealed through comparative venom-gland transcriptomics. BMC Genomics 14:394 86. Margres MJ, Walls R, Suntravat M, Lucena S, Sanchez EE, Rokyta DR (2016) Functional characterizations of venom phenotypes in the eastern diamondback rattlesnake (Crotalus adamanteus) and evidence for expressiondriven divergence in toxic activities among populations. Toxicon 119:28–38 87. Aggarwal S, Yadav AK (2016) False discovery rate estimation in proteomics. Methods Mol Biol 1362:119–128 88. Elias JE, Gygi SP (2010) Target-decoy search strategy for mass spectrometry-based proteomics. Methods Mol Biol 604:55–71 89. Modahl CM, Mrinalini FS, Mackessy SP (2018) Adaptive evolution of distinct preyspecific toxin genes in rear-fanged snake venom. Proc Biol Sci 285(1884):20181003 90. Neilson KA, Keighley T, Pascovici D, Cooke B, Haynes PA (2013) Label-free quantitative shotgun proteomics using normalized spectral abundance factors. Methods Mol Biol 1002:205–222 91. Paoletti AC, Parmely TJ, Tomomori-Sato C, Sato S, Zhu D, Conaway RC, Conaway JW, Florens L, Washburn MP (2006) Quantitative proteomic analysis of distinct mammalian mediator complexes using normalized spectral abundance factors. Proc Natl Acad Sci 103 (50):18928–18933

127

92. Rokyta DR, Margres MJ, Calvin K (2015) Post-transcriptional mechanisms contribute little to phenotypic variation in snake venoms. G3 5(11):2375–2382 93. Fabre B, Lambour T, Bouyssie´ D, Menneteau T, Monsarrat B, Burlet-Schiltz O, Bousquet-Dubouch M-P (2014) Comparison of label-free quantification methods for the determination of protein complexes subunits stoichiometry. EuPA Open Proteom 4:82–86 94. Ishihama Y, Oda Y, Tabata T, Sato T, Nagasu T, Rappsilber J, Mann M (2005) Exponentially modified protein abundance index (emPAI) for estimation of absolute protein amount in proteomics by the number of sequenced peptides per protein. Mol Cell Proteomics 4(9):1265–1272 95. Calvete JJ (2013) Snake venomics: from the inventory of toxins to biology. Toxicon 75:44–62 96. Pahari S, Mackessy SP, Kini RM (2007) The venom gland transcriptome of the Desert Massasauga Rattlesnake (Sistrurus catenatus edwardsii): towards an understanding of venom composition among advanced snakes (Superfamily Colubroidea). BMC Mol Biol 8:115 97. Casewell NR, Huttley GA, Wuster W (2012) Dynamic evolution of venom proteins in squamate reptiles. Nat Commun 3:1066 98. Schroeder A, Mueller O, Stocker S, Salowsky R, Leiber M, Gassmann M, Lightfoot S, Menzel W, Granzow M, Ragg T (2006) The RIN: an RNA integrity number for assigning integrity values to RNA measurements. BMC Mol Biol 7:3–3

Chapter 7 Three-Dimensional Structure Determination of Peptides Using Solution Nuclear Magnetic Resonance Spectroscopy Christina I. Schroeder and K. Johan Rosengren Abstract Nuclear magnetic resonance (NMR) spectroscopy has over the last few decades proven to be an extremely useful technique for, and indeed an integral part of, investigating the structural features of peptides and small proteins directly in solution, without the need for crystallization. This advantage over X-ray methods is important when dealing with peptides and small proteins that do not readily form crystals. In this chapter we outline what specific NMR experiments are useful, considerations about how to acquire and interpret these experiments, and how information derived from the NMR data can be used to determine solution structures of small peptides. Key words Peptide, Nuclear Magnetic Resonance spectroscopy, Homonuclear NMR, Total correlation spectroscopy, Nuclear Overhauser effect spectroscopy

1

Introduction Small proteins of around 10–60 amino acids tend to be difficult to crystallize, resulting in X-ray crystallography not being a suitable technique for structural studies. Homonuclear solution NMR spectroscopy is instead generally the method of choice for elucidating the three-dimensional structure of peptides. Although some are amenable to bacterial protein expression, many peptides, including ones that are posttranslationally modified, disulfide-rich (requiring correct folding), or venom-derived that can be toxic to the expressing cell, are difficult to produce in 15N and/or 13C-labeled form, prohibiting heteronuclear NMR to be used. Therefore, despite recent advances in heteronuclear NMR using sophisticated three(or higher) dimensional techniques and automated assignment protocols, homonuclear and natural abundance heteronuclear two-dimensional (2D) NMR remain an important technique that can be generally applied to peptides extracted from natural sources or chemically synthesized [1–3]. This chapter will not provide any background on the theory of nuclear magnetic resonance. Instead

Avi Priel (ed.), Snake and Spider Toxins: Methods and Protocols, Methods in Molecular Biology, vol. 2068, https://doi.org/10.1007/978-1-4939-9845-6_7, © Springer Science+Business Media, LLC, part of Springer Nature 2020

129

130

Christina I. Schroeder and K. Johan Rosengren

it is intended to serve as a practical guide to sample preparation, experimental setup, spectral analysis, structure calculations, and structure validation.

2

Materials

2.1 NMR Spectrometer

There are many NMR spectrometers on the market with various magnetic field strengths. As a general rule, the NMR spectrometer with the strongest magnetic field available to you will provide highest resolution and likely give you the best data. However, due to the characteristics of the various models of probes available and your sample conditions, the NMR spectrometer with the strongest magnetic field may not give you the best signal-to-noise ratio due to, e.g., inadequate solvent suppression. In particular, for low-concentration samples using a spectrometer equipped with a cryoprobe will provide significant improvement in signal-to-noise.

2.2 Computing and Software

Many different software packages are available for the different computational steps required, including data processing, data analysis, and structure calculations. Software programs for analysis of NMR data are often written as part of research projects rather than for commercial purposes and are free to use for academic research, although some of the common program packages do require a license. In terms of computational power, a modern PC is generally fully sufficient but different program packages may be written for different platforms. Linux or MacOSX operating systems are popular and provide the most flexibility for the range of applications utilized during the different steps of the process.

3

Methods

3.1 Sample Preparation and Solvents

1. Prepare peptide samples with a minimum of 5% deuterated solvent. For standard 5 mm NMR tubes, dissolve water-soluble peptides in ~500 μL 90% H2O/10% D2O or 99.96% D2O to a concentration of ~1 mM. Poorly soluble samples may require addition of organic cosolvent (see Note 1). Samples from HPLC purification tend to have a pH ~3–4 (uncorrected for isotope effects) when dissolved in pure water, which often give excellent spectra with sharp lines, but this can be adjusted and samples buffered if more physiological pH and salt conditions are required (see Note 2). The following protocol will assume that a sample dissolved in water or aqueous buffer is used. 2. To acquire data suitable for three-dimensional (3D) structure determination of a protein a spectrometer of >500 MHz is typically required. Access to a NMR spectrometer with a cryoprobe will, as noted above, provide significantly improved quality of data and also shorten required acquisition time.

Peptide 3D Structure Determination by NMR

3.2 Data Acquisition and Processing

131

Modern-day spectrometers tend to have highly automated setups allowing users with limited experience and background in NMR theory to set up standard experiments. The following basic protocol will highlight the steps required to go through if you manually set up your experiments. For users of the most common Bruker NMR instruments the commands utilized by the TopSpin software for each step are provided in square bracket. 1. Insert the sample into the spectrometer and adjust the temperature to 298 K [edte], or an alternative desired temperature if your sample is unstable at room temperature or requires specific conditions (see Note 3), and let the temperature equilibrate. 2. Lock on the deuterium signal of your chosen solvent system [lock—chosen H2O/D2O or other solvent if appropriate]. Perform gradient shimming [topshim] to ensure a homogeneous magnetic field over your entire sample (see Note 4). Tune and match the sample [atma—for automated; atmm or wobb—for manual systems] (see Note 5) and calculate the p1 90
pulse [pulsecal] (see Note 6). Adjust your offset frequency defining the center of 1H spectral window to coincide with the water resonance [o1, can be optimized to minimize the water signal in gs mode]. An optimal offset will ensure optimal water suppression by your chosen suppression scheme; excitation sculpting using gradients is generally preferred [4] (see Note 7). These parameters can be copied across to subsequent experiments using the same sample to reduce experiment setup time. 3. Set up a one-dimensional (1D) 1H NMR experiment (see Note 8) with a 15 ppm sweep width [sw] (see Note 9) and 32 k data points in the F1 dimension using 8–64 scans (depending on sample concentration) (see Note 10). Adjust all required pulse lengths and powers according to the determined 90
1H pulse at the specified power level [pl1] for each new experiment [getprosol 1H 90 pulse power level], and determine the receiver gain [rga] (see Note 11). Start your experiment [zg] and after completion process the data by Fourier transformation [ft]. From the appearance of the 1D spectrum, ensure that sufficient solvent suppression has been achieved and that the peptide is behaving as expected under the conditions used. Protein signals should be sharp and easily distinguishable from noise and, for a structured peptide, are expected to be well dispersed (Fig. 1). If the peptide signals are broad or weak further optimization of the sample conditions may be required before two 2D data sets can be recorded (see Note 12). 4. For 2D experiments adjust the sweep width to cover the frequencies of all signals observed in the 1D spectrum. 2D experiments (see Note 12) required for full 1H spectral assignment

132

Christina I. Schroeder and K. Johan Rosengren

Fig. 1 1D 1H NMR spectra showing the amide region of two disulfide bond isomers of [E17K]ProTx-II. (a) Isomer 1 displays native (desired) disulfide bond connectivity and (b) isomer 2 displays misfolded (non-native) disulfide connectivity. The comparison highlights the difference in the dispersion of amide proton chemical shifts in folded and misfolded peptides [47]

and for generation of distance restraints for structure calculations include TOCSY [5] (see Note 13) typically recorded using a spin-lock sequence with a mixing time of ~80 ms and 4–16 scans (depending on spectrometer and sample concentration), and NOESY [6, 7] (see Note 14) typically recorded with a mixing time of 100–300 ms and 24–80 scans. 5. For small peptides, in cases where the NOESY cross peaks are weak due to the fast tumbling of the peptide in solution, a ROESY experiment [8, 9] with a mixing time of 200–350 ms may be preferred (see Note 15). 6. For dihedral restraints, to confirm assignments and to compile a complete chemical shift list, run additional experiments including DQF-COSY [10] (see Note 16), E.COSY [11] (see Note 17), natural abundance 1H–15N HSQC [12], and natural abundance 1H–13C HSQC [12, 13] (see Note 18). The natural abundance data will require high sample concentration or access to a cryoprobe. Experiments like E.COSY and 1 H–13C HSQC, which are exclusively for analyzing non-exchangeable protons, are preferably run in D2O following the D2O exchange experiments (see below) in order to minimize interference of the water signal during peak assignment. 7. To determine amide proton temperature coefficients that can be indicative of hydrogen bonds, record a series of TOCSY experiments in 90% H2O/10% D2O at 280–310 K, in 5 K temperature increments, and monitor the changes in amide proton chemical shifts (see Note 19).

Peptide 3D Structure Determination by NMR

133

Fig. 2 Chemical shift of H2O as a function of temperature. Eq. (1) (see Note 22) produces the solid line which deviates from the empirical measurements by less than 1 ppb [30]

8. To identify amides involved in potential hydrogen bonds by monitoring D2O exchange, run a series of TOCSY spectra over 24 h immediately upon dissolution of a fully protonated lyophilized peptide in 100% D2O and monitoring the exchange (disappearance) of HN proton signals over time (see Note 20). 9. 2D spectra are typically acquired with 4096 data points in the F2 dimension and 512 increments in the F1 dimension for TOCSY, NOESY, DQF-COSY, and E.COSY experiments, 2048  256 for 1H–13C HSQC and 2048  128 for 1H–15N HSQC data points in the F2 dimension and increments in the F1 dimension, respectively. 10. Process all spectra using TopSpin (Bruker) [ft or efp for 1D; xfb for 2D] (see Note 21). Baseline corrections can be performed in both the f2 [abs2.water—specify position of water signal] and f1 [abs1] dimensions. 11. Reference spectra to 0.00 ppm using an internal standard such as DSS or TMS or to water at 4.77 ppm at 298 K (see Note 22, Fig. 2). 13C and 15N dimensions can be referenced indirectly to the referenced proton spectrum [xiref]. 3.3 Resonance Assignments and Derivation of Structural Restraints

1. Homonuclear resonance assignments are achieved using Wu¨thrich’s sequential assignment protocol [14]. Use TOCSY and NOESY data acquired at the same temperature and identify amino acid spin systems in the TOCSY (Fig. 3) that can be linked by sequential NOEs in the NOESY (see Note 23, Figs. 4, 5, and 6). Following completion of the resonance assignments, export distance restraints “peaks” and chemical shift “prot” information to be used as input files during the structure calculations.

134

Christina I. Schroeder and K. Johan Rosengren

0.5

1

1.5

chemical shift (ppm)

2

2.5 α β γ

3

δ ε

3.5

4

4.5

Lys Hz

Arg HE

u Ly s M et Ph e Se r Th r Tr p Ty r Va l

Le

Ile

Al

a Ar g As n As p C ys G ln G lu G ly H is

5

Fig. 3 TOCSY spin systems for the different amino acids based on the proton random coil shifts. The differences in number of scalar-coupled protons and their chemical shifts result in distinct patterns for most amino acids. Arg and Lys residues contain side-chain HN groups that are observable at lower pH, when not in fast exchange with the solvent. These residues thus can be identified based on the presence of two identical spin system originating from the backbone and side-chain amides, which both are part of the same scalar-coupling pathway. Pro lacks an amide proton, and thus does not appear in the amide region of the TOCSY spectrum. The cross peaks are color coded based on their position in the side chain

2. Complete 1H–13C HSQC assignments using the now known 1 H chemical shifts. Distinct 13C chemical shifts [15] can be used to confirm assignments of ambiguous 1H resonances with similar 1H chemical shifts, such as Pro Hβ and Hγ resonances (see Note 24, Fig. 7). 3. Complete 1H–15N HSQC assignments using the known 1H amide proton shifts. 3.4 Secondary Structure Predictions

1. Calculate and plot the secondary Hα (and Cα/Cβ) chemical shifts and evaluate distance restraint patterns to gain a first insight into potential secondary structure elements present in the peptide (see Note 25, Figs. 8 and 9).

Peptide 3D Structure Determination by NMR

135

Fig. 4 Expected cross peaks for the tripeptide fragment Ser-Val-Asp in different types of NMR experiments. The DQF-COSY HNi-Hαi cross peaks are shown in blue and yellow. The TOCSY HNi-Hαi, HNi-Hβ1/2i, and HNiHγ∗i cross peaks are shown in green. The “imprint” of the spin systems on the following amino acid, which is observed in the form of sequential Hαi-HNi + 1, Hβi-HNi + 1, and Hγ∗i-HNi + 1 cross peaks in the NOESY, is shown in red and highlighted by arrows. The majority of the TOCSY cross peaks are also observed in the NOESY spectrum because of the close proximity of the Hα and side-chain protons of an amino acid to its own HN, and the HN of the following amino acid

136

Christina I. Schroeder and K. Johan Rosengren

Fig. 5 Fingerprint region of an 80 ms TOCSY spectrum of [A2]GVIA (2 mM, 293 K, 90% H2O/10% D2O, pH 3.5, Bruker 750 MHz). Amino acid spin systems are labeled according to standard one-letter amino acid codes and residue numbers

3.5 Dihedral-Angle Restraints

1. Measure 3JHN-Hα coupling constants from the 1D spectra or DQF-COSY spectra to obtain experimentally derived ϕ dihedral angles (see Note 26, Fig. 10). 2. Compile a list of all chemical shifts for Hα, Cα, Cβ, HN, and N derived from NOESY, 1H–13C HSQC, and 1H–15N HSQC spectra to generate constraints for ϕ and ψ backbone dihedral

Peptide 3D Structure Determination by NMR

137

Fig. 6 Fingerprint region of a 300 ms NOESY spectrum of GVIA (2 mM, 293 K, 90% H2O/10% D2O, pH 3.5, Bruker 750 MHz). Intra-residual cross peaks are labeled with residue numbers. The sequential walk described by Wu¨thrich [14], including HNi–Hαi and Hαi–Hαi + 1 connectivities, as well as O4α–G5 HN and O21α–Y22 HN connections (O—hydroxyproline), is illustrated. Green and red lines correspond to intra- and inter-residual steps, respectively, linking spin systems during the assignment process

138

Christina I. Schroeder and K. Johan Rosengren

Fig. 7 Natural abundance 1H–13C-HSQC spectrum of a conotoxin peptide, acquired at 0.3 mM, 298 K, 100% D2O, and pH 4 on a Bruker Avance 600 NMR spectrometer equipped with a cryoprobe. The typical clustering of different types of peptide 1H–13C cross peaks, useful for confirming side-chain assignments, is highlighted by boxes and resonance identifiers

angles based on database matching using TALOS-N (see Note 27) [16–18]. 3. To determine χ 1 dihedral angles, measure 3JHα-Hβ coupling constants from the E.COSY spectrum and use these coupling constants together with intensities of intra-residual NOEs to stereospecifically assign Hβ methylene pairs [19] (see Note 28, Fig. 11). TALOS-N is also able to make prediction of χ 1 dihedral angles [18], and extending the chemical shift analysis to disulfide bonds, the program DISH is able to also predict χ 2 angles for cystine residues [20].

Peptide 3D Structure Determination by NMR 2

A

139

C

1.5 Δ δ/ppm

1 0.5 0

-0.5 -1 -1.5

1

6

11

16

21

26

31

36

41

46

51

56

Ma-2 residue number

ƒ TM/ppm

B 0.6

D

0.4 0.2 0 -0.2 -0.4 1

2

3

4

5

6 7 8 9 10 11 12 13 14 15 16 LvIA residue number

Fig. 8 Comparison of backbone secondary Hα chemical shifts for (a) mambalgin-2 (PDB ID: 2MFA) [3], and (b) LvIA (PDB ID: 2MDQ) [48]. The secondary Hα chemical shifts are the difference between the observed chemical shifts and random coil values [49]. Stretches of positive secondary shifts suggest β-sheet formation and stretches of negative secondary chemical shifts suggest α-helix or turns (black arrows). (c) NMR solution structure of mambalgin-2 highlighting β-sheets corresponding to the stretches of positive secondary Hα chemical shifts and (d) NMR solution structure of LvIA highlighting the α-helix motifs corresponding to stretches of negative secondary Hα chemical shift

3.6 Structure Calculation and Analysis

1. Prepare data files for structure calculations according to the input format required for your choice of computational software. For structure determination using CYANA [21, 22] the minimum requirements are the following: a “seq” file defining the amino acid sequence of the studied system, a “prot” file comprising the chemical shifts of all assigned atoms, and a “peaks” file containing the positions and intensities or volumes of all NOE cross peaks (see Note 29). In addition, a dihedral restraint list, including experimentally or TALOS-N-derived dihedral angles, may be included following conversion to an “aco” file. Restraints for disulfide bonds are required if disulfide bonds are present and disulfide connectivity is known (see Note 30) and distance restraints for cyclization need to be included if the peptide is cyclic (see Note 29). For a full list of files required

140

Christina I. Schroeder and K. Johan Rosengren

a-helix

310-helix

b-strand

1234567

1234567

1234567

4 4 4 4 4 4 4*

4 4 4 4 4 4 4*

9 9 9 9 9 9 9*

dNN(i,i+1) daN(i,i+1) daN(i,i+3) daβ(i,i+3) daN(i,i+2) dNN(i,i+2) daN(i,i+4) 3

JHNa(Hz)

Fig. 9 Summary of sequential and medium-range NOE patterns and 3JHN–Hα useful for identifying secondary structures. NOEs involving HN, Hα, and Hβ protons observed for ideal α-helices, 310-helices, and parallel or antiparallel β-strands are shown as bars. The thickness of the bar indicates the intensity of the NOE. ∗3JHN–Hα values are approximate (adapted from Sutcliffe [50])

Fig. 10 Backbone dihedral angles phi (ϕ, green), psi (ψ, blue), and the sidechain dihedral angle chi1 (χ1, red) define the conformation around key rotatable bonds, and can be experimentally determined to define the protein fold

and individual file format, see the CYANA manual: (http:// www.cyana.org/wiki/index.php/CYANA_3.0_Reference_ Manual). 2. 3D solution structure calculations are carried out with the goal of finding the lowest energy minimum satisfying all the restraints available from the input files (see Note 31). 3. Use the AUTO macro in CYANA to complete NOESY assignments followed by CALC or ANNEAL protocols to perform torsion angle dynamics calculations using all available structural information (see Note 32).

Peptide 3D Structure Determination by NMR

141

Fig. 11 Newman projections for the three possible χ1 dihedral angles staggered around the Cα–Cβ bond. Gauche and trans are indicated as g and t, respectively. By analysis of the coupling constants and NOE patterns the correct conformation can be identified, allowing both stereospecific assignment of the geminal protons and side-chain dihedral-angle restraints to be introduced. Adapted from Wagner et al. [38]

4. When all violations of experimental data have been resolved, and a high-quality structural family has been achieved, either this can be chosen as the final representation of the solution structure of the studied peptide or further refinement including energy minimization in water can be done in the program CNS [23]. A number of useful scripts for translating the CYANA inputs and preliminary structure into a molecular template, as well as running CNS calculations, are available from the RECOORD database (http://www.ebi.ac.uk/pdbe/ recalculated-nmr-data) [24]. Within CNS, calculate an ensemble of 50–100 structures and refine these further in a water shell [25] (see Note 33). 5. Out of the 50–100 structures calculated in CNS, choose a set of 20 structures that are of high quality and can be considered representative of the solution structure of your peptide. Final structures should be consistent with the experimental data with ˚ or dihedral-angle no or few NOE violations greater than 0.2 A
violations greater than 2 , and have low computational “energy.” Furthermore, they should have good covalent

142

Christina I. Schroeder and K. Johan Rosengren

Table 1 NMR distance and dihedral statistics for NKR-5-3B [2] Distance constraints: NOE

Total

1048

Intra-residual

(|i-j| ¼ 0) 297

Sequential

(|i-j| ¼ 1) 274

Medium range

(|i-j|  4) 294

Long range

(|i-j|  5) 183

Hydrogen bonds (for 48 H-bonds)

96

Dihedral angles: ϕ

54

ψ

54

χ1

28

Structure statistics Violations Distance constraints (>0.2 A˚)

0

Dihedral-angle constraints (>3º)

0

Maximum distance constraints ˚) violation (A

0.176

Maximum dihedral-angle constraints (º)

2.3

Energies (kcal/mol, mean  standard deviation) Overall

2137  19.3

Bond

16.1  0.81

Angles

42.4  2.58

Improper

20.3  2.16

Van der Waals

237.4 6.15

NOE (experimental)

0.260  0.017

Constrained dihedral (experimental)

0.619  0.32

Dihedral

279.1  1.15

Electric

2259  15.8

RMS deviation from idealized geometry ˚) Bond length (A

0.00831  0.00023

Bond angles (º)

0.878  0.021

Impropers (º)

1.12  0.081 (continued)

Peptide 3D Structure Determination by NMR

143

Table 1 (continued) Average pairwise root mean square deviationa (A˚) Heavy

0.77  0.10

Backbone

0.51  0.14

MolProbity Clashscore, all atomsb

14.4  2.15

Poor rotamers

00

Ramachandran favored (%)

95.2  1.57

Ramachandran allowed (%)

4.84  1.57

Ramachandran outliers (%)

00 1.97  0.095

MolProbity score MolProbity score percentile

c

CB deviations, bad backbone bonds/angles

76.9  4.46 00

a

Pairwise root mean square deviation from 20 refined structures over amino acids 1–64 Number of steric overlaps (>0.4 A˚)/1000 atoms c 100% is the best among structures of comparable resolution. 0% is the worst b

structure and nonbonded interactions as judged by a MolProbity analysis [26] (see Note 34, Table 1). 6. For visualization and analysis of the calculated structures use a suitable graphics program such as MOLMOL (see Note 35, Fig. 12) [27] or PyMOL (the PyMOL Molecular Graphics System, Version 1.7.4 Schro¨dinger, LLC). 7. Generate files for submissions of the atom coordinates, experimental restraints, and chemical shifts to the Protein Data Bank (PDB) and Biological Magnetic Resonance Bank (BMRB). Information about the submission procedure and file formats are available at http://www.rcsb.org/pdb/.

4

Notes 1. Prior to acquiring NMR spectra, the solubility of your peptide needs to be considered. During the production of the peptide of interest, whether it is produced chemically using solid-phase peptide synthesis or bacterial protein expression, a basic understanding of the solubility of the peptide will be gained. This will guide the choice of solvent suitable for the subsequent NMR experiments. To reduce acquisition time NMR samples are generally of high concentration. Up to 2 mM of peptide in

144

Christina I. Schroeder and K. Johan Rosengren

Fig. 12 Superposition of the 20 final structures of the backbone cyclic enterocin NKR-5-3B [2], showing the protein backbone atoms (a) and all heavy atoms (b). The presentation of a NMR structure as an ensemble rather than a single model allows the assessment of what parts of the peptide are well defined and which parts are poorly defined by the data, with the latter resulting in several conformations being equally plausible. Poor resolution may be a result of genuine flexibility, as is often the case around chain termini, exposed loops, and solvent-exposed side chains, but may also be a result of lack of key experimental data

~550 μL of solvent is not uncommon, and indeed may be preferred if your sample allows. With access to NMR spectrometers with increased magnetic field and technical advances including cryoprobes, sample concentration can be significantly reduced and a few 100 μM is often sufficient for a full structure determination. However, lower concentration requires increased acquisition time. As a general rule, the more concentrated the sample is, the better, as long as the sample is soluble in a suitable solvent and does not have a tendency to aggregate. The most obvious choice of solvent for peptides is water. However, if your sample has a tendency to aggregate in aqueous buffer systems, a titration of various solvents may be necessary to achieve optimal solubility to produce the best data possible. All samples prepared in water should be dissolved using ultrapure Milli-Q (18.2 MΩ at 25
C) or 99.96% deuterium oxide (D2O). If a buffer is preferred, choose a system without protons that will interfere with the 1H data, e.g., phosphate buffer. If organic cosolvent is required the most common NMR organic solvents include acetonitrile-d3 (CD3CN), dimethyl sulfoxide-d6 (DMSO), 2,2,2trifluoroethanol (TFE), ethanol-d5, and methanol-d3, which all should be of analytical spectroscopic grade. A minimum of 5% of deuterated solvent needs to be present to ensure a sufficient lock signal. NMR tubes should be clean and of a quality suitable for the instrument used.

Peptide 3D Structure Determination by NMR

145

2. Sample pH influences the peptide sample and the quality of data in several ways. First, the pH of the sample will affect the protonation state of amino acid side chains. At an acidic pH range carboxylic acids on Glu and Asp will be protonated, while at pH >4.5 they will be deprotonated and negatively charged. This will affect the solubility of your peptide and may or may not affect also the conformation of your peptide, depending on what interactions these side chains participate in. Second, the pH importantly affects the chemical exchange between solvent molecules and exchangeable protons in your peptide, e.g., the critical backbone amide protons. This exchange results in line broadening and loss of signal at more neutral or basic pH. Most peptide samples are exposed to acid during their HPLC purification resulting in a pH of 3–4 when dissolved in water, which tends to result in sharp line. Increasing the pH to ionize the carboxylic acids, by adjusting the pH to 5.0–6.0, may be desirable; however, if HN signals disappear assigning the spectra using the sequential walk will become difficult and information for structure determination is lost. If a solvent other than water is to be used, adjust the pH in water, re-lyophilize the peptide, and then dissolve in the desired solvent. 3. Experiments are typically first run at room temperature (298 K). Increased temperature leads to faster molecular tumbling, which may be advantageous for sharper lines, in particular for larger proteins. For small peptides however, a low temperature may be of advantage, as slower tumbling will favor the buildup of NOEs that are critical for the structure determination. The chemical shift of an amide proton is temperature dependent and by adjusting temperature peak overlap in the amide region may be resolved. Additional TOCSY and NOESY experiments run at different temperatures (275–320 K) may therefore be helpful for complete sequential assignments. 4. In order to achieve the narrowest lines possible and thus a wellresolved spectrum, shimming of the sample is required. Shimming adjusts the resolution of the signal by optimizing the homogeneity of the magnetic field. Shim coils surrounding the sample each produces a magnetic field with a particular spatial profile that can be used to cancel out any inhomogeneities in the main magnetic field. The current of these magnetic fields is adjusted until homogeneity has been achieved which can be observed through a spectrum possessing sharp lines. 5. Tuning and matching optimize the band-pass filters of the circuits of the NMR probe, which is achieved by changing the capacitance of tuning and matching capacitors, altering the band and efficiency of the filters. It is essential for optimizing the pulses and signals of the NMR experiment, especially for

146

Christina I. Schroeder and K. Johan Rosengren

higher frequencies. This can be done automatically with newer instruments or manually with older instruments. The tuning optimizes the frequency until it coincides with the frequency of the pulses transmitted, much like tuning a radio station to minimize the noise. The matching alters the efficiency and minimizes the use of power transmitted. Tuning and matching are not independent of each other and need to be adjusted iteratively for best results. Correct tuning and matching lead to shorter pulse lengths and better signal to noise. 6. The length of the p1 90
1H pulse should be calculated for each sample in order to maximize the transfer of the net magnetization from Z into the X, Y plane during the experiment. A p1 90
pulse too short or too long will lead to loss in observable signal. The p1 90
pulse only needs to be calculated once for NMR spectra run in sequence on the same sample and can be copied to all the other experiments within that sequence. With the high performance of modern amplifiers it is also possible to calculate the appropriate length and power of other required pulses such as inversions or “soft” selective pulse based on the knowledge of the length and power of the 90
pulse. 7. If the solvent of choice contains non-deuterated atoms, it is necessary to use solvent suppression, as the solvent signal will overpower the signal of the sample of interest. Therefore, when running a sample in 90% H2O/10% D2O, it is crucial to suppress the water signal in order to observe signal from the sample. Water suppression is temperature dependent as both the position and line shape of the water signal are affected by increased or decreased temperature. Accurate setup of the frequency to be suppressed, guided by the O1 offset, is therefore required to ensure adequate water suppression. The introduction of cryoprobes has been a revolution in terms of gain in sensitivity; however, using a cryoprobe can lead to difficulties with solvent suppression. A few of the common water suppression pulse sequences used are presaturation, Watergate [28], and excitation sculpting with gradients [4]. Modern pulse programs typically use gradient-based suppression schemes. 8. A 1D NMR experiment can provide useful information, including the number of spin systems, resonance frequencies, line widths, and coupling constants. If every peak in the 1D spectrum is resolved, the number of peaks will be the same as the number of protons in the peptide. The resonance frequency of each of the protons can provide information about how shielded the nucleus is and also give an initial indication of the environment in which the proton exists. For example, a widely dispersed HN region suggests the presence of ordered secondary and tertiary structure (Fig. 1). Line widths can provide information about chemical and conformational

Peptide 3D Structure Determination by NMR

147

exchange in the peptide since regions of conformational exchange give rise to line broadening. General broadening can also be a sign of aggregation of the peptide. An obvious problem with 1D experiments for peptides larger than a few residues is that there is often signal overlap. Therefore 2D experiments are needed to resolve this overlap using a second frequency domain. Despite problems with signal overlap for larger peptides and proteins, an initial 1D spectrum can still be useful in providing a quick indication of the sweep width required, whether the peptide is correctly folded or not and if there are more than one conformation present, as this will increase the number of signals, and therefore remain a valuable first tool before undertaking 2D experiments. 9. The sweep width determines the range of chemical shifts for which data are collected. In the first instance, run a 1D with a sweep width of at least 15 ppm to ensure that any unusually downfield shifted amides, Trp indole protons, or upfield shifted methyl groups are not excluded. Once the maximum sweep width required has been established by running a 1D 1H NMR, this can be transferred to 2D experiments. For TOCSY and NOESY experiments, 12  12 ppm is usually sufficient. For 1 H–13C HSQC and 1H–15N HSQC the sweep width of the indirect dimension becomes important as the larger the sweep width, the more increments are required for sufficient resolution, which leads to long experiments, in particular if recorded at natural abundance of 13C or 15N. For 15N data typically a sweep width of 35 ppm (~100–135 ppm) will cover the range of backbone amides and Asn/Gln side chains, while for 13C data a sweep width of 80 ppm (~3–77 ppm) will cover any internal reference standard and the Cα and aliphatic side-chain carbons. Aromatic carbons are outside this range and will appear folded in the spectra; however, their chemical shifts are generally not of interest for structure determination of peptides. 10. The choice of number of scans for each experiment is determined by factors such as sample concentration, length of experiment, and availability of instrument time. If working with a concentrated sample, then fewer scans are required, whereas a dilute sample would need more scans to improve the signal to noise. Obviously, the more scans used, the longer the experiment time, which requires longer access to the spectrometer. A good way to determine how many scans are needed for an experiment is to run a couple of 1D experiments and analyze the signal to noise. A doubling of scans increases the signal to noise by a factor of √2. Note that when increasing the number of scans it is important to take note of pulse program requirements as phase cycles may require the number of scans to be a multiple of an integer n.

148

Christina I. Schroeder and K. Johan Rosengren

11. To maximize the signal-to-noise ratio, the receiver gain, which represents the window of the analog-to-digital converter (ADC), needs to be optimized. At low receiver gain settings only a fraction of the digitization levels are used, leading to that signal and noise are poorly distinguished. However, at a too high setting the signal will exceed the level of the ADC leading to clipping and severe distortions in the FID. Thus for dilute samples that need a higher receiver gain a strong solvent suppression is required. An optimal receiver gain can be determined automatically by the spectrometer but needs to be determined for each individual NMR experiment acquired on the same sample, and thus cannot be copied from one experiment to the next. 12. In 2D NMR spectroscopy the spectral information is effectively spread in two frequency domains, thereby reducing but not always eliminating overlap. Different types of pulse sequences allow different types of interactions between nuclei to be detected. 2D experiments can be divided into different categories. Experiments such as correlation spectroscopy (COSY) and total correlation spectroscopy (TOCSY) utilize scalar spinspin connectivities (Fig. 13), while experiments such as nuclear Overhauser effect spectroscopy (NOESY) and rotating frame Overhauser effect spectroscopy (ROESY) detect dipolar spinspin connectivities. Both categories are necessary for assignment and structure determination. Heteronuclear single quantum coherence/correlation (HSQC) experiments are also common. These are highly sensitive experiments due to the strong one-bond scalar coupling but their use is limited by the fact that the natural abundance of 13C and 15N is only ~1% and ~0.5%, respectively. However, with access to higher field spectrometers and more sensitive probes, natural abundance 15N and 13C data are now generally achievable. 13. Total correlation spectroscopy (TOCSY) is a valuable experiment to define the whole “spin systems” for individual amino acids in a peptide or protein [5]. The TOCSY mixing scheme allows for multiple coherence transfer steps and thus all the protons within an amino acid that are within a scalar-coupled pathway are correlated to each other in a TOCSY spectrum. The presence of quaternary carbons or carbonyl groups prevents scalar coupling. Due to the fact that carbonyl groups separate adjacent amino acids, each amino acid will have an isolated coupling pattern, and aromatic protons are also generally separated from the rest of their amino acid resonances. 14. The nuclear Overhauser effect spectroscopy (NOESY) provides information of the relative distances of protons in a molecule [6, 7]. The Overhauser effect decreases at a rate proportional to 1/r6 and the observed cross peak intensity

Peptide 3D Structure Determination by NMR

149

can therefore be related to a distance between two protons that are 5 A˚ apart. The NOESY experiment is an integral part of structural determination of peptides and proteins since it provides crucial information about sequential connectivities as well as inter-proton distances from atoms that are distant in the primary sequence. These inter-proton distances together with additional information such as dihedral angles and hydrogen bonds form the basis of the 3D structure calculations. 15. Both the sign and strength of the NOE are dependent on the molecular tumbling, or correlation time, and thus at a crossover point where the NOE changes from being positive to negative, referred to as the extreme narrowing limit, cross peak intensities in NOESY spectrum are weakened or even nulled. This is commonly observed for small peptides (1–2 kDa) in aqueous solution. The rotating frame Overhauser effect spectroscopy (ROESY) [8, 9] overcomes this problem by incorporating a spin-lock sequence into the mixing time, while transverse relaxation occurs. The spin-lock field is kept weak to avoid scalar connectivity breakthrough. The sign of the ROE is in contrast to the NOE always positive. 16. The DQF-COSY can provide information about connectivities of atoms via scalar couplings and, by allowing the measurement of the strength of the couplings, about dihedral angels. Cross peaks in a DQF-COSY spectrum only appear between protons that are separated by three or fewer covalent bonds (Fig. 13), and thus it is very useful for the assignment of individual spins within longer or aromatic amino acid side chains in peptides. Due to its anti-phase multiplet pattern, which reports on, for

Fig. 13 Scalar spin-spin connectivities seen around aromatic rings in (a) correlation spectroscopy (COSY) data and (b) total correlated spectroscopy (TOCSY) data

150

Christina I. Schroeder and K. Johan Rosengren

example the 3JHN-Hα coupling constant, the DQF-COSY can also be used to derive dihedral restraints. 17. In a DQF-COSY spectrum it is difficult to directly determine the 3JHα–Hβ coupling constants due to the complexity of the multiplet pattern resulting from the Hα–Hβ1–Hβ2 spin interactions; however it is possible to de-convolute this by an exclusive COSY (E.COSY) spectrum [11]. Given that the Hα protons are the key resonances to be analyzed in this spectrum it is preferred to run the experiment in D2O to avoid interference of the water signal. The E.COSY separates and allows for the measurement of both passive and active 3JHα–Hβ coupling constants, which together with a short-mixing-time NOESY can be used to obtain stereospecific assignments and χ 1 restraints (Fig. 11). 18. A HSQC experiment provides a 2D spectrum correlating protons (1H) on one axis with their attached heteronucleus (typically 13C or 15N) on the second axis. A HSQC spectrum will contain one peak for each unique proton attached to the heteronucleus. In the 1H–15N HSQC each amino acid will have one peak resulting from the amide group of the peptide bond, except Pro, which lacks an amide proton. Nitrogenbound side-chain protons can also be observed. For isotopically enriched samples, a 1H–15N HSQC experiment is usually the first and the most common heteronuclear experiment run as it can provide information about whether the peptide is folded or have disordered regions as this will result in a welldispersed spectrum or clusters of overlapped peaks, respectively. A 1H–13C HSQC provides correlation between protons attached to carbons. A 1H–13C HSQC experiment is best run in 99.96% D2O as the water signal tends to overlap with Hα–Cα signals. A 1H–13C HSQC is very useful as a complementary experiment to the DQF-COSY in confirming assignments of side chain protons with similar chemical shifts but where the 13C shifts of their respective heteroatoms are distinct (Fig. 7). The chemical shifts from both 1H–15N HSQC and 1 H–13C HSQC can also be used for deriving dihedral restraints using TALOS-N (see below). 19. Monitoring the temperature dependence is another method to delineate HN protons that are protected from the solvent and that may participate in hydrogen bonds. In a peptide that does not experience any type of fold change as a result in temperature the amide proton chemical shift should be linearly dependent on the temperature. When the chemical shift is plotted as a function of temperature, amide protons that move less than 4.6 ppb/K are considered protected and with a probability of ~85% are involved in hydrogen bonds [29] (Fig. 14).

Peptide 3D Structure Determination by NMR

0 -2

151

Shielded

TC D ppb/K

-4 -6 -8 -10 -12 -14

GVIA [A2]GVIA

Non-shielded

-16 2

5

7

9

12

14

16

19

20

23

25

27

Fig. 14 Amide proton temperature coefficient values (ppb/K) for omega-conotoxin GVIA and [A2] GVIA derived from variable temperature experiment (280–320 K). HN protons with a temperature dependence of 95%) of all picked peaks should be able to be appropriately assigned for the structure output to be reliable. During the AUTO process seven iterative cycles of calculations are performed. Initially only peaks that can be unambiguously assigned based on unique chemical shifts are included, but after each round the preliminary structures are used to exclude assignment possibilities based on observed proton distances, thus allowing more assignments to be included. Final assigned “peak” lists are provided as an output that can be reanalyzed in the spectral data and used as input for additional CYANA calculations. Generally, an initial family of 50–100 structures is calculated, from which 20 structures with the lowest target function are chosen for further analysis. An iterative approach where structures are analyzed and the data revisited in order to resolve violations of the experimental data, unfavored covalent structure, or poor resolution in certain regions is often required. Once NOE assignments have been completed and a convergent structure achieved further refinement can be done by incorporation of additional structural information such as hydrogen bonds and χ 1 dihedral restraints, which generally also involve introducing stereospecific assignments of sidechain protons. For hydrogen bonds that are identified in preliminary structures, restraints can be included for hydrogen bonding pairs where the amide proton shows slow exchange with the solvent (present after >4 h following dissolution in D2O), or if the amide temperature coefficient suggests a hydrogen bond (Fig. 14). These restraints are included as upper limit distances in a separate hydrogen bond file. χ 1 dihedral restraints are added directly to the dihedral-angle restraint file (aco), while stereospecific assignments are defined in the CALC or AUTO input protocols resulting in that pseudoatom treatments are not applied for these stereopairs. 33. The CNS program allows for simulated annealing calculations using either torsion angle or Cartesian coordinate-based molecular dynamics. The advantage with the Cartesian approach is that is allows inclusion of water molecules and other molecular entities that are not part of the covalent system defined exclusively by torsion angles. A flowchart outlining the different steps involved in the CNS structure calculations strategy applied in our laboratories is presented in Fig. 15. Initially information regarding amino acid sequence, disulfide bonds, or other modifications is combined with the molecular information from the topology file (details about atoms, bonds, angle, dihedrals for different residues, and different features such as peptide and disulfide bonds) in order to create a

158

Christina I. Schroeder and K. Johan Rosengren

Fig. 15 Overview of the simulated annealing protocol used for NMR structure calculation and refinement in CNS. Information about the protein sequence, any posttranslational modifications, and the nature of the different amino acids are combined to create a molecular template file defining the system, including information about the presence of atoms, bonds, and angles. The force field parameter file provides the ideal values and force constants defining “energies” of covalent and nonbonded interactions. A random low-energy starting structure is generated and used for the simulated annealing during which the protein is heated and then slowly cooled while experimental restraints are applied. After the initial calculations a water shell is added to the structure, which is refined, and energy is minimized through further restrained molecular dynamics. Final structures are analyzed for quality and structural features

Peptide 3D Structure Determination by NMR

159

molecular template file. After addition of empirical parameters defining these covalent geometries (bond lengths, bond angles, dihedrals, etc. and the force constant applied to them) a low-energy extended random conformation starting structure is computed. The topology and parameter files provided within the CNS program can if required be modified to include unnatural amino acids, such as hydroxyproline. Using the energy-minimized extended starting structure together with the validated NMR restraints the CNS-simulated annealing and energy minimization protocols can be employed [23, 24]. For the initial high-temperature part generally a torsion-angle approach is preferred. Simulated annealing uses a simplified energy function (one term describes all nonbonded interactions) and a low cutoff distance (energies are only calculated between atoms within a certain distance of each other) to reduce computational time. The energy of the system is initially raised (controlled by a temperature parameter) which allows for increasing kinetic energy of the simulated atoms, followed by slow cooling of the system. This aims to ensure that local energy minima are overcome to locate the global energy minimum Etot (Eq. 2). After the initial structure calculations a second stage of refinement and energy minimization is performed. For this part the preliminary structures are enclosed in a box of explicit solvent molecules. Energy is again put into the system in the form of mild heating and Cartesian-based molecular dynamics calculations performed at elevated temperature and during a slow cooling phase, which is followed by Powell minimization. The inclusion of solvent allows for inclusion of more accurate electrostatic interactions during the calculations. 50–100 structures are typically calculated and analyzed for errors and ambiguities. NOE and dihedral violations may have to be resolved through structural analysis and additional rounds of calculations. Structures are “accepted” when no or few NOE and ˚ and 2.0
, respectively, are present. dihedral violations of 95%). 2. 24-Well cell culture plates (manual hanging-drop vapor diffusion crystallization method). 3. 96-Well crystallization plates (sitting-drop method). 4. Siliconized glass cover slides, vacuum grease, crystallization screens. 5. Micropipettes with 10 μL and 1 mL pipette tips. 6. Buffers (0.02 M Tris–HCl pH 8.0, 1 M NaCl), (0.02 M glycine-HCl, pH 3.0), and (1 M Tris–HCl pH 9.0). 7. Columns for molecular size exclusion (Sephacryl S-100 and S200), ion-exchange (Mono Q and Mono S, Source Q and Source S, etc.), and affinity (Ni-sepharose, Benzamidine Sepharose 4 Fast Flow) chromatographies. 8. Akta Pure, Akta Avant, etc. 9. Centricon concentrator from Amicon or GE Healthcare Life Sciences.

3

Methods

3.1 Protein Purification

For snake venom toxins (crude venom) the individual proteins are purified by using molecular size exclusion, ion-exchange, and affinity chromatographies at 18  C usually in a cold room to avoid the formation of bubbles in the chromatographic columns [5, 7]. These methods are usually modified as the snake venom mostly contains proteins with isoforms [14]. The purification steps involved are the following: 1. Solubilize the crude venoms (10–100 mg/mL) in buffer (0.02 M Tris–HCl, 0.1 M NaCl). 2. Centrifuge it at 10,000  g for 10 min in a refrigerated centrifuge. Set the temperature to 4  C. 3. Apply the supernatant to molecular size-exclusion chromatography using column Sephacryl S 100 or 200 attached to a purification system. 4. Set the method to run the chromatography, i.e., flow rate at 0.2–0.5 mL/min, sample collection volume at 1 mL/tube, and UV absorbance at 280 nm. 5. Once the chromatography has been completed check the purity of the proteins under each peak using the SDS-PAGE. 6. If the protein of interest is obtained in pure form (at least >95% for crystallization), well and good.

X-rays Crystallography of Snake and Spider Venom Toxins

165

7. If the protein of interest is still contaminated with other proteins, move to the next step of purification, i.e., ion-exchange or affinity chromatography. 8. For ion-exchange chromatography pool (mix) all the tube under the peak that contains the protein of interest contaminated with the other proteins. 9. Remove the salt (desalting by using the desalting column or dialysis membrane). 10. Prepare two buffers, namely A and B: buffer A (0.02 M Tris–HCl), B (buffer A + 1 M NaCl). 11. Load the sample to ion-exchange column using the buffer A. Once all the sample has been applied, change the buffer to buffer B and elute the bound sample/proteins with linear or nonlinear concentration gradient of NaCl (0–100%). 12. After completion of the method, analyze the purity of the proteins under each peak using the SDS-PAGE. 13. The affinity chromatography is usually carried out for the proteins that bind to specific ligands like benzamidine bound to the resin, which specifically bind serine proteinases. 14. The sample is applied to the column using a buffer with high pH (usually 0.02 M Tris–HCl, pH 7.5), and the bound protein (s) is (are) eluted with low pH (usually 0.02 M glycine-HCl, pH 3.0). 15. To avoid the eluted sample from denaturation by low pH adjust the pH immediately to neutral by adding 100 μL of 1 M Tris–HCl pH 9.0. 16. If the protein obtained from affinity chromatography is not pure or contains isoforms, ion-exchange chromatography is again performed in order to get it in the desired level of purity. 17. Most often, a final step of molecular size-exclusion chromatography is performed on the pure protein to remove the salt and traces of impurities and make it suitable for crystallization. 18. The spider venom toxins are usually cloned, expressed, and purified using the affinity chromatography on Ni-sepharose, which specifically binds proteins with His-tag [9, 11]. The elution of the bound protein is carried out using a buffer containing imidazole (0.5 M). The method is the same as described for the purification of serine proteinases from snake venom discussed above. 3.2

Crystallization

1. Concentrate (using Centricon concentrator) or lyophilize the pure protein (purity >95%) to at least 10 mg/mL. Add 0.02% sodium azide solution to protect your protein solution from bacterial and fungal growth and 3–5% 1,2-propanediol in the protein buffer to stabilize it.

166

Anwar Ullah et al.

2. Set up the crystallization experiment. 3. Put 200–500 μL of the crystallization solution in the well. 4. Take 0.5–2 μL of the concentrated pure protein and mix it with an equal amount of crystallization solution, on a siliconized coverslip. 5. Invert this coverslip and place it over well with the solution. 6. Seal the crystallization plate and keep it at a temperature from 4 to 18  C. 7. Observe the plate after 24 h and check if the crystals have been formed. 8. If not then check it after 3 days, 1 week, and then 1 month. 9. Usually, small crystals or precipitates are observed in the first trial, which needs to be improved by changing the crystallization solution properties such as the buffer pH, concentration of precipitant, and temperature. 3.3 Data Collection, Processing, and Structure Solution

1. Harvest and mount a single crystal with a loop. 2. Use a cryoprotectant like polyethylene glycol (PEG) (40%) [18] or glycerol (25–50% v/v) solution to protect the crystal from radiation damage. 3. Flash cool the crystal in a 100 K nitrogen gas stream. 4. Collect the X-ray diffraction data on a synchrotron radiation beamline. 5. Expose the crystal for 30–60 s per 1–2 degree rotation in ø with the crystal-to-detector distance set to 50–100 mm. 6. Index, integrate, and scale the data by using DENZO and SCALEPACK programs from the HKL-2000 package [19]. 7. This indexing, integration, and scaling can also be completed using Mosflm [20] or XDS [21]. 8. The initial structural model can be determined by molecular replacement using the program MOLREP [22] from CCP4. 9. Model refinement can be carried out by alternating cycles of REFMAC5 from CCP4 [22]. 10. The visual inspection of the electron density maps and manual rebuilding can be done with COOT [23]. 11. The structure is analyzed using various molecular visualization software like PyMOL [24] and Chimera [25].

4

Notes 1. Purification of proteins from the snake and the spider venoms may sometime become a challenge, due to the presence of

X-rays Crystallography of Snake and Spider Venom Toxins

167

isoforms of these proteins. However, using a modified method in each step of purification can be useful to solve these problems [7]. 2. In molecular size-exclusion chromatography always use a slow flow rate (maximum 0.2 mL/min) and apply the minimum sample volume (0.2–1 mL) (Fig. 1a, b) [7, 8]. 3. In ion exchange (both anion and cation exchange) the resolution of peaks and the protein purification can be improved by giving a nonlinear gradient of NaCl (5%, 10%, 15%, 20%, 30%, 40%, 50%, 70%, and final 100%) (Fig. 1c, d) [7, 14] than a linear one. The option hold in the chromatographic system can be used to achieve this. For example, use the hold option at 5% NaCl concentration (buffer B) and wait till the UV baseline becomes stable. Then continue the method up to 10% and hold the method at this percentage as described previously. In manual separation method the technique can be done using buffer with 50 mM, 100 mM, 150 mM, 200 mM, 250 mM, 300 mM, 400 mM, 500 mM, 600 mM, 700 mM, and 1 M, each one column volume for elution purposes. These modified methods significantly improve the purification of the target protein. 4. The above method is also applicable to affinity chromatography (both pH and imidazole elution) (Fig. 1e, f). Hold option can be used for example to give a change in pH 7, 6, 5, 4, and 3. The method for imidazole elution is the same as discussed above for the ion-exchange chromatography [5]. 5. Protein purification can be improved in affinity chromatography using 500 mM NaCl in the loading buffer (Fig. 1e, f). This solution is applied to the column after all the sample has been loaded to the column prior to the elution of target protein. This method removes the proteins nonspecifically bound to the affinity resin. If isoforms of the protein are present after the affinity column purification, then another step of ion-exchange chromatography may be carried out as discussed above (Fig. 1g, h) [14]. 6. It is useful to check the homogeneity of your pure protein by dynamic light scattering if the facility is available (Fig. 2a, b) [7]. 7. For crystallization of the pure protein, the desired concentration varies from 10 to 20 mg/mL for the snake and spider venom toxins. The actual concentration required however can be found by using some kits called pre-crystallization solutions (Hampton Research). These kits used some very common precipitant like ammonium sulfate and polyethylene glycol (PEG) to check the precipitation properties of the concentrated proteins. Light precipitate is preferred over heavy one. No precipitation after 24 h means the concentration of protein

168

Anwar Ullah et al.

Fig. 1 Chromatographic profile for purification of toxins from Bothrops venom. (a) Molecular size-exclusion chromatography on Sephacryl S 200 column. (b) SDS-PAGE analysis of peak fractions, lane M: molecular weight markers (kDa),

X-rays Crystallography of Snake and Spider Venom Toxins

169

is insufficient for crystallization and its need to be further concentrated [8]. 8. In crystallization trails most often the crystals obtained are small (Fig. 2c, e, f) and not suitable for diffraction data collection. It is desirable to optimize the properties (pH, concentration of salt, and other precipitant like PEG) in order to get good and large crystals (Fig. 2d, g) for good diffraction data collection. It is usually achieved by changing the pH (0.1–0.3 point difference below and above of the original one); for example if the original pH was 5 then in the optimization processes use 4.0, 4.1, 4.2, 4.4, 4.6, 4.8, 4.9, 5.1, 5.2, 5.4, 5.7, 6.0, 6.2, and 6.5 and the precipitant concentration difference by 5%: for example if the original precipitant concentration was 25%, use 15%, 20%, 25%, 30%, 35%, 40%, 45%, and 50%. These modifications considerably improve the quality of the crystals [11]. 9. In the crystallization experiment, it is not a good idea to mix both the solution (precipitation solution and protein solution) thoroughly; rather they are placed just near to each other in order to diffuse to one another slowly. This practice increases the chances of crystallization (Fig. 3a–k) [12]. 10. If the crystal is suitable for diffraction data collection (large enough) and it is not diffracting well, then annealing can be done, in which the cooled crystal is kept at room temperature for approximately 30–60 s, and then rapidly transfer it to cryogenic temperature [13].

ä Fig. 1 (continued) TP: total proteins (crude venom) 1a to 4a corresponding peak fractions. Lanes 3a and 4a represent pure phospholipases A2 (dimer and monomer, respectively). (c) Ion-exchange (anion-exchange) chromatography profile for purification of metalloproteinase (class III) and L-amino acid oxidase. (d) SDS-PAGE analysis of peak fractions, lane M: molecular weight markers (kDa), 1b to 4b corresponding peak fractions. Lanes 3b and 4b represent pure metalloproteinase III and L-amino acid oxidase, respectively. (e) Affinity chromatography (benzamidine sepharose 4B (Fast Flow) profile of serine proteinase. (f) SDS-PAGE analysis of peak fractions, lane M: molecular weight markers (kDa), 1c to 2c corresponding peak fractions. Lane 3c represents serine proteinase isoforms. (g) Ion-exchange (cation-exchange) chromatography profile for purification of serine proteinase isoforms. (h) SDS-PAGE analysis of peak fractions, lane M: molecular weight markers (kDa), TP total proteins (eluted from affinity chromatography), 1d and 2d corresponding peak fractions. Lane 1d represents pure serine proteinase (M. weight ~32 kDa) and lane 2d pure serine proteinase (M. weight ~28 kDa), respectively

170

Anwar Ullah et al.

Fig. 2 Dynamic light scattering correlogram for (a) an aggregated (polydisperse) protein. (b) Monodisperse (nonaggregated) protein. (c) Hairlike crystals of an L-amino acid oxidase (before optimization). (d) Well-formed crystals of an L-amino acid oxidase after optimization. (e) Hairlike crystals of serine proteinase (before optimization). (f) Small crystals of serine proteinase after optimization (first round). (g) Well-formed crystals of serine proteinase after optimization (second round)

X-rays Crystallography of Snake and Spider Venom Toxins

171

Fig. 3 Schematic representation of protein structure determination starting from the crude sample (toxins). a–k represent steps involved in the process

11. Once the diffraction data is collected the structure of the desired protein is determined as discussed above and shown in Fig. 3a–k in flowchart. References 1. Kang TS, Georgieva D et al (2011) Enzymatic toxins from snake venom: structural characterization and mechanism of catalysis. FEBS J 278:4544–4576 2. Ullah A, Masood R, Ali I, Ullah K, Ali H, Akbar H, Betzel C (2018) Thrombin-like enzymes from snake venom: structural characterization and mechanism of action. Int J Biol Macromol 114:788–811 3. Gremski LH, Trevisan-Silva D et al (2014) Recent advances in the understanding of

brown spider venoms: from the biology of spiders to the molecular mechanisms of toxins. Toxicon 83:91–120 4. Fernandes-Pedrosa MF, Junqueira-de-Azevedo ILM, Gonc¸alves-de-Andrade RM et al (2008) Transcriptome analysis of Loxosceles laeta (Araneae, Sicariidae) spider venomous gland using expressed sequence tags. BMC Genomics 9:279 5. Ullah A, Souza TA et al (2013) Crystal structure of Jararacussin-I: the highly negatively

172

Anwar Ullah et al.

charged catalytic interface contributes to macromolecular selectivity in snake venom thrombin-like enzymes. Protein Sci 22:128–132 6. Zhu Z, Liang Z et al (2005) Crystal structures and amidolytic activities of two glycosylated snake venom serine proteinases. J Biol Chem 280:10524–10529 7. Ullah A, Souza TA, Abrego JR, Betzel C, Murakami MT, Arni RK (2012) Structural insights into selectivity and cofactor binding in snake venom L-amino acid oxidases. Biochem Biophys Res Commun 421:124–128 8. Ullah A, Souza TA, Betzel C, Murakami MT, Arni RK (2012) Crystallographic portrayal of different conformational states of a Lys49 phospholipase A2 homologue: insights into structural determinants for myotoxicity and dimeric configuration. Int J Biol Macromol 51:209–214 9. de Giuseppe PO, Ullah A et al (2011) Structure of a novel class II phospholipase D: catalytic cleft is modified by a disulphide bridge. Biochem Biophys Res Commun 409:622–627 10. Murakami MT, Fernandes-Pedrosa MF et al (2006) Structural insights into the catalytic mechanism of sphingomyelinases D and evolutionary relationship to glycerophosphodiester phosphodiesterases. Biochem Biophys Res Commun 342:323–329 11. Ullah A, Magalha˜es GS, Masood R, Mariutti RB, Coronado MA, Murakami MT, Barbaro KC, Arni RK (2014) Crystallization and preliminary X-ray diffraction analysis of a novel sphingomyelinase D from Loxosceles gaucho venom. Acta Crystallogr F Struct Biol Commun 70:1418–1420 12. Ullah A, Masood R, Spencer PJ, Murakami MT, Arni RK (2014) Crystallization and preliminary X-ray diffraction studies of an Lamino-acid oxidase from Lachesis muta venom. Acta Crystallogr F Struct Biol Commun 70:1556–1559 13. Ullah A, de Souza Tde A, Masood R, Murakami MT, Arni RK (2012) Purification, crystallization and preliminary X-ray diffraction analysis of a class P-III metalloproteinase (BmMP-III) from the venom of Bothrops moojeni. Acta Crystallogr Sect F Struct Biol Cryst Commun 68:1222–1225

14. Fernandes de Oliveira LM, Ullah A, Masood R, Zelanis A, Spencer PJ, Serrano SM, Arni RK (2013) Rapid purification of serine proteinases from Bothrops alternatus and Bothrops moojeni venoms. Toxicon 76:282–290 15. Coronado MA, Georgieva D, Buck F, Gabdoulkhakov AH et al (2012) Purification, crystallization and preliminary X-ray diffraction analysis of crotamine, a myotoxic polypeptide from the Brazilian snake Crotalus durissus terrificus. Acta Crystallogr Sect F Struct Biol Cryst Commun 68:1052–1054 16. Coronado MA, Ullah A, da Silva LS et al (2015) Structural insights into substrate binding of brown spider venom class II phospholipases D. Curr Protein Pept Sci 16:768–774 17. Masood R, Ullah K, Ali H, Ali I, Betzel C, Ullah A (2018) Spider’s venom phospholipases D: a structural review. Int J Biol Macromol 107 (Pt A):1054–1065 18. Berejnov V, Husseini NS, Alsaied OA, Thorne RE (2006) Effects of cryoprotectant concentration and cooling rate on vitrification of aqueous solutions. J Appl Crystallogr 39:244–251 19. Otwinowski Z, Minor W (1997) Macromolecular crystallography. Part A. Academic Press, New York, pp 307–326 20. Battye TGG, Kontogiannis L, Johnson O, Powell HR, Leslie AGW (2011) iMosflm: a new graphical interface for diffraction-image processing with MOSFLM. Acta Cryst D67:271–281 21. Kabsch W (2010) XDS. Acta Cryst D66:125–132 22. Vagin AA, Teplyakov A (1997) Molrep: an automated program for molecular replacement. J Appl Crystallogr 30:1022–1025 23. Emsley P, Lohkamp B, Scott WG, Cowtan K (2010) Features and development of Coot. Acta Crystallogr D Biol Crystallogr 66:486–501 24. DeLano WL (2002) The PyMOL molecular graphics system. DeLano Scientific, San Carlos 25. Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM, Meng EC, Ferrin TE (2004) UCSF Chimera-a visualization system for exploratory research and analysis. J Comput Chem 25:1605–1612

Chapter 9 MALDI-TOF Mass Spectrometric Profiling of Spider Venoms Ondrej Sˇedo, Stano Peka´r, and Zbyneˇk Zdra´hal Abstract Fingerprinting by means of matrix-assisted laser desorption-ionization time-of-flight mass spectrometry (MALDI-TOF MS) represents a tool for rapidly detecting proteinaceous compounds from spider venoms. Here we describe an optimized protocol and discuss methodological details with the aim of providing a platform for obtaining the most informative and reproducible mass spectral data. Key words Spider venom, MALDI-TOF MS, Proteins, Fingerprinting

1

Introduction The rapid development of separation and mass spectrometric techniques over the past three decades has revealed new opportunities for gaining deeper insights into the protein composition of various types of biological materials [1]. Because a major fraction of bioactive compounds present in spider venoms is attributed to proteinaceous compounds, including both peptide and protein components [2], current methodologies offer a platform for correlating compositional data with properties of particular venoms. This may facilitate further exploitation of these compounds in pharmacology and other disciplines. Generally, the coverage of proteinaceous compounds in samples depends strongly on the methodical setup selected. On the one hand, a combination of multidimensional separation/enrichment steps combined with supervised analytical approaches may result in the identification of thousands of proteins from a single sample [3]. In the case of spider venom analyses, the application of these advanced proteomics tools, based mainly on a final analytical step using liquid chromatography-tandem mass spectrometry (LC-MS/ MS), is frequently limited due to the absence of known spider protein sequences in available databases. As routine MS/MS data treatment is based on their comparison with known protein sequences, the identification success rate may be low in contrast

Avi Priel (ed.), Snake and Spider Toxins: Methods and Protocols, Methods in Molecular Biology, vol. 2068, https://doi.org/10.1007/978-1-4939-9845-6_9, © Springer Science+Business Media, LLC, part of Springer Nature 2020

173

174

Ondrej Sˇedo et al.

to the rich amount of mass spectral information obtained from the analysis. Without enhancement of the protein database or additional information on expected protein sequences from comprehensive transcriptomic or genomic analyses, the identification of proteins relies on de novo peptide sequencing directly from the MS/MS data. Despite the development of software tools, this is based mostly on manual evaluation of the mass spectra, which is very time consuming. The relevancy of de novo sequencing outputs is limited due to the absence of firm criteria for validating the results. Therefore, the outputs of MS/MS-based de novo sequencing should be accompanied by sequence data obtained using another independent method. On the other hand, there are other, much simpler, and more rapid methods that provide limited, but relevant information based on particular peptide/protein compositions that enables characterization of the samples. Among them, fingerprinting of compounds by means of matrix-assisted laser desorption-ionization time-offlight mass spectrometry (MALDI-TOF MS) represents a highthroughput and easy-to-use approach. This method has found use in routine microbial diagnostics, where thousands of clinical laboratories worldwide employ MALDI-TOF MS as the first method of choice for identification of bacterial and fungal pathogens [4, 5]. Principally, this method can also be applied to other types of biological materials [6, 7] and, with some limitations, also to the detection of compounds other than proteins. Due to ionization suppression effects taking place during the MALDI process, only a restricted number of compounds, mostly corresponding to those of lower molecular weight, higher basicity, and higher abundance, are detected among the entire set of compounds from the sample. Nevertheless, valuable biological information can be correlated with various sample properties obtained by this analytical approach, as in the case of spider venoms. Primarily, the MALDI-TOF MS fingerprinting analysis of spider venoms provides molecular weights for tens to hundreds of venom components [8]. This information from venom profiling has been used to distinguish species of spiders [9, 10] and ontogenetic shifts [11], or to characterize adaptation to prey [12, 13]. The sample preparation protocols for MALDI-TOF MS profiling analysis used in the literature differ in three main ways: the MALDI matrix compound, the matrix solvent composition, and the deposition of the sample onto the sample plate. For that reason, we recently conducted a study aimed at establishing a sample preparation protocol providing the most informative and reproducible MALDI-TOF mass spectra, thereby providing a platform for relevant interlaboratory comparative studies [14]. We have described this protocol in detail, together with a preceding approach to sample venom from the glands.

MALDI-TOF MS Profiling of Spider Venoms

2

175

Materials 1. MALDI matrix: Prepare saturated solution of alpha-cyano-4hydroxycinnamic acid (CHCA) in water:acetonitrile:TFA (47.5:50:2.5, v/v). Water of purity equal to doubly distilled and analytical grade chemicals should be used (see Note 1). 2. Select a mass calibration standard covering the m/z range of 2–20 kDa, e.g., Bacterial Test Standard (BTS) provided by Bruker Daltonics, Escherichia coli whole-cell protein extract, or hen lysozyme. 3. Use plastics (tips, vials) declared to be compatible with proteomics analysis (see Note 2). 4. Use a common stainless steel MALDI target compatible with the MALDI-TOF mass spectrometer. There are no specific requirements for the number or the size of sample wells. 5. Any type of MALDI-TOF mass spectrometer capable of linear positive detection mode can be used for the purpose of venom fingerprinting analysis (see Note 3). 6. Electrostimulation device (AC 1–12 V, 3–30 mA), 0.9% NaCl solution, and capillaries should be used for venom sampling.

3

Methods

3.1 Venom Sample Collection

The venom gland is paired and is situated in the basal segment of the chelicerae and often extends to the prosoma. There are two ways to sample venom. In species whose prosoma is at least 5 mm long and the cheliceral fangs are long, the venom can be sampled by means of electrostimulation [15]. In the case of smaller species or with very short cheliceral fangs, the venom cannot be obtained by electrostimulation; therefore the whole gland must be dissected.

3.1.1 Venom Sample Collection from Larger Species with Long Chelicelar Fangs

1. Anesthetize the spider with CO2 for at least 1 min (see Note 4). 2. Place it upside down on a metal stub that is connected by a wire to the electrical source. 3. Place a piece of gauze moistened with concentrated NaCl solution on the top of the stub. 4. Trigger the venom by touching the spider’s chelicera with the other electrode and sample it with the capillary (see Note 5). 5. Store the sample at
20  C prior to analysis.

3.1.2 Venom Sample Collection from Small Species with Short Chelicelar Fangs

1. Anesthetize the spider with CO2 for at least 1 min (see Note 4). 2. Pin it to the substrate, upside down, via the prosoma. 3. Remove the chelicera and take the glands out by means of fine pincers.

176

Ondrej Sˇedo et al.

4. Place the glands immediately in a drop of 0.9% NaCl, e.g., on a glass slide. 5. Move the glands through a series of drops for 5 min to wash away hemolymph and epithelial cells and other tissues on the surface of the glands. 6. Place the glands into 10 μL of physiological solution, compress with a pair of pincers, and then remove the tissue. 7. Store the sample at
20  C prior to analysis. 3.2 MALDI-TOF MS Analysis

1. Mix 0.6 μL of the venom sample with 2.4 μL of the MALDI matrix solution by pipetting up and down several times in a 200 μL sample vial. 2. Deposit 0.6 μL of this mixture onto three wells of the stainless steel MALDI target (see Note 6). 3. Allow to dry at room temperature. 4. Prepare a spot for mass calibration located close to the samples on the sample plate. Apply 0.6 μL of BTS (see Note 6) and, after being dried at room temperature, overlay with an equal volume of the MALDI matrix (see Note 7). 5. Introduce the sample plate into the MALDI-TOF mass spectrometer and set voltages to provide analysis in the linear positive detection mode with optimum signal quality in the mass range of 2–10 kDa (see Notes 8 and 9). Set the mass detection range to 2–30 kDa and ion deflection to 500 Da. 6. Carry out mass calibration by applying at least 200 laser shots onto the well with the calibrant. Calculate the calibration function by using known proteins of E. coli providing signals at m/ z ¼ 3637.8; 4346.3; 5095.8; 5380.4; 6254.4; 7273.5; and 10,299.1 (see Note 10). In the case of BTS, also add signals corresponding to protein standards at m/z ¼ 13,683.2 and 16,952.2 Da to the calibrant list. In the case of lysozyme as a mass calibration standard, m/z values used for calibration should be the following: 4769.3; 7153.5; and 14,306.2. After calculating the calibration function, the mass errors should be lower than 100 ppm; otherwise the calibration step should be repeated (see Note 11). 7. Determine the threshold laser power (the power at which the peaks start to appear) by a preliminary analysis using 200 shots on each sample with varying laser energies. 8. Carry out the remaining analyses at a laser power corresponding to approximately 120% of the determined threshold laser power. Most spider venoms provide intense signals within the m/z range of 3000–6000 (see Fig. 1).

MALDI-TOF MS Profiling of Spider Venoms

Intens. [a.u.]

Intens. [a.u.]

Intens. [a.u.]

Intens. [a.u.]

Intens. [a.u.] Intens. [a.u.]

4

⫻10 6 4 2 0 8000 6000 4000 2000 0 ⫻104 2 0 6000 4000 2000 0 ⫻104 1.0 0.5 0.04 ⫻10 0.5 0.0 2000

177

4735 3000

5063 5455

3781 4265 3431

7904

4006 3000 3835 2724 32493576

6306

4373

4719 5063 5455

6293 6090

4698 3000

2999

3677 4006

5455

3575 4331 4006 3576 3952 3781

2881

5064 4436

3572

4735 5063 4895

5455

6292 7904

6293 5986 6293

4331 4735 5063 5456

4416 5063 3910 4230 4698 5455 3151 3572 3000

5987

4000

5000

6090 6307 6103 6000

7000

8000

9000

10000

11000

m/z

Fig. 1 MALDI-TOF mass spectral fingerprints of venoms obtained from six Zodarion rubidum Simon individuals

9. The mass spectral acquisition can be carried out in an automatic arrangement with the laser aimed in a pre-defined serpentine or other raster. At least three independent mass spectra, comprising 1000 laser shots each, should be acquired from each of the sample wells. Within an individual well, a minimum of 200 and a maximum of 400 shots should be obtained from one raster position (see Note 12). 10. If available, the storage of mass spectra should be controlled by a software mechanism automatically accepting/rejecting the mass spectral accumulations on the basis of monitoring of selected mass spectral features (if this function is available in the acquisition software). These features should involve the intensity and the mass resolution, which should be monitored for the second highest signal in the mass spectrum within the m/z range of 4000–7000 (see Note 13). 11. If automatic mass spectral acquisition fails to provide mass spectra fulfilling the acceptance criteria specified above, manual mass spectral acquisition should be carried out using higher laser power and finding positions on the sample well that provide acceptable signals. 12. Transfer the data to suitable peak-picking software (for example Flex Analysis or Biotyper RUO from Bruker Daltonics). Only repeatable signals should be taken into account for further evaluation. For example, from nine mass spectra per sample (three accumulations per three wells) we recommend to take into account only signals present in at least seven (>67%)

178

Ondrej Sˇedo et al.

of the mass spectra. As relevant signals, those exceeding the signal-to-noise ratio S/N > 3 should be accepted. We do not recommend using any absolute or relative intensity thresholds that would decrease the informativeness of the output. 13. If noisy signals dominate in the low-mass range (usually at m/ z < 3000), the low-mass limit in the peak list should be restricted appropriately. There is no practical reason for setting any such limit for the high-mass range. 14. The resulting data format derived from the mass spectra is a table, containing a list of m/z values and intensities for each of the relevant signals, and which is easily transferable to any common statistics software (see Notes 14 and 15).

4

Notes 1. Use a CHCA preparation specified by the manufacturer as being suitable for MALDI-TOF mass spectrometry. Other commercially available preparations of CHCA may not provide good peptide/protein ionization or may need to be recrystallized for this purpose. The performance of CHCA slowly decreases after being repeatedly exposed to air and humidity. Use always small or aliquoted stocks of this compound. The CHCA solution is stable for 1 week when being stored at +4  C. 2. Leaching or possible solubilization of compounds from any components of the pipetting/sample storage material may have a negative impact on the MALDI-TOF mass spectral quality. 3. The peptide/protein fingerprinting does not require a state-ofthe art MALDI-TOF mass spectrometric instrumentation. Older instruments also provide sufficient mass spectral quality for this type of analysis. 4. To maximize the venom volume, it is recommended to sample the venom preferably from adult females. As the gland size is shaped by allometry, adult males and juveniles have smaller glands and therefore provide a small volume of venom. Furthermore, the venom should be sampled a few days after feeding to ensure that the content was not depleted. In species with a very thick cuticle of the prosoma, the gland is very difficult to dissect, and milking by means of electrostimulation is preferable. 5. The electric shock also triggers the production of saliva, which should not contaminate the venom sample. For this reason, it is essential that the distal fang is inserted inside a capillary of an appropriate size (which is possible only if it is long enough).

MALDI-TOF MS Profiling of Spider Venoms

179

6. The volume of the sample/matrix mixture may be modified according to the size of the sample well, which may differ between types of instruments and MALDI targets. 7. A cheaper calibration standard can be prepared from E. coli cells cultivated on a common growth medium. Harvest approximately 5–10 mg of the cellular material and inactivate the bacteria by vortexing for 1 min in 1.2 mL of an ethanol:water (75:25, v/v) mixture. After centrifugation, remove the supernatant. Add 100 μL of an acetonitrile:formic acid:water, 50:35:15 v/v, mixture to the pellet and vortex for 1 min. After centrifugation, the supernatant can be used as a standard for mass calibration. The pellets after bacterial inactivation can be stored at +4  C for months, and the extracts at
20  C for weeks without any decrease in mass spectral quality. Preferably, E. coli strain DH5α should be used; however, other E. coli strains usually provide very similar signals. As another alternative, hen egg lysozyme signals can be used for mass calibration. While the bacterial standards need to be overlaid by the matrix solution on the MALDI target to provide high-quality data, lysozyme should be deposited onto the MALDI target by using the same method as is used for the venoms (see Subheading 3.2, steps 1 and 2). 8. When using our Ultraflextreme instrument, voltage settings were the following: ion source 1: 20.0 kV; ion source 2: 18.7 kV; lens: 6.0 kV; and pulsed ion extraction: 130 ns. These values should be set specifically by skilled personnel or a company technician, as their optimal values differ between individual instruments, even those of the same brand. 9. Apart from linear positive arrangements, different ion detection modes are available with common MALDI-TOF mass spectrometers. However, despite much higher mass spectral resolution obtained, using reflectron detection mode significantly decreases the number and the mass range of compounds detected. Using negative ion detection modes greatly decreases the signal intensity of proteinaceous compounds, especially when using the MALDI matrix solution used in the protocol described. 10. Application of relatively higher laser power is frequently needed to obtain valuable calibration signals from E. coli extracts at m/z ¼ 10,299.1. 11. When using an older instrument, the limit of 100 ppm may not be achievable, most frequently for the signal at m/z ¼ 3637.8. 12. Due to the homogeneity of CHCA matrix crystals, the mass spectra can be accumulated from any position in the well without the need to find “sweet spots” (positions providing good signals). After applying more than 400 laser shots to the

180

Ondrej Sˇedo et al.

same position, the mass spectral quality may decrease. The setting of automatic runs may depend on parameters of individual instruments. For older instruments allowing only low laser frequency, the number of shots can be decreased to improve the analysis throughput. 13. For an Ultraflextreme instrument, the thresholds for acceptance of the mass spectra were set to mass resolutions greater than 300 and signal intensities greater than 600 a.u. The mass spectra were evaluated on the basis of these parameters after the accumulation of each 200 shots. These values should be set according to the performance of individual instruments. 14. An important parameter to be specified for further data evaluation is the mass error tolerance (the maximal m/z difference between two signals from different mass spectra that correspond to the same compound). In the case of the Ultraflextreme instrument, this value was set to 500 ppm. Older types of MALDI-TOF mass spectrometers may be expected to provide a lower mass accuracy that should be then reflected in the data evaluation algorithm. On the other hand, an overly increased mass tolerance may deteriorate the discriminatory power of the method. 15. The relative and absolute signal intensities in the fingerprints have no direct relationship to the true abundance of individual compounds in the sample. This is due to different ionization efficiencies of venom components and their mutual ionization suppression. To obtain relevant quantitative information, different proteomics methods should be used, mostly based on LC-MS/MS analysis.

Acknowledgments This work was carried out with support of CEITEC 2020, project LQ1601, with financial support from the Ministry of Education, Youth and Sports of the Czech Republic (MEYS CR). CIISB research infrastructure project LM2015043 funded by MEYS CR is gratefully acknowledged for financial support for the MALDIMS measurements at the Proteomics Core Facility. References 1. Aebersold R, Mann M (2016) Massspectrometric exploration of proteome structure and function. Nature 537:347–355 2. Robinson SD, Undheim EAB, Ueberheide B et al (2017) Venom peptides as therapeutics: advances, challenges and the future of venom-

peptide discovery. Expert Rev Proteomics 14:931–939 3. Kim MS, Pinto SM, Getnet D et al (2014) A draft map of the human proteome. Nature 509:575–581

MALDI-TOF MS Profiling of Spider Venoms 4. Freiwald A, Sauer S (2009) Phylogenetic classification and identification of bacteria by mass spectrometry. Nat Protoc 4:732–742 5. Zhou Y, Shen N, Hou HY et al (2017) Identification accuracy of matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) for clinical pathogenic bacteria and fungi diagnosis: a metaanalysis. Int J Clin Exp Med 10:4057–4076 6. Sˇedo O, Ma´rova´ I, Zdra´hal Z (2012) Beer fingerprinting by Matrix-Assisted Laser Desorption-Ionisation-Time of Flight Mass Spectrometry. Food Chem 135:473–478 7. Sˇedo O, Korˇa´n M, Jakesˇova´ M et al (2016) Rapid assignment of malting barley varieties by matrix-assisted laser desorption-ionisation Time-of-flight mass spectrometry. Food Chem 206:124–130 8. Escoubas P, Ce´le´rier ML, Nakajima T (1997) High-performance liquid chromatography matrix-assisted laser desorption/ionization time-of-flight mass spectrometry peptide fingerprinting of tarantula venoms in the genus Brachypelma: chemotaxonomic and biochemical applications. Rapid Commun Mass Spectrom 11:1891–1899 9. Peka´r S, Sˇmerda J, Hrusˇkova´ M et al (2012) Prey-race drives differentiation of biotypes in ant-eating spiders. J Anim Ecol 81:838–848

181

10. Li J, Li D, Zhang F et al (2014) A comparative study of the molecular composition and electrophysiological activity of the venoms from two fishing spiders Dolomedes mizhoanus and Dolomedes sulfurous. Toxicon 83:35–42 11. Ca´rdenas M, Sˇedo O, Peka´r S (2014) Is there ontogenetic shift in the capture traits of a preyspecialized ant-eating spider? J Zool 293:234–242 12. Peka´r S, Sˇedo O, Lı´znarova´ E et al (2014) David and the Goliath: potent venom of an ant-eating spider (Araneae) enables capture of a giant prey. Naturwissenschaften 101:533–540 13. Peka´r S, Petra´kova´ L, Sˇedo O et al (2018) Trophic niche, capture efficiency, and venom profiles of six sympatric ant-eating spider species (Araneae: Zodariidae). Mol Ecol 27:1053–1064 14. Bocˇa´nek O, Sˇedo O, Peka´r S et al (2017) Evaluation of sample preparation protocols for spider venom profiling by MALDI-TOF MS. Toxicon 133:18–25 15. Bascur L, Yevenes I, Adrian H (1980) An electric method to obtain Loxosceles spider venom. Toxicon 18:224

Part IV Characterization of Toxins Function

Chapter 10 Methods for Evaluation of a Snake Venom-Derived Disintegrin in Animal Models of Human Cancer Stephen D. Swenson, Catalina Silva-Hirschberg, and Francis S. Markland Abstract Integrin targeting has been shown to be an effective approach for anticancer therapy. We engineered a recombinant disintegrin, vicrostatin (VCN), that binds with high affinity and specificity to the Arg-Gly-Asp (RGD) class of integrins, including αvβ3, αvβ5, and α5β1, involved in tumor invasion and metastasis. We used three different delivery modalities to examine anticancer activity of VCN in mouse models of human ovarian cancer, glioma, and prostate cancer. A female mouse model was used to examine the treatment of established ovarian cancer (OC) using VCN delivered intraperitoneally (IP) weekly either in saline or impregnated in a viscoelastic gel. SKOV3luc cells (a human OC cell line) were directly injected IP into immunodeficient mice. We also examined the antitumor activity of radioiodinated VCN delivered intravenously in a human glioma model in nude mice. We evaluated the effectiveness of 131I-VCN in combination with the DNA alkylating agent temozolomide in limiting glioma growth. Finally, treatment of a bone metastatic model of human prostate cancer (PC) in immunodeficient mice was examined using a liposomal formulation of VCN (LVCN) delivered intravenously. Human PC cells were suspended in a solution of Matrigel and injected into the left tibia of immunodeficient mice. Diameters of both the left and right (control) tibias were measured by caliper repeatedly after VCN treatment was initiated. Key words Ovarian cancer, Glioma, Prostate cancer, Integrins, Disintegrins, Animal models

1

Introduction One of the components in the venom of the southern copperhead snake (Agkistrodon contortrix contortrix) is a disintegrin we named contortrostatin (CN). CN is presumably present in the venom to prevent blood clot formation, thus allowing toxic components of the venom to spread rapidly throughout the body of the envenomated species. The mechanism employed by CN to prevent blood clot formation is mediated through blocking an integrin receptor (αIIbβ3) on the platelet surface [1] that inhibits platelet aggregation. Integrins similar to those found on platelets are also present on the surface of many types of cancer cells and facilitate tumor growth and metastasis. We have been interested in the use of snake venom disintegrins as cancer therapeutics, based on the important

Avi Priel (ed.), Snake and Spider Toxins: Methods and Protocols, Methods in Molecular Biology, vol. 2068, https://doi.org/10.1007/978-1-4939-9845-6_10, © Springer Science+Business Media, LLC, part of Springer Nature 2020

185

186

Stephen D. Swenson et al.

role integrins play in cancer invasion and dissemination, as well as in angiogenesis. Disintegrins are a class of disulfide-rich peptides, originally isolated from snake venom, many of which contain an RGD motif [2–4]. These peptides hold a significant translational potential with desirable pharmacological attributes based on their high-affinity/ high-specificity interaction with tumor integrins. The integrinbinding activity of disintegrins depends on the appropriate pairing of several cysteine residues responsible for the disintegrin fold, a mobile 11-amino acid loop protruding from the polypeptide core displaying the tripeptide motif RGD in many disintegrins [5, 6]. Disintegrins bind only to the activated conformations of integrins on motile cells such as cancer cells and angiogenic endothelial cells [4, 7], making them attractive vehicles to specifically target cancerous tissues. We and others showed that disintegrins are well tolerated and can be infused without toxicity or detrimental effect on blood pressure, body temperature, or other physiological parameters [8–10]. Subsequently, the Markland laboratory designed a sequence-engineered recombinant RGD disintegrin, vicrostatin (VCN), that can be reliably produced in large quantity (~200 mg/L purified VCN from bacterial cell lysate) in a robust recombinant bacterial system [11, 12]. VCN, a single polypeptide chain of 69 amino acids, retains the binding properties of the natural disintegrin from which it originated, the homodimer contortrostatin (CN), while showing 13-fold improved binding affinity, compared to CN, for an important integrin in angiogenesis, integrin α5β1. VCN inhibits ADP-induced platelet aggregation in a dose-dependent manner with an IC50 identical to that of native CN (~60 nM). Moreover, our recombinant disintegrin, VCN, inhibits tumor cell adhesion, endothelial and tumor cell invasion, and endothelial cell tube formation in a manner indistinguishable from CN. Interaction of VCN with integrins on cancer cells leads to disruption of integrin-mediated signaling through PI3K, Src, FAK, and MAPK [13], ultimately leading to decreased invasion, migration, and survival of cancer cells. Integrins are heterodimeric cell-surface adhesion receptors (containing α and β chains) operating at the interface between the extracellular matrix (ECM) and cytoskeletal apparatus [7, 14, 15]. A subgroup of integrins recognize the Arg-Gly-Asp (RGD) sequence present in key extracellular matrix proteins [16] involved in tumor development, angiogenesis, and progression. Integrins link cytoskeletal proteins to the extracellular matrix and are involved in bidirectional signaling to alter cellular functions. Among these interactions are the adhesion of both endothelial and cancer cells to extracellular matrix proteins and additional interactions that are integral to tumor growth, metastasis, and angiogenesis. Integrins allow cells to mechanically sense their environment [17] by integrating multiple signaling pathways initiated

Evaluation of a Snake Venom Peptide in Animal Cancer Models

187

by extracellular cues with the cell’s cytoskeleton. Integrins exhibit structural diversity and undergo conformational changes that are central to the regulation of their function. They exist in three major conformational states: an inactive or low-affinity state, a primed or activated high-affinity state, and a ligand-bound or occupied state [7]. Activated integrins are not usually present on quiescent tissues, but some, such as αvβ3, play an important role in neoplastic processes including cancer progression [18, 19], invasion (via invadosome formation) [20], angiogenesis [21], and metastasis. The role of αvβ3 in enhancing invasion by cancer cells [22, 23] underscores its potential for therapeutic strategies aimed at inhibiting pathways involved in cancer dissemination [24]. The complexity of the cancer microenvironment dictates that for optimal efficacy an anti-integrin therapeutic must target at least two members of the RGD-binding integrin class and preferably more [25] given the ability of cancer cells to change their integrin repertoire in response to drug treatment. VCN, which targets multiple integrin members including (but not limited to) αvβ3, αvβ5, α5β1, and αIIbβ3, admirably meets this criterion, accounting for its potent anticancer activity in preclinical studies in multiple human cancers including prostate [26], breast [11, 12], ovarian [27], and glioma [28]. Importantly, toxicity studies in rodents using a dose of VCN much higher than that used in our cancer therapeutic studies demonstrated a complete lack of toxicity. Relevant pharmacotoxicological properties of new chemical entities need to be extensively studied in laboratory animals before human administration. Although the validity of animal testing to predict efficacy and safety in humans has been questioned, it is generally believed that pharmacokinetic (PK) data can be extrapolated to humans reasonably well using the appropriate PK principles [29]. Animal cancer models are imperfect representations of the complex, diverse, and multifaceted spectrum of diseases that encompass human cancer. Cancer by its very nature exhibits considerable intra- and inter-tumor heterogeneity both genotypically and phenotypically that are dynamic and variable in nature. This must be considered in the interpretation and extrapolation of experimental data generated in preclinical models of human cancer and their potential relevance in evaluating new therapeutic agents for the cancer patient [30]. Animal models of cancer include a wide spectrum: (1) ectopic xenografts {subcutaneous (SC), intraperitoneal (IP), intravenous (IV), intramuscular (IM)} of tumor-derived cell lines or tissue explants implanted into syngeneic or immune-compromised rodent hosts; (2) orthotopic models in which explants of tumor tissues or established tumor lines are implanted within the organ or tissue of origin, thus recapitulating the intricacies and cell-cell interactions of the local microenvironment in which a primary tumor grows and from which it invades and disseminates;

188

Stephen D. Swenson et al.

(3) germ-line transgenic and conditional transgenic models in which the expression patterns of specific oncogenes or tumorsuppressor genes can be regulated systemically or in a tissue- and temporal-specific manner; (4) primary human tumor grafts maintaining the genotypic and phenotypic profile of a primary tumor from which they are derived (patient-derived xenografts, PDXs); and (5) various carcinogen-induced models that recapitulate the time-dependent and multistage progression of tumor pathogenesis in response to environmental carcinogens and tumor-promoting agents [30]. Orthotopic cancer models in syngeneic and immunecompromised rodents involve the implantation of tumor cells or primary tumor tissue explants into the originating tissue site of the cancer in rodents, resulting in much higher metastatic rates and a pathological phenotype that more closely recapitulates the human clinical course of metastatic disease [31]. Stable, reliable, and reproducible orthotopic animal models are critical, as they provide an opportunity for studying the mechanisms of pathogenesis as well as malignant progression, i.e., localized invasion and distal metastatic spread of the primary tumor, processes important to elucidate in the discovery and development of novel therapeutic agents. Orthotopic implantation combined with subsequent harvesting at metastatic sites can generate cancer cell variants that are clinically relevant to the metastasis process [32]. Thus, in contrast to ectopic subcutaneous tumor xenograft models, orthotopic models can more accurately reconstitute an organ microenvironment that dictates the tumor cell phenotype. The nonmalignant cellular components residing in the stromal microenvironment of tumors including inflammatory cells, macrophages, fibroblasts, and endothelial cells as well as soluble cytokines, chemokines, and cell matrix components and adhesion proteins not only influence the natural history of tumor proliferation, angiogenesis, invasion, and metastasis, but are also capable of altering the response of tumors to distinct antitumor and antiangiogenic therapeutic agents. Consequently, orthotopic models enable the assessment of the effects of specific targeted or cytotoxic agents on primary tumor growth in the appropriate microenvironment, as well as their impact on localized tumor invasion, metastatic spread, and emergence of acquired therapeutic resistance. These models can also be utilized effectively to evaluate survival endpoints preclinically in relation to continuous versus intermittent treatment with novel therapeutic agents or dosing combinations, for example, using Kaplan–Meier analyses of tumor growth inhibition versus tumor regression in the context of survival of tumor-bearing animals [30].

Evaluation of a Snake Venom Peptide in Animal Cancer Models

2

189

Materials

2.1 Ovarian Cancer: Intraperitoneal Model

1. Human ovarian cancer cell lines OVCAR 3 and SKOV3 (ATCC, Manassas, VA) are grown as monolayers in vitro. They form both subcutaneous and intraperitoneal tumors in immunocompromised hosts. 2. Lentivirus-containing genes coding for luciferase (luc) and green fluorescent protein (GFP) was obtained from Virus Production Core laboratory at Children’s Hospital Los Angeles. 3. Immunocompromised hosts: female athymic nude mice (Balb/ c/nu/nu) 20 g each (Simonsen Laboratories, Gilroy, CA). 4. FACS instrumentation. 5. 1 μg/g Mouse body weight of D-luciferin (Perkin Elmer, Waltham, MA) in PBS (Corning, Manassas, VA) 90 s prior to imaging. 6. PolyHEMA {poly(2-hydroxyethyl methacrylate)} (SigmaAldrich, St. Louis, MO). 7. Imaging with charge-coupled device (CCD) camera exposure times 1–60 s, binning 2–8, FOV 14.6 cm, f/stop 1, no filter (IVIS 200, Xenogen-Caliper, Alameda, CA). 8. Anesthetic: Isoflurane (2% vaporized in O2) (VetOne, Fluriso Boise, ID). 9. Living Image software (Perkin Elmer, Waltham, MA). 10. VCN (laboratory prepared). 11. Oxiplex, a hydrogel containing carboxymethyl cellulose and polyethylene glycol with calcium (FzioMed, San Luis Obispo, CA), is used to deliver VCN to the peritoneal cavity.

2.2 Glioma: Intracranial Model

1. Radioinert iodine and radioactive (131I) iodine (Perkin Elmer, Waltham, MA). 2. Mice (Simonsen Laboratories, Gilroy, CA). 3. Gel electrophoresis and mass spectrometry (BioRad, Hercules, CA, and Agilent, Santa Clara, CA, respectively). 4. U251 and U87 glioblastoma cell lines (ATCC, Manassas, VA) are infected with genes for luciferase and GFP. 5. Luciferin (PerkinElmer, Waltham, MA). 6. Luminescent imaging: VEVO200 imaging instrument (IVIS 200, Xenogen-Caliper, Alameda, CA). 7. Temozolomide (Merck, Kenilworth, NJ).

190

Stephen D. Swenson et al.

2.3 Prostate Cancer: Subcutaneous and Bone Metastasis Model

1. PC-3 cells (ATCC, Manassas, VA). 2. Liposomal VCN is prepared by Molecular Express, Inc. (Rancho Dominguez, CA). 3. PBS (Corning Life Sciences, Tewksbury, MA) and empty liposomes (Molecular Express Inc., Rancho Dominguez, CA). 4. Caliper (VWR, Radnor, PA). 5. Polyclonal rat anti-mouse CD31/PECAM immunohistochemistry (Santa Cruz Biotechnology, Dallas, TX). 6. CWR22rV1 prostate carcinoma cells (ATCC, Manassas, VA). 7. Solution of Matrigel (Corning Life Sciences, Tewksbury, MA). 8. Immunodeficient mice, Balb c nu/nu (Simonsen Laboratories, Gilroy CA). 9. Testosterone and DMSO (Sigma-Aldrich, St. Louis, MO).

3

Methods

3.1 Ovarian Cancer: Intraperitoneal Delivery of a Gel Formulation of VCN

In our ovarian cancer studies (using an orthotopic model) our goal is to significantly delay or prevent the dissemination of advanced ovarian cancer (OC) by delivery of a novel noncytotoxic, antiinvasive peptide directly to the peritoneal cavity. OC is a devastating disease with close to 14,000 deaths expected to occur in the USA alone in 2019 [33]. The standard treatment is an aggressive surgery followed by adjuvant platinum-taxane chemotherapy. However, no efficient therapeutic method aimed at stopping OC dissemination is currently available in the clinic. Our goal is to develop a more efficient and less toxic form of therapy aimed at addressing metastatic residual disease in order to prolong the survival and improve the quality of life of OC patients. The mode of OC dissemination is unique among solid tumors. Exfoliated OC cells from primary tumors as well as from macroand micrometastases assemble into free-floating multicellular aggregates (called spheroids) and are carried away by the peritoneal fluid to secondary sites in the abdominal cavity where they attach, invade into submesothelial connective tissue, and establish peritoneal micrometastases [34]. These spheroids, unlike individual OC cells, are known to be significantly more resistant to chemotherapy. Unlike primary tumors and macrometastases, OC micrometastases evade detection at the time of surgery, and failure to eradicate them is the main reason for recurrence [35]. One effective way of addressing this clinical problem is by the means of a broad-spectrum antiinvasive therapeutic solution. Toward this end, the effect of VCN delivered intraperitoneally (IP) is being studied. What complicates the management of OC is that chemotherapy administered intravenously (IV) penetrates only poorly in the peritoneal compartment

Evaluation of a Snake Venom Peptide in Animal Cancer Models

191

where the disease is spreading. Consequently, significantly higher concentrations of chemotherapy agents than could normally be achieved by IV administration would be required for disease control. One solution to the poor penetration problem following IV administration is the use of IP administration of chemotherapy after debulking surgery, and in advanced OC a 21.6% decrease in the risk of death was reported in patients undergoing combined IP and IV therapy versus those undergoing IV therapy alone [36]. In 2006 the NCI issued a statement: “On the basis of the results of these randomized phase III clinical trials, a combination of IV and IP administration of chemotherapy conveys a significant survival benefit among women with optimally debulked ovarian cancer, compared to IV administration alone” [37]. Although efficacious in the long run, the regimens of currently used IP platinum and taxane agents are accompanied by dramatic off-target toxic effects [38–40]. This is the main reason why a significant number of women opt out of IP treatment despite its potential benefits [41, 42]. Wright et al. [43] stressed that increasing IP/IV chemotherapy use in clinical practice is an important and underused strategy to improve ovarian cancer outcome. They observed that the use of IP/IV chemotherapy in clinical practice is not only feasible but results in improved survival when compared to IV chemotherapy alone [43]. Thus, there is increasing support for the use of IP chemotherapy, and with the application of a noncytotoxic anti-invasive agent, such as VCN, this would address the issue of toxicity of present agents and the high level of patient noncompliance. 1. To ensure confirmation of tumor growth the OC cell lines have been stably transfected by a lentivirus-containing gene coding for luciferase (luc) and green fluorescent protein (GFP), which are positioned upstream and downstream of an internal ribosome entry site (IRES) sequence, respectively. Prior to implantation and to ensure homogeneity of the labeled cells they are grown and FACS sorted three times. 2. Spheroids are multicellular aggregates that can be easily grown in vitro from any human OC cell line by simply seeding a defined number of individual cells in appropriate media and forcing the cells to grow in suspension by plating them on extremely low-binding surfaces. To generate spheroids from SKOV3GFP/LUC cells, we coat multi-well plates with polyHEMA {poly(2-hydroxyethyl methacrylate)}, onto which we seed a defined number of cells in their regular medium and allow them to form spheroids over a defined time interval. 3. For the dispersed cell model both OVCAR3 and SKOV3 cells growing as a monolayer in culture are harvested and injected IP

192

Stephen D. Swenson et al.

into 5-week-old female athymic nude mice weighing approximately 20 g. Each animal receives 2
106 cells, either dispersed cells or cell equivalents in the spheroid model, and the tumors are allowed to implant for 4–14 days before treatment is initiated. 4. To evaluate the presence of tumor prior to initiating treatment, and biweekly throughout the study, mice are imaged in the Molecular Imaging Center at the Keck School of Medicine, USC. To carry out the imaging mice are injected IV with 1 μg/ g mouse body weight of D-luciferin in PBS 90 s prior to imaging. Imaging is performed using a charge-coupled device (CCD) camera. Mice are anesthetized with isoflurane prior to and during imaging. Total photon flux (photons/sec) is measured from a fixed region of interest (ROI) over the peritoneal space of the mice using Living Image software as previously described [44]. 5. A preliminary dose range-finding in vivo evaluation of VCN in a murine model of human ovarian cancer is performed, using the SKOV3 line, which is an intensely studied OC line [45–47]. These studies are aimed at evaluating the effects on tumor dissemination of the administration of 1, 2.5, and 5 mg VCN delivered once per week embedded in 1 mL Oxiplex versus Oxiplex alone (vehicle control) (see Note 1). 6. All animals are sacrificed when the control animals reach the endpoint set for the study (i.e., excessive tumor growth with ascites buildup, visible signs of emaciation, and a loss of 10% or more of body weight). 7. Animals are necropsied, and gross examination reveals a dosedependent effect of VCN administration on tumor growth; this is determined subjectively by visual scoring of the number of macroscopic tumor foci present. Upon careful dissection, control animals show extensive and widespread macroscopic carcinomatosis throughout the peritoneal cavity, while the treated groups show a significant decrease in the total number of macroscopic foci that correlates with the administration of increasing amounts of drug (e.g., the animals that received 5 mg/mL of VCN weekly display virtually no visible tumor). 8. To demonstrate the reliability of the bioluminescent signal from the luciferase-expressing tumors, the SKOV3GFP/LUC model is utilized, an inoculum of 2
106 cells is injected IP, and the tumors are allowed to grow for 2 weeks prior to the initiation of therapy. Therapy in this study consisted of 5 mg VCN delivered in 1 mL Oxiplex/week versus vehicle control with 1 mL Oxiplex alone/week for a 4-week course (see Note 2).

Evaluation of a Snake Venom Peptide in Animal Cancer Models

193

9. Even though direct IP injection of suspended OC cells produces an appropriate animal model, the spheroid model of OC growth is increasingly regarded as a more physiologically relevant model [48]. Therefore, once an optimally efficacious dose of the VCN-Oxiplex formulation was identified, this dose was tested in a spheroid model of SKOV3GFP/LUC. Previous in vivo models employed the injection of a large number of dissociated cells harvested from OC cell monolayers grown in culture. In this study, OC spheroids are generated as described above and implanted IP. As a control for spheroid growth a group injected with inocula consisting of 2
106 standard dispersed cells is included. Following inoculation, the animals are imaged weekly to follow tumor growth and spread. The spheroids are allowed to grow until the animals progress to sacrifice criteria, reached in about 30 days from the time of inoculation (see Note 3). 3.2 Glioma: Intravenous Delivery of 131I-VCN, a MultiIntegrin Targeting Agent for Glioblastoma

Glioblastoma multiforme (GBM) is the most malignant type of glioma and a devastating brain cancer affecting both male and female with no age barrier [49, 50]. Despite advances in surgery, chemotherapy, and radiation therapy, the average life span of patients with GBM is ~15 months from the time of diagnosis [51]. It has been shown that the invasiveness of GBM is linked to integrins; tumor cells display a subset of activated integrins on their surface, whereas surrounding normal brain cells do not display these integrins [28, 52–54]. GBM also has numerous angiogenic blood vessels [55–59], which facilitate dissemination of GBM into normal brain tissue, making treatment of the tumor a challenge. Integrins are also involved in GBM angiogenesis and are associated with GBM stem cell function. Disrupting integrin activity reduces tumor formation and increases animal survival in animal model studies [54]. Thus, therapy against GBM based on integrin involvement is particularly attractive, producing antiangiogenic, antiinvasive, and possibly anti-stem cell effects [54]. Integrin targeting should be applicable to all GBM subtypes irrespective of oncogene and tumor-suppressor status. The design of VCN includes a carboxyl-terminal sequence alteration that produces much higher affinity for integrin α5β1, which is known to be involved in angiogenesis and GBM invasion [11, 60]. Recent in vivo studies in our lab indicate that VCN specifically induces apoptosis of angiogenic endothelial and cancer cells, supporting its role as an antiangiogenic and anticancer agent [11, 12]. Importantly, via integrin ligation, VCN appears to elicit a cascade of signaling events leading to massive disruption of the actin cytoskeleton in human umbilical vein endothelial cells (HUVEC) and cancer cells. Further, the molecular mechanism of CN (and VCN) involves the dissociation of talin from the cytoplasmic domain of β1 integrin, which leads to a dramatic impairment of

194

Stephen D. Swenson et al.

cell invasiveness. The binding of talin to integrin β1 tails represents the final common step in integrin activation [61]. This presumed activity of VCN accounts for the impressive inhibition of invasive ability of tumor cells and angiogenic vasculature in a human breast cancer model [11, 12]. We had previously shown that it is possible to deliver a concentrated dose of radiation to the tumor site with the use of 131ICN [28, 62]. However, we demonstrated recently that 131I-VCN is more useful as a therapeutic agent for brachytherapy and more practical due to its availability from the high-level bacterial recombinant expression system. The brachytherapy strategy is encouraged by results from others who are conducting clinical trials using a humanized murine monoclonal antibody (mAb) targeting tenascin, an extracellular protein expressed in large amounts by the stroma of GBM but not by normal brain. Intralesional radioimmunotherapy is aimed at controlling the progression of GBM. 131Iantitenascin mAb was injected directly into the tumor mass to maximize the mAb concentration in the tumor and to irradiate the GBM cells [63]. A clinical study evaluated efficacy and toxicity in GBM [64–66]. Median overall survival time was greater than historic controls in these studies and there was minimal toxicity. We will take advantage of the dual activity of 131I-VCN. It will be used as a therapeutic agent based on its ability to bind specifically to integrins expressed by GBM tumor and endothelial cells and disrupt the invasive potential of GBM. Furthermore, it will also act as a brachytherapy agent following radioiodination of the single tyrosine residue in VCN (without inhibition of integrin-binding activity) [28, 62]. Brachytherapy is specifically localized to the tumor since 131I-VCN will bind to GBM and angiogenic endothelial cells, but does not bind to normal brain. 1. For VCN iodination we employ both radioinert and radioactive iodine using the chloramine-T procedure [67] as we described earlier [62], and then confirm retention of biological activity of the iodinated VCN. Additionally, the labeled material is analyzed both by gel electrophoresis and mass spectrometry; we identified an ~70:30 mixture of mono-iodinated to di-iodinated VCN in the reaction, which yields VCN with full bioactivity. Allowing the reaction to proceed to form more di-iodinated material results in reduced biological activity and attempting to produce only mono-iodinated material results in a high level of un-iodinated species which defeats the ability to deliver radioactivity to the tumor. 2. We evaluated the ability of radioiodinated VCN to be delivered IV and be retained in the brain and/or blood vessels supplying the brain. In these studies, two groups are used: non-treated control animals, and tumor-bearing animals in which the tumor was allowed to grow for 14 days post-implantation. In both the tumor-bearing and control animals a single dose of

Evaluation of a Snake Venom Peptide in Animal Cancer Models

195

I-VCN (100 μCi) is administered IV. 131I-VCN is allowed to circulate for 18 h at which time the animals are sacrificed, and brains removed and counted (see Note 4). 131

3. We then determined the efficacy of 131I-VCN in limiting tumor progression. In two distinct models of glioma we evaluated the efficacy of 131I-VCN in extending progression-free survival of study animals. Tumors are implanted 3 mm deep in the midline of mice brains using stereotactic injections. Following surgical exposure of the skull and drilling a small-bore hole with a dental drill, either U251 or U87 (2
105/5 μL) cells are injected using a slow controlled injection followed by a 1-min rest period with syringe in place to minimize leakage from the injection track. The syringe is slowly removed, and the surgical incision is then closed with two stitches (3.0 silk). Seven days following implantation tumor take is evaluated by luminescent imaging. All cells have been stably infected with genes for luciferase and GFP. 4. Cells can be imaged by IV injection of luciferin followed by a 90-s wait period at which time the animal is placed in the VEVO200 imaging instrument and image acquired. Positive image indicates tumor take. In our experience >90% of mice will be positive after 7 days. Positive, U87-tumor-bearing mice are then randomized into treatment and control groups. The following groups are employed: (1) untreated control, (2) treatment with 131I alone (nuclide alone), and (3) 131IVCN. The dose of radioactivity administered is 200 μCi either 131 I alone or 1 mg VCN labeled with 200 μCi 131I. The mice are treated with a single dose and monitored daily for changes in weight and physical status (see Note 5). 5. We also determined the efficacy of 131I-VCN in combination with a clinically utilized therapeutic in limiting tumor progression. We examined the efficacy of 131I-VCN when delivered with the DNA alkylating agent, temozolomide (TMZ). TMZ functions by alkylating/methylating DNA, which most often occurs at the N-7 or O-6 positions of guanine residues. The result of this DNA modification is damage to the DNA which triggers death of the tumor cells. In this study we implant U251 tumors into the brain as described above and confirm tumor growth 8 days postimplantation. At this time the tumorpositive mice are randomized into five groups of ten mice each. The groups are (1) untreated control, (2) 131I alone, (3) 131I plus TMZ, (4) 131I-VCN, and (5) 131I-VCN plus TMZ. The radioactivity is delivered as 200 μCi either iodine alone or 1 mg VCN labeled with 200 μCi 131I given in two doses separated by 10 days (each with 200 μCi 131I-VCN). The groups receiving TMZ have the agent delivered as a daily 25 mg/kg gavage in two 10-day cycles spaced by a 7-day off rest period (see Note 6).

196

Stephen D. Swenson et al.

3.3 Prostate Cancer: An Intravenous Liposomal Formulation of VCN for Cancer Therapy

Prostate cancer (PC) is the most prevalent and the second most lethal cancer among American men [68]. Duration of benefit from androgen deprivation therapy, the mainstay of treatment of metastatic disease [69], is highly variable and difficult to predict [70]. Metastatic PC invariably progresses to a castrate-resistant state in spite of androgen-signaling networks remaining vitally active [71]. Several potent agents targeting this network are associated with improved survival including androgen receptor antagonists such as enzalutamide [72], and agents targeting inhibition of androgen synthesis via the 17α-hydroxy/17,20-lyase (CYP17) pathway, such as abiraterone [73, 74]. The chemotherapeutic agents, docetaxel and cabazitaxel, also have survival benefits. Several studies have indicated that different members of the integrin family are expressed in various PC cell lines, as well as human PC and benign prostatic hypertrophy [13, 19]. Similar to normal cells, these integrins can be presumed to be of importance in the motility of PC cells as well as the interaction of these cells with the local milieu. Of interest is the finding that different PC cell lines express different profiles of integrins, correlating with the ability of these PC cells to progress and metastasize. 1. All liposomal formulations of VCN (LVCN) were prepared by a proprietary method by Molecular Express, Inc. 2. To evaluate the in vivo activity of LVCN in animal models, we examined the effect of LVCN treatment on a xenograft model of human prostate cancer. PC-3 (human PC) cells (2
106) are implanted subcutaneously in the hip flank of 5-week-old male nude mice. Drug administration is initiated when tumors become palpable (14 days), at which time LVCN and unencapsulated VCN are administered twice weekly via i.v. injection (100 μg VCN equivalent per dose for 5 weeks); PBS and empty liposome inoculations serve as controls. Tumors are measured weekly via caliper in a blinded fashion. Tumor volume is determined from the formula V ¼ (W2
L)/2, where V is volume, W is width, and L is length (see Note 7). 3. We also determined the anti-angiogenic effect of LVCN in the PC-3 model using polyclonal rat anti-mouse CD31/PECAM (platelet endothelial cell adhesion molecule) immunohistochemistry (IHC). Quantification of blood vessel density is performed in a blinded fashion as previously described [11] (see Note 8). 4. In addition, we evaluated the therapeutic efficacy of LVCN in treating a PC bone metastasis model. CWR22rV1 prostate carcinoma cells (5
104 cells) are suspended in a solution of Matrigel and directly injected into the proximal left tibia of immunodeficient mice. Following injection, the diameter of both the left and right (control) tibias is measured by caliper

Evaluation of a Snake Venom Peptide in Animal Cancer Models

197

twice weekly. As the CWR22rV1 cell line is androgen dependent, animals receive daily intraperitoneal injections of testosterone suspended in DMSO (1 mg/kg). Tumors grow slowly over the initial 20 days, but at this point tumors are evident, and treatment begins. LVCN and unencapsulated VCN are administered twice weekly via i.v. injection (100 μg per dose) for 5 weeks. LVCN enters and is retained in the tumor due to the enhanced permeability and retention (EPR) effect, and to the “leaky” tumor vasculature (see Note 9). 3.4

4

Conclusions

Animal models are essential for the development of new drugs, especially those in the oncology field. Having models that are simple, economic, and representative of the disease under study is fundamental. In this work we present different models used for the development of therapies for ovarian cancer, glioma, and prostate cancer. Snake venom components have been of great scientific interest for many years, and the Markland lab is mainly interested in a class of peptides, called disintegrins, found in the venom of many different snake species. Currently, two disintegrin-based antiintegrin therapeutics are approved and in clinical use in the USA in the cardiovascular arena. Eptifibatide and tirofiban are based on snake venom disintegrins (barbourin and echistatin, respectively), and bind to platelet integrin GPIIb/IIIa (αIIbβ3). These agents act to reduce the risk of ischemic events post-balloon angioplasty [75] and demonstrate the effectiveness of anti-integrin therapy in the clinical setting with minimal off-target side effects. In our investigations, we have developed VCN, a recombinant disintegrin, which inhibits tumor cell adhesion, and endothelial and tumor cell invasion, and demonstrates significant potential for a promising future in the cancer therapy arena.

Notes 1. At the completion of the dose-dependent study of the effect of VCN administration on ovarian cancer growth, we concluded that a dose of 5 mg of VCN administered once per week is more efficacious than the same dose divided in multiple smaller doses and administered more frequently. 2. In the study of the bioluminescent signal reliability (Fig. 1), a comparison of representative 2-week- and 4-week-treated animals demonstrates dramatic differences between the groups. Treated animals are essentially devoid of visible macroscopic tumor foci by gross examination (not shown). By contrast, Oxiplex-alone (vehicle control) animals show extensive tumor dissemination throughout the peritoneal cavity. Bioluminescent imaging (BLI) quantitation shows greater than 95%

198

Stephen D. Swenson et al.

Fig. 1 Tumor-bearing mice (SKOV3GFP/LUC) treated with VCN-Oxiplex or Oxiplex alone. Ovarian cancer cells (2
106 SKOV3GFP/LUC cells) were injected IP in female nude mice and allowed to grow for 2 weeks. At this time treatment was initiated with the vehicle Oxiplex alone (1 mL administered once weekly, left two panels) or VCN-Oxiplex (5 mg/mL VCN, 1 mL administered once weekly, right two panels). Weekly bioluminescent imaging of the luciferase-containing tumor cells was used to follow tumor growth and dissemination, and representative animals after 2 and 4 weeks of treatment are shown. Euthanasia was carried out after 4 weeks of treatment due to massive tumor burden in the vehicle-treated animals. Upon visual examination, extensive tumor growth was observed in the control animals, but no visible macroscopic tumor foci were observed in treated mice. The colored bar to the right of the figure shows the color scale corresponding to the intensity of luminescence from low (blue, at the bottom) to high (red, at the top)

inhibition of tumor spread in the treated group after 4 weeks of treatment. 3. In the spheroid model of OC, BLI revealed that unlike dissociated cells which form a smaller number of macroscopic foci, the inoculated spheroids reproducibly form significantly higher numbers of smaller size tumor foci starting as early as 4 days postinoculation. This is consistent with a much different tumor growth behavior compared to the use of dissociated cells as the initial inoculum. Nonetheless, bioluminescent imaging (BLI) quantitation in the spheroid model showed greater than 95% inhibition of tumor spread in the treated group after 4 weeks of treatment. 4. We found that a significant portion of the IV-delivered radioiodinated VCN is found in the brain and associated vasculature in tumor-bearing animals while non-tumor-bearing animals display minimal uptake to the brain. In the control group no counts above background are observed while in the tumor-

Evaluation of a Snake Venom Peptide in Animal Cancer Models

199

U87 Survival

Percent survival

150

U87 Control U87 131 Iodine U87 131I-VCN

100

50

0

0

10

20

30

40

50

Days

Fig. 2 Survival of U87 glioma-bearing animals following treatment with 131I-VCN. Treatment with 131I-VCN extends survival by greater than 13 days, a 60% extension of time to death. Animals that survive generally have a good appearance and stable weight until 24 h before succumbing to the disease

bearing animals an average of 1.4% of the injected counts are identified in the brain. 5. In another study U87-tumor-bearing mice in the control groups began dying on day 20 and all mice in the control and 131 I alone groups are dead by day 25. Treated mice in this group displayed no adverse effects of treatment and survived until day 34 with all mice dead by day 44. The median survival for the controls (untreated and 131I alone) is 22 and 21 days, respectively, while with the 131I-VCN-treated animals, the median survival is 35 days (Fig. 2). 6. In the combination therapy approach using the U251 glioma model, control-treated animals began dying immediately after the second round of TMZ and radioactivity, with an average median survival for these groups of 30 days. The 131I-VCNand 131I-VCN plus TMZ-treated animals remained vibrant and slowly began dying on day 42 with the final survivor in the 131IVCN group surviving 69 days. Three of the ten 131I-VCN plus TMZ mice died prior to day 57 with another two dying by day 78 and the remainder surviving beyond 90 days (Fig. 3). One of the long-surviving mice began displaying signs of neurological impairment and regrowth of the tumor was confirmed by optical imaging. This mouse had its tumor harvested and cells placed in culture as TMZ-resistant cells. 7. In the first prostate cancer study there is a 75% reduction in tumor growth of LVCN-treated animals compared to controls (Fig. 4). Unencapsulated VCN displayed no antitumor activity, most likely due to its low dose (100 μg) and the lack of an

Stephen D. Swenson et al.

U251 Survival

Percent survival

150

U251 Control 131I 131i + TMZ 131I-VCN 131I-VCN +TMZ

100

50

0

0

20

40

60

80

100

Days

Fig. 3 Survival of U251 glioma-bearing animals following treatment with 131IVCN  TMZ. Treatment of U251 tumor-bearing mice (ten animals per group) with the combination of 131I-VCN plus TMZ greatly enhances survival. The untreated control animals plus animals treated with 131I alone and 131I plus TMZ had median survival times of 30, 29, and 31 days, respectively. The 131IVCN and 131I-VCN plus TMZ groups had median survival of 50 and greater than 75 days. Four mice in the 131I-VCN plus TMZ group displayed no sign of tumor at the conclusion of the experiment. The combination of 131I-VCN plus TMZ appears to be a highly effective therapeutic strategy for gliomas such as U251 PBS

Empty Liposomes

VCN

LVCN

3000

Tumor Volume (mm^3)

200

2500 2000 1500 1000 500 0 Week 0

Week 1

Week 2

Week 3

Week 4

Week 5

Fig. 4 PC-3 xerografts treated with LVCN. Tumor volumes were determined in athymic nude mice (N ¼ 5) bearing subcutaneous xerografts (2
106 cells implanted) of the prostate cancer cell line PC-3 during treatment with LVCN or VCN administered twice weekly via i.v. injection (100 μg/dose). Control animals were infused with PBS or empty liposomes. Treatment lasted for 4 weeks. Vertical bars indicate 95% confidence intervals. There are statistically significant differences between the LVCN treatment group and PBS, empty liposomes, and VCN control groups ( p values adjusted for multiple comparisons 1
106 cells) as well as further time for analysis. The type of fluorescent dye also has its limitations. For example, calcein-AM, a compound that is hydrolyzed by intracellular esterase to release fluorescent calcein, is more suited for short experiments of a few hours. For increased duration of the signal, the carbocyanine (DiI) [38] or carboxyfluorescein dyes (CDFDA-SE and CFDA-SE) can be used [39]. These dyes are lipophilic compounds which can be incorporated into the cell membrane [40]; however, they may also affect cellular electron transport, therefore compromising cell viability [41]. Carboxyfluorescein dyes are stable for longer periods and act by covalently binding to intracellular amino groups, thus being required for use in amine-free buffers. These compounds are also sensitive to changes in pH [42]. To overcome these disadvantages, we have recently adapted for cell staining in the cell adhesion assay the indocyanine green (ICG), a water-soluble tricarbocyanine dye that absorbs and emits in the near-infrared (NIR) range with limited autofluorescence, enabling a high signal-to-noise ratio [43]. At the end of the cell adhesion experiment, the tissue culture plates are placed inside an Odyssey NIR laser scanner (Li-Cor; Lincoln, Nebraska, USA) and the stained cells scanned with fixed intensity at an excitation of 785 nm and emission of 800 nm. Thereafter, for data analysis, the intensity of the obtained images is quantified using the Li-Cor imaging program provided by the scanner. We employed this method to visualize the adherence of NCI-60 panel of colorectal carcinoma cell lines [44, 45]. This protocol provides a rapid method for measuring cell adhesion and does not interfere with matrix components. In addition, this method is rapid, can be performed with living cell cultures, and does not involve laborintensive cell labeling and standardization per run, thus reducing handling time (Figs. 4 and 5).

Integrin-mediated Cell Adhesion

213

Another important consideration is the cell type used in the adhesion assay. Primary cells are often preferred over long-term established cell lines because they are considered a better replicate of “real cells” by expressing physiological (low) levels of integrins. However, primary cells are variable and often can only be cultured for a few passages before they reach senescence or undergo undesirable phenotypic changes. Established cloned cells have the benefit of phenotypic stability and ease of culture and are used extensively for cell-based adhesion assays. Primary human vein endothelial cells (HUVEC) [46] and primary or cell lines of human skin fibroblasts [47] which express physiological levels of different integrins are commonly used for investigations of endothelial and fibroblast cell adhesion. Current integrin-mediated cellbased adhesion research is also employing the erythroleukemia cell line K562, transfected and overexpressing at high levels a single integrin subunit [26, 48, 49].

2 2.1

Materials Equipment

1. 96-Well polystyrene plates (Fisher Scientific, catalog number: 07–200-91; Corning Incorporated, catalog number: 3598) 2. Multichannel pipettes Fisherbrand™).

(e.g.,

Fisher

Scientific,

model:

3. Water-jacketed incubator 37  C, 5% CO2 (e.g., Thermo Fisher Scientific, Thermo Scientific™, model: Series 8000 WaterJacketed CO2 Incubators). 4. Cell culture laminar hood (e.g., NuAire CellGard™). 5. Tabletop centrifuge for 15 mL and 50 mL conical tubes (e.g., Eppendorf, model: 5804). 6. Light microscope (e.g., Olympus, model: CKX41 or ZEISS Axio Vert). 7. Spectrophotometer to measure absorbance at 570 nm or fluorescence at 492/517 nm with 96-well format (such as TECAN, Infinite® 200 PRO spectrofluorimeter), an Odyssey NIR laser scanner (Li-Cor; Lincoln, Nebraska, USA) that measures excitation at 785 nm and emission at 800 nm. 8. Software: Microtiter plate reader software, Li-Cor Image Studio™ Software Work Area, Microsoft Excel, GraphPad Prism 6. 2.2

Reagents

1. ECM molecule of choice, e.g., fibronectin (bovine plasma) (Sigma-Aldrich, catalog number: 341631 or F1141); collagens (from calf skin) (Sigma-Aldrich, catalog number: C8919, C7661), etc.

214

Philip Lazarovici et al.

2. The integrin antibodies anti-α1 (CD49a), anti-α2 (CD49b), anti-α4 (CD49d), anti-α5 (CD49e), anti-αv(CD51), and antiβ3 (CD61) were purchased from BD Biosciences (San Diego, CA, USA). The anti-β1 integrin (CD29) antibody was obtained from Chemicon International (Temecula, CA, USA). 3. Disintegrins were isolated from different snake venoms using high-pressure liquid chromatography or purchased commercially (Obtustatin-α1β1 non-RGD antagonist from Creative Peptides Co, USA, catalog number: R0985; Echistatin-α Vβ3–RGD antagonist from SmarTox Biotechnology, Saint Egre`ve, France). 4. 0.5 M EDTA solution (pH 8.0) (Life Technologies, Invitrogen™/Ambion; catalog number: AM9260G). 5. Trypsin-EDTA for cell culture (Sigma-Aldrich, catalog number: T4049) or HyClone™ HyQTase cell detachment reagent (GE Healthcare, HyClone™, catalog number: SV30030.01). 6. Bovine serum albumin (BSA) (Life Technologies, Invitrogen™, catalog number: 15561020). 7. Phosphate-buffered saline (PBS) (Life Technologies, Invitrogen™, catalog number: 14190–144): Prepare 1
PBS containing 2 mM CaCl2 and 2 mM MgCl2. 8. Hanks’ balanced salt solution (HBSS) (Sigma-Aldrich, catalog number: 55037C). 9. Crystal violet (Serva catalog number: 27335) or CellTracker™ Green CMFDA (5-chloromethylfluorescein diacetate, Thermo Fisher Scientific, catalog number: C2925) or indocyanine green (Toronto Research Chemicals, North York, ON, Canada, catalog number: I697000) or other cell-staining dye. 10. Acetic acid. 11. 4% Paraformaldehyde or 1% glutaraldehyde in PBS. 12. 0.2% Triton X in HBSS or 1% SDS in H2O for cell lysis. 2.3 Culture Media and Other Solutions

1. Dulbecco’s modified eagle medium (DMEM) (Life Technologies, Invitrogen™, catalog number: 10313–021) or RPMI 1640 medium (Gibco-Thermo Fisher Scientific, catalogue number: 11875093) or other appropriate growth medium for culturing cell line of interest. 2. Fetal bovine 30–2020™).

serum

(FBS)

(ATCC,

catalog

number:

3. Serum-free washing medium, such as DMEM (or RPMI) containing 0.1–0.5% BSA, 2 mM CaCl2, and 2 mM MgCl2. 4. Blocking medium, 0.5% or 1% BSA in DMEM or RPMI medium or HBSS with 5 mM MgCl2. 5. Trypsin-EDTA for cell culture (Sigma-Aldrich, catalog number: T4049).

Integrin-mediated Cell Adhesion

3 3.1

215

Methods Cell Growth

3.2 Coating Plates with ECM Protein

1. Cells were grown in tissue culture standard conditions using DMEM supplemented with 10% fetal bovine serum (FBS) and antibiotics (see Note 1). K562 cells transfected with α1, α2, or other integrin subunits were cultured in RPMI 1640 supplemented with 10% FBS, antibiotics, and 0.5 mg/mL of geneticin G418 (used in selection to screen for resistant cells which express the integrin subunit under the control of neo gene) [24]. Primary adult dermal human microvascular and other endothelial cells (dHMVEC) were purchased from Cambrex (Walkersville, MD) or Lonza Inc. (Allendale, NJ, USA) or Clonetics (San Diego, CA) and were cultured in complete endothelial cell basal media-2 (EBM-2, Cambrex) supplemented with EGM-2 SingleQuots (Clonetics) and used in experiments between passages 3 and 8. 1. Prepare a stock of 10 mg/mL collagen I or 1 mg/mL collagen IV in 0.02 M acetic acid and other ECM proteins in H2O at 4  C. Prepare 40 μg/mL collagen I solution in PBS, and store at 4  C; prepare 0.1% BSA solution in DMEM (see Note 2). 2. Add ECM solution (30–100 μL/well according to desired concentration) with multichannel pipette to wells of a 96-well tissue culture plate, cover, and place at 4  C overnight. Coat in triplicate or quadruplicate for each sample. Leave wells uncoated as negative control. Consider including the following reference control groups for monitoring each step of the procedure: wells uncoated with ECM protein; wells unwashed with DMEM; wells untreated with cells; and wells with no added dye (background). 3. Invert plate and shake out coating solution. Remove the remaining coating solution from each well with a tip pipette (see Note 3). 4. Dilute BSA to 1% (w/v) in H2O or PBS or HBSS containing 5 mM MgCl2. Add 100 μL/well with multipipette, cover, and place for blocking nonspecific binding at 4  C for 4 h (or at room temperature for 1 h). 5. Invert plate and shake out BSA-blocking solution. Remove the remaining blocking solution from each well with a tip pipette. Wash with washing buffer twice.

3.3

Cell Labeling

1. Thaw 10
trypsin/EDTA (then dilute to 1
in PBS), and warm PBS and serum-free medium (see Note 4). 2. Deprive cells at 75% confluence of serum for 4–6 h before the adhesion assay. To do so, wash cells three times with serum-free DMEM and cultivate them in DMEM.

216

Philip Lazarovici et al.

3. During the last 45 min of operation, remove medium from cells in T75 flask, and add serum-free washing medium. Subsequently, remove medium; wash with PBS; add 5 mM EDTA in HBSS without calcium, magnesium, and sodium bicarbonate or 1
trypsin/EDTA to detach the cells; and then observe under a microscope to confirm complete dissociation of the cells; time needed is ~1–20 min (depending on cell type). 4. Remove released cells, wash flask with 20 mL of serum-free medium, pellet cells in 40 mL of serum-free medium, prepare in 6 mL of serum-free medium, and count (15 μL of suspended cells plus 15 μL of trypan blue; add 15 μL to each side of hemocytometer; cell#/mL ¼ combined count from both sides x 104). Dilute cells to 4
105 (for crystal violet or ICG staining) or 1
106/mL (for CMFDA staining) in DMEM with 0.1% BSA. 3.4 Adhesion Assay Using Crystal Violet Staining

1. Count cell to 4
105/mL, add 50 μL cell suspension to each well, and perform the adhesion assay for 30 min in a CO2 incubator at 37  C (see Note 5). 2. Examine plate in inverted microscope. Photograph selected wells if desired. Invert plate gently onto an absorbent pad. Remove the remaining cell solution from each well with pipette (see Note 6). 3. With multipipette, slowly add 100 μL/well of serum-free medium down the side of each well (tilt plate; PBS is not recommended for this wash). Invert plate gently onto an absorbent diaper pad. Remove the remaining wash solution from each well with pipette (see Note 7). 4. Slowly add 100 μL/well of serum-free medium down the side of each well. Examine plate in inverted microscope. Cells in BSA negative control wells should be rare (if not, repeat wash). Adherent cells in ECM-coated wells should all be adhered and spread out evenly. Invert plate gently onto an absorbent diaper pad. Remove the remaining wash solution from each well with pipette (see Note 7). 5. The disintegrins or snake venom fractions to be tested need to be dissolved in HBSS at a concentration of 2 μg protein/well (see Note 8) and incubated with cells at room temperature for 30 min. As a control for this experiment, the same number of cells should be allowed to bind in parallel to ECM proteincoated wells, in the absence of the disintegrins. 6. At the end of experiment, with multichannel pipette, slowly add 100 μL/well of freshly diluted 1% glutaraldehyde in PBS. Fix for 10 min at room temperature. Alternatively, fix with 4% paraformaldehyde at room temperature for 10–15 min. Invert

Integrin-mediated Cell Adhesion

217

plate gently onto an absorbent diaper pad. Remove the remaining fix solution from each well with pipette (see Note 7). 7. With multipipette, add 100 μL/well of freshly filtered (use 0.2 μm syringe filter) crystal violet (0.1% in H2O from a solution of 5 mg/mL in 2% ethanol). Stain for 30 min at room temperature. Invert plate onto an absorbent diaper pad, and then wash plate gently by immersion in a plastic tray containing tap water. Invert plate onto an absorbent diaper pad. Remove the remaining wash from each well with pipette. Reimmerse in fresh tap water. Invert plate onto an absorbent diaper pad. Remove the remaining wash from each well with pipette. Repeat if required. Allow to dry for 5–10 min at room temperature. 8. With multipipette, add 50 μL/well of 0.5% Triton X-100 (diluted in PBS) and allow to solubilize the adhered cells overnight at room temperature in a drawer or use 2% SDS at room temperature for 30 min. Read on a spectrophotometer the absorbance at 550 nm (SDS) or 595 nm (Triton X-100). BSA background should be less than 0.1 OD (see Note 9). ECM protein-coated wells’ optical density values should be about 1.0–2.0 OD. In the presence of snake venom integrin antagonists (HPLC fractions) the values decrease in a dosedependent fashion and usually represent about 80–100% of the values in wells without disintegrins. A typical ECM proteinmediated cell adhesion experiment using this method is presented in Fig. 1. 3.5 Adhesion Assay Using CMFDA Staining

For CMFDA staining, the cells were incubated at 37  C for 30 min with 12.5 μM CMFDA or another cell tracker dye. The labeled cells were then centrifuged at 1000 rpm and washed twice with HBSS containing 1% BSA to remove excess CMFDA. 1. 100 μL of the cell suspension was added to each well (1.1
105 cells/well) and the plate was incubated at 37  C in the presence or absence of snake venom integrin antagonist for 30 or 60 min (see Note 8). 2. The wells were gently washed three times with 1% (w/v) BSA in HBSS to remove unattached or loosely bound cells (see Note 3). 3. Cells that had firmly adhered to the wells were lysed by the addition of 0.5% Triton X-100 (diluted in HBSS or doubledistilled water). 4. The fluorescence in each well was quantified with a SPECTRAFluor Plus plate reader (Tecan, San Jose, CA) at λex ¼ 485 nm and λem ¼ 530 nm. To determine the number of adhered cells from the fluorescence values, a standard curve was generated by serial dilutions of known numbers of CMFDA-labeled cells

218

Philip Lazarovici et al.

(Fig. 2a). Typical CTL or disintegrin-mediated or anti-integrin subunit monoclonal antibody-mediated cell adhesion experiments using CMFDA method are presented in Fig. 2. A typical HPLC fraction adhesion assay using CMFDA method, to identify members of C-lectin-type proteins, antagonists of α2β1 integrin, in the Echis sochureki viper’s venom is presented in Fig. 3. 3.6 Adhesion Assay Using Indocyanine Green (ICG) Staining

1. Measure out 1 mg of the indocyanine green powder. Dissolve the ICG powder in 100 μL of dimethyl sulfoxide (DMSO). Add 400 μL of Dulbecco’s modified Eagle’s medium (DMEM + 10% fetal calf serum) media to the mixture and shake it well. This results in a final concentration of 2 mg/mL of ICG. 2. For ICG staining, the cells are incubated at 37  C for 60 min with 100 μL of ICG solution. The labeled cells are then centrifuged at 1000 rpm and washed twice with HBSS containing 1% BSA to remove excess ICG. 3. Count cells to 1
105/mL, add 50 μL cell suspension to each well, and perform the adhesion assay for 30 min in a CO2 incubator at 37  C (see Note 5) in the presence and absence of snake venom fraction or disintegrin. Examine the plate in an inverted microscope. Photograph selected wells if desired. Invert plate gently onto an absorbent diaper pad. Remove the remaining solution from each well with pipette (see Note 6). 4. With the multipipette, slowly add 100 μL/well of serum-free medium or HBSS containing 1% BSA, down the side of each well. Invert plate gently onto an absorbent diaper pad. Remove the remaining wash solution from each well with pipette (see Note 7). Repeat this step three times. 5. The living cells are ready to be imaged. Alternatively add 100 μL/well of freshly diluted 1% glutaraldehyde in PBS. Fix for 10 min at room temperature. Remove the remaining fix solution from each well with pipette (see Note 7). 6. Cell-associated NIR intensity is estimated using Odyssey Infrared Imager (LI-COR Biosciences) under the following conditions: resolution, 170–340 microns; pixel area, 0.03 mm2; quality, medium-low; focus offset, 1–3; channels, 800 nm; and intensity, 1–3. NIR intensity in control groups measures the ICG binding to ECM-coated wells (noise, background) in the absence of cells. The NIR intensity is estimated semiquantitatively using Odyssey software (see Note 10). Specific adherence by NIR imaging is defined as the difference between the NIR intensity of cells on ECM-coated wells and NIR intensity of the ECM-coated wells without cells (see Note 11).Typical experiments of disintegrin (viperistatin)- and CTL (VP-12)mediated adhesion comparing the CMFDA and NIR methods are presented in Figs. 4 and 5.

Integrin-mediated Cell Adhesion

4

219

Notes 1. Perform all steps under sterile conditions. 2. The choice of ECM coating will depend on the cells being used. For cells which adhered poorly (such as HEK293), we used a mixture of fibronectin and collagen type 1. 3. When removing solutions, tilt the plate slightly, place pipette at the side of the well, and gently draw the solution up. 4. Pre-warm trypsin or HyQTase solution (for detaching cells) and medium to 37  C in a water bath prior to the experiment. If using trypsin to detach cells, it is advised to subsequently inhibit trypsin activity either by using trypsin inhibitors or by adding 5 mL growth medium containing serum. We recommend using HyQTase, or a similar product, rather than digestive enzymes such as trypsin to detach cells for short-term assays where integrin function is under investigation. Integrins are sensitive to trypsin cleavage and therefore the use of trypsin can affect the time taken for cells to adhere to the ECM-coated wells. In some cases, when cells are extremely adherent, the use of trypsin may be unavoidable and necessary for ECM degradation to break the cell-ECM binding. 5. Avoid errors in cell counting and/or cell clumping by thoroughly mixing the cell suspension and ensuring sufficient cellcell dissociation prior to counting/seeding. Use multichannel pipettes where appropriate, to ensure an equal amount of cells to each well. Optimize the number of cells plated for every cell line being tested; for monitoring cell adhesion to the ECM, it is best to avoid overcrowding of cells/confluent monolayers. 6. To achieve consistency, always add/remove DMEM gently with a multichannel pipette for multiple wells. 7. Do not allow wells to dry! 8. Snake venom fractions to be tested should be used immediately following dissolution in water or medium. Activity may decrease with storage, even when frozen. A reliable approach is to weigh out 100 μg aliquots of lyophilized dry fractions with a microbalance for use in individual experiments. 9. Usually all negative controls should have a very low optical density but in some cases they are too high or even denser than the testing wells. From our experience, the red in the medium interfered with the optical density and therefore, we used phenol red-free culture medium for the adhesion assay. 10. Image with an Odyssey Family Imaging System (https://www. licor.com/). If signal is too strong or too weak, adjust the

220

Philip Lazarovici et al.

imaging parameters to optimize the image. Use the AutoScan function to improve the dynamic range of the image. 11. To achieve optimal NIR measurements, adherent cells should consistently cover the complete surface of each well. If this cannot be done with the quantity of cells recommended, adjust the quantity of cells in order to cover the well. Additionally, once cells have been seeded, we have found that gentle agitation of the plate on a surface in a clockwise followed by an anticlockwise direction helps to disperse the cells more evenly across the well. Nevertheless, if performing the assay for the first time, or if using a new cell line, in particular epithelial or endothelial cells which prefer to have cell-cell contacts, we recommend using a normal tissue culture-treated 96-well plate to optimize cell numbers and the technique for cell addition to the ECM-coated well.

Acknowledgments Philip Lazarovici holds the Jacob Gitlin Chair in Physiology and is affiliated with, and partially supported by, the Grass Center for Drug Design and Synthesis of Novel Therapeutics, David R. Bloom Center of Pharmacy, and the Adolph and Klara Brettler Medical Research Center at the Hebrew University of Jerusalem, Israel. Peter I Lelkes is the Laura H. Carnell Professor of Bioengineering.Cezary Marcinkiewicz and Peter I Lelkes acknowledge support through a grant from Temple University’s Moulder Center for Drug Discovery. References 1. Gumbiner BM (1996) Cell adhesion: the molecular basis of tissue architecture and morphogenesis. Cell 84:345–357 2. Akiyama SK (1996) Integrins in cell adhesion and signaling. Hum Cell 9:181–186 3. Hynes RO (1987) Integrins: a family of cell surface receptors. Cell 48:549–554 4. Humphries JD, Chastney MR, Askari JA, Humphries MJ (2018) Signal transduction via integrin adhesion complexes. Curr Opin Cell Biol 56:14–21. https://doi.org/10.1016/j. ceb.2018.08.004 5. Manninen A, Varjosalo M (2017) A proteomics view on integrin-mediated adhesions. Proteomics 17. https://doi.org/10.1002/pmic. 201600022

6. Gehler S, Ponik SM, Riching KM, Keely PJ (2013) Bi-directional signaling: extracellular matrix and integrin regulation of breast tumor progression. Crit Rev Eukaryot Gene Expr 23:139–157 7. Patarroyo M, Tryggvason K, Virtanen I (2002) Laminin isoforms in tumor invasion, angiogenesis and metastasis. Semin Cancer Biol 12:197–207. https://doi.org/10.1016/ S1044-579X(02)00023-8 8. Juliano RL, Varner JA (1993) Adhesion molecules in cancer: the role of integrins. Curr Opin Cell Biol 5:812–818 9. Ruoslahti E (1996) RGD and other recognition sequences for integrins. Annu Rev Cell Dev Biol 12:697–715. https://doi.org/10. 1146/annurev.cellbio.12.1.697

Integrin-mediated Cell Adhesion 10. Calvete JJ, Marcinkiewicz C, Monleon D, Esteve V, Celda B, Juarez P, Sanz L (2005) Snake venom disintegrins: evolution of structure and function. Toxicon 45:1063–1074. https://doi.org/10.1016/j.toxicon.2005.02. 024 11. Marcinkiewicz C (2013) Applications of snake venom components to modulate integrin activities in cell-matrix interactions. Int J Biochem Cell Biol 45:1974–1986. https://doi.org/10. 1016/j.biocel.2013.06.009 12. David V, Succar BB, de Moraes JA, SaldanhaGama RFG, Barja-Fidalgo C, Zingali RB (2018) Recombinant and Chimeric Disintegrins in Preclinical Research. Toxins 10. https://doi.org/10.3390/toxins10080321 13. Fujii Y, Okuda D, Fujimoto Z, Horii K, Morita T, Mizuno H (2003) Crystal structure of trimestatin, a disintegrin containing a cell adhesion recognition motif RGD. J Mol Biol 332:1115–1122 14. Swenson S, Ramu S, Markland FS (2007) Antiangiogenesis and RGD-containing snake venom disintegrins. Curr Pharm Des 13:2860–2871 15. Kisiel DG, Calvete JJ, Katzhendler J, Fertala A, Lazarovici P, Marcinkiewicz C (2004) Structural determinants of the selectivity of KTS-disintegrins for the alpha1beta1 integrin. FEBS Lett 577:478–482. https://doi.org/10. 1016/j.febslet.2004.10.050 16. Moreno-Murciano MP, Monleon D, Calvete JJ, Celda B, Marcinkiewicz C (2003) Amino acid sequence and homology modeling of obtustatin, a novel non-RGD-containing short disintegrin isolated from the venom of Vipera lebetina obtusa. Protein Sci 12:366–371. https://doi.org/10.1110/ps. 0230203 17. Walsh EM, Marcinkiewicz C (2011) NonRGD-containing snake venom disintegrins, functional and structural relations. Toxicon 58:355–362. https://doi.org/10.1016/j. toxicon.2011.07.004 18. Takeda S (2016) ADAM and ADAMTS family proteins and snake venom metalloproteinases: a structural overview. Toxins 8. https://doi. org/10.3390/toxins8050155 19. Huang TF (1998) What have snakes taught us about integrins? Cell Mol Life Sci 54:527–540. https://doi.org/10.1007/s000180050181 20. Lu X, Lu D, Scully MF, Kakkar VV (2005) Snake venom metalloproteinase containing a disintegrin-like domain, its structure-activity relationships at interacting with integrins.

221

Curr Med Chem Cardiovasc Hematol Agents 3:249–260 21. Apte SS (2009) A disintegrin-like and metalloprotease (reprolysin-type) with thrombospondin type 1 motif (ADAMTS) superfamily: functions and mechanisms. J Biol Chem 284:31493–31497. https://doi.org/10. 1074/jbc.R109.052340 22. Arlinghaus FT, Eble JA (2012) C-type lectinlike proteins from snake venoms. Toxicon 60:512–519. https://doi.org/10.1016/j. toxicon.2012.03.001 23. Momic T, Cohen G, Reich R, Arlinghaus FT, Eble JA, Marcinkiewicz C, Lazarovici P (2012) Vixapatin (VP12), a c-type lectin-protein from Vipera xantina palestinae venom: characterization as a novel anti-angiogenic compound. Toxins 4:862–877. https://doi.org/10. 3390/toxins4100862 24. Staniszewska I, Walsh EM, Rothman VL, Gaathon A, Tuszynski GP, Calvete JJ, Lazarovici P, Marcinkiewicz C (2009) Effect of VP12 and viperistatin on inhibition of collagen-receptor-dependent melanoma metastasis. Cancer Biol Ther 8:1507–1516 25. Arlinghaus FT, Momic T, Ammar NA, Shai E, Spectre G, Varon D, Marcinkiewicz C, Heide H, Lazarovici P, Eble JA (2013) Identification of alpha2beta1 integrin inhibitor VP-i with anti-platelet properties in the venom of Vipera palaestinae. Toxicon 64:96–105. https://doi.org/10.1016/j.toxicon.2013.01. 001 26. Jakubowski P, Calvete JJ, Eble JA, Lazarovici P, Marcinkiewicz C (2013) Identification of inhibitors of alpha2beta1 integrin, members of C-lectin type proteins, in Echis sochureki venom. Toxicol Appl Pharmacol 269:34–42. https://doi.org/10.1016/j.taap.2013.03.002 27. Momic T, Katzehendler J, Benny O, Lahiani A, Cohen G, Noy E, Senderowitz H, Eble JA, Marcinkiewicz C, Lazarovici P (2014) Vimocin and vidapin, cyclic KTS peptides, are dual antagonists of alpha1beta1/ alpha2beta1 integrins with antiangiogenic activity. J Pharmacol Exp Ther 350:506–519. https://doi.org/10.1124/jpet.114.214643 28. Momic T, Katzhendler J, Shai E, Noy E, Senderowitz H, Eble JA, Marcinkiewicz C, Varon D, Lazarovici P (2015) Vipegitide: a folded peptidomimetic partial antagonist of alpha2beta1 integrin with antiplatelet aggregation activity. Drug Des Devel Ther 9:291–304. https://doi.org/10.2147/DDDT.S72844

222

Philip Lazarovici et al.

29. Humphries MJ (2009) Cell adhesion assays. Methods Mol Biol 522:203–210. https://doi. org/10.1007/978-1-59745-413-1_14 30. Pimton P, Sarkar S, Sheth N, Perets A, Marcinkiewicz C, Lazarovici P, Lelkes PI (2011) Fibronectin-mediated upregulation of alpha5beta1 integrin and cell adhesion during differentiation of mouse embryonic stem cells. Cell Adhes Migr 5:73–82 31. Souza DH, Iemma MR, Ferreira LL, Faria JP, Oliva ML, Zingali RB, Niewiarowski S, Selistre-de-Araujo HS (2000) The disintegrinlike domain of the snake venom metalloprotease alternagin inhibits alpha2beta1 integrinmediated cell adhesion. Arch Biochem Biophys 384:341–350. https://doi.org/10.1006/ abbi.2000.2120 32. Cominetti MR, Ribeiro JU, Fox JW, Selistrede-Araujo HS (2003) BaG, a new dimeric metalloproteinase/disintegrin from the Bothrops alternatus snake venom that interacts with alpha5beta1 integrin. Arch Biochem Biophys 416:171–179 33. Vitale M, Di Matola T, Fenzi G, Illario M, Rossi G (1998) Fibronectin is required to prevent thyroid cell apoptosis through an integrinmediated adhesion mechanism. J Clin Endocrinol Metab 83:3673–3680. https://doi.org/ 10.1210/jcem.83.10.5175 34. Amarante-Mendes JG, Starobinas N, Macedo MS (1994) Colorimetric assay for the measurement of thymocyte/thymic stromal cell adhesion. Brazilian J Med Biol Res 27:1321–1330 35. Bonnekoh B, Wevers A, Jugert F, Merk H, Mahrle G (1989) Colorimetric growth assay for epidermal cell cultures by their crystal violet binding capacity. Arch Dermatol Res 281:487–490 36. Dutt MK (1980) Staining of depolymerised DNA in mammalian tissues with methyl violet 6B and crystal violet. Folia Histochem Cytochem 18:79–83 37. Poole CA, Brookes NH, Gilbert RT, Beaumont BW, Crowther A, Scott L, Merrilees MJ (1996) Detection of viable and non-viable cells in connective tissue explants using the fixable fluoroprobes 5-chloromethylfluorescein diacetate and ethidium homodimer-1. Connect Tissue Res 33:233–241 38. Heilingoetter CL, Jensen MB (2016) Histological methods for ex vivo axon tracing: A systematic review. Neurol Res 38:561–569. https://doi.org/10.1080/01616412.2016. 1153820

39. Gallagher SJ, Shank JA, Bochner BS, Wagner EM (2002) Methods to track leukocyte and erythrocyte transit through the bronchial vasculature in sheep. J Immunol Methods 271:89–97 40. Ragnarson B, Bengtsson L, Haegerstrand A (1992) Labeling with fluorescent carbocyanine dyes of cultured endothelial and smooth muscle cells by growth in dye-containing medium. Histochemistry 97:329–333 41. Anderson WM, Trgovcich-Zacok D (1995) Carbocyanine dyes with long alkyl side-chains: broad spectrum inhibitors of mitochondrial electron transport chain activity. Biochem Pharmacol 49:1303–1311 42. Lu W, McCallum L, Irvine AE (2009) A rapid and sensitive method for measuring cell adhesion. J Cell Commun Signal 3:147–149. https://doi.org/10.1007/s12079-009-00528 43. Portnoy E, Lecht S, Lazarovici P, Danino D, Magdassi S (2011) Cetuximab-labeled liposomes containing near-infrared probe for in vivo imaging. Nanomedicine 7:480–488. https://doi.org/10.1016/j.nano.2011.01. 001 44. Cohen G, Lecht S, Arien-Zakay H, Ettinger K, Amsalem O, Oron-Herman M, Yavin E, Prus D, Benita S, Nissan A, Lazarovici P (2012) Bio-imaging of colorectal cancer models using near infrared labeled epidermal growth factor. PLoS One 7:e48803. https:// doi.org/10.1371/journal.pone.0048803 45. Cohen G, Lecht S, Oron-Herman M, Momic T, Nissan A, Lazarovici P (2013) Near infrared optical visualization of epidermal growth factor receptors levels in COLO205 colorectal cell line, orthotopic tumor in mice and human biopsies. Int J Mol Sci 14:14669–14688. https://doi.org/10.3390/ ijms140714669 46. Lee H-J, Lee J-S, Hwang SJ, Lee H-Y (2015) Insulin-like growth factor binding protein-3 inhibits cell adhesion via suppression of integrin beta4 expression. Oncotarget 6:15150–15163. https://doi.org/10.18632/ oncotarget.3825 47. Unger C, Felldin U, Rodin S, Nordenskjold A, Dilber S, Hovatta O (2016) Derivation of human skin fibroblast lines for feeder cells of human embryonic stem cells. Curr Protoc Stem Cell Biol 36:1C.7.1–1C.711. https:// doi.org/10.1002/9780470151808. sc01c07s36

Integrin-mediated Cell Adhesion 48. Delwel GO, Hogervorst F, Kuikman I, Paulsson M, Timpl R, Sonnenberg A (1993) Expression and function of the cytoplasmic variants of the integrin alpha 6 subunit in transfected K562 cells. Activation-dependent adhesion and interaction with isoforms of laminin. J Biol Chem 268:25865–25875

223

49. Kawaguchi S, Hemler ME (1993) Role of the alpha subunit cytoplasmic domain in regulation of adhesive activity mediated by the integrin VLA-2. J Biol Chem 268:16279–16285

Chapter 12 Using C. elegans to Study the Effects of Toxins in Sensory Ion Channels In Vivo Valeria Va´squez Abstract Caenorhabditis elegans is a powerful animal model in which transgenesis, behavior, and physiology can be merged to study in vivo the effect of natural and synthetic agonists in sensory ion channels. Worms have polymodal sensory neurons (like the ASH pair) that couple ion channel activation with a robust and easily scorable aversive-like behavior. We expressed the transient receptor potential vanilloid 1 (TRPV1) channel from rat (r) in worms’ ASH neurons and determined its sensitivity to the tarantula double-knot toxin (DkTx) and the active component of chili peppers (capsaicin). This chapter describes protocols for generating and maintaining transgenic rTRPV1 worms to determine dose-dependent behavior. The goal is to provide an efficient tool to characterize the function of sensory channels (wild type and mutants) in vivo. Key words C. elegans, Worms, Toxins, DkTx, Capsaicin, Sensory ion channels, Transient receptor potential channels, TRPV1, Behavior, Pain

1

Introduction The transient receptor potential vanilloid 1 (TRPV1) is a polymodal excitatory ion channel expressed in somatosensory neurons activated by heat, proalgesic agents, and natural toxins [1]. Permanent activation of TRPV1 by inflammatory mediators contributes to neuronal hyperactivity and chronic pain [2]; hence, TRPV1 is a promising target for the management of neurogenic inflammation. TRPV1 is also the target of toxins produced by plants (e.g., cactuslike plant, chili pepper) and animals (e.g., tarantula, centipede, scorpion, snake, anemone), generally causing severe and prolonged pain, accompanied by inflammation [3]. Unlike other toxins, the double-knot toxin (DkTx) from the tarantula Ornithoctonus huwena binds irreversibly to TRPV1 and traps it in its open state by associating with the pore-forming region of the channel [4]. This irreversibility could translate into persistent neuronal depolarization, prolonged pain, and ultimately neuronal

Avi Priel (ed.), Snake and Spider Toxins: Methods and Protocols, Methods in Molecular Biology, vol. 2068, https://doi.org/10.1007/978-1-4939-9845-6_12, © Springer Science+Business Media, LLC, part of Springer Nature 2020

225

226

Valeria Va´squez

degeneration. Unlike mammals, worms’ life cycle is short and provides the opportunity to quickly engineer transgenic animals that carry large transgene constructs integrated in their genomes. We used a transgenic worm that carries a codon-optimized rat (r) rTRPV1 gene with artificial introns to study how the tarantula DkTx and the active component of chili peppers (capsaicin) modulate TRPV1 channel function in vivo. The nonparasitic nematode Caenorhabditis elegans is an excellent model system for studying nociception. Like polymodal multidendritic neurons in mammals, amphid ASH neurons in C. elegans also respond to chemical, osmotic, and mechanical stimuli, with transient increases in cytoplasmic Ca2+ [5]. Moreover, because ASH neurons face the external environment, one could use this feature to expose them to different chemical agonists and evoke aversive-like behaviors (Fig. 1). Outstandingly, it has been shown that activation of non-native worm ion channels, such as rat TRPV1 and rat TRPV4 [6–9] expressed in ASH neurons, elicits neuronal depolarization and avoidance responses typical of wild-type ASH receptors. There are many tools that make this animal model very efficient for characterizing the modulation of sensory ion channels by chemical agonists, including expression of transgenes in specific neurons, mid- and high-throughput behavioral assays, Ca2+ imaging, and electrophysiology. Here, we present protocols to characterize how toxins modulate mammalian sensory ion channels using mid-throughput behavior assays.

Fig. 1 Behavioral tests are performed by placing drops in front of moving young adult worms on plates without food. For sensitization assays, worms are incubated for 10–15 min in control solution with 5 μM DkTx before the capsaicin challenge. Assays are performed after 15 min of transferring worms from OP50-containing NGM plates. Top panel shows the lack of response of rTRPV1-transgenic worms to control solution. Bottom panel shows aversive-like responses from rTRPV1-trangenic worms exposed to agonist-containing drops. Trials that elicit aversive-like responses are scored as positive withdrawal responses

The Effect of DkTx and Capsaicin in Worms Expressing rTRPV1

2

227

Materials – Prepare solutions with ddH2O, unless otherwise indicated. – Use low-binding pipette tips to reduce surface tension on the tip wall and avoid losing worms and eggs throughout the transfer procedures.

2.1 Worm Growth Media

1. Cholesterol: 5 mg/mL Stock in ethanol (molecular biology grade). Store in a glass bottle at 4
C. 2. CaCl2: 1 M Stock in water. Filter solution (with a 0.22 μm filter) and store at room temperature. 3. MgCl2: 1 M Stock in water. Filter solution (with a 0.22 μm filter) and store at room temperature. 4. KH2PO4: 1 M Stock in water. Weigh 136 g and add water up to 900 mL. Adjust pH to 6.0 with KOH pellets and fill cylinder up to 1000 mL. Autoclave and store at room temperature. 5. LB broth microbiology media: Weigh 25 g of LB powder, add up to 1 L of water, and mix. Autoclave keeping the bottle lids loose, and store at room temperature. 6. LB plates: Add 7.5 g of agar to glass media storage bottles containing 500 mL of LB media (see Note 1). Autoclave, keeping the bottle lids loose. Subsequently put the bottles in a water bath at 55
C until the bottles can be handled with bare hands. Pour approximately 20 mL of LB agar into 100 mm Petri dishes. 7. A starter culture of E. coli strain OP50 from the Caenorhabditis Genetic Center (CGC).

2.2 Nucleic Acid Isolation

1. M9 buffer: 3 g of KH2PO4, 6 g of Na2HPO4, 5 g of NaCl, 1 mL of 1 M MgSO4 in ffi800 mL of water. Complete volume with water up to 1 l. Sterilize by autoclaving. 2. KCl: 1 M Stock in water. Dissolve 3.73 g of KCl in 25 mL of water. Fill tube with water up to 50 mL. Store at room temperature. 3. Tris–HCl: 1 M Stock in water. Dissolve 6.06 g of Tris base in 20 mL of water. Adjust pH to 8 with 1 M HCl. Fill tube with water up to 50 mL. Store at room temperature. 4. Gelatin: 2% Stock in water. Dissolve 0.01 g of gelatin in 100 mL of water while stirring on a hot plate. Fill tube with water up to 50 mL. Store at 4
C. 5. 2 Worm lysis buffer: 2.5 mL of 1 M KCl, 0.5 mL of 1 M Tris–HCl pH 8, 1.25 mL of 1 M MgCl2, 225 μL of 100% Tergitol-type NP-40, 225 μL of 100% Tween 20, and 250 μL of 2% gelatin. Fill tube with water up to 50 mL. Filter solution (with a 0.22 μm filter) and store at 4
C.

228

2.3

Valeria Va´squez

Genotyping

1. Thermocycler. 2. Commercial polymerase chain reaction reagents: Polymerase enzyme, corresponding buffer, oligos (forward and reverse), and deoxynucleotide mix. 3. Commercial reverse transcriptase reaction reagents: Reverse transcriptase enzyme, corresponding buffer, and deoxynucleotide mix. 4. TAE buffer: 10 Stock. Dissolve 48.4 g of Tris base and 3.7 g of ethylenediaminetetraacetic acid (EDTA) in 11.4 mL of glacial acetic acid and 800 mL of water. Complete final volume to 1 l and store at room temperature. 5. One percent agarose gel: Add 0.5 g of agarose (molecular biology grade) to 50 mL of 1 TAE buffer in a glass narrowmouth Erlenmeyer flask. Microwave at 50% power for 5 min, or until the agarose is completely dissolved. Pour agarose in a 7  10 cm DNA gel tray with a fixed-height comb. 6. DNA ladder. 7. DNA gel-loading dye.

2.4

Transgenesis

1. Vibration isolation table. 2. Inverted DIC microscope, with 5 and 40 water-dipping objective. 3. Micromanipulator to control and position the glass (needle) pipette. 4. Microinjection glass pipette. 5. Pressurized injection system. 6. Injection oil. 7. Worm pick. 8. Two percent agarose pad: Add 1 g of agarose to 50 mL of M9 buffer in a glass narrow-mouth Erlenmeyer flask. Microwave at 50% power for 7 min, or until the agarose is completely dissolved. Aliquot agarose in clear glass test tubes with push caps. Store at 4
C. 9. Worm strain carrying ttTi5605 II; unc-119(ed3) III. 10. Site-specific MosSCI addgene.org/19330).

vectors

(pCFJ151—https://www.

11. Co-injection marker plasmids (pGH8—Prab-3::mCherry:unc54utr; pCFJ90—Pmyo-2::mCherry::unc-54utr UTR; pCFJ104—Pmyo-3::mCherry::unc-54 UTR). 12. Negative selection marker plasmid (pMA122—Phsp-16.41:: peel-1::tbb-2utr).

The Effect of DkTx and Capsaicin in Worms Expressing rTRPV1

2.5

Behavior

229

1. M13 buffer: 30 mM Tris–HCl, 100 mM NaCl, 10 mM KCl. Adjust pH to 7.0. Filter solution (with a 0.22 μm filter), wrap the lid with parafilm, and store at 4
C (see Note 2). 2. Control solution: M13 plus 1% v/v ethanol, reagent grade. 3. Standard dissecting microscope. 4. Borosilicate glass capillaries: L, 3.5 in. (9 cm); ID x OD ¼ 0.53  1.14 mm. 5. Micropipette puller. 6. KOH: 5 M Stock in water. Dissolve 14 g of KOH pellets in 25 mL of water. Fill tube with water up to 50 mL. Store at room temperature. 7. DkTx 50 μM stock: Dissolve 21 μg of DkTx in 50 μL of water. 8. Capsaicin 10 mM stock: Dissolve 3 mg of in 1 mL of DMSO. 9. Capsaicin 125 μM stock: Add 18.75 μL of capsaicin 10 mM stock to 1481.25 μL of control solution.

3 3.1

Methods Worm Food

1. Prepare nematode growth media (NGM) [10] as follows: Dissolve 3 g of NaCl and 2.5 g of Bacto-peptone in 900 mL of water. Fill tube with water up to 975 mL. Aliquot 488 mL into two 500 mL glass media storage bottles and add 8.5 g of agar to each bottle. Autoclave, keeping the bottle lids loose. Subsequently put the bottles in a water bath at 55
C until the bottles can be handled with bare hands. 2. Once the media has cooled down to 55
C, in a sterile environment (close to a flame) add to each 488 mL media bottle the following reagents in the listed order: 500 μL of 5 mg/mL cholesterol, 500 μL of 1 M CaCl2, 500 μL of 1 M MgSO4, and 25 mL of 1 M KH2PO4, pH 6.0. Swirl the bottles until the media becomes homogenous. 3. Immediately pour 12 mL of media into 60 mm plastic culture dishes (see Note 3). Once the media has solidified the NGM plates are ready for further use. 4. Isolate single colonies of OP50 on LB agar plates from the CGC starter culture. 5. Pick a single OP50 colony to inoculate 250 mL of sterile LB media in a 500 mL glass narrow-mouth Erlenmeyer flask. Grow the cultures at 37
C for 16 h at 250 rpm. The OP50enriched media is now ready to inoculate the NGM plates and/or to store at 4
C, for several months.

230

Valeria Va´squez

6. Add 700 μL of OP50-enriched media to each NGM plate. Incubate the food plates at 37
C for 2 days. Subsequently, let the plates reach room temperature before further use or storage at 4
C (see Note 4). 3.2 Quick DNA Isolation and PCR

This protocol yields DNA of sufficient quality to amplify up to 3000 bp. 1. Add approximately 1.5 mL of water to a food plate with 30 gravid hermaphrodites and pipette the water up and down several times to get the worms and eggs that stick to the bacterial lawn. 2. Collect the worm-containing liquid and transfer it to a 1.5 mL tube. 3. Let the worms settle by gravity (approximately 10 min) on the bottom of the tube without using a centrifuge. 4. Remove the supernatant and add 1.4 mL of water. 5. Repeat steps 3–4 three times. In the final step, leave approximately 40 μL of water. 6. Add 40 μL of 2 worm lysis solution to a PCR tube. Using the same pipette tip, remove 40 μL of water that contains worms. 7. Add 1 μL of proteinase K and shake vigorously. 8. Keep the PCR tube at 80
C for at least 10 min (see Note 5). 9. Take the PCR tube out of the freezer and immediately place it in a thermocycler. 10. Run a program with the following parameters: 60
C for 1 h, 80
C for 15 min, and 4
C on hold. 11. Immediately after the program has concluded, transfer 40–50 μL of the supernatant to a clean microcentrifuge tube and store at 80
C until further use (see Note 6). 12. Use 1–2 μL of the supernatant as the DNA template for the PCR reaction. 13. Perform PCR using a TAQ polymerase kit and following the manufacturer’s protocol.

3.3 RNA Isolation and RT-PCR

1. Add approximately 1.5 mL of M9 buffer to a food plate with 100 gravid hermaphrodites and pipette the water up and down several times to get the worms and eggs that stick to the bacterial lawn. 2. Collect the worm-containing buffer and transfer it to a 15 mL conical tube. 3. Spin down the tube at 2000  g for 1 min, remove the supernatant, and add M9 buffer up to approximately 2 mL. Repeat this step and carefully remove the supernatant.

The Effect of DkTx and Capsaicin in Worms Expressing rTRPV1

231

4. Add 1 mL of Trizol (inside a chemical hood). 5. Vortex and freeze in liquid nitrogen while swirling the conical tube. 6. Thaw the frozen worms completely while swirling it in a 37
C water bath. 7. Repeat steps 5–6 six times (see Note 7). 8. Transfer the Trizol-worm microcentrifuge tube.

mixture

to

a

2

mL

9. Vortex for 30 s and place on ice for 30 s. Repeat this step three times. 10. Add 200 μL of chloroform (inside a chemical hood). 11. Vortex for 15 s and let the mixture sit at room temperature for 3 min. 12. Spin down the tube at 12,000  g for 15 min at 4
C. 13. After spinning down, there should be an approximately 400 μL bottom layer (pink), an approximately 100 μL middle layer (yellowish), and an approximately 700 μL top layer (transparent). 14. Transfer only the top transparent layer into a new 2 mL microcentrifuge tube. 15. Add 500 μL of isopropanol and mix by inverting the tube three times. Let the tube sit at room temperature for 10 min. 16. Keep the tube at 20
C for at least 1 h. 17. Spin down the tube at 12,000  g for 15 min at 4
C. 18. Carefully remove the supernatant without removing the RNA, which must be left on a side of the tube. 19. Add 1 mL of 75% ethanol (molecular biology grade). 20. Spin down the tube at 7500  g for 5 min at room temperature. 21. Carefully remove as much supernatant as possible and air-dry the pellet inside a chemical hood until it is completely dry, and the ethanol smell is gone. 22. Add 10 μL of RNase-free water and put the tube inside a 60
C water bath for 10 min. 23. Measure the RNA concentration at 260 nm. 24. Spin down at 2000  g and use the supernatant in further steps. 25. Use 1 μg of RNA as a template for the RT-PCR reaction. 26. Perform the RT-PCR using a one-step kit and following the manufacturer’s protocol.

232

3.4

Valeria Va´squez

Transgenesis

1. Synthesize the rTRPV1 gene as codon-optimized for C. elegans protein expression machinery and with three synthetic introns and avoidance of splice sites in coding regions [11]. 2. Clone TRPV1 construct transgenes into a MosSCI targeting vector (pCFJ151) [12]. 3. Mix in microfuge tube the TRPV1 constructs with co-injection array marker plasmids (pGH8, pCFJ90, and pCFJ104), and negative selection marker plasmid (pMA122—Phsp-16.41:: peel-1::tbb-2utr), and Mos1 transposase plasmid (pCFJ601— Peft-3::Mos1-transposase). 4. Purify all plasmids with a commercial kit. 5. Centrifuge mixture for 5 min at 12,000  g. 6. Fill the microinjection pipette with the DNA mixture and mount on the manipulator. 7. Locate and position the pipette tip in the middle of the field of view. 8. Add a drop of oil to a 2% agarose pad. 9. Use a worm pick to transfer young adult hermaphrodites of appropriate strain (i.e., ttTi5605 II; unc-119(ed3) III) from the food plate to the oil. 10. Move the worm gently until the ventral side opposes the tip of the microinjection pipette. 11. Inject the DNA mix into the distal arm of the gonad [13] of a worm strain. 12. Recover the injected worms with 20 μL of M9 buffer and transfer to food plate. 13. At day 5 screen injected worm for rescued “crawling” progeny. Collect five crawlers from each crawling plate and set aside for backup use. 14. The remaining plate is heat shocked (34
C) for 60 min in an air incubator to remove arrays. 15. Surviving crawlers are examined for the absence of array markers. 16. Candidates are confirmed by PCR across insertion locus to detect site-specific insertion at MosSCI site.

3.5 Worm Synchronization

1. Add approximately 1.5 mL of water to a food plate with gravid hermaphrodites and pipette the water up and down several times to get the worms and eggs that stick to the bacterial lawn. 2. Collect the worm-containing liquid and transfer it to a 1.5 mL tube.

The Effect of DkTx and Capsaicin in Worms Expressing rTRPV1

233

3. Spin down the tube at 2000  g for 45 s, remove the supernatant, and add water up to approximately 1.4 mL. Repeat this step twice and leave the worms in 1 mL of water. 4. Add 120 μL of household bleach and 20 μL of 5 M KOH and vortex or shake vigorously. 5. Incubate the tube for 10–15 min while shaking it periodically. Monitor in a dissecting scope and stop when 70% of the worms are degraded and the eggs are released. 6. Spin down the tube at 2000  g for 45 s, keep the eggs, and carefully remove the supernatant. 7. Add water up to approximately 1 mL and spin down again; remove the supernatant and leave approximately 40 μL of water and resuspend. 8. Plate the desired amount of eggs on the food plates (see Note 8). 9. Incubate the plates at the desired culture temperature (e.g., 20
C) and measure behavior after 3 days (see Note 9). 3.6

Drop Assay

This assay should be performed on three separate days. During each day, test between 10 and 25 worms per condition (i.e., strain, toxin class, toxin concentrations). 1. Make microinjection needles out of borosilicate glass capillaries using a micropipette puller following the manufacturer’s instructions. 2. Mount the microinjection needles in a holder with rubber tubing and operate by mouth pipetting. The average drop size should be about 5 nL [14] (see Note 10). 3. Wash off young adult hermaphrodites from food plates with 500 μL of control solution (see Note 11) and transfer the solution to a 1.5 mL tube. Spin down the tube at 2000  g for 45 s and carefully remove the supernatant. 4. Add 40–50 μL of control buffer; transfer the worms to an empty NGM plate, and let them acclimate for 15 min (see Note 12). 5. Perform behavioral trials under a dissecting scope by placing a drop of the control solution in front of moving young adult hermaphrodites [14] (Fig. 1, top panel, and Fig. 2, no DkTx) (see Note 13). Trials that elicit aversive-like responses are scored as positive withdrawal responses (Fig. 1, bottom panel). 6. After determining that the control solution does not elicit withdrawal responses (