Plant Cold Acclimation: Methods and Protocols [2nd ed.] 9781071606599, 9781071606605

Table of contents :
Front Matter ....Pages i-xii
Introduction: Plant Cold Acclimation and Winter Survival (Dirk K. Hincha, Ellen Zuther)....Pages 1-7
Measuring Freezing Tolerance of Leaves and Rosettes: Electrolyte Leakage and Chlorophyll Fluorescence Assays (Anja Thalhammer, Majken Pagter, Dirk K. Hincha, Ellen Zuther)....Pages 9-21
Differential Thermal Analysis: A Fast Alternative to Frost Tolerance Measurements (Andrey V. Malyshev, Ilka Beil, Juergen Kreyling)....Pages 23-31
Infrared Thermal Analysis of Plant Freezing Processes (Gilbert Neuner, Edith Lichtenberger)....Pages 33-41
Conducting Field Trials for Frost ToleranceBreeding in Cereals (Luigi Cattivelli, Cristina Crosatti)....Pages 43-52
A Whole-Plant Screening Test to Select Freezing-Tolerant and Low-Dormant Genotypes (Annick Bertrand, Annie Claessens, Josée Bourassa, Solen Rocher, Vern S. Baron)....Pages 53-60
Mapping of Quantitative Trait Loci (QTL) Associated with Plant Freezing Tolerance and Cold Acclimation (Evelyne Téoulé, Carine Géry)....Pages 61-84
Identification of Arabidopsis Mutants with Altered Freezing Tolerance (Carlos Perea-Resa, Rafael Catalá, Julio Salinas)....Pages 85-97
Cryo-Scanning Electron Microscopy to Study the Freezing Behavior of Plant Tissues (Seizo Fujikawa, Keita Endoh)....Pages 99-117
Using Pixel-Based Microscope Images to Generate 3D Reconstructions of Frozen and Thawed Plant Tissue (David P. Livingston III, Tan D. Tuong)....Pages 119-139
Phenotyping Plant Cellular and Tissue Level Responses to Cold with Synchrotron-Based Fourier-Transform Infrared Spectroscopy and X-Ray Computed Tomography (Ian R. Willick, Jarvis Stobbs, Chithra Karunakaran, Karen K. Tanino)....Pages 141-159
Proteomic Approaches to Identify Proteins Responsive to Cold Stress (Anna M. Jozefowicz, Stefanie Döll, Hans-Peter Mock)....Pages 161-170
Proteomic Approaches to Identify Cold-Regulated Plasma Membrane Proteins (Md Mostafa Kamal, Daisuke Takahashi, Takato Nakayama, Yushi Miki, Yukio Kawamura, Matsuo Uemura)....Pages 171-186
A Lipidomic Approach to Identify Cold-Induced Changes in Arabidopsis Membrane Lipid Composition (Yu Song, Hieu Sy Vu, Sunitha Shiva, Carl Fruehan, Mary R. Roth, Pamela Tamura et al.)....Pages 187-202
Multiplexed Profiling and Data Processing Methods to Identify Temperature-Regulated Primary Metabolites Using Gas Chromatography Coupled to Mass Spectrometry (Alexander Erban, Federico Martinez-Seidel, Yogeswari Rajarathinam, Frederik Dethloff, Isabel Orf, Ines Fehrle et al.)....Pages 203-239
Determining the ROS and the Antioxidant Status of Leaves During Cold Acclimation (Andras Bittner, Thomas Griebel, Jörn van Buer, Ilona Juszczak-Debosz, Margarete Baier)....Pages 241-254
Analysis of Changes in Plant Cell Wall Composition and Structure During Cold Acclimation (Daisuke Takahashi, Ellen Zuther, Dirk K. Hincha)....Pages 255-268
Subcellular Compartmentation of Metabolites Involved in Cold Acclimation (Imke I. Hoermiller, Thomas Nägele, Arnd G. Heyer)....Pages 269-275
Mathematical Modeling of Plant Metabolism in a Changing Temperature Regime (Lisa Fürtauer, Thomas Nägele)....Pages 277-287
Characterization of Ice-Binding Proteins from Sea-Ice Microalgae (Maddalena Bayer-Giraldi, EonSeon Jin, Peter W. Wilson)....Pages 289-302
Isolation and Characterization of Ice-Binding Proteins from Higher Plants (Melissa Bredow, Heather E. Tomalty, Laurie A. Graham, Audrey K. Gruneberg, Adam J. Middleton, Barbara Vanderbeld et al.)....Pages 303-332
Back Matter ....Pages 333-336

Citation preview

Methods in Molecular Biology 2156

Dirk K. Hincha Ellen Zuther Editors

Plant Cold Acclimation Methods and Protocols Second Edition

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.

Plant Cold Acclimation Methods and Protocols Second Edition

Edited by

Dirk K. Hincha and Ellen Zuther Max-Planck-Institut für Molekulare Pflanzenphysiologie, Potsdam, Germany

Editors Dirk K. Hincha Max-Planck-Institut fu¨r Molekulare Pflanzenphysiologie Potsdam, Germany

Ellen Zuther Max-Planck-Institut fu¨r Molekulare Pflanzenphysiologie Potsdam, Germany

ISSN 1064-3745 ISSN 1940-6029 (electronic) Methods in Molecular Biology ISBN 978-1-0716-0659-9 ISBN 978-1-0716-0660-5 (eBook) https://doi.org/10.1007/978-1-0716-0660-5 © 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, expressed 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: 1 New York Plaza, New York, NY 10004, U.S.A.

Preface It has frequently been remarked in the scientific literature that plants as sessile organisms are forced to deal with changes in environmental conditions by rapid acclimation, because they obviously cannot simply move to a more favorable location. Over evolutionary time scales, this has provided plants with complex molecular, physiological, and morphological mechanisms to thrive and survive under stressful conditions. Prevalent stresses that plants are exposed to both in natural and agricultural environments are low temperatures and freezing, i.e., the crystallization of ice in living tissues. Not surprisingly, many plant species are able to survive tissue freezing. In addition, plants from temperate and boreal climate zones are able to adapt to low temperatures prior to an actual freezing event. This phenomenon of plant cold acclimation, i.e., the increase of freezing tolerance under conditions of low, but non-freezing, temperatures, has been the subject of intensive scientific study for many decades. More recently, phenomena such as deacclimation (the loss of freezing tolerance when plants are exposed to warmer temperatures after cold acclimation) and cold memory (when plants react differently to a second cold exposure compared to the first) have gained interest in the research community. Under natural conditions, cold acclimation helps plants to survive seasonal low winter temperatures. Cold acclimation and freezing tolerance are quantitative traits, and cold acclimation is accompanied by complex changes in gene expression, enzyme activities, and the contents of a large number of proteins, primary and secondary metabolites, and lipids. In addition, even a few days under cold conditions can trigger irreversible morphological changes, particularly in growing plant tissues. Therefore, research on plant cold acclimation, deacclimation, and cold memory can (and quite often has to) be performed at different organizational levels, from populations to single genes and molecules. This requires experimental expertise in scientific disciplines such as ecology, plant breeding, biophysics, computational biology, genetics, physiology, or molecular biology, and strongly favors interdisciplinary approaches. This volume of Methods in Molecular Biology combines a wide selection of experimental methods ranging from the whole plant level of ecology and breeding to molecular profiling and the detailed analysis of specific proteins, with many levels of investigated complexity in between. This second edition provides updates on chapters describing methods already included in the first edition but also a number of chapters dealing with new methodology that has found its way into the field in recent years. We hope that this collection of detailed experimental protocols will help researchers, both new and experienced, to enter this exciting field of research or broaden the scope of their investigations. Potsdam, Germany

Dirk K. Hincha Ellen Zuther

v

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

v ix

1 Introduction: Plant Cold Acclimation and Winter Survival . . . . . . . . . . . . . . . . . . . Dirk K. Hincha and Ellen Zuther 2 Measuring Freezing Tolerance of Leaves and Rosettes: Electrolyte Leakage and Chlorophyll Fluorescence Assays . . . . . . . . . . . . . . . . . . . . Anja Thalhammer, Majken Pagter, Dirk K. Hincha, and Ellen Zuther 3 Differential Thermal Analysis: A Fast Alternative to Frost Tolerance Measurements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Andrey V. Malyshev, Ilka Beil, and Juergen Kreyling 4 Infrared Thermal Analysis of Plant Freezing Processes . . . . . . . . . . . . . . . . . . . . . . . Gilbert Neuner and Edith Lichtenberger 5 Conducting Field Trials for Frost Tolerance Breeding in Cereals . . . . . . . . . . . . . . Luigi Cattivelli and Cristina Crosatti 6 A Whole-Plant Screening Test to Select Freezing-Tolerant and Low-Dormant Genotypes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Annick Bertrand, Annie Claessens, Jose´e Bourassa, Solen Rocher, and Vern S. Baron 7 Mapping of Quantitative Trait Loci (QTL) Associated with Plant Freezing Tolerance and Cold Acclimation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Evelyne Te´oule´ and Carine Ge´ry 8 Identification of Arabidopsis Mutants with Altered Freezing Tolerance . . . . . . . . Carlos Perea-Resa, Rafael Catala´, and Julio Salinas 9 Cryo-Scanning Electron Microscopy to Study the Freezing Behavior of Plant Tissues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Seizo Fujikawa and Keita Endoh 10 Using Pixel-Based Microscope Images to Generate 3D Reconstructions of Frozen and Thawed Plant Tissue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . David P. Livingston III and Tan D. Tuong 11 Phenotyping Plant Cellular and Tissue Level Responses to Cold with Synchrotron-Based Fourier-Transform Infrared Spectroscopy and X-Ray Computed Tomography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ian R. Willick, Jarvis Stobbs, Chithra Karunakaran, and Karen K. Tanino 12 Proteomic Approaches to Identify Proteins Responsive to Cold Stress . . . . . . . . . Anna M. Jozefowicz, Stefanie Do¨ll, and Hans-Peter Mock 13 Proteomic Approaches to Identify Cold-Regulated Plasma Membrane Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Md Mostafa Kamal, Daisuke Takahashi, Takato Nakayama, Yushi Miki, Yukio Kawamura, and Matsuo Uemura

1

vii

9

23 33 43

53

61 85

99

119

141

161

171

viii

14

15

16

17

18

19

20 21

Contents

A Lipidomic Approach to Identify Cold-Induced Changes in Arabidopsis Membrane Lipid Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yu Song, Hieu Sy Vu, Sunitha Shiva, Carl Fruehan, Mary R. Roth, Pamela Tamura, and Ruth Welti Multiplexed Profiling and Data Processing Methods to Identify Temperature-Regulated Primary Metabolites Using Gas Chromatography Coupled to Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Alexander Erban, Federico Martinez-Seidel, Yogeswari Rajarathinam, Frederik Dethloff, Isabel Orf, Ines Fehrle, Jessica Alpers, Olga Beine-Golovchuk, and Joachim Kopka Determining the ROS and the Antioxidant Status of Leaves During Cold Acclimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Andras Bittner, Thomas Griebel, Jo¨rn van Buer, Ilona Juszczak-Debosz, and Margarete Baier Analysis of Changes in Plant Cell Wall Composition and Structure During Cold Acclimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Daisuke Takahashi, Ellen Zuther, and Dirk K. Hincha Subcellular Compartmentation of Metabolites Involved in Cold Acclimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Imke I. Hoermiller, Thomas N€ a gele, and Arnd G. Heyer Mathematical Modeling of Plant Metabolism in a Changing Temperature Regime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ¨ rtauer and Thomas N€ Lisa Fu a gele Characterization of Ice-Binding Proteins from Sea-Ice Microalgae . . . . . . . . . . . . Maddalena Bayer-Giraldi, EonSeon Jin, and Peter W. Wilson Isolation and Characterization of Ice-Binding Proteins from Higher Plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Melissa Bredow, Heather E. Tomalty, Laurie A. Graham, Audrey K. Gruneberg, Adam J. Middleton, Barbara Vanderbeld, Peter L. Davies, and Virginia K. Walker

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

187

203

241

255

269

277 289

303

333

Contributors JESSICA ALPERS • Applied Metabolome Analysis Research Group, Max Planck Institute of Molecular Plant Physiology, Potsdam, Germany MARGARETE BAIER • Plant Physiology, Dahlem Center of Plant Sciences, Freie Universit€ at Berlin, Berlin, Germany VERN S. BARON • Que´bec Research and Development Centre, Agriculture and Agri-Food Canada, Que´bec, QC, Canada MADDALENA BAYER-GIRALDI • Helmholtz Centre for Polar and Marine Research, Alfred Wegener Institute, Bremerhaven, Germany ILKA BEIL • Institute of Botany and Landscape Ecology, Experimental Plant Ecology, University of Greifswald, Greifswald, Germany OLGA BEINE-GOLOVCHUK • Applied Metabolome Analysis Research Group, Max Planck Institute of Molecular Plant Physiology, Potsdam, Germany; Nuclear Pore Complex and Ribosome Assembly, Biochemie-Zentrum, Universit€ a t Heidelberg, Heidelberg, Germany ANNICK BERTRAND • Que´bec Research and Development Centre, Agriculture and Agri-Food Canada, Que´bec, QC, Canada; Lacombe Research and Development Centre, Agriculture and Agri-Food Canada, Lacombe, AB, Canada ANDRAS BITTNER • Plant Physiology, Dahlem Center of Plant Sciences, Freie Universit€ at Berlin, Berlin, Germany JOSE´E BOURASSA • Que´bec Research and Development Centre, Agriculture and Agri-Food Canada, Que´bec, QC, Canada MELISSA BREDOW • Department of Biology, Queen’s University, Kingston, ON, Canada RAFAEL CATALA´ • Departamento de Biotecnologı´a Microbiana y de Plantas, Centro de Investigaciones Biologicas Margarita Salas (CIB-CSIC), Madrid, Spain LUIGI CATTIVELLI • CREA, Research Centre for Genomics and Bioinformatics, Fiorenzuola d’Arda, Italy ANNIE CLAESSENS • Que´bec Research and Development Centre, Agriculture and Agri-Food Canada, Que´bec, QC, Canada CRISTINA CROSATTI • CREA, Research Centre for Genomics and Bioinformatics, Fiorenzuola d’Arda, Italy PETER L. DAVIES • Department of Biology, Queen’s University, Kingston, ON, Canada; Department of Biomedical and Molecular Sciences, Queen’s University, Kingston, ON, Canada FREDERIK DETHLOFF • Applied Metabolome Analysis Research Group, Max Planck Institute of Molecular Plant Physiology, Potsdam, Germany; Proteomics and Biomarkers, Max Planck Institute of Psychiatry, Mu¨nchen, Germany STEFANIE DO¨LL • Leibniz Institute of Plant Biochemistry (IPB), Halle (Saale), Germany; Molecular Interaction Ecology, German Centre for Integrative Biodiversity Research (iDiv) Halle-Jena-Leipzig, Leipzig, Germany KEITA ENDOH • Forestry and Forest Products Research Institute, Forest Tree Breeding Center, Hitachi, Japan ALEXANDER ERBAN • Applied Metabolome Analysis Research Group, Max Planck Institute of Molecular Plant Physiology, Potsdam, Germany

ix

x

Contributors

INES FEHRLE • Applied Metabolome Analysis Research Group, Max Planck Institute of Molecular Plant Physiology, Potsdam, Germany CARL FRUEHAN • Division of Biology, Kansas Lipidomics Research Center, Kansas State University, Manhattan, KS, USA SEIZO FUJIKAWA • Faculty of Agriculture, Hokkaido University, Sapporo, Japan LISA FU¨RTAUER • Evolution€ a re Zellbiologie der Pflanzen, Ludwig-Maximilians-Universit€ at Mu¨nchen, Planegg, Germany CARINE GE´RY • Institut Jean-Pierre Bourgin, Equipe VAST, Versailles Cedex, France LAURIE A. GRAHAM • Department of Biomedical and Molecular Sciences, Queen’s University, Kingston, ON, Canada THOMAS GRIEBEL • Plant Physiology, Dahlem Center of Plant Sciences, Freie Universit€ at Berlin, Berlin, Germany AUDREY K. GRUNEBERG • Department of Biomedical and Molecular Sciences, Queen’s University, Kingston, ON, Canada ARND G. HEYER • Department of Plant Biotechnology, Institute of Biomaterials and Biomolecular Systems, University of Stuttgart, Stuttgart, Germany DIRK K. HINCHA • Max-Planck-Institut fu¨r Molekulare Pflanzenphysiologie, Potsdam, Germany IMKE I. HOERMILLER • Department of Plant Biotechnology, Institute of Biomaterials and Biomolecular Systems, University of Stuttgart, Stuttgart, Germany EON SEON JIN • Hanyang University, Seoul, Republic of Korea ANNA M. JOZEFOWICZ • Department of Physiology and Cell Biology, Leibniz Institute of Plant Genetics and Crop Plant Research (IPK), Stadt Seeland, Germany ILONA JUSZCZAK-DEBOSZ • Plant Physiology, Dahlem Center of Plant Sciences, Freie Universit€ at Berlin, Berlin, Germany; Molecular Physiology, Rheinische Friedrich-WilhelmsUniversit€ at Bonn, Bonn, Germany MD MOSTAFA KAMAL • United Graduate School of Agricultural Sciences, Faculty of Agriculture, Iwate University, Morioka, Iwate, Japan CHITHRA KARUNAKARAN • Canadian Light Source, Saskatoon, SK, Canada YUKIO KAWAMURA • United Graduate School of Agricultural Sciences, Faculty of Agriculture, Iwate University, Morioka, Iwate, Japan; Department of Plant-Bioscience, Faculty of Agriculture, Iwate University, Morioka, Iwate, Japan JOACHIM KOPKA • Applied Metabolome Analysis Research Group, Max Planck Institute of Molecular Plant Physiology, Potsdam, Germany JUERGEN KREYLING • Institute of Botany and Landscape Ecology, Experimental Plant Ecology, University of Greifswald, Greifswald, Germany EDITH LICHTENBERGER • Department of Botany, Unit Functional Plant Biology, Stress Physiology, University of Innsbruck, Innsbruck, Austria DAVID P. LIVINGSTON III • USDA-ARS and North Carolina State University, Raleigh, NC, USA ANDREY V. MALYSHEV • Institute of Botany and Landscape Ecology, Experimental Plant Ecology, University of Greifswald, Greifswald, Germany FEDERICO MARTINEZ-SEIDEL • Applied Metabolome Analysis Research Group, Max Planck Institute of Molecular Plant Physiology, Potsdam, Germany ADAM J. MIDDLETON • Department of Biology, Queen’s University, Kingston, ON, Canada; Department of Biochemistry, University of Otago, Dunedin, New Zealand YUSHI MIKI • Department of Plant-Bioscience, Faculty of Agriculture, Iwate University, Morioka, Iwate, Japan

Contributors

xi

HANS-PETER MOCK • Department of Physiology and Cell Biology, Leibniz Institute of Plant Genetics and Crop Plant Research (IPK), Stadt Seeland, Germany € GELE • Plant Evolutionary Cell Biology, Department Biology I, LudwigTHOMAS NA Maximilians-Universit€ at Mu¨nchen, Planegg-Martinsried, Germany TAKATO NAKAYAMA • Department of Plant-Bioscience, Faculty of Agriculture, Iwate University, Morioka, Iwate, Japan GILBERT NEUNER • Department of Botany, Unit Functional Plant Biology, Stress Physiology, University of Innsbruck, Innsbruck, Austria ISABEL ORF • Applied Metabolome Analysis Research Group, Max Planck Institute of Molecular Plant Physiology, Potsdam, Germany; Owlstone Medical Ltd, Cambridge, UK MAJKEN PAGTER • Department of Chemistry and Bioscience, Aalborg University, Aalborg East, Denmark CARLOS PEREA-RESA • Departamento de Biotecnologı´a Microbiana y de Plantas, Centro de Investigaciones Biologicas Margarita Salas (CIB-CSIC), Madrid, Spain YOGESWARI RAJARATHINAM • Applied Metabolome Analysis Research Group, Max Planck Institute of Molecular Plant Physiology, Potsdam, Germany SOLEN ROCHER • Que´bec Research and Development Centre, Agriculture and Agri-Food Canada, Que´bec, QC, Canada MARY R. ROTH • Division of Biology, Kansas Lipidomics Research Center, Kansas State University, Manhattan, KS, USA JULIO SALINAS • Departamento de Biotecnologı´a Microbiana y de Plantas, Centro de Investigaciones Biologicas Margarita Salas (CIB-CSIC), Madrid, Spain SUNITHA SHIVA • Division of Biology, Kansas Lipidomics Research Center, Kansas State University, Manhattan, KS, USA; Department of Biochemistry and Molecular Biophysics, Kansas State University, Manhattan, KS, USA YU SONG • Division of Biology, Kansas Lipidomics Research Center, Kansas State University, Manhattan, KS, USA JARVIS STOBBS • Canadian Light Source, Saskatoon, SK, Canada DAISUKE TAKAHASHI • United Graduate School of Agricultural Sciences, Faculty of Agriculture, Iwate University, Morioka, Iwate, Japan; Graduate School of Science and Engineering, Saitama University, Saitama, Japan; Max-Planck-Institut fu¨r Molekulare Pflanzenphysiologie, Potsdam, Germany PAMELA TAMURA • Division of Biology, Kansas Lipidomics Research Center, Kansas State University, Manhattan, KS, USA KAREN K. TANINO • Department of Plant Sciences, University of Saskatchewan, Saskatoon, SK, Canada EVELYNE TE´OULE´ • Sorbonne-Universite´, Paris Cedex 05, France; Institut Jean-Pierre Bourgin, Equipe VAST, Versailles Cedex, France ANJA THALHAMMER • Physikalische Biochemie, Universit€ at Potsdam, Potsdam, Germany HEATHER E. TOMALTY • Department of Biomedical and Molecular Sciences, Queen’s University, Kingston, ON, Canada TAN D. TUONG • USDA-ARS and North Carolina State University, Raleigh, NC, USA MATSUO UEMURA • United Graduate School of Agricultural Sciences, Faculty of Agriculture, Iwate University, Morioka, Iwate, Japan; Department of Plant-Bioscience, Faculty of Agriculture, Iwate University, Morioka, Iwate, Japan JO¨RN VAN BUER • Plant Physiology, Dahlem Center of Plant Sciences, Freie Universit€ at Berlin, Berlin, Germany BARBARA VANDERBELD • Department of Biology, Queen’s University, Kingston, ON, Canada

xii

Contributors

HIEU SY VU • Division of Biology, Kansas Lipidomics Research Center, Kansas State University, Manhattan, KS, USA; Children’s Medical Research Institute at University of Texas-Southwestern, Dallas, TX, USA VIRGINIA K. WALKER • Department of Biology, Queen’s University, Kingston, ON, Canada; Department of Biomedical and Molecular Sciences, Queen’s University, Kingston, ON, Canada RUTH WELTI • Division of Biology, Kansas Lipidomics Research Center, Kansas State University, Manhattan, KS, USA IAN R. WILLICK • Department of Plant Sciences, University of Saskatchewan, Saskatoon, SK, Canada PETER W. WILSON • Southern Cross University, Lismore, NSW, Australia ELLEN ZUTHER • Max-Planck-Institut fu¨r Molekulare Pflanzenphysiologie, Potsdam, Germany

Chapter 1 Introduction: Plant Cold Acclimation and Winter Survival Dirk K. Hincha and Ellen Zuther Abstract This introductory chapter provides a brief overview of plant freezing tolerance, cold acclimation, including subzero acclimation, and the subsequent deacclimation when plants return to warm conditions favoring growth and development. We describe the basic concepts and approaches that are currently followed to investigate these phenomena. We highlight the multidisciplinary nature of these investigations and the necessity to use methodologies from different branches of science, such as ecology, genetics, physiology, cell biology, biochemistry, and biophysics to gain a complete understanding of the complex adaptive mechanisms ultimately underlying plant winter survival. Key words Cold acclimation, Experimental approaches, Freezing tolerance, Global climate change

The phenomenon of plant cold acclimation, that is, the increase of freezing tolerance during exposure to low but nonfreezing temperatures, has already been described in the nineteenth century (see ref. 1 for references and details) and has been the subject of intensive scientific study ever since (see refs. 1–4 for comprehensive reviews). The basic phenotypic readout for cold acclimation is the increased survival of plants, tissues, or cells after a freeze–thaw cycle through a damaging temperature range. Unfortunately, not only in common usage but also in the scientific literature, the term freezing tolerance is often used synonymously with cold or low temperature tolerance. This is very unfortunate because freezing and cold/low temperature denote completely different concepts. While freezing is a clearly defined physical process (i.e., the crystallization of ice), cold is a completely subjective term, not only for humans but also for plants. As an example, many tropical and subtropical plants suffer severe damage at temperatures below approximately 15  C, while Antarctic algae show a heat-shock response already at 5  C [5]. In addition, mechanisms of injury are very different. Low temperature (or chilling) damage is a direct temperature effect. Freezing damage, on the other hand, is mainly the result of osmotic dehydration triggered by extracellular ice crystallization that leads Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_1, © Springer Science+Business Media, LLC, part of Springer Nature 2020

1

2

Dirk K. Hincha and Ellen Zuther

to the diffusion of water from the cells to the growing ice crystals [1–3]. Under natural conditions, cold acclimation is a response of plants to cope with seasonal low temperature. Since flowering plants were initially only adapted to the tropical climate that was prevalent in the Mesozoic era and only became exposed to temperate climates after severe climate cooling events in the Eocene and Oligocene, the ability to survive in temperate climates with seasonal cold had to evolve from tropical species. It has recently been estimated that less than half of all angiosperm families have members that are adapted for survival under seasonal cold conditions (see ref. 6 for a review) and that this trait has developed several times independently [7]. Cold acclimated plants are additionally able to increase their freezing tolerance further when they are exposed to mild, nondamaging temperatures below 0  C in a process called subzero acclimation [8–12]. While the molecular basis of subzero acclimation is still unclear, it has been shown that the transcriptional responses are distinct from those observed during cold acclimation [13] but it requires previous cold acclimation and involves specific modifications of cell wall structure and composition [14]. In support of strong natural selection for freezing tolerance in plants, natural variation in this trait has been reported in both herbaceous and tree species. This variation is usually related to latitudinal and/or climatic gradients that are obvious candidates as potential driving forces of selection [15–19]. From a more practical point of view, such natural variation can be used to investigate the potential function of metabolic and physiological trait variation in determining the differential freezing tolerance of different genotypes within a species through correlation analysis [19, 20], quantitative trait locus (QTL) [21–23], or genomewide association (GWA) mapping studies [24]. In addition, in crop plant species the corresponding variability may be used for breeding purposes, and for the identification of molecular markers that may be used for marker-assisted breeding. Obviously, global climate change has a major impact on the winter survival of plants and this impact is going to increase in the coming decades. In general, winters in the cold regions of the Earth are getting milder and naively one might assume that in future freezing tolerance will become less important for plant survival and geographical distribution, and also for crop yields. However, warmer winters are accompanied by reduced snow fall and a higher incidence of erratic early or late season frost. Since snow cover is a very effective insulator, plants under snow are exposed to significantly less severe freezing temperatures than in the absence of snow. This can lead to the counterintuitive effect that plants are exposed to lower freezing temperatures as global warming progresses. In addition, premature warming in early spring can lead to loss of

Plant Freezing Tolerance and Winter Survival

3

freezing tolerance (deacclimation) [25, 26], making plants more prone to damage during later cold spells [27]. Therefore, research on all aspects of plant survival in winter, including, in addition to cold acclimation, much neglected areas such as subzero acclimation and deacclimation, will be relevant more than ever in times of global climate change (see ref. 28 for a recent review). There are generally three strategies available to plants to survive winter. The most obvious strategy that is used by many herbaceous annuals is seasonal avoidance by surviving winter as seeds or as roots or rhizomes buried sufficiently deeply in the soil to evade freezing. Deep supercooling (i.e., preventing ice crystallization even at temperatures significantly below 0  C) is a frequently observed winter survival strategy for example in cold acclimated trees [29]. The third strategy that is embraced by all plants that acclimate to cold is to modify their cellular constituents in a way that allows for survival at lower subzero temperatures in the presence of extracellular ice. The term supercooling usually refers to a solution (cellular or otherwise) that shows no ice crystallization when cooled below its melting point. It should be emphasized here that the freezing point of pure water is, contrary to common perception, not identical to its melting point at 0  C. Ice crystallization at or slightly below 0  C requires impurities that can serve as seeds. Pure water only crystallizes at about 42  C, the homogeneous nucleation temperature. In any biological system it is probably impossible to remove all molecules and structures that could serve to seed ice crystallization. In addition, colligative freezing point depression from common metabolites such as sugars or amino acids only has a very limited potential to achieve significant supercooling as a concentration of 1 M of any (nondissociating) solute will depress the freezing/ melting point only by 1.84  C and the ability of living cells to accumulate solutes is obviously limited. Therefore, supercooling relies on biological antifreezes, such as specific proteins [30] and more complex solutes such as flavonoids and tannins [31]. These substances allow some plants and algae to remain unfrozen at subzero temperatures, similar to Arctic and Antarctic fish and some species of insects that produce highly efficient antifreeze or thermal hysteresis proteins [30]. In addition to their thermal hysteresis activity, many such proteins also exhibit ice recrystallization inhibition activity, that is, the ability to suppress the growth of large ice crystals at the expense of smaller crystals. In particular in the plant proteins, thermal hysteresis activity is very low (0.1–0.5  C) and they are mainly referred to as ice binding proteins (IBP) [32]. Their main function is probably the regulation of ice crystal size in the intercellular spaces of plant tissues [33], or in the case of sea-ice algae binding of the algal cells to ice surfaces [34]. An additional strategy that has been discovered in trees is the sealing of particularly sensitive plant organs, such as flower buds,

4

Dirk K. Hincha and Ellen Zuther

from ice crystals that form in other parts of the plant [35, 36]. Specific anatomical structures have been identified in these studies that prevent the spreading of ice into such organs. Cold acclimation and freezing tolerance are genetically complex, quantitative traits. The increase in freezing tolerance during cold acclimation is accompanied by complex physiological changes that are, at least to a large extent, based on complex changes in gene expression (see ref. 4 for a review). Understanding of temperature perception in plants is still quite fragmentary [37], but recent work has suggested that phytochromes may be involved as primary sensors [38, 39]. On the other hand, work from many laboratories primarily in Arabidopsis thaliana has shed light on the signal transduction cascades regulating the expression of important target genes encoding potential protective proteins such as the COR/LEA proteins [40]. In addition, genes encoding enzymes either directly involved in reactive oxygen species (ROS) detoxification or the biosynthesis of low molecular weight antioxidants, and enzymes responsible for the biosynthesis of compatible solutes are upregulated during cold acclimation [41]. For instance, in Arabidopsis and Thellungiella compatible solutes such as sugars and proline are strongly accumulated during plant cold acclimation and their content is highly correlated with freezing tolerance across diverse genotypes [19, 42, 43]. In fact, compatible solutes can have strong predictive value for freezing tolerance in Arabidopsis [20]. In addition to compatible solutes, many other primary and secondary metabolites are accumulated during cold acclimation, but for most of these solutes no specific role in cellular freezing tolerance has been proposed or experimentally shown. However, for flavonoids correlations of the expression levels of biosynthetic genes and of some key compounds with freezing tolerance in different Arabidopsis genotypes was found [44] and mutant studies established a functional role for both flavonols and anthocyanins in Arabidopsis cold acclimation [45]. Also, cellular lipid composition is strongly modified during cold acclimation, both for membrane (diacyl) and storage (triacyl) lipids [46, 47]. Of course, all these metabolic changes are also evident on the gene expression (e.g., [48–52]) and protein abundance [53–55] levels. Investigation of knockout mutants with altered freezing tolerance or cold signal transduction behavior is an invaluable tool to clarify the function of specific genes/proteins or key metabolic pathways or compounds in the process of cold acclimation (see refs. 56–58 for reviews). The brief overview given above should give an idea about the many different organizational levels (from populations to single genes and molecules) and the corresponding scientific disciplines that are involved in research of plant cold acclimation and freezing tolerance, with their respective focus on ecology, breeding, genetics, cell biology, physiology, or molecular biology, or any

Plant Freezing Tolerance and Winter Survival

5

combination of these specialties. Obviously, this area of research strongly favors interdisciplinary approaches. At the same time this means that researchers very often have to combine experimental methods and concepts from different areas of science. We therefore hope that this second edition of the Methods in Molecular Biology volume on freezing tolerance and cold acclimation, that comprises a large range of experimental protocols covering all the mentioned organizational levels and disciplines, will help not only researchers new to this exciting field but also those already working in a particular area of cold acclimation and freezing tolerance research who are looking to expand their range of experimental approaches. References 1. Steponkus PL (1984) Role of the plasma membrane in freezing injury and cold acclimation. Annu Rev Plant Physiol 35:543–584 2. Guy CL (1990) Cold acclimation and freezing stress tolerance: role of protein metabolism. Ann Rev Plant Physiol Plant Mol Biol 41:187–223 3. Levitt J (1980) Responses of plants to environmental stresses. Volume I: chilling, freezing, and high temperature stresses. Physiological ecology, 2nd edn. Academic, Orlando, FL 4. Hincha DK, Espinoza C, Zuther E (2012) Transcriptomic and metabolomic approaches to the analysis of plant freezing tolerance and cold acclimation. In: Tuteja N, Gill SS, Toburcio AF, Tuteja R (eds) Improving crop resistance to abiotic stress, vol 1. Wiley-Blackwell, Berlin, pp 255–287 5. Vayda ME, Yuan M-L (1994) The heat shock response of an antarctic alga is evident at 5 C. Plant Mol Biol 24:229–233 6. Preston JC, Sandve SR (2013) Adaptation to seasonality and the winter freeze. Front Plant Sci 4:167 7. Schubert M, Gronvold L, Sandve SR et al (2019) Evolution of cold acclimation and its role in niche transitions in the temperate grass subfamily Pooideae. Plant Physiol 180:404–419 8. Herman EM, Rotter K, Premakumar R et al (2006) Additional freeze hardiness in wheat acquired by exposure to 3 C is associated with extensive physiological, morphological, and molecular changes. J Exp Bot 57:3601–3618 9. Le MQ, Engelsberger WR, Hincha DK (2008) Natural genetic variation in acclimation capacity at sub-zero temperatures after cold acclimation at 4 C in different Arabidopsis thaliana accessions. Cryobiology 57:104–112

10. Livingston DP III (1996) The second phase of cold hardening: freezing tolerance and fructan isomer changes in winter cereal crowns. Crop Sci 36:1568–1573 11. Livingston DP III, Van K, Premakumar R et al (2007) Using Arabidopsis thaliana as a model to study subzero acclimation in small grains. Cryobiology 54:154–163 12. Olien CR (1984) An adaptive response of rye to freezing. Crop Sci 24:51–54 13. Le MQ, Pagter M, Hincha DK (2015) Global changes in gene expression, assayed by microarray hybridization and quantitative RT-PCR, during acclimation of three Arabidopsis thaliana accessions to sub-zero temperatures after cold acclimation. Plant Mol Biol 87:1–15 14. Takahashi D, Gorka M, Erban A et al (2019) Both cold and sub-zero acclimation induce cell wall modification and changes in the extracellular proteome in Arabidopsis thaliana. Sci Rep 9:2289 15. Holliday JA, Ritland K, Aitken SN (2010) Widespread, ecologically relevant genetic markers developed from association mapping of climate-related traits in Sitka spruce (Picea sitchensis). New Phytol 188:501–514 16. Kreyling J, Thiel D, Simmnacher K et al (2012) Geographic origin and past climatic experience influence the response to late spring frost in four grass species in central Europe. Ecography 35:268–275 17. Kreyling J, Wiesenberg GLB, Thiel D et al (2012) Cold hardiness of Pinus nigra Arnold as influenced by geographic origin, warming, and extreme summer drought. Environ Exp Bot 78:99–108 18. Zhen Y, Ungerer MC (2008) Clinal variation in freezing tolerance among natural accessions of Arabidopsis thaliana. New Phytol 177:419–427

6

Dirk K. Hincha and Ellen Zuther

19. Zuther E, Schulz E, Childs LH et al (2012) Natural variation in the non-acclimated and cold-acclimated freezing tolerance of Arabidopsis thaliana accessions. Plant Cell Environ 35:1860–1878 20. Korn M, G€artner T, Erban A et al (2010) Predicting Arabidopsis freezing tolerance and heterosis in freezing tolerance from metabolite composition. Mol Plant 3:224–235 21. Gery C, Zuther E, Schulz E et al (2011) Natural variation in the freezing tolerance of Arabidopsis thaliana: effects of RNAi-induced CBF depletion and QTL localisation vary among accessions. Plant Sci 180:12–23 22. Meissner M, Orsini E, Ruschhaupt M et al (2013) Mapping quantitative trait loci for freezing tolerance in a recombinant inbred line population of Arabidopsis thaliana accessions Tenela and C24 reveals REVEILLE1 as negative regulator of cold acclimation. Plant Cell Environ 36:1256–1267 ˚ gren J, Atchison RA et al (2014) 23. Oakley CG, A QTL mapping of freezing tolerance: links to fitness and adaptive trade-offs. Mol Ecol 23:4304–4315 24. Horton MW, Willems G, Sasaki E et al (2016) The genetic architecture of freezing tolerance varies across the range of Arabidopsis thaliana. Plant Cell Environ 39:2570–2579 25. Pagter M, Arora R (2013) Winter survival and deacclimation of perennials under warmer climate: physiological perspectives. Physiol Plant 147:75–87 26. Zuther E, Juszczak I, Lee YP et al (2015) Time-dependent deacclimation after cold acclimation in Arabidopsis thaliana accessions. Sci Rep 5:12199 27. Augspurger CK (2013) Reconstructing patterns of temperature, phenology, and frost damage over 124 years: spring damage risk is increasing. Ecology 94:41–50 28. Vyse K, Pagter M, Zuther E et al (2019) Deacclimation after cold acclimation—a crucial, but widely neglected part of plant winter survival. J Exp Bot 70:4595–4604 29. Kuroda K, Kasuga J, Arakawa K et al (2003) Xylem ray parenchyma cells in boreal hardwood species respond to subfreezing temperatures by deep supercooling that is accompanied by incomplete desiccation. Plant Physiol 131:736–744 30. Bar-Dolev M, Braslavsky I, Davies PL (2016) Icebinding proteins and their function. Annu Rev Biochem 85:515–542 31. Kuwabara C, Wang D, Endoh K et al (2013) Analysis of supercooling activity of tanninrelated polyphenols. Cryobiology 67:40–49

32. Bredow M, Walker VK (2017) Ice-binding proteins in plants. Front Plant Sci 8:2153 33. Griffith M, Yaish MWF (2004) Antifreeze proteins in overwintering plants: a tale of two activities. Trends Plant Sci 9:399–405 34. Guo S, Stevens CA, Vance TDR et al (2017) Structure of a 1.5-MDa adhesin that binds its Antarctic bacterium to diatoms and ice. Sci Adv 3:e170440 35. Kuprian E, Munkler C, Resnyak A et al (2017) Complex bud architecture and cell-specific chemical patterns enable supercooling of Picea abies bud primordia. Plant Cell Environ 40:3101–3112 36. Neuner G, Kreische B, Kaplenig D et al (2018) Deep supercooling enabled by surface impregnation with lipophilic substances explains the survival of overwintering buds at extreme freezing. Plant Cell Environ 42:2065–2074 37. Knight MR, Knight H (2012) Low-temperature perception leading to gene expression and cold tolerance in higher plants. New Phytol 195:737–751 38. Jung J-H, Domijan M, Klose C et al (2016) Phytochromes function as thermosensors in Arabidopsis. Science 354:886–889 39. Legris M, Klose C, Burgie ES et al (2016) Phytochrome B integrates light and temperature signals in Arabidopsis. Science 354:897–900 40. Ding Y, Shi Y, Yang S (2019) Advances and challenges in uncovering cold tolerance regulatory mechanisms in plants. New Phytol 222:1690–1704 41. Guy CL, Kaplan F, Kopka J et al (2008) Metabolomics of temperature stress. Physiol Plant 132:220–235 42. Korn M, Peterek S, Mock H-P et al (2008) Heterosis in the freezing tolerance, and sugar and flavonoid contents of crosses between Arabidopsis thaliana accessions of widely varying freezing tolerance. Plant Cell Environ 31:813–827 43. Lee YP, Babakov A, de Boer B et al (2012) Comparison of freezing tolerance, compatible solutes and polyamines in geographically diverse collections of Thellungiella spec. and Arabidopsis thaliana accessions. BMC Plant Biol 12:131 44. Schulz E, Tohge T, Zuther E et al (2015) Natural variation in flavonol and anthocyanin metabolism during cold acclimation in Arabidopsis thaliana accessions. Plant Cell Environ 38:1658–1672 45. Schulz E, Tohge T, Zuther E et al (2016) Flavonoids are determinants of freezing

Plant Freezing Tolerance and Winter Survival tolerance and cold acclimation in Arabidopsis thaliana. Sci Rep 6:34027 46. Degenkolbe T, Giavalisco P, Zuther E et al (2012) Differential remodeling of the lipidome during cold acclimation in natural accessions of Arabidopsis thaliana. Plant J 72:972–982 47. Wang X, Li W, Li M et al (2006) Profiling lipid changes in plant responses to low temperatures. Physiol Plant 126:90–96 48. Fowler S, Thomashow MF (2002) Arabidopsis transcriptome profiling indicates that multiple regulatory pathways are activated during cold acclimation in addition to the CBF cold response pathway. Plant Cell 14:1675–1690 49. Hannah MA, Heyer AG, Hincha DK (2005) A global survey of gene regulation during cold acclimation in Arabidopsis thaliana. PLoS Genet 1:e26 50. Hannah MA, Wiese D, Freund S et al (2006) Natural genetic variation of freezing tolerance in Arabidopsis. Plant Physiol 142:98–112 51. Kreps JA, Wu Y, Chang HS et al (2002) Transcriptome changes for Arabidopsis in response to salt, osmotic, and cold stress. Plant Physiol 130:2129–2141 52. Oono Y, Seki M, Satou M et al (2006) Monitoring expression profiles of Arabidopsis genes

7

during cold acclimation and deacclimation using DNA microarrays. Funct Integr Genomics 6:212–234 53. Amme S, Matros A, Schlesier B et al (2006) Proteome analysis of cold stress response in Arabidopsis thaliana using DIGE-technology. J Exp Bot 57:1537–1546 54. Kawamura Y, Uemura M (2003) Mass spectrometric approach to identifying putative plasma membrane proteins of Arabidopsis leaves associated with cold acclimation. Plant J 36:141–154 55. Kjellsen TD, Shiryaeva L, Schro¨der W et al (2010) Proteomics of extreme freezing tolerance in Siberian spruce (Picea obovata). J Proteome 73:965–975 56. Chinnusamy V, Zhu J, Zhu J-K (2007) Cold stress regulation of gene expression in plants. Trends Plant Sci 12:444–451 57. Medina J, Catala R, Salinas J (2011) The CBFs: three Arabidopsis transcription factors to cold acclimate. Plant Sci 180:3–11 58. Thomashow MF (2010) Molecular basis of plant cold acclimation: insights gained from studying the CBF cold response pathway. Plant Physiol 154:571–577

Chapter 2 Measuring Freezing Tolerance of Leaves and Rosettes: Electrolyte Leakage and Chlorophyll Fluorescence Assays Anja Thalhammer, Majken Pagter, Dirk K. Hincha, and Ellen Zuther Abstract Quantitative assessment of freezing tolerance is essential to unravel plant adaptations to cold temperatures. Not only the survival of whole plants, but also impairment of detached leaves or small rosettes after a freeze–thaw cycle can be used to accurately quantify plant freezing tolerance in terms of LT50 values. Here we describe two methods to determine the freezing tolerance of detached leaves or rosettes using a full or selected set of freezing temperatures and an additional method using chlorophyll fluorescence as a different physiological parameter. Firstly, we illustrate how to assess the integrity of (predominantly) the plasma membrane during freezing using an electrolyte leakage assay. Secondly, we provide a chlorophyll fluorescence imaging protocol to determine the freezing tolerance of the photosynthetic apparatus. Key words Freezing tolerance, LT50, Electrolyte leakage, Chlorophyll fluorescence, Fv/Fm

1

Introduction Robust experimental approaches for the precise quantification of plant freezing tolerance are of fundamental importance to understand the genetic and molecular mechanisms underlying and determining this complex trait. Cellular membranes are widely accepted as primary sites of freezing damage (see ref. 1 for a comprehensive review). Therefore, next to plant survival, methods assessing cellular membrane integrity are frequently used to determine plant freezing tolerance. The two methods we describe here provide stable and highly reproducible LT50 values, defined as the temperatures at which 50% of damage occurs. Moreover, combining the two protocols allows to discriminate the site of freezing damage between plasma membrane and chloroplasts. This constitutes a powerful tool to investigate the mechanistics of plants altered in their freezing tolerance by breeding or genetic engineering and to test the site of activity of known or novel cellular protectants [2]. In addition, we have used both methods also to assess the natural variation in the freezing tolerance of different A. thaliana

Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_2, © Springer Science+Business Media, LLC, part of Springer Nature 2020

9

10

Anja Thalhammer et al.

accessions [3–6], to distinguish between cold acclimation and deacclimation of different Arabidopsis accessions [7] and to investigate subzero acclimation [8, 9]. Moreover, both protocols can easily be adapted for use with other plant species, such as electrolyte leakage assays in Eutrema salsugineum (previously Thellungiella salsuginea) [10]. The first part of the protocol we describe concerns the controlled-rate freezing of detached leaves or rosettes. Subsequently, the thawed leaves can be used for conductivity measurements to assess electrolyte leakage. Various forms of this method have been used for many years, not only to determine freezing damage but also in the context of many other stresses that may have an impact on cellular membranes, such as drought or reactive oxygen species. The method we use here was originally described in [11]. It not only reports on the intactness of the plasma membrane as a semipermeable barrier for intracellular ions but also on the integrity of the vacuole as the major storage compartment for inorganic ions [3, 12]. In addition to the plasma membrane, also chloroplast membranes are susceptible to freezing damage. Linear electron transport is interrupted, which finally results in the inactivation of photosynthesis [13]. Therefore, chlorophyll a fluorescence measurements are a suitable tool to study freezing damage. Values of Fv/Fm determined with dark-adapted leaves reflect the potential quantum use efficiency of photosystem II (PSII) and have been widely used for assessing stress damage to the photosynthetic apparatus [14– 16]. The second protocol therefore describes the use of chlorophyll fluorescence imaging [17, 18] to quantify leaf freezing damage. Its use is advantageous over classical chlorophyll fluorescence measurements because it is not limited to single point measurements but allows integration of Fv/Fm over the whole leaf area. This turned out to be extremely important, as in Arabidopsis the basal leaf parts are damaged at milder freezing temperatures than the upper parts [12]. Here, we give a comprehensive description of experimental design, plant cultivation and freezing procedures, which are identical for both assays. From this point on, the freeze-thawed leaves or rosettes can be either used for electrolyte leakage measurement or subjected to chlorophyll fluorescence imaging.

2 2.1

Materials Equipment

1. Programmable cooling bath thermostats (CC-K20 with control system “Pilot One,” E-grade “Exclusive” and lid; Huber, Offenburg, Germany, or similar) with a large opening to allow for handling of three metal racks for 48 glass tubes each. If possible, the thermostats should be placed in a 4
C chamber to

Freezing Tolerance of Leaves and Small Rosettes

11

prevent overheating during cooling of the baths to low sub-zero temperatures. 2. Silicon oil (Thermal HY; Julabo, Seelbach, Germany, or similar) to fill the cooling baths. 3. Lab 960 Conductometer equipped with LF613T Conductivity electrode (SI Analytics, Mainz, Germany). 4. Water bath with the capacity for boiling. 5. Imaging Pam Chlorophyll Fluorometer M-series (Maxi version) with IMAG-MAX/L measuring head and IMAG-K6 CCD camera (the less sensitive IMAG-K7 is also possible to use and a new development is the IMAG-3D with a MAXI Imaging head expanded by a 3D scanner) (Walz, Effeltrich, Germany). 6. Bottletop dispenser, volume range 1–10 ml (optional). 7. Automatic pipette suitable for 300 μl volumes. 2.2

Software

1. GraphPad Prism version 6 to 8 (GraphPad, La Jolla, USA), SAS (SAS Institute, Cary, USA), or R. 2. ImagingWinGigE (Walz, Effeltrich, Germany).

2.3

Consumables

1. Glass tubes in metal racks (10 cm height, 1.5 cm diameter, 48 tubes per rack, 3 or 6 racks per experiment)—can be washed and reused. 2. ddH2O. 3. Metal lids for glass tubes—can be washed and reused. 4. Razor blades or scalpels. 5. Blunt end forceps. 6. Round bottom 15 ml falcon tubes without lid—can be washed and reused. 7. Standard microscopy glass slides.

3

Methods

3.1 Freezing Experiment [3, 19, 20] 3.1.1 Experimental Design

The use of two cooling baths in parallel will provide you with sufficient space to process 288 samples in one experiment. The freezing temperature course of each experiment should be carefully planned. For nonacclimated (6-week-old) leaves of the moderately freezing tolerant Arabidopsis thaliana accession Col-0 a range of 1 to 16
C and for cold acclimated leaves a range of 1 to 20
C is recommended. It is advisable to take samples in steps of 1 to 2
C and to condense the temperature interval to 0.5 to 1
C steps in the range of the expected LT50 value in order to reach an optimal resolution. To monitor the comparability between single

12

Anja Thalhammer et al.

experiments an internal control line (e.g., A. thaliana Col-0) should be used in each independent experiment. Each plant line in an experiment should be covered by four to five technical replicates, so using the setup given here will enable you to assess three to four independent plant lines in a single experiment. For the screening of a larger number of lines in one experiment younger (4-week-old) rosettes together with a reduced number of temperature steps (short temperature protocol—STP) can be used. These adapted settings are suitable to investigate seven lines together with the wild type in five replicates when two cooling baths are available. 3.1.2 Plants Growth

1. Seeds are sown in soil and vernalized in a phytotron for 1 week under cold-night conditions (12:12 day–night cycle, 20
C:6
C). After 1 additional week in short day conditions (8:16 h day–night cycle, 20
C:16
C), plants are pricked (3 plants per pot (Ø 10 cm)) and kept in short day conditions for 2 more weeks. Four weeks after sowing, plants are transferred to long day conditions (16:8 h day–night cycle, 20
C:18
C) with light supplementation to reach at least 200 μmol quanta/m2/s. Nonacclimated plants are thus used when 6 weeks old. When young whole rosettes are used with the STP protocol, more plants have to be pricked (7–10 plants per pot (Ø 10 cm)) and should be kept in short day conditions for only 1 week after pricking. After transfer to long day conditions for another week, 4-week-old plants can be used for the experiment [20]. Plants should be kept at short day conditions for the whole growth period if the cold chamber runs at short day conditions. 2. For cold acclimation, the same quantity of plants is transferred to a 4
C growth cabinet (16:8 h day–night cycle, 4
C) with 90 μmol quanta/m2/s for an additional 14 days [3]. Three days of cold treatment will be sufficient for 4-week-old plants [20]. The cold treatment will minimize developmental processes, so relative leaf age will be hardly changed between nonacclimated and cold acclimated leaves. Ideally, for the assay using detached leaves, plants should be in a developmental stage where the rosette is fully expanded, but the plants are not yet bolting. Avoid the use of already flowering plants, as they do not cold acclimate properly. Roughly, 40–45 plants will provide enough leaf material to cover all technical replicates of one plant line. For the STP protocol exactly 40 rosettes are needed for five replicates including the controls.

3.1.3 Freeze–Thaw Experiment

1. Prepare and label a sufficient number of glass tubes in metal racks and put them on ice. Additionally, prepare a rack containing tubes for unfrozen control leaves which will be kept on ice during the entire experiment.

Freezing Tolerance of Leaves and Small Rosettes

13

Fig. 1 Schematic overview of sample handling during the preparation for a freezing experiment. Stacking of three leaves from different leaf rosettes and cutting of petioles (a), insertion of leaf stacks into glass tubes containing water (b), and addition of small ice crystals to initiate ice nucleation in the leaves (c)

2. Fill 300 μl of ddH2O into each glass tube for freezing treatment and 600 μl of ddH2O into the control tubes. For STP 200 μl of ddH2O are sufficient to avoid covering the whole rosette with water. 3. Cut three to six rosettes and arrange the single fully expanded leaves in a way to generate stacks of three leaves stemming from different rosettes each (see Note 1). The total amount of leaf material should be about the same in each stack. Exclude apparently damaged or senescent leaves (see Note 2). For chlorophyll fluorescence experiments carefully remove soil crumbs from the leaves as these will block the fluorescence signal. Prepare a sufficient number of stacks to cover all temperature steps of one technical replicate and the control. Work quickly to avoid wilting of the leaves. Whole rosettes have to be cut for the STP protocol and soil has to be removed. The following step 4 is not necessary for this protocol, so continue directly with step 5. 4. Cut the petioles with a sharp razor blade to generate a common base of each stack (Fig. 1a). 5. Put each leaf stack or small rosette (STP) to the bottom of the appropriate test tube with a forceps so that the petioles or the basis of the rosettes are enclosed by water (Fig. 1b). Take care not to press or fissure the leaves. Put the tubes on ice immediately. 6. Prepare all replicates of all plant lines correspondingly. 7. Put lids on the tubes with the control samples and store them on ice for the duration of the experiment.

14

Anja Thalhammer et al.

8. Put the other tubes into the respective racks in the precooled cooling baths (1
C) and allow the samples to equilibrate to 1
C for 30 min. 9. Grind some ice (produced from ddH2O) with a mortar and pestle to obtain small ice crystals. 10. Carefully add a small ice crystal to the bottom of each tube with a small spatula (Fig. 1c) to initiate freezing, close the tubes with lids and incubate for another 30 min at 1
C to allow for ice nucleation in the leaves and temperature equilibration. 11. Start the program with a cooling rate of 4
C/h, take out respective samples at the desired temperatures and store them on ice. If needed, lower cooling rates, such as 2
C/h can be used as well, but higher cooling rates should be avoided, as sample temperature will then not be able to follow bath temperature. For the STP protocol with A. thaliana Col-0 use the following temperature steps, for nonacclimated 2
C, 4
C, 6
C, 8
C, 10
C, 12
C, 16
C; and for cold acclimated 4
C, 6
C, 8
C, 10
C, 12
C, 16
C, 20
C [20]. 12. Remove samples from the bath at the predetermined steps in the temperature protocol and transfer them immediately to an ice bath. Leave all samples on ice in the 4
C chamber overnight to thaw. Samples are then ready to process either for electrolyte leakage or chlorophyll fluorescence measurements. 3.2 Electrolyte Leakage [3, 11, 19] 3.2.1 Experimental Setup

1. Add 7 ml or 4 ml (STP) of ddH2O to each tube, including the control tubes, so that the leaves are completely immersed. If you are using large leaves it might be necessary to use more water. 2. Incubate the samples on a shaker at 150 rpm at 4
C for approximately 24 h (see Note 3). 3. Prepare metal racks with 15 ml falcon tubes in the same set-up as in your freezing experiment. Wash the tubes twice with deionized water before the experiment and let them dry. This is especially important for the STP protocol with small rosettes resulting in low conductivity levels. Fill 4.5 ml of ddH2O into each tube and add 1 ml of the respective sample after careful mixing with a pipette. 4. Insert the electrodes carefully into the sample tubes. Measure the electrical conductivity of each sample after mixing the solution thoroughly by moving the electrodes up and down for 10–12 times (see Note 4). Wait until a stable value is displayed (several seconds). Measure each sample twice and note the higher value. Before transferring the electrodes to the next sample, clean them by swaying in a beaker with ddH2O and tap them dry on a pile of paper towels (see Note 5).

Freezing Tolerance of Leaves and Small Rosettes

15

5. Empty the falcon tubes and rinse them 2–3 times with deionized water. 6. Incubate the glass tubes containing the leaves in a boiling water bath for 30 min for determination of the total electrolyte content. 7. Allow the samples to cool down to room temperature and repeat steps 1–4. Use the same tube setup as for the unboiled samples. Since conductivity is temperature dependent, it is important that unboiled and boiled samples are measured at the same temperature. 3.2.2 Data Analysis

1. Collect the data in a suitable data processing software (e.g., MS Excel). Calculate the percentage of electrolyte leakage (% EL) relative to the conductivity of the boiled samples: %EL ¼

conductivity ½unboiled sample  100 conductivity ½boiled sample

2. As the leakage of electrolytes in the control samples is not caused by freezing, normalize the % EL of each sample (% ELsample) to the average % EL of the control samples (% ELcontrol) within each replicate line. For a better comparability of the graphs, normalize your data to a maximum electrolyte leakage of 100%. For this purpose, use the % EL of the lowest freezing temperature (% ELmax). %ELnormalized


%ELsample  %ELcontrol ¼  100 ð%ELmax  %ELcontrol Þ

3. Further data analysis of the resulting sigmoidal curves to calculate LT50 values can be done either with GraphPad Prism or SAS (both commercially available) or with the freely available drc package in R. LT50 values of nonacclimated and coldacclimated plants (3 days) of two Arabidopsis thaliana as well as two Eutrema salsugineum accessions calculated with the three different methods are very similar (Table 1). GraphPad Prism: Import the % ELnormalized values into the GraphPad Prism software indicating the appropriate number of replicates. Analyze the data using nonlinear regression (curve fit) with a sigmoidal dose-response (variable slope). This will give you the LT50 value over all technical replicate lines as LogEC50 (Fig. 2) [3] (see Note 6). For control purposes we strongly recommend to compare the leakage curves of all single replicates of one line before calculating the combined LT50 value

16

Anja Thalhammer et al.

Table 1 LT50 values with standard errors (SE) of leaves from plants before (nonacclimated, na) and after 3 days of cold-acclimation (acc) of two Arabidopsis thaliana (Col-0, N14) as well as two Eutrema salsugineum (Tuva, Jiangsu) accessions calculated with three different software packages Col-0

N14

Tuva

Jiangsu

SE Col-0

SE N14

SE Tuva

SE Jiangsu

GraphPad Prism

5.36

5.73

4.83

4.57

0.11

0.07

0.07

0.08

drc package in R

5.33

5.71

4.77

4.52

0.12

0.10

0.08

0.08

SAS

5.34

5.73

4.81

4.55

0.11

0.08

0.07

0.08

GraphPad Prism

7.75

8.56

8.87

8

0.23

0.19

0.22

0.16

drc package in R

7.83

8.56

8.88

8.04

0.23

0.19

0.20

0.17

SAS

7.69

8.56

8.86

8.00

0.25

0.19

0.20

0.16

na

3d acc

from all replicates. If a leakage curve of a replicate does not have a sigmoidal pattern due to an outlier it should be excluded from the analysis. SAS: Import the %ELnormalized values into SAS. Analyze the data for all technical replicates by nonlinear regression analysis (PROC NLIN) specifying the following sigmoid function with programming statements: %EL ¼

ELmin þ ðELmax  ELmin Þ 1 þ e ðc ðdT ÞÞ

where ELmin is the minimum electrolyte leakage, ELmax is the maximum electrolyte leakage, c is the slope of the function at the inflection point d, and T is the treatment temperature. The temperature (d) at the inflection point is the LT50 value over all technical replicates. Estimating parameters in a nonlinear model in SAS is an iterative process that commences from starting values. Therefore, supplying the initial values of ELmin, ELmax, c and d for the NLIN procedure is required. These values can be estimated from the leakage curves. R: Import the %ELnormalized values into R. Analyze the data for all technical replicates using the four-parameter log-logistic function in the drc package [21]. The four-parameter logistic function is given by the formula: f ðx Þ ¼ c þ

dc

1þ

e ðb ð log ðx Þ log ðe ÞÞÞ

where d is the upper limit (maximum electrolyte leakage), c is the lower limit (minimum electrolyte leakage), x is the treatment

Freezing Tolerance of Leaves and Small Rosettes

17

Fig. 2 Electrolyte leakage from nonacclimated (na) and cold-acclimated (acc) A. thaliana Col-0 leaves frozen to either 19 (a) or eight (STP) (b) different temperatures. Curves were fitted using a logistic regression model and LT50 values were calculated as the temperatures at which 50% electrolyte leakage occurred using the GraphPad Prism software. Error bars represent SE from at least four replicate measurements, with each replicate including either three leaves from different plants (a) or one small rosette (b). LT50 na was 5.43
C and LT50 acc was 8.82
C after 2 weeks of cold acclimation (a). For the STP protocol, LT50 na was 5.34
C and LT50 acc was 7.88
C after 3 days of cold acclimation (b)

18

Anja Thalhammer et al.

temperature, e is the LT50 value over all technical replicates, and b denotes the relative slope around e. 3.3 Chlorophyll Fluorescence Imaging [12, 22] 3.3.1 Experimental Setup

1. Keep samples on ice during the whole experiment. 2. Dark-adapt samples, for example, by completely covering with a black sheet or cloth for at least 20 min in order to bring all PSII reaction centres into the open state. 3. Set up the measuring facilities in a dark room to keep the samples dark adapted during the measurement and start the ImagingWin GigE software for data acquisition. 4. Arrange one of the three leaves of one replicate sample carefully on a microscopy slide in the Live Video mode (NIR-measuring light pulses, Frequency 1 Hz) (see Note 7). Focus the leaf and fix the distance of the camera to the sample. This distance must be kept the same during all measurements. 5. Quit the Live Video mode by pressing Exit and press the Fo, Fm button to trigger a saturation pulse. This measurement will give you the variable (Fv) and maximal (Fm) chlorophyll a fluorescence of the leaf. 6. Select the Fv/Fm image, reflecting the maximal PSII quantum yield of a dark-adapted leaf. 7. Save the picture as PAM Image (PIM) file for later analysis, press the New Record button, and continue with the next sample.

3.3.2 Data Analysis

1. Use the ImageWin GigE software for data processing. You can extract the Fv/Fm pictures displayed in a false color scale representation as shown in Fig. 3. 2. In addition, you can quantify Fv/Fm values by integration over the whole leaf area or desired leaf sections. For this purpose, open the respective PIM file with the ImageWin GigE software and select Fv/Fm. By using the AOI (Area of Interest) routine, you can set user defined limits within which the average Fv/Fm value will be calculated. This will be displayed in the Report menu (see Note 8). 3. Determination of the LT50 values is done analogous to the electrolyte leakage procedure (see Note 9). Determine relative Fv/Fm (Fv/Fm rel) values by normalizing the Fv/Fm values of each replicate line (Fv/Fm sample) to that of the unfrozen control sample (Fv/Fm control) as well as to the Fv/Fm value of the sample exposed to the lowest freezing temperature (Fv/Fm max).

Freezing Tolerance of Leaves and Small Rosettes

19

Fig. 3 Chlorophyll fluorescence imaging of A. thaliana Col-0 leaves. Detached leaves of nonacclimated (na— upper panel) and cold acclimated (acc—lower panel) plants were frozen to different temperatures and thawed slowly. Maximum quantum yield of photosystem II (Fv/Fm) is shown in false color images (as specified by the scale bar) at the indicated freezing temperatures. The LT50 values calculated from the numeric Fv/Fm-values derived from logistic regression models using the GraphPad Prism software are indicated by arrows

 F v =F m

rel

¼

F v =F msample  F v =F mcontrol



ðF v =F mmax  F v =F mcontrol Þ

 100

4. Import the Fv/Fm rel values into one of the analysis software tools described above and continue as described for % ELnormalized values.

4

Notes 1. We found no significant difference in freezing tolerance between old and young fully expanded leaves from plants of the same age [3]. This may, however, be different for small, very young leaves. 2. You should avoid to work with leaves that are too long and stick out above the surface of the cooling fluid (silicon oil), as these parts will not reach the same temperatures as the lower, fully immersed parts. This will lead to a gross distortion of the LT50 values. If you have to work with such leaves, you need to cut them to a suitable length. 3. If you work with other plants than Arabidopsis, this may need to be modified, as thicker leaves may need longer incubation times to reach equilibrium with the surrounding water (see for example ref. 23).

20

Anja Thalhammer et al.

4. For a higher sample throughput it is recommended to use two programmable baths and two conductivity meters simultaneously. Measuring the conductivity can then be done in two racks in parallel. Make sure to use the same electrode on a respective sample before and after boiling to exclude systematic errors caused by differences in the electrodes. 5. From time to time the conductivity electrodes should be cleaned by washing in 1% SDS in a glass beaker under gentle stirring for 1.5 h. Be careful to fix the electrodes in a way that they will not slip into the beaker and get broken by the stirring bar. 6. For orientation purposes, it is helpful to know that in A. thaliana Col-0, the LT50 of nonacclimated leaves derived from the electrolyte leakage assay is about 5.5
C and for cold-acclimated leaves about 9.5
C [3, 6]. The LT50 of 4-week-old rosettes reaches values of about 5.3
C for nonacclimated and about 7.8
C for cold acclimated leaves [20]. These numbers reflect the ability of the plant to adapt to freezing temperatures by exposure to low but nonfreezing temperatures. 7. Measuring the chlorophyll fluorescence of all three leaves in a tube is not feasible with the number of samples generated in these experiments. However, the measurement of one leaf is sufficient to provide stable results, when you stick to the suggested number of four to five technical replicates per sample. However, if one leaf is providing an obviously unreliable picture, you should assess another leaf from the same tube. Do not measure the same leaf twice, since the saturating pulse will close all reaction centres of PSII. A second measurement will therefore not provide the Fv/Fm anymore. 8. In healthy, nonstressed leaves Fv/Fm usually has a value of around 0.83 [24] and declines with increasing stress levels. 9. Although there is a tight linear correlation between LT50 values from both assays, those derived from chlorophyll fluorescence imaging are generally lower than those from electrolyte leakage measurements. This is mainly due to secondary damage resulting from the preincubation of samples for electrolyte leakage measurements in distilled water. However, biological reasons in terms of a higher freezing tolerance of photosynthetic membranes in comparison to the plasma membrane cannot be completely excluded [12]. References 1. Steponkus PL (1984) Role of the plasma membrane in freezing injury and cold acclimation. Annu Rev Plant Physiol 35:543–584

2. Thalhammer A, Bryant G, Sulpice R et al (2014) Disordered cold regulated15 proteins protect chloroplast membranes during freezing

Freezing Tolerance of Leaves and Small Rosettes through binding and folding, but do not stabilize chloroplast enzymes in vivo. Plant Physiol 166:190–201 3. Rohde P, Hincha DK, Heyer AG (2004) Heterosis in the freezing tolerance of crosses between two Arabidopsis thaliana accessions (Columbia-0 and C24) that show differences in non-acclimated and acclimated freezing tolerance. Plant J 38:790–799 4. Hannah MA, Wiese D, Freund S et al (2006) Natural genetic variation of freezing tolerance in Arabidopsis. Plant Physiol 142:98–112 5. Korn M, Peterek S, Mock HP et al (2008) Heterosis in the freezing tolerance, and sugar and flavonoid contents of crosses between Arabidopsis thaliana accessions of widely varying freezing tolerance. Plant Cell Environ 31:813–827 6. Zuther E, Schulz E, Childs LH et al (2012) Clinal variation in the non-acclimated and cold-acclimated freezing tolerance of Arabidopsis thaliana accessions. Plant Cell Environ 35:1860–1878 7. Zuther E, Juszczak I, Lee YP et al (2015) Time-dependent deacclimation after cold acclimation in Arabidopsis thaliana accessions. Sci Rep 5:12199 8. Le MQ, Pagter M, Hincha DK (2015) Global changes in gene expression, assayed by microarray hybridization and quantitative RT-PCR, during acclimation of three Arabidopsis thaliana accessions to sub-zero temperatures after cold acclimation. Plant Mol Biol 87:1–15 9. Takahashi D, Gorka M, Erban A et al (2019) Both cold and sub-zero acclimation induce cell wall modification and changes in the extracellular proteome in Arabidopsis thaliana. Sci Rep 9:2289 10. Lee YP, Babakov A, de Boer B et al (2012) Comparison of freezing tolerance, compatible solutes and polyamines in geographically diverse collections of Thellungiella sp. and Arabidopsis thaliana accessions. BMC Plant Biol 12:131 11. Ristic Z, Ashworth EN (1993) Changes in leaf ultrastructure and carbohydrates in Arabidopsis thaliana L. (Heyn) cv. Columbia during rapid cold acclimation. Protoplasma 172:111–123

21

12. Ehlert B, Hincha DK (2008) Chlorophyll fluorescence imaging accurately quantifies freezing damage and cold acclimation responses in Arabidopsis leaves. Plant Methods 4:12 13. Krause GH, Grafflage S, Rumich-Bayer S et al (1988) Effects of freezing on plant mesophyll cells. Symp Soc Exp Biol 42:311–327 14. Woo NS, Badger MR, Pogson BJ (2008) A rapid, non-invasive procedure for quantitative assessment of drought survival using chlorophyll fluorescence. Plant Methods 4:27 15. Maxwell K, Johnson GN (2000) Chlorophyll fluorescence-a practical guide. J Exp Bot 51:659–668 16. Lawson T, Vialet-Chabrand S (2018) Chlorophyll fluorescence imaging. Methods Mol Biol 1770:121–140 17. Oxborough K (2004) Imaging of chlorophyll a fluorescence: theoretical and practical aspects of an emerging technique for the monitoring of photosynthetic performance. J Exp Bot 55:1195–1205 18. Lichtenthaler HK, Miehe´ JA (1997) Fluorescence imaging as a diagnostic tool for plant stress. Trend Plant Sci 2:316–320 19. McKhann HI, Gery C, Be´rard A et al (2008) Natural variation in CBF gene sequence, gene expression and freezing tolerance in the Versailles core collection of Arabidopsis thaliana. BMC Plant Biol 8:105 20. Zuther E, Schaarschmidt S, Fischer A et al (2019) Molecular signatures associated with increased freezing tolerance due to low temperature memory in Arabidopsis. Plant Cell Environ 42:854–873 21. Ritz C, Streibig JC (2005) Bioassay analysis using R. J Stat Software 12:1–22 22. Schreiber U, Bilger W (1987) Rapid assessment of stress effects on plant leaves by chlorophyll fluorescence measurements. In: Plant response to stress. Springer, Berlin, pp 27–53 23. Hincha DK, Pfu¨ller U, Schmitt JM (1997) The concentration of cryoprotective lectins in mistletoe (Viscum album L.) leaves is correlated with leaf frost hardiness. Planta 203:140–144 24. Hunt S (2003) Measurements of photosynthesis and respiration in plants. Physiol Plant 117:314–325

Chapter 3 Differential Thermal Analysis: A Fast Alternative to Frost Tolerance Measurements Andrey V. Malyshev, Ilka Beil, and Juergen Kreyling Abstract Frost tolerance is an important factor influencing plant growth, plant species distribution and competitive balance among plant species in the face of climate change. Traditional methods for estimating frost tolerance are often time consuming and require a large sample size, limiting the temporal and spatial resolutions. Differential thermal analysis (DTA) can be advantageous compared to other methods used to determine frost tolerance, most importantly by (1) increasing the number of tested species, tissue types and sampling dates, (2) allowing to test frost tolerance in situ, and (3) more realistically testing the influence of freezing rate and duration. Here, we discuss a typical procedure for DTA, compare its use to other frost tolerance methods and point out its limitations. Key words DTA, Freezing tolerance, Plant organ, Lethal temperature, Cell death, Exothermic reaction, Intracellular freezing

1

Introduction Plant frost tolerance continues to be important in the context of global climate change, because it is expected to increase temperature variability [1]. Highly fluctuating temperatures can disrupt plant cold acclimation, making plants less able to withstand unexpected freezing events, which are predicted to occur at a decreased frequency [2]. Cold acclimation (also known as cold hardening) is a suite of changes in gene expression and physiology that increase plant freezing tolerance [3, 4]. In fall, a reduced photoperiod and declining temperatures trigger cold acclimation in perennial plants. Species-specific cold acclimation rates, maximum frost tolerance during cold spells and deacclimation under warm conditions can influence plant survival and growth performance in a future climate with potentially far-reaching consequences such as reduced primary productivity and changes in species composition [5, 6]. Frost susceptibility also varies substantially within species. An ecotype is a plant population that displays genetic differentiation

Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_3, © Springer Science+Business Media, LLC, part of Springer Nature 2020

23

24

Andrey V. Malyshev et al.

and morphological and/or physiological characteristics attributable to particular environmental parameters within the ecotype’s habitat [7]. Local adaptation creates ecotypes that may differ as much in their acclimation and freezing tolerance as different plant species [8]. It is therefore often necessary to test a larger number of ecotypes, spanning a species’ distributional range, often over a temporal range as well. For such tests, fast methods to quantify frost tolerance using minimal plant material are necessary which makes DTA attractive. Here, we describe the methodology of DTA using various plant organs and how it can be used to track among- and within-species differences in frost tolerance. Limitations of the method are also presented, as well as guidelines on how to fix common methodological issues. The guidelines are intended to be general, and to apply to most vascular plants. The actual experimental design needs will likely require modification based on the specific species used, taking into account factors such as plant organ morphology and variation in frost tolerance among plant individuals.

2

Using DTA to Determine Plant Freezing Tolerance

2.1 Advantages and Suitability of DTA

Methods to determine frost tolerance can be divided into two broad categories; those subjecting whole organisms to frost and those using only a part of an organism. Freezing whole plants gives the most realistic frost tolerance estimates but is extremely limited in its application, as the number of plants needed is large and plant size is restricted to the size of the freezing chamber/room. Therefore, most methods to determine frost tolerance use parts of plants. Most of these methods involve preparing replicates of the desired plant organs (leaf buds, leaves, twigs, or roots) and freezing them at controlled rates. At set freezing temperatures, samples are taken out of freezing chambers, thawed and analyzed for frost damage. A temperature at which 50% of the cells/samples die, referred to as lethal temperature (LT50), is then used to rank species and ecotypes with respect to their frost tolerance (see Chapter 2). Electrolyte leakage measurements, visual assessment and chlorophyll fluorescence measurements are the most common methods used to quantify frost damage when set freezing temperatures are used [9]. Such methods share common drawbacks, however: the number of samples needed is large (at least 3 samples per temperature  7–10 different temperatures), tests are time-consuming (2–7 days) and the LT50 estimation is only accurate if enough set freezing temperatures are used below and above the LT50, being dependent on accurate estimation of LT50 temperature prior to testing. DTA bypasses many of the mentioned drawbacks by detecting frost damage in real time, enabling determination of LT50 within hours. Alternatives such as nuclear magnetic resonance and

Frost Tolerance via Differential Thermal Analysis

25

Fig. 1 Decision tree to determine if the use of DTA is appropriate

cryomicroscopy are more complex, laborious and expensive, being better suited for studying the underlying mechanisms of frost damage at cellular and subcellular resolution [9]. Despite its advantages in terms of time and sample size economy, there are cases where DTA may not work. Figure 1 shows a decision tree which aids in determining if using the DTA method is likely to be useful.

26

Andrey V. Malyshev et al.

How DTA Works

DTA is a calorimetric method which measures temperature changes in living tissue. A temperature sensor is attached to or inserted into a particular plant tissue and temperature changes during freezing are recorded. All plants can supercool, which means water can remain unfrozen in their tissues below 0  C. Plants capable of deep supercooling to much lower freezing temperatures (approx. – 40  C) have advanced frost tolerance mechanisms, allowing them to prevent intracellular water from freezing. Mechanisms of deep supercooling include eliminating ice nucleating centers and production of antifreeze proteins [10]. It should be noted that most frost tolerant plants (down to 196  C) [11], with a distribution that reaches the subarctic, do not utilise deep supercooling but are able to withstand most intracellular water exiting cells and freezing extracellularly [12]. Such freezing-induced cell desiccation tolerance cannot be tested with DTA, which otherwise works for most temperate species that show deep supercooling. Water freezing inside cells is lethal as without a fluid internal environment the functioning of the living protoplasm ceases [13]. Distinct temperature increases are observed as samples are frozen: the high temperature exotherm (HTE) and the low temperature exotherm (LTE) [14]. HTE occurs as water freezes extracellularly, such as apoplastic water in the xylem [15]. In plants where intracellular water supercools to a lower temperature than extracellular water, HTE is followed by the LTE, with LTE indicating the temperature at which frost damage occurs in the tissue. Sometimes the HTE is the only temperature increase that occurs, because intracellular water does not supercool to a lower temperature than extracellular water. In such cases, freezing occurs simultaneously in the intracellular and extracellular compartments and indicates the LT50. The median of LTE across several samples is used to accurately estimate LT50. DTA is most easily implemented and most robust when used in plant tissues which only have the HTE, as a larger amount of water freezes in this case, resulting in an easily recognizable sudden temperature increase.

2.3 How to Perform DTA

1. Plant organ selection: After choosing the target plant species, the plant organ to be tested for frost tolerance needs to be carefully selected in the context of the research question as freezing tolerance of different organs can vary by more than 5  C [16]. Additionally, the same organs can have different frost tolerance at different developmental stages at a particular time. Frost tolerance of expanding leaves on a single tree can vary by up to approximately 8  C depending on whether a leaf is beginning to open or is fully expanded [17]. If realistic frost tolerance of whole plants is to be determined, pilot trials are needed to establish a correlation between LT50 values of specific organs determined via DTA and other organs, which are most limiting to plant survival. If a comparison among ecotypes

2.2

Frost Tolerance via Differential Thermal Analysis

27

with respect to frost tolerance is most important, relative frost tolerance of the same organs (e.g., buds, leaves, or twigs) across species is used. In this case, it is suggested to use the organ with the highest water content, providing exotherms which are easier to detect. In addition, if changes in freezing tolerance across time are to be investigated, an organ needs to be selected which will provide sufficient material to be repeatedly destructively sampled. For each sampling date, a minimum of four samples is recommended for a robust estimate per plant, with six to ten being ideal, depending on the size of plants and variability within an individual. The physical size of each sample needs to be kept as constant as possible across species as sample size does influence freezing tolerance, with larger samples being more likely to have more nucleation and hence lower frost tolerance [18]. 2. Setup of a DTA test: Miniature Type T Thermocouples (e.g., Omega 5SRTC-TT-TI-36-1M) and Thermocouple Data Acquisition Modules (e.g., Omega TC-08) need to be acquired, with the number needed depending on the number of samples to be tested simultaneously. The solder junction of each thermocouple is attached to the sample. Various attachment methods are possible: thermocouples can be inserted into slits cut into woody tissues or taped to leaf surfaces with thermally conductive adhesive tape. Leaves can be folded or wrapped around thermocouples. Temperature changes in each sample are recorded and logged, for example, with freely downloadable software (e.g., PicoLog 6 software; https:// www.picotech.com/downloads). Samples are often wrapped in aluminum and foil inserted into holes drilled into an aluminum block, which ensures a more uniform temperature across samples, reducing temperature variability among sensors. Alternatively, test tube racks can be used (Fig. 2). For approximately every five to ten samples a reference thermocouple should be attached to a dead sample of the same organ, to be used as a reference temperature profile. The reference sample is usually killed by boiling the sample prior to the test. Temperature differences between the tested sample and the reference are plotted against the ambient air temperature. It is also possible to attach the thermocouples to plants in-situ, in which case appropriate battery and rain protection is required. More information on such a set up can be found in [18]. Samples are then cooled at a uniform rate in a climate chamber (or outside temperature is simply recorded). Temperature is recorded every 6–10 s. Unrealistic freezing rates faster than 4  C/h are not recommended, as they can cause increased cell damage due insufficient time for water to migrate to favorable freezing sites [9]. Sample surfaces need to be uniformly

28

Andrey V. Malyshev et al.

Fig. 2 Examples of possible DTA tissue temperature profiles. Temperature differences between the live and dead samples are plotted as a function of changes in ambient air temperature inside a freezing chamber. Plants A, B, and C have LT50 temperatures of 22  C, 12  C, and 7  C. Air temperatures down to 30  C are sufficient to determine LT50, provided that no extremely frost tolerant (e.g., subarctic) species are tested which survive much colder temperatures due to intracellular desiccation. Note that no LTE was observed for plants B and C, as only shallow supercooling took place, causing simultaneous freezing of extracellular and intracellular water (only HTE, no LTE)

moistened across samples as moistened and dry samples differ in the timing of HTE, with higher moisture content resulting in earlier freezing [19]. If an additional test is to be run in parallel, such as pausing at each freezing temperature to take samples out for visual inspection, the time spent at each temperature step should be constant (at least 20 min for temperature stabilization). Freezing is continued down to a temperature where all samples have experienced a low

Frost Tolerance via Differential Thermal Analysis

29

Fig. 3 Method comparison of visual and DTA frost tolerance tests. Colored temperature profiles are temperature changes detected by thermocouples attached to leaves of Syringa vulgaris. Photos show plants that had been taken out of the chamber for visual assessment (checked 5 days later) at the respective temperatures. DTA estimated the lethal temperatures for three leaves as 6.8  C, 6.5  C, and 5.9  C for leaves 1, 2, and 3, respectively. The corresponding average of these LT50 values ( 6.4  C) corresponded well with visual test results (LT50 around 6  C)

temperature exotherm or until the minimum temperature possible with respect to the climate chamber is reached. 3. DTA verification analysis: A pilot test should always be performed for every new species and organ to be tested to establish how well HTE and LTE can be detected with the specific sample preparation technique. Ideally, DTA should be performed once in concordance with another method to determine frost tolerance to verify the accuracy of the acquired LT50 values. Simultaneous to DTA, a separate sample set is removed at predefined freezing temperatures and inspected for damage by visual inspection, electrolyte leakage, or chlorophyll fluorescence analysis, to determine LT50 with an alternative method. Figure 3 shows how LT50, determined via a visual test, corresponds to the HTE detected via DTA. As shown in Fig. 3, the temperature of lethal tissue damage is estimated as the difference between the beginning and end of the temperature increase of the HTE (only HTE, no LTE in this species at this time), seen as a distinct peak within tissue temperature changes. It is useful to test attaching thermocouples to different parts of an organ/plant sample to check if the withinsample frost tolerance variability warrants attaching multiple

30

Andrey V. Malyshev et al.

thermocouples to each sample. Figure 4 shows an example of thermocouples attached to two adjacent leaves differing in frost tolerance. 4. Issues may arise where the exact lethal temperature of a sample is not detectable either due to high fluctuation in the recorded temperatures or due to device malfunction. Most common sources of error and/or causes of malfunction include suboptimal connection of the sensors to the sample and insufficient intracellular water content of the investigated tissue. Troubleshooting tips are summarized in Table 1.

Fig. 4 An example of thermocouple sensors being connected to Syringa vulgaris leaves with adhesive tape

Table 1 Common issues causing difficulty in detecting the exact lethal temperature of a sample Issue

Likely cause

Solution

High temperature oscillations make detecting temperature exotherms difficult

Thermocouple is not securely attached to tissue

Insert thermocouple into tissue slits, fold tissue over thermocouple

Solder junction breaks in thermocouple

Thermocouple was attached to Solder wires together in the junction sample too strongly

Temperature exotherms are too Too little water available in small to be reliably detected. sample to freeze.

Increase size of sample (e.g., longer/ thicker twigs)

There is very high variability in HTE among samples

Sample surface differs in moisture content

Moisten samples evenly and quickly wrap samples in aluminum

LTE not detected (HTE only)

Too little intracellular water available/species does not have LTE

Test frost tolerance with another method; if lower than HTE, DTA not possible.

Thermocouples give unrealistic Short circuiting, interrupted values ( 200  C, +200  C) connection between pc and logger or broken often with very high thermocouple fluctuation

Avoid connection between the conducting end of the thermocouple and metal block or aluminum foil.

Dead control samples also show Samples are not sufficiently exotherms isolated from each other

Wrap samples in additional tape and/or aluminum foil and place in holes of aluminum blocks with greater separation between samples.

Frost Tolerance via Differential Thermal Analysis

31

References 1. Augspurger CK (2013) Reconstructing patterns of temperature, phenology, and frost damage over 124 years: spring damage risk is increasing. Ecology 94:41–50 2. Kodra E, Steinhaeuser K, Ganguly AR (2011) Persisting cold extremes under 21st-century warming scenarios. Geophys Res Lett 38: L08705 3. Kalberer SR, Wisniewski M, Arora R (2006) Deacclimation and reacclimation of coldhardy plants: current understanding and emerging concepts. Plant Sci 171:3–16 4. Thomashow MF (1999) Plant cold acclimation: freezing tolerance genes and regulatory mechanisms. Annu Rev Plant Physiol Plant Mol Biol 50:571–599 5. Richardson AD, Black A, Delbart N et al (2010) Influence of spring and autumn phenological transitions on forest ecosystem productivity. Phil Trans R Soc B Biol Sci 365:3227–3246 6. Kindermann J, Wu¨rth G, Kohlmaier GH et al (1996) Interannual variation of carbon exchange fluxes in terrestrial ecosystems. Global Biogeochem Cycles 10:737–755 7. Turrill WB (1946) The ecotype concept: a consideration with appreciation and criticism, especially on recent trends. New Phytol 45:34–43 8. Malyshev AV, Henry HAL, Bolte A et al (2018) Temporal photoperiod sensitivity and forcing requirements for budburst in temperate tree seedlings. Agric For Meteorol 248:82–90 9. Burke MJ, Gusta LV, Quamme HA et al (1976) Freezing and injury in plants. Annu Rev Plant Physiol 27:507–528 10. Kuwabara C, Wang D, Endoh K et al (2013) Analysis of supercooling activity of tanninrelated polyphenols. Cryobiology 67:40–49

11. Sakai A (1966) Studies of frost hardiness in woody plants. II. Effect of temperature on hardening. Plant Physiol 41:353–359 12. Lindstrom OM, Anisko T, Dirr MA (2019) Low-temperature exotherms and cold hardiness in three taxa of deciduous trees. J Am Soc Hortic Sci 120:830–834 13. Kuroda K, Ohtani J, Fujikawa S (1997) Supercooling of xylem ray parenchyma cells in tropical and subtropical hardwood species. Trees 12:97–106 14. von Fircks HA (1993) Exothermic responses of dormant Salix stems during exposure to subzero temperatures. Physiol Plant 87:271–278 15. Aslamarz AA, Vahdati K, Rahemi M et al (2010) Supercooling and cold-hardiness of acclimated and deacclimated buds and stems of persian walnut cultivars and selections. HortScience 45:1662–1667 16. Sakai A, Larcher W (1987) Frost survival of plants: responses and adaptation to freezing stress. Springer-Verlag, Berlin, 321 pp 17. Taschler D, Beikircher B, Neuner G (2004) Frost resistance and ice nucleation in leaves of five woody timberline species measured in situ during shoot expansion. Tree Physiol 24:331–337 18. Pramsohler M, Hacker J, Neuner G et al (2012) Freezing pattern and frost killing temperature of apple (Malus domestica) wood under controlled conditions and in nature. Tree Physiol 32:819–828 19. Ishikawa M, Oda A, Fukami R et al (2015) Factors contributing to deep supercooling capability and cold survival in dwarf bamboo (Sasa senanensis) leaf blades. Front Plant Sci 5:791

Chapter 4 Infrared Thermal Analysis of Plant Freezing Processes Gilbert Neuner and Edith Lichtenberger Abstract Infrared thermal analysis is an invaluable technique to study the plant freezing process. In the differential mode, infrared thermal analysis (IDTA) allows to localize ice nucleation and ice propagation in whole plants or plant samples at the tissue level. Ice barriers can be visualized and supercooling of cells, tissues, and organs can be monitored. Places where ice masses are accommodated in the apoplast can be identified. Here, we describe an experimental setting developed in our laboratory, give detailed information on the practical procedure and preconditions, and give additionally an idea of the problems that would be encountered and how they may be overcome. Key words Deep supercooling, Freezing stress, Frost, Ice barriers, Ice nucleation, Ice propagation

1

Introduction While symplastic (intracellular) ice formation always has lethal consequences for plant cells, apoplastic (extracellular) ice formation can safely be survived down to a certain freezing temperature. Where ice forms, how ice propagates and is accommodated within the apoplast but also how ice formation can be avoided in certain cells, tissues, and organs are essential aspects of plant freezing tolerance and are all important factors that affect the ability of a plant to survive freezing. These traits may even be as important as the ability to withstand the dehydration stresses associated with ice formation [1]. As far as known, the response to the presence of ice within a plant’s tissue can be complex and quite diverse [2]. A deeper understanding of the control of apoplastic ice formation processes will lead to new strategies and technologies for improving plant freezing tolerance [2]. Below 0  C without ice nucleation water remains in the supercooled state. Once ice has nucleated, it readily propagates at high rates of up to 27 cm/s [3–5] throughout all plant parts that are not protected by an ice barrier. A precondition for ice nucleation is the formation of an active ice crystal nucleus. An ice crystal nucleus is a

Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_4, © Springer Science+Business Media, LLC, part of Springer Nature 2020

33

34

Gilbert Neuner and Edith Lichtenberger

cluster of regularly ordered water molecules. The cluster size is temperature dependent. With decreasing temperature the critical size decreases successively, increasing the probability of ice nucleation [6, 7]. There are principally two ways how ice crystal nuclei can form: (1) autonomously, that is, homogeneous ice nucleation, or (2) water molecules are forced to form an ice crystal nucleus by an ice nucleation active substance (heterogeneous ice nucleation). Apoplastic ice nucleation is considered to occur heterogeneously, symplastic ice nucleation very likely is of homogeneous nature, as it can be close to the homogeneous ice nucleation temperature of water at 38.5  C [6, 7]. When crystallization of water occurs below 0  C, liberation of heat (enthalpy of fusion) produces the so called freezing exotherm, that is, a sudden rise in sample temperature. Depending on the amount of water that is freezing, this can be several K (e.g., 6K in Rhododendron ferrugineum leaf [8]) or can be close to the thermometric resolution limit of current instrumentation, as for example observed during the so-called low temperature freezing exotherms produced by symplastic freezing of deeply supercooled cells (e.g., xylem parenchyma cells [8]). Measurement of freezing exotherms can principally be conducted punctually by the use of thermocouples. By differential thermal analysis (DTA [9]) very small freezing events become detectable at low noise, as the temperature of a dry and dead reference sample is subtracted from the measured living plant sample. More sophisticated is the employment of high-resolution infrared thermography as a two-dimensional thermal image of the plant sample can be obtained, which allows to localize ice nucleation sites and to monitor ice propagation throughout the plant [10– 17]. Again, by using the infrared camera in the differential imaging mode (IDTA, Infrared differential thermal analysis), the resolution of the thermal images obtained during the freezing process can be significantly increased [3–5, 18–26]. IDTA is based on the subtraction of a reference image, captured just before the occurrence of freezing, from the sequence of images during freezing that then show only the changes in temperature during freezing of water in the plant sample [3]. By this process, background temperature fluctuations and thermal gradients in the images are canceled out and even small changes in temperature caused by freezing of water in the plant tissue can be visualized.

2

Materials For measurement of IDTA during the freezing process, a high resolution infrared camera is necessary that is suitable for operation at temperatures in the subzero temperature range. Additionally, a temperature controlled freezing chamber large enough to

Infrared Thermal Analysis

35

accommodate the infrared camera together with the plant sample is necessary. Otherwise, the camera can also be positioned outside of the freezer, given that monitoring of samples is possible through a hole in the lid of the freezing chamber that fits to the close-up lense of the camera. In the following, the measurement procedure as developed in our laboratory in Innsbruck is described.

2.1

Infrared Camera

A digital infrared camera model ThermaCAM S60 (FLIR Systems AB, Danderyd, Sweden) is employed for measurement of sequences of two-dimensional infrared images of plant samples. The thermal resolution lies between 0.08 and 0.12 K. The camera is equipped with a close-up lens (LW64/150). In this way, a spatial resolution of the thermal images of 200 μm is achieved. The infrared camera is connected by a FireWire/IEEE1394 interface with a control computer. The ThermaCAM S60 has a maximum time resolution of 25 images/s. Recording at maximum time resolution results in large amounts of data (3.8 MB/s, 13.7 GB/h). For the original infrared video records a powerful PC providing sufficiently high disk space is necessary. IDTA images can be extracted by subtraction of a reference image from the original infrared image which is performed during data analysis with the ThermaCAM Researcher software package (FLIR Systems AB). Absolute tissue temperature is recorded with thermocouples (see Note 1) placed close to the surface of the analyzed plant samples. Thermocouples are connected to a data logger (CR10X, Campbell Scientific, Loughborough, UK).

2.2 Thermally Insulated Camera Housing

When used inside the freezing compartment, the lowest freezing temperature that can be studied is currently set by the minimum operation temperature of the ThermaCAM, which is 25  C. In order to protect the infrared camera from fast temperature changes and potential water condensation (see Note 2), we use a thermally insulated housing for the camera. The housing is made of a 21  25  31 cm Plexiglas box which is thermally insulated on the inside with 3 cm thick Styrofoam plates. The box has a hole that fits to the close-up lens of the infrared camera, and a lid on top to insert and remove the infrared camera. A second hole opposite to the opening for the close-up lens allows for inserting FireWire connection and electric cables.

2.3 Temperature Controlled Freezing Chamber

For freezing treatment of the plant samples, a sufficiently spacious fully temperature controlled freezing chamber is necessary (see Note 3). This is realized in Innsbruck by a computer-controlled commercial chest freezer [27]. The freezing compartment has a volume of 141.9 L (43  50  66 cm). The freezing device needs to be fully temperature controlled and should allow a precise setting of cooling and thawing rates (see Note 4). The control technique

36

Gilbert Neuner and Edith Lichtenberger

should keep temperature oscillations at a minimum, preferentially less than 0.2 K (see Note 5). 2.4

3

Sample Holder

The entire plant sample investigated must be in a focus plane to ensure that ice formation can be monitored with high resolution. The sample holder plate (approximately 15  15 cm) should be impermeable for infrared radiation and provide a homogenous thermal background. Usually an ordinary plastic plate is sufficient. For convenient distance adjustments this plate may be placed on a laboratory lifting plate. The actual measurement area of the infrared camera equipped with the close-up lens is approximately 3.7  4.6 cm.

Methods

3.1 Sample Preparation

1. In case that detached plant parts have to be used, it must be considered that the sample size has strong effects on ice nucleation temperature. The size of detached plant samples should not be too small (see Note 6). 2. For identification of action sites of intrinsic ice nucleation active substances, plant samples should have a dry surface. Water on the plant surface can freeze extrinsically at first and can then potentially trigger consequent intrinsic freezing of the sample [28]. 3. For control of ice nucleation temperature the use of INA bacteria (see Note 7) is recommended. 4. Plant samples must be mounted on the sample holder in such a way that all plant parts investigated are exposed in the focal plane of the infrared camera, particularly when the infrared camera is used with the close-up lens. For small plant parts such as single leaves this is not problematic (Fig. 1a). For leafy twigs or whole herbaceous plants this can be done by fanning out the plant samples on the sample holder plate (see Note 8). 5. For recording of the absolute temperature it is advisable to use a set of at least four thermocouples as a reference and additional temperature control. The thermocouples are fixed onto the sample holder close to the investigated plant samples (see Note 1) with double sided adhesive tape and additionally outside the measurement frame by further pieces of adhesive tape to securely hold them in place (see Note 9). 6. The sample holder plate, with the mounted plant samples and thermocouples on top of the laboratory lifting plate, is then placed at the bottom of the freezing compartment.

Infrared Thermal Analysis

37

Fig. 1 (a) On the digital image of a single, detached leaf of Senecio incanus a 50 μL droplet of INA bacteria suspension positioned on the upper leaf surface can be seen (horizontal black bar width is 0.5 cm). Ice nucleated initially in the droplet on the surface at 4.3  C but this surface ice did not enter the leaf. Freezing is visualized by a brightening, while unfrozen areas remain black. At 4.5  C a second ice nucleation event in the petiole of the leaf initiated an ice wave via the vascular system into the leaf blade. The original sequence of infrared images during this second freezing event. (b–g) gives only a blurred picture. When these infrared images are referenced to the image immediately before the leaf freezing event (h), IDTA images (i–n) are obtained that show more details of the freezing process. The numbers in the bottom left corner of each image indicate the time in seconds after ice nucleation in the petiole of the leaf 3.2 Data Acquisition and Analysis

1. The infrared camera is inserted into the camera housing and the lid is closed. The infrared camera is then turned on to record the live image. 2. The camera in the housing is then put upside down into the freezing compartment and is positioned in a distance of approximately 9 cm to the plant sample, using two additional lifting plates. 3. To bring the target area into focus, fine adjustments of the distance between the close-up lens of the infrared camera and the plant sample can be made by the use of the lifting plate supporting the sample holder plate. 4. A single digital image of the sample setting should be recorded by the infrared camera. 5. The infrared camera is switched into the infrared mode. The recording conditions (time resolution, storage space for

38

Gilbert Neuner and Edith Lichtenberger

recordings) are set (see Note 10). The lid of the freezing compartment is closed. 6. The parameters of the freezing treatment have to be set. The experimental settings, particularly the cooling rate, strongly affect ice nucleation temperature and should be selected to come close to natural night frost conditions (see Notes 3 and 4). After setting all parameters the freezing treatment can be started. 7. As soon as 0  C is reached, the recording mode of the infrared camera is manually started (see Note 11). 8. After the programmed freezing treatment is finished the infrared recording is stopped and the lid of the freezing compartment is opened and the whole equipment is allowed to thaw to room temperature. 9. The recorded video sequences of infrared images are scanned by the use of the ThermaCAM Researcher software package to elucidate video sequences that show freezing events. On the original infrared images freezing processes can be detected but they are often blurred (Fig. 1b–g). Once identified, IDTA images can be extracted. IDTA is based on the subtraction of a reference image, captured just before the occurrence of freezing (Fig. 1h), from the sequence of images during freezing that then show only the changes in temperatures during freezing of water in the plant sample, revealing much more details of the ice formation processes (Fig. 1i–n). For an unambiguous assignment of freezing processes to certain tissues an overlay of the digital image and the IDTA images [24] can be done by After Effects (Adobe Systems Inc., San Jose, CA, USA).

4

Notes 1. Thermocouples themselves can be a source of ice nucleation [9]. In case ice nucleation sites in the plant sample should be identified it is advisable to avoid close contact of thermocouple solder junctions to the sample. 2. The housing ensures that the infrared camera is kept at a higher temperature than the surroundings. This prevents water condensation on the close-up lens which could lead to erroneous measurements. 3. Currently commercially available freezing devices are all convective systems. Convective freezing devices cannot simulate radiant heat loss and do not produce temperature gradients in plants during freezing as in nature [29]. As plants are not colder than air in these devices, dew and ice deposits on the

Infrared Thermal Analysis

39

Fig. 2 (a) Digital image of an infructescence of Primula veris before exposure to a controlled freezing treatment (horizontal black bar width is 0.5 cm). Initial ice nucleation in the supporting stem occurred at 3.2  C. Single fruits and developing seeds inside the fruits froze separately at various lower freezing temperatures: IDTA images of the infructescence show freezing of (b) two single fruits at 10.7  C and (c, d) separate freezing of developing seeds inside these fruits at around 12.0  C (white arrows). Freezing is visualized by a brightening, while unfrozen areas remain black. The time span after ice nucleation in the supporting stem is indicated in the bottom left corner of each image

plant surface are absent. These conditions favor supercooling which should be kept in mind. 4. Moderate cooling rates below 0  C should be selected in order to simulate natural night frosts. In nature, cooling rates below 0  C often are not faster than 2 K/h [26, 29]. Higher cooling rates tend to provoke supercooling in the sample. Additionally, if frost survival is investigated, exposure times and thawing rates have to be controlled. For plant samples with unknown frost survival mechanisms it is advisable to use a freezing protocol that eventually extends down to freezing temperatures where, after the initial apoplastic ice wave, additional freezing events at lower temperatures may be recorded. These are either further apoplastic freezing events occurring later due to ice barriers (Fig. 2) or are freezing exotherms that originate from symplastic freezing. 5. Higher temperature oscillations can be buffered by placing the sample holder and the camera inside of a thermally insulated box inside the freezing compartment. 6. Sample size [20] and detachment [29] significantly influence the ice nucleation temperature. In general, the smaller the sample size, the lower temperature samples tend to supercool down to. For apple twigs only if sized 40 cm in length ice nucleation temperatures recorded in the laboratory on detached twigs were roughly equivalent to those measured in the field on intact trees [20]. Several plant species have structural ice barriers between different organs [8], and in some cases their role may only be studied in whole plants. Thermal ice barriers as found in alpine cushion plants [19] can only be measured on whole intact plants in as natural experimental settings as possible. 7. INA bacteria (e.g., strain 5176 of Pseudomonas syringae) have a high ice nucleating activity. They can be obtained for example

40

Gilbert Neuner and Edith Lichtenberger

from the DSMZ Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Braunschweig, Germany and can be applied as a bacterial suspension in 50 μL droplets (see Fig. 1). Depending on the question investigated, droplets can be deposited either on intact, scratched or cut plant surfaces. Likewise, the leaf petiole or cut stem surfaces can be wrapped in moist cotton wool that is soaked with a suspension of INA bacteria [26]. 8. For fanning out plant samples double sided adhesive tape (5 cm width) has proven to be most convenient. 9. For documentation of thermocouple positioning it is advisable to take a photo of the experimental setting. 10. A measurement frequency of 10 images/s is usually sufficient. When the maximum of 25 images/s is recorded this results in a large amount of data (3.8 MB/s, 13.7 GB/h). Sufficient disk space on the computer must be provided. 11. During the recording it is advisable to regularly (at least every 30 min) conduct an image matching. For easier handling of the video material during subsequent analysis it is recommended to produce several consecutive short videos rather than one single long video. References 1. Wisniewski ME, Gusta LV, Fuller MP et al (2009) Ice nucleation, propagation and deep supercooling: the lost tribes of freezing studies. In: Gusta LV, Wisniewski ME, Tanino KK (eds) Plant cold hardiness: from the laboratory to the field. CAB International, Cambridge, pp 1–11 2. Gusta LV, Wisniewski ME, Trischuk RG (2009) Patterns of freezing in plants: the influence of species, environment and experimental procedures. In: Gusta LV, Wisniewski ME, Tanino KK (eds) Plant Cold Hardiness: from the laboratory to the field. CAB International, Cambridge, pp 214–223 3. Hacker J, Neuner G (2007) Ice propagation in plants visualized at the tissue level by infrared differential thermal analysis (IDTA). Tree Physiol 27:1661–1670 4. Hacker J, Neuner G (2008) Ice propagation in dehardened alpine plant species studied by infrared differential thermal analysis (IDTA). Arc Antarc Alp Res 40:660–670 5. Hacker J, Spindelbo¨ck J, Neuner G (2008) Mesophyll freezing and effects of freeze dehydration visualized by simultaneous measurement of IDTA and differential imaging chlorophyll fluorescence. Plant Cell Environ 31:1725–1733

6. Franks F (1985) Biophysics and biochemistry at low temperatures. Cambridge University Press, Cambridge 7. Chen S-H, Mallamace F, Mou C-Y et al (2006) The violation of the Stokes-Einstein relation in supercooled water. Proc Natl Acad Sci U S A 103:12974–12978 8. Sakai A, Larcher W (1987) Frost survival of plants. Responses and adaptation to freezing stress. In: Billings WD, Golley F, Lange OL, Olson JS, Remmert H (eds) Ecological studies, vol. 62. Springer, Berlin 9. Burke MJ, Gusta LV, Quamme HA et al (1976) Freezing and injury in plants. Annu Rev Plant Physiol Plant Mol Biol 27:507–528 10. Le Grice P, Fuller MP, Campbell A (1993) An investigation of the potential use of thermal imaging technology in the study of frost damage to sensitive crops. In: Proceedings of the international conference on biological ice nucleation, university of wyoming, Laramie, WY, USA, 4–6 Aug 1993 11. Ceccardi TL, Heath RL, Ting IP (1995) Low-temperature exotherm measurement using infrared thermography. HortSci 30:140–142 12. Wisniewski M, Lindow SE, Ashworth EN (1997) Observations of ice nucleation and

Infrared Thermal Analysis propagation in plants using infrared video thermography. Plant Physiol 113:327–334 13. Lutze JL, Roden JS, Holly CJ et al (1998) Elevated atmospheric [CO2] promotes frost damage in evergreen tree seedlings. Plant Cell Environ 21:631–635 14. Wisniewski M, Fuller M (1999) Ice nucleation and deep supercooling in plants: new insights using infrared thermography. In: Margesin R, Schinner F (eds) Cold adapted organisms. Ecology, physiology, enzymology and molecular biology. Springer, Berlin, pp 105–118 15. Pearce RS, Fuller MP (2001) Freezing of barley studied by infrared video thermography. Plant Physiol 125:227–240 16. Ball MC, Wolfe J, Canny M et al (2002) Space and time dependence of temperature and freezing in evergreen leaves. Funct Plant Biol 29:1259–1272 17. Sekozawa Y, Sugaya S, Gemma H (2004) Observations of ice nucleation and propagation in flowers of Japanese Pear (Pyrus pyrifolia Nakai) using infrared video thermography. J Jpn Soc Hort Sci 73:1–6 18. Neuner G, Xu BC, Hacker J (2010) Velocity and pattern of ice propagation and deep supercooling in woody stems of Castanea sativa, Morus nigra and Quercus robur measured by IDTA. Tree Physiol 30:1037–1045 19. Hacker J, Ladinig U, Wagner J et al (2011) Inflorescences of alpine cushion plants freeze autonomously and may survive subzero temperatures by supercooling. Plant Sci 180:149–156 20. Pramsohler M, Hacker J, Neuner G (2012) Freezing pattern and frost killing temperature of apple (Malus domestica) wood under controlled conditions and in nature. Tree Physiol 32:819–828

41

21. Kuprian E, Briceno V, Wagner J et al (2014) Ice barriers promote supercooling and prevent frost injury in reproductive buds, flowers and fruits of alpine dwarf shrubs throughout the summer. Environ Exp Bot 106:4–12 22. Charrier G, Pramsohler M, Charra-Vaskou K et al (2015) Ultrasonic emissions during ice nucleation and propagation in plant xylem. New Phytol 207:570–578 23. Kuprian E, Tuong TD, Pfaller K et al (2016) Persistent supercooling of reproductive shoots is enabled by structural ice barriers being active despite an intact xylem connection. PLoS One 11:e0163160 24. Kuprian E, Munkler C, Resnyak A et al (2017) Complex bud architecture and cell-specific chemical patterns enable supercooling of Picea abies bud primordia. Plant Cell Environ 40:3101–3112 25. Kuprian E, Munkler C, Resnyak A et al (2018) Does winter dehydration account for seasonal increase in supercooling ability of Norway spruce bud primordia? Tree Physiol 38:591–601 26. Stegner M, Sch€afernolte T, Neuner G (2019) New insights in potato leaf freezing by infrared thermography. Appl Sci 9:819 27. Neuner G, Buchner O (1999) Assessment of foliar frost damage: a comparison of in vivo chlorophyll fluorescence with other viability tests. J Appl Bot 73:50–54 28. Pearce RS (2001) Plant freezing and damage. Ann Bot 87:417–424 29. Neuner G, Hacker J (2012) Ice formation and propagation in alpine plants. In: Lu¨tz C (ed) Plants in alpine regions: cell physiology of adaptation and survival strategies. Springer, Wien, pp 163–174

Chapter 5 Conducting Field Trials for Frost Tolerance Breeding in Cereals Luigi Cattivelli and Cristina Crosatti Abstract Cereal species can be damaged by frost either during winter or at flowering stage. Frost tolerance per se is only a part of the mechanisms that allow plants to survive during winter, while winter-hardiness also considers other biotic or physical stresses that challenge the plants during the winter season, limiting their survival rate. While frost tolerance can also be tested in controlled environments, winter-hardiness can only be determined with field evaluations. Post-heading frost damage occurs from radiation frost events in spring during the reproductive stages. A reliable evaluation of winter-hardiness or of post heading frost damage should be carried out with field trials replicated across years and locations to overcome the irregular occurrence of natural conditions which satisfactorily differentiate genotypes. The evaluation of postheading frost damage requires a specific attention to plant phenology. The extent of frost damage is traditionally determined with a visual score at the end of the winter, although, recently an image-based phenotyping coupled with unmanned aerial vehicles (UAVs) has been proposed. Key words Barley, Frost tolerance, Post-heading frost damage, Vernalization, Wheat, Winterhardiness

1

Introduction Cereal species can be damaged by frost either during winter or at flowering stage. The tolerance of winter cereals to low temperatures depends on the physiological process known as hardening or cold acclimation that occurs when plants are exposed to temperatures ranging from 0 to 5  C prior to winter freezing. There is large genetic variation for the ability to survive freezing temperatures among the cereal species, with winter-habit rye cultivars having the best freezing tolerance followed by hexaploid winter wheat and winter barley and oat [1]. Nevertheless, it should be noticed that rye, wheat, barley, and oat genotypes are all capable to cold acclimate, to some extent, in response to low temperatures. Frost tolerance is intimately connected with vernalization. The cereal genotypes have traditionally been classified into three main

Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_5, © Springer Science+Business Media, LLC, part of Springer Nature 2020

43

44

Luigi Cattivelli and Cristina Crosatti

groups: spring types, which pass to the reproductive phase quickly, without vernalization and even in short days; winter types, which display a strong vernalization requirement and sensitivity to short days; intermediate or alternative (also called facultative) types which flower quickly in long days but in which floral induction is more or less inhibited by short days. In the light of the knowledge achieved by molecular genetics three major vernalization loci, Vrn-1, Vrn-2, and Vrn-3, have been identified as the determinants of the vernalization response [2]. Since no allelic variation at the Vrn-3 locus was observed within the cultivated germplasm, a two genes epistatic model was proposed in barley [3] as well as in wheat [4]. A higher level of frost tolerance is generally associated with the winter growth habit. Nevertheless, some studies have also reported a high level of frost tolerance in facultative genotypes without vernalization response [5, 6]. Plants with facultative growth habit are more ready to react to changes in environmental factors (light intensity, temperature) assuring flowering under a wide range of climatic conditions, a trait associated with a high adaptation capacity. Although generally less resistant than winter genotypes, a significant variation for frost tolerance has been detected also within spring genotypes [7]. The general association between winter habit and frost tolerance is explained by the genetic linkage between theVrn-1 locus and the two loci controlling frost tolerance, Frost Resistance-1 (Fr-1) and Fr-2 all located on the long arm of chromosome 5A in wheat [8, 9] and 5H in barley [10]. Fr-1 has a pleiotropic effect of Vrn-1 (or it cosegregates with Vrn-1), while Fr-2 maps about 30cM proximal from Vrn-1/Fr-1. Fr-2 contains a cluster of CBF genes, a family of cold inducible transcription factors known to control the expression of a large part of the cold-regulated genes. Allelic variations at the Fr-1/CBF locus are known to modify cold acclimation capacity and, in turn, frost tolerance [11]. Frost tolerance per se is only a part of the mechanisms that allow plants to survive during winter and to synchronize their life cycle with the seasonal cycle. From an agricultural (and economical) point of view, winter hardiness (or winter survival) is a more relevant and broader concept than frost tolerance, although frost tolerance often represents the main factor for winter survival. Winter hardiness considers the plant within the whole ecosystem, where other organisms (e.g., pathogens specifically adapted to low temperature) or physical conditions (e.g., anoxia, limited soil fertility) can challenge the plants, thus limiting their survival rate [12]. Winter hardiness can be determined only with field evaluations. Barley and wheat crops can also experience frost damage at the reproductive stage. Post-heading frost damage is a main problem in southern Australia where barley and wheat are planted in autumn with the majority of the growing season over winter. Winter is usually mild and winter frost damage is virtually absent. The predominant frost damage occurs from radiation frost events in spring

Frost Tolerance Field Trials

45

during the reproductive stages. Radiation frosts occur under clear night skies, where more heat is radiated away from the crop canopy than it receives. The loss of radiant energy causes the temperature to fall, which can damage sensitive reproductive tissues at subzero temperatures. These frost events can cause floret and spike abortion as well as damage to the developing grain, which can have a significant impact on yield and quality [13–15].

2 2.1

Methods Winter Hardiness

The intensity and the frequency of frost events during winter are unpredictable and vary significantly depending on locations and years, limiting the effectiveness of winter hardiness field trials (see Note 1). A reliable evaluation of winter hardiness is carried out with field trials replicated across years and locations. In each location, a randomized design with three or more replicates is recommended. Small plots (2–3 m2) are usually sufficient to evaluate winter hardiness. The sowing date, being determinant for growth stage, plays a decisive role in frost tolerance and winter hardiness [16]. Late sowing limits the plant development before frost events and does not allow the complete deployment of the plant acclimation potential. Therefore, the application of different sowing dates might reveal differences in the frost tolerance of the genotypes under evaluation. No-till management practice has been shown to improve winter survival since the stubble leftover can help maintain snow cover, which insulates the plants from the cold [17]. Overall, a complete design to test winter hardiness considers three or more locations in different climatic regions, in two or more years and, if possible, two sowing dates with a replicated field design. A weather station in the proximity of the field trial is used to record the temperature throughout the winter season. The extent of winter damage is usually assessed at the end of the winter by visual scoring. A frequently used scoring system is based on a 0–9 scale [16, 18] with: l

0: no damage.

l

1: slightly yellowed leaf tips.

l

2: half yellowed basal leaves.

l

3: fully yellowed basal leaves.

l

4: whole plants slightly yellowed.

l

5: whole plants yellowed and some plants withered.

l

6: whole plants yellowed and 10% plant mortality.

l

7: whole plants yellowed and 20% plant mortality.

l

8: whole plants yellowed and 50% plant mortality

l

9: all plants killed.

46

Luigi Cattivelli and Cristina Crosatti

Some examples of field experiments for the evaluation of frost tolerance and of frost damage are given in Figs. 1, 2, 3, and 4. Visual rating, although easy to apply, is subject to bias and human error which may lead to a reduction in the precision and the accuracy of the data. To overcome these limitations and make an objective evaluation, the application of image-based methods and vegetation indices [19, 20] have been recently proposed [21] to evaluate the status of the vegetation in the field after winter that,

Fig. 1 Barley field trial after winter. The cultivar on the left has no damage (score 0), while the cultivar on the right shows all plants completely yellowed, but with no dead plants (score 5)

Fig. 2 A close-up of a barley plant with clear symptoms of frost damage on the older leaves, but with the crown still alive

Frost Tolerance Field Trials

47

Fig. 3 Barley field trial after a cold winter. Some cultivars are completely killed by frost (an example is given in Fig. 4), while other cultivars show extended leaf damage, but limited plant mortality. Border plots (dark green) are sown with bread wheat, a cereal species with a higher frost tolerance than barley

Fig. 4 A barley genotype showing a high level of frost damage, where only a few plants are still alive after winter (score 9)

48

Luigi Cattivelli and Cristina Crosatti

in turn, reflects the extent of winter damage. Images can be captured at the ground level by modified vehicles or via unmanned aerial vehicles (UAV) equipped with sensors. Obviously, the use of UAV does not require access to the field and has a much greater processivity. The protocol to score for winter hardiness using image-based methods requires two evaluations of the vegetation in the field, the first one at the beginning of the winter after plant establishment to set the baseline and the second after snow melting at the end of the frost period. Multispectral cameras can be used to capture spectral reflectance at specific wavelengths (blue 480 nm, green 560 nm, red 670 nm, red edge 720 nm, and near-infrared 840 nm) to generate georeferenced orthomosaics of the flight for each wavelength. Orthomosaic images of respective bands are used to generate a normalized difference vegetation index (NDVI) map [22]. Variation in light conditions can be adjusted by using an image of a calibration panel with known reflectance. NDVI considers the greenness of the vegetation within each experimental unit and the comparison between the NDVI before and after winter allows to estimate the winter damage. As an alternative way to measure the amount of green vegetation, the fractional green canopy cover can be estimated using a normal camera and software for image processing [23] that classifies pixels to vegetation and nonvegetation based on the ratio of red/green, ratio of blue/green, excess green index, and pixel continuity [21]. A specific Winter Survival Index was proposed by Chen et al. [21] to estimate winter survival avoiding the confounding effect of poor emergence: (NDVIbefore – NDVIafter)/NDVIafter. In environments where snow is frequent, a long-lasting snow cover can interfere significantly with the evaluation of winter hardiness (see Note 2). Under these conditions, the field evaluation can be integrated with a parallel experiment, where plants are grown in boxes in an open-air space and protected from snow with a shelter [5]. Under the shelter the plants are exposed to natural temperature variations, but without snow cover. A comparison between shelter-protected and field-grown plants allows the estimation of the snow and other winter-related stress factors on winter hardiness. The progress of cold acclimation during winter in field grown plants can be monitored taking leaf or crown samples from some representative plants and using them for standard frost evaluation tests, such as the assessment of electrolyte leakage or chlorophyll fluorescence (photosystem II maximal quantum yield, Fv/Fm) after freezing under controlled conditions (see Chapter 2 for details). It has been demonstrated that cereal leaves cut from field-grown plants are a viable system to test frost tolerance using Fv/Fm, given that the photosynthetic machinery maintains its full activity for several days when leaves are kept in the dark or under low light at low temperature [24].

Frost Tolerance Field Trials

49

A useful detail to achieve a perfect evaluation of frost damage is represented by a set of plots grown in the same field as those exposed to natural cold condition, but excluded from frost, and therefore acting as control. An effective modular and mobile plotheater, able to prevent frost damage at plot level, while not impacting microclimate or yield, has been designed by Stutsel et al. [15]. 2.2 Post-heading Frost Damage

3

As for winter hardiness, the evaluation of post-heading frost damage requires a multilocations field trial design to overcome the unpredictability of late frost events. Furthermore, the evaluation of post-heading frost damage has a high risk of escape due to differences in plant phenology and morphology, and spatial temperature variations [25]. To control for plant phenology two main strategies are used: (1) test genotypes with similar heading time; (2) use different sowing dates to compare plants at a similar phenological stage deriving from different sowings (for additional techniques see Note 3). Typically, an experimental design for the evaluation of post-heading frost damage is run on several locations and with two to four sowing dates. Usually, early sowing encourages early flowering during the period of highest frost risk. Different sowing dates are required to allow for maturity differences between genotypes and to collect data from multiple frost events during the same season. Small plots are used to give the maximum number of genotypes in a small area to reduce the effect of spatial temperature variation. Thermometers distributed in the field are used to monitor spatial temperature differences. After each frost event, a number of tillers at the same developmental stage are tagged to allow a comparison at similar developmental stages. Frost-induced sterility is assessed 10–20 days later for each spike and is expressed as percentage of total florets [13]. Additional frostinduced grain damage is scored at maturity.

Notes 1. Field evaluation of frost tolerance has been the first and simplest method used to select for frost tolerant genotypes. Quite often, the irregular occurrence of natural conditions which satisfactorily differentiate genotypes results in large experimental errors and complicates the detection of small but meaningful differences among cultivars. This limitation has prompted the development of a number methods for the assessment of frost tolerance under controlled conditions. Nevertheless, field evaluation offers the unique opportunity to assess the overall winter hardiness capacity, that is more than a simple evaluation of frost tolerance measured under controlled temperature conditions in a growth chamber. To overcome the intrinsic limitations of a field evaluation of winter hardiness, replication of the

50

Luigi Cattivelli and Cristina Crosatti

experiments across locations and years must always be considered. A fundamental question concerns what kind of frost tolerant plants we will need in the future, when the foreseen global climate changes will generally increase the temperature, reducing the frequency and the extent of harsh winters [12]. At first sight, global warming may reduce frost damage in crops, but this is not likely to happen. A consequence of climate change is the fluctuation of winter temperatures, with winter warm spells becoming more frequent than in the past [26, 27]. This will have a strong impact on the frost tolerance both in crops and natural flora. Vernalization is generally saturated or partially saturated during the first month of the winter. As a consequence, during a winter warm spell plants start active growth, thereby losing most of their freezing tolerance (deacclimation). This condition leads to the exposition of not hardened plants to subsequent frost events. Since frost tolerance is intimately associated with a reduction in plant growth [28], it is unlikely that actively growing plants retain their hardening capacity. Therefore, the ability to rapidly change the physiological conditions in response to a fluctuation of temperature might become more important than absolute frost tolerance capacity. To test the adaptation of cereal species and genotypes to the new winter climate under natural conditions will be an essential aspect of cereal breeding for temperate and cold regions in the coming years. Therefore, although in the last decades the evaluation of winter hardiness has often been substituted with tests for frost tolerance in controlled conditions, the field evaluation will remain and will acquire even more relevance in the future, since there will be a need to breed new varieties for the new climatic conditions. 2. In deep snow regions, the plants survive winter under a longlasting snow cover. Snow provides a protection from deep frost keeping soil temperature at crown level between 0 and 10  C despite the very low air temperature. Prolonged snow cover prevents photosynthesis, reduces plant metabolism and exposes the plants to psychrophilic pathogenic fungi known as snow mould fungi [29] as well as to some risks of anoxic stress when the snow melts [30]. The most relevant snow mould fungi are snow scald (Sclerotinia borealis), speckled snow mold (Typhula ishikariensis), and pink snow mold (Microdochium nivale) [29]. When snow mould infections occur during an experiment for the evaluation of winter hardiness, it should be considered that the performance of the plants will be mainly determined by their susceptibility to snow mould, which, to some extent, is determined by the cold acclimation state of the plants [31]. 3. A different method for the evaluation of post-heading frost damage has recently been proposed. The application of

Frost Tolerance Field Trials

51

supplementary artificial light in the field, aiming to modify the natural photoperiod, together with a specific field design allows the generation of a “photoperiod gradient.” Thereby, same genotypes, depending on the light conditions, can be stimulated to flower at different times. This will allow for the screening of genotypes with different phenology under natural field frost conditions at matched developmental stages [32]. References 1. Fowler DB, Carles RJ (1979) Growth, development, and cold tolerance of fall-acclimated cereal grains. Crop Sci 19:915–922 2. Trevaskis B, Hemming MN, Dennis ES et al (2007) The molecular basis of vernalizationinduced flowering in cereals. Trends Plant Sci 12:352–357 3. von Zitzewitz J, Szucs P, Dubcovsky J et al (2005) Molecular and structural characterization of barley vernalization genes. Plant Mol Biol 59:449–467 4. Yan L, Loukoianov A, Tranquilli G et al (2003) Positional cloning of the wheat vernalization gene VRN1. Proc Natl Acad Sci U S A 100:6263–6268 5. Rizza F, Pagani D, Gut M et al (2011) Diversity in the response to low temperature in a set of representative barley genotypes cultivated in Europe. Crop Sci 51:2759–2779 6. von Zitzewitz J, Cuesta-Marcos A, Condon F et al (2011) The genetics of winter-hardiness in barley: perspectives from genome-wide association mapping. Plant Genome 4:76–91 7. Tondelli A, Francia E, Visioni A et al (2014) QTLs for barley yield adaptation to Mediterranean environments in the ‘Nure’‘Tremois’ biparental population. Euphytica 197:73–86 8. Va´gu´jfalvi A, Galiba G, Cattivelli L et al (2003) The cold regulated transcriptional activator Cbf3 is linked to the frost-tolerance gene Fr-A2 on wheat chromosome 5A. Mol Gen Genomics 269:60–67 9. Va´gu´jfalvi A, Aprile A, Miller A et al (2005) The expression of several Cbf genes at the Fr-A2 locus is linked to frost resistance in wheat. Mol Gen Genomics 274:506–514 10. Francia E, Rizza F, Cattivelli L et al (2004) Two loci on chromosome 5H determine low temperature tolerance in the new ‘winter’ x ‘spring’ (‘Nure’ x ‘Tremois’) barley map. Theor Appl Genet 108:670–680 11. Tondelli A, Barabaschi D, Francia E et al (2011) Inside the CBF locus in Gramineae. Plant Sci 180:39–45

12. Cattivelli L (2011) More cold tolerant plants in a warmer world. Plant Sci 180:1–2 13. Reinheimer JL, Barr AR, Eglinton JK (2004) QTL mapping of chromosomal regions conferring reproductive frost tolerance in barley (Hordeum vulgare L.). Theor Appl Genet 109:1267–1274 14. Fuller MP, Fuller AM, Kaniouras S et al (2007) The freezing characteristics of wheat at ear emergence. Eur J Agron 26:435–441 15. Stutsel BM, Callow JN, Flower K et al (2019) An automated plot heater for field frost research in cereals. Agronomy 9:96 16. Crosatti C, Pagani D, Cattivelli L et al (2008) Effects of the growth stage and hardening conditions on the association between frost resistance and the expression of the cold induced protein COR14b in barley. Environ Exp Bot 62:93–100 17. Fowler DB (2012) Wheat production in the high winter stress climate of the Great Plains of North America—an experiment in crop adaptation. Crop Sci 52:11–20 18. Rizza F, Crosatti C, Stanca AM et al (1994) Studies for assessing the influence of hardening on cold tolerance of barley genotypes. Euphytica 75:131–138 19. Grieder C, Hund A, Walter A (2015) Image based phenotyping during winter: a powerful tool to assess wheat genetic variation in growth response to temperature. Funct Plant Biol 42:387–396 20. Humplı´k JF, Laza´r D, Husicˇkova´ A et al (2015) Automated phenotyping of plants shoots using imaging methods for analysis of plant stress responses—a review. Plant Methods 11:29 21. Chen Y, Sidhu HS, Kaviani M et al (2019) Application of image-based phenotyping tools to identify QTL for in-field winter survival of winter wheat (Triticum aestivum L). Theor App Genet 132:2591–2604 22. Dvorak V, Selbeck J, Dammer KH et al (2013) Strategy for the development of a smart NDVI camera system for outdoor plant detection and

52

Luigi Cattivelli and Cristina Crosatti

agricultural embedded systems. Sensors 13:1523–1538 23. Patrignani A, Ochsner TE (2015) Canopeo: a powerful new tool for measuring fractional green canopy cover. Agron J 107:2312–2320 24. Badeck FW, Rizza F (2015) A combined field/ laboratory method for assessment of frost tolerance with freezing tests and chlorophyll fluorescence. Agronomy 5:71–88 25. Frederiks TM, Christopher JT, Sutherland MW et al (2015) Post-head-emergence frost in wheat and barley: defining the problem, assessing the damage, and identifying resistance. J Exp Bot 66:3487–3498 26. Beniston M (2005) Warm winter spells in the Swiss Alps: strong heat waves in a cold season? A study focusing on climate observations at the Saentis high mountain site. Geophys Res Lett 32:L01812 27. Shabbar A, Bonasal B (2003) An assessment of changes in winter cold and warm spells over Canada. Nat Hazards 29:173–188 28. Achard P, Gong F, Cheminant S et al (2008) The cold-inducible CBF1 factor-dependent

signaling pathway modulates the accumulation of the growth-repressing DELLA proteins via its effect on gibberellin metabolism. Plant Cell 20:2117–2129 29. Gaudet DA (1994) Progress towards understanding interaction between cold hardiness and snow mould resistance and development of resistant cultivars. Can J Plant Pathol 16:241–246 30. Andrews CJ, Pomeroy MK (1979) Toxicity of anaerobic metabolites accumulating in winter wheat seedlings during ice encasement. Plant Physiol 64:120–125 31. Gaudet DA, Wang Y, Frick M et al (2011) Low temperature induced defence gene expression in winter wheat in relation to resistance to snow moulds and other wheat diseases. Plant Sci 180:99–110 32. Frederiks TM, Christopher JT, Harvey GL et al (2012) Current and emerging screening methods to identify posthead-emergence frost adaptation in wheat and barley. J Exp Bot 63:5405–5416

Chapter 6 A Whole-Plant Screening Test to Select Freezing-Tolerant and Low-Dormant Genotypes Annick Bertrand, Annie Claessens, Jose´e Bourassa, Solen Rocher, and Vern S. Baron Abstract Winter survival is a determinant factor for the persistence of perennials grown in northern climates. High winter survival cultivars, however, have lower yield due to their early transition into a dormant state in the fall. Here we describe a whole plant assay entirely performed indoor in growth chambers and walk-in freezers to identify low-dormant genotypes with superior freezing tolerance within populations of open pollinated species. Three successive freezing stresses are applied to a broad base of 3000 genotypes to progressively eliminate 97% of the population and to retain only the 3% best performing genotypes. This approach can be used to generate recurrently selected populations in different species. Key words Freezing tolerance, Dormancy, Recurrent selection, Controlled conditions, Freezing stress

1

Introduction Winter survival is a determinant factor for the persistence of perennials grown in northern climate. High winter survival cultivars, however, have lower yield due to their early transition into a dormant state in the fall. Breeding for low fall dormancy is an efficient approach to increase the annual yield of perennials by extending their growing seasons from late summer through early winter. While the phenotypic correlation between winter survival and fall dormancy is high [1], the low genotypic correlation between these two traits suggests that they can be improved simultaneously [2]. Winter survival and fall dormancy are quantitative traits with low to moderate heritability. Success in breeding freezing-tolerant plants using conventional plant breeding methodologies has been limited in spite of the presence of large genetic diversity for this trait within populations of open-pollinated species [3]. Improvement of plant winter hardiness has historically been based on field selection of genotypes that survive winters [4] while selection for reduced fall

Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_6, © Springer Science+Business Media, LLC, part of Springer Nature 2020

53

54

Annick Bertrand et al.

dormancy is based on the height of genotypes in the fall [5]. However, the unpredictability of test winters and of fall conditions due to large variations between and within locations and the environmental conditions to which the plants are exposed severely limits the predictability of these approaches [6]. As a result, costly assessment of the genetic material at multiple locations over many years is often used to minimize environmental effects and to increase the likelihood to accurately discriminate plants with regard to their winter hardiness and dormancy potentials. New approaches are needed by breeding programs to accelerate and reduce the cost of assessment of freezing tolerance and fall dormancy. In addition, the efficiency of gene discovery studies is highly determined by the availability of assays for high throughput and reliable screening of plant phenotypes. To address issues associated with the uncertainty and resource intensive assessment of winter hardiness in the field and based on cumulative evidence indicating that tolerance to low freezing temperatures plays a central role with regard to adaptation to winter [7], we devised a method of selection entirely performed indoor under controlled conditions. Using that approach, large numbers of genotypes from initial genetic backgrounds are subjected to successive freezing stresses to progressively eliminate freezing sensitive plants and ascertain the phenotype of the hardiest plants. To simultaneously select for freezing tolerance and reduced dormancy, the regrowth conditions of the plants after each freezing stress is set to photoperiod and temperature conditions allowing to discriminate between dormant (short) and nondormant (tall) genotypes within a population. At the completion of the selection process, plants comprising the 3% superior genotypes can be intercrossed to generate a new population putatively improved for its tolerance to freezing (TF) and reduced dormancy (RD). Up to now, we successfully applied this approach in alfalfa [8]. Other perennial species such as red clover [9] and perennial ryegrass [10] were improved for freezing tolerance using a similar approach [11].

2

Materials Start with open pollinated species with characterized extensive genetic variability to avoid genetic bottlenecks eventually leading to lower plasticity for adaptation and restricted fitness.

2.1

Plant Material

1. Seeds: 3000 viable seeds of an initial genetic background ready for germination (see Notes 1 and 2). 2. Soil: potting soil plus slow release fertilizer (14-14-14, 300 mL per 100 L of soil; Osmocote, Scotts, Marysville, OH, USA). 3. Fertilizer: 20-20-20 + micronutrients (1 g/L).

A Screening Test for Freezing Tolerance and Low Dormancy

55

4. Pots: the pot system consists of 3000 individual cells arranged in trays: 164 mL-volume Ray Leach Cone-tainers (SC10 Super Cell, low density) and RL98 trays as well as IPL Rigi-pots IP110 (Stuewe and Sons, Inc., Tangent, OR, USA). 5. Plastic labels. 2.2 Controlled Environment Chambers and Programmable Freezer

3

1. Controlled environment chambers: growth chambers with a temperature range from 2 to 25  C and irradiance from 150 to 600 μmol photons/m2/s PPFD, and controlled photoperiod. A surface of 13.38 m2 is required for seeding one population. 2. Large walk-in programmable freezer with a temperature range from 2 to 30  C (see Note 3).

Methods

3.1 Seeding and Growth

1. Fill the pots (Cone-tainers), which have been previously placed in IP110 Rigi-Pots, with uniformly humid soil. Slightly compact the soil up to 1.5 cm below the top. 2. Cover the seeds with 0.5 cm of sieved soil, pack the soil to ensure a good contact between seeds and soil, and water the pots uniformly by hand. 3. Identify each IP110 tray with a plastic label. 4. Place trays in growth chambers under the following environmental conditions: 16 h photoperiod with an irradiance of 400–600 μmol/m2/s PPFD and a day–night temperature regime of 22:17  C. 5. Water the plants when needed and fertilize twice a week for the first 3 weeks of growth (see Note 4) with a 1 g/L solution of a commercial fertilizer (20-20-20 plus micronutrients). 6. After 4 weeks of growth, transfer the plants to cold acclimation conditions.

3.2

Cold Acclimation

1. Place trays in growth chambers set to the following environmental conditions: 8 h photoperiod (200–250 μmol/m2/ s PPFD) under a constant temperature of 2  C. 2. Water the plants when needed. Do not fertilize. 3. After 2 weeks of acclimation at 2  C, transfer the tubes containing the plants in RL98 racks leaving a free row between each plant row. This pattern will facilitate even temperature distribution in each pot (see Note 5). 4. Transfer the plants to a programmable freezer set at 2  C after an adequate irrigation of the pots (see Note 6). 5. After the soil is well frozen (1–2 days), cover the racks with tarpaulins to avoid plant desiccation.

56

Annick Bertrand et al.

6. Acclimate the plants at nonlethal freezing temperature ( 2  C) in the dark for 2 weeks (see Note 7). 7. Remove the tarpaulin the day before application of freezing stress. 3.3 Application of Freezing Stresses (Fig. 1)

1. After 2 weeks at 2  C, progressively decrease the temperature in the freezer to the expected lethal temperature for 50% of the plants (LT50) using the following stepwise decrease: decremental steps of 2  C during a 30 min period followed by a 90-min plateau at each temperature (see Note 8). 2. When the expected LT50 is reached, withdraw the plants from the freezer after a 90-min exposure.

Fig. 1 Schematic illustration of the procedure used for the selection of populations with improved freezing tolerance (TF) and reduced dormancy (RD)

A Screening Test for Freezing Tolerance and Low Dormancy

57

Fig. 2 Regrowth of alfalfa genotypes following exposure to four different photoperiods (300 genotypes for each photoperiod). Under a 12-h photoperiod, 50% of the genotype remained short (dormant) while the other 50% were tall. Under an 8-h photoperiod all plants remained dormant while none remained dormant under a photoperiod of 14 h

3. Let the plants slowly thaw overnight at 4  C in the dark. 4. Cut the plants to 3–4 cm height. 5. Transfer the plants to the following environmental conditions: 12 h photoperiod (see Note 9, Fig. 2) with an irradiance of 400–600 μmol/m2/s PPFD and a day–night temperature regime of 18:15  C for 4 weeks. 6. Select tall vigorous genotypes that survived the freezing stress (see Note 10). 7. Expose the selected genotypes to a second round of cold acclimation followed by exposure to a freezing stress down to a temperature close to the initial test temperature to eliminate another group of genotypes (see Note 11). 8. Transfer the plants to 12 h photoperiod with an irradiance of 400–600 μmol/m2/s PPFD and a day–night temperature regime of 18:15  C for 4 weeks. 9. Repeat the entire cycle of acclimation–freezing stress–regrowth a third time. 10. Among the plants remaining after the third cycle of selection, retain the 90 tallest and most vigorous genotypes (see Note 12). 11. Transplant these genotypes into 15 cm-diameter pots and transfer the plants to optimal growth conditions until flowering. 3.4

Intercrossing

Intercross the selected plants randomly using hand pollination or confined enclosure, depending on plant species and pollination type (Fig. 3).

58

Annick Bertrand et al.

Fig. 3 Crossing 90 superior genotypes of alfalfa in a greenhouse 3.5

Harvesting Seeds

1. Harvest the seeds of each genotype separately when they are ripened. 2. Pool seeds into a bulked sample using an equal representation of each genotype (see Note 13).

4

Notes 1. Depending on the plant species tested, seeds could need a pretreatment such as scarification or imbibition to optimize germination. 2. Proceed with a germination test before seeding to ensure having sufficient genetic material to complete three freezing stresses and end up with a sufficient number of genotypes to reduce the risk to create a genetic bottleneck during the recurrent selection process. 3. The lowest temperature of the freezer has to be lower or equal to the LT50 of the population under selection. In cases when the expected LT50 is lower than the lowest temperature of the freezer, stress intensity can be varied by increasing exposure time at the last plateau.

A Screening Test for Freezing Tolerance and Low Dormancy

59

4. Fertilisation should be stopped 1 week before transferring the plants to cold acclimation because it could interfere with the cold acclimation process. 5. Thermocouples measuring the actual temperature reached in the soil could be installed 1-cm deep in the soil. 6. The day before transferring the plants to 2  C, it is important that the plants have been uniformly well watered and well drained. Soil water content has a major impact on the process of plant cold acclimation and freezing tolerance [12]. 7. Acclimation at 2  C in the dark is very important to reach the maximum cold acclimation of the plants [13]. 8. Preliminary assays with a subset of plants are highly recommended to accurately target the LT50 and avoid insufficient or too intensive selection pressure that will markedly affect the screening process. LT50 varies with species, germplasm, and the number of cycles of selection. 9. The photoperiod that discriminates between dormant and nondormant cultivar could vary according to the origin (latitude) and the plants species under study. A preliminary test to determine this photoperiod is highly recommended (Fig. 2). 10. Each cycle of acclimation–freezing stress–regrowth should eliminate 60% of the plants: 50% should be killed by the frost stress and, among the survivors, the 10% most dormant genotypes should be eliminated. If you start with 3000, around 1200 plants should remain after a first cycle, 480 after two cycles and 192 after three cycles. After the third cycle, select the 90 tallest most vigorous genotypes among the 192 surviving plants. 11. The level of temperature of each freezing stress could be adjusted according to the number of surviving plants that are obtained after each cycle. If too many plants are killed after one cycle, set the second stress 1–2  C above the temperature of the previous stress. If not enough plants are killed, set the following stress 1–2  C below the previous temperature. If after three stresses too many undamaged plants are obtained, choose the 90 most vigorous individuals of the population to undergo the next cycle of selection. 12. For a better estimate of the vigour of the genotypes, it is recommended to make this last screening after 2–3 weeks of regrowth. 13. This pool of seeds represents the next TF-RD population that will be seeded to undergo another cycle of selection.

60

Annick Bertrand et al.

Acknowledgments This research is supported in part by Agriculture and Agri-Food Canada, and by additional contributions from Canadian Cattlemen’s Association under the Agri-Science Clusters Initiative. References 1. Schwab PM, Barnes DK, Sheaffer CC (1996) The relationship between field winter injury and fall growth score for 251 alfalfa cultivars. Crop Sci 36:418–426 2. Brummer EC, Cassler MD (2014) Cool-season forages. In: Smith S, Diers B, Specht J, Carver B (eds) Genetic gain in major U.S. field crops. CSSA Spec. Publ. 33. ASA, CSSA, and SSSA, Madison, WI 3. Castonguay Y, Dube´ M-P, Cloutier J et al (2013) Molecular physiology and breeding at the crossroads of cold hardiness improvement. Physiol Plant 147:64–74 4. Castonguay Y, Laberge S, Brummer EC et al (2006) Alfalfa winter hardiness: a research retrospective and integrated perspective. Adv Agron 90:203–265 5. Cunningham SM, Volenec JJ, Teuber LR (1998) Plant survival to detect differences in gene expression that result from and root and bud composition of alfalfa populations selected for contrasting fall dormancy. Crop Sci 38:962–969 6. Limin AE, Fowler DB (1991) Breeding for cold hardiness in winter wheat-problems, progress and alien gene expression. Field Crops Res 16:190–197 7. Volenec JJ, Cunningham SM, Haagenson DM et al (2002) Physiological genetics of alfalfa

improvement: past failures, future prospects. Field Crop Res 75:97–110 8. Bertrand A, Claessens A, Rocher S (2018) An indoor screening method for reduced dormancy in alfalfa. In: Brazauskas G, Statkeviciute G, Janaviciene K (eds) Breeding grasses and protein crops in the era of genomics. Springer, New York, pp 209–214 9. Bertrand A, Bipfubusa M, Castonguay Y et al (2016) A proteome analysis of freezing tolerance in red clover (Trifolium pratense L.). BMC Plant Biol 16:65 10. Iraba A, Castonguay Y, Bertrand A et al (2013) Characterization of populations of turf-type perennial ryegrass recurrently selected for superior freezing tolerance. Crop Sci 53:2225–2238 11. Castonguay Y, Michaud R, Nadeau P et al (2009) An indoor screening method for improvement of freezing tolerance in alfalfa. Crop Sci 49:809–818 12. Be´langer G, Castonguay Y, Bertrand A et al (2006) Winter damage to perennial forage crops in eastern Canada: causes, mitigation, and prediction. Can J Plant Sci 86:33–47 13. Dionne J, Castonguay Y, Nadeau P et al (2001) Freezing tolerance and carbohydrate changes during cold acclimation of green-type annual bluegrass (Poa annua L.) ecotypes. Crop Sci 41:443–451

Chapter 7 Mapping of Quantitative Trait Loci (QTL) Associated with Plant Freezing Tolerance and Cold Acclimation Evelyne Te´oule´ and Carine Ge´ry Abstract Most agronomic traits are determined by quantitative trait loci (QTL) and exhibit continuous distribution in natural or especially built segregating populations. The genetic architecture and the hereditary characteristics of these traits are much more complicated than those of oligogenic traits and need adapted strategies for deciphering. The model plant Arabidopsis thaliana is widely studied for quantitative traits, especially via the utilization of genetic natural diversity. Here we describe a QTL-mapping protocol for analyzing freezing tolerance after cold acclimation in this species, based on its specific genetic tools. Nevertheless, this approach can be applied for the elucidation of complex traits in others species. Key words QTL mapping, Arabidopsis thaliana, Freezing tolerance, Cold acclimation, Natural variability

1

Introduction Low temperatures, besides drought and salt stress, are among the most important abiotic environmental factors affecting the geographical distribution of plant species, as well as growth and yield of crop plants. In cold or temperate regions, the ability of plants to survive freezing temperatures is largely dependent of their ability to acclimate to cold. The acclimation process triggers an increase in freezing tolerance after an exposition to low but nonfreezing temperatures by inducing wide modifications in the physiology and metabolism of plants, such as changes in lipid composition of membranes, increase in the concentration of cryoprotectant molecules such as sugars and proline. These changes are largely associated with extensive variations of the transcriptome [1]. Due to its major agronomic importance, this trait has been extensively studied not only in crops but also in model plants such as Arabidopsis thaliana, and numerous cold induced genes have been identified [2, 3]. Moreover, uncertainties associated with global climate change and the observation of recurring frost periods in spring

Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_7, © Springer Science+Business Media, LLC, part of Springer Nature 2020

61

62

Evelyne Te´oule´ and Carine Ge´ry

pose new questions about the efficiency of acclimation and the risks associated with deacclimation. Although approaches based on monogenic mutant analyses have been very successful, freezing tolerance is a complex trait and specific genetic approaches, such as QTL mapping [4–6], have appeared as powerful tools to identify novel genes involved in freezing tolerance. These strategies have already been applied successfully to crop plants [7–9] and more recently to model species [10]. The model species Arabidopsis thaliana is chilling tolerant and capable of cold acclimation. It is widely distributed all over the Northern hemisphere and therefore natural accessions are submitted to various climatic conditions. Natural variation for freezing tolerance after cold acclimation has been demonstrated in several studies [11] and this trait appears to be correlated with habitat winter temperatures [12–14]. This suggests that this trait could be under natural selection and that natural genetic variability could be larger than previously predicted. Deciphering this complex trait via QTL mapping in Arabidopsis thaliana, where numerous genetic tools are available, appears particularly pertinent. These approaches could increase understanding of the gene networks and molecular mechanisms underlying the response to cold in A. thaliana and then identify new targets for breeding in crop plants. Several fruitful studies confirmed this working hypothesis [4–6, 11]. The main idea in QTL analysis and mapping is the identification of linkage disequilibrium or significant associations between genetic markers and quantitative phenotypic data in a segregating population issued from a biparental cross: this implies first, available fine genetic maps and also efficient and reproducible phenotypic tests. Phenotyping has proved to be a crucial part of the work. Considering freezing tolerance, this trait is continuous, which could be assessed through freezing tests in the field or under controlled conditions. Integrative measures consist of an evaluation of freezing injuries by means of grades on a standard scale, and/or evaluation of recovery by estimating the percentage of viability [11]. This way of evaluation is especially interesting in natural diversity studies, because it can be easily adapted to medium throughput and large sample numbers. Electrolyte leakage tests, based on the evaluation of plasmalemma alterations [15] also allow for the evaluation of freezing tolerance and are used in various species. In particular, the method of LT50 determination, that is, the temperature corresponding to 50% lethality measured via electrolyte leakage, is very reliable and has been used in several studies [4, 11, 16]. In this chapter, we describe protocols that are currently used in our team for mapping of quantitative trait loci (QTL) associated with plant freezing tolerance after cold acclimation in A. thaliana. To summarize, protocols are developed for three main steps: primary QTL detection in classical segregating populations and

Mapping Freezing Tolerance QTL

63

localization of wide candidate regions, validation of the QTL by genetic approaches using more specific material, then fine mapping of positive regions. A strategy based on mutants and genetic complementation will be described for the step of validation of potential candidate gene(s). Yet the context could be largely different for other species, for example in crop species with less available genetic tools or where the state of the art is less developed. Here, it will be advantageous to take all points outlined below into account before initiating an experimental strategy for QTL detection and mapping.

2 2.1

Materials Plant Material

2.2 DNA Extraction for Genotyping of Plant Material

Seed stocks of Arabidopsis thaliana accessions and mapping populations such as recombinant inbred lines (RILs) and highly inbred families (HIFs) are obtained from the Arabidopsis thaliana Biological Resources Centre (BRC) at the IJPB in INRA/Versailles. All seed stocks are described and can be ordered at http:// publiclines.versailles.inra.fr/. 1. Extraction buffer: 200 mM Tris–HCl (pH 7.5), 250 mM NaCl, 25 mM EDTA, 0.5% (w/v) SDS. 2. Isopropanol. 3. 75% ethanol. 4. TE buffer: 10 mM Tris–HCl (pH 8.0), 1 mM EDTA. 5. Ball mill MM30 (Retsch, Haan, Germany). 6. Table top centrifuge.

2.3 Evaluation of Freezing Tolerance

1. Plant growth and freezing tests: temperature-controlled greenhouse, growth chambers for acclimation and freezing periods (chambers with a 12 h photoperiod, 70 μmol quanta/m2/ s light intensity) able to stay stable at 4
C for acclimation and at temperatures between 4
C and 8
C for freezing test. 2. To ensure homogeneous germination, stratification of seeds is necessary: seeds are put into 0.1% agarose and kept at 4
C in the dark for 3 days. Sowing is then done by pipette.

2.4 QTL Analysis Software

MAPMAKER 3.0 [17] (https://www.mapmaker.com/) has been used to establish genetic maps. The Unix version of QTL CARTOGRAPHER 1.14 [18, 19] (statgen.ncsu.edu/qtlcart/) has been used to performed QTL analyses.

64

3

Evelyne Te´oule´ and Carine Ge´ry

Methods

3.1 Identification of Efficient QTL Mapping Populations

1. The first step consists of choosing the parental accessions of the mapping population. The parental lines must exhibit significant genetic distance from each other. Sometimes they are chosen for their phenotypic contrast, but this is not absolutely necessary. Effectively, despite a small difference in phenotype between accessions, large variability may be observed in the mapping populations derived from the cross (Fig. 1). 2. Several mapping populations are already available in BRC: F2 and RILs (Fig. 2) (see Note 1). RILs are often preferred to other populations when preexisting and available. If not, F2 populations are more rapid to obtain, especially in agronomic species, and could be used in primary QTL detection. Yet specific RIL populations can be built. The general strategy to produce this material is described in Fig. 2 (see Note 2). Mapping populations are genotyped with molecular markers: single nucleotide polymorphisms (SNPs), microsatellites and/or indels chosen to obtain regular spacing of markers on the genome. The distance between markers can vary depending on the plant material (see Note 3). MAPMAKER 3.0 is used to establish the genetic map and the Kosambi [20] mapping function is used to convert recombination data into map distances, as this has been shown to be optimal for A. thaliana [21, 22].

Fig. 1 Example of the distribution of the scores for freezing tolerance in a core population of RILs. This is an example of proximal parental phenotypes as seen from the distribution. Both parental phenotypes are indicated by arrows. This population is derived from a cross of the accessions Bur-0  Col-0 and in spite of a small difference in freezing tolerance phenotype between the two parents, transgressions and a strong variability are observed in the RIL population and freezing tolerance QTL were detected with a significant LOD score

Mapping Freezing Tolerance QTL

65

Fig. 2 Schematic illustration of the generation of mapping populations for QTL detection. Crossing homozygous parental accessions produces fully heterozygous F1 plants. In A. thaliana the parents are homozygous due to the predominant self- pollination of this species. F1 plants can be backcrossed with one parent, generating a BC population or self-pollinated to produce a F2 population. The F2 population itself can be used as a mapping population or lines can be propagated through single seed descent (SSD) until the F6 generation without selection. Then one plant per line is chosen again for selfing to obtain F7 seeds, which are used as a bulk for genotyping. For illustration, only one chromosome pair is shown 3.2 Evaluation of Freezing Tolerance Phenotypes

Seeds of all RILs to be investigated in a chosen population (see Note 4) are put in 0.1% agarose at 4
C in the dark for 3 days to ensure homogenous germination. They are then sown with a pipette in square pots containing organic substrate and irrigated with mineral nutrient solution for the first time and then only with water. Plants are sown in small bunches, 12 lines per pot in a random design allowing for blind notation. Each line is sown in five replicates, and mean scores are calculated. Plants are grown in the greenhouse for 14 days at which time they reach the 6–8 leaf stage. They are then transferred for cold acclimation to a growth chamber at 5
C at a 12 h photoperiod (70 μmol quanta/ m2/s) for 7 days. Acclimated plants are then exposed to freezing temperatures varying from 4
C to 8
C, depending upon germplasm, for 48 h. At this time, plants are removed from freezing conditions and put back in the greenhouse. These conditions have

66

Evelyne Te´oule´ and Carine Ge´ry

been optimized in a first round of experiments to maximize the variation in the response. A week later, freezing tolerance is determined by evaluating dead or alive plants, leaf damage and capacity for continued growth. A method favored by agronomists [23, 24] is used for evaluation: damage to leaves and percentage of survival are evaluated by noting on a scale ranging from 0 (no damage) to 6 (all plants dead) [25]. 3.3

QTL Analysis

1. The phenotype scores are collected in an Excel file, the mean and standard deviation are calculated for each RIL line to check the quality of scoring. Outlier values can be removed at this step. 2. Matrix data for QTL detection is built by combining the mean phenotypic values with genotype data for each line from the RIL population (Fig. 3).

Fig. 3 Example of freezing tolerance data treatment for QTL detection. The successive transformations of raw data during the QTL detection process with QTLCartographer are shown. (a) In the primary data file, all RIL lines available in BRC must appear in column X, even if the experiment was only performed on a subset of RILs. Missing data are noted with “-”. (b) Transformed file in txt format, data are organized in columns. (c) Transposed data file exhibiting the complete set of RILs. (d) Treated data file ready for use in QTLCartographer. Transposed phenotypic data are pasted at the end of the genotypic data

Mapping Freezing Tolerance QTL

67

3. QTL analyses are performed with a Unix version of QTL Cartographer [18, 19], which is a suite of programs for mapping QTL onto a genetic linkage map, but other software are available as well (see Note 5). First, IM (Interval Mapping) [26] is used to determine putative QTL involved in the variation of the trait. However, the localization and genetic effects of such detected QTL can be confused because of the presence of other linked QTL or by nonrandom segregation of other QTL in the population. Composite interval mapping (CIM) allows for addressing these limitations: the significant markers in the population, detected by regression analysis, are chosen as cofactors to estimate maximum likelihood for QTL. Effects of possible other QTL are then taken into account. CIM, Model 6 of QTL Cartographer, is then performed on the same data: the closest marker to each local LOD (Logarithms of Odds) score peak, a putative QTL, is used as a cofactor to control the genetic background while testing at a position of the genome [18, 19]. When a cofactor is also a flanking marker of the tested region, it is excluded from the model. The number of cofactors involved in our models varied between 1 and 6 for a full RIL set (400–500 lines) and a maximum of 4 for a core population (164 lines). The LOD significance threshold (2.3 LOD) is estimated from one thousand permutation tests [27, 28]. 4. To better illustrate the process, a concrete example of the procedure is depicted below. When using the successive programs proposed in QTLCartographer, some values or names are proposed by default. Here we show an example of how to manage the procedure (Fig. 4) (see Note 6). The parameters used with our mapping populations for detecting QTL associated with plant freezing tolerance are listed. We use the RILs of a core population of 19RV, derived from a cross between Can-0, an accession issued from the Canary Islands, used as female, and with a low acclimation potential, and, the reference accession Col-0, used as male, which has a good acclimation potential. 3.4 The QTLCartographer Process

1. In the original Excel file with all the phenotype data, put one trait per column, use “.” instead of “,” for numbers, use “-” for missing data, use « X » as title of the first column corresponding to RIL lines number, use the name of phenotypic trait as title corresponding to measures, here we have used only one: “note” for cold damage score (Fig. 3a). 2. Save this file as a text file (Windows). Named it popnumbertrait.txt (here: 19RVcold.txt). 3. Transpose lines/columns using functions of Splus (Fig. 3b) (http://www.insightful.com/products/splus/). 4. In the transposed file, add a “∗” before the trait name without space and save (Fig. 3c).

Fig. 4 Example of program workflow. When opening QTL Cartographer, in the successive screens values are proposed by default and some parameters need to be changed. In this example, the created directory to work with is 19RVcold and all file names must be identical. The proposed name for input file is RV20CIM.mps. To modify it select line 1 and change the name to 19RVcold.mps. When all parameters are correct, enter 0 to execute the program. This procedure can be applied in all programs of QTLCartographer (see text)

Mapping Freezing Tolerance QTL

69

5. In the QTL Cartographer directory, create a new directory for each population/trait studied (19RVcold in this example). 6. In this directory, add the “.raw” and “.maps” files, issued from MAPMAKER and downloaded from the Arabidopsis BRC website. Change “.maps” to “.mps”. 7. In the “.raw” file, change 0 (at the beginning of line 2) to the number of traits you have to analyze (here 1). Paste, after the genotyping data, the phenotype values from the transposed file with “∗” before the trait name (Fig. 3d). Add all trait values in the same file if there is more than one trait. 8. In the software, use the following commands: cd QTLCartographer Rmap -e 19RVcold.log -W 19RVcold -X 19RVcold 1 -> 19RVcold.mps 5 -> 2 (Kosambi) don’t change the resource file (qtl.cart) 0 (= execute) Rcross 1 -> 19RVcold.raw 0 Qstats 0 LRmapqtl (->more affected markers by trait) 7 -> 99 (number higher than number of traits -> all the traits) 0 SRmapqtl 6 -> 2 (FB model) 7 -> 99 0 Zmapqtl to do for each trait then take only one Eqtl for all 8 -> 3 9 -> 1 then 2, 3... 0 Eqtl to extract data 6 -> 3 10 -> 1 (LOD score) 0 Preplot 8 -> postscript (file.ps) 10 -> 1 0

Evelyne Te´oule´ and Carine Ge´ry

70

Then go back to the directory, in file 19RVcold.plt: – At line 10, replace set output’s name “19RVcold/19RVcold.ps” by set output’s name “19RVcold.ps”. – Adjust Y limits (LOD score): set Y range, replace 100 by 30. Go back to the application cd 19RVcold (directory name) gnuplot (postscript file is created) load “19RVcold.plt” quit

The most useful files created in the directory to represent the detected QTL are as follows: – “.ps” a pdf file with the graphs (Fig. 5). – “.eqt” will give you Marker position for each significant LOD score, LOD score value and the associated additive value (Fig. 6). – “.z3” will give you LOD score values all along the chromosome. Use this data if you want to make a graph with Excel. 30

LOD score

25

‘c1t1.s’ ‘c1t1.lr’ ‘c1t1.z3’ ‘c1t1.z6’

a

‘c1t1.s’ ‘c1t1.lr’ ‘c1t1.z3’ ‘c1t1.z6’

‘c1t1.s’ ‘c1t1.lr’ ‘c1t1.z3’ ‘c1t1.z6’

‘c1t1.s’ ‘c1t1.lr’ ‘c1t1.z3’ ‘c1t1.z6’

‘c1t1.s’ ‘c1t1.lr’ ‘c1t1.z3’ ‘c1t1.z6’

20 15 10 5 0 0

0,2

0,4

0,6

0,8

1

0

0,2

0,4

0,6

0

0,2

0,4

0,6

0,8

0

Position in Morgans on each chromosome

0,2

0,4

0,6 0

0,2

0,4

0,6

0,8

1

25

LOD score

20

b

15 10 5 0

LOD score

20 15

c

10 5 0

1

2

3 Chromosome

4

5

Fig. 5 Illustration of the localization of detected QTL in the 19RV population for the freezing tolerance trait. The five chromosomes of A. thaliana are represented individually and positions of the markers are on the abscissa, levels of LOD scores on the ordinate. (a) Graphs issued from QTLCartographer in “.ps” files, IM and CIM curves are superimposed. (b) Graphs issued from QTLCartographer in “.z3” files, only the IM curve is shown (c) Graphs obtained by analysis of the same data with the software R/qtl (compare a), IM and CIM curves are superimposed

Mapping Freezing Tolerance QTL

71

Fig. 6 Characteristics of the detected QTL explaining freezing tolerance variation in the Can-0  Col0 population (19RV). (a) Raw output Eqt file, raw file can be converted to an Excel file to work with data. Position: position of the QTL is expressed in cM from the first marker of the chromosome. Additive: represents the mean effect of the replacement of the non-Col-0 alleles by Col-0 at the locus. R2: represents the contribution of identified QTL (or interaction QTLQTL when significant) to the total phenotypic variation. (b) The .sr file is modified to process the CIM calculation. Statistical analysis classified the markers by rank and the highlighted Marker 11 with Rank 1 will be kept as a cofactor, allowing for detection of intervals to localize QTL more precisely

72

Evelyne Te´oule´ and Carine Ge´ry

It is then necessary to choose cofactors to run composite interval mapping (CIM): – Use the markers closest to the higher peak, even if it is not this one who has the higher LOD score value. – Avoid markers at the extremity of the chromosome. Modify the “.sr” file to let only markers used as cofactors. If there are only 3 cofactors, keep rank 1, 2, and 3 (Fig. 6). Go back to the application (cd ..) Zmapqtl to do for each trait then apply only one Eqtl for all 17 -> use the good stem 19RVcold 19 -> use the good working directory 19RVcold 8 -> 6 (CIM) 9 -> 1 then 2, 3... 12 -> maximum number of cofactors retained 13 -> Ws=5 0 Preplot 8 -> postscript (file.ps) 10 -> 1 0

Then go back to the directory, in file 19RVcold.plt: – At line 10, replace set output’s name “19RVcoldREP/ 19RVcold.ps” by set output’s name “19RVcold.ps”. – Adjust Y limits (LOD score): set Y range, replace 100 by 30. Go back to the application cd 19RVcold gnuplot (postscript file is created) load "19RVcold.plt" cd .. Eqtl extract data 6 -> 6 10 -> 1 (LOD score) 0

In the 19RVcold.eqt file, R2  100 explain x% of the phenotype variation and additivex2 correspond to 2a. – “.z6” will give you LOD score values all along the chromosome. Use these data if you want to make a graph with Excel.

Mapping Freezing Tolerance QTL

73

After running the program, detected QTL are represented in graphs (“.ps” files, or converted “.z3 “in Excel files). Figure 5 shows an example of the 19RV population analysis. All characteristics of detected QTL are summarized in the Eqt.out files (Fig. 6). For each detected QTL, the position on the chromosome and relative to the closest marker is indicated. Additive effects represent the mean effect of the replacement of the non-Col-0 alleles by Col-0 alleles at the locus. The R2 value represents the contribution of each QTL to the total phenotypic variation for the trait (see Note 7). 3.5

Validation of QTL

1. Residual heterozygosity in RILs is used to validate a potential QTL in a candidate region, via Highly Inbred Families (HIFs). This material is produced by self-fertilization of selected RILs exhibiting residual heterozygosity in the detected candidate region. Twenty-four seeds are sown and plants are grown individually. DNA of each plant is extracted and genotyping is realized with markers localized on the borders of the heterozygous region (and one in the middle if the region is wide). Five plants fixed A, which is the allele of one parent, five fixed B, which is the allele of the other parent and five were heterozygous. All the recombinants are kept and their selfed progenies are collected (see Note 8). 2. Three of these fixed progenies are phenotyped. Then, interpretation is based on the difference in the level of freezing tolerance between individuals that had fixed allele A or allele B: if their phenotype is significantly different, the HIF is retained (Fig. 7) (see Note 9). 3. The global analysis consists of searching for a correlation between genotype and phenotype. Due to the characteristics of phenotyping tests using small bunches of plants, fixed progeny testing (described above) is only appropriate for freezing tests and is used for this step, but other techniques are also available (see Note 10).

3.6

Fine Mapping

3.6.1 Screening Recombinant HIFs

1. These suitable HIFs identified as described in Subheading 3.5 are then submitted to a strategy aiming at the reduction of the size of the initial QTL region, which is often large (several Mb) and can encompass more than 1000 genes. Reducing the size of the region allows for efficient progress in identifying the causal polymorphism of the QTL. This step is based on production and analysis of recombinants. Recombinant plants (rHIF for recombinant HIF) are identified by genotyping among progeny of self-fertilized HIF (Fig. 8). To be efficient at this step, the number of plants to screen has to be estimated: in for this purpose, this theoretical formula can be used:

74

Evelyne Te´oule´ and Carine Ge´ry

Fig. 7 QTL validation with HIFs. The Can-0  Col-0 population (19RV) is shown as an example. (a) RIL 46 is shown as an example. It is self-fertilized and the progeny is genotyped. For easier reading, chromosome pairs are represented as one block only with colors depending upon the genotype: black is homozygous for Col-0, white is homozygous for Can-0, and gray is a heterozygous region. In RIL46, the residual heterozygous region is located on chromosome 4. The fixed progenies are phenotyped and if there is a significant difference in freezing tolerance between the HIF with the fixed Col-0 region and the HIF with the fixed Can-0 region, the QTL is validated (see pictures on the left). If there is no difference in freezing tolerance (pictures on the right), the RIL is not used for further analysis. (b) Localization of residual heterozygous regions along chromosome 4 in nine RIL lines. Five RILs, including RIL 46, showing a segregating phenotype in HIFs, are retained for further analysis, while the other four lines are not used further

number of expected recombinants ¼ ½ðsize of the regionÞ=ðsize of the chromosomeÞ  2  ½number of genotyped plants: The result constitutes an approximation and this step can be submitted to practical constraints (see Note 11) and necessitates a rapid

Mapping Freezing Tolerance QTL

75

Fig. 8 Example of genotyping for fine mapping. (a) RIL 46 is chosen for the residual heterozygote region between SNPlex 11 and SNPlex 15 corresponding to a peak of the LOD score. The initial size of the heterozygous region is 5.80 Mb. The initial RIL 46 has been self-fertilized, then progenies were genotyped.

76

Evelyne Te´oule´ and Carine Ge´ry

method to extract DNA because several thousand plants have to be analyzed in parallel. 3.6.2 96-Well Format DNA Isolation from Arabidopsis

1. Collect samples (one or two leaves from 2-week-old plants) in 1.2 ml 8-strip collection tubes containing two metal balls. 2. Add 120 μl of extraction buffer and cap the tubes. 3. Disrupt tissue for 5 min in a ball mill at maximum speed (30 vibrations/s). Turn racks over and disrupt again for 5 min. 4. Centrifuge for 10 min at 6000 rpm. 5. In a 96-well assay plate (polypropylene), add 100 μl of isopropanol in each well and then transfer 100 μl of supernatant (pipet slowly). Mix by pipetting. Cover with a flexible assay plate cover. 6. Incubate samples at room temperature for at least 10 min. 7. Centrifuge at 6000 rpm. 8. Decant supernatant: turn plate upside down carefully and then tap gently on paper. 9. Add 100 μl of 75% ethanol. Cover samples with a flexible assay plate cover. 10. Centrifuge for 5 min at 6000 rpm. 11. Decant supernatant as described before and allow drying. 12. Add 50–100 μl (depending upon starting material) of TE buffer and let DNA suspend. 13. Use 1 or 2 μl of DNA for genotyping PCR.

3.6.3 Narrowing Down the QTL Region

As in the first round of detection, fixed progeny is the only suitable strategy with our criteria of scoring freezing damage. The analysis of phenotyping results at this step allows for retaining or excluding tested regions: if the freezing tolerance of a fixed rHIF differs significantly from the freezing tolerance of lines outside the QTL, the gene involved in the QTL is probably localized in the region. If there is no difference, this zone can be excluded from the analysis. This phenotyping/genotyping loop is repeated until the region of

ä Fig. 8 (continued) Twenty lines are shown as examples: some are fixed for one parent or the other, some are still heterozygous and some are recombinants. (b) The recombinant line 46-6 has been self-fertilized then progenies were genotyped. The heterozygous region in line 46-6 is reduced to 2.8 Mb. Again, among the 28 progenies shown here some are fixed for one or the other parent, some are heterozygous and others are recombinant. (c) Same analysis for the recombinant line 46-12. This line was chosen to split the initial region in two candidate regions of 2.8 Mb and 3.12 Mb, respectively, as fixed in complementary regions of the initial heterozygous region. The same strategy is pursued until the size of the candidate region is sufficiently reduced. A Col-0, B Can-0, H heterozygote, ? missing genotype. SNPlex SNP initially used for mapping the RIL population, MSAT microsatellite marker, IND insertion/deletion marker

Mapping Freezing Tolerance QTL

77

interest is reduced sufficiently in size to contain a small number of genes that can be tested directly (see Subheading 3.7). Below a certain size, around 10 kb in Arabidopsis, it is not feasible to continue with this procedure: the number of plants that need to be genotyped and tested becomes too large (several thousands) to be practical. A better strategy is to use specific rHIFs to cross them and get advanced rHIFs: arHIF [29] (Fig. 9). 3.7 Validation of Candidate Gene(s)

1. When the candidate region is sufficiently reduced to include a workable number of genes, go to specific databases to identify the genes precisely and check if knock-out (KO) mutants or knock-down lines are available (T-DNA or chemically induced mutants, RNAi lines, or others). If there are no such lines available, an alternative strategy would be to create KO lines using CRISPR/Cas9 [30, 31]. 2. Phenotype the mutants for their freezing tolerance and compare to the phenotype of the RILs. If the phenotype of a KO mutant is close to that of the fixed HIF progeny, the gene is a real good candidate. 3. Check the sequences of the candidate gene in the parental accessions to search for specific polymorphisms and enlarge the analysis to a significant number of accessions. A mutation, recurrent in natural collections and associated with an analogous phenotype, is a good candidate for causality. 4. Complement the mutant(s) via classical transgenesis, using the allele of each parent: if even a partial functional complementation is observed, the gene is confirmed as a candidate. 5. After this step, numerous analyses such as comparison of transcriptomes [1] and quantitative complementation by crosses with others mutants can be performed to reinforce and further characterize the identified candidate gene.

3.8

Conclusion

Cold is a major abiotic stress affecting crop productivity all over the world. Increasing the knowledge of genetic mechanisms controlling the tolerance to freezing and the adaptation to unstable climate conditions would be of great help for agronomic purposes. The QTL mapping strategy is a powerful tool that allows for the identification of new genes involved in molecular pathways without a priori candidates, in particular when based on natural genetic variability. All steps described here for mapping and cloning are summarized in Fig. 10. This protocol is directly usable for Arabidopsis and is adaptable to other species.

78

Evelyne Te´oule´ and Carine Ge´ry

Fig. 9 Creation of advanced rHIFs (arHIFs). Two complementary rHIFs are chosen in the restricted candidate zone. They are fixed by self-fertilization to produce on a small genomic candidate region. One is fixed for one parent and the other for the second parent. Then, they are crossed, constituting an arHIF, which is heterozygous only for a small part of the candidate region. Validation is done by progeny or fixed progeny testing. This material is free of interaction with borders and is very valuable for further analysis

Mapping Freezing Tolerance QTL

79

Fig. 10 General chart to identify QTL associated with freezing tolerance in Arabidopsis. Main steps are summarized in this figure and the repeated loop to reduce the initial candidate region is shown by an arrow

4

Notes 1. The Versailles Arabidopsis Stock Center provides seeds useful to the international research community. Numerous seed stocks are available: more than 600 natural accessions, 262 F2 mapping populations, 16 RIL populations, and 3 near-isogenic line sets (HIF: Heterogeneous Inbred Family). Available RILs are all derived from the same type of crossing: the female parent is one accession of the core collection (maximizing the diversity of the whole collection) and the male parent is Col-0 (first reference for Arabidopsis genome sequence). For each RIL population are available: a minimal set, a core population of 164 RILs and the complete set. The minimal set includes 20 lines representative of the whole RIL population, allowing for easy checking of the phenotypic diversity in the selected cross. The core population is a set of 164 RILs optimized for QTL detection. The complete set includes all RIL lines issued

80

Evelyne Te´oule´ and Carine Ge´ry

from the initial cross, derived from the 500 starting F2 plants. All genotyping data are also available and can be downloaded as “.raw” and “.maps” files. 2. RIL lines are homozygous plants with a low residual heterozygosity (around 3%). RIL populations are built as follows: F1 seeds are produced by crossing the selected parental accessions, plants are grown and self-fertilized. Five hundred of the resulting F2 lines are propagated through single seed descent without selection until the F6 generation is reached. Then, one plant per line is chosen again for selfing to obtain F7 seeds, which are used as a bulk for genotyping. Each F7 line in the population contains a unique mosaic of the two parental genomes. It has retained specific recombination events along the selfing cycles. A great advantage of RIL populations is that the genotypes of the lines are fixed, and then, once genotypes for a population are identified, the population can be used for any number of replications or analyzed and measured under specific conditions for any number of traits. The main weakness of this material is that, in some species, there are no preexisting RIL populations, and that building such material is time consuming and expensive. In this case, backcrosses or F2 plants can be used. The genetic complexity of this kind of material could be bypassed by increasing the number of tested lines to keep sufficient power of QTL detection. F2 populations are developed by selfing F1 plants. Each individual in the F2 generation receives recombinant chromosomes from each parent so that at each locus the genotype is AA, AB, or BB. This is a powerful tool for QTL detection, as all genotypes are represented. This structure also allows for analyzing additive and dominant effects at the detected QTL. The main weakness of F2 material is the impossibility of replication in scoring the phenotype, so some traits cannot be analyzed with such a population. For example, in the Arabidopsis freezing tolerance project, integrative scores for freezing damage after acclimation cannot be evaluated on individual plants and therefore F2 populations are unusable. Backcross progenies are genetically intermediate structures: when two inbred lines, denoted A and B, are crossed the resulting F1 generation is fully heterozygous (AB). These F1 individuals are then crossed with one of the parental lines, producing the backcross progeny. This strategy is easy to set up, but at each locus only one of the homozygous parental genotypes is present. Therefore, only one side of the cross can be exploited for QTL detection, and the detection of dominant traits is excluded. Moreover, it is very difficult to detect epistasis effects in such a population.

Mapping Freezing Tolerance QTL

81

3. SNP, microsatellite and indel markers are used for genotyping. The selected markers must be evenly distributed along the chromosomes, a mean distance of 5–10 cM is sufficient, but distance must not exceed 20 cM. In our experiments, 80–90 markers are used on 200–300 individuals. The proportion of each parental allele has to be around 50% in the RIL populations to avoid biased analysis. At last, due to the large amount work required for genotyping, the selected markers have to be easy to use, that is, high specificity and easy identification of heterozygotes (SNPs are generally typed with multiplex system). 4. The definition of the number of lines to test is a complex task because the question of the best size of the population has no easy answer: the researcher must find a consensus between theoretical rules for the power of QTL detection and practical constraints combining genotyping of individuals and collecting phenotypic data. The general rule is “the more RILs the better” [20]. This could be particularly important when smalleffect QTL are supposed to interfere in the genetic architecture of a complex trait. Therefore, it appears as a good choice to increase the number of lines rather than the number of replicates if the effort to collect phenotypic data is a limiting constraint [21]. Our phenotypic analysis is based on an integrative parameter, easy to measure and nondestructive for the plant. It is an efficient way to analyze a large number of RILs in a population to evaluate natural variability for the character. When using time-consuming or destructive tests, or a very expensive analysis, such as metabolomics for example, it could be better to adapt the sample collection strategy. Nevertheless, in order to reduce the phenotypic task it is possible to define “core populations” in a selected RIL population. The core population is an optimal subset of lines (164/500 in the case of our Arabidopsis populations) based on genetic diversity and allowing the user to phenotype only a reduced number of lines without losing QTL detection power. 5. Besides QTL Cartographer, presented here as an example, numerous other software for QTL detection is available, each with specific capacities or tools and it could be useful to explore the different possibilities to optimize the choice for each specific case. Here is a nonexhaustive list: MapQTL (https://www. kyazma.nl/index.php/MapQTL/), R/qtl (http://www.rqtl. org/), QGene (http://www.qgene.org/), QTLNetwork (https://doc.qt.io/qt-5/qtnetwork-index.html), GridQTL (http://www.gridqtl.org.uk), and QTL IciMapping (https:// www.integratedbreeding.net/).

82

Evelyne Te´oule´ and Carine Ge´ry

6. Figure 4 illustrates the practical use of the QTL Cartographer software: what options are available and how the parameters can be changed. It is very important to check that all files have the same name. Here, 19RVcold has been chosen to create the directory and all the following files must have exactly this name, otherwise the program will abort. 7. In quantitative trait analysis, the genetic model includes dominant and additive effects of alleles. Depending upon the genetic structure of the mapping population it is possible to calculate the additive value: here, and mentioned as additive in the raw file columns of Eqtl, the additive values are calculated as the effect of replacement of the allele from the Can-0 parent by the allele of Col-0. The negative values are due to the score scale: the higher the score is, the more damaged the plants were by freezing. On the other hand, the effects of dominance can be estimated only if analysis has been run on F2 population data. In all other cases, values are equal to zero. 8. Five plants of each type will be retained to be able to replace unexpected losses (dead plants, sterility, delayed development, etc.) and recover enough (three individuals) plants of each type without the need to generate a new generation. 9. At this step, only positive HIFs can be selected. Effectively, as the genome of a RIL is a unique mosaic of parental genomes, the absence of validation can be due to epistatic interactions between nonidentified regions. That is the reason why, when possible, it is better to test several HIFs covering the candidate region to avoid local phenomena or specific genetic interactions. At last, easily “workable lines” must be chosen with normal development and fertility and low sensitivity to pathogens, to increase the efficiency of the following steps. 10. When screening and phenotyping are possible on the same individual plant, progeny testing can be used. It consists in screening large amount of material to correlate individual phenotype with genotype. Practically, around 100 seeds of a HIF are sown individually, grown in phenotyping conditions, and genotyped with markers located at the extremities of heterozygote zone. The interpretation is based on the same scheme that of the fixed progeny. 11. For example, when genotyping 1000 plants for a region of around 50 kb on chromosome 4 (which is 19,000 kb long), around 5 recombinants are expected. This can vary a lot depending on the location of the region on the chromosome and also due to chance events. The general rule is to divide the candidate region into regular intervals with molecular markers. It is not useful to analyze a large population of recombinant lines at the same time, because in this material the genetic

Mapping Freezing Tolerance QTL

83

background is globally homogeneous between lines and phenotyping fixed progenies allows for excluding regions at each step. Therefore, it is better to search for regularly spaced markers to be able to refine the candidate genome region as quickly as possible. References 1. Zuther E, Lee YP, Erban A et al (2018) Natural variation in freezing tolerance and cold acclimation response in Arabidopsis thaliana and related species. Adv Exp Med Biol 1081:81–98 2. Liu Y, Dang P, Liu L et al (2019) Cold acclimation by the CBF-COR pathway in a changing climate: lessons from Arabidopsis thaliana. Plant Cell Rep 38:511–519 3. Pareek A, Khurana A, Sharma AK et al (2017) An overview of signalling regulons during cold stress tolerance in plants. Curr Genomics 18:498–511 4. Meissner M, Orsini E, Ruschhaupt M et al (2013) Mapping quantitative trait loci for freezing tolerance in a recombinant inbred line population of Arabidopsis thaliana accessions Tenela and C24 reveals REVEILLE1 as negative regulator of cold acclimation. Plant Cell Environ 36:1256–1267 5. Tayeh N, Bahrman N, Sellier H et al (2013) A tandem array of CBF/DREB1 genes is located in a major freezing tolerance QTL region on Medicago truncatula chromosome 6. BMC Genomics 14:814 6. Oakley CG, Savage L, Lotz S et al (2018) Genetic basis of photosynthetic responses to cold in two locally adapted populations of Arabidopsis thaliana. J Exp Bot 69:699–709 7. Zhang Z, Li J, Pan Y et al (2017) Natural variation in CTB4a enhances rice adaptation to cold habitats. Nat Commun 8:14788 8. Jha UC, Bohra A, Jha R (2017) Breeding approaches and genomics technologies to increase crop yield under low-temperature stress. Plant Cell Rep 36:1–35 9. Li J, Pan Y, Guo H et al (2018) Fine mapping of QTL qCTB10-2 that confers cold tolerance at the booting stage in rice. Theor Appl Genet 131:157–166 10. Marchadier E, Hanemian M, Tisne´ S et al (2019) The complex genetic architecture of shoot growth natural variation in Arabidopsis thaliana. PLoS Genet 15:e1007954 11. Ge´ry C, Zuther E, Schulz E et al (2011) Natural variation in the freezing tolerance of Arabidopsis thaliana: effects of RNAi-induced CBF

depletion and QTL localisation vary among accessions. Plant Sci 180:12–23 12. Zuther E, Schulz E, Childs LH et al (2012) Clinal variation in the non-acclimated and cold-acclimated freezing tolerance of Arabidopsis thaliana accessions. Plant Cell Environ 35:1860–1878 13. Gehan MA, Park S, Gilmour SJ et al (2015) Natural variation in the C-repeat binding factor cold response pathway correlates with local adaptation of Arabidopsis ecotypes. Plant J 84:682–693 14. Kang J, Zhang H, Sun T et al (2013) Natural variation of C-repeat-binding factor (CBFs) genes is a major cause of divergence in freezing tolerance among a group of Arabidopsis thaliana populations along the Yangtze River in China. New Phytol 199:1069–1080 15. Whitlow TH, Bassuk NL, Ranney TG et al (1992) An improved method for using electrolyte leakage to assess membrane competence in plant tissues. Plant Physiol 98:198–205 16. Burr KE, Tinus RW, Wallner SJ et al (1990) Comparison of three cold hardiness tests for conifer seedlings. Tree Physiol 6:351–369 17. Lander ES, Green P, Abrahamson J et al (1987) MAPMAKER: an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics 1:174–181 18. Basten CJ, Weir BS, Zeng ZB (2000) QTL CARTOGRAPHER version 1.14. North Carolina State University, Raleigh, NC 19. Da Costa L et al (2012) Composite interval mapping and multiple interval mapping: procedures and guidelines for using windows QTL Cartographer. Methods Mol Biol 871:75–119 20. Kosambi DD (1944) The estimation of map distances from recombinant values. Ann Eugenics 12:172–175 21. Loudet O, Chaillou S, Camilleri C et al (2002) Bay x Shadara recombinant inbred line population: a powerful tool for the genetic dissection of complex traits in Arabidopsis. Theor Appl Genet 104:1173–1184

84

Evelyne Te´oule´ and Carine Ge´ry

22. Simon M, Loudet O, Durand S et al (2008) Quantitative trait loci mapping in five new large recombinant inbred line populations of Arabidopsis thaliana genotyped with consensus single-nucleotide polymorphism markers. Genetics 178:2253–2264 23. Andaya VC, Mackill DJ (2003) Mapping of QTLs associated with cold tolerance during the vegetative stage in rice. J Exp Bot 54:2579–2585 24. Mugabe D, Coyne CJ, Piaskowski J et al (2019) Quantitative trait loci for cold tolerance in chickpea. Crop Sci 59:573–582 25. McKhann HI, Gery C, Be´rard A et al (2008) Natural variation in CBF gene sequence, gene expression and freezing tolerance in the Versailles core collection of Arabidopsis thaliana. BMC Plant Biol 8:105 26. Monir MM, Khatun M, Mollah MNH (2018) β-composite interval mapping for robust QTL analysis. PLoS One 13:e0208234

27. Churchill GA, Doerge RW (1994) Empirical threshold values for quantitative trait mapping. Genetics 138:963–971 28. Neto EC, Keller MP, Broman AF et al (2012) Quantile-based permutation thresholds for Quantitative Trait Loci hotspots. Genetics 191:1355–1365 29. Krymann J, Mitchell-Olds T (2005) Epistasis and balanced polymorphism influencing complex trait variation. Nature 435:95–98 30. Wolter F, Schindele P, Puchta H (2019) Plant breeding at the speed of light: the power of CRISPR/Cas to generate directed diversity at multiple sites. BMC Plant Biol 19:176 31. Rothan C, Diouf I, Causse M (2019) Trait discovery and editing in tomato. Plant J 97:73–90

Chapter 8 Identification of Arabidopsis Mutants with Altered Freezing Tolerance Carlos Perea-Resa, Rafael Catala´, and Julio Salinas Abstract Low temperature is an important determinant in the configuration of natural plant communities and defines the range of distribution and growth of important crops. Some plants, including Arabidopsis thaliana, have evolved sophisticated adaptive mechanisms to tolerate freezing temperatures. Central to this adaptation is the process of cold acclimation. By means of this process, many plants from temperate regions are able to develop or increase their freezing tolerance in response to low, nonfreezing temperatures. The identification and characterization of factors involved in freezing tolerance is crucial to understand the molecular mechanisms underlying the cold acclimation response and has a potential interest to improve crop tolerance to freezing temperatures. Many genes implicated in cold acclimation have been identified in numerous plant species by using molecular approaches followed by reverse genetic analysis. Remarkably, however, direct genetic analyses have not been conveniently exploited in their capacity for identifying genes with pivotal roles in that adaptive response. In this chapter, we describe a protocol for evaluating the freezing tolerance of both nonacclimated and cold acclimated Arabidopsis plants. This protocol allows for the accurate and simple screening of mutant collections for the identification of novel factors involved in freezing tolerance and cold acclimation. Key words Freezing temperature, Freezing tolerance, Cold acclimation, Arabidopsis thaliana

1

Introduction Plants are sessile organisms continuously adapting to the environmental changes to ensure an appropriate development. Low temperatures are one of the most important environmental constraints that limit the development and survival of plants, and determine their geographical distribution [1]. The stress induced by low temperatures also produces important economic losses, reducing the yield of agricultural crops every year. It is known that modest increases in the freezing tolerance of crop species would positively affect agricultural production [2]. Plants from temperate regions have evolved an adaptive response, known as a cold acclimation [1, 3], whereby they develop or increase their freezing tolerance after being exposed during several days to low, nonfreezing

Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_8, © Springer Science+Business Media, LLC, part of Springer Nature 2020

85

86

Carlos Perea-Resa et al.

temperatures (0–10  C). Understanding the molecular mechanisms underlying this response is essential to conceive how plants grow and develop under adverse conditions originated by abiotic stresses, and to generate new biotechnological strategies to improve crop tolerance to freezing temperatures and other related stresses such as drought and high soil salinity. Genetic analysis is a classical and powerful tool for identifying genes implicated in a given physiological process. In response to freezing temperatures, the identification and characterization of mutant plants with altered freezing tolerance before and/or after cold acclimation has been carried out essentially in Arabidopsis, a model plant that is able to acclimate to low temperature, increasing its constitutive freezing tolerance. Its small genome, the first to be sequenced in plants, together with its physiological characteristics, facilitates the subsequent molecular identification and characterization of the mutated genes. The most commonly used mutants in the screenings were generated by ethyl methanesulfonate (EMS), an organic compound that randomly produces nucleotide substitutions in DNA [4–6]. For instance, Warren et al. [7] identified several Arabidopsis EMS mutants, termed sensitivity to freezing (sfr), that showed reduced freezing tolerance compared with wildtype (WT) plants. Consistent with the expectations that sfr should be loss-of-function mutations, most of them were recessive. Seven sfr mutants were nonallelic and only acquired partial freezing tolerance after cold acclimation. A preliminary study revealed that four sfr mutations, sfr3, sfr4, sfr6, and sfr7, reduced or blocked anthocyanin accumulation during this adaptive response. The sfr4 mutant was also impaired in cold-induced accumulation of sucrose and glucose, and both sfr4 and sfr7 mutants showed abnormal fatty acid composition when cold acclimated [8]. In another study [9], Xin and Browse identified several Arabidopsis EMS mutants with increased freezing tolerance. One of them, eskimo1 (esk1), presented an increase in both constitutive freezing tolerance and cold acclimation capacity. esk1 was originated by a single recessive mutation in the AT3G55990 locus that produced elevated proline levels but did not generate constitutive expression of cold-regulated genes. Finally, Llorente and colleagues [10] identified freezing sensitive 1 ( frs1), an Arabidopsis EMS mutant that exhibited decreased constitutive freezing tolerance and capacity to cold acclimate. Complementation analysis revealed that the frs1 mutation was a new allele of ABA3, supporting that ABA is essential for full development of cold acclimation and for constitutive freezing tolerance in Arabidopsis. In this chapter, we describe a simple and precise protocol for evaluating the freezing tolerance of both, nonacclimated and cold acclimated Arabidopsis plants. The protocol is suitable for screening mutants generated by EMS, Fast Neutron (FN) or T-DNA insertions, and can be carried out with plants grown on media, in Petri

Screening for Freezing Tolerant and Sensitive Mutants in Arabidopsis

87

dishes, or on soil, in pots. Important aspects, depending on searching for mutants with increased (tolerant) or decreased (sensitive) freezing tolerance, are also detailed. In addition, this protocol can also be used in reverse genetic studies to determine the involvement of a gene of interest in freezing tolerance, and to assess the effect that different treatments may produce on the tolerance of Arabidopsis to freezing temperatures.

2

Materials

2.1

Plant Material

2.2

Plate Assay

WT seeds of the appropriate ecotype, mutagenized M2 families or M2 pools, depending if screening for sensitive or tolerant mutants respectively (see Note 1), and seeds of previously reported tolerant and/or sensitive freezing mutants to be used as controls in the screenings [7, 9]. EMS and FN mutagenized seeds from different ecotypes are commercially available at Lehle seeds (www.ara bidopsis.com), while T-DNA mutant collections can for example be ordered at the Nottingham Arabidopsis Stock Centre (NASC) and can be searched online through The Arabidopsis Information Resource (TAIR; https://www.arabidopsis.org). 1. 1.5 mL Eppendorf or 50 mL Falcon tubes. 2. Germination medium (GM) (0.5 Murashige and Skoog basal salt mix; 2.5 mM Morpholino ethanesulfonic acid (MES), 1% sucrose, pH 5.7; 0.8% agar). 3. Amphotericin B (final concentration 2.5 mg/L). 4. Round plates (Ø 15 cm). 5. 3M Micropore tape. 6. Filter paper or nylon mesh. 7. Bell jar. 8. Bleach. 9. HCl. 10. Forceps. 11. Liquid nitrogen. 12. Mortar and pestle. 13. Spoon.

2.3

Soil Assay

1. Peat substrate. 2. Vermiculite. 3. Clay pots (Ø 10 cm). 4. Trays. 5. Plastic film.

88

Carlos Perea-Resa et al.

2.4 Growth Chambers and Other Equipment

1. Plant growth chamber set at 20–22  C with cool-white light (100 μmol/m2/s). 2. Plant growth chamber set at 4  C with cool-white light (50 μmol/m2/s). 3. Plant growth chamber with a range of programmable temperatures from 4 to 14  C with lights off. 4. Cold room. 5. Autoclave. 6. Fume hood. 7. Sterile work bench.

3

Methods

3.1 Seed Mutagenesis

Protocols to obtain EMS or FN mutagenized seeds have been previously described [5, 11]. T-DNA mutant collections can be generated as reported [12]. Generation of M2 families and M2 pools from M1 mutagenized seeds has already been communicated [4] (see Note 1).

3.2 Screening Using Plates

We strongly recommend vapor phase seed sterilization using chlorine gas (see Note 2).

3.2.1 Seed Sterilization and Plating

1. Put WT, control and mutagenized seeds into appropriate tubes depending on number (Eppendorf or Falcon tubes are suitable). 2. Open the tubes into a hermetic bell jar placed in a fume hood. 3. Generate chlorine gas by combining 100 mL of bleach and 3 mL of HCl in a 200 mL glass placed inside the jar. 4. Close the jar and let the seeds exposing to chlorine gas for 3 h. 5. Open the jar inside the fume hood and air-ventilate the seeds for 15 min before closing the tubes. 6. Cut filter paper pieces according to the plate size and autoclave (see Note 3). 7. Place sterile filter papers on MS plates by using sterilized forceps. 8. Distribute sterilized seeds over the sterile papers (see Note 4). 9. Seal plates with 3M Micropore tape and transfer them to a cold room (4  C) under darkness during 2 days for seed stratification. 10. Transfer plates to a growth chamber and let seeds germinate and develop for 12 days at 20–22  C under long-day conditions (16 h light:8 h darkness).

Screening for Freezing Tolerant and Sensitive Mutants in Arabidopsis 4 °C, 1 h

89

4 °C, 1 h

-3 °C, 6 h (S) -7 °C, 6 h (T)

Non-Acclimated seedlings/plants

+1 °C/30 min

-1 °C/30 min

-9 °C, 6 h (S) -14 °C, 6 h (T)

Cold-Acclimated seedlings/plants

Fig. 1 Schematic representation of the freezing program used for the screenings. In all cases, before being subjected to freezing temperatures, seedlings and plants are exposed for 1 h to 4  C in the freezing chamber. Then, temperature is progressively decreased ( 1  C/30 min) until reaching the desired freezing temperature. As an example, the two different temperatures we generally use to screen nonacclimated seedlings or plants (Col-0) ( 3 and 7  C), and the two temperatures we generally use to screen cold-acclimated seedlings or plants (Col-0) ( 9 and 14  C) are shown. 3 and 9  C are employed when looking for sensitive mutants (S), while 7 and 14  C when looking for tolerant mutants (T). After exposing plants to the appropriate freezing temperature for 6 h, temperature is gradually increased to 4  C (+1  C/ 30 min). One hour later, plants are transferred to 20  C under long-day light regime for recovery and subsequent survival evaluation 3.2.2 Freezing Assay

1. Transfer plates to the growth chamber for freezing assay (see Notes 5 and 6) and expose seedlings to freezing temperatures under dark conditions. Appropriate temperatures should be empirically established depending on the accessions employed, the type of seedlings used (nonacclimated or cold acclimated) in the screening and the searched mutants (tolerant or sensitive to freezing) (see Notes 7 and 8) (Fig. 1). 2. Prepare fresh ice chips by vaporizing sterile distilled water in a mortar containing liquid nitrogen. Grind the ice to a fine powder and gently apply over the seedlings homogeneously when temperature decreases to 2  C. Close the plates and let the program finish (see Note 9). 3. After the freezing assay, when seedlings are exposed to 4  C and medium is still frozen, move plates to a sterile work bench and, using sterilized forceps, carefully transfer the filter papers with

90

Carlos Perea-Resa et al.

Fig. 2 Identification of Arabidopsis mutants with increased freezing tolerance before and after cold acclimation. (a) Nonacclimated seedlings 1 week after being exposed to 7  C for 6 h. Col-0 (WT), tolerant control (TC), and two M4 homozygous families for two tolerant mutations are shown. (b) Cold acclimated plants 1 week after being exposed to 14  C for 6 h. Col-0 (WT), tolerant control (TC), and a M3 segregating family for a tolerant mutation are shown

the frozen seedlings to new plates containing GM medium supplemented with amphotericin B (see Notes 10 and 11). 4. Move seedlings to a growth chamber and let them to recover at 20–22  C for 1 week under long-day light regime. An example of recovered seedlings is shown in Fig. 2a (see Note 12). 5. When screening for tolerant mutants, transfer surviving M2 seedlings to soil and allow them to reproduce for phenotype confirmation in a secondary screening with the corresponding M3 families (see Subheading 3.4). 6. If screening for sensitive mutants, repeat the freezing assays with additional seeds from the selected M2 lines (see Note 13), confirm the survival rate, and proceed to phenotype confirmation (see Subheading 3.4). 3.3 Freezing Using Pots

1. Mix peat substrate with vermiculite in a 3:1 ratio and add one volume of water per three volumes of mix.

3.3.1 Pot Preparation and Growth Conditions

2. Fill pots homogeneously with soil avoiding leaving air bubbles (see Note 14). 3. Distribute the seeds on the soil as separated as possible (see Note 15).

Screening for Freezing Tolerant and Sensitive Mutants in Arabidopsis

91

4. Place the pots into trays and cover it with plastic film for humidity maintenance during the first days after germination, and transfer them to a cold room (4  C) in darkness for 2 days for stratification. 5. Move the trays containing the pots to a growth chamber set at 20–22  C with a long-day light regime, and allow seeds to germinate and develop for 2 weeks. 3.3.2 Freezing Assay

1. Transfer individual pots to the growth chamber for freezing assay (see Notes 6 and 16) and expose plants to freezing temperatures. 2. Screening conditions are essentially the same as those described for seedlings on plates (see Notes 7, 8, and 17) (Fig. 1). 3. After the freezing assay, move plants to a growth chamber and let them to recover at 20–22  C for 1 week under long-day light regime (Fig. 2b). 4. When screening for tolerant mutants, allow surviving M2 plants to reproduce for phenotype confirmation in a secondary screening with the corresponding M3 families (see below). 5. If screening for sensitive mutants, repeat the freezing assays with additional seeds from the selected M2 lines (see Note 13), confirm the survival rate, and proceed to phenotype confirmation in a secondary screening (see below).

3.4 Confirmation of Mutant Phenotypes

1. When screening for freezing tolerant mutants, screen about 100 seedlings or plants from each generated M3 family for their freezing tolerance and calculate the survival rate. 2. M3 families from freezing tolerant M2 seedlings or plants showing survival rates of 100% are likely produced by a single mutation in homozygosis. Families showing a 3:1 or 1:3 tolerant–sensitive ratio would be generated by a single dominant or recessive mutation in heterozygosis, respectively (see Note 18). In these cases, select and reproduce at least ten freezing tolerant seedlings or plants to obtain the corresponding M4 families. The screening of these will allow for the identification of the homozygous mutant lines (see Note 19). 3. If screening for freezing sensitive mutants, collect seeds from at least ten M2 plants from each selected M2 family to obtain the corresponding M3 plants. 4. The screening of around 100 seedlings or plants from each generated M3 family will allow for the identification of homozygous mutant lines for the phenotype selected (those showing survival rates of 0%). 5. Once the mutant homozygous lines are identified, the tolerance phenotypes should also be corroborated by means of full

92

Carlos Perea-Resa et al. 4 °C, 1 h .5 -0

-4 °C

/h °C

-5 °C

-8 °C

h C/

° -1

-6 °C -7 °C

Non-Acclimated plants

h C/

° -1

-9 °C

-7 °C -8 °C

-9 °C -10 °C -11 °C

Cold-Acclimated plants

-12 °C

Fig. 3 Schematic representation of the freezing program used for establishing the freezing–survival curve. In all cases, before being subjected to freezing temperatures, plants growing on soil are exposed for 1 h to 4  C in the freezing chamber. Then, temperature is progressively decreased ( 0.5  C/h for nonacclimated and 1  C/h for cold-acclimated plants) until reaching the desired freezing temperature. As an example, the starting temperatures we generally use for nonacclimated ( 4  C), and cold-acclimated ( 7  C) WT (Col-0) plants are shown. Once the starting temperatures are reached, they are sequentially decreased ( 1  C/h) until reaching the last temperatures tested for the freezing–survival curves ( 9  C and 12  C for nonacclimated and cold acclimated plants, respectively, in our standard experiments). Every hour, a pot of WT plants and another one of mutants are transferred to 4  C under long-day light regime for slow thawing for 1 day, and subsequent survival evaluation after 1 week at 20  C

freezing–survival curves, in which the survival rates of mutant and WT plants to different freezing temperatures are determined (see Note 20). 6. To perform the freezing–survival curves, pots containing WT and mutant plants, growing as described in Subheading 3.3, are transferred to the growth chamber for freezing assays (see Notes 6, 16, and 21). 7. Program the freezing chamber with the protocol described in Fig. 3, which allows to study the freezing tolerance to different temperatures in a single assay. In this protocol, one pot of WT plants and another one of mutants are removed from the chamber once it reaches each freezing temperature that will be used to make the freezing–survival curve. Transfer pots to a growth chamber set at 4  C and let them thaw slowly overnight. Then, allow them to recover at 20–22  C for 1 additional week under long-day light regime.

Screening for Freezing Tolerant and Sensitive Mutants in Arabidopsis

3.5 Molecular Identification of Selected Mutations

93

The number of mutations that a selected homozygous mutant line contains in its genome varies depending on the method employed for seed mutagenesis. Before proceeding to the molecular identification of the mutations, mutant lines should be backcrossed to WT several times in order to clean up their genetic background of mutations not related to the freezing phenotype of interest. Before the development of next-generation sequencing (NGS) technologies, the molecular identification of EMS- or FN-produced mutations causing a phenotype of interest in Arabidopsis was performed through a map-based cloning approach. This method is based in crossing a mutant plant of interest with a WT of a different accession. Subsequent polymorphism analysis determines the chromosome region where the mutation is located and, finally, by complementation experiments, the mutagenized locus is identified [13]. Unfortunately, the enormous amount of time required to identify a mutation causing a phenotype of interest by this method constitutes an important obstacle when planning a forward genetic screening. Recently, NGS technologies have been applied for the rapid and precise molecular identification of mutations in different Arabidopsis ecotypes, a method named mappingby-sequencing. For instance, Austin and colleagues [14] have described a rapid and robust method for mapping mutations, independent of their chromosomal location, by sequencing only a small pooled M2 population. The protocol was able to identify a highly restricted region containing very few SNPs that can be easily validated using standard reverse genetic techniques. On the other hand, Uchida et al. [15] have been able to identify a mutated locus in a nonreference Arabidopsis accession, that is, whose genome is not publicly available, by only one round of genome sequencing. Nowadays, different publicly available web-based platforms (i.e., NGM [http://bar.utoronto.ca/ngm/] or GALAXY [https://usegalaxy.org/]) and pipelines (i.e., SIMPLE described by Wachsman et al. [16]) provide appropriate mapping tools and all the information necessary to perform a full analysis of the NGS data. When screening mutant collections generated by T-DNA insertions, identifying the site of insertion in the genome is commonly performed using an adapter ligation-mediated PCR protocol [17]. This method consists of three steps and takes about 3 weeks to be completed. First, an adapter is ligated to genomic DNA after digestion with a restriction enzyme. Then, by using specific primers to the adapter and T-DNA, the T-DNA/genomic DNA junction is amplified by PCR. Finally, sequencing the T-DNA/genomic junction allows for mapping the T-DNA location in the genome.

94

4

Carlos Perea-Resa et al.

Notes 1. M2 families by pedigreeing are the recommended material to screen for sensitive mutants. Since sensitive mutants will not survive the screening, you must ensure a high (>200) number of mutant seeds for each family. When screening for tolerant mutants, M2 pools are the material of choice. Generation of M2 families and M2 pools has been reported in detail [4]. 2. Seeds can also be sterilized with bleach as described [18]. 3. Wrap paper pieces with aluminum foil. For a correct sterilization, do not autoclave more than ten paper pieces together. Other supports, such as a nylon mesh, can also be used. 4. The seed number per plate depends on the germination rate and on the type of screening that is going to be performed. When screening for freezing tolerant mutants, a high number of pooled M2 seeds can be plated (~300 seeds/Ø 15 cm plate). When the screening is performed to identify freezing sensitive mutants, the number of seeds plated from each M2 family should allow for establishing a significant sensitive/tolerant segregation (~100 seeds/Ø 15 cm plate). 5. For evaluation of freezing tolerance after cold acclimation, plates should be transferred to a growth chamber set at 4  C with cool-white light (50 μmol/m2/s) for 5 days under longday conditions before freezing, to ensure the full development of the adaptive response. 6. We strongly recommend transferring seedlings or plants to the appropriate growth chamber for the freezing assay always at the same time of the day, since the expression of several coldregulated genes involved in cold acclimation is subjected to circadian regulation [19, 20]. 7. Appropriate freezing temperatures to evaluate the tolerance of nonacclimated or cold acclimated seedlings depends on the type of screening to be performed (searching for freezing tolerant or sensitive mutants), and should be previously established in each case using WT seedlings and previously reported tolerant and/or sensitive mutants that will act as positive controls. If searching for freezing tolerant mutants, the highest temperature that produces 0% surviving seedlings should be used. On the contrary, when screening for freezing sensitive mutants, the lowest temperature that allows for 100% seedling survival is the convenient one. The time that seedlings should be exposed to the appropriate freezing temperatures must be determined at the same time as freezing temperatures. Different temperatures are used for the evaluation of nonacclimated and cold acclimated seedlings, the latter always requiring lower

Screening for Freezing Tolerant and Sensitive Mutants in Arabidopsis

95

temperatures (~2–3  C). Optimally, freezing temperatures must be gradually reached ( 1  C/30 min) starting from 4  C (Fig. 1). 8. During the freezing assay, it is critical that the temperature inside the chamber is homogeneous in such a way that all seedlings are exposed to the same conditions. At the beginning of the screening, we recommend to perform a temperature map of the chamber to verify such homogeneity. 9. Ice chips are ice nucleation sites that favor freezing homogeneity in the GM medium and of all seedlings of the plate. 10. The paper containing the seedlings should be transferred to new plates since freezing temperatures depolymerize the growth medium. 11. Amphotericin B is a polyene antifungal drug that helps to minimize plate contaminations. 12. In our experience, 1 week of recovery is enough to establish whether seedlings have survived the freezing treatment. Particularly, we consider that a plant has survived the treatment when it develops new leaves. Fungal contaminations usually appear when longer recovery times are allowed. 13. We recommend selecting M2 lines with a 3:1 or 1:3 sensitive– tolerant segregation, which would indicate that the sensitive phenotype is produced by a single dominant or recessive mutation, respectively. Phenotypes produced by single mutations are preferred because these mutations are easy to map and identify molecularly. 14. In our hands, clay pots allow for water transpiration and work better than plastic pots. 15. The number of seeds per pot depends on the germination rate and on the type of screening that is going to be performed. When screening for freezing tolerant mutants, a high number of pooled M2 seeds can be sown (~60 seeds/Ø 10 cm pot). When the screening is performed to identify freezing sensitive mutants, the number of seeds sown from each M2 family should allow to establish a significant sensitive/tolerant segregation (~40 seeds /Ø 10 cm pot). In all cases, it is important to remove all seedlings showing a growth delay or developmental defects, or that are too close to other seedlings. 16. When screening for mutants with altered freezing tolerance after cold acclimation, plants should be previously exposed to 4  C with cool-white light (50 μmol/m2/s) during 1 week to ensure the adaptive response. 17. When freezing plants grown on soil, the addition of ice chips is not necessary because freezing occurs very homogeneously on the surface of the pot.

96

Carlos Perea-Resa et al.

18. Other segregations will suggest that the mutant phenotype originates from more than one mutation. 19. When screening for freezing tolerant mutants by pooling, mutant seedlings or plants containing the same mutation may be selected. Allelism tests should then be performed between the identified mutant lines. 20. For tolerant mutants the curve should start at temperatures allowing for 100% survival of WT plants and finish at temperatures in which no mutant plant survives. In the case of sensitive genotypes, the curve should start at temperatures allowing for 100% survival of mutant plants and finish at temperatures ensuring 0% survival of WT plants. These recommendations are applicable to both nonacclimated and cold acclimated plants. 21. A pot of each genotype analyzed (i.e., WT and mutant) for each temperature tested is used to generate the freezing–survival curves. References 1. Levitt J (1980) Responses of plants to environmental stresses: chilling, freezing and high temperatures stresses, 2nd edn. Academic, New York 2. Steponkus PL, Uemura M, Joseph RA et al (1998) Mode of action of the COR15a gene on the freezing tolerance of Arabidopsis thaliana. Proc Natl Acad Sci U S A 95:14570–14575 3. Guy CL (1990) Cold acclimation and freezing stress tolerance: role of protein metabolism. Annu Rev Plant Physiol Plant Mol Biol 41:187–223 4. Lightner J, Caspar T (1998) Seed mutagenesis of Arabidopsis. In: Martı´nez-Zapater JM, Salinas J (eds) Methods in molecular biology, vol 82. Humana, Totowa, NJ, pp 91–103 5. Kim YS, Schumaker KS, Zhu JK (2006) EMS mutagenesis of Arabidopsis. In: Salinas J, Sanchez-Serrano JJ (eds) Methods in molecular biology, vol 323. Humana, Totowa, NJ, pp 101–103 6. Weigel D, Glazebrook J (2006) EMS mutagenesis of Arabidopsis seed. CSH Protoc 28. https://doi.org/10.1101/pdb.prot4621 7. Warren G, McKown R, Marin AL et al (1996) Isolation of mutations affecting the development of freezing tolerance in Arabidopsis thaliana (L.) Heynh. Plant Physiol 111:1011–1019 8. McKown R, Kuroki G, Warren G (1996) Cold responses of Arabidopsis mutants impaired in freezing tolerance. J Exp Bot 47:1919–1925

9. Xin Z, Browse J (1998) eskimo1 mutants of Arabidopsis are constitutively freezing-tolerant. Proc Natl Acad Sci U S A 95:7799–7804 10. Llorente F, Oliveros JC, Martı´nez-Zapater JM et al (2000) A freezing-sensitive mutant of Arabidopsis, frs1, is a new aba3 allele. Planta 211:648–655 11. Koornneef M, Dellaert LWM, van der Veen JH (1982) EMS- and radiation-induced mutation frequencies at individual loci in Arabidopsis thaliana (L.) Heynh. Mutat Res 93:109–123 12. Alonso JM, Stepanova AN (2003) T-DNA mutagenesis in Arabidopsis. In: Grotewold E (ed) Methods in molecular biology, vol 236. Humana, Totowa, NJ, pp 177–188 13. Jander G (2006) Gene identification and cloning by molecular marker mapping. In: Salinas J, Sanchez-Serrano JJ (eds) Methods in molecular biology, vol 323. Humana, Totowa, NJ, pp 115–126 14. Austin RS, Vidaurre D, Stamatiou G et al (2011) Next-generation mapping of Arabidopsis genes. Plant J 67:715–725 15. Uchida N, Sakamoto T, Kurata T et al (2011) Identification of EMS-induced causal mutations in a non-reference Arabidopsis thaliana accession by whole genome sequencing. Plant Cell Physiol 52:716–722 16. Wachsman G, Modliszewski JL, Valdes M, Benfey PN (2017) A SIMPLE pipeline for mapping point mutations. Plant Physiol 174:1307–1313

Screening for Freezing Tolerant and Sensitive Mutants in Arabidopsis 17. O’Malley RC, Alonso JM, Kim CJ et al (2007) An adapter ligation-mediated PCR method for high-throughput mapping of T-DNA inserts in the Arabidopsis genome. Nat Protoc 2:2910–2917 18. McCourt P, Keith K (1998) Sterile techniques in Arabidopsis, Methods in molecular biology, vol 82. Humana, Totowa, NJ, pp 13–17

97

19. Mikkelsen MD, Thomashow MF (2009) A role for circadian evening elements in coldregulated gene expression in Arabidopsis. Plant J 60:328–339 20. Dong MA, Farre´ EM, Thomashow MF (2011) Circadian clock-associated 1 and late elongated hypocotyl regulate expression of the C-repeat binding factor (CBF) pathway in Arabidopsis. Proc Natl Acad Sci U S A 108:7241–7246

Chapter 9 Cryo-Scanning Electron Microscopy to Study the Freezing Behavior of Plant Tissues Seizo Fujikawa and Keita Endoh Abstract A cryo-scanning electron microscope (cryo-SEM) is a valuable tool for observing bulk frozen samples to monitor freezing responses of plant tissues and cells. Here, the essential processes of a cryo-SEM to observe freezing behaviors of plant tissue cells are described. Key words Cryo-scanning electron microscope (cryo-SEM), Ice crystal, Extracellular freezing, Intracellular freezing, Cryo-fixation, Recrystallization, Cooling rate

1

Introduction A cryo-scanning electron microscope (cryo-SEM) or a low-temperature SEM has been used to observe bulk biological samples under a freezing condition. Since the early reports on the development of a simple cryo-SEM, in which the SEM was equipped with only a cold stage in the SEM column [1], the instruments have been improved in a step-by-step fashion [2–4]. Various cryo-SEMs with their own unique combination of features are now available. A cryo-SEM has been used to observe hydrated structures of biological materials including plant tissues, in which water in samples is kept from conversion to ice by cryo-fixation using very rapid freezing (for recent review, see ref. 5). A cryo-SEM has also been used to observe the distribution of water in plant tissues [6–11] as well as the distribution of contents dissolved in water [12]. Moreover, a cryo-SEM has been used to analyze effects of the freezedrying process in relation to morphological changes of plant tissues [13]. A cryo-SEM is particularly useful for observing responses in plant tissue cells to freezing. Studies using a cryo-SEM have shown interaction between extracellular ice crystals and cells in the fruiting body of mushrooms [14, 15], in leaves of cereals [16, 17], in leaves

Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_9, © Springer Science+Business Media, LLC, part of Springer Nature 2020

99

100

Seizo Fujikawa and Keita Endoh

of Arabidopsis [18], in evergreen leaves of trees [19–21], in leaves of freezing-sensitive plant species [22, 23], and in bark tissues of fruit trees [24]. Direct observation of frozen tissue cells by a cryoSEM has provided information on freezing responses of xylem parenchyma cells in several tree species which adapt to subfreezing temperatures by deep supercooling [25–33] and information on freezing behavior of tissue cells in dormant buds in trees that adapt to subfreezing temperatures by extraorgan freezing [34, 35]. Here, we introduce our manuals in addition to general manuals for observing freezing responses of plant tissue cells by cryo-SEM. Additionally, we recommend the use of freeze-fracture replica electron microscopy in combination with cryo-SEM for understanding freezing responses of plant tissue cells in more detail (for the manual of diverse freeze-replica technique, see ref. 36).

2 2.1

Materials Main Apparatus

1. Cryo-SEM: Although there are many different types of commercially available cryo-SEM, a cryo-SEM generally consists of a specimen preparation chamber and SEM column equipped with cold stages. A preevacuation chamber is connected to the specimen preparation chamber in order to transfer frozen samples from outside to the specimen preparation chamber. The specimen preparation chamber is equipped with not only a cold stage but also a cold knife for fracturing frozen samples and a metal-coating system for coating freeze-fracture faces to facilitate radiation of secondary electrons as well as to inhibit electric charging. The specimen preparation chamber is directly connected to the SEM column with another cold stage on which frozen samples are observed with cover by a cold trap for decontamination of samples. In different types of cryoSEM, the cold stage is cooled by a connected copper braid cooled by liquid nitrogen (LN2) [37], by a piped system for circulating LN2 [3] or by a Joule–Thomson refrigerator principle [38]. Metal coating is done by resistance heating [2, 39], sputter coating [3, 37], or using electron-beam guns [40, 41]. 2. Specimen carrier (standard apparatus of cryo-SEM): The carrier is used for transferring frozen samples from outside to inside the cryo-SEM and within the cryo-SEM by a rod (Fig. 1). A frozen sample is firmly fixed to the specimen carrier under LN2. The specimen carrier with the sample is transferred to the cold stage in the specimen preparation chamber through the preevacuation chamber, and after treatment in the preparation chamber, the sample is transferred to a cold stage in the SEM column. The specimen carrier with the sample is

Cryo-Scanning Electron Microscopy

101

Fig. 1 A specimen carrier (right) and sample holders (left) with cavities of different sizes. Sample holders are made of aluminum and are 5 mm in diameter and 5 mm in height with cavities of different sizes (depth and diameter) in the center in which samples are put. The specimen carrier has a hole in which a specimen holder is fixed by a screw under LN2

temporarily fixed on both cold stages, and sophisticated temperature control of the samples is done through the carrier. 3. Sample holders: Samples to be frozen and observed by the cryo-SEM are put in sample holders. The samples in holders can be firmly fixed, without physical stresses to frozen samples themselves, by using a screw to a hole of the specimen carrier under LN2. Since samples have various sizes and shapes, diverse kinds of sample holders are convenient. We usually have more than 100 sample holders made of aluminum with different sizes of holes in which samples are put (Fig. 1). Different sizes of sample holders are also useful. When size of the sample holder is changed, the size of holes in the specimen carrier must also be changed. 4. LN2 Dewar for fixing samples to the specimen carrier under LN2: Proper size of Dewar filled with LN2 is used to fix frozen samples into holes in the specimen carrier under LN2. Most types of cryo-SEM are equipped with such a Dewar as a standard apparatus. 5. LN2 container for stock frozen samples: A large container is filled with LN2 to stock frozen samples until use. We use 20-L containers that have several partitions for keeping samples. When stocked samples are used for cryo-SEM observation, they are transferred into a small Dewar vessel (1-L) filled with LN2. 6. Programmed freezer: In addition to cryo-SEM related apparatus, a programmed freezer is necessary to perform the described experiments. The cooling rate of samples is controlled from 0
C to temperatures lower than 50
C. There are many types of programmed freezer. We use stock freezers

102

Seizo Fujikawa and Keita Endoh

for cooling down to 60
C. The stock freezer is equipped with a temperature control heater operated by a personal computer. 2.2

Materials

1. Use fresh plants under control (warm) growth conditions, under cold-acclimation conditions or in the dormant state. 2. In the case of dormant trees, stocked samples can also be used. For stock samples in dormant trees, remove 1–5-year-old twigs with length of about 30–50 cm from standing trees during winter, put about 20 twigs in a polystyrene bag together with a sufficient amount of snow, close the bag and store in a cold room kept at 10
C for 6 months at maximum until use (see Note 1).

3

Methods

3.1 Preparation of Samples Before Freezing

In order to firmly fix frozen samples during fracturing and cryoSEM observation, put fresh samples in a hole of the sample holder before freezing (see Note 2). During sample preparation, maintain the desired temperatures depending on samples. Do all the processes quickly to maintain the freshness of samples. 1. Write appropriate numbers by oil-based ink (see Note 3) indicating sample conditions, including name of the plant species, name of tissues, sampling conditions, and freezing conditions, on the bottom or side walls of the sample holders. 2. Remove tissues from fresh or stocked plants. 3. Cut the removed tissues with a sharp knife (see Note 4) to appropriate sizes that can be placed in a hole of the specimen holder with slight protrusion from the hole (see Note 5). 4. Put a small amount of distilled water in each of the holes in sample holders using a pipette (see Note 6). 5. Using forceps, put samples in a hole of the sample holder in contact with water in the hole. Arrange the samples so that small areas protrude from the top surface of the sample holder. Since the protruding parts are cut (fractured) with a cold knife that moves in parallel with the top surface of sample holder, arrange the samples to make the fractured plane correspond to the desired plane to be observed. 6. Prepare all of the samples (hereafter, samples in holders will be simply called “samples”) just before freezing.

3.2 Cryo-Fixation of Reference Samples

For comparison of structures with those after controlled-freezing, structures before freezing need to be observed for reference samples. Although such reference structures can be provided by various

Cryo-Scanning Electron Microscopy

103

kinds of cryo-fixation using rapid freezing, cryo-fixation by plunging samples into cooled Freon 22 is recommended for the study of freezing behavior of plant tissue cells (see Note 7). The process of cryo-fixation by plunge freezing with cooled Freon 22 is described below. 1. Fill a small Dewar vessel (1-L) with LN2. 2. Insert a copper-made well into LN2 in the Dewar vessel by grasping the well firmly with crucible tongs. Avoid getting any LN2 in the well. 3. Inject Freon 22 slowly into the cooled well in LN2. 4. When about half of the well is filled with Freon 22 (about 5 mL), stop injection. 5. Freon 22 in the well is soon frozen. 6. Insert metal sticks into the frozen Freon 22 in order to make a mixture of solid and liquid Freon 22 (160
C). 7. Quickly plunge a fresh sample (provided by the process described under Subheading 3.1) by grasping the sample holder with forceps and putting it into the liquid part of Freon 22 for at least 5 s. 8. Quickly transfer the sample to a basket filled in the LN2 and release it from the forceps. 9. Store samples in a LN2 stock container until use (see Note 8). 3.3 Controlled Freezing of Samples

In order to understand freezing responses of plant tissue cells, it is necessary to freeze samples at controlled cooling rates (generally by slow cooling to mimic temperature reduction in the field) to the desired temperature, and finally the frozen samples are cryo-fixed to keep conditions at the final freezing condition. Here, the general processes for controlled slow freezing are described. 1. Put several fresh samples (provided as described under Subheading 3.1) in a petri dish and cover with a lid at 4
C or other desired nonfreezing temperature (see Note 9). 2. Previously cool an empty petri dish in a programmed freezer kept at the starting temperature of freezing for samples and put a small amount of ice chips (originating from frost) obtained by scratching from side walls of the programmed freezer with cooled forceps in the cooled petri dish. 3. Transfer the petri dish with samples to a programmed freezer kept at 3
C or other desired freezing temperature depending on the purpose (see Note 10). 4. Wait for more than 30 min to obtain a temperature equilibrium of samples.

104

Seizo Fujikawa and Keita Endoh

5. After a temperature equilibrium at 3
C or desired freezing temperature has been reached, open the lid of the petri dish and put ice chips on the samples using cooled forceps as above. 6. Confirm freezing of samples (see Note 11). 7. Keep samples at 3
C or desired freezing temperature for a further 30 min or more to reach temperature equilibrium after freezing. 8. Start cooling at a controlled rate depending on the purpose (see Note 12) to the desired final freezing temperature (see Note 13). 9. After the desired final freezing temperature has been reached, cryo-fix samples immediately or after the desired time at the final temperature by plunge-freezing with cooled Freon 22 (by the process described under Subheading 3.2) (see Note 14). 10. Store samples in a LN2 stock container until use. 3.4 Cryo-Scanning Electron Microscopy

Frozen samples including both reference and controlled-freezing samples are observed by a cryo-SEM. Although each cryo-SEM has a different construction, common processes for cryo-scanning electron microscopy are described here. For details of the methods for specimen preparation and observation, see the procedures described in the manufacturer’s instruction manual for each cryoSEM. 1. With the SEM in operation (at fully high vacuum), cool a cold stage and a cold trap in the SEM column as well as a cold stage and a cold knife in the preparation chamber. 2. Wait for about 1 h to achieve full cooling (lower than 160
C) of the abovementioned cryo-SEM apparatuses. 3. Set the temperature of a cold stage in the preevacuation chamber to 105
C. Set maximum cooling in other apparatuses. 4. Transfer stocked frozen samples in the LN2 stock container into a small Dewar vessel filled with LN2. 5. Cool a specimen carrier by dipping into LN2 in a LN2 Dewar until bubbling has stopped (about 30 s). 6. Transfer frozen samples to the LN2 Dewar together with the cooled specimen carrier in LN2. 7. Put a sample in a hole of the specimen carrier using forceps and firmly fix samples in the hole of the carrier by a screw using a screwdriver under LN2. 8. Cover samples with the cold trap for decontamination, if there is one, under LN2.

Cryo-Scanning Electron Microscopy

105

9. Quickly transfer the covered specimen carrier with the sample to the preevacuation chamber. 10. Quickly evacuate the preevacuation chamber. 11. After completion of full evacuation in the preevacuation chamber, open the gate to the preparation chamber, transfer the specimen carrier with the sample, and fix the specimen carrier on a cold stage of the preparation chamber kept at 105
C (see Note 15). 12. After the temperature has reached equilibrium at 105
C, confirm removal of frost that may have covered samples by the naked eye or by using binoculars (see Note 16). 13. Fracture protruding areas of samples by moving a cold knife in parallel with the top surface of the sample holder (see Note 17). 14. After fracturing the sample, cover the fractured plane with a cold knife kept at 160
C as a cold trap for decontamination. 15. Keep fractured samples for a few minutes at 105
C for making slight etching (see Note 18). 16. To stop further etching, cool the cold stage to a temperature lower than 120
C (see Note 19). 17. Remove the cold knife covering samples. 18. Coat the fractured face by metal evaporation. Depending on the apparatus, refer to the manufacturer’s instructions for appropriate coating (see Note 20). 19. Transfer the carrier with samples from the cold stage of the preparation chamber to a cold stage in the SEM column kept at 160
C, fix the carrier, and cover the periphery of samples by a cold trap. 20. Observe fracture planes by using secondary emissions by the SEM (see Note 21). 21. Take photographs of desired areas following the procedure described in the manufacturer’s instruction manual. 3.5 Methods for Analyzing Effects of Freezing

Several examples of specific methods to determine effects of freezing on plant tissue cells are described. In most cases, prepare samples from at least three separate freezing experiments with more than three samples in each separate experiment and observe more than 100 cells in total from more than six different samples in order to determine the tendency of freezing-induced changes.

3.5.1 Detection of Extracellular Freezing

The usual method for observing structural changes induced by extracellular freezing under equilibrium freezing conditions is described using wheat leaves under control (warm) growth conditions (Fig. 2).

106

Seizo Fujikawa and Keita Endoh

Fig. 2 Cryo-SEM photographs showing parts of freeze-fractured wheat leaves (a) in a reference sample cryo-fixed at room temperature and (b) in a sample slowly frozen from 3 to 10
C at a cooling rate of 0.2
C/min. Equilibrium slow freezing to 10
C caused distinct shrinkage of cells (arrows) with occupation of extracellular spaces by large extracellular ice crystals (EI). Bars ¼ 10 μm

1. Prepare samples at room temperature by the process described under Subheading 3.1 and cryo-fix samples at room temperature as reference samples by the process described in Subheading 3.2. 2. For equilibrium freezing of samples, provide samples prepared at room temperature by the process described in Subheading 3.1 and slowly freeze the samples by the process described in Subheading 3.3, in which samples are kept at 3
C for 30 min, inoculated with ice, kept at 3
C for 30 min, cooled at a rate of 0.2
C/min to 10
C, and cryo-fixed immediately after reaching 10
C. 3. Take photographs of cryo-fixed reference and cryo-fixed slowly frozen samples using a cryo-SEM as described in Subheading 3.4. 4. Compare the structures of a reference sample (Fig. 2a) and a sample slowly frozen to 10
C (Fig. 2b) (see Note 22).

Cryo-Scanning Electron Microscopy

107

Fig. 3 Cryo-SEM photographs showing temperature limit of deep supercooling in xylem parenchyma cells of birch harvested in summer. (a) Reference samples cryo-fixed at room temperature. (b) Cryo-fixed samples after very slow cooling to 15
C at a rate of 5
C/day. (c) Cryo-fixed samples after very slow cooling to 20
C at a rate of 5
C/day. While very small intracellular ice crystals (which are difficult to detect or are seen as numerous small holes due to sublimation of ice) are produced in reference samples (a) and samples slowly frozen to 15
C (b), large intracellular ice crystals (arrows) are produced in cells slowly frozen to 20
C (c). The sizes of intracellular ice crystals differ depending on whether they are produced by rapid freezing due to cryo-fixation of liquid water (a, b) or produced during very slow cooling (5
C/day) (c). Distinct differences in sizes of intracellular ice crystals indicate temperature limit of supercooling between 15 and 20
C. Bars ¼ 2 μm 3.5.2 Detection of Temperature Limit of Deep Supercooling

The method for examining the temperature limit of deep supercooling is described using xylem parenchyma cells in mulberry trees harvested in summer (Fig. 3). 1. Prepare samples at room temperature by the process described in Subheading 3.1 and cryo-fix samples from room temperature as described in Subheading 3.2. 2. To prepare frozen samples, put samples in a petri dish in a programmed freezer kept at 5
C overnight and then lower the temperature very slowly in a stepwise manner by 5
C/day to 50
C (see Note 23). 3. At each 5
C decline in temperature, cryo-fix samples by the process described in Subheading 3.2. 4. Take photographs of cryo-fixed reference and treated samples cooled from 5 to 40
C by the cryo-SEM (see Subheading 3.4). 5. Compare sizes of intracellular ice crystals (see Note 24) in the reference sample (Fig. 3a) and samples very slowly cooled to different freezing temperatures (Fig. 3b, c).

108

Seizo Fujikawa and Keita Endoh

Fig. 4 Cryo-SEM photographs showing a recrystallization experiment in tissue cells of dormant buds of katsura tree in order to determine the presence or absence of freezable water after slow freezing to desired temperatures. (a–c) Cells in inner scales. (d–f) Cells in flower primordia. (a, d) Reference samples cryofixed at 4
C. (b, e) Cryo-fixed samples after very slow freezing from 3 to 30
C at a cooling rate of 5
C/ day. (c, f) Cryo-fixed samples after a recrystallization experiment in which samples in (b, e) were rewarmed at 20
C. Formation of large intracellular ice crystals (asterisk) in flower primordia cells after the recrystallization experiment indicates the presence of freezable water even after slow cooling to 30
C (f), whereas the absence of such intracellular change in cells in inner scales indicates complete crystallization of freezable water by slow freezing to 30
C (c). Freezable water remaining at 30
C after very slow cooling produced very small intracellular ice crystals by cryo-fixation and rewarming of such cells produced large intracellular ice crystals as a result of recrystallization. Bars ¼ 1 μm 3.5.3 Detection of Freezable Water Under the Condition of Slow Freezing by a Recrystallization Experiment (Fig. 4)

In order to examine whether freezable water has remained in plant cells under the condition of slow freezing to the desired temperature, the method for a recrystallization experiment is described using tissue cells of dormant buds in katsura tree [31, 34]. 1. Prepare samples at 4
C as described in Subheading 3.1 and cryo-fix reference samples at 4
C (Subheading 3.2).

Cryo-Scanning Electron Microscopy

109

2. To prepare slowly frozen samples, prepare samples at 4
C as described in Subheading 3.1 and slowly freeze the samples from 3
C at a cooling rate of 5
C/day to 30
C (Subheading 3.3) and then cryo-fix the samples at 30
C as in Subheading 3.2. 3. For the recrystallization experiment, transfer the above samples which have been slowly frozen to 30
C and cryo-fixed to a petri dish in a freezer kept at 20
C, keep overnight to allow recrystallization, and again cryo-fix (Subheading 3.2). 4. Take photographs of cryo-fixed reference samples, cryo-fixed samples slowly frozen to 30
C and recrystallization experiment samples using a cryo-SEM (Subheading 3.4). 5. Compare cell shapes and especially the sizes of intracellular ice crystals (see Note 24) among reference samples (Fig. 4a, d), samples slowly frozen to 30
C (Fig. 4b, e) and recrystallization experiment samples (Fig. 4c, f). 3.5.4 Confirmation of Equilibrium Freezing

The method to determine whether the cooling rates or freezing conditions used produced equilibrium dehydration is described using nonacclimated leaves of Arabidopsis thaliana (Fig. 5) [18]. 1. Prepare reference samples at room temperature (Subheading 3.1) and cryo-fix the reference samples at room temperature (Subheading 3.2). 2. To prepare one type of controlled-freezing samples, prepare samples at room temperature (Subheading 3.1), keep the

Fig. 5 Cryo-SEM photographs showing equilibrium freezing of Arabidopsis leaves. (a) Reference sample cryofixed at room temperature. (b) Cryo-fixed sample after freezing at 2
C for 1 h. (c) Cryo-fixed sample after freezing at 2
C for 3 days. Compared with the reference samples (a), samples frozen at 2
C exhibited shrinkage by dehydration and far smaller intracellular ice crystals as a result of cryo-fixation (b, c). The sizes of intracellular ice crystals (shown between arrowheads in the maximum diameter) are similar in cryo-fixed samples after freezing at 2
C for 1 h (b) and 3 days (c). Since the size of intracellular ice crystals depends on the degree of cellular hydration, the result indicates that equilibrium dehydration occurred with freezing at 2
C for 1 h. EI extracellular ice. Bars ¼ 10 μm. (Reproduced from Nagao et al. [18] with permission from Springer)

110

Seizo Fujikawa and Keita Endoh

samples at 2
C for 30 min, inoculate with ice, keep at 2
C for 1 h, and immediately cryo-fix as described in Subheading 3.2. 3. For a second type of controlled-freezing samples, prepare samples at room temperature and then keep them at 2
C for 30 min as above. Inoculate the samples with ice and keep at 2
C for 3 days, then cryo-fix (Subheading 3.2; see Note 25). 4. Take photographs of cryo-fixed reference samples and cryofixed samples at 2
C that have been kept for 1 h and 3 days. 5. Compare sizes of intracellular ice crystals in reference samples (Fig. 5a) and samples kept at 2
C for 1 h (Fig. 5b) and for 3 days (Fig. 5c). 3.5.5 Evaluation of the Cell Wall as a Barrier Against Penetration of Extracellular Ice Crystals (In the Case of Chilling-Sensitive Plant Tissue Cells)

The method to estimate the effect of cell walls as a barrier against extracellular ice crystals is described using chilling-sensitive leaves of Saintpaulia grotei Engl. (Fig. 6) [23]. 1. Prepare reference samples from room temperature (Subheading 3.1) and cryo-fix the samples at room temperature (Subheading 3.2). 2. Supercooled samples are prepared as described in Subheading 3.1, put in a programmed freezer kept at 2
C for 1 h under the condition of supercooling (see Note 26) and then cryofixed at 2
C (Subheading 3.2). 3. Frozen samples are prepared as described in Subheading 3.1, put into a petri dish in a programmed freezer kept at 2
C for 30 min, frozen by ice inoculation, kept for a further 30 min at 2
C, and cryo-fix at 2
C (Subheading 3.2).

Fig. 6 Cryo-SEM photographs showing the effect of the cell wall against the presence of neighboring extracellular ice crystals in chilling-sensitive Saintpaulia grotei Engl. leaves. (a) Reference samples cryofixed at room temperature. (b) Cryo-fixed samples with supercooling at 2
C for 1 h. (c) Cryo-fixed samples frozen at 2
C for 1 h. While reference cells (a) and supercooled cells (b) exhibited small intracellular ice crystals (shown between arrows in the minimum diameter), frozen samples produced large intracellular ice crystals due to penetration of extracellular ice crystals through the cell walls. EI extracellular ice. Bar ¼ 20 μm. (Reproduced from Yamada et al. [23] with permission from Springer)

Cryo-Scanning Electron Microscopy

111

4. Take photographs of cryo-fixed reference samples, supercooled samples, and samples frozen at 2
C using a cryo-SEM (Subheading 3.4). 5. Compare intracellular ice crystal sizes in the reference sample (Fig. 6a), sample supercooled at 2
C (Fig. 6b), and sample frozen at 2
C. 3.5.6 Evaluation of the Cell Wall as a Barrier Against Penetration of Extracellular Ice Crystals (In the Case of Chilling-Resistant Plant Tissue Cells)

The method to estimate the effect of cell walls as a barrier against extracellular ice crystals is described using cells with freezedamaged protoplasts in leaves of an orchid (Fig. 7) [23]. 1. Prepare samples at room temperature as described in Subheading 3.1. 2. For breaking protoplasts together with plasma membranes, freeze samples by direct immersion in LN2 for about 1 min

Fig. 7 Cryo-SEM photographs showing the effect of the cell wall against the presence of neighboring extracellular ice crystals in chilling-resistant orchid leaves. Before examination, the samples were freeze-thawed to destroy protoplasts. (a) Protoplast-broken cells cryo-fixed at room temperature. (b) Cryo-fixed protoplast-broken cells after freezing at 2
C for 30 min. Freezing at 2
C caused distinct cell shrinkage with small intracellular ice crystals by dehydration through cell walls, indicating that cell walls even after protoplast destruction act as a complete barrier against penetration of extracellular ice crystals. EI extracellular ice. Bar ¼ 20 μm. (Reproduced from Yamada et al. [23] with permission from Springer)

112

Seizo Fujikawa and Keita Endoh

and then thaw at room temperature for 30 min and cryo-fix (Subheading 3.2; see Note 27). 3. To produce frozen samples, put frozen-thawed samples in a petri dish (see Note 28), transfer to a programmed freezer kept at 2
C for 30 min, freeze by ice inoculation, keep for a further 30 min at 2
C, and cryo-fix (Subheading 3.2; see Note 29). 4. Take photographs of cryo-fixed freeze-thawed samples for reference and cryo-fixed freeze-thawed samples after further freezing at 2
C using a cryo-SEM (Subheading 3.4). 5. Compare cell shapes before (Fig. 7a) and after slow freezing at 2
C (Fig. 7b).

4

Notes 1. Add snow every month to prevent drying of twigs during the storage period. Long-term preservation of twigs exceeding 6 months may cause physiological changes. 2. Since loose fixation of frozen samples may result in drifting of samples under SEM observation, especially at higher magnification, close contact of samples in the sample holder is required. When samples are in the holder, physically firm fixation of the holders to the carrier holes by a screw is possible. If frozen plant tissues are directly fixed to the specimen carrier using a screw under LN2, strong physical stress by fixation of the samples may cause crushing of the samples. On the other hand, loose fixation of samples to the specimen carrier may produce drifting of samples during observation due to elimination of small ice crystals (originating from frost on surfaces of the samples) by sublimation during processes of cryo-SEM observation. 3. Do not use a magic marker with water-soluble ink, which may disappear after thawing. 4. Physical damage caused by cutting of samples may result in structural changes. Observation should be done in areas further away from the cutting. 5. If changes in survival of frozen samples are concomitantly measured in parallel with cryo-SEM observation, use the same sizes of samples in cryo-SEM observation and survival assays for comparison between them. Furthermore, if freezing responses of samples are measured by different methods, such as differential thermal analysis, also use the same sizes of samples for comparison.

Cryo-Scanning Electron Microscopy

113

6. Water is used to firmly fix samples to the holes of holders after freezing. If necessary, use an appropriate buffered solution instead of distilled water. For samples that need to be kept frozen for long periods (a few to several days), the samples should be completely embedded in water in order to prevent drying due to sublimation. For samples that need to avoid contact with water, use a small amount of glycerol for fixation of the samples under freezing to the sample holder instead of water. 7. Cryo-fixation by plunge-freezing with cooled Freon 22 (9000
C/s) produces very small intracellular ice crystals in many plant tissue cells, though comparatively large intracellular ice crystals are often detected in highly hydrated samples by cryo-SEM observation. However, for detection of freezing responses, it is necessary to use the same cryo-fixation method for reference and controlled-freezing samples. For cryo-fixation of samples previously frozen by controlled freezing, only plunge-freezing can be used. When reference samples without effects of ice crystals need to be obtained, other techniques including propane jet freezing [42], spray freezing [43], impact freezing [44], and high-pressure freezing [45] are useful. In these techniques, vitrification of sample water is possible, though this is restricted to very small volumes. See ref. 36 for manuals of these cryo-fixation techniques. 8. With a constant supply of LN2 in the stock container, frozen samples can be stored for more than 1 year. 9. Keep the temperature at 0–4
C for preparation of cold-acclimated samples or samples from dormant plants. For samples from growing plants, use room temperature for preparation. 10. Change the temperatures for the start of freezing depending on samples. For freeze-sensitive plant tissue cells, starting temperature for freezing of 1
C or 2
C is necessary because a lower temperature may cause intracellular freezing. In plant tissues with moderate freezing tolerance both before and after cold acclimation, a starting temperature for freezing of 3
C is used. In very cold hardy plant tissue cells, starting cooling at 5
C is possible. 11. Freezing of samples can be confirmed by the naked eye or by physical fixation of samples to the specimen holder by touching with cooled forceps. 12. For most samples, a cooling rate of 0.2
C/min is appropriate to achieve equilibrium freezing. In some samples, equilibrium freezing is obtained at far lower cooling rates. We sometime use a cooling rate of 5
C/day as the lowest cooling rate. All cooling rates can be programmed in a programmable freezer.

114

Seizo Fujikawa and Keita Endoh

13. By changing the final freezing temperature, the process of freezing in parallel with reduction of freezing temperature can be observed. 14. For cryo-fixation of samples after controlled freezing, put a small Dewar with cooled Freon 22 in a programmed freezer just before cryo-fixation, quickly grasp the controlled-frozen samples with cooled forceps, plunge into partially melted Freon 22 and transfer to an LN2 Dewar. When it is necessary to exactly maintain the final freezing temperature, we use a programmed freezer placed in a cold room in which the temperature is also controlled to be the same as that in the programmed freezer. 15. A sample temperature of 105
C under high vacuum in a cryo-SEM allows for slight sublimation from frozen samples in order to remove frost that has unintentionally covered samples during transfer of the samples from outside the cryo-SEM. 16. In a cryo-SEM, indicated temperatures for samples are generally measured in a specimen carrier. Keep samples for several minutes after temperature equilibrium at 105
C with visual confirmation of removal of frost from samples. 17. Fracture by one cut is favorable for reducing areas where the cold knife had contact with the fracture plane. Contact areas of knife should be avoided for observation. For cryo-SEM observation, making a macroscopically flat fracture plane is favorable for prevention of charging during cryo-SEM observation. If fractured pieces remain near fractured samples, remove such debris by reversing the specimen carrier using the rod of the specimen carrier, because the presence of such debris may become a source of contamination and charging. 18. For cryo-SEM observation of frozen samples, a slight degree of etching is recommended to make structures clear. Freezefracture without etching is not recommended for specimen preparation of a cryo-SEM due to high risk for contamination of fractured faces with frost. Vapor pressure of ice differs in relation to temperature [46]. When the temperature of samples has a higher vapor pressure than the vapor pressure of the SEM (i.e., vacuum in the SEM), etching by sublimation of ice occurs. On the other hand, when the temperature of samples has a lower vapor pressure than vacuum, no etching occurs. 19. In a vacuum of a general cryo-SEM, etching does not occur in samples kept at temperatures lower than 120
C. 20. In all types of metal coaters, even coating of fracture faces is necessary. Coating with three-dimensional rotation of samples is recommended (except for sputter coating).

Cryo-Scanning Electron Microscopy

115

21. Use appropriate accelerating voltages. Observation of frozen samples by a cryo-SEM should be completed within 2 h in order to avoid contamination, increased charging and temperature rise in samples. Images by other emissions such as reflection electrons are also used. 22. By changing the final freezing temperature under equilibrium slow freezing, the process of extracellular freezing can be observed. 23. Without ice inoculation, most samples kept at 5
C overnight are frozen. Even though a few samples may remain unfrozen at 5
C, transfer of samples to 10
C will ensure that all samples are frozen. In deep supercooling cells, such differences of starting temperature of freezing do not make a difference in the freezing behavior of cells. 24. Measure minimum or maximum diameter of ice (between eutectics) that appeared in fracture faces depending on the purpose. When minimum diameter of ice is measured, the difference among treatments becomes smaller (see Fig. 6), whereas when maximum diameter is measured, the difference becomes larger (see Fig. 5). 25. Although freezing at 2
C is shown here, it is also possible to cool samples to lower temperatures. 26. Under this condition at 2
C for 1 h, no samples are frozen without ice inoculation. 27. Freezing by LN2 causes intracellular freezing that results in distinct destruction of protoplasts together with plasma membranes. Thus, the effect of only cell walls as a barrier of penetration of extracellular ice crystals is analyzed. 28. Remove water from outside wall of the sample holders in freeze-thawed samples. The presence of water and resultant formation of ice crystals makes it difficult to fix the sample holders in the holes of specimen carriers. 29. Although freezing at 2
C is shown here, it is also possible to cool samples to lower temperatures.

Acknowledgments One of the authors (S.F.) sincerely appreciates the strong support by JEOL Co. Ltd. for improvement and development of a cryoSEM since he started his studies using a cryo-SEM in 1974. The authors also appreciate the excellent works by Mr. K. Shinbori, Institute of Low Temperature Science, Hokkaido University, for making many apparatuses of a cryo-SEM.

116

Seizo Fujikawa and Keita Endoh

References 1. Echlin P (1971) The examination of biological material at low temperatures. Scanning electron microscopy. IITRI, Chicago, pp 225–232 2. Nei T, Yotsumoto H, Hasegawa Y, Nagasawa Y (1973) Development of new cryo-unit attached to scanning electron microscope. J Electron Microsc 22:169–182 3. Sargent JA (1988) Low temperature scanning electron microscopy: advantages and application. Scanning Microsc 2:835–849 4. Fujikawa S, Suzuki T, Ishikawa T et al (1988) Continuous observation of frozen biological materials with cryo-scanning electron microscope and freeze-replica by a new cryo-system. J Electron Microsc 37:315–322 5. McCully ME, Canny MJ, Huang CX (2009) Cryo-scanning electron microscopy (CSEM) in the advancement of functional plant biology: morphological and anatomical applications. Funct Plant Biol 36:97–124 6. Utsumi Y, Sano Y, Ohtani J et al (1996) Seasonal changes in the distribution of water in the outer growth rings of Fraxinus mandshurica var. Japonica: a study by cryo-scanning electron microscopy. IAWA J 17:113–124 7. Utsumi Y, Sano Y, Fujikawa S et al (1998) Visualization of cavitated vessels in winter and refilled vessels in spring in diffuse-porous trees by cryo-scanning electron microscopy. Plant Physiol 117:1463–1471 8. Sano Y, Fujikawa S, Fukazawa K (1995) Detection and features of wetwood in Quercus mongolica var. grosseserrata. Trees 9:261–268 9. Johnson DM, Meinzer FC, Woodruff DR et al (2009) Leaf xylem embolism, detected acoustically and by cryo-SEM, corresponds to decreases in leaf hydraulic conductance in four evergreen species. Plant Cell Environ 32:828–836 10. Firon N, Nepi M, Pacini E (2012) Water status and associated processes mark critical stages in pollen development and functioning. Ann Bot 109:1201–1213 11. Umebayashi T, Morita T, Utsumi Y et al (2016) Spatial distribution of xylem embolism in the stems of Pinus thunbergii at the threshold of fatal drought stress. Tree Physiol 36:1210–1218 12. Cochard H, Bodet C, Ame´glio T et al (2000) Cryo-scanning electron microscopy observations of vessel content during transpiration in walnut petioles. Facts or artifacts? Plant Physiol 124:1191–1202 13. Nei T, Fujikawa S (1977) Freeze-drying process of biological specimens observed with a

scanning electron microscope. J Microsc 111:137–142 14. Fujikawa S, Miura K (1986) Plasma membrane ultrastructural changes caused by mechanical stress in the formation of extracellular ice as a primary cause of slow freezing injury in fruitbodies of basidiomycetes (Lyophyllum ulmarium (Fr.) Kuhner). Cryobiology 23:371–382 15. Fujikawa S (1990) Cryo-scanning electron microscope and freeze-replica study on the occurrence of slow freezing injury. J Electron Microsc 39:80–85 16. Pearce RS (1988) Extracellular ice and cell shape in frost-stressed cereal leaves: a low temperature scanning-electron-microscopy study. Planta 175:313–324 17. Pearce RS, Ashworth EN (1992) Cell shape and localization of ice in leaves of overwintering wheat during frost stress in the field. Planta 188:324–331 18. Nagao M, Arakawa K, Takezawa D et al (2008) Long- and short-term freezing induce different types of injury in Arabidopsis thaliana leaf cells. Planta 227:477–489 19. Ball MC, Canny MJ, Cheng X et al (2004) Structural changes in acclimated and unacclimated leaves during freezing and thawing. Funct Plant Biol 31:29–40 20. Roden JS, Canny MJ, Huang CX et al (2009) Frost tolerance and ice formation in Pinus radiata needles: ice management by the endodermis and transfusion tissues. Funct Plant Biol 36:180–189 21. Endoh K, Fujikawa S, Arakawa K (2012) Freezing behavior of cells in evergreen needle leaves of fir (Abies sachalinensis). Cryobiol Cryotechnol 58:125–134 22. Ashworth EN, Pearce RS (2002) Extracellular freezing in leaves of freezing-sensitive species. Planta 214:798–805 23. Yamada T, Kuroda K, Jitsuyama Y et al (2002) Roles of the plasma membrane and the cell wall in the responses of plant cells to freezing. Planta 215:770–778 24. Ashworth EN, Echlin P, Pearce RS et al (1988) Ice formation and tissue response in apple twigs. Plant Cell Environ 11:703–710 25. Fujikawa S, Kuroda K, Ohtani J (1996) Seasonal changes in the low-temperature behaviour of xylem ray parenchyma cells in red osier dogwood (Cornus sericea L.) with respect to extracellular freezing and supercooling. Micron 27:181–191 26. Kuroda K, Ohtani J, Fujikawa S (1997) Supercooling of xylem ray parenchyma cells in

Cryo-Scanning Electron Microscopy tropical and subtropical hardwood species. Trees 12:97–106 27. Fujikawa S, Kuroda K, Ohtani J (1997) Seasonal changes in dehydration tolerance of xylem ray parenchyma cells of Stylax obassia twigs that survive freezing temperatures by deep supercooling. Protoplasma 197:34–44 28. Kuroda K, Ohtani J, Kubota M et al (1999) Seasonal changes in the freezing behavior of xylem ray parenchyma cells in four boreal hardwood species. Cryobiology 38:81–88 29. Fujikawa S, Kuroda K, Jitsuyama Y et al (1999) Freezing behavior of xylem ray parenchyma cells in softwood species with differences in the organization of cell walls. Protoplasma 206:31–40 30. Fujikawa S, Kuroda K (2000) Cryo-scanning electron microscopic study on freezing behavior of xylem ray parenchyma cells in hardwood species. Micron 31:669–686 31. Kuroda K, Kasuga J, Arakawa K et al (2003) Xylem ray parenchyma cells in boreal hardwood species respond to subfreezing temperatures by deep supercooling that is accompanied by incomplete desiccation. Plant Physiol 131:736–744 32. Kasuga J, Arakawa K, Fujikawa S (2007) High accumulation of soluble sugars in deep supercooling Japanese white birch xylem parenchyma cells. New Phytol 174:569–579 33. Kasuga J, Endoh K, Yoshiba M et al (2013) Roles of cell walls and intracellular contents in supercooling capability of xylem parenchyma cells of boreal trees. Physiol Plant 148:25–35 34. Endoh K, Kasuga J, Arakawa K et al (2009) Cryo-scanning electron microscopic study on freezing behaviors of tissue cells in dormant buds of larch (Larix kaempferi). Cryobiology 59:214–222 35. Endoh K, Kuwabara C, Arakawa K et al (2014) Consideration of the reasons why dormant buds of trees have evolved extraorgan freezing as an adaptation for winter survival. Environ Exp Bot 106:52–59

117

36. Fujikawa S (1991) Freeze-fracture techniques. In: Harris JR (ed) Electron microscopy in biology: a practical approach. IRL Press, Oxford, pp 173–201 37. Robards AW, Crosby P (1979) A comprehensive freezing, fracturing and coating system for low temperature scanning electron microscopy. Scanning Electron Microsc 2:325–343 38. Pawley J, Norton JT (1978) A chamber attached to the SEM for fracturing and coating frozen biological samples. J Microsc 112:169–182 39. Bastacky J, Hook GR, Finch GL et al (1987) Low-temperature scanning electron microscopy of frozen hydrated mouse lung. Scanning 9:57–70 40. Fujikawa S, Suzuki T, Sakurai S (1990) Use of micromanipulator for continuous observation of frozen samples by cryo-scanning electron microscopy and freeze replicas. Scanning 12:99–106 41. Fujikawa S, Suzuki T, Ishikawa T et al (1988) Continuous observation of frozen biological materials with cryo-scanning electron microscope and freeze-replicas by a new cryo-system. J Electron Microsc 37:315–322 42. Muller M, Meister N, Moor H (1980) Freezing in a propane jet and its application in freeze-fracturing. Mikroskopie 36:129–140 43. Bachmann L, Schmitt WW (1971) Improved cryofixation applicable to freeze etching. Proc Natl Acad Sci U S A 68:2149–2152 44. Heuser JE, Reese TE, Dennis MJ et al (1979) Synaptic vesicle exocytosis captured by quick freezing and correlated with quantal transmitter release. J Cell Biol 81:275–300 45. Moor H, Bellin G, Sandri C et al (1980) The influence of high pressure freezing on mammalian nerve tissue. Cell Tissue Res 209:201–216 46. Umrath W (1983) Calculation of the freezedrying time for electron-microscopical preparations. Mikroskopie 40:9–34

Chapter 10 Using Pixel-Based Microscope Images to Generate 3D Reconstructions of Frozen and Thawed Plant Tissue David P. Livingston III and Tan D. Tuong Abstract Histological analysis of frozen and thawed plants has been conducted for many years but the observation of individual sections only provides a two-dimensional representation of a three-dimensional phenomenon. Currently available optical sectioning techniques for viewing internal structures in three dimensions are either low in resolution or the instrument cannot penetrate deep enough into the tissue to visualize the whole plant. Methods using higher resolution equipment are expensive and often require time-consuming training. In addition, conventional stains cannot be used for optical sectioning techniques. We present a relatively simple and less expensive technique using pixel-based (JPEG) images of conventionally stained histological sections of an Arabidopsis thaliana plant. The technique uses commercially available software to generate a 3D representation of internal structures. Key words Histology, Paraffin, Pixel-based images, Adobe After Effects, MRI, CT, Confocal microscopy

1

Introduction Since the development of X-rays, imaging internal structures of biological systems has provided information to help explain a variety of external symptoms of biological organisms. However, many of the techniques used for human and animal analysis such as magnetic resonance imaging (MRI) and positron emission tomography/computed tomography (PET/CT) are limited in resolution and cannot provide enough detail to be useful in plant systems. While synchrotron radiation imaging (SRI) produces images at considerably better resolution [1], it is rarely used for plants because of its expense and the limited number of facilities available for routine use. Micro-CT has been used to image internal structures in plants [2] but the low contrast between tissues within the

Electronic supplementary material: The online version of this chapter (https://doi.org/10.1007/978-1-07160660-5_10) contains supplementary material, which is available to authorized users. Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_10, © Springer Science+Business Media, LLC, part of Springer Nature 2020

119

120

David P. Livingston III and Tan D. Tuong

plant makes this technique a challenge. Confocal laser scanning microscopy (CLSC) is used for imaging fluorescent compounds within tissues at higher resolution than MRI or CT (see ref. 3 for a review). However, the use of CLSC is limited by the ability to penetrate tissues beyond about 0.4 mm [4]. Software used in conjunction with MRI, PET/CT, SRI, and CLSC all allow for optical sectioning of the organism and provide excellent 3D reconstructions in a nondestructive manner. Livingston et al. [4] review the positive and negative aspects of these and other techniques for viewing and reconstructing biological systems. The limitations of resolution, tissue penetration, and expense make these techniques impractical for routine 3D reconstruction of plants. The 3D reconstruction technique described here and elsewhere [4], will not replace the aforementioned techniques (MRI, CAT, CLSC). It is meant to be used on samples too small for MRI and CAT and too big or too thick for CLSC such as crown tissue of grasses. Advantages of this technique are that it is inexpensive, requiring only a digital camera, a microscope and the Adobe System software, After Effects. Another advantage is that tissues of any size, stained with any conventional stain, including fluorescently labeled antibodies (unpublished data) and images of any magnification can be used in the reconstruction. Disadvantages are that tissues must be sampled destructively to collect images and a certain degree of manual manipulation is required to align images that have been captured. Also, since images are pixel-based (not vector), boundaries of individual images will be visible and will create somewhat of a ladder effect in the final reconstruction. For most purposes this will not interfere with visualization of the final reconstruction. This chapter is not meant to be a tutorial on histological techniques nor in the use of Adobe After Effects. A basic knowledge of histology is required to apply this technique. After Effects (AE) is a somewhat complicated program used for postproduction in the film industry. Despite its complexity, anyone with a familiarity of Adobe Photoshop can learn what is needed to apply the 3D reconstruction technique described here. The reader is referred to http:www.linda.com for comprehensive online tutorials to learn how to use AE.

2

Materials 1. Chemicals and equipment for fixing, dehydrating, and embedding plant tissue in paraffin [5]. Equipment for sectioning, staining, and covering tissues on microscope slides.

3-D Reconstruction of Plant Tissue

121

2. Microscope or light table on which to photograph sequential images. 3. Tripod or camera mount for light table photography. 4. Cannon Rebel T5i or another consumer grade camera, with live view software. 5. Adapter to allow the camera to mount to the microscope (available, e.g., from Martin Microscope, Easley, SC, USA). 6. Mac or PC computer with at least 4 Gb of memory. 7. Adobe After Effects CS6. A newer version is available but CS6 will be the version used in this chapter. 8. Downloaded AE Plugin which allows AE to generate JPEG2000 images. Go to http://www.fnordware.com/j2k/ download the file “j2k.plugin” and place it into the Applications (programs) folder AfterEffects > Plug-ins > Format. See Note 1 for why this download is important. 9. Downloaded AE Script called “AlignLayers” at http://www. james-cheetham.com/Downloads/Tools/AESCRIPTS/. Select AlignLayers from the list on this site. Once the file is unzipped, move it to your Applications (programs) folder AE CS6 > scripts > ScriptUI Panels. Now open AE and go to file > scripts > run script file and find the .jsx file called AlignLayers. Click on it and it will open a panel in AE. To make the panel open automatically every time you open AE, go to “window” and the script, AlignLayer.jsx will be at the bottom of the window. Make sure a check mark is next to it. 10. The script allows you to evenly distribute your images in Z space. It also allows you to distribute images in x and y but you will use only z-alignment for this reconstruction.

3

Methods

3.1 Preparation of Sections

1. The color of the stain used to highlight objects of interest in sections should stand out from the rest of the tissue either in hue or in color saturation or both. Consistency of stain throughout the entire sequence is important since all routines that depend on color recognition will use the same hue/saturation setting for each image. 2. Consistently undistorted sequential sections have a direct impact on the quality of the reconstruction. Great care should be taken in transferring ribbons/sections to slides and drying sections on the slide. 3. Sections can be as thin as possible. For high magnification reconstructions (200
or higher) 10 μm or thinner will give better images due to a reduced focal length. However, if the

122

David P. Livingston III and Tan D. Tuong

tissue being sectioned is thick (1 or more cm), consideration must be given to the number of sections necessary to section the whole piece of tissue. We routinely section at 20 μm for a winter cereal crown which gives from 200 to 300 total sections. While theoretically there is no restriction in the number of images that can be processed, attempting to use more than 500 images that are 3–5 MB each will reduce the performance of the After Effects software significantly. 3.2

Photography

1. Any sequential set of digital images can be used in this 3D reconstruction technique. For example, a sequence of images can be photographed on a light table and assembled and processed in AE. A microscope is only necessary if tissues need to be magnified to view them adequately. 2. The Cannon Rebel T5i camera with software allows on-screen real-time imaging. The EFS 55–250 mm telephoto zoom is needed in combination with an adaptor made by Martin Microscope (see Subheading 2). The autofocus feature on the lens should be disabled since it does not perform very well with microscope images. This means each image will need to be focused manually. All camera settings can be made on the computer.

3.3 Importing Images and Converting Them to JPEG2000

1. Open AE and double click anywhere in the Project panel on the upper left of the screen (Fig. 1). This will open the File Manager and allow you to select the images to import. Find the folder with the images and click on a single image; all images in that folder will be imported. Make sure “JPEG Sequence” has a check mark next to it in the menu in the bottom-middle of the file manager screen (Fig. 2); this will ensure images are imported as a sequence (see Note 2 for difference between importing images as “Footage” or “Sequence”). Click “Open.” 2. A single line showing the file names of all the images will be listed in the Project panel. Click/hold and drag the sequence to the composition icon at the bottom of the project panel (the composition icon looks like a piece of film (see Fig. 1)). This will bring all the images to the Composition panel (Fig. 1) at the bottom of the screen and show the first image in the preview panel at the center of the screen. To auto-scroll through all the images, press the space-bar. Press the space-bar again to stop scrolling. 3. At the AE main menu (Fig. 1) select File > Save As > Save As and save the project in the same folder with the JPEG source images with an appropriate name (see Notes 3 and 4). Be aware that saving the project is not the same as saving images to a folder. You will save images by using the “render queue.”

3-D Reconstruction of Plant Tissue

123

Fig. 1 The opening screen after starting Adobe After Effects. The different panels used in this demonstration are indicated as are various icons mentioned in the text

4. Rendering: To save (render) the images in JPEG2000 format, highlight the sequence in the Composition panel (Fig. 1), go to AE main menu and select Composition > Add to Render Queue. The Render-Queue panel will open within the Composition panel at the bottom of the screen (Fig. 3). The composition containing the sequence of JPEG source images will become a tab at the top of the panel. In the Render panel click “Best Settings” (in blue) to the right of “Render Settings.” 5. In the window that opens, go to the box to the right of “Resolution” and make sure you are rendering at full resolution. Leave all other settings the same and click OK. 6. Back at the Render Queue panel click “Lossless” (in blue) next to “Output Module.” In the window that opens, go to the Output Module Settings > Main Options window, click the box to the right of “Format” and select JPEG2000. Click the box to the right of “Channels” and select “RGB + Alpha”. Leave all other options as they are and click OK.

124

David P. Livingston III and Tan D. Tuong

Fig. 2 The Import File window that opens (Mac) when the Project panel is double-clicked. Notice that only one file in the folder is selected and the JPEG Sequence option (arrow) is selected

Fig. 3 The Render Queue panel that opens when Composition > Add to Render Queue is selected from the AE main menu. Note the text in blue. All three regions in blue text must be opened and the parameters adjusted as needed. Notice the Render button at the far right (arrow)

3-D Reconstruction of Plant Tissue

125

Fig. 4 The window (Mac) that opens when the “output to” is clicked in the Render Queue panel. Notice the New Folder button in the lower left corner (arrow)

7. At the Render Queue, click the blue file name to the right of “Output To.” In the window that opens (Fig. 4) create a new subfolder (use the button at the lower left of the screen) in the same folder as the JPEG source files and AE Project. It is recommended that you name the folder and each image something similar to “source JPEG2000”; AE will automatically number each image sequentially. 8. Click “Save” and on the far right side of the Render Queue panel, click the “Render” button (Fig. 3). It will take a few minutes for the images to render into the new folder. A blue line across the top of the Composition panel will indicate the progress. 3.4 Aligning Images with Each Other

1. After images have been saved, double click anywhere in the project panel and click on one file in the JPEG2000 folder to bring all the images back into AE as a sequence of JPEG2000 images. Drag the new sequence to the Composition icon (Fig. 1).

126

David P. Livingston III and Tan D. Tuong

Fig. 5 The Effects panel with Warp Stabilizer opened and the three main parameters twirled down (arrows). The parameters in the boxes to the right of “Method” and “Framing” must be changed (see Subheading 3.4)

2. With the new composition in the Composition panel highlighted, go to the AE main menu Effect > Distort > Warp Stabilizer. The Effects panel will open (Fig. 5) within the Project panel and the Project panel will revert to a tab. 3. Twirl down “advanced” in the Warp Stabilizer window by clicking the small gray triangular arrow on the left (Fig. 5) and be sure “Detailed Analysis” is checked. In the box to the right of “Method” select “Position, Scale, Rotation.” In the box to the right of “Framing” select “Stabilize Only.” Leave all other settings as they are. 4. The stabilization will take from 15 to 30 min depending on the number of images. Review the stabilization by placing the time needle at the start of the composition and press the spacebar. See Note 5 and Subheading 3.13 if the alignment is not satisfactory. 3.5 Clearing the Back Ground

1. Highlight the sequence of aligned images in the Composition panel and go to the AE Main Menu, Effects > Keying > Color Key. In newer versions of EA “color key” is in the folder Effects > Obsolete. Click the eyedropper to the right of the “Key Color” line and move it to the color of the background you wish to remove, ideally somewhere near the tissue (see Note 6). Click the left button when you want to select a particular color. In the Effects panel, hover over the blue “0” to the right of “Color Tolerance” until the two-way arrow

3-D Reconstruction of Plant Tissue

127

appears and slide it to adjust the amount of color removed. Toggle the Color Keying effect on and off by clicking “fx” to the left of “Color Key” in the Effects panel. Do not worry about remaining background at the corners of the image. A mask will be applied to the images to remove vignetting in the corners as well as most of the noise in the background. 2. Color Keying will be applied automatically to all images in the sequence in the same way. The color will not be removed in the same way for any image or images within the sequence that was stained either darker or lighter. This underscores the need to insure that all images were stained identically (see Subheading 3.1). 3.6 Cleaning Images Using a Mask

1. Even the best color-keying routine will not be able to remove all noise from individual images due to dust and other particles adhering to slides. To clean images a “garbage matte” will be drawn around the region of interest, which will hide aberrant colors resulting from unwanted particles on the image. 2. At the top of the AE window click on the fountain pen icon (the “pen tool” in Fig. 1); also make sure the box to the right of the pen tool icon, labeled “RotoBezier” is checked. 3. Now the cursor (pen) in the preview window will be a small pen tip. Click the new cursor to one side of the tissue in the image in the preview panel, then move a little further and click again. Two small boxes will be drawn with a line connecting them. Continue this process until the tissue is completely outlined with about 5 or 6 boxes connected with lines. When you finally close the circle, all objects outside the line will disappear. 4. Now go to the Composition panel and click the gray triangular arrow to the far left (Fig. 1) to twirl down the layer. Twirl down “masks” and look for a box that says “Add.” Click the arrow to the right and select “none.” This will allow you to visualize everything in the image as the mask is adjusted. At the end of the masking process reselect “add” to again remove (hide) everything outside the mask. 5. Now make adjustments to the mask so it will remain just outside the tissue for each image and hide as much background noise as possible. 6. In the composition panel twirl down (click the gray triangular arrow to the left) “Masks” and then twirl down “Mask 1.” Click the stopwatch to the left of the line labeled “Mask Path.” This will put a key frame (small diamond) on your time line in your composition. Now click-hold the time needle and move it forward in your sequence until you see tissue beginning to migrate outside the mask. In the Preview panel command-click the mask tool while it is inside the mask, then

128

David P. Livingston III and Tan D. Tuong

move the arrow (cursor) over one of the boxes, click-hold and move the box so the tissue is once again inside the mask (see Note 7). 7. Work your way around the mask to ensure all the tissue is surrounded by the mask. Move the time needle ahead about 10 or 20 images and repeat this process. You should only have to do this about six or eight times over the course of the entire sequence unless the tissue is uneven or distorted. You should not have to reposition the mask for each image since masking with key frames will reposition the mask gradually between each key frame. 8. At the end of the sequence, move the time needle back to the beginning and press the spacebar to scroll through the sequence. Reposition the mask at any image where it was improperly placed. Keep in mind that a key frame will automatically be added to your timeline each time you reposition the mask. 9. To see what the final image series will look like, change the mask type to “Add” in the box to the right of “Mask 1” under the Mask menu on the Composition panel. To hide the mask as you scroll through the sequence click the Mask Visibility icon (Fig. 1) at the bottom of the Preview panel. 10. When you are satisfied that the mask will remove as much noise as is practical then render the aligned/cleared/cleaned images into a new folder as described in Subheading 3.3. Then save the project. 3.7 Outlining Tissue Using Emboss

1. To provide depth cues for the viewer, shading and texture should be added to the edges on each image. When the images are combined and rotated, the adjusted surface of the object infers a true 3D shape to the viewer. 2. Import the aligned and cleaned images as a sequence (double click the Project panel) and then drag the sequence to your composition icon. With the composition highlighted, go to the AE main menu, Layer > Layer Styles > Bevel and Emboss. 3. In the Preview panel notice that a thin gray line with a slight depth has appeared around all objects in the image. You will need to adjust this line somewhat and make it darker. So, in the Composition panel, twirl down the sequence and then twirl down “Layer Styles” at the bottom. Then twirl down “Bevel and emboss.” Next to “Style” twirl down the box, and change this to “Outer Bevel.” Next to “Technique” change “smooth” to “chisel hard.” Change the Size to 3.0. Change the “Highlight Color” to black and the opacity of both highlight color and shadow color to 100%. 4. Render the sequence (as described in Subheading 3.3) into a new folder with an appropriate name.

3-D Reconstruction of Plant Tissue

3.8 Distribute Aligned, Cleared Images in Z-Space

129

Now the aligned and cleaned images with the beveled and embossed edges must be distributed in Z space to allow 3D manipulation. 1. Double click in the Project panel and highlight the folder containing the aligned and cleaned images with emboss. This time, do not select an individual image. By highlighting the folder, all images will be imported as footage within the folder (see Note 2 for a description of the difference between footage and sequence). To see the individual images once the folder is imported into the Project panel, click the folder. 2. To speed up the next few steps, press the “Caps-Lock” key. This will disable AE from refreshing the preview screen each time a change is made. 3. Select all the images in the folder (click the top image and then shift-click the last one) and drag them to the Composition icon. A window will open in the middle of the screen called “New Composition from Selection.” In the “Still Duration” box type 20,000 for a 2 min duration. Leave other settings as is and click OK. 4. In the AE Main Menu go to Layer > New > Null Object. Then select all the images in the composition window, except the Null Layer. To quickly select all layers click one image and then type “Command (control on a PC) A” to de-select the Null Layer, Command-Click on it. 5. Click the coil icon just under the heading “Parent” in the composition panel. Drag the coil to the layer labeled “Null 1.” Now all the selected images will be under the control of the Null Layer. 6. Click the 3D Cube box (Fig. 1) for the null layer and one of the selected image layers. The box will automatically be checked on all selected layers. 7. Next, with all layers, except the null selected, check the box next to “Z-alignment” in the AlignLayers Panel to the right of the preview panel (Fig. 1). In the rectangular box to the far right change the “0” to about 6 and press enter. There are no units to this number so you will need to experiment with it to get the proper overall height (or depth) of your 3D object.

3.9 Animation to Manipulate the Object in 3D Space

1. In the Composition panel twirl down the Null Layer and twirl down “Transform.” On the line labeled “X-Rotation” change the number in blue to the far right to about 120. This will orient your 3D object at a slight angle. 2. Then click the stop-watch next to the line labeled Z-rotation to set a key frame at the start of your composition. Move the time needle to the end of the composition and change “0X” to

130

David P. Livingston III and Tan D. Tuong

about “10X.” This will allow ten complete rotations of your 3D object within the 2 min timeframe of your composition. 3. Now turn off the caps-lock button and your 3D object will appear in the Preview panel after a short wait. You will likely need to reposition the image to center it, or to reduce the size. 4. Twirl down “Transform” in the Null Layer and adjust the numbers in blue in the line labeled “Position.” To change the size of the object adjust the blue numbers in the line labeled “Scale.” 5. You should experiment with the different angles and speed of rotation to create the most effective animation to visualize your 3D object. 6. Press the space bar to watch your object rotate. It will be slow since AE has to redraw each layer for every shift in the view. Once the animation has played through, a green line will appear at the top of the composition indicating that this portion of your animation has been rendered into RAM memory and will now play much faster. Reposition your time needle at the beginning of the green line and press the spacebar again to observe the animation in real time. Reducing the resolution (Fig. 1) of the Preview Panel will speed up the playback. 7. To hide the gray lines in the Preview Panel go to the AE Main Menu. Under “View,” uncheck “Show Layer Controls.” Save the project. 3.10 Adding Layers to View Internal Structures

1. To view internal structures, repeat Subheadings 3.5, 3.7, and 3.8 using a different color to perform clearing as described in Subheading 3.5. It will be necessary to clear all unwanted tissue and leave behind only the structures that are stained a color that is distinct from surrounding tissues (Fig. 6). This works best with stains that are specific for specialized cells, such as Safranin for xylem vessels or immunohistological stains that are specific for certain proteins. 2. Begin by importing the aligned and cleaned images from Subheading 3.6 as a sequence. Drag the sequence to the Composition icon and highlight the sequence in the Composition Panel. Go to Effect > Obsolete > Color Key and use the dropper to select the color to remove that will leave behind the structures you wish to reconstruct. You may need to experiment with precisely how much of the surrounding color to remove. A stroke can be applied to the sequence (see Subheading 3.7) but may not be necessary. Render the images into a new folder. 3. Repeat Subheading 3.8 for the images showing structures inside the tissue (see Note 8).

3-D Reconstruction of Plant Tissue

131

Fig. 6 A single section of Arabidopsis stained with Safranin and Fast Green. The background of the image has been cleared (Subheading 3.5) and is shown on top of the transparency grid in the Preview panel (Fig. 1). This image shows the narrow black outline around all the tissues called a “stroke” (Subheading 3.7)

4. Now superimpose the images showing the whole plant with the images showing only the inside. The inside will be made visible by gradually reducing the opacity of the whole plant images during the animation. 5. Click the tab for the 3D composition you made in Subheading 3.9. Under the “Transform” menu of the Null Layer click the blue “Reset.” Click the stopwatch on all the lines where key frames were set to remove them. 6. In the Project panel select both 3D compositions and drag them to the Composition icon (this process is called “precomposing” or just “pre-comping”). Now you will have two layers in the Composition panel that have all your 3D layers embedded in them. Rename the layer “3D master” or something similar, by left-clicking the composition in the Project panel and selecting “Rename.” 7. Add a Null layer to the new composition and then after selecting the two compositions (in the Composition panel) pull the coil to the null layer so both layers will be subject to

132

David P. Livingston III and Tan D. Tuong

Fig. 7 The composition panel with three footage compositions nested (pre-comped—see Subheading 3.10) and under the control of a Null layer. Note the coil described in Subheading 3.8 and that the “little sun” icon (Collapse Transformations) is turned on for the three compositions. Keyframes have been added (small diamonds on the right side of the panel) to signal when changes to the various parameters (positional parameters for the Null layer and opacity to the 3D layers) are to begin

manipulations from the null (see Subheading 3.9). Click the box with the icon that looks like a little sun (Fig. 1, “Collapse Transformations”) on each layer (Fig. 7, arrow). 8. Twirl down the null layer as described in Subheading 3.9 and experiment with different rotations. 9. To reduce the opacity of the surrounding tissue and visualize internal structures (Fig. 8b), twirl down the whole tissue layer, then twirl down “Transform.” Make sure the Time Needle is at the beginning of your time line. In the line labeled “Opacity,” click the stopwatch to the left to set a key frame on the Time Line. Move the time needle ahead to where you want to begin viewing internal structures. Click the gray diamond to the far left to set a key frame. This will let the program know where to begin the reduction in opacity. Then move the time needle forward a few seconds and change the opacity from 100 to about 5% or 10% depending on how transparent you want the external tissue to be (Fig. 7). 10. When the composition is played, the first few seconds will show the whole tissue rotating and when the time needle reaches the opacity key frame, the external tissue will gradually become transparent showing the internal structure in 3D (Fig. 8a, b). Key frames can be repositioned anywhere in the composition by click-hold and dragging them or they can be removed by clicking and deleting. 3.11 Rendering the 3D Animation as a Movie

1. You can preview your animation by rendering it as a movie at a much lower resolution than you will in your final video. Sometimes this is faster than waiting for AE to load the animation into RAM memory.

3-D Reconstruction of Plant Tissue

133

Fig. 8 3D reconstruction of Arabidopsis (cv Columbia-0) 7 days after freezing at 14  C. (a) Reconstruction of the whole plant. Note the tattered appearance of the dead leaves. (b) The same reconstruction as in a except that tissue surrounding the putative freeze-response (in red) has been digitally cleared. Note the circular shape of the red staining region in the center of the crown (arrow). Its appearance as a donut-shape is more obvious in the supplemental video

2. Follow the directions for rendering (Subheading 3.3) except in the Render Settings reduce the resolution to Quarter or Half. Click OK. In the Output Module make sure Format is set to “QuickTime” and Channels are set to RGB + Alpha. Click OK. In the Output To window have your video rendered to the same folder as all your other files related to the project. 3. When the video is finished rendering, you can view it outside of AE in QuickTime or some other video viewing program that will play .mov files. 4. Use this reduced resolution version of your animation to decide the best rotations and opacity changes. Make appropriate changes to your composition and render it at full resolution. 5. A disadvantage of this 3D reconstruction technique is that the final video is not interactive. Once the animation is rendered to a video, the 3D object cannot be manipulated. It is therefore important to select as many angles in the animation process as may be important to clearly understand the tissue in 3D. 3.12 Adding a Background and Drop-Shadow

1. To add a background color to your video import the full resolution .mov file rendered in Subheading 3.11. Drag it to the Composition icon. In the AE main menu go to Layer > New > Solid. A solid colored layer will be placed on top of your .mov layer in the Composition panel hiding your

134

David P. Livingston III and Tan D. Tuong

video. Click and drag the solid layer below the .mov layer to allow you to see the video on top of the background (Fig. 8). Change the color of the background by highlighting the solid layer in your Composition panel and in the AE main menu go to Layer > Solid Settings. Click the color box to change the color. 2. To add a drop shadow, highlight the Video layer and go to the AE main menu Layer > Layer Styles > Drop Shadow. In the Composition panel, a Layer Styles menu will be added to the composition. Twirl down “Drop Shadow” and experiment with the settings. We have found good settings to be Opacity ¼ 50, Distance ¼ 50, and size ¼ 50. Leave the other settings as they are. 3. Text can also be added to your video and the text can be animated in a variety of ways. Descriptions of how to do this are complicated and are best described in video tutorials. See www.Linda.com. 4. When you are ready, render the 3D video at full resolution. 3.13 Manual Image Alignment

During manual alignment there is no reason to worry that you will drift in one direction or another as you manipulate each image. It is not known why, but if a center point is used in each image as a kind of visual fiducial, the resulting 3D volume will be anatomically accurate. 1. Double click in the Project panel and import the JPEG2000 images from Subheading 3.3 as a sequence. While not required, it will be easier to manually align images that have been cleared of background. Follow the procedure above (Subheading 3.5) for digitally clearing the background. Render the cleared images into a new folder using composition > add to render > queue as described above in Subheading 3.3. 2. Double click in your project panel and select the folder (i.e., not an individual image this time) where the cleared images were rendered and click “open.” All images will be imported within the folder as footage (i.e., not as a sequence). 3. Click on the folder in the project panel and all the individual images will be listed. Select the topmost image and shift > click the last one. Click and drag the entire set of images to the composition icon to create a footage composition in the Composition panel (Fig. 1). This will allow all images to be viewed simultaneously on top of each other. To view only a single image select all the images in the composition (command or control A) and click on the small eyeball (Fig. 1) in one image. This will hide all images from view. With the Composition panel highlighted press the “t” key to open the transparency

3-D Reconstruction of Plant Tissue

135

parameters of all the images. Change the transparency number from 100% to about 60%, then press “t” again to close the transparency parameter window. 4. To begin manual alignment, click anywhere in the Composition panel to deselect all the layers. Then scroll to the bottom of the list of images and click the eyeball of the bottom two images to turn on only those images. You will be able to see the bottommost image through the second one and can now adjust the position and rotation of the second image from the bottom so that it falls directly on top of the bottom image (which remains stationary). Be sure the eyeball for both images is showing but only highlight the top image in the Composition panel. Whichever image is highlighted will be the image that will be manipulated (see Note 9). 5. Press the “w” key to turn on the rotation parameter within the Preview pane (see Note 10). Click-hold the left mouse button in the Preview panel to rotate the image. If you hold down the “command” (control on a PC) button and mouse click/hold you will be able to move the top image positionally. By alternating rotation and positional manipulation you will be able to place the top image directly over the bottom one. 6. To see the progress of your adjustments, click the eyeball on the top image on and off. When you are satisfied with your progress, save the project (command > s) and click the eyeball of the bottom image to turn it off, then click the eyeball of the next image above to turn it on. Be sure and highlight the new image, then begin your adjustment with the mouse. After some practice you should be able to make manual adjustments to about 60 images in an hour. 7. When you are finished aligning, select all images, click the letter “t” and bring the transparency of all images back to 100%. 8. When all images have been adjusted, and the transparency is at 100% you will render the aligned images. Because all images are currently on top of each other, they must be distributed as a sequence before rendering. 9. Select all the images (the image you click first will determine the order in which the next process will occur) and then in the AE main menu go to Animation > Keyframe Assistant > Sequence layers. A small window will open asking if you want any overlap between images. Make sure the overlap box is not checked and click OK. 10. In the Composition panel at the bottom all your images will be distributed automatically in a “stair-step” pattern. If you click the space-bar, the sequence will automatically play through one image at a time.

136

David P. Livingston III and Tan D. Tuong

11. Once you are satisfied with the alignment, render the images to a different folder (see Subheading 3.3, step 4) with a name such as “aligned.” 12. After images have been aligned and rendered go to Subheading 3.6 and proceed with the remaining steps. 3.14 Images Used in the Demonstration

Arabidopsis thaliana (cv Columbia-0) is frozen at 14  C for 3 h and, after thawing, was allowed to recover from freezing for 7 days. Plants are fixed, dehydrated and embedded in paraffin, as described elsewhere [5]. The paraffin-embedded plants are cut into cross sections at a thickness of 20 μm. Sections are photographed at 40
with the camera described above and imported into Adobe AE on a Mac Pro computer. Except for the bottom 1 or 2 mm of the root, it takes 330 sections to photograph the entire plant used here. Freezing Arabidopsis plants at 14  C under our conditions [6] resulted in a survival of 40% of the population. The plant shown in Fig. 9 is one that would eventually fully recover and regrow because meristematic tissues in the center of the rosette survived freezing. This is the region of the plant from which new growth appears, even though leaves that have been frozen are dead. Dead leaves in this case resulted in a somewhat wrinkled view of the top part of the plant in the 3D reconstruction (Fig. 8a). The crown region of an Arabidopsis plant is analogous to the apical meristems of winter cereals that are also the hardiest tissues in the plant [7, 8]. Individual 2D cross-sectional images show sporadic Safraninstained tissue in seemingly random regions of the crown (not shown). When a sequential series of those images was aligned and viewed in 3D, a ring structure resembling a donut became visible in the center of the crown region of the plant (Fig. 8a, the “donut”

Fig. 9 A surviving Arabidopsis plant after being frozen at 14  C for 3 h and allowed to recover for 7 days. Note the dead leaves surrounding the live tissue regrowing from the center

3-D Reconstruction of Plant Tissue

137

structure is easier to see in the Supplemental Video 1 included on SpringerLink with this chapter). This structure was not apparent when individual images were viewed in 2D. It was only when viewed in 3D that the donut-like structure became apparent. The ring structure is similar to that found in oats recovering from freezing [9] but dissimilar in that the ring in a 3D reconstruction of oat forms a somewhat spherical shape. The Arabidopsis 3D reconstruction also shows continuity throughout most of the taproot for vessel plugging (see ref. 5 for 2D longitudinal view), which was also similar to that observed in oat. Safranin commonly stains lignin/phenolic compounds [10] a deep red color as shown here. It is not unusual for specific cells within plants to produce phenolic compounds in response to abiotic or biotic stress [11]. Because similar Safranin staining regions in unfrozen plants did not occur it is reasonable to assume that the Safranin shown here stained tissue/cells that were responding in some way to freezing stress. Alternatively, since the Safranin staining regions do not become visible until about 3 days after freezing, the plant could be responding to secondary infection by bacteria or fungi [12, 13] as described by Beckman [11]. Whatever the composition of the red-staining material, this example illustrates how 3D reconstruction can be used to demonstrate continuity in 3D space of substances that are distributed in a seemingly random manner when viewed in 2D. The technique is well suited for in situ hybridization analysis to detect mRNA in an anatomical context and show its continuity in 3D space. In addition, it has been used successfully to 3D image cancerous and healthy tissues in veterinary pathology studies [4].

4

Notes 1. The reason for converting to JPEG2000 (J2K) is that J2K images are better quality in a smaller file and will be easier to manipulate in AE than either original JPEG or TIFF images. 2. Images can be imported as either a sequence or as footage. Sequences can be thought of as a horizontal line of images. One change will be applied to the entire series of images. To Color Key all your images simultaneously they must be imported as a sequence. Footage can be thought of as a vertical column of images. Changes can be made to individual images without affecting any other. Manual Alignment requires importing your images as footage so that each image can be separately manipulated. 3. It is important to decide on an organizational structure before starting. For example, the project file as well as all folders containing images generated by the project should be in the

138

David P. Livingston III and Tan D. Tuong

same folder. If folders and/or files are rearranged or deleted, reopening a project associated with the deleted or rearranged files/folders will cause a conflict, that is, AE will not know where to find the files. If that happens go to the AE main menu File > replace footage and select the folder where you moved the images. 4. Save the project regularly. Under After Effects > Preferences > Auto-save specify a time for AE to save automatically. This may cause some annoyance but it is well worth it, if a major problem causes AE to freeze or crash. This will be a rare occurrence but it does happen. 5. Warp stabilizer will produce an alignment automatically but a small amount of resolution will be compromised in each image. Warp Stabilization works well with most crown tissue sections from winter cereals but the Arabidopsis plant used in this demonstration had to be aligned manually. If Warp Stabilization does not produce an adequate alignment, images can be aligned manually. This is a somewhat time-consuming process since each image has to be manipulated individually. With practice, about 60 images can be manually aligned in an hour. See Subheading 3.13 for instructions on manual alignment. 6. When Color keying, as many color keys as needed can be added to the Effects panel to remove all of an unwanted color or several different colors in the same sequence. 7. To change the preview window from a black background to a transparent background click on the transparency grid at the bottom of the preview window (Fig. 1). 8. Use command (control) > Z to back up, redo. 9. It is not unusual to lose a panel due to a slip of the mouse or click of a button. To get back to your original panel structure click the “workspace” box at the top right side of the AE window (Fig. 1) and select “all panels.” 10. The size of the Preview panel can be increased or decreased using the > and < keys. To move the Preview panel spacebar, click in the preview panel to reveal a hand. Use the mouse to reposition the window. References 1. Peyrin F (2009) Investigation of bone with synchrotron radiation imaging: from micro to nano. Osteoporos Int 20:1057–1063 2. Wang Z, Verboven P, Nicolai B (2017) Contrast-enhanced 3D micro-CT of plant tissues using different impregnation techniques. Plant Methods 13:105–121

3. Ferrando M, Spiess WEL (2000) Review: Confocal scanning laser microscopy. A powerful tool in food science. Food Sci Technol Int 6:267–284 4. Livingston DP III, Tuong TD, Gadi SRV et al (2010) 3D reconstructions with pixel-based images are made possible by digitally clearing

3-D Reconstruction of Plant Tissue plant and animal tissue. J Microsc 240:122–129 5. Livingston DP III, Tuong TD, Haigler CH et al (2009) Rapid microwave tissue processing of winter cereals for histology allows identification of separate zones of freezing injury in the crown. Crop Sci 49:1837–1842 6. Livingston DP III, Herman EM, Premakumar R et al (2007) Using Arabidopsis thaliana as a model to study subzero acclimation in small grains. Cryobiology 54:154–163 7. Tanino KK, McKersie BD (1985) Injury within the crown of winter wheat seedlings after freezing and icing stress. Can J Bot 63:432–435 8. Olien CR, Marchetti BL (1976) Recovery of hardened barley from winter injuries. Crop Sci 16:201–204 9. Livingston DP III, Henson CA, Tuong TD et al (2013) Histological analysis and 3D

139

reconstruction of winter cereal crowns recovering from freezing: a unique response in oat (Avena sativa L.). PLoS One 8:e53468 10. Johansen DA (1940) Stains. In: Plant microtechnique. McGraw-Hill Book Company, New York, pp 49–64 11. Beckman CH (2000) Phenolic-storing cells: keys to programmed cell death and periderm formation in wilt disease resistance in general defense responses in plants? Physiol Mol Plant Pathol 57:101–110 12. Marshall D (1988) A relationship between icenucleation-active bacteria, freeze damage, and genotype in oats. Phytopathology 78:952–957 13. Olien CR, Smith MN (1981) Extension of localized freeze injury in barley by acute postthaw bacterial disease. Cryobiology 18:404–409

Chapter 11 Phenotyping Plant Cellular and Tissue Level Responses to Cold with Synchrotron-Based Fourier-Transform Infrared Spectroscopy and X-Ray Computed Tomography Ian R. Willick, Jarvis Stobbs, Chithra Karunakaran, and Karen K. Tanino Abstract Despite the extensive use of synchrotron radiation in material and biomedical sciences, it has only recently been utilized to expand our understanding of plant responses to environmental stress. Recent advances have led to the development of phenotyping platforms to identify chemical and morphological differences in breeding plant material. While these methodologies are applicable for and tested with a variety of abiotic and biotic stresses, they are particularly useful as a first step to identify cold-induced chemical and morphological changes in plants. Here, we describe two methods to determine cold acclimation-induced changes at the cellular and tissue levels. First, we illustrate how to quantify and visualize changes in tissue chemistry using Fourier-transform infrared spectroscopy. Second, we describe how to nondestructively prepare, analyze, and interpret X-ray phase contrast images and render this data as two- or threedimensional models. While these techniques utilize synchrotron radiation, the methodology and standard practices are applicable for handheld and laboratory bench-top equipment operating with conventional light sources. Key words Attenuated total reflectance (ATR), Cold acclimation, Cell walls, Focal plane array (FPA), Fourier-transform Infrared (FTIR), Phase-contrast X-ray imaging, Synchrotron

1

Introduction Overwintering cereal crops cope with a variety of freezing-related stresses including desiccation, osmotic stress, hypoxia, flooding, ice encasement, soil heaving, and/or smothering [1]. Winter habit cereals enhance their freezing tolerance and avoidance through preexposure to low, nonfreezing temperatures. Without this cold acclimation period, spring and winter cereals have a similar LT50, or the temperature at which half the plants recover, of
5  C [1]. Once cold acclimated, winter wheat (Triticum aestivum L.)

Electronic supplementary material: The online version of this chapter (https://doi.org/10.1007/978-1-07160660-5_11) contains supplementary material, which is available to authorized users. Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_11, © Springer Science+Business Media, LLC, part of Springer Nature 2020

141

142

Ian R. Willick et al.

and rye (Secale cereale L.) can survive exposure to temperatures as low as
24 and
34  C [1]. This increase in freezing survival is the result of significant modifications to the biochemistry and physiology of the plants. Cold acclimation promotes the accumulation of cryoprotectant sugars [2], dehydrins [3], antifreeze proteins [4], and ice-nucleating and supercooling-promoting substances [5]. On a cellular level, cold acclimation promotes the reorganization of the plasma membrane [6] and cell wall [7, 8]. In plant organs, modification of tissues promotes controlled sites for ice nucleation [9, 10], while simultaneously developing tissues to form structural and/or chemical barriers to prevent the propagation of ice from these ice sinks, into nonfrozen freezing-sensitive tissues. In order to select superior cold hardiness germplasm, we require a more complete spatial understanding of the physiological and biochemical mechanisms underpinning freezing survival. Various analytical techniques can identify biochemical modifications to the plant. Mass spectrometry coupled to either gas or liquid chromatography is the classical method for quantifying precise lipid, metabolite, or protein profiles. However, these techniques destroy critical complementary information such as the spatial localization of compounds at the tissue, cellular or subcellular level [11]. Traditional bulk sampling averages the intrinsic heterogeneity of complex plant tissues and organs resulting in masked novel tissue or cell level phenotypes. Furthermore, traditional biochemical techniques require milligram or gram quantities of freeze-dried tissue to isolate a compound of interest. Analysis of inherently small samples such as the shoot apical meristem of a wheat crown [8] or fresh pollen from field pea (Pisum sativum L.) [12] requires harvesting hundreds of plants to accrue enough biomass for analysis. Fourier-transform infrared (FTIR) spectroscopy was developed with conventional light sources to study chemical constituents within whole cells [7, 8, 13] or plant tissues [12, 14–16]. For example, Japanese bunching onion (Allium fistulosum L.) was used as a model system to map chemical changes with a focal plane array (FPA) in the fractions of carbohydrates carrying methylesters and of fractions of α-helical and β-sheet secondary structures in proteins in the epidermal cells of nonacclimated and cold acclimated plants [7]. In winter wheat, differences in peaks related to carbohydrates and methylesters were spatially distinct and could be quantified in nonacclimated and cold acclimated cells from the shoot apical meristem and the vascular tissue at the base of the crown [8]. This research led to the identification of tissue specific mechanisms of freezing in the cold acclimated winter cereal crown [10]. Focal plane array mapping with FTIR spectroscopy localized relatively pure protein areas within tissues and revealed protein secondary structure explaining the biological differences noted among varieties. From an applied perspective, quick measurements of basic chemical constituents, such as lignin or the degree of methyl-esterification in the

Phenotyping Plant Cold Responses with FTIR and μCT

143

cell wall or modifications to epicuticular wax biochemistry are useful tools for phenotyping larger plant populations. With the advent of handheld Attenuated Total Reflectance (ATR)-FTIR devices, this FTIR method will provide breeders with additional tools to measure such modifications in the field [15]. While this chapter uses cereals as model systems and focuses on responses in cold acclimated plants, the described FTIR techniques are applicable to a variety of different stresses and plant species. This includes characterizing the effect of high temperatures on the epicuticular wax of pea pollen [12], chilling injured corn (Zea mays L.) [15] or drought stressed wheat flag leaves [17]. It has also been used to investigate cold-induced modifications of the cell walls of Arabidopsis (unpublished data) and Japanese bunching onion [7] and for screening wheat lines for biochemical barriers in the spikelet to Fusarium Head Blight [14] or dormancy induction through the formation of biochemical barriers in overwintering tree buds [16]. In the future, breeding programs incorporating FTIR spectroscopy into field level analysis could detect early injury or infection in field-grown plants that is not visible to the naked eye and ultimately differentiate susceptible and resistant breeding material. From a basic research perspective, FTIR spectroscopy is a useful tool to identify critical regions of interest at the cellular and tissue levels for subsequent in-situ biochemical and molecular characterization. In combination with FTIR chemical analysis, three-dimensional reconstructions provide an additional layer of detail on how plants tolerate and avoid freezing. A similar system for observing complex organs in winter cereal crowns was developed by imaging successive stained axial sections with a brightfield microscope and stacking the images with corrective software. Three-dimensional structures were digitally rendered and provided important physiological information at the tissue scale [18]. However, this technique requires successive days to fix, dehydrate and embed tissues and needs significant skill to process and stain each successive section. Here, we will discuss an alternative imaging technique using X-ray imaging and computed tomography. This facilitates the nondestructive imaging of live plants within 15–20 min. In conjunction with synchrotron radiation, three-dimensional renderings of anatomical structures can provide resolutions of micron or nano scales. This technique does require access to sophisticated equipment. Fortunately, synchrotron facilities allow free use of beamline equipment through a competitive application process. In brief, X-rays passing through an object attenuate and phase shift depending on the refractive index and density of the material [19]. Phase contrast imaging relies on the refraction of X-rays to produce edge-enhanced images [20]. Soft materials, such a plant tissue, have a significantly higher phase shift as compared with attenuation. As a result, a lower dose of X-rays is required for

144

Ian R. Willick et al.

good image contrast of plant materials as opposed to polymers or sedimentary materials [21]. As compared with conventional light sources, X-rays produced from synchrotron radiation have a higher flux density (photons/s/mm2) which increases the signal-to-noise ratio. At energy levels higher than 5 keV produced by synchrotron radiation, the phase signal is three times higher than the absorption signal. This facilitates the selection of specific wavelengths for imaging of soft tissues as well as selecting partial coherence to improve image contrast in low-density plant organs [21]. Synchrotron radiation phase contrast X-ray imaging and computed tomography has identified and quantified rachis node densities in wheat heads and correlated this trait with increased tolerance to Fusarium Head Blight disease [14]. This technique would be useful in identifying similar mechanisms of tissue specific freezing injury in complex organs such as overwintering tree buds or delicate structures such as the tissues connecting the Arabidopsis leaf petiole and meristem. This chapter acts as a primer for sample collection, processing, and analysis. We will describe general practices on how to analyze spectra and capture two-dimensional X-ray images to produce three-dimensional renderings of an organ of interest. For FTIR analysis, there are numerous platforms available for the analysis and interpretation of infrared spectra including OPUS, OriginPro, and MatLab plugins. We have chosen to use Orange [22] because of its recommended use by the Canadian Light Source Synchrotron (Saskatoon, Saskatchewan, Canada) and open access availability to the public. Detailed information on the analysis and interpretation of spectra with Orange is available on their website (https:// orange.biolab.si).

2 2.1

Materials Equipment

1. FTIR spectrometer with a light microscope and a 64  64 pixel focal plane array detector or an ATR head accessory. 2. Cryomicrotome (e.g., 3050 S, Leica Biosystems, Nussloch, Germany). 3. Dewar. 4. Freeze-dryer. 5. Mortar and pestle. 6. Tongs.

2.2

Software

1. Aviso version 9.7 (Thermo Fisher Scientific and Konrad-ZuseZentrum fur Informationstechnik Berlin, Germany). 2. Orange version 3.2 (University of Ljubljana, Slovenia). 3. OPUS version 7.2 (Bruker Optics Inc., Billerica, MA). 4. UFO-KIT (https://ufo.kit.edu/dis/index.php/software/).

Phenotyping Plant Cold Responses with FTIR and μCT

2.3

Consumables

2.3.1 Bulk and ATR-FTIR

145

1. Acetone (100%). 2. Double distilled water. 3. Glass slides (25 mm  75 mm  1.0 mm). 4. Paper towel. 5. Potassium bromide (FTIR grade, >99% trace metals basis). 6. Laboratory grade razor blades. 7. Transparent one-sided tape.

2.3.2 Focal Plane Array FTIR

1. Barium Fluoride (BaF2) polished circular windows (2.25 mm diameter, 1 mm thick; e.g., Crystran Ltd, Poole, UK). 2. Brush. 3. Cryostat glass insert (70 mm, e.g., from Leica Bio-Systems; Denmark). 4. Disposable base molds (7 mm  7 mm  5 mm). 5. Double distilled water. 6. Liquid Nitrogen.

3

Methods

3.1 Experimental Design 3.1.1 Growing Winter Wheat

1. Imbibe seeds on moist filter paper in petri dishes and hold at 4  C in the dark for 3 days and then at 22  C until roots have grown to 5–10 mm. Transfer germinating seeds to 14  24 cm plexiglass trays with holes backed by a 1.8 mm mesh screen. Affix trays over tanks filled with aerated modified half-strength Hoagland’s solution [23]. Refill tank solution with distilled water as needed. Replace tank solution every 2 weeks with new modified Hoagland’s solution (see Note 1). 2. Tanks are transferred to a constant 20  C growth chamber, with a 16 h day and photosynthetic photon flux density (PPFD) of 400 μmol/m2/s until seedlings reach the threeleaf stage. 3. For cold acclimation, plants are held at 4  C with a 16 h photoperiod and PPFD of 400 μmol/m2/s (see Note 2).

3.2 Sample Preparation for FTIR 3.2.1 Bulk FTIR

1. Freeze-dry tissue for a minimum of 48 h. Length of time may vary depending on the tissue in question. Loosely packed leaves or roots require about 48 h. Complex organs such as tree buds or cereal crowns require 72 h or more. 2. Dried sample are transferred to a clean 1.5 mL microfuge tube with two acid-washed stainless steel ball bearings and are ground to a fine powder with a ball mill shaker. Alternatively, samples can be ground in a mortar and pestle.

146

Ian R. Willick et al.

3. A 0.93 g sample of solid KBr is added to an acetone washed mortar and is ground to a fine powder. Add 2 mg of the ground plant sample and homogenize it with the KBr. For each set of samples, one pure KBr pellet (0.95 g) serves as a control. 4. Load the powdered mixture into a die and then transfer to the pellet press. Apply 7 tons of pressure to the die for 5 min (see Note 3). 5. The following are optimized pellet parameters: sample concentration in KBr pellet: 1%. Particle size of sample: 2 μm. Pellet diameter: 13 mm. 3.2.2 ATR-FTIR

1. If using live samples (see Note 4), lightly blot the material with a wet paper towel prior to analysis to remove any surface contaminants. 2. For samples to be freeze-dried, place leaves in the correct orientation of interest. Either the adaxial or the abaxial side facing away from the glass slide. 3. Place a piece of tape along the sample’s edges. This ensures freeze-dried samples are flat. Avoid touching the area of interest on the leaf surface. This can result in the disruption of surface structures and chemistry. 4. Samples can then be arranged in a slide box and freeze-dried for 72 h. Samples stored in a desiccator in a cool, dark room are stable for up to 4 months.

3.2.3 Focal Plane Array (FPA) FTIR

1. Trim sample to a maximum size of 7 mm  7 mm with a clean razor blade. Place samples in the well of the disposable plastic mold (7 mm  7 mm  5 mm) so that the tissue surface of interest is facing toward the base of the mold (see Note 5). This is the sectioning plane. Decant water into the mold (see Note 6). 2. Grip the mold with a pair of metal tongs and lower to the surface of the liquid nitrogen. The mold well, sample and water are in direct contact with the liquid nitrogen for 2 min. Transport samples in liquid nitrogen to the cryostat (see Notes 7 and 8). 3. Set the cryostat to
20  C. Place brushes in cryostat, attach blade, and screw glass plate into place approximately 15 min prior to sectioning (see Note 9). 4. Set the cryostat to cut 12 μm thick sections. 5. Place OCT (optical cutting temperature) solution on a prechilled circular cryostat block and immediately affix frozen sample block with tweezers. The flat surface of the sample (sectioning surface) needs to face away from the cryostat block. Allow samples to equilibrate for 10 min.

Phenotyping Plant Cold Responses with FTIR and μCT

147

6. Attach block with sample onto the block mount. Screw the block into place and orient the sample parallel to the blade. Adjust the working distance between the blade and the sample. 7. When sectioning, double-check if the sample is evenly cut. If tearing occurs, adjust the humidity or temperature in the cryostat. If sections do not slide underneath the glass plate, adjust the glass plate to be out further from blade and closer to the block. 8. Use a fine brush to transfer sections to a BaF circular disc. 9. To avoid section tearing post-sectioning, samples should be sandwiched between two discs, wrapped in tissue paper and then placed in labelled envelopes for transport. Store slides at
20  C until ready for FTIR analysis. 3.3 Acquisition of FTIR Spectra

1. Open OPUS. Holding the mouse over each icon will reveal its function.

3.3.1 Bulk and ATR-FTIR

2. Before collecting data, confirm that the glowbar source is on. Turn the scanner on. 3. Use the “Measure” button to acquire new data. Clicking on this button will open a new multitab. 4. Under the “Basic Tab” verify the Experiment file, Operator, Sample, and Sample Form. Including this information will facilitate easier data organization. 5. Under the “Advanced Tab,” modify the following parameters: resolution 2 cm
1; 64 or 128 coadded scans; 64 background coadded scans; Spectrum type: Absorbance; Frequency limit: 4000–600 cm
1; Scanner Velocity: 6:10 kHz; Detector: deuterated triglycine sulphate; Aperture size: 12 mm; Operating vacuum: 75% of the variance in the dataset (Fig. 4c). 4. Connect the “PCA” widget with the “Scatterplot” widget to visualize the principal components plot (Fig. 4c). 5. To export raw data to create figures, download the raw data associated with the score and loading plots by connecting the “PCA” widget to a “Save Data” widget. Double-click on the line connecting the two widgets to open an edit links window.

152

Ian R. Willick et al.

Fig. 3 Example of an Orange (version 3.21) widget workflow for the visualization, analysis, and extraction of spectra data

6. To export the score plot raw data (Fig. 4c), select the “Transformed Data” to send to the “Save Data” widget. 7. To export the loading plot raw data (Fig. 4d), select the “Components” data to send to the “Save Data” widget. 3.4.4 Semiquantification of Spectra Peak Area with Orange

1. Connect the “Dataset” widget to the “Preprocess Spectra” widget. Include only the “baseline correction” preprocess step. Add the “Cut (keep)” preprocessing step with the high and low limits corresponding with the edges of your peak of interest. Then add “baseline correction.” 2. Connect the “Preprocess Spectra” widget to the “Integrate Spectra” widget. Double-click on the “Integrate Spectra” widget. Click on “add integral” and then “integral from baseline.” The low and high limits should correspond with the edges of your peak of interest. Click Ok. 3. Review integrated peak areas (arbitrary units) with the “Data Table” widget or export the data to Excel for statistical analysis using the “Save Data” widget.

Phenotyping Plant Cold Responses with FTIR and μCT

153

Fig. 4 Analysis of the fingerprint region (1800–900 cm
1) of absorbance spectra collected from the cell walls of nonacclimated (NA) and cold acclimated (ACC) winter wheat cv. Norstar (n ¼ 20). (a) Averaged absorbance spectra. (b) Second derivative of absorbance spectra to illustrate the multiple peaks. Principal component analysis representing 95.8% of the variance in the data. (c) The score plot illustrates that 87.2% of the variance in the data is explained by the first principal component (PC1) and would therefore have a larger influence on the dataset than PC2. (d) The most positive regions in the loading plot correspond to the arabinoxylan (1050 cm
1), amide I (1650 cm
1) and amide II (1540 cm
1) peaks. The separation between the nonacclimated and cold acclimated groups are strongly influenced by differences in these three peaks

3.5

X-Ray Imaging

3.5.1 Phase-Contrast Micro-computed Tomography (PC-μCT) Experimental Set-Up

1. The X-ray source for our experiments is located at the BMITBM beamline (05B1-1) at the Canadian Light Source Synchrotron (CLS). The 05B1-1 bending magnet beamline generates X-rays between 12.6 and 40 keV (0.8–0.3 A˚) with the beam size of 240 mm (H)  7 mm (V). Similar beamlines are available at other facilities worldwide.

154

Ian R. Willick et al.

2. We use a Hamamatsu C9300-124 (4000  2600 pixel) CCD detector with a 9 μm pixel size and 36.0 mm (H)  24.0 mm (V) field of view (FOV) at 4.5 frames/s. 3. Crop the usable detector FOV from 24 mm (H)  7 (mm) to 8 mm (H)  4.8 mm (V). 3.5.2 PC-μCT Image Collection

1. Samples are placed inside falcon centrifuge tubes and sealed to prevent sample movement and dehydration. 2. Place samples 70 cm from the detector. 3. For imaging, use a monochromatic beam with an X-ray energy ˚ ) with a 0.5 mm thick aluminum filter of 20.4 keV (0.608 A before the monochromator. 4. To create a three dimensional or computed tomography (CT) data set, record a series projection (two dimensional images) at 0.0999-degree step size (per projection image) for a rational angle of 180.096 . Use an exposure time of 0.92 s per image. 5. Record ten flat images (with the X-ray beam and no sample) and dark images (without beam and no sample) before and after recording the projection images of the sample.

3.5.3 Image Processing and Three-Dimensional Reconstruction

1. Reconstruct the CT data using UFO-KIT software (https:// bmit.lightsource.ca/user-guide/software/) [24] (see Note 14). 2. To begin reconstruction for absorption, correlate first and last projections to find the center of rotation in “ez_ufo”. 3. To remove ring artifacts (concentric rings in reconstructed images) apply the “ring removal 5” filter. 4. The histogram using the minimum and maximum values in the 32-bit histogram and select “convert to 8bit” before saving (see Note 15). 5. Select the “input” and “output” directories. 6. Assure raw tomograms are separated in “dark,” “flat,” and “tomo” folders (see Note 16). 7. Click “reconstruct.” The two-dimensional reconstructed images are based on absorption (Fig. 5a, e). 8. To reconstruct with phase retrieval, repeat the steps 2–6 and select enable Paganin/TIE phase retrieval using the following parameters: Energy ¼ 20.4 keV and pixel size ¼ 8.75 μm. 9. For two-dimensional phase contrast images (Fig. 5b, f): use sample to detector distance ¼ 70 cm, δ/β ¼ 100 and beam energy ¼ 20.4 keV. 10. The reconstructed images produce an 8-bit depth TIFF image stack with 544 images per stack.

Phenotyping Plant Cold Responses with FTIR and μCT

155

Fig. 5 Two- and three-dimensional X-ray images of a nonacclimated wheat crown. A (a) phase retrieved and (b) absorption XZ plane orthogonal slice from a wheat crown. The same crown, volume rendered (c) phase retrieved and (d) absorption images of a wheat crown digitally sliced to observe the shoot apical meristem. (e) Phase retrieved and (f) absorption images in the XY orthogonal slice of the same nonacclimated wheat crown. 3.5.4 CT Image Visualization and Volume Rendering

1. Image stacks are opened, visualized, and volume rendered in Avizo 9.7. 2. Use the “volume edit” tool to crop the falcon tube out of the image stack. First, place a cylindrical object around the outside of the winter wheat crown. Then cut the outside volumes. Only

156

Ian R. Willick et al.

the wheat crown structure (or tissue of interest) should remain in the image stack. 3. Create an “ortho slice” object to visualize and compare slices setting the histogram min and max values to 115 and 165, respectively, for the absorption images and 135 and 278, respectively, for the phase contract images to remove the air pixel from the images. 4. Create a “volume rendering” object and adjust the histogram min and max values. 5. On the same sample, create three-dimensional cutaways with the “volume edit” object and cut away using a box object and cutting away the inside volume (Fig. 5c, d). 6. Save images and video with the snapshot and movie maker tool. 7. Demonstration videos (Videos 1 and 2) are available with this chapter at link.springer.com. These videos compare phase retrieval and absorbance of: (a) 3D cutaway of Fig. 5c, d; and (b) cross-sectional images from the upper to the lower winter wheat crown of Fig. 5e, f. 3.5.5 Tradeoffs Between Absorbance and Phase-Contrast Imaging

Absorbance images are produced primarily from density changes in a medium. This is advantageous for relatively low energy experiments where lower energy photons are more readily absorbed. Phase contrast is technically present in the absorption data. However, algorithms are required to “retrieve” the phase, especially at lower energies. At lower energies, absorbance dominates the images and phase cannot be seen without these algorithms. As the sample to detector distance increases so does the phase contrast. At higher energies, phase remains constant while absorption decreases. In our case, 20.4 keV is considered a low energy. Increasing the sample to detector distance increases the phase. Additionally, phase is produced from a difference in refractive indices and highlights the interface between two media such as pores in samples, vascular tissue, the interface between air and the sample or subtle variations in tissue density. The phase algorithm applies a mean filter. Data is smoothed in comparison to the absorption images which facilitates visualization of homogenous tissues. In the process, we lose some pixel resolution. It is a trade-off to observe the subtle variations in plant material (in phase images) in comparison with pixel resolution (in absorbance images). As a result, there is less noise in the phase images and this is apparent in the orthogonal cross sections. Phase contrast also has an advantage of making segmentation more efficient as material boundaries have more prominent borders. These trade-offs should be evaluated depending on the beam energy used, the type of sample, as well as the type of morphological information one hopes to glean from the data set.

Phenotyping Plant Cold Responses with FTIR and μCT

4

157

Notes 1. Based on the first author’s personal observations with hydroponic systems, cereals grow optimally in aerated half-strength Hoagland solution with the pH ranging from 6.5 to 7.5. Tank solution pH should be tested every 4 days. 2. Temperature and duration of exposure will significantly affect the plant’s capacity to tolerate and avoid freezing. Conduct a time response experiment to determine the length of time required to reach maximum freezing tolerance for the species and cultivar of choice. Forty-two days at 4  C are required to reach maximum freezing tolerance for Norstar winter wheat. 3. The stainless steel die consists of a base, plunger, and optically polished anvils. 4. Fresh samples allow for nondestructive sampling, but timing of sample analysis is dependent on the availability of plant material. Ideally, use intact plants and not cuttings. Wilting and cut samples can result in oxidative stress responses and subsequent artifacts in absorbance spectra. 5. Sample tilting in complex organs such as wheat crowns or vegetative buds by as little as 10 can result in artifacts. Avoid fixing samples in paraffin wax or hard resin. These substances introduce artifacts in the resulting FTIR spectra. 6. Water surrounding the mold when frozen will hold the sample in place during sectioning. Quick submergence of the entire mold and sample will result in the loss of water surrounding the sample and the formation of air pockets. Both will increase the difficulty of removing an intact frozen block from the mold. 7. Fix and section samples on the same day. Storing samples for more than 2 h at
20 or
80  C can damage tissue morphology and alter sample chemistry. 8. Avoid fixing samples in ethanol, formaldehyde or its derivatives, methanol, glacial acetic acid etc. Fixation and dehydration irreversibly modify plant tissue composition creating artifacts in the resulting spectra. 9. This allows all material to equilibrate to the chamber temperature. Warm instruments can tare sections. 10. Clean the ATR probe with pure acetone and a Kim-wipe prior to and between samples. 11. Most ATR probes have a penetration depth between 2 and 4 μm. Resulting spectra may be representative of both the epicuticular and cuticular layers. If this is an issue, remove wax layers with a methanol-based solvent. Collected wax can

158

Ian R. Willick et al.

then be analyzed using bulk FTIR spectroscopy to confirm ATR-FTIR results. 12. Depending on the spectromicroscope, cross-sectional dimensions of the beam will vary between 2 and 16 μm. Synchrotron radiation will also vary in intensity across the cross-sectional area. Spectra analysis should not occur until images are standardized. To avoid uniformity issues, use FPA mapping with a glowbar light source. Collect spectra with a single point array detector using synchrotron radiation to produce spectra with a high signal-to-noise ratio. 13. Poor sample preparation due to inadequate grinding time, contaminated samples/slides, or contaminated background spectra result in sloped spectra. Experience in preparing samples and attention to detail will minimize this issue. 14. UFO is a multithreaded, GPU-enabled and distributed data processing framework for CT data reconstruction. It uses a filtered back projection algorithm and Paganin phase retrieval using the transport of intensity equation (TIE). 15. The images are collected as 16-bit tiff files. UFO converts the files to 32-bit tiff files upon initial reconstruction. The 32-bit tiffs are large and do not contain more data than the original 16-bit images. A conversion to 8-bit saves memory and no information is lost. Clipping the histogram to only include the values with grey scale further saves memory and saves all required information. 16. UFO uses the average of the flat and dark images to correct and normalize the images.

Acknowledgments The authors acknowledge funding support from the National Science and Engineering Council CRD and Discovery grants, the Saskatchewan Ministry of Agriculture and the CanadaSaskatchewan Growing Forward 2 bilateral agreement as well as the Saskatchewan Wheat Development Commission. Part of the research described in this paper was performed at the Canadian Light Source, a national research facility of the University of Saskatchewan, which is supported by the Canada Foundation for Innovation (CFI), the Natural Sciences and Engineering Research Council (NSERC), the National Research Council (NRC), the Canadian Institutes of Health Research (CIHR), the Government of Saskatchewan, and the University of Saskatchewan.

Phenotyping Plant Cold Responses with FTIR and μCT

159

References 1. Gusta LV, Chen THH, Fowler DB (1983) Factors affecting the cold hardiness of winter wheat. In: Fowler DB, Gusta LV, Slinkard AE et al (eds) New frontiers in winter wheat production. The University of Saskatchewan Printing Services, Saskatoon, Canada, pp 1–25 2. Livingston DP III, Henson CA (1998) Apoplastic sugars, fructans, fructan exohydrolase, and invertase in winter oat: responses to second-phase cold hardening. Plant Physiol 116:403–408 3. Trischuk RG, Schilling BS, Low NH et al (2014) Cold acclimation, de-acclimation and re-acclimation of spring canola, winter canola and winter wheat: the role of carbohydrates, cold-induced stress proteins and vernalization. Environ Exp Bot 106:156–163 4. Griffith M, Yaish MW (2004) Antifreeze proteins in overwintering plants: a tale of two activities. Trends Plant Sci 9:399–405 5. Kasuga J, Hashidoko Y, Nishioka A et al (2008) Deep supercooling xylem parenchyma cells of katsura tree (Cercidiphyllum japonicum) contain flavonol glycosides exhibiting high anti-ice nucleation activity. Plant Cell Environ 31:1335–1348 6. Uemura M, Steponkus PL (1994) A contrast of the plasma membrane lipid composition of oat and rye leaves in relation to freezing tolerance. Plant Physiol 104:479–496 7. Tanino KK, Kobayashi S, Hyett C et al (2013) Allium fistulosum as a novel system to investigate mechanisms of freezing resistance. Physiol Plant 147:101–111 8. Willick IR, Takahashi D, Uemura M et al (2018) Tissue-specific changes in apoplastic proteins and cell wall structure during cold acclimation of winter wheat crowns. J Exp Bot 69:1221–1234 9. Ishikawa M, Sakai A (1981) Freezing avoidance mechanisms by supercooling in some Rhododendron flower buds with reference to water relations. Plant Cell Physiol 22:953–967 10. Willick IR, Gusta LV, Fowler DB et al (2019) Ice segregation in the crown of winter cereals: evidence for extraorgan and extratissue freezing. Plant Cell Environ 42:701–716 11. McCann MC, Hammouri M, Wilson R et al (1992) Fourier transform infrared microspectroscopy is a new way to look at plant cell walls. Plant Physiol 100:1940–1947 12. Lahlali R, Jiang Y, Kumar S et al (2014) ATR–FTIR spectroscopy reveals involvement of lipids and proteins of intact pea pollen grains to heat stress tolerance. Front Plant Sci 5:747 13. McCann MC, Chen L, Roberts K et al (1997) Infrared microspectroscopy: sampling

heterogeneity in plant cell wall composition and architecture. Physiol Plant 100:729–738 14. Lahlali R, Karunakaran C, Wang L et al (2015) Synchrotron based phase contrast X-ray imaging combined with FTIR spectroscopy reveals structural and biomolecular differences in spikelets play a significant role in resistance to Fusarium in wheat. BMC Plant Biol 15:24 15. Tanino K, Willick IR, Hamilton K et al (2017) Chemotyping using synchrotron mid-infrared and X-ray spectroscopy to improve agricultural production. Can J Plant Sci 97:982–996 16. Lee Y, Karunakaran C, Lahlali R et al (2017) Photoperiodic regulation of growth-dormancy cycling through induction of multiple bud–shoot barriers preventing water transport into the winter buds of Norway Spruce. Front Plant Sci 8:2109 17. Willick IR, Lahlali R, Vijayan P et al (2018) Wheat flag leaf epicuticular wax morphology and composition in response to moderate drought stress are revealed by SEM, FTIRATR and synchrotron X-ray spectroscopy. Physiol Plant 162:316–332 18. Livingston DP III, Henson CA, Tuong TD et al (2013) Histological analysis and 3D reconstruction of winter cereal crowns recovering from freezing: a unique response in oat (Avena sativa L.). PLoS One 8:e53468 19. Karunakaran C, Lahlali R, Zhu N et al (2015) Factors influencing real time internal structural visualization and dynamic process monitoring in plants using synchrotron-based phase contrast X-ray imaging. Sci Rep 5:12119 20. Mayo SC, Chen F, Evans R (2010) Micronscale 3D imaging of wood and plant microstructure using high-resolution X-ray phasecontrast microtomography. J Struct Biol 171:182–188 21. Mooney SJ, Pridmore TP, Helliwell J et al (2012) Developing X-ray computed tomography to non-invasively image 3-D root systems architecture in soil. Plant Soil 352:1–22 22. Demsar J, Curk T, Erjavec A et al (2013) Orange: data mining toolbox in python. J Mach Learn Res 14:2349–2353 23. Gauch HG (1972) Inorganic plant nutrition. Dowden, Hutchinson and Ross Inc., Stroudburg, PA, 488 pp 24. Vogelgesang M, Farago T, Morgenever TF et al (2016) Real-time image content based beamline control for smart 4D X-ray imaging. J Synchrotron Radiat 23:1254–1263 25. Kacurakova M, Capek P, Sasinkova V et al (2000) FT-IR study of plant cell wall model compounds: pectic polysaccharides and hemicelluloses. Carbohydr Polym 43:195–203

Chapter 12 Proteomic Approaches to Identify Proteins Responsive to Cold Stress Anna M. Jozefowicz, Stefanie Do¨ll, and Hans-Peter Mock Abstract Changing environmental conditions greatly affect the accumulation of many proteins; therefore, the analysis of alterations in the proteome is essential to understand the plant response to abiotic stress. Proteomics provides a platform for the identification and quantification of plant proteins responsive to cold stress and taking part in cold acclimation. Here, we describe the preparation of proteins for LC-MS measurement to monitor the changes of protein patterns during cold treatment in Arabidopsis thaliana. In our protocol, proteins are precipitated using TCA/acetone, quantified with 2D Quant Kit and digested with trypsin using a filter-based method and analyzed using an LC-MS approach. The acquired results can be further applied for label-free protein quantification. Key words Cold stress, Soluble proteins, Mass spectrometry, Protein digestion, Protein quantification

1

Introduction Environmental stresses such as high light, drought, or low temperature reduce growth and yield of plants. Cold stress affects plant metabolism and physiology and has therefore great impact on plant performance. In order to survive low temperatures plants originating from temperate zones developed a broad range of physiobiochemical responses and defense mechanisms [1], which are reflected by the changes in the protein abundance. Proteins that are differentially regulated in response to cold stress are involved in many vital processes such as signal transduction, translation, reactive oxygen species scavenging, photosynthesis, photorespiration, C/N metabolisms, and energy production. Cold acclimation involves synthesis of proteins such as dehydrins, antifreeze proteins, heat shock proteins, and cold shock domain proteins [2]. Proteome analysis by mass spectrometry (MS) has greatly contributed to the characterization of new proteins involved in cold stress response in various plant species. The complexity of the

Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_12, © Springer Science+Business Media, LLC, part of Springer Nature 2020

161

162

Anna M. Jozefowicz et al.

Sample preparation

Cultivation

proteomic mixtures is the reason that a separation method prior to analysis by mass spectrometry is required. The application of the classical 2D electrophoresis and 2D DIGE resulted in identification and quantification of cold responsive proteins in Arabidopsis [3, 4], rice [5], wheat [6], and barley [7]. Disadvantages of this technique are the high protein amounts needed, the limited dynamic range, and difficulties to resolve low abundant and hydrophobic proteins [8]. Gel-free methods were established to overcome some of these restrictions. The bottom-up approach is based on the digestion of a complex protein mixture into peptides. The peptides are separated by nanoscale reversed-phase liquid chromatography (nano-LC) and eluted into a tandem mass spectrometer. The LC-MS technique was recently applied to resolve the proteome of tobacco [9], rice [10], and Arabidopsis [11]. Here, we present the protocol for protein TCA/acetone extraction [12], 2D Quant based quantification, tryptic digestion [13], and MS analysis of proteins from coldtreated Arabidopsis thaliana, as presented in Fig. 1. For protein

10 °C

20 °C

20 °C - control Sowing

10 °C - stress

Repotting

Treatment

Leaf material

Protein mixture

Peptides

TCA/acetone

Tryptic

precipitation

digestion

LC-MS acquisition

nanoLC separation Cy5

Intens. x107 3

MS and MS/MS acquisition

2

Intens. x106

1 0

C_o.Phase_Tryp_RB3_01_3952.d: +MS, 37.6-43.4min #11781-14564 2+ 583.3017

2+ 441.2425

8

2+ 404.2054

6 2+ 506.7586

0

20

40

60

80

Protein identification (e.g. Mascot)

100

4

120 2

0

2+ 610.8051

2+ 718.3729

Time [min]

3+ 784.3688

1+ 559.2901

2+ 324.7074

300

2+ 669.8367

400

500

600

700

800

m/z

Fig. 1 Workflow for protein identification from cold treated Arabidopsis thaliana plants including plant cultivation, sample preparation, and LC-MS data acquisition

Cold Responsive Proteins

163

identification, we describe protocols applied for the instruments present in our lab. They can be easily adapted to the instruments from other vendors.

2

Materials

2.1 Plant Cultivation and Treatment

2.2 Protein Precipitation

Equipment for controlled plant growth (standardized substrate, pots or multiwell trays, plant incubators with light and temperature controls). 1. Precipitation solution: 10% (w/v) trichloroacetic acid (TCA), 0.07% (w/v) 2-mercaptoethanol (2-ME) in acetone. For one sample: mix 1 g of TCA, fill to 10 mL with cold acetone and add 6.3 μL of 2-ME. Prepare directly before use. 2. Washing solution: 0.07% (w/v) 2-ME in acetone. For one sample: take 20 mL of cold acetone and add 12.6 μL 2-ME. Prepare directly before use. 3. Liquid nitrogen.

2.3 Protein Resuspension and Quantification

1. Ultrapure water. 2. 1 M dithiothreitol (DTT) stock: resuspend 1.54 g in 10 mL of water. Aliquot in 1 mL portions and store at 20  C. 3. Lysis buffer: 7 M urea, 2 M thiourea, 5 mM DTT, 2% CHAPS, pH 8.0. For 50 mL, mix 21 g urea, 7.6 g thiourea, and 1 g of CHAPS and fill up to 40 mL. To dissolve urea prepare a 25  C water bath. When all the compounds are dissolved, add 250 μL of 1 M DTT and check the pH. If needed adjust to pH 8. Fill with water to 50 mL. 4. Amicon Ultrafree-MC 0.45 μm Filter. 5. 2-D Quant Kit (GE Healthcare Life Sciences, USA). 6. Spectrophotometer.

2.4

Protein Digestion

1. Microcon-10 kDa Centrifugal Filter Unit with Ultracel-10 membrane (Merck Millipore; Germany). 2. Urea buffer: 7 M urea, 100 mM Tris–HCl, 5 mM EDTA, pH 8.0. Buffer can be stored at room temperature. For 50 mL mix: 21 g urea, 307 mg Trizma hydrochloride, 370 mg Trizma base, and 29 mg of EDTA and fill up to 40 mL. To dissolve urea prepare a 25  C water bath. After all the compounds are dissolved check pH. If needed adjust to pH 8 and fill with water to 50 mL. 3. 1 M DTT in urea buffer: Mix 1.53 g of DTT with 10 mL urea buffer. Aliquot and store at 20  C.

164

Anna M. Jozefowicz et al.

4. 1 M 2-iodoacetamide (2-IAA) in urea buffer: Mix 1.85 g of 2-iodoacetamide with 10 mL urea buffer. Aliquot and store at 20  C. 5. Solution 1: 20 mM DTT in urea buffer. For 1 mL mix 20 μL of 1 M DTT with 980 μL of urea buffer. 6. Solution 2: 80 mM 2-IAA in urea buffer. For 1 mL mix 80 μL of 1 M 2-IAA with 920 μL of urea buffer. 2-IAA is light sensitive and should be kept in the dark. 7. Solution 3: 50 mM ammonium bicarbonate. Dissolve 0.2 g of ammonium bicarbonate in 50 mL of LC-MS grade water. 8. Trypsin stock solution: Prepare 200 ng/μL solution, aliquot and store at 80  C (e.g., Trypsin Gold, Mass Spectrometry Grade (Promega, USA)). 9. Tabletop centrifuge. 2.5 Sample Resuspension for LC-MS 2.6 Protein Identification by LC-MS

1. Sample resuspension solution: 1% (v/v) acetonitrile (ACN), 0.1% (v/v) formic acid (FA) in LC-MS grade water.

All solutions should be prepared using LC-MS grade solvents. 1. Solvent A: 0.1% FA in water. 2. Solvent B: 0.1% FA in ACN. 3. Loading Buffer: 0.1% trifluoroacetic acid (TFA) in water. 4. 10 mM sodium formate. 5. Nano LC instrument: Dionex UltiMate 3000 RSLCnano System (Thermo Fisher Scientific, Dreieich, Germany). 6. MS instrument: Impact II ESI-QTOF mass spectrometer (Bruker Daltonics, Bremen, Germany) with a CaptiveSpray nanoBooster source. 7. Nano Trap Column: Acclaim PepMap100 C18, 5 μm, 100 A˚ (Thermo Fisher Scientific). 8. Analytical column: 50 cm  75 μm (Acclaim PepMap RSLC C18, Thermo Fisher Scientific). 9. LC-MS certified vials. 10. Software DataAnalysis v4.4, ProteinScape v4.0.

3

Methods

3.1 Plant Cultivation and Cold Treatment

1. Germinate seeds under standard conditions (e.g., for Arabidopsis: 9 h light with 140 μmol/m2/s light intensity, 20:18  C day–night). 2. After 2 weeks, transfer the seedlings into single pots and grow under conditions as described in step 1 (see Note 1).

Cold Responsive Proteins

165

3. Four weeks after sowing move half of the plants to 10  C (12 h light with 140 μmol/m2/s light intensity). The other half of the plants should be grown as described in step 1 but with 12 h of light period. Pay attention to identical conditions. Grow all plants together in one growing device and randomize positions. Ideally, use two identical growth devices for cold treatment and control (see Note 2 for additional hints for plant cultivation). 4. Sample the plants after 7 days of treatment. Cut off the aerial parts, freeze in liquid nitrogen and store at 80  C until the analysis. 3.2 Protein Precipitation

Wear gloves and clean lab coat while performing all steps of the experiment to avoid contamination with keratin and dust. For all glassware, growth containers, mortar and pestle, avoid the standard dishwasher and cleaning detergents for reduction of PEG contamination. 1. Grind material to a fine powder using a mortar and pestle precooled in liquid nitrogen. 2. Immediately after grinding, resuspend the powder with tenfold volume of cold Precipitation Solution (e.g., 1 g of plant powder with 10 mL of Precipitation Solution) in a 15 mL Falcon tube. 3. Aliquot the suspension equally to 2 mL reaction tubes using a pipette with a precut 1 mL tip. Keep the tubes on ice. Cool down the samples in liquid nitrogen for 15 s. 4. Incubate samples at 20  C for 45 min (this time might be prolonged up to 2 h). Mix the samples by inversion after 5, 10 and 15 min of incubation. 5. Centrifuge for 5 min, 36,000  g at 4  C (see Note 3). 6. Remove supernatant with a fine syringe coupled to a Bu¨chner flask and vacuum pump. 7. Resuspend the pellets in 1.5 mL of Washing Solution and vortex. 8. Incubate for 5 min in ultrasonic bath (see Note 4). 9. Cool the samples in liquid nitrogen for 2 min. 10. Incubate at

20  C for 30 min.

11. Centrifuge for 5 min, 36,000  g at 4  C (see Note 5). 12. Remove supernatant with the fine syringe coupled to a Bu¨chner flask and vacuum pump. 13. Repeat the washing procedure (steps 7–12). 14. Dry the pellets in the vacuum centrifuge for 15 min until the acetone is no longer perceptible.

166

Anna M. Jozefowicz et al.

3.3 Protein Resuspension and Quantification

1. Determine the weight of pellet and resuspend it in Lysis Buffer. For each 1 mg of pellet, add 50 μL of Lysis Buffer. Vortex until the pellet is soaked with buffer (see Note 6). 2. Keep in ultrasonic bath for 15 min. 3. Incubate samples at 37  C for 1 h with shaking. 4. Centrifuge for 15 min at 36,000  g at room temperature (RT) (see Note 5). 5. Transfer the supernatant to a new tube and discard the pellet. If the pellet leftovers are still in the supernatant clarify it by centrifugation through a 0.45 μm filter unit. 6. Samples can be directly used for protein quantification or stored at 20  C. 7. Prepare the working solution by mixing 100 parts of Color Reagent A and 1 part of Color Reagent B. For each single measurement 250 μL is required. Each sample should be measured with two different volumes and in duplicate (it means four repetitions for one sample). Additional 4 mL of color reagent are needed to prepare the calibration curve. 8. To prepare the standard curve use a Bovine Serum Albumin (BSA) 2 mg/mL solution. Set up 1.5 mL reaction tubes and pipet 0, 1.25, 2.5, 3.75, 5, and 6.25 μL of BSA solution in duplicate (see Note 7). 9. Prepare tubes containing 1–5 μL of sample (see Note 8). 10. Add 125 μL of precipitant to each tube, vortex, and incubate at RT for 2–3 min. 11. Add 125 μL of coprecipitant and vortex. 12. Centrifuge for 5 min at 36,000  g. 13. Aspire the supernatant completely (see Note 9). 14. Add 25 μL of cooper solution and 100 μL of ultrapure water. Vortex each sample about 10–20 s (see Note 10). 15. Centrifuge shortly to collect liquid on the bottom of the tube. 16. Add 250 μL of working solution to each tube, mix it by inversion, and centrifuge shortly. 17. Incubate at RT for 20 min. 18. Shortly before the end of incubation time, transfer 200 μL to a 96-well plate. 19. Read the absorbance at 480 nm using water as a blank. Calculate protein concentration in the sample using the BSA standard curve.

3.4

Protein Digestion

1. Fill the YM-10 filter unit with 200 μL of ultrapure water and centrifuge it at 14,000  g, 5 min, at RT. Carefully remove remaining water (see Note 11).

Cold Responsive Proteins

167

2. Mix 10 μg of protein with Lysis Buffer to a final volume of 200 μL and transfer it to the filter. 3. Centrifuge until there is no more than 10 μL of sample on the filter (about 45–60 min). 4. Apply 100 μL of Urea Buffer and centrifuge it until there is no more than 10 μL of liquid on the filter (about 25 min). To remove all detergents from the sample repeat this step three times. 5. Apply 300 μL of Solution 1. Incubate for 60 min with gentle shaking at 60  C. 6. Cool down the samples at RT and add 100 μL of Solution 2. Incubate at 37  C, 30 min in the darkness with gentle shaking (see Note 12). 7. Centrifuge the samples until no liquid remains (about 40–60 min). 8. Add 200 μL of Solution 3 and centrifuge it for about 20 min. Repeat this step three times (see Note 13). 9. For each sample, mix 200 μL of Solution 3 with trypsin stock solution. The recommended ratio of enzyme to protein is 1:50, therefore for 10 μg of protein 1 μL of Trypsin stock solution should be used. 10. Fill the filter with Trypsin solution and mix several times by gentle pipetting. Digest the proteins overnight (about 16 h) at 37  C. 11. Next day: Elute peptides into fresh reaction tube by centrifuging for about 30 min. 12. Wash the filter by adding 50 μL of Solution 3 and centrifuge it for about 10 min, until no liquid remains. Repeat this step three times to elute remaining peptides. 13. Dry the peptides in the vacuum centrifuge until no liquid remains. 14. Resuspend the pellet in sample resuspension solution to obtain a concentration suitable for loading onto a reverse-phase column. If 10 μg was taken for the digestion, you might use 50 μL of solution to obtain 0.2 μg/μL peptide concentration. Vortex and centrifuge for 5 min by 36,000  g. 15. Transfer the supernatant to the LC vials. 3.5 Protein Identification with LC-MS

Samples can be analyzed using the LC-MS/MS system of choice. Here we present the method developed for the Dionex UltiMate 3000 RSLCnano System coupled to a Bruker Impact II UHR-ESIQTOF mass spectrometer. 1. Use 0.6–0.8 μg of sample (3–5 μL) for an injection onto the nano-flow liquid chromatography system.

168

Anna M. Jozefowicz et al.

Table 1 A multistep gradient for separation of peptides using nano-LC reversedphase separation Time (min) % Mobile phase B Remarks 0

2

Initial conditions

0

2

Start of sample loading on the precolumn

5

2

Start of gradient

125

40

End of gradient

130

95

150

95

151

2

160

2

Column washing Column equilibration

2. Calibrate the instrument using 10 mM sodium formate. 3. Peptides are first desalted and concentrated on the trap column using loading buffer at a flow rate of 10 μL/min for 5 min. 4. Peptides are eluted to the analytical column using multistep gradient of ACN in water applying a flow rate of 300 nL/min (Table 1, see Note 14). The length of the gradient is 120 min, but due to the loading and column washing steps, the total time for an LC-MS/MS run is 160 min. 5. The eluent is sprayed into Impact II using the CaptiveSpray nanoBooster source. The source conditions are 150  C dry temperature, 1300 V capillary voltage, 0.2 bar nanoBooster, and a dry gas flow rate of 0.3 L/min. 6. For MS and MS/MS acquisition, use the predefined “Instant Expertise” method (Compass v. 1.9, Bruker). Acquire the m/z data in the range of 150–2200 and set the fixed total cycle time to 3.0 s. For the MS spectra the acquisition speed is 2 Hz with a collision energy of 7 eV. For MS/MS the acquisition speed is dependent on the precursor signal intensities and is set to 4 Hz for lower (2500 cts) and 16 Hz for higher (25,000 cts) intensities with linear adjustment for the precursors between low and high intensity. The collision energy is adjusted between 23 and 65 eV as a function of the m/z value. 7. Process raw data using appropriate software (e.g., DataAnalysis v4.4) and export the MS/MS spectra as XML file. 8. Upload the XML file to ProteinScape v.4.0. Using the Mascot search engine (v2.5.1) perform the database searches against an Arabidopsis thaliana database (TAIR10_pep_20101214). The search parameters applied are as follows: trypsin as protease, 15 ppm peptide mass tolerance, 0.05 Da fragment mass

Cold Responsive Proteins

169

tolerance, one missed cleavage allowed, carbamidomethylation (C) as fixed modification, oxidation (M) as variable modification, and peptide charges +1, +2, +3. Use a Mascot integrated peptide decoy database search. The false discovery rate was < 1% and ion score cutoff 15 and significance threshold p < 0.05. 9. Data can be further used for label-free quantification using software of choice (e.g., Progenesis QI for Proteomics [Nonlinear Dynamics, Newcastle upon Tyne, UK] or MaxQuant [14]).

4

Notes 1. Plan a sufficient number of biological replicates and experimental repetitions. Single plants offer more biological variability; pools can be used, but the number of replicates should not be lower than five. It is common to perform three independent experiments. 2. 10  C is appropriate to provoke a cold stress acclimation response in Arabidopsis thaliana; however, 4  C might also be used. As light stress is the major threat for the plants under cold stress, make sure that the light amount is the same for all plant replicates. Measure the light intensity at plant height with a light meter for both 10 and 20  C conditions and adjust it to the same value. Check also temperature at plant height as lamps may produce excessive heat. In addition, the choice of growing substrate (low- or high-nutrient) does strongly affect proteome composition. The other important factor for a consistent response toward cold stress is the plant age. The ideal precultivation time for Arabidopsis is 4–6 weeks when the plants have developed 8–10 leaves. Apply the 12 h light period for at least 1 week before onset of the low temperature treatment. 3. The 36,000  g is the maximum speed of our centrifuge. The centrifugation speed can be adapted to the available equipment, but the time of centrifugation might change. 4. The resuspension of pellets might be aided by stirring with a glass rod. 5. The time of centrifugation might be prolonged if the material is not pelleting easily. 6. A glass rod can be used to break the pellet. 7. For higher reproducibility of results, the use of Positive-Displacement pipettes (e.g. Microman, Gilson, USA) is advised. 8. Typically, we take 2.5 and 5 μL of sample, but a test assay for any new sample type is necessary to ensure that the protein

170

Anna M. Jozefowicz et al.

amount is appropriate to fall within the range of the standard curve. 9. Liquid must be removed completely. A cut gel-loading tip attached to the end of the tubing and coupled to a Bu¨chner flask and vacuum pump can be used. After solubilization samples can be stored up to 1 h. 10. The exact amount of cooper solution must be added to each sample. Use a multistep pipette or Microman pipette if available. 11. All centrifugation steps must be performed at room temperature as the solubility of urea drops with temperature. The centrifugation speed should not be higher than 14,000  g. High speed might break the filter and cause sample loss. 12. Iodoacetamide is light sensitive. A thermoshaker can be covered with aluminum foil to ensure proper conditions for alkylation. 13. It is necessary to remove urea from the samples as it suppresses ionization in the ESI source. 14. Always prepare fresh solutions for nano-LC-MS analyses. References 1. Heidarvand L, Amiri RM (2010) What happens in plant molecular responses to cold stress? Acta Physiol Plant 32:419–431 2. Yan SP, Zhang QY, Tang ZC et al (2006) Comparative proteomic analysis provides new insights into chilling stress responses in rice. Mol Cell Proteomics 5:484–496 3. Fanucchi F, Alpi E, Olivieri S et al (2012) Acclimation increases freezing stress response of Arabidopsis thaliana at proteome level. Biochim Biophys Acta 1824:813–825 4. Amme S, Matros A, Schlesier B et al (2006) Proteome analysis of cold stress response in Arabidopsis thaliana using DIGE-technology. J Exp Bot 57:1537–1546 5. Lee DG, Ahsan N, Lee SH et al (2009) Chilling stress-induced proteomic changes in rice roots. J Plant Physiol 166:1–11 6. Gharechahi J, Alizadeh H, Naghavi MR et al (2014) A proteomic analysis to identify cold acclimation associated proteins in wild wheat (Triticum urartu L.). Mol Biol Rep 41:3897–3905 7. Golebiowska-Pikania G, Kopec P, Surowka E et al (2017) Changes in protein abundance and activity involved in freezing tolerance acquisition in winter barley (Hordeum vulgare L.). J Proteomics 169:58–72

8. Zhu WH, Smith JW, Huang CM (2010) Mass spectrometry-based label-free quantitative proteomics. J Biomed Biotechnol 2010:840518 9. Hu RS, Zhu XX, Xiang SP et al (2018) Comparative proteomic analysis reveals differential protein and energy metabolisms from two tobacco cultivars in response to cold stress. Acta Physiol Plant 40:19 10. Lee J, Lee Y, Kim M et al (2017) Quantitative shotgun proteomic analysis of cold-stressed mature rice anthers. Plant Biotechnol Rep 11:417–427 11. Ma J, Wang DH, She J et al (2016) Endoplasmic reticulum-associated N-glycan degradation of cold-upregulated glycoproteins in response to chilling stress in Arabidopsis. New Phytol 212:282–296 12. Schlesier B, Mock HP (2006) Protein isolation and second-dimension electrophoretic separation. Methods Mol Biol 323:381–391 13. Distler U, Kuharev J, Navarro P et al (2014) Drift time-specific collision energies enable deep-coverage data-independent acquisition proteomics. Nat Methods 11:167–170 14. Tyanova S, Temu T, Cox J (2016) The MaxQuant computational platform for mass spectrometry-based shotgun proteomics. Nat Protoc 11:2301–2319

Chapter 13 Proteomic Approaches to Identify Cold-Regulated Plasma Membrane Proteins Md Mostafa Kamal, Daisuke Takahashi, Takato Nakayama, Yushi Miki, Yukio Kawamura, and Matsuo Uemura Abstract Plasma membrane is the primary determinant of freezing tolerance in plants because of its central role in freeze–thaw cycle. Changes in plasma membrane protein composition have been one of the major research areas in plant cold acclimation. To obtain comprehensive profiles of the plasma membrane proteomes and their changes during the cold acclimation process, a plasma membrane purification method using a dextran–polyethylene glycol two polymer system and a mass spectrometry-based shotgun proteomics method using nano-LC-MS/MS for the plasma membrane proteins are described. The proteomic results obtained are further applied to label-free protein semiquantification. Key words Cold acclimation, Plasma membrane, Nano-LC-MS/MS, Shotgun proteomics, Labelfree semiquantification, In-solution digestion, In-gel digestion

1

Introduction Many plants that grow in temperate and subarctic regions increase in freezing tolerance when exposed to a nonfreezing, low temperature [1], which is known as cold acclimation. Cold acclimation results in diverse alterations in plant cell physiology, morphology, and molecular biology [2, 3]. In many cases, freezing results in ice formation extracellularly (extracellular freezing) and plant cells must keep ice crystals from entering into the cytoplasm. Because the plasma membrane (PM) plays a central role in water transport between the inside and the outside of the cell and functions as a barrier for separation of the cytoplasm from the extracellular region, there is a consensus that stabilization of the PM is a prerequisite for survival under freezing stress [4–6]. Thus, it is reasonable to consider that the PM composition responds to low temperature and changes during cold acclimation in order to withstand the upcoming stresses incurred during a freeze–thaw cycle.

Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_13, © Springer Science+Business Media, LLC, part of Springer Nature 2020

171

172

Md Mostafa Kamal et al.

In fact, there are a number of reports that describe dynamic changes in PM components, both in protein and lipid compositions [7–12]. In addition, several studies demonstrated that specific proteins in the PM were functionally involved in cold acclimation and cold signaling [13–18]. With a recent advance of protein and peptide separation and identification techniques (such as nano– liquid chromatography), increased availability of protein analysis equipment (such as mass spectrometers) and development of user-friendly software and excellent genome/protein databases (such as MASCOT, SEQUEST, X! Tandem, and PEAKS), it became possible to reveal proteomic responses to cold acclimation on a large scale in a relatively short time. In this chapter, we introduce the procedures for cold acclimation, plasma membrane isolation, and protein mass analysis to identify cold-regulated proteins associated with the PM. The procedures described below are easily able to be adapted to Arabidopsis plants [19, 20] but, in general, it will be applied for other plants as experimental subjects. We have been using the procedures slightly modified for monocotyledonous plants such as rye and oat [21, 22] and Brachypodium [23]; woody plants such as poplar [24]; Arabidopsis suspension cultured cells [25]; and dicotyledonous plants including tomato [26], pea [27], and beet [28]. Lastly, detailed protocols for protein identification of the PM and microdomains in the PM using nano-LC MS/MS were described elsewhere [29] (Fig. 1).

2

Materials Prepare all solutions using ultrapure water (prepared by purifying deionized water by Millipore apparatus to attain a resistance of 18.2 MΩ cm at 24  C) except where noted otherwise and analytical grade reagents. Prepare and store all reagents at room temperature (unless indicated otherwise). Carefully follow all waste disposal regulations determined by local authorities when disposing of waste materials.

2.1

Plant Growth

1. Plant seeds: Arabidopsis seeds can be obtained from Arabidopsis stock centers such as ABRC, NASC, and SASSC (http:// www.arabidopsis.org/portals/mutants/stockcenters.jsp) or purchased from Lehle Seeds (Round Rock, TX, USA). Several accessions have been used but Columbia-0 (Col-0) is one of the most popular ecotypes for cold acclimation studies. 2. Plant bedding mix: two parts of vermiculite and one part of perlite. 3. Nutrient solution A (10 stock): 60 mM KNO3, 40 mM Ca (NO3)·4H2O, 20 mM NH4H2PO4, 10 mM MgSO4·7H2O.

Cold Acclimation and Plasma Membrane Proteins

CA plants Arabidopsis grown at 2rC

In-solution Tryptic digestion

or

In-gel Tryptic digestion

Up to 7 days

1 day

2-3 days

Peptide purification with SPE C-TIP

Days to weeks

Label-free quantification of proteins identified

1-3 days

Data mining with several tools

Days to months

Data processing

Data acquisition with nano-LC-MS/MS

Sample preparation

Plasma membrane purification

3 weeks

Plant cultivation

NA plants Arabidopsis grown at 23rC

173

Fig. 1 A representative workflow for shotgun proteomics of Arabidopsis PM proteins. It consists of three parts; plant cultivation, sample preparation, and data processing. All plants are grown at 23  C for 3 weeks (NA plants) and some of the NA plants are then transferred to a cold chamber for cold acclimation treatments (CA plants). Plasma membrane (PM) fractions are isolated using a two-phase partition system, and proteins in the PM fractions are digested with trypsin. Subsequently, the obtained peptides are purified and concentrated with SPE-C-TIP. Peptides are then subjected to a nano-LC-MS/MS system and software for label-free identification and semiquantification of PM proteins is used. Data obtained are processed with several tools for mining of novel and/or known factors

4. Nutrient solution B (50 stock): 25 mM KCl, 12.5 mM H3BO3, 1 mM MnSO4·5H2O, 1 mM ZnSO4·7H2O, 0.25 mM CuSO4·5H2O, 0.25 mM H2MoO4. 5. Nutrient solution C (50 stock): 8 mM Na2 EDTA, 8 mM FeSO4·7H2O. 6. Working nutrient solution: mix the nutrient solution A (1 L), B (200 mL) and C (200 mL) and add tap water (8.6 L) to make 10 L working solution.

174

Md Mostafa Kamal et al.

2.2 Plasma Membrane Purification

Several items, including 2 L of ultrapure water, a Polytron homogenizer, centrifuge rotors, and ultracentrifuge rotors, should be precooled in a refrigerator. 1. Homogenizing medium: 0.5 M sorbitol, 50 mM Mops-KOH (pH 7.6), 5 mM EGTA, 5 mM EDTA, 2 mM phenylmethanesulfonyl fluoride (PMSF), 4 mM salicylhydroxamic acid (SHAM), 2.5 mM 1,4-dithiothreitol. Store at 4  C (see Note 1). 2. Polytron generator (PT10SK, Kinematica Inc., Lucerne, Switzerland). 3. Cheesecloth. 4. Microsome (MS)-suspension medium: 10 mM KH2PO4/ K2HPO4 (K-P) buffer (pH 7.8), 0.3 M sucrose. Store at 4  C (see Note 2). 5. Electric Teflon–glass homogenizer. 6. NaCl medium: 100 mM NaCl in MS-suspension medium. Store at 4  C. 7. Plasma membrane (PM)-suspension medium: 10 mM MopsKOH (pH 7.3), 1 mM EGTA, 0.3 M sucrose. Store at 4  C. 8. Two-phase medium: weigh 1.4 g of polyethylene glycol 3350 and 1.4 g dextran in a 40 mL centrifuge tube (5.6% [w/w] polymers in final solution with microsomal suspensions). Add 9.4 mL MS-suspension medium and 7.3 mL NaCl medium (30 mM NaCl in final solution) to the centrifuge tube and mix well by shaking. Prepare three tubes per sample. Store in a refrigerator overnight to completely dissolve the polymers. 9. Bio-Rad Protein Assay Kit (Bio-Rad Laboratories, Hercules, CA): store at 4  C.

2.3

Tryptic Digestion

2.3.1 In-Solution Digestion

All preparations must be carefully performed in a clean bench with gloves and a clean lab coat to avoid contamination from keratin, dust, and other exogenous proteinaceous materials. 1. UTU buffer: 6 M urea, 2 M thiourea, 10 mM Tris–HCl, pH 8.0. Weigh 3.6 g urea, 1.52 g thiourea, and 12.1 mg Tris base, dissolve in 9 mL water and adjust the pH to 8.0 with 1 N HCl. Fill to 10 mL with water. 2. Reduction buffer: dissolve 7.7 mg dithiothreitol (DTT) in 10 mL of water (see Note 3). 3. Alkylation buffer: dissolve 50 mg iodoacetamide (IAA) in 10 mL of water (see Note 3). 4. Tris–HCl: 10 mM Tris–HCl, pH 8.0. Dissolve 0.121 g of Tris base in 90 mL water and adjust the pH to 8.0 with 1 N HCl. Adjust the final volume to 100 mL.

Cold Acclimation and Plasma Membrane Proteins

175

5. Lys-C stock solution: prepare 0.2 μg/μL mass spectrometry grade lysyl endopeptidase (Wako, Japan) stock solution in 10 mM Tris–HCl, pH 8. Aliquot in small volumes in microtubes and store at 20  C. 6. Trypsin stock solution: prepare 0.2 μg/μL modified sequencing grade trypsin (Promega) stock solution in 10 mM Tris– HCl, pH 8.0. Aliquot in small volumes in microtubes and store at 20  C. 7. Acidification solution: 10% trifluoroacetic acid (TFA). Add 10 μL TFA to 90 μL water, mix well (see Note 4). 2.3.2 In-Gel Digestion

1. SDS gel-loading buffer (2): 50 mM Tris–HCl (pH 6.8), 2% (w/v) SDS (sodium dodecyl sulfate), 3% (w/v) dithiothreitol (DTT), 0.2% (w/v) bromophenol blue (BPB), 10% (v/v) glycerol. Store at 4  C for instant use or at 20  C for long term use (see Note 3). 2. Fixation buffer: add 40 mL ethanol, 10 mL acetic acid to 50 mL of water and mix well. Store at 4  C. 3. Digestion buffer/100 mM ammonium bicarbonate (ABC) buffer: dissolve 0.79 g of ammonium bicarbonate in 90 mL water and adjust the final volume to 100 mL. Store at 4  C (see Note 5). 4. Acetonitrile (ACN): LC-MS grade acetonitrile. Always keep out of light and prevent evaporation. Caution! ACN is volatile and highly toxic. Always use under a fume hood. 5. ACN (50:50) buffer: 25 mM ABC and 50% (v/v) ACN. Add 25 mL ABC buffer to 25 mL of water. Store at 4  C (see Note 5). 6. Reduction buffer: dissolve 7.7 mg DTT in 5 mL of water (see Note 3). 7. Alkylation buffer: dissolve 51 mg in iodoacetamide (IAA) in 5 mL of water (see Note 3). 8. Trypsin solution: add 2 mL ABC to a vial containing 20 μg of trypsin (sequence grade modified, Promega). Mix well and aliquot in small volume in microtube and store at 20  C. 9. TFA-ACN solution: add 450 μL ACN to 450 μL of water in a 1.5 mL tube. Quickly add 50 μL TFA and mix well (see Note 4). 10. Acidification solution (0.1% TFA): add 1 μL TFA to 999 μL of water and mix well (see Note 4).

2.4 Peptide Purification

1. SPE C-TIP T-300 (Nikkyo Technos Co., Ltd., Tokyo, Japan) or any C-18 stop and go extraction (Stage) tip. 2. 1.5 mL microtubes: make a hole of 3 mm in diameter in the cap with a soldering iron. Prepare two tubes per sample.

176

Md Mostafa Kamal et al.

3. Solution A for C-18 stage-tips: 5% ACN and 0.1% TFA: add 50 μL ACN, 1 μL TFA, and 949 μL water, mix well (see Note 4). 4. Solution B for C18 stage tips: 80% ACN and 0.1% TFA: Add 800 μL ACN, 1 μL TFA, and 199 μL of water, mix well (see Note 4). 5. 0.1% TFA solution: quickly add 1 μL TFA into 999 μL of water, mix well (see Note 4). 2.5 Instruments and Software for Nano-LC-MS/MS

1. LC instrument: ADVANCE UHPLC system (MICHROM Bioresources, Auburn, CA, USA). 2. MS instrument: LTQ Orbitrap XL mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). 3. Ion source: ADVANCE Bioresources).

spray

source

(MICHROM

4. Trap column for peptide concentration: L-column Micro 0.3  5 mm (CERI, Japan). 5. Column for peptide separation: Magic C18 AQ nanocolumn (0.1  150 mm; MICHROM Bioresources). 6. Data conversion software: Proteome Discoverer (ver. 2.1, Thermo Fisher Scientific). 7. Search engine for protein identification: MASCOT search engine (version 2.6.2, Matrix Science, London, UK).

3

Methods

3.1 Plant Growth and Cold Acclimation

1. Plant seeds in a moist vermiculite–perlite mix (2:1) in plastic pots and place the pots in a controlled-environment chamber at 23  C under continuous light (100 μmol/m2/s). 2. Add nutrient solution occasionally from the bottom of the pot (usually twice a week). 3. Grow for approximately 3 weeks to obtain nonacclimated plants (see Note 6). 4. Cold acclimate by transferring nonacclimated plants to a cold growth chamber at 2  C under 12-h light condition (100 μmol/m2/s) for up to 7 days.

3.2 Plasma Membrane Purification (Fig. 2)

Wear gloves and a clean lab coat throughout the experiments to avoid contamination by keratin, dust and other exogenous proteinaceous materials. It is preferable to use low protein absorption microtubes at all stages. Perform all steps on crushed ice (unless indicated otherwise). Centrifuges should be prechilled at 4  C. 1. Cut off the aerial parts of Arabidopsis seedlings and weigh them (10 g or more in fresh weight is desirable for the plasma

Cold Acclimation and Plasma Membrane Proteins

homogenize

filter through four layers of cheesecloth

centrifuge

177

collect supernatant and ultracentrifuge two times

collect pellet and subject a twophase partition system

collect the pellet

collect upper phase and ultracentrifuge two times

repeat two-phase partition (three times in total)

collect the upper phase and mix with newly prepared lower phase

Fig. 2 A schematic overview of PM extraction from plants. Leaves are homogenized in a homogenizing medium and then passed through four layers of cheesecloth to remove debris. Subsequently, the filtrates are centrifuged at 5000  g and subsequently at 231,000  g to obtain microsomal fractions. Microsomal fractions are suspended in MS-suspension medium and then recentrifuged twice for washing. The resultant microsomal fractions are subjected to a two-phase partition system that consists of polyethylene glycol 3350 and dextran T500 in MS-suspension medium with NaCl. After repeating two phase partitioning three times to increase the purity of the PM in the upper phase, PM fractions are recovered, diluted with PM-suspension medium and centrifuged (231,000  g) twice to remove polymers

membrane purification). Put the harvested plant material in a plastic container and wash with chilled distilled water. Wash twice and then drain on paper towels. Keep the harvested plants wrapped with paper towels on crushed ice. 2. Put plant samples into four volumes of chilled homogenizing medium and cut into small pieces with a pair of scissors. 3. Homogenize with a chilled Polytron generator until the samples are broken into tiny pieces (speed 5–6 for 60–90 s). Filter the homogenates through four layers of cheesecloth and squeeze thoroughly. Put the filtrate into 40 mL centrifuge tubes. 4. Centrifuge at 5000  g for 15 min with a chilled rotor to remove debris and heavy membrane fractions. Transfer the supernatants into ultracentrifuge tubes by decantation. Discard precipitates. 5. Centrifuge at 231,000  g for 50 min with a chilled ultracentrifuge rotor to precipitate microsome fractions. Discard supernatants by decantation.

178

Md Mostafa Kamal et al.

6. Add appropriate volume of MS-suspension medium to each tube (usually 2–3 mL per tube) and homogenize the pellets with a Teflon–glass homogenizer. Collect the microsomal suspensions with a large-aperture Pasteur pipette into ultracentrifuge tubes. Balance ultracentrifuge tubes in pairs with MS-suspension medium. 7. Ultracentrifuge at 231,000  g for 50 min as described in step 5. After centrifugation, discard the supernatant with an aspirator. 8. Put 5 mL of MS-suspension medium in a Teflon–glass homogenizer and mark the solution surface on the glass homogenizer as an indication of 5 mL volume. Discard the medium. 9. Add 2 mL of MS-suspension medium onto microsomal pellets in the ultracentrifuge tubes. Break up the precipitated pellets with a glass rod. Transfer into a Teflon–glass homogenizer using a large-aperture Pasteur pipette. Put 2 mL of MS-suspension medium into the same ultracentrifuge tubes and break-up the remaining pellets by pipetting. Transfer into the Teflon–glass homogenizer already containing the first part of the resuspended pellet and add MS-suspension medium up to 5 mL. Homogenize well with an electric Teflon–glass homogenizer (moving up and down five times) on ice (see Note 7). 10. Put all of the homogenate in a centrifuge tube containing two-phase partition medium (tube A). Add 5 mL of MS-suspension medium to the other two two-phase partition mixtures (tubes B and C). Chill on crushed ice for 10 min. During this time, mix well every 2 min. 11. Centrifuge tubes A and B at 650  g for 5 min in a chilled rotor. Two phases should be observed to have settled in both tubes. Discard the upper phase of tube B with a Pasteur pipette and transfer the upper phase of tube A into tube B. Chill on crushed ice for 10 min. During this time, mix well every 2 min (see Note 8). 12. Centrifuge tubes B and C at 650  g for 5 min in a chilled rotor. Discard the upper phase of tube C with a Pasteur pipette and transfer the upper phase of tube B into tube C. Balance tube C with another centrifuge tube filled with water. Chill on crushed ice for 10 min. During this time, mix well every 2 min (see Note 8). 13. Centrifuge at 440  g for 5 min and split the resultant upper phase of tube C into two ultracentrifuge tubes. Fill up the tubes with PM-suspension medium and balance them. Ultracentrifuge at 231,000  g for 50 min, as described in step 5 (see Note 8).

Cold Acclimation and Plasma Membrane Proteins

179

14. Discard the supernatant with an aspirator. Add 1 mL of PM-suspension medium to each tube. Homogenize the pellets with a glass rod. Transfer into an electric Teflon–glass homogenizer and homogenize well (moving up and down five times) with cooling on ice. Collect the plasma membrane suspensions with a Pasteur pipette into ultracentrifuge tubes. Balance ultracentrifuge tubes in pairs with PM-suspension medium. Ultracentrifuge again at 231,000  g for 35 min. 15. Discard the supernatant with an aspirator. Add minimal volume of PM-suspension medium to the plasma membrane pellets. Homogenize the pellets with a glass rod. Transfer into an electric Teflon–glass homogenizer and homogenize well (moving up and down five times) with cooling on ice. Transfer into a 1.5 mL microtube. 16. Measure protein content using the Bradford assay (Bio-Rad Protein Assay Kit). Use 100 μg of protein for in-solution tryptic digestion and LC-MS/MS analysis. Part of the remaining PM fractions should be divided into aliquots for in-gel tryptic digestion, other part should be frozen in liquid nitrogen immediately and stored at 80  C. 3.3 In-Solution Tryptic Digestion

All of these procedures must be performed at a clean bench whenever possible and at room temperature unless otherwise specified. 1. Precipitate 100 μg of PM protein by ultracentrifugation (231,000  g, 4  C, 50 min). 2. Discard supernatant by decantation. Resuspend the final pellet with minimal volume of UTU buffer. Homogenize the pellets with a glass rod. Transfer into an electric Teflon–glass homogenizer and homogenize well (moving up and down five times) with cooling on ice. Transfer to 1.5 mL microtubes. 3. Solubilize samples and measure protein concentration with Bradford assay (Bio-Rad) kit according to the instruction manual. 4. Transfer 5–10 μg of PM protein to another 1.5 mL microtube. Make up to 20 μL with UTU buffer. Store the remaining PM protein at 80  C. 5. Reduction: add 0.15–0.3 μL of reduction buffer (0.15 μL reduction buffer for every 5 μg protein; not necessary to be a precise ratio) and incubate for 30 min at room temperature. 6. Alkylation: add 0.15–0.3 μL alkylation buffer and incubate in the dark for 20 min at room temperature. 7. Predigestion: add lysyl endopeptidase in an enzyme to protein ratio of 1:100 (w/w) and incubate for 3 h at room temperature in the dark.

180

Md Mostafa Kamal et al.

8. Digestion: dilute the sample 4- to 4.5-fold using 10 mM Tris– HCl and digest with modified sequencing grade trypsin at a 1:100 (w/w) ratio of enzyme to protein and incubate overnight at room temperature (see Note 9). 9. Stopping the digestion: add TFA to a final concentration of 0.2–0.5% (v/v) to bring the pH  3.0 using 10% TFA stock solution (see Note 10). 10. Vacuum concentration: Concentrate the digest to 20 μL using a vacuum concentrator at room temperature. 3.4 In-Gel Tryptic Digestion

1. Loading protein into gel: Resuspend 5–10 μg of PM protein in an equal volume of SDS sample loading buffer (keep the total volume within 20 μL) and vortex. Spin the tubes briefly and heat at 95  C for 5 min and spin again. Cool the tubes to room temperature. Load the protein sample in a precast 7–10% gradient polyacrylamide gel (e.g., PAGEL by ATTO). Keep one well empty in between two samples to prevent samples from mixing. Keep one well empty in between two samples to prevent samples from mixing. Electrophorese at 100 V with constant current until the sample dye (bromophenol blue (BPB)) enters 2 mm in the well (Fig. 3). The purpose of this brief electrophoresis is to concentrate and keep all proteins in a gel band and to remove unwanted contaminants (i.e., nonproteinaceous substances). Electrophorese at 100 V with constant current until the sample dye (BPB) enters 2 mm in the well (Fig. 3). The purpose of this brief electrophoresis is to concentrate the protein and remove unwanted contaminants. 2. Excising gel bands: Open the gel cassette and cut out individual gel bands, approximately 2 mm above and below the BPB band, with a scalpel. Chop the gel piece into four to eight pieces and put in a 1.5 mL microtube (Fig. 3). Dicing of the gel allows the gel pieces to occupy less volume at the bottom of the microtube, allows trypsin to reach the protein efficiently, and reduces the volume of required chemical reagents to process the protein digestion. 3. Fixing the gel pieces: Agitate the gel pieces in 200 μL of fixation solution for 10 min. Spin down briefly and discard the supernatant. For better fixation repeat this step once more. 4. Washing the gel pieces: Wash the gel pieces with 200 μL of water by agitating for 10 min. Spin briefly and discard the supernatant. Add 400 μL ABC-ACN buffer and agitate for 10 min. Spin briefly and discard the supernatant. Add 200 μL ACN and incubate at room temperature for 5 min, then discard the ACN (see Note 11). Add 100 μL ABC buffer and spin briefly. Incubate at room temperature for 5 min (see Note 12). Add 100 μL ACN and spin down the gel pieces. Incubate at

Cold Acclimation and Plasma Membrane Proteins

181

Fig. 3 A step-by-step guideline for in-gel digestion. Protein loaded into the SDS-PAGE (7–10%) wells by one empty lane interval. BPB indicates the localization of the protein inside the gel. The electrophoresis is stopped when BPB reaches approximately 2 mm from the edge of the well. Gel piece for each of the protein replicates is cut out approximately 2 mm above and below the BPB and diced into four to eight pieces. The diced gels are then transferred to a microtube for further treatment for in-gel tryptic digestion. The dicing reduces the occupied volume at the bottom of the tube

room temperature for 15 min. Spin briefly and discard the ACN (see Note 13). Dry the gel pieces in a vacuum centrifuge for 45 min (see Note 14). 5. Reduction-alkylation: Add 100 μL of reduction buffer and spin briefly. Incubate the tubes at 56  C for 45 min. Spin briefly and discard the supernatant. Add 100 μL of alkylation buffer and spin briefly. Wrap the tubes in aluminum foil and incubate in darkness for 30 min at room temperature. Spin the tubes and discard the supernatant. Repeat the washing steps from the previous step to wash the gel pieces. 6. Digestion: Cool the tubes to room temperature and put on ice. Add 25–50 μL of trypsin (at a ratio of 1:25 (v/v) of trypsin: protein) solution to each tube. Spin the tubes briefly and incubate on ice for 45 min to rehydrate the gels. Discard the supernatant and spin briefly. Add 100 μL of ABC and spin briefly. Incubate the tubes overnight (18 h) at 37  C.

182

Md Mostafa Kamal et al.

7. Stopping digestion: Add TFA to each sample to a final concentration of 0.5% to stop the enzymatic reaction and agitate the tubes for 15 min. 8. Extract digested peptides: Add 100 μL of ACN and agitate for 15 min. Spin briefly to collect the supernatant into another tube. Add 50 μL of TFA-ACN buffer and agitate for 15 min. Spin the tubes briefly and collect the supernatant into the tube containing the ACN supernatant. Dry the collected supernatant to dryness in a vacuum concentrator for 45 min (see Note 14). Add 30 μL of 0.1% TFA and store at 20  C for future use or move to the next step for peptide purification. 3.5 Purification (After Both In-Solution and In-Gel Digestion)

1. Insert the stage-tip into the holed lid of microtube. 2. Preconditioning: Load 30 μL of solution B inside the stage-tip and centrifuge at 1000  g for 1 min. Make sure there is no solution inside the stage-tip. Load 30 μL of solution A inside the stage-tip and centrifuge at 1000  g for 1 min for complete removal of the solution A. 3. Load the digested peptide sample in the stage-tip and centrifuge for 1–2 min at 1000  g. The centrifugation duration should be adjusted based on the sample complexity to make sure complete removal of the peptide solution from the stagetip. 4. Wash the adsorbed peptide in C18 membrane by solution A. Spin at 1000  g for 1 min. Make sure no solution left inside the stage-tip. 5. Transfer the stage-tip to another clean holed-microtube. Load 30 μL of solution B and centrifuge at 1000  g for 1 min to elute the peptide. Repeat this step once more and collect the eluted peptide in the same microtube. 6. Dry the eluted peptide in a vacuum concentrator for 15 min. Add 15 μL of 0.1% (v/v) TFA and transfer to a suitable liquid chromatography vial. Store the desalted peptide at 80  C for future use (see Note 15).

3.6 Nano-LC-MS/MS Analysis 3.6.1 Settings of Nano-LC-MS/MS

An example of a nano-LC-MS/MS run and the database search settings for Arabidopsis PM proteins are described below. 1. Mobile phase for peptide elution from trap column: 0.1% (v/v) formic acid in acetonitrile. 2. Mobile phase for peptide separation: linear gradient of acetonitrile from 5% (v/v) to 45% (v/v). 3. Flow rate and analysis time: 500 nL/min for 120 min. 4. Spray voltage for peptide ionization: 1.6 kV.

Cold Acclimation and Plasma Membrane Proteins

183

5. Mass spectrometer control settings: scan range, 400–1800 m/z; resolution, 60,000; Collision induced dissociation, ten most intense ions with a threshold above 500. 3.6.2 Settings for Data Conversion, Protein Identification and Quantification

1. Parameters for conversion from raw files to mascot generic format (mgf) files (Proteome Discoverer software, ver. 2.1): precursor mass range, m/z 350–5000; highest and lowest charge state, 0; lower and upper RT limit, 0; the minimum total intensity of a spectrum, 0; and the minimum number of peaks in a spectrum, 1. 2. Parameters for identification of proteins (Mascot search engine, ver. 2.6.2): database, Arabidopsis Araport11 protein database; allowed number of missed cleavage, 2; fixed modification, carbamidomethylation (C); variable modification, oxidation (M); peptide mass tolerance, 5 ppm; MS/MS tolerance, 0.6 Da; peptide charges, +2, +3, +4, include decoy database.

3.6.3 Protein Localization and Gene Ontology Study

4

Localization, biological process, and molecular function of identified proteins play vital roles for proper validation of the genes encoding identified proteins. There are several web-based resources available to analyze the molecular characteristics of the proteins (Table 1).

Notes 1. Mops-KOH (pH 7.6), EGTA (pH 8.0), and EDTA (pH 8.0) should be prepared as 0.5 M stock solutions and stored at 4  C. EGTA and EDTA can be dissolved by adding KOH but the pH of the solutions should be adjusted at 8.0. When BSA is dissolved, BSA powder should be preequilibrated at room temperature. PMSF and SHAM should be separately prepared as 1 M and 1.6 M stock solutions, respectively, in DMSO and stored at 4  C. DTT should be stored at 20  C as a 1 M stock solution. PMSF, SHAM, and DTT should be diluted only as needed just before use. 2. KH2PO4/K2HPO4 (K-P) buffer (pH 7.8) should be prepared as a 0.5 M stock solution that can then be used to make the MS-suspension medium. First, 200 mL of 0.5 M K2HPO4 and 30 mL of 0.5 M KH2PO4 are prepared. The pH of the 0.5 M K2HPO4 is adjusted to 7.8 by adding 0.5 M KH2PO4. 3. It is recommended to aliquot DTT and iodoacetamide (IAA) in small volumes and store at 20  C for long term use. 4. TFA evaporates quickly. Thus, solutions containing TFA should be freshly prepared immediately before use. 5. Ammonium bicarbonate solution should be stored 4  C and used within 1 month.

184

Md Mostafa Kamal et al.

Table 1 Web-based resources available to analyze the molecular characteristics of identified proteins Database

Description

Website

References

ARAMEMNON

Plant membrane protein database. Prediction of transmembrane domains, subcellular localization

http://aramemnon.botanik. uni-koeln.de/

[30]

SUBA4

Subcellular localization, GFP localization, protein–protein interaction

http://suba.live/

[31]

AgriGO

Gene ontology analysis toolkit

http://systemsbiology.cau. edu.cn/agriGOv2/

[32]

WoLF PSORT

Subcellular localization sites of protein https://wolfpsort.hgc.jp/

[33]

Plant-mPLoc

Prediction of subcellular localization of http://www.csbio.sjtu.edu. proteins with multiple sites cn/bioinf/plant-multi/

[34]

PPDB

Protein function, protein properties, and subcellular localization

MIND

Membrane protein–protein interaction https://associomics.dpb. [36] network carnegiescience.edu/ Associomics/MIND.html

KEGG

Pathway mapping

https://www.genome.jp/ kegg/

[37]

DAVID

Gene ontology

https://david.ncifcrf.gov/ home.jsp

[38, 39]

http://ppdb.tc.cornell.edu/ [35]

6. Nonacclimated plants should be harvested before bolting. It may be necessary to adjust how long plants are kept before harvesting. 7. In this step, homogenization should not be too long or too vigorous because harsh homogenization can severely disrupt membrane integrity. 8. Two-phase partitioning is the most important step for preparing highly purified PM. When the upper phase of the two-phase partition medium is removed, the Pasteur pipette should be moved from left to right near the boundary of the two phases to prevent taking lower phase. 9. The concentration of UTU in the predigested sample should be diluted to 2 M to make sure the trypsin works properly. 10. At this stage green precipitation might occur. If the precipitation occurs the digest should be centrifuged at 5000  g for 2–3 min and supernatant should be collected in a fresh microtube.

Cold Acclimation and Plasma Membrane Proteins

185

11. At this stage, dehydrated, compressed and completely bleached gels should be observed. If the gels do not change, repeat this step twice. 12. At this stage rehydrated, swollen gel pieces should be observed. If gels do not change repeat this step twice. 13. At this stage partly bleached opaque gels might appear, which is acceptable. 14. Time might change based on the heating system/properties of the vacuum concentrator, and do not over dry the gel pieces which can cause loss of pieces by electrostatic force. 15. It is recommended to analyze the digested and purified peptides by nano-LC-MS/MS within 1 week.

Acknowledgments This work was supported in part by a Research Fellowship for Young Scientists (#247373 to D.T.) and Grants-in-Aid for Scientific Research (#22120003, #24370018, #17H03961 to M.U. and Y.K.) from JSPS, Japan. References 1. Levitt J (1980) Response of plants to environmental stresses. J Range Manag 5:188–228 2. Guy CL (1990) Cold acclimation and freezing stress tolerance: role of protein metabolism. Annu Rev Plant Physiol Plant Mol Biol 41:187–223 3. Thomashow MF (1999) Plant cold acclimation: freezing tolerance genes and regulatory mechanisms. Annu Rev Plant Physiol Plant Mol Biol 50:571–599 4. Steponkus PL (1984) Role of plasma membrane in cold acclimation and freezing injury in plants. Annu Rev Plant Physiol 35:543–584 5. Webb MS, Uemura M, Steponkus PL (1994) A comparison of freezing injury in oat and rye: two cereals at the extremes of freezing tolerance. Plant Physiol 104:467–478 6. Uemura M, Tominaga Y, Nakagawara C et al (2006) Responses of the plasma membrane to low temperatures. Physiol Plant 126:81–89 7. Yoshida S, Uemura M (1984) Protein and lipid compositions of isolated plasma membranes from orchard grass (DactyIis glomerata L.) and changes during cold acclimation. Plant Physiol 75:31–37 8. Uemura M, Yoshida S (1984) Involvement of plasma membrane alterations in cold acclimation of winter rye seedlings (Secale cereale L. cv Puma). Plant Physiol 75:818–826

9. Lynch DV, Steponkus PL (1987) Plasma membrane lipid alterations associated with cold acclimation of winter rye seedlings (Secale cereale L. cv Puma). Plant Physiol 83:761–767 10. Uemura M, Joseph RA, Steponkus PL (1995) Cold acclimation of Arabidopsis thaliana: effect on plasma membrane lipid composition and freeze-induced lesions. Plant Physiol 109:15–30 11. Kawamura Y, Uemura M (2003) Mass spectrometric approach for identifying putative plasma membrane proteins of Arabidopsis leaves associated with cold acclimation. Plant J 36:141–154 12. Minami A, Fujiwara M, Furuto A et al (2009) Alterations in detergent-resistant plasma membrane microdomains in Arabidopsis thaliana during cold acclimation. Plant Cell Physiol 50:341–359 13. Mazars C, Thion L, Thuleau P et al (1997) Organization of cytoskeleton controls the changes in cytosolic calcium of cold-shocked Nicotiana plumbaginifolia protoplasts. Cell Calcium 22:413–420 14. Orvar BL, Sangwan V, Omann F et al (2000) Early steps in cold sensing by plant cells: the role of actin cytoskeleton and membrane fluidity. Plant J 23:785–794

186

Md Mostafa Kamal et al.

15. Sangwan V, Foulds I, Singh J et al (2001) Cold-activation of Brassica napus BN115 promoter is mediated by structural changes in membranes and cytoskeleton, and requires Ca2+ influx. Plant J 27:1–12 16. Welti R, Li W, Li M et al (2002) Profiling membrane lipids in plant stress responses: role of phospholipase Dα in freezing-induced lipid changes in Arabidopsis. J Biol Chem 277:31994–32002 17. Li W, Li M, Zhang W et al (2004) The plasma membrane-bound phospholipase Dδ enhances freezing tolerance in Arabidopsis thaliana. Nat Biotechnol 22:427–433 18. Yamazaki T, Kawamura Y, Minami A et al (2008) Calcium-dependent freezing tolerance in Arabidopsis involves membrane resealing via synaptotagmin SYT1. Plant Cell 20:3389–3404 19. Kondo M, Takahashi D, Minami A et al (2012) Function of Arabidopsis dynamin-related proteins IE during cold acclimation. Cryobiol Cryotechnol 58:105–111. (In Japanese with English summary) 20. Miki Y, Takahashi D, Kawamura Y et al (2018) Temporal proteomics of Arabidopsis plasma membrane during cold- and de-acclimation. J Proteome 197:71–81 21. Takahashi D, Kawamura Y, Yamashita T et al (2012) Detergent-resistant plasma membrane proteome in oat and rye: similarities and dissimilarities between two monocotyledonous plants. J Proteome Res 11:1654–1665 22. Takahashi D, Li B, Nakayama T et al (2013) Plant plasma membrane proteomics for improving cold tolerance. Front Plant Sci 4:90 23. Nakayama T, Takahashi D, Kawamura Y et al (2013) Compositional changes in plasma membrane proteins in Brachypodium distachyon during cold acclimation. Cryobiol Cryotechnol 59:61–65. (In Japanese with English summary) 24. Kasuga J, Takahashi D, Kawamura Y et al. (2012) Proteomic analysis of seasonal colddeacclimation process in poplar phloem and xylem tissues. Abstract of Plant and Microbe Adaptation to Cold 2012 (O-18) 25. Li B, Takahashi D, Kawamura Y et al (2012) Comparison of plasma membrane proteomic changes of Arabidopsis suspension cells (T87 line) after cold and abscisic acid treatment in association with freezing tolerance development. Plant Cell Physiol 53:542–554 26. Ceballos-Laita L, Gutierrez-Carbonell E, Takahashi D et al (2018) Effects of Fe and Mn deficiencies on the protein profiles of tomato (Solanum lycopersicum) xylem sap as revealed by shotgun analyses. J Proteome 170:117–129

27. Gutierrez-Carbonell E, Takahashi D, Lu¨thje S et al (2016) A shotgun proteomic approach reveals that Fe deficiency causes marked changes in the protein profiles of plasma membrane and detergent-resistant microdomain preparations from Beta vulgaris roots. J Proteome Res 15:2510–2524 28. Gutierrez-Carbonell E, Takahashi D, Lattanzio G et al (2014) The distinct functional roles of the inner and outer chloroplast envelope of Pea (Pisum sativum) as revealed by proteomic approaches. J Proteome Res 13:2941–2953 29. Takahashi D, Li B, Nakayama T, Kawamura Y et al (2014) Shotgun proteomics of plant plasma membrane and microdomain proteins using nano-LC-MS/MS. In: Novo JVJ, Komatsu S, Weckwerth W, Wjienkoopeds S (eds) Methods in molecular biology (plant proteomics: methods and protocols), 2nd edn. Springer Science + Business Media, LLC, New York, pp 481–498 30. Schwacke R, Schneider A, van der Graaff E, Fischer K et al (2003) ARAMEMNON, a novel database for Arabidopsis integral membrane proteins. Plant Physiol 131:16–26 31. Hooper CM, Castleden IR, Tanz SK et al (2016) SUBA4: the interactive data analysis centre for Arabidopsis subcellular protein locations. Nucleic Acids Res 45:D1064–D1074 32. Tian T, Liu Y, Yan H et al (2017) agriGO v2.0: a GO analysis toolkit for the agricultural community. Nucleic Acids Res 45:W122–W129 33. Horton P, Park KJ, Obayashi T et al (2007) WoLF PSORT: protein localization predictor. Nucleic Acids Res 35:W585–W587 34. Chou KC, Shen HB (2010) Plant-mPLoc: a top-down strategy to augment the power for predicting plant protein subcellular localization. PLoS One 5:e11335 35. Sun Q, Zybailov B, Majeran W et al (2008) PPDB, the Plant Proteomics Database at Cornell. Nucleic Acids Res 37:D969–D974 36. Lalonde S, Sero A, Pratelli R et al (2010) A membrane protein/signaling protein interaction network for Arabidopsis version AMPv2. Front Physiol 1:24 37. Ogata H, Goto S, Sato K et al (1999) KEGG: Kyoto encyclopedia of genes and genomes. Nucleic Acids Res 27:29–34 38. Huang DW, Sherman BT, Lempicki RA (2009) Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 4:44–57 39. Huang DW, Sherman BT, Lempicki RA (2009) Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res 37:1–13

Chapter 14 A Lipidomic Approach to Identify Cold-Induced Changes in Arabidopsis Membrane Lipid Composition Yu Song, Hieu Sy Vu, Sunitha Shiva, Carl Fruehan, Mary R. Roth, Pamela Tamura, and Ruth Welti Abstract Lipid changes that occur in leaves of plants (e.g., Arabidopsis thaliana), during cold and freezing stress can be analyzed with electrospray ionization triple quadrupole mass spectrometry, using high-throughput multiple reaction monitoring (MRM). An online tool, LipidomeDB Data Calculation Environment, is employed for mass spectral data processing. Key words Cold, Freezing, Postfreezing recovery, Lipidomics, Mass spectrometry, Multiple reaction monitoring (MRM), Lipidomics data processing

1

Introduction Cold and freezing stresses limit crop yield and the arability of land. Thus, the development of more cold- and freezing-resistant crop species can increase crop production. The first step in development of cold- and freezing-resistant plants is understanding the biochemistry and genetics behind plant response to cold and freezing. Arabidopsis thaliana is a moderately cold-tolerant plant that can be readily grown and manipulated in the lab, making it possible to identify biochemical changes that occur during cold acclimation, freezing, and the postfreezing recovery period under controlled conditions. At the cellular level, freezing temperature causes extracellular ice formation, followed by severe cellular dehydration, finally leading to loss of membrane functionality and cell death [1–3]. Cold acclimation contributes to the development of plant freezing tolerance. Exposure to cold, but nonfreezing, temperature for 1 day or

Electronic supplementary material: The online version of this chapter (https://doi.org/10.1007/978-1-07160660-5_14) contains supplementary material, which is available to authorized users. Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_14, © Springer Science+Business Media, LLC, part of Springer Nature 2020

187

188

Yu Song et al.

more drops the lethal temperature for Arabidopsis (Columbia accession) from
2 to
8  C [3, 4]. Lipid changes are part of plant response to cold and freezing stress and may play roles in maintaining membrane structure, in signaling, or both. Lipid changes include modifications of membrane and neutral lipids, such as fatty acid desaturation, fatty acid oxidation, hydrolysis of head group components and acyl chains, head group acylation, and head group glycosylation. Some lipid alterations during cold or freezing stress can modulate plant damage [5–12]. For example, fatty acid desaturases, including FAD2, FAD5, and FAD6, act to increase fatty acid unsaturation in membrane lipids and are critical for normal growth at low temperatures [5, 6]. Two phospholipase Ds, which hydrolyze phospholipids to form phosphatidic acid, act during freezing and postfreezing recovery but play opposite (i.e., positive and negative) roles to each other during these processes [7–9]. Phospholipase activity may be modulated by lipid binding to acyl-CoA-binding protein 4, which may mediate the expression of phospholipase Dδ, the phospholipase with a positive effect on freezing tolerance [13]. Diacylglycerol acyltransferase1 (DGAT1) is also proposed to contribute to freezing tolerance in a close relative of Arabidopsis, Boechera stricta. In this plant, DGAT1 expression is highly induced during cold acclimation, resulting in accumulation of triacylglycerols in response to freezing stress [14]. Additionally, processive glycosylation of monogalactosyldiacylglycerol to form oligogalactosyldiacylglycerols by the freezing-activated galactolipid:galactolipid galactosyltransferase, encoded by SENSITIVE TO FREEZING2, increases freezing tolerance by stabilizing the chloroplast membrane [15]. Oxophytodienoic (OPDA) and jasmonic (JA) acids are synthesized from fatty acids originating in the membrane. These signaling compounds have been demonstrated to be capable of signaling and modulating gene expression [16], and JA biosynthesis has been shown to have a positive effect on Arabidopsis freezing tolerance [17, 18]. The roles of some other lipid changes are less clear, and these changes, as well as those previously mentioned, deserve additional investigation. For example, lipidomic and transcriptomic approaches have shown that, under low temperature stress, allocation of fatty acids and diacylglycerol (DAG) moieties between the endoplasmic reticulum and chloroplast is altered [10]. Freezing and postfreezing stress also induce the synthesis of galactoseacylated monogalactosyldiacylglycerols (acMGDGs), with primarily unoxidized fatty acyl chains attached to the galactose moiety [12]. The function of these galactolipid modifications needs further clarification. There is still much to be discovered about the roles of lipids, lipid-related enzymes, and regulated genes under cold and freezing stress. In the past, analysis of lipids was hampered by the poor

Arabidopsis Leaf Lipidomics

189

sensitivity and resolution of “traditional” analytical technology. Mass spectrometry-based lipidomics offers many advantages in monitoring cold- and freezing-induced lipid changes. Large numbers of lipid molecular species can be analyzed in a relatively short time, with higher sensitivity and resolution than traditional methods. Lipid extracts from cold- or freezing-treated plants can be introduced to a mass spectrometer by direct infusion or by liquid chromatography, both of which have been utilized in Arabidopsis cold stress studies [19, 20]. Vu, Shiva, and coworkers [21, 22] established a comprehensive lipidomic approach including lipid extraction, mass spectral analysis, and data processing. They developed a direct infusion method using electrospray ionization (ESI) triple quadrupole mass spectrometry to analyze plant phospholipids and glycolipids by a series of precursor and neutral loss scans. The method was optimized and applied to detect changes in levels of phospholipids, galactolipids, and others, including oxidized and head group-acylated monogalactosyldiacylglycerols, under freezing and postfreezing stress [21, 22]. Here we update and extend the analysis by improving leaf lipid extraction efficiency by using a single-extraction protocol and employing multiple reaction monitoring (MRM) to measure a greater number of analytes in a shorter time. Data processing has been improved. The updated lipid extraction method is much less labor-intensive than most comparable methods and provides good lipid recovery [23]. The direct infusion MRM method measures selected phospholipids, galactolipids, other glycerolipids, sphingolipids, and sterol derivatives, with most lipids designated by lipid class, total acyl carbons, and total double bonds (i.e., total acyl carbons: total acyl carbon–carbon double bonds). Each MRM transition is based on the mass–charge ratio (m/z) of the intact ionized lipid and the m/z of one fragment formed in the mass spectrometer. Acquisition parameters, as well as MRM acquisition times, are summarized, based on Vu et al. [24]. Lipid amounts are determined as normalized mass spectral signal/plant dry mass. The intensities of peaks in each sample are compared to those of added internal standards. A value of 1 represents the same intensity as 1 nmol of the relevant internal standard (or 1 nmol of an average intensity of two relevant internal standards). For the most common membrane lipid classes, internal standards are non-naturally occurring compounds in the same class. However, for some less studied analytes, such as oxidized and head group-acylated galactolipids, there are no readily available internal standards that are good structural matches. Thus, such compounds may vary in mass spectral intensity per mole, compared to their (poorly matched) internal standards; in most cases, response factors have not been determined. Consequently, calculated analyte levels may not reflect the absolute content of each lipid analyte, but sample-to-sample, relative-amount comparisons are valid. A recent update of

190

Yu Song et al.

LipidomeDB Data Calculation Environment (DCE) extends its functionality to process data acquired in direct-infusion MRM mode [25]. The newly added algorithm for MRM data includes isotopic deconvolution, based on the available data. Additionally, in the method described herein, to assure consistency of data for each analyte throughout long periods of mass spectral data acquisition, a data correction strategy utilizing quality control samples (QC), based on work by Dunn et al. [26], is employed. Normalization to QC values can increase the analytical precision of lipid quantification. The protocol for cold and freezing treatment described herein contains descriptions of both nonlethal freezing (mild,
2  C for 16 h) and more severe freezing (
8  C for 2 h) treatments. Although Arabidopsis thaliana (Columbia-0) was used, other natural accessions or mutant lines can also be analyzed via this protocol. More details on cold acclimation, freezing, and postfreezing treatments can be found elsewhere in this volume.

2

Materials (See Note 1)

2.1 Cold Acclimation and Freezing Treatment

1. Arabidopsis thaliana seeds. 2. Soil, such as Pro-Mix “PGX” (Hummert International). 3. 3½00 Kord square pots (Hummert International), or 72-cell plug trays (International Greenhouse Company) (see Note 2). 4. Fertilizer (e.g., Peters 20:20:20, Hummert International, or Miracle-Gro 20:20:20, Scotts). 5. Refrigerator at 4  C. 6. Light meter. 7. Waxed paper. 8. Scissors. 9. Ice chips. 10. Growth chamber, such as a Conviron ATC26. 11. Walk-in cold room at 4  C. 12. Portable light cart, such as Hummert International. 13. Programmable freezing chamber, such as ESU-3CA cold temperature chamber (Espec Corporation, Hudsonville, MI).

2.2 For Sampling, Lipid Extraction, and Dry Mass Measurement

1. Scissors. 2. Optima-grade isopropanol with 0.01% butylated hydroxytoluene (BHT) (w/v). 3. Optima-grade chloroform, methanol, and water, combined to make extraction solvent, chloroform–methanol–water (40:55:5, v/v/v).

Arabidopsis Leaf Lipidomics

191

4. Glass vials, 8 mL with Teflon-lined screw caps. 5. Dry block heater that accepts 8 mL vials (and other sizes). 6. Orbital shaker. 7. Oven, vented to hood. 8. Balance that determines mass, preferably to micrograms (e.g., AX26 DeltaRange microbalance, Mettler-Toledo, Greifensee, Switzerland). 9. Ionizer antistatic system (see Note 3). 2.3 Mass Spectrometry

1. Optima-grade chloroform. 2. Internal standard mix, containing LPC(13:0), LPC(19:0), LPE(14:0), LPE(18:0), LPA(14:0), LPA(18:0), LPG(14:0), LPG(18:0), PA(28:0) [di14:0], PA(40:0) [diphytanoyl], PA (40:0) [diphytanoyl], PC(24:0) [di12:0], PC(48:2) [di 24:1], PE(24:0) [di12:0], PE(40:0) [diphytanoyl], PG(28:0) [di14:0], PG(40:0) [diphytanoyl], PI(34:0) [16:0/18:0], PI (36:0) [di18:0], PS(28:0) [di14:0], PS(40:0) [diphytanoyl], DGDG(34:0) [18:0/16:0], DGDG(36:0) [di18:0], MGDG (34:0) [18:0/16:0], MGDG(36:0) [di18:0], and TAG(51:3) [17:1/17:1/17:1] (see Notes 4 and 5). 3. Vacuum concentrator, such as CentriVap (Labconco Corp., Kansas City, MO), vented to a fume hood, or nitrogen gas stream evaporator, in a fume hood. 4. Rotary evaporator with water bath, such as Bu¨chi Re121 with Model 461 water bath (Bu¨chi Labortechnik AG, Switzerland). 5. Preslit, target Snap-it 11 mm Snap Caps. 6. Amber vials, 12  32 mm. 7. Autosampler, such as CTC PAL HTC-xt (LEAP). 8. Sample trays to hold vials, such as VT54 (LEAP). 9. Chloroform–methanol–water (30:66.5:3.5, v/v/v) to fill the wash reservoirs on the autosampler for washing the syringe and sample loop. 10. Methanol–acetic acid (9:1, v/v) for washing PEEKsil tubing. 11. “MS solvent” mixture: chloroform–methanol–300 mM ammonium acetate in water (30:66.5:3.5, v/v/v) to dissolve the internal standards/lipid extract. 12. PEEKsi tubing 1/3200 OD  50 μm ID  5 cm, 1/3200 OD  50 μm ID  15 cm, 1/3200 OD  50 μm ID  50 cm (IDEX Health & Science, USA). 13. LC pump, such as LC-30AD (Shimadzu, Japan). 14. Triple quadrupole mass spectrometer, such as Sciex 6500+ system equipped with an ESI source and Analyst and MultiQuant software programs (Sciex, Concord, Ontario, Canada).

192

3

Yu Song et al.

Methods

3.1 Cold Acclimation and Mild/Severe Freezing Treatment

1. Sow four Arabidopsis thaliana seeds per well in 72-well plug trays filled with loosely packed, Pro-Mix “PGX” soil saturated with 0.01% 20-20-20 fertilizer. Place trays in refrigerator or cold room at 4  C for 2 days for stratification of seeds. Place trays in a growth chamber under a 14/10 h light/dark cycle at 21  C with 60% humidity. Maintain light intensity at 90 μmol/ m2/s with cool white fluorescent lights. Cover trays with propagation domes for the first 9 days to maintain high humidity. Water trays every 4 days. On day 14 after sowing, thin plants to one plant per well. On day 20, fertilize plants with 0.01% 20-20-20 fertilizer. 2. On day 26, transfer soil-grown Arabidopsis plants to the portable light cart. Put the light cart into the cold room with desired temperature (4  C) for cold acclimation. Measure the light intensity with a light meter and adjust it and the day–night cycle to match the ongoing growing conditions. Acclimate plants by placing in the cold room for 3 days. 3. For plants that will undergo severe freezing stress at
8  C (see Note 6 for mild freezing conditions), add pieces of waxed paper that cover half of the soil around each plant. Gently place waxed paper under Arabidopsis rosettes and on top of soil as shown in Fig. 1. Transfer plants to be frozen to the programmable freezing chamber. Program the freezing chamber so that the temperature drops from the cold acclimation

Fig. 1 Arabidopsis plants prepared to undergo freezing at
8  C. Two half circles of waxed paper were placed under each rosette. The purpose of the waxed paper is to eliminate freezing of leaves to the soil, which makes it difficult to obtain clean leaf or rosette samples when the plants are frozen

Arabidopsis Leaf Lipidomics

193

point (4  C) to
2  C at
2  C per hour (see Note 7). Plants may be held at
2  C for 1 h for ice crystal formation before the temperature is dropped directly from
2  C to the final temperature (
8  C). Ice chips may be added on soil (under or around waxed paper) at this step to prevent supercooling (see Note 8). After the freezing treatment (
8  C for 2 h), plants may be thawed at 4  C or other desired temperature (see Note 9). 3.2 Sampling, Lipid Extraction, and Dry Mass Measurement

The extraction method is based on Shiva et al. [23]. This singleextraction protocol is less labor-intensive, reduces reagent volumes, and has comparable extraction efficiency to previously described methods. 1. Sample plants at desired time points before cold acclimation, during acclimation, before and after freezing, and during postfreezing recovery period (see Note 10). Cut leaves, or other desired tissues, directly above a vial containing isopropanol with 0.01% BHT preheated on the heating block to 75  C. Submerge tissue quickly and heat the vial containing the tissue at 75  C for 15 min (see Note 11). For single leaves, use an 8-mL vial containing 1.5 mL of isopropanol with 0.01% BHT. Add 4.5 mL extraction solvent, chloroform–methanol–water (40:55:5, v/v/v). Different sizes of vials and extraction solvent volumes may be used, depending on the size of plant tissues (see Note 12). 2. Shake the vials at 100 rpm on an orbital shaker for 24 h at room temperature. Transfer intact, extracted leaf tissue to a new vial using forceps. Evaporate any remaining solvent and water from the tissue in the new vial, first in the hood at room temperature and then in an oven overnight at 105  C. Weigh the dried leaf tissues using the microgram balance. The original vial contains the extracted lipids. 3. Store the extracted lipids (in their solvent) at
20  C or colder.

3.3 Mass Spectrometry

1. For each analytical sample, add 20 μL of internal standard mix (details are described in Note 5 and composition is summarized in Supplemental Table 1) to a 2-mL amber glass vial. Add a volume of lipid extract equivalent to 0.085 mg dry tissue mass to the vial and place in the vacuum concentrator (CentriVap) to evaporate the solvent. Finally, add 300 μL MS solvent, chloroform–methanol–300 mM ammonium acetate in water (30:66.5:3.5, v/v/v). 2. Make “standards-only” (“IS,” internal standard) samples by adding 20 μL internal standard mix to a 2-mL amber glass vial, removing the solvent using the nitrogen gas stream evaporator, and adding 300 μL MS solvent (see Note 13).

194

Yu Song et al.

3. Determine the number of quality control (QC) samples needed. A good estimate is to prepare a number equal to half the number of analytical samples. Prepare QC samples by pooling 1 mL of lipid extract (of the original 6 mL) from each sample from different treatments to make a QC stock solution. Calculate the concentration of QC stock solution in terms of the amount of dry leaf mass of sample used to make the combined extract per volume (see Note 14). To make a solution for 100 QC samples, add 2 mL internal standard mixture and a volume of the QC pool equivalent to 8.5 mg of dry leaf mass. Evaporate the solvent using a rotary evaporator with water bath set at 40  C, the vacuum concentrator, or nitrogen evaporator. Add 30 mL MS solvent. Aliquot 300 μL QC mixture into 2-mL amber glass vials for working QCs. Each working QC sample contains the same amount of lipid extract (0.085 mg dry leaf mass equivalent), internal standard mix (20 μL), and MS solvent (300 μL) as each analytical sample. Label the QC samples, store at
80  C, and bring to room temperature 1 h before analysis. 4. Program the autosampler and pump to infuse 75 μL sample at a flow rate of 25 μL/min for 3 min and then a flow rate of 70 μL/min for 2 min with a total acquisition time of 4 min. Wash the sample syringe and the injection port of the autosampler with chloroform–methanol–water (30:66.5:3.5, v/v/ v) from the two wash reservoirs after the sample loop is filled. Wash the LC/MS system after each analysis with methanol– acetic acid (9:1, v/v) at a flow rate of 70 μL/min from 5 to 5.6 min after the start of the infusion. PEEKsil tubing is used to infuse the samples into the mass spectrometer (see Note 15). 5. In a VT54 rack, arrange analytical samples, QC samples, and IS vials for mass spectral analysis in this order: QCs 1–6, samples 1–3, IS 1, QC 7, samples 4–6, QC 8, samples 7–10, QC 9, samples 11–14, QC 10, samples 15–18, IS 2, QC 11, and so on (i.e., the first 6 samples should be QCs, then a QC sample should be inserted every 3–4 analytical samples, and IS samples should be inserted every 20 samples). 6. Establish an MRM method on the mass spectrometer using the parameters for each MRM transition, including intact ion m/z (quadrupole 1 or Q1), fragment m/z (second analyzer, that is, quadrupole 3 or Q3), collision energy (CE), and dwell time, as listed in Supplemental Tables 2 (positive mode) and 3 (negative mode) for the plant lipid analytes and internal standards (see Note 16). The data acquisition methods (.dam files for the Sciex 6500+ mass spectrometer) are available for download at https://www.k-state.edu/lipid/analytical_laboratory/analy sis_components/data_acquisition_methods/index.html. Global mass spectrometry parameters are indicated in Note 17.

Arabidopsis Leaf Lipidomics

3.4 Data Processing and Normalization

195

1. Data export can vary depending on the mass spectrometer. For the Sciex 6500+ mass spectrometer, use the data processing software MultiQuant to process and export MRM data (combined and averaged over the infusion) to Excel (see Note 18). 2. Prepare a template for MRM data processing at LipidomeDB Data Calculation Environment (DCE) from that found in Supplemental Table 4 or from the “MRM example data upload file” available at http://lipidome.bcf.ku.edu:8080/ Lipidomics/. For explanations on each row and column of the upload file (see Note 19). 3. Use the updated LipidomeDB DCE at http://lipidome.bcf.ku. edu:8080/Lipidomics/ for identification and quantification of lipids. After logging in, select “Add MRM Experiment,” upload the Excel file, and continue to process. The output data appear directly to the right of the Input intensities in the same units as the internal standards (typically nmol). The output data are isotopically deconvoluted and normalized to the internal standards (see Note 20). 4. Remove the background from each lipid analyte by subtracting the average of the appropriate internal standard samples from that tray. As desired, use an adaptation of the method of Dunn et al. [26] to reduce the variability caused by instability of the instrument and assure the consistency of the data throughout the entire acquisition period (see Notes 21 and 22). 5. Calculate amounts of lipid analyzed (in values equivalent to nmol) by dividing the calculated amounts by 0.085 mg dry tissue mass. Resulting data will be in “normalized intensity per extracted dry mass (nmol),” where a value of 1 equals the same intensity as 1 nmol of internal standard (see Note 23).

4

Notes 1. Many of the materials indicated here are the same as listed in ref. 21. The methods extend those described in ref. 24, adapt them to a different mass spectrometer, and apply them to the analysis of lipids derived from cold and freezing experiments. 2. We typically use 27-day-old Arabidopsis plants, from which we sample leaves or rosettes. However, plants at other developmental stages may be used. The current extraction and analysis protocols are appropriate for any aboveground vegetative tissue, flowers, or siliques. 3. Using an antistatic system with a microgram-accurate balance will increase the stability of mass measurements.

196

Yu Song et al.

4. Lipid class abbreviations are: DGDG, digalactosyldiacylglycerol; LPA, lysophosphatidic acid; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; LPG, lysophosphatidylglycerol; MGDG, monogalactosyldiacylglycerol; PA, phosphatidic acid; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PG, phosphatidylglycerol; PI; phosphatidylinositol; PS, phosphatidylserine; TAG, triacylglycerol. 5. If samples are being analyzed on a Sciex 6500+ mass spectrometer, a volume of standard stock solution (20 μL) is added to the 2-mL amber glass vial. The internal standard amounts in 20 μL stock solution are listed in Supplemental Table 1. It is best to determine the concentration of most phospholipids for the stock solution by phosphate assay [27]. Concentrations of MGDG, DGDG, and PI, which are hydrogenated mixtures of 16-carbon/18-carbon and di18-carbon species, and TAG are best determined by gas chromatography of fatty acid methyl esters derived from these lipids. 6. For plants that will undergo mild freezing stress, no waxed paper is needed. Before moving into the freezing chamber, saturate soil with water and then add ice chips (see Note 8). Program the freezing chamber to drop the temperature from the cold acclimation point (4  C) to
2  C in 1 h. Hold the temperature of the freezing chamber at
2  C overnight (16 h, e.g., from 5:00 pm to 9:00 am of next day). After freezing treatment, move plants into a growth chamber at 21  C for 1 h for postfreezing recovery. As shown in Fig. 2, plants won’t show an obvious change in appearance during cold acclimation, after the mild (
2  C for 16 h) freezing treatment, or after postfreezing recovery from mild freezing treatment. 7. Although transferring directly to the low freezing temperature may not perfectly mimic natural freezing, a freezing regimen without gradual temperature change can be employed. If plants are to be placed directly at the low freezing temperature, ice chips can be added immediately before placing the plants in the freezing chamber. 8. For any freezing regimen, soil should be saturated with water prior to adding ice chips. An alternative approach to placing ice chips on the soil is to partly submerge the 3½00 square pots or the 72-well plug tray in an ice slurry. 9. Plants may be thawed at 4  C or at the growing temperature. Although plants may sustain more damage with recovery at the growing temperature, recovery characteristics of acclimated plants are clearly distinguishable from those of nonacclimated plants (unpublished data) when plants are subjected to freezing at
8  C.

Arabidopsis Leaf Lipidomics

197

Fig. 2 Plant appearance before cold acclimation, after acclimation, after mild-freezing treatment, and after postfreezing recovery. Photographs depict different natural Arabidopsis accessions: untreated (before cold acclimation) grown at 21  C, after 3 day cold acclimation at 4  C, after 16 h mild freezing treatment at
2  C, and after 1 h postfreezing recovery at 21  C

10. Depending on the particular experimental goal, plants can be sampled early or late in cold acclimation (to measure early or late cold-induced molecular changes), immediately after freezing treatment (to measure freezing-induced changes), and/or during the recovery phase (to measure thawing-related changes). During the cold acclimation period, it is best to sample inside the cold room, and the temperature of the heating block may need to be closely monitored to maintain 75  C. To sample right after freezing, it is critical to collect the plant tissues quickly without allowing them to thaw. Especially when handling a large number of plants with a reach-in freezing chamber, avoid letting plants wait outside of the chamber; instead, pull out only the number of plants that can be sampled in less than 30 s. Within 30 s, two workers typically can sample four Arabidopsis rosettes. If using a 72-well plug tray, the tray can be cut (before treatment) in sections of four plants for sampling by two workers. 11. It is critical to drop harvested leaves into isopropanol at 75  C immediately to prevent activation of phospholipase D, a damage-activated enzyme, which will degrade membrane lipids and produce phosphatidic acid.

198

Yu Song et al.

12. Should a different volume of isopropanol be required (to fully submerge plant tissues when harvesting), the volumes of extraction solvent can be varied accordingly. The final solvent composition should be chloroform–isopropanol–methanol– water in the ratio 30:25:41.5:3.5 (v/v/v/v). For rosettes or other large plant tissues, use a 50-mL glass tube (25  150 mm) with a Teflon-lined screw cap containing 4 mL of isopropanol with 0.01% BHT. Add 12 mL extraction solvent, chloroform–methanol–water (40:55:5, v/v/v). 13. “Standards-only” spectra are used to correct instrument background signal and assess sample carryover. Internal standard peaks in “standards-only” spectra will likely have higher intensities than in sample spectra because of low ion suppression. Intensities of plant lipid peaks in “standards-only” spectra should be very low and may be subtracted from the intensities of the same mass spectral transitions in plant lipid spectra to remove background signal. 14. The concentration of the QC pool ¼ (the sum of the masses of all samples)/(total volume of all samples). 15. PEEKsil tubing reduces carryover between infusions compared to other types of tubing. Using PEEKsil, most lipid analytes are washed out within 4 min. 16. Adapting the method described in ref. 24 to the Sciex 6500+ mass spectrometer required the removal of several analytes with intact ion m/z > 1250, so that the instrument could be operated in the “low mass” range. 17. Acquire the mass spectral data for the samples on a triple quadrupole mass spectrometer (Sciex 6500+ mass spectrometer) equipped with an ESI probe, using MRM loops containing transitions for analytes measured in positive and/or negatives modes. Analytes should be grouped together based on the analysis mode (positive or negative) to avoid excessive mode switching. Acquisition time is 4 min, with a single injection for each sample. Sample acquisition begins after a 15-s delay. In positive mode, the ion spray voltage is 5500 V, the curtain gas, 35 psi; the source temperature, 100  C; the ion source gas (GAS1), 45 psi; the ion source gas (GAS2), 45 psi; the declustering potential, 100 V; and the entrance potential, 10 V. In negative mode, the ion spray voltage is
4500 V, the curtain gas, 35 psi; the source temperature, 100  C; GAS1, 45 psi; GAS2, 45 psi; the declustering potential,
100 V; and the entrance potential,
10 V. Acquire 27 cycles of the MRM list (from Supplemental Tables 2 and 3) in 4 min. The collision gas is nitrogen. 18. Establish a quantitation method. Build the new quantitation method based on a QC sample which is likely to contain all the analytes of interest. Review component information and set

Arabidopsis Leaf Lipidomics

199

Integration & Regression parameters (adjust for each compound or set default values for all compounds) to make sure the method has suitable integration for each targeted analyte. For direct infusion data on the Sciex 6500+ mass spectrometer, choose the “Summation” integration algorithm, and set the retention time(s) and summation window(s) so that the entire time (4 min) that you would like to sum is covered in all samples. Save the quantitation method as a “name.qmethod” file. For data processing, build a new result table and select all the samples you want to process. Choose the quantitation you established before and finish the processing. To export the results, select the “Result Table-Metric,” “Area,” and “Transpose” formats for the exported results. Finally, a text file that can be opened in Excel is generated. Verify that compounds are listed in the order that you have in your MRM data upload file. Remove rows 2 and 3 from the output and paste the intensity data from column B into cell AA2 and to its right and down in the MRM data upload file. For illustrations of these steps, visit http://lipidome.bcf.ku.edu:8080/Lipidomics/ ExampleFiles/Directions%20for%20Sciex%206500+%20direct %20infusion%20multi-sample%20processing%20and% 20export%20in%20MultiQuant.pdf. 19. For column A, indicate an arbitrary number unique to each lipid analyzed. Lipid formulas go in column B and lipid names in column C. Formulas should be for the uncharged (M) version of the lipid. Column D holds information on the adduct used in the mass spectrometry experiment from this list: [M + H]+, [M + NH4]+, [M + Na]+, [M
H]
, [M
CH3]
, [M + OAc]
, or [M + C2H3O2]
. The M mass, plus or minus the indicated ion, should correspond to the m/z used for intact ion data acquisition (column N). Column E holds the formula of the charged fragment. This needs to correspond to the m/z used for fragment ion data acquisition (column O). Column F indicates which data will be considered as a group for isotopic deconvolution. If this column is empty or if all the entries are identical, all intensity data will be considered in the isotopic deconvolution algorithm. If both positive and negative modes are used, MRM pairs in the two modes should also have different designations. In columns G, I, and K, internal standards to be used for each analyte are designated based on the “arbitrary number” in column A. Columns H, J, and L should indicate the amounts of the standards in columns G, I, and K, respectively. The molar units used here (typically nmol) will be the units in the output file. For column M, the possible entries are “Line” or “Average.” “Average” is most commonly used with MRM data and is recommended. “Average” will average the intensity/nmol of the internal standards and use this value

200

Yu Song et al.

to calculate the amount of analyte (in nmols) from the observed intensity. Entries for the experimental intact ion m/ z and charged fragment m/z used in the acquisition go in columns N and O. Leave columns P and Q blank; after processing these columns will contain the values for intact ion m/z and charged fragment m/z calculated from the compound formulas, adduct information, and fragment formulas, which can be compared with the experimental values of these parameters. Leave columns R to Z blank or enter other compoundspecific information that you would like to have associated with the data. Place input data (from MultiQuant) to the right and downward starting in cell AA2, overwriting the “test sample” data in Supplemental Table 4. Sample labels should be in row 2, starting at AA2 going to the right, overwriting “test sample” names, and analyte intensities should be in row 3, starting at AA3 and going to the right and downward. Over each column of sample data, the word “Input” should appear in row 1. 20. More details about the function of LipidomeDB DCE are available via a “tutorial” linked to its home page, http:// lipidome.bcf.ku.edu:8080/Lipidomics/. 21. Remove the data for the first five QC samples in each set, due to potential instrument instability at the beginning. To correct for any variability across different sample sets (days), multiply the value of each lipid in each sample by the average of its QC values from the entire acquisition process divided by the average in its QC values in the sample’s own set. 22. If the QC values (other than the first five in each set) are normalized along with the other data on the same tray as described in Note 21, the coefficient of variation for each analyte can be calculated as the standard deviation of the remaining normalized QC samples/the average of those samples. Coefficient of variation values of less than ~20–30% represent analytes with reasonable analytical precision [24, 26]. 23. Data may also be calculated in “percentage of total normalized signal” by multiplying each value times 100 and dividing by the sum of the normalized analyte values for that sample.

Acknowledgments The authors would like to thank lab member Libin Yao for her contributions to plant stress experiments in our laboratory, Mark Ungerer for use of his lab’s freezing chamber, and Ari Jumpponen for use of his lab’s light cart. This work was supported by the USDA National Institute of Food and Agriculture, Hatch/Multi-State project 1013013, and National Science Foundation MCB

Arabidopsis Leaf Lipidomics

201

1413036. Instrument acquisition at KLRC was supported by National Science Foundation (EPS 0236913, DBI 0521587, DBI 1228622, DBI 1726527), K-IDeA Networks of Biomedical Research Excellence (INBRE) of National Institute of Health (P20GM103418), and Kansas State University. Contribution no. 20-008-B from the Kansas Agricultural Experiment Station. References 1. Thomashow MF (1999) Plant cold acclimation: freezing tolerance genes and regulatory mechanisms. Annu Rev Plant Biol 50:571–599 2. Xin Z, Browse J (2000) Cold comfort farm: the acclimation of plants to freezing temperatures. Plant Cell Environ 23:893–902 3. Uemura M, Joseph RA, Steponkus PL (1995) Cold acclimation of Arabidopsis thaliana (effect on plasma membrane lipid composition and freeze-induced lesions). Plant Physiol 109:15–30 4. Gilmour SJ, Hajela RK, Thomashow MF (1988) Cold acclimation in Arabidopsis thaliana. Plant Physiol 87:745–750 5. Miquel M, James D, Dooner H et al (1993) Arabidopsis requires polyunsaturated lipids for low-temperature survival. Proc Natl Acad Sci U S A 90:6208–6212 6. Hugly S, Somerville C (1992) A role for membrane lipid polyunsaturation in chloroplast biogenesis at low temperature. Plant Physiol 99:197–202 7. Welti R, Li W, Li M et al (2002) Profiling membrane lipids in plant stress responses. Role of phospholipase Dα in freezing-induced lipid changes in Arabidopsis. J Biol Chem 277:31994–32002 8. Li W, Wang R, Li M et al (2008) Differential degradation of extraplastidic and plastidic lipids during freezing and post-freezing recovery in Arabidopsis thaliana. J Biol Chem 283:461–468 9. Degenkolbe T, Giavalisco P, Zuther E et al (2012) Differential remodeling of the lipidome during cold acclimation in natural accessions of Arabidopsis thaliana. Plant J 72:972–982 10. Li Q, Zheng Q, Shen W et al (2015) Understanding the biochemical basis of temperatureinduced lipid pathway adjustments in plants. Plant Cell 27:86–103 11. Barnes AC, Benning C, Roston RL (2016) Chloroplast membrane remodeling during freezing stress is accompanied by cytoplasmic acidification activating SENSITIVE TO FREEZING2. Plant Physiol 171:2140–2149

12. Vu HS, Roth MR, Tamura P et al (2014) Head-group acylation of monogalactosyldiacylglycerol is a common stress response, and the acyl-galactose acyl composition varies with the plant species and applied stress. Physiol Plant 150:517–528 13. Chen QF, Xiao S, Chye ML (2008) Overexpression of the Arabidopsis 10-kilodalton acylcoenzyme A-binding protein ACBP6 enhances freezing tolerance. Plant Physiol 148:304–315 14. Arisz SA, Heo JY, Koevoets IT et al (2018) DIACYLGLYCEROL ACYLTRANSFERASE1 contributes to freezing tolerance. Plant Physiol 177:1410–1424 15. Moellering ER, Muthan B, Benning C (2010) Freezing tolerance in plants requires lipid remodeling at the outer chloroplast membrane. Science 330:226–228 16. Taki N, Sasaki-Sekimoto Y, Obayashiet T et al (2005) 12-Oxo-phytodienoic acid triggers expression of a distinct set of genes and plays a role in wound-induced gene expression in Arabidopsis. Plant Physiol 139:1268–1283 17. Sharma M, Laxmi A (2016) Jasmonates: emerging players in controlling temperature stress tolerance. Front Plant Sci 6:1129 18. Hu Y, Jiang L, Wang F et al (2013) Jasmonate regulates the inducer of CBF expression–crepeat binding factor/DRE binding factor1 cascade and freezing tolerance in Arabidopsis. Plant Cell 25:2907–2924 19. Burgos A, Szymanski J, Seiwert B et al (2011) Analysis of short-term changes in the Arabidopsis thaliana glycerolipidome in response to temperature and light. Plant J 66:656–668 20. Vu HS, Tamura P, Galeva NA et al (2012) Direct infusion mass spectrometry of oxylipincontaining Arabidopsis membrane lipids reveals varied patterns in different stress responses. Plant Physiol 158:324–339 21. Vu HS, Shiva S, Hall AS et al (2014) A lipidomic approach to identify cold-induced changes in Arabidopsis membrane lipid composition. In: Hincha DK, Zuther EZ (eds) Plant cold

202

Yu Song et al.

acclimation, Methods in molecular biology, vol 1166. Humana, New York, pp 199–215 22. Shiva S, Vu HS, Roth MR et al (2013) Lipidomic analysis of plant membrane lipids by direct infusion tandem mass spectrometry. In: Munnik T, Heilmann I (eds) Plant lipid signaling protocols, Methods in molecular biology, vol 1009. Humana, Totowa, NJ, pp 79–91 23. Shiva S, Enninful R, Roth MR et al (2018) An efficient modified method for plant leaf lipid extraction results in improved recovery of phosphatidic acid. Plant Methods 14:14 24. Vu HS, Shiva S, Roth MR et al (2014) Lipid changes after leaf wounding in Arabidopsis thaliana: expanded lipidomic data form the basis for lipid co-occurrence analysis. Plant J 80:728–743

25. Fruehan C, Johnson D, Welti R (2018) LipidomeDB Data Calculation Environment has been updated to process direct-infusion multiple reaction monitoring data. Lipids 53:1019–1020 26. Dunn WB, Broadhurst D, Begley P et al (2011) Procedures for large-scale metabolic profiling of serum and plasma using gas chromatography and liquid chromatography coupled to mass spectrometry. Nat Protoc 6:1060–1083 27. Ames BN (1966) Assay of inorganic phosphate, total phosphate and phosphatases. In: Neufeld E, Ginsburg V (eds) Complex carbohydrates, Methods in enzymology, vol 8. Academic, New York, pp 115–118

Chapter 15 Multiplexed Profiling and Data Processing Methods to Identify Temperature-Regulated Primary Metabolites Using Gas Chromatography Coupled to Mass Spectrometry Alexander Erban, Federico Martinez-Seidel, Yogeswari Rajarathinam, Frederik Dethloff, Isabel Orf, Ines Fehrle, Jessica Alpers, Olga Beine-Golovchuk, and Joachim Kopka Abstract This book chapter describes the analytical procedures required for the profiling of a metabolite fraction enriched for primary metabolites. The profiling is based on routine gas chromatography coupled to mass spectrometry (GC-MS). The generic profiling method is adapted to plant material, specifically to the analysis of plant material that was exposed to temperature stress. The method can be combined with stable isotope labeling and tracing experiments and is equally applicable to preparations of plant material and microbial photosynthetic organisms. The described methods are modular and can be multiplexed, that is, the same sample or a paired identical backup sample can be analyzed sequentially by more than one of the described procedures. The modules include rapid sampling and metabolic inactivation protocols for samples in a wide weight range, sample extraction procedures, chemical derivatization steps that are required to make the metabolite fraction amenable to gas chromatographic analysis, routine GC-MS methods, and procedures of data processing and data mining. A basic and extendable set of standardizations for metabolite recovery and retention index alignment of the resulting GC-MS chromatograms is included. The methods have two applications: (1) The rapid screening for changes of relative metabolite pools sizes under temperature stress and (2) the verification by exact quantification using GC-MS protocols that are extended by internal and external standardization. Key words Gas chromatography, Time-of-flight mass spectrometry, GC-MS, TOF-MS, Metabolomics, Metabolite profiling, Metabolism, Relative quantification, Absolute quantification, Stable isotope labeling

1

Introduction Metabolite profiling methods are the basis of modern metabolomic approaches that aim for comprehensive analyses of biological systems [1, 2]. Targeted and nontargeted metabolic profiling methods that are in part automated and technically robust have been developed to investigate various parts of metabolism. A GC-MS–based

Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_15, © Springer Science+Business Media, LLC, part of Springer Nature 2020

203

204

Alexander Erban et al.

method that covers a wide range of primary metabolism has made a strong impact. The analytical procedures and the means to identify metabolites within the generated complex GC-MS data were easily transferable between labs [3]. The method covers, among others, sugars, amino acids, amines, organic acids, and phosphorylated metabolites and includes small secondary metabolites. The molecular size of the covered metabolites ranges between small 2-carbon metabolites, such as glycolate, glyoxylate, or glycine; and 18-carbon trisaccharides, such as raffinose. The metabolite coverage of the GC-MS profiling method that is described in the following has been thoroughly studied. The metabolite structures and compound identifications are accessible online via the Golm Metabolome Database (http://gmd.mpimp-golm.mpg.de/). Libraries that provide mass spectra and retention indices of more than 1000 metabolites and tools for mass spectral analysis have been made publicly available [4–7]. Metabolite profiling methods are fast and efficient postgenomic tools that screen for relative changes of metabolite pool sizes and can be extended to be fully quantitative. Comparisons of the effects of environmental changes, for example low temperature stress [8– 11], or genetic modifications are typically made relative to control plants of a wild type genotype that is cultivated under optimal standard conditions. The speed of profiling allows for extended experimental designs that involve typically more than 100 samples and may comprise more than 1000 samples. As a consequence, many independent events of the same genetic manipulation or breeding populations can be studied with high replication. Experimental designs that eliminate or suppress the influence of noncontrolled factors can be applied, and highly resolved time courses or dosage dependencies of metabolic responses explored. Whole experiments can easily be independently repeated. Thus, the bottleneck of metabolic physiological studies is moved back to the sound performance of well-designed physiological experiments. For most physiological questions information on the changes of relative pool sizes or on patterns of metabolic changes is sufficient to diagnose the effect of the experimental intervention. But some questions, not least those raised by the demands of systems modeling, require information on exact metabolite concentrations. Exact quantification requires extension of metabolite profiling by additional experiments, such as sample specific standardization of compound recovery. In addition, compound specific quantitative calibration samples are necessary, sometimes in numbers that can easily exceed the number of samples to be quantified. For this reason, exact quantification should never be performed before metabolite profiling has shown that the metabolite of interest is indeed among the most relevant within the screened metabolic fraction.

Multiplexed GC-MS based Metabolite Profiling

205

A second requirement after obtaining information on relevant candidate metabolites by profiling methods is the verification of metabolite identity [12]. This is a basic, but indispensable, requirement because metabolite profiles are complex and may contain hundreds of known metabolites, but also a large fraction of still nonidentified metabolites. Metabolic products also comprise a large number of chemical isomers, such as the different mono-, di-, and trisaccharides. Such isomers can be hard to distinguish and to selectively quantify. To identify compounds by GC-MS, authenticated reference substances are needed [4–7], which are employed in standard addition experiments to test for match of mass spectrum and chromatographic retention. Such standard addition experiments can be efficiently used for two purposes as they also allow for the determination of quantitative recovery. They are a prerequisite of exact quantification that tests so called matrix effects caused by the physicochemical nature of the sample. In the following protocol we describe the basic analytical modules that encompass both relative and absolute quantification by GC-MS based analysis of a metabolite fraction enriched for primary metabolites. The method is modular, serves the demand for standardized reporting [13, 14], and uses a generalizable structure that can also describe other MS based metabolite profiling methods or other variants of the GC-MS method, such as GC-MS methods that use different extraction procedures, GC capillary columns, GC settings, or GC-MS instrumentation [1, 2, 15–17]. The method starts with one example of a plant cultivation protocol and ends with the identification of relevant metabolites by mass spectral analysis and standard addition. Additional actions that are required to upgrade the basic profiling method from analysis of relative pool size changes to the quantification of absolute pool sizes are added. The described plant cultivation protocol is a hydroponic system that enables combination of routine metabolite profiling with the stable isotope analysis of metabolites by feeding of 13C- or 15 N-labeled metabolic precursors. The methods are generally applicable to plant material and to microbial photosynthetic organisms.

2

Materials Use ultrapure or bidistilled water at approximately 0.055 μS/cm, and purchase analytical grade reagents and all chemicals in best available purity. Buy small packages to avoid contaminations and loss of reagent reactivity. Buy authenticated reference substances for internal and external quantitative calibration in highest available purity and in amounts suitable for the accurate gravimetric determination of stock solutions. For some key compounds we include the CAS (Chemical Abstracts Service) number for unambiguous identification. Diligently follow recommended procedures for the safe handling of chemicals and for waste disposal.

206

Alexander Erban et al.

2.1 Plant Hydroponic Culture

1. Circular stainless steel meshes of 11 cm diameter. Stainless steel wire mesh MW: 1.40 mm—DS: 0.25 mm; Material: stainless steel V2A/1.4301, mesh size: 1.40 mm; wire thickness: 0.25 mm. 2. Hydroponic slightly tapered cultivation jars with loosely fitting, nonairtight lids. Glassware dimensions: top diameter 10 cm, bottom diameter 7.5 cm. 3. Open bottom seedling starter tray with 15 jar positions. Tray dimensions: 50 cm
30 cm. Cell diameter: 10 cm. 4. Closed bottom seedling starter tray. Tray dimensions: 54 cm
34 cm. 5. Adhesive tape 2.5 cm
5.0 cm. 6. Murashige and Skoog (MS) medium prepared according to the published protocol [18] or without the compound or inorganic elemental nutrient that will be used for stable isotope labeling. 7. Phytotronic plant growth chamber(s) adjustable to various temperature regimes (e.g., 10, 20, or 30  C day temperatures) at ~50 μmol quanta/cm2/s, higher light intensity with long day conditions at 16 h:8 h (day–night), or alternative conditions depending on experimental purposes. 8.

13

C- or 15N-labeled compounds and respective nonlabeled equivalents at highest chemical and isotopic purity available to supplement MS media.

9. Ethanol. 10. Triton X-100. 2.2 Extraction and Standardization

1. Methanol. 2. Chloroform. 3.

13

C6-Sorbitol (CAS 121067-66-1), nonadecanoic methyl ester (CAS 1731-94-8) and other authenticated reference substances for quantitative internal and external calibration as required.

4. 1.5 mL safe-lock, tapered bottom plastic microvials. 5. 2.0 mL safe-lock, round bottom plastic microvials. 6. Centrifuge for microvials. 7. Oscillating ball mill with adaptors that hold 5 or more 2 mL microvials. 8. 5 mm stainless steel balls. 9. Vacuum concentrator system with rotors for microvials. 10. Silica gel. 11. Argon.

Multiplexed GC-MS based Metabolite Profiling

207

12. Calibrated regularly checked pipetting devices in adequate volume ranges and respective disposable pipette tips. 13. Calibrated and regularly checked balance with at least 0.1 mg precision. 14. Heated shakers for microvials. 15. Vortex mixer. 2.3 Chemical Derivatization

1. Methoxyamine hydrochloride (CAS 593-56-6). 2. Pyridine (CAS 110-86-1). 3. 4-(Dimethylamino)pyridine (CAS 1122-58-3). 4. N,O-Bis(trimethylsilyl)trifluoroacetamide (CAS 24589-78-4). 5. n-alcanes: n-decane (CAS 124-18-5), n-dodecane (CAS 112-40-3), n-pentadecane (CAS 629-62-9), n-octadecane (CAS 593-45-3), n-nonadecane (CAS 629-92-5), n-docosane (CAS 629-97-0), n-octacosane (CAS 630-02-4), n-dotriacontane (CAS 544-85-4), n-hexatriacontane (CAS 630-06-8). 6. GC glass vials with crimp or screw caps and chemically inert septa. 7. Crimp-cap sealer.

2.4

GC-MS

2.4.1 Gas Chromatography

GC-MS system with electron impact and/or chemical ionization. The MS system can have nominal mass resolution or better. Electron impact ionization and atomic mass unit resolution is preferred for use with conventional GC-MS mass spectral libraries. Other systems will require the establishment of custom spectral libraries based on authenticated reference compounds. 1. We use an Agilent 6890N gas chromatograph with split/splitless injector and electronic pressure control up to 150 psi (Agilent, Bo¨blingen, Germany) to exemplify the analytical procedures and to describe the details that should be considered and reported when publishing GC-MS profiling data. Other GC systems are equally amenable to metabolite profiling analyses. 2. Low-bleeding septa. 3. Inert conical single taper liner with glass wool for split/less injection. 4. Low-bleeding GC capillary column suitable for hyphenation to mass spectrometry systems. The stationary phase needs to be stable in the presence of trimethylsilylation and methoxyamination reagents. A 5% phenyl–95% dimethylpolysiloxane fused silica capillary column with 30 m length, 0.25 mm inner diameter, 0.25 μm film thickness and an integrated 10 m precolumn are preferred for the separation of chemically derivatized

208

Alexander Erban et al.

primary metabolites and for use with retention index libraries that are exchangeable between metabolite proofing laboratories [16, 17]. Alternatively, a 35% phenyl–65% dimethylpolysiloxane fused silica capillary column with 30 m length, 0.32 mm inner diameter, and 0.25 μm film thickness may be used [15, 17]. The use of more polar or other stationary phases will require the establishment of custom retention index libraries based on authenticated reference compounds [1, 2, 4–7, 13–17]. 5. Helium 5.0 carrier gas. 6. n-hexane. 7. Ethyl acetate. 2.4.2 Mass Spectrometry

1. A Pegasus III time-of-flight mass spectrometer (LECO Instrumente GmbH, Mo¨nchengladbach, Germany), to exemplify the details of this module that should be considered and reported when publishing GC-MS profiling data. Other MS systems are equally amenable to metabolite profiling analyses. 2. A micrOTOF-Q II hybrid quadrupole time-of-flight mass spectrometer (Bruker Daltonics, Bremen, Germany) equipped with a multipurpose APCI source and an ESI source and otofControl software 4.0, to exemplify the details of this module that should be considered and reported when publishing GC-MS profiling data. Other MS systems are equally amenable to metabolite profiling analyses. 3. Perfluorotributylamine (PFTBA) to calibrate the MS system. 4. Na-formate clusters: Mix 12.5 mL H2O, 12.5 mL isopropyl alcohol, 50 μL concentrated formic acid, and 250 μL 1 M NaOH. Mix severely. Consider safe lab work, using and mixing these chemicals.

3

Methods

3.1 Plant Cultivation and Labeling

The described hydroponic system allows for separation of shoot and root tissue and is amenable to Arabidopsis thaliana (Arabidopsis) or other plants that fit the dimensions of the selected glass jars. Keep the environmental conditions of the phytotronic chambers as reproducible and controlled as possible. Environmental factors that may introduce noise to the growth experiment are the orientation of the light bulbs, light intensity gradients, temperature gradients, noncontrolled light- or temperature-fluctuations during day and night. Additional confounding factors may arise from the use of multiple independent phytotronic chambers.

Multiplexed GC-MS based Metabolite Profiling

209

1. Fold stainless steel meshes to a ~8 cm diameter before placing them inside the glassware. The diameter of the folded mesh and the slightly tapered glassware define the height level of the hydroponic solution. For example, a total diameter of 8 cm results in 5 cm height of hydroponic solution. 2. Dry-sterilize the assembled glassware with the metal meshes inside before sowing. 3. Fill each container with autoclaved media of interest. For example, 250 mL of MS medium [18] that may be adjusted to specific labeling purposes or feeding (e.g., 2% sucrose for cultivation in the presence of a carbohydrate source). 4. Surface-sterilize seeds using 400 seeds per aliquot inside 1.5 mL microvials covered with 1 mL of 70% ethanol + 0.5% Triton X-100. Shake at 600 rpm for 20 min. Remove supernatant. 5. Wash with 100% ethanol for 5 min to remove traces of detergent. 6. Pour the contents of the microfuge vials on autoclaved filter paper inside a laminar flow cabinet and let the ethanol evaporate. 7. Sow the dry seeds in petri dishes filled with solid MS-agar medium [18]. Leave the seeds imbibed [19] on petri dishes for 5 days at 4  C in the dark to synchronize germination. 8. Shift the sown seeds after 5 days of imbibition and stratification to a 20  C phytotron chamber. Seeds will germinate within 24 h. After 5 days of germination in petri dishes Arabidopsis seedlings reach stage 1.0 with fully opened cotyledons [20]. 9. Transfer plants 5 days after germination or less to avoid root material reaching the bottom of the dish, which would imply additional mechanical stressors upon transfer to the metallic mesh inside the glass jars. 10. Place each seedling on the mesh in a 1 cm3 piece of the surrounding solid MS-agar medium (Fig. 1a). Avoid overcrowding of Arabidopsis plants on the meshes; for example, if the desired harvesting stage is ~1.10 to 1.12, up to 15 plants per jar may be sown [20]. Less number of plants are recommended for sampling at later developmental stages to prevent early plant-to-plant competition and shading, cf. assembled glass jar (Fig. 1b) and transferred plantlets. 11. Place eight glass jars in each seedling starter tray leaving positions empty between each jar to prevent misbalance. Use an 8.5 cm diameter petri dish to fill the empty wells of the seedling starter tray (Fig. 1c). The dishes act as size adaptors between glass jars and rack. The adaptors align the top edge of the seedling starter tray to the mesh height in each glass jar and

210

Alexander Erban et al.

Fig. 1 Hydroponic plant cultivation. (a) Arabidopsis thaliana plants (Col-0) at developmental stage ~1.10. Plantlets were transferred with agar blocks (white arrow), precultivated hydroponically at 20  C:18  C (day–night) with 16 h:8 h light–darkness in liquid MS medium with 2% sucrose. The medium was substituted at stage 1.10 by 15N-ammonium nitrate. (b) Sterilized glass cultivation jars with stainless steel metal mesh inserts (white arrow) arrayed within the starter tray setup. (c) 15-position seedling starter tray with petri dish height adaptors (white arrow) in half of the wells. The black open-bottom insert tray is positioned in a green nonpartitioned closed bottom seedling starter tray

thereby limit light exposure of the root systems. Further reduction of root light exposure is achieved by a second closed bottom seedling starter tray, cf. assembled seedling starter trays (Fig. 1c). 12. Reduce within chamber variability by rotating the glass jar positions daily, thereby compensating and minimizing influences of light and temperature gradients in the phytotronic chambers. 13. Shift plants at defined developmental stages to altered temperature and/or to altered medium compositions, for example, by substitution of nonlabeled MS-medium components by stable isotope precursors at identical concentration without chemical changes to the medium composition (see Note 1). 14. Minimize block/tray effects by pooling material across jars and different trays when assembling biological replicates. 3.2 Sampling and Gravimetric Determination of the Sample Amount

Sampling methods for metabolic profiling must be performed in situ with minimal disturbance of the plant’s environment. Specifically, the temperature and illumination of leaves must not be changed prior to or during sampling. Metabolic inactivation must be immediate and needs to be maintained during subsequent sample processing. All other factors that influence metabolism must be controlled by the experimental design. The experimental design must include a randomization or arraying strategy to account for residual experimental factors that cannot be completely controlled. For relative quantification of metabolite pool sizes, the amount of

Multiplexed GC-MS based Metabolite Profiling

211

all samples must be in the same range with a defined maximal tolerance. The amount can be defined either by fresh or dry weight, number of cells or other quantitative properties that describe the amount of sample. For absolute quantification, the exact amount of the sample must be determined and the specific recovery of each analyte has to be evaluated. 1. Prepare 1.5 or 2.0 mL microvials prior to sampling. 2. Number vials using a permanent marker. 3. Precool and keep vials in liquid nitrogen. 4. Determine the empty weight of each vial while frozen including the hoarfrost that may form during the process. Return vial to liquid nitrogen. 5. Take a vial from liquid nitrogen, cut a sample from a plant and seal sample into vial. Return the loaded vial to liquid nitrogen within 10 s from cutting or faster. 6. Determine the weight of the loaded vial while frozen including the hoarfrost that may form (see Note 2) (Fig. 2). Return loaded vial to liquid nitrogen. Samples can be stored at this step at 80  C. 7. The sample amount of a single leaf for absolute quantification must not be lower than 2.5 mg and must not exceed 125.0 mg fresh weight for extraction in 2 mL microvials. The sample amount for relative quantification should be kept constant, ideally with a tolerance of 5–10% (see Note 3). 8. Add a precooled stainless steel ball to each vial. The curvature of the steel ball must not exceed the curvature of the bottom of the microvial. Load microvials into a precooled mounting adaptor of an oscillating ball mill. Homogenize samples to a fine powder by 1.0 min bursts at 15 s1 frequency (see Note 4). Keep samples below 60  C throughout the process. If necessary, return the loaded adaptors to liquid nitrogen between bursts. Store homogenized samples without removing the steel balls at 80  C until further processing (see Note 4). 9. Keep, store, and continue to process a set of 5–10 empty microvials from the batch used for each respective experiment in parallel through all subsequent steps. These vials are processed as nonsample controls and allow for the determination of the specific chemical laboratory background of each experiment (see Note 5). 10. Prepare a set of quality control samples that is used to control the performance of the GC-MS profiling system. Specifically control for system drifts when comparing between independently repeated experiments that cannot be processed together. For this purpose, prepare a large representative leaf

212

Alexander Erban et al.

Fig. 2 Precision of gravimetric weight determination of samples that were shock frozen in liquid nitrogen. The relative standard deviation (RSD) of the direct weighing of test weights was determined at room temperature using a balance with a tolerance of 0.1 mg. Test weights were made of aluminum foil and covered the weight range of ~1.0 to 50.0 mg. Replication number per weight equaled 6. The statistical test applied to define the absence of significantly changed weight means was Tukey HSD after ANOVA. Aluminum foil was folded to brick shape as reference weight. (a–c) Test weights of ~1, ~10 and 50 mg (top to bottom). (d) Averaged relative standard deviation of combined deepfrozen conditions 5–6. The conditions were left to right, (1, 2) room temperature

Multiplexed GC-MS based Metabolite Profiling

213

sample from immediately frozen material that combines all sample types of the experimental design. Homogenize thoroughly with a mortar and pestle under liquid nitrogen, avoid hoarfrost during homogenization. Generate equal frozen aliquots of the average sample fresh weight of the experiment with a tolerance of 5–10%. Keep, store, and continue to process in parallel a set of 5–10 quality control samples through all subsequent steps of the analysis. Store surplus quality control samples at 80  C until processing of the next and subsequent independently repeated experiments. Extraction methods determine and delimit the chemical nature and coverage of metabolite profiling methods. The extraction method that is described in the following generates a fraction that is enriched for polar primary metabolites while volatiles, highly lipophilic metabolites, and complex lipids are removed. Extraction efficiencies can vary between metabolites depending on the choice of solvent. Extraction efficiency can also depend on variations of the physical properties or chemical composition between the compared sample types. So far we did not observe variable matrix effects in temperature stress experiments. However, frequent control experiments that test the recovery of metabolites are advised. Moreover, knowledge of metabolite recovery is required to verify temperature regulated metabolites by exact quantification. Isotope-labeled internal standards can be added during extraction to determine the specific recovery of each quantified metabolite from each investigated sample. Alternatively, nonlabeled authenticated reference compounds can be added at a constant concentration to representative samples. This process estimates specific constant factors of metabolite recovery for each type of profiled sample. Both types of recovery experiments will test the overall recovery of the method including matrix effects that may occur at the subsequent steps of the method. Isotopic tracing experiments are best not combined with quantitative internal standardization using labeled authenticated reference compounds. ä

3.3 Metabolite Extraction and Standardization

Fig. 2 (continued) weighting of the test weights (two independent series of measurements), (3, 4) test weight inside 2.0 or 1.5 mL vials determined at room temperature, (5, 6) test weight inside liquid N2 frozen 2.0 or 1.5 mL vials. For this purpose, empty vials were precooled with liquid nitrogen. Empty vials were placed on the scales without aluminum foil to tare the scale. Empty vials were cooled down again in liquid nitrogen, then put back to the scale with aluminum foil bricks added to determine the differential weight. Note that the precision of weighting while frozen was below 15% relative standard deviation. The precision of weighting, besides reproducible timing and sequencing of weighting events, highly depends on avoiding air-humidity that may condensate on cold surfaces

214

Alexander Erban et al.

1. For the extraction of a set of approximately 150 samples, prepare 50 mL fresh 90% methanol: water (v/v) extraction solvent and add methanol soluble internal standards. For routine profiling dissolve 0.02 mg/mL 13C6-sorbitol in this solvent. 2. Prepare 35 mL fresh chloroform solvent with chloroform soluble internal standards. Dissolve 0.25 mg/mL nonadecanoic acid methyl ester in chloroform to test for the presence or absence of lipophilic compounds. 3. Prepare 60 mL fresh bidistilled water. 4. Dissolve other additional internal standards (e.g., stable isotope labeled or nonlabeled xenobiotic internal standards) according to their solubility either in the methanol solvent, the chloroform solvent, or the bidistilled water solvent. Adjust the concentration of internal standards to concentrations that approximate the expected endogenous metabolite concentrations (see Notes 6 and 7). 5. Prepare dilution series for quantitative external calibration of metabolites. Prepare a stock mixture of authenticated reference substances at up to 10 mg/mL in 90% methanol, in chloroform or in bidistilled water according to compound solubility. Choose the relative amount of each metabolite in a stock mixture to mimic the composition that is expected within the analyzed samples. Prepare dilution series to cover the highest and lowest expected amount. Perform prior test experiments to adjust the calibration series appropriately. Process calibration series in parallel through all subsequent steps of the method. 3.3.1 Methanol Extraction with Liquid Partitioning into Chloroform

Methanol extracts may contain an excess of complex lipids and chlorophyll. These compound classes are nonvolatile and can cause frequent maintenance of the GC-MS system unless liquid partitioning into chloroform removes these compound classes. 1. Add 330 μL methanol solvent containing the internal standards to a microvial with 100 mg  5–10% frozen sample powder (see Note 8). Do not remove the steel ball during extraction. 2. Mix thoroughly using a vortex mixer and shake all samples simultaneously for 15 min at 70  C. Vent microvials after 1 min at 70  C to release excess vapor pressure. 3. Cool to room temperature. 4. Add 230 μL chloroform solvent containing internal standards. 5. Mix thoroughly using a vortex mixer and shake all samples simultaneously for 5 min at 37  C. 6. Add 400 μL bidistilled water containing internal standards, if added.

Multiplexed GC-MS based Metabolite Profiling

215

7. Mix thoroughly using a vortex mixer and separate liquid and solid phases by centrifugation for 5 min at 14,000 rpm or 20,800
g (see Note 9). 8. Transfer an 80 μL aliquot of the upper phase, which contains the polar metabolic complement of the sample, to a 1.5 mL safe-lock microvial or directly to a GC vial and dry in a vacuum concentrator. Prepare microvials with two or more backup samples of each extract (e.g., of a second 80 μL and a 160 μL aliquot). For external calibration series take identical aliquots and prepare microvials to run at least two calibration series in parallel to each experimental sample set and prepare multiple backups for use in subsequent experiments. 9. Seal dried samples with closed caps under inert gas (e.g., argon), in plastic bags with silica gel. Store at 20  C or colder until further processing. Before reopening the plastic bags allow equilibration to room temperature and remove condensed water. 3.3.2 Methanol Extraction Without Liquid Partitioning into Chloroform

This method is used to include small lipophilic metabolites and may require frequent maintenance of the GC-MS system. This method can be applied to small, dilute samples or to samples that contain a smaller proportion of complex lipids and chlorophyll. 1. Prepare methanol extraction solvent with methanol soluble internal standards. Dissolve 0.005 mg/mL 13C6-sorbitol in 100% methanol. 2. Dissolve other stable isotope labeled or nonlabeled xenobiotic internal or external standards in methanol at concentrations that approximate the expected endogenous metabolite concentrations (see Note 6). 3. Extract samples that vary in fresh weight with a constant 40:1 (v/w) solvent to fresh weight ratio. In detail, add 400 μL methanol solvent to a microvial with 10 mg frozen sample powder. Scale up extraction volume in proportion to sample fresh weight (see Note 10). Do not remove the steel ball during extraction. 4. Mix thoroughly using a vortex mixer and shake all samples simultaneously for 15 min at 70  C. Vent microvials after 1 min at 70  C to release excess vapor pressure. 5. Cool to room temperature and centrifuge for 5 min at 14,000 rpm or 20,800
g. 6. Transfer the complete supernatant, that is, a 350 μL aliquot of the liquid phase, to a 1.5 mL safe-lock microvial or directly to a GC vial and dry in a vacuum concentrator. 7. Seal dried samples with closed caps under inert gas (e.g., argon) in plastic bags with silica gel. Store at 20  C or colder until

216

Alexander Erban et al.

further processing. Before reopening the plastic bags allow equilibration to room temperature and remove condensed water. 3.3.3 Methanol– Chloroform Extraction Without Liquid Partitioning

This method is used to include a higher fraction of small lipophilic metabolites and may require more frequent maintenance of the GC-MS system. To include more lipophilic compounds compared to the method described in Subheading 3.3.2, chloroform is added. This method can be applied to small, diluted samples or to samples that contain an even smaller proportion of complex lipids and chlorophyll. 1. Perform all steps described in Subheading 3.3.1 but do not add 400 μL bidistilled water. Centrifuge 5 min at 14,000 rpm or 20,800
g to separate solids from liquid supernatant.

3.4 Retention Index Standardization for GC Analysis

Gas chromatography is subject to drifts of retention time [17]. Such drifts interfere with the retention time alignment of GC-MS chromatograms, especially if an experiment comprises a high number of samples. Retention index standards are used to align GC chromatograms and to improve compound identification by matching to retention index libraries of authenticated reference compounds. 1. Prepare retention index (RI) standard mixture of n-alkanes in pyridine. Combine n-decane (RI 1000), n-dodecane (RI 1200), n-pentadecane (RI 1500), n-octadecane (RI 1800), nnonadecane (RI 1900), n-docosane (RI 2200), n-octacosane (RI 2800), n-dotriacontane (RI 3200), and n-hexatriacontane (RI 3600) at a final concentration of 0.22 mg/mL each, except n-decane and n-hexatriacontane which are added at 0.44 mg/ mL. Alternatively, a mixture of fatty acid methyl esters can be used [15, 17]. 2. To calculate retention indices use the following definition: RI of an n-alkane equals the number of carbons multiplied by 100. Apply the method for linear temperature programmed gas chromatography [21].

3.5 Chemical Derivatization for GC Analysis

Chemical derivatization reactions are required to modify the structure of nonvolatile compounds to form volatile products that can be analyzed by GC. The choice of derivatization reactions determines and delimits the coverage and sensitivity of metabolite profiling methods [22]. The chemical derivatization method that is described in the following is essentially the same as described previously [1, 2, 15, 16]. The reactions have low specificity, high yields for almost complete conversion and generate volatile derivatization products (i.e., the analytes) of most stable primary metabolites (see Note 11).

Multiplexed GC-MS based Metabolite Profiling

217

1. Prepare fresh methoxyamine reagent daily. Dissolve first 5 mg/ mL 4-(dimethylamino)pyridine in pyridine, then add methoxyamine hydrochloride to a final concentration of 40 mg/mL (see Note 12). 2. Prepare trimethylsilylation reagent. Mix fresh N,O-bis(trimethylsilyl)trifluoroacetamide and retention index standard mixture dissolved in pyridine (see Subheading 3.4) in a 7:1 (v/v) ratio. Avoid humidity and do not store opened bottles of N,O-bis(trimethylsilyl)trifluoroacetamide or preparations of trimethylsilylation reagent (see Note 13). 3. Mix dried extract thoroughly with 40 μL methoxyamine reagent using a vortex mixer and shake all samples simultaneously for 90 min at 30  C (see Note 14). 4. Add 80 μL trimethylsilylation reagent and mix thoroughly using a vortex mixer. Incubate for 30 min at 37  C (see Note 14). 5. Perform reactions either in GC vials or transfer 80 μL to a GC vial. Avoid humidity, close vial immediately, and keep vials at room temperature on a GC autoinjector system until injection (see Note 15). 6. If a chemically derivatized sample is reanalyzed by GC-MS or used for multiplexed analysis, either exchange the septum of the used GC vial after first analysis or perform multiplexed analysis immediately, for example, by sequential split and splitless injections (see Subheadings 3.6.1 and 3.6.2) or EI and APCI (see Subheading 3.8) (Fig. 3). If immediate multiplexed analysis is not be possible, do not store GC vials with punctured septa for extended periods, but rather use a not yet derivatized backup sample (see Note 13). 3.6 Injection for GC Analysis

3.6.1 Preparing Sample Injection

The available gas chromatography settings for GC-MS systems are typically inbuilt features of the employed GC-MS system. The choice of injection technology for gas chromatography modifies the amount of the chemically derivatized sample that is transferred onto the capillary column. The temperature, pressure and gas flow during injection may influence peak shape. 1. Mount a 10 μL syringe on the GC injection system. 2. Mount a new conical single taper split/splitless liner with glass wool into the injector port of the GC system before analysing a new set of samples. 3. Before each sample injection, clean syringe by full volume draws of pure ethyl acetate and n-hexane.

218

Alexander Erban et al.

A

Intensity (arbitrary units)

Intensity (arbitrary units)

C

73

EI

103

217 147 307

time Exact masses of molecular ions and fragments

Si

307.1581 O

M: 569.2876 M+H: 570.2954

Si O Si

O O

Si O

103.0579

Si

N

O 103.0579

D

100

200

300

400

500

m/z

M+H: 570.2927

APCI

Intensity (arbitrary units)

B

307.1559 217.1059

103.0560

100

200

300

400

500

m/z

Fig. 3 Multiplexing gas chromatography–mass spectrometry (GC-MS) based metabolite profiling. (a) Overview scheme of multiplexing GC-MS profiling. The same sample is measured by electron impact ionization (EI) and atmospheric pressure chemical ionization (APCI) GC-MS systems to obtain an extended range of detection. Splitless and split measurements with optimized split ratios may be additionally implemented to extend the range towards highly abundant metabolites. (b) Fructose after methoxyamination and trimethylsilylation yields the analyte Fructose (5TMS) (1MEOX). Exemplary fragments with exact masses and expected molecular ions and molecular weights are indicated. (c) EI mass spectrum with decreasing sensitivity for high molecular weight fragments and molecular ions. (d) APCI mass spectrum with decreasing sensitivity of low molecular weight fragments. Note that both ionization technologies yield largely overlapping sets of fragments from EI and in-source fragmentation after APCI. In combination, complex data-rich mass spectra are available for detailed interpretation of fragmentation reactions of known and unknown analytes. Mass accuracy of APCI spectra can be improved by m/z recalibration using known and validated exact masses of coeluting compounds from the same chromatogram and chromatographic region

4. Perform at least five injections of N,O-Bis(trimethylsilyl)trifluoroacetamide after a change of syringe, liner, or GC column (see Subheading 3.7).

Multiplexed GC-MS based Metabolite Profiling 3.6.2 Splitless Injection for GC Analysis

219

1. Inject 1 μL of chemically derivatized sample at 230–250  C in splitless mode with helium carrier gas flow set to 0.6 mL/min. 2. Adjust purge time to 1 min with purge flow set to 20 mL/min. 3. Keep the flow rate constant and electronic pressure control enabled (see Note 16).

3.6.3 Split Injection for GC Analysis

1. Inject 1 μL chemically derivatized sample at 230–250  C in split mode at a split flow ratio of at least 1:30 with helium carrier gas flow set to 0.6 mL/min. 2. Adjust purge time to 1 min with purge flow set to 20 mL/min. 3. Keep the flow rate constant and electronic pressure control enabled (see Note 16).

3.7

GC Analysis

3.7.1 5% Phenyl–95% Dimethylpolysiloxane Fused Silica Capillary GC Column

The type and dimensions of capillary GC columns, the temperature ramping and the flow/pressure settings of the carrier gas determine the scope of the profiling method and the elution sequence of the analytes. Minor shifts of retention indices between GC-MS systems using identical column types can be mathematically compensated [17, 21]. 1. Mount a 5% phenyl–95% dimethylpolysiloxane fused silica capillary column, with 30 m length, 0.25 mm inner diameter, 0.25 μm film thickness, and an integrated 10 m precolumn. Use a low-bleeding column suitable for mass spectrometry. 2. Operate system with helium carrier gas set to 0.6 mL/min constant flow. 3. Start the temperature program isothermal for 1 min at 70  C, ramp to 350  C at 9  C/min, keep at 350  C for 5 min. Cool and return to initial conditions as fast as instrument specifications allow for. Attention: to avoid damage of the transferline coupled to the micrOTOF-Q do only ramp up temperature to 330  C! 4. Set the transfer line temperature to 250  C, in case of the micrOTOF-Q GC-APCI System 290  C.

3.7.2 35% Phenyl–65% Dimethylpolysiloxane Fused Silica Capillary GC Column

1. Alternatively, mount a 35% phenyl–65% dimethylpolysiloxane fused silica capillary column, with 30 m length, 0.32 mm inner diameter, and 0.25 μm film thickness or use other low-bleeding columns suitable for mass spectrometry. 2. Operate system with helium carrier gas set to 2.0 mL/min constant flow. 3. Start the temperature program isothermal for 2 min at 85  C, ramp to 350  C at 15  C/min, keep at 360  C for 0 min. Cool and return to initial conditions as fast as instrument specifications allow for. Attention: to avoid damage of the transferline

220

Alexander Erban et al.

coupled to the micrOTOF-Q do only ramp up temperature to 330  C! 4. Set the transfer line temperature to 250  C, in case of the micrOTOF-Q GC-APCI System to 290  C. 3.8

MS Analysis

3.8.1 Mass Calibration for Mass Spectrometry: Electron Impact Ionization (EI)

The mass spectrometry settings for GC-MS systems are typically inbuilt features of the employed system. The mass calibration of GC-MS systems is typically part of inbuilt auto-tuning processes. Mass calibration and instrumental limitations define the mass range which can be chosen for analysis. 1. Trigger the auto-tuning process with PFTBA to calibrate the MS system before processing a set of samples comprising a metabolite profiling experiment. 2. Set the mass range to m/z ¼ 70–600. GC-EI-MS systems typically operate at or close to nominal mass resolution and accuracy.

3.8.2 Mass Calibration for Mass Spectrometry: Atmospheric Pressure Chemical Ionization (APCI)

1. The mass spectrometer is externally tuned via direct injection for mass-calibration coupled to an ESI source. Inject Na-formate clusters (see Note 17) for external tuning with a flow rate of 3 μL/min until a Score of 100% at mass quality (zooming) 0.001% with enhanced quadratic calibration mode is attained. Tune at a mass range of 50–1500 m/z, use rolling average of three spectra at 1 Hz at positive mode with a given list of expected fragments resulting from Na-formate clusters. Use default mass spectrometer settings for direct infusion of small molecules. Settings for “Source” in the otofControl software: End Plate Offset 500 V; Capillary 4500 V; Nebulizer 0.4 Bar; Dry Gas 4.0 L/min; Dry Temp 180  C. Settings for “Tune” in the otofControl software: Transfer Funnel 1 RF 150.0 Vpp; Transfer Funnel 2 RF 200.0 Vpp; Transfer isCID Energy 0.0 eV; Transfer Hexapole 50.0 Vpp; Quadrupole Ion Energy 4.0 eV; Quadrupole Low Mass 90.00 m/z; Collision Cell Collision Energy 7.0 eV; Collision Cell Collision RF 650 Vpp; Collision Cell Transfer Time 80.0 μs; Collision Cell Pre Pulse Storage 5.0 μs. 2. Running the system after external calibration with the APCI source for GC data acquisition, PFTBA is injected prior to the solvent peak into each chromatogram for later internal calibration of each file. Other analytes that are part of the sample (e.g., internal standards or always occurring known compounds) may be used for this purpose as well if exact atomic compositions of resulting molecular ions and fragments are known and the mass spectra cover the mass range of interest.

Multiplexed GC-MS based Metabolite Profiling

3.9 Ionization for MS Analysis 3.9.1 Ionization for Mass Spectrometry: Electron Impact Ionization (EI)

3.9.2 Ionization for Mass Spectrometry: Atmospheric Pressure Chemical Ionization (APCI)

221

The ionization process determines the mass spectrum of compounds and thereby the availability and chemical nature of molecular ions and mass fragments that can be used for specific and selective analysis of compounds in complex mixtures. Use electron impact ionization, set the ion source temperature to 250  C and the filament bias current to 70 eV. Optimize detector voltage depending on detector age to approximately 1500–1950 V (see Note 18), to obtain the same sensitivity in all measurements. Details of this procedure must be considered and reported when publishing GC-MS profiling data (see Note 19). 1. Settings for “Source” in the otofControl software: End Plate Offset 500 V; Capillary 2000 V; Corona 300 nA; Nebulizer 3.5 Bar; Dry Gas 2.5 L/min; Dry Temp 250  C. 2. Settings for “Tune” in the otofControl software: Transfer Funnel 1 RF 300.0 Vpp; Transfer Funnel 2 RF 300.0 Vpp; Transfer CID Energy 0.0 eV; Transfer Hexapole 60.0 Vpp; Quadrupole Ion Energy 4.0 eV; Quadrupole Low Mass 50.00 m/z; Collison Cell Collision Energy 8.0 eV; Collison Cell Collision RF 750 Vpp; Collison Cell Transfer Time 60.0 μs; Collison Cell Pre Pulse Storage 5.0 μs.

3.10 Mass Spectrometric Analysis

The type of mass spectrometric analysis determines speed of mass spectral scanning and the accuracy and precision of mass recordings. The latter may determine the availability of specific ions for the selective monitoring of compounds in complex mixtures. A scanning rate of 20 scans/s is recommended (see Note 20).

3.11 Data Mining: Electron Impact Ionization (EI)

Already at typical nominal mass resolution of GC-EI-MS systems the algorithms used to process GC-MS chromatogram data files and those that are applied to pick peak apices and respective peak responses, so called mass features, can influence relative and absolute quantification.

3.11.1 Chromatogram Data Processing and Generation of Comprehensive Peak Lists and Data Matrices

1. Check the quality of raw chromatogram files without preprocessing using the manufacturer’s file format and respective manufacturer’s software. Avoid the following analytical artifacts through system maintenance: Column bleeding and chemical background caused by laboratory contaminations; chromatography artifacts (e.g., unusual peak shapes, peak tailing, overall retention drifts, and shifts of peak position); mass spectral artifacts (e.g., absence of baseline responses, presence of positive or negative electronic spikes, drift of mass calibration); and quantitative artifacts (e.g., peak overloading, drifts of overall recovery or changes of recovery of compound classes, and fast loss of detector sensitivity) (see Note 21).

222

Alexander Erban et al.

2. Eliminate single deviant chromatograms before subsequent data processing, reanalyze full experimental sets by split injection in cases of peak overload (see Subheading 3.6.2) or with a more sensitive system in case of low abundance (see Subheading 3.8.2) (Fig. 3). 3. Perform baseline correction above noise, apply smoothing algorithm set to five scans, with expected peak width of ten scans, and export a chromatogram file in an interchange format (e.g., CDF-format) using the manufacturer’s software options. It is best to perform baseline correction with baselines orthogonal to peak heights. 4. Generate a comprehensive peak list of each chromatogram that contains all observed mass features above a signal-to-noise ratio of 2. Mass feature information contains monitored mass (m/z ratio), retention index (arbitrary RI units) (see Subheading 3.4), and respective detected response (arbitrary intensity units). For this purpose, use, for example, TagFinder software [23, 24] or other equivalent software (see Note 22). Convert manufacturer’s peak response values, if necessary, into integers, round manufacturer’s m/z values according to the instrument’s mass precision or into integers representing nominal mass units and calculate retention indices according to given preciseness of the time-scale (Subheading 3.4). 5. Archive the raw chromatogram files in manufacturer’s data format. Archive the processed chromatogram files in the chosen interchange format. Archive the comprehensive peak list obtained from each chromatogram. 3.11.2

Mass Alignment

Comprehensive peak lists of multiple chromatograms are combined into tables that represent a matrix of detected responses of all mass features across all chromatograms of an experimental data set. The process requires a mass alignment step, more precisely alignment according to m/z ratio. This alignment ensures that the mass features that are merged across the different individual chromatograms have identical mass within the limits of the instrument’s mass precision. 1. Round the m/z values from the preciseness provided by the GC-MS. 2. Align according to equal mass (see Note 23). 3. Store mass of the merged mass feature in the resulting tabular matrix of an experimental data set.

3.11.3 Chromatographic Alignment

Chromatographic alignment of comprehensive peak lists is required for the generation of tables that contain the detected responses of observed mass features across all chromatograms of an experimental data set. The chromatographic alignment ensures that the mass

Multiplexed GC-MS based Metabolite Profiling

223

features that are merged across the different individual chromatograms elute within an identical chromatographic window. 1. Use retention indices for chromatographic alignment [17]. Calculate retention indices of all mass features within the peak lists of single chromatograms. Use the recorded retention times of standards that are added to each individual chromatogram (see Subheading 3.4). 2. Sort observed mass features with identical nominal mass from all peak lists constituting an experimental set according to retention indices. Group into common retention index intervals by defining the gap width between neighboring intervals and by defining the number of tolerated random observations within the gap; for details refer to previous publications [23, 24]. 3. Store the alignment of merged mass features with information on minimal, average, and maximal observed retention index including retention index width of the interval in the resulting tabular matrix of an experimental data set (see Note 24). 3.11.4 Grouping of Aligned Mass Features

Mass features from mass spectrometric analyses carry redundant quantitative information, due to induced fragmentation of the molecular ion after ionization and due to the presence of mass isotopomers that result from the presence of naturally occurring stable isotopes. Such redundancies can be reduced by grouping of mass features that represent the same analyte. 1. Group all mass features with overlapping retention index intervals, that is, a time group of mass features. 2. Group all mass features within a time group that have correlated responses across all chromatograms of an experimental data set, that is, a cluster of mass features. This correlation approach does not work for instable fragment ratios introduced by flux experiments. In flux experiments the sums of individual isotopomer traces need to be correlated to each other. 3. Store all mass features with group assignments within tabular data set for nontargeted data mining (see Note 25).

3.12 Data Mining: Atmospheric Pressure Chemical Ionization (APCI) 3.12.1 Chromatogram Data Processing and Generation of Comprehensive Peak Lists

Using high mass resolution of GC-APCI-MS requires more accurate mass alignment and the algorithms used to process GC-MS chromatogram data files and those applied to pick peak apices and respective peak responses, the so-called mass features, may influence relative and absolute quantification. 1. Check the quality of raw chromatogram files without preprocessing and eliminate single deviant chromatograms before subsequent data processing as detailed in Subheading 3.11.1.

224

Alexander Erban et al.

2. If necessary, reanalyze full experimental sets by split injection in cases of peak overload (see Subheading 3.6.2) or with a more sensitive system in case of low sensitivity (see Subheading 3.8.2) (Fig. 3). 3. Archive the raw chromatogram files in manufacturer’s data format, for potentially mandatory upload into public databases (e.g., https://www.ebi.ac.uk/metabolights as in [25]). 3.12.2

Mass Alignment

Comprehensive peak lists of multiple chromatograms are combined into tables that represent a matrix of detected responses of all mass features across all chromatograms of an experimental data set. The process requires a mass alignment step, alignment according to m/ z ratio at instrument mass precision. This alignment ensures that the mass features that are merged across the different individual chromatograms have identical mass within the instrument’s limits. Round the m/z values according to the preciseness provided by the GC-MS system. For the given system, round either at 1 mDa for a single chromatogram evaluation or at 50 mDa for sets of chromatograms; be sure this preciseness can be obtained for each analyte in each chromatogram within a dataset.

3.12.3 Chromatographic Alignment and Analysis of Mass Features

We recommend multitargeted data analysis of GC-APCI-MS at the high mass resolution and the use of vendor software. Multitargeted analysis can also be established using nonvendor software. Alternatively, round to nominal mass precision and export in CDF file format for the rapid nontargeted screening using, for example, the TagFinder procedures (see Subheading 3.11.1).

3.13 Response Normalization of Aligned Mass Features After Data-Matrix Generation

Obtained data matrices of mass features from mass spectrometric analyses need to be normalized prior to relative quantification to account for variation in sample amount and for variation due to differential loss in the course of the analytical process. Exact quantification requires in addition a metabolite-specific recovery factor and co-processed external calibration series to calculate metabolite concentrations. 1. Normalize responses of all remaining mass features, for example, to sample fresh weight and to the response of the internal standard (e.g., 13C6-sorbitol). 2. Use normalized responses directly for statistical analyses or transform normalized responses according to the experimental design and upcoming statistical necessities. For example, calculate ratios of each normalized mass feature to the median of the nontreated control group or to the median response of all chromatograms of an experimental set per mass feature

Multiplexed GC-MS based Metabolite Profiling

225

individually, subsequently calculate logarithms of the ratios to equally represent fold increases and fold decreases. 3. Apply tools for data visualization, for example, principal component analysis (PCA), hierarchical cluster analysis (HCA), or clustered heat map display of the tabular normalized data to analyze the variation between replicate samples representing identical conditions compared to the variation between conditions. Visualize data to analyze the groups of response patterns of mass features across all studied conditions. 4. Apply statistical procedures, such as ANOVA and post hoc tests or their nonparametric counterparts, with correction for multiple testing or use more elaborate approaches (e.g., [11, 25]), to discover experimentally relevant mass features or clusters of mass features (see Note 26). During the discovery phase focus on clusters that comprise at least three mass features. Also consider mass features that are absent in control conditions and present in other conditions or vice versa (see Note 27). For the targeted exact quantification of specific metabolites, select at least one selective mass feature (unique and quantitative) and use external calibration curves to calibrate the apparent concentration. Perform quantification within the range of the upper and lower limits of quantification of the calibration curve. Avoid peak overloading by reanalysis using split injection (see Subheading 3.6.2). Never extrapolate beyond the limits of quantification. Correct for the recovery of an isotopically labeled or other internal standard within each chromatogram or for the recovery factor determined in preceding recovery experiments to obtain the concentration within the sample. 3.14 Nontargeted Data Mining of Relevant Mass Features

Nontargeted data mining is a prerequisite for nonbiased discovery of relevant mass features from metabolite profiling experiments. Many options exist, as for example reviewed by Wolfender and coauthors [26]. The following approach and workflow includes previous suggestions but is specialized and can be used to find mass features that differentially accumulate (e.g., stable isotope incorporation). The focus lies on statistical relevance and explanation of datasets. This workflow equally weighs large and small changes of relative metabolite pool sizes, reducing bias of large variances. This implementation becomes necessary, if we consider that small changes in the abundance of signaling molecules can be amplified, triggering larger signaling cascades [27]. Untargeted data mining can be divided into four steps. Preprocessing entails the mathematical transformation of data matrices that result from chromatogram data processing steps

226

Alexander Erban et al.

(see Subheading 3.11) to meet the prerequisites and enable multivariate testing. Class comparison by supervised statistical tests and procedures is meant to find mass features that significantly change according to a predefined set of experimentally tested treatments. Typical examples of class comparison procedures are applications of generalized linear models (GLMs), false discovery rate (FDR) correction, analysis of variance (ANOVA), Bonferroni correction for multiple testing, or the Tukey honest significant difference test. Class discovery methods are aiming to uncover hidden and unexpected patterns in data matrices. Class discovery algorithms are unsupervised and entail analyses of the general variance and trends. Examples are bootstrapped HCA, K-means clustering, PCA and mutual information or correlation matrices for network inference. Class prediction finds potential metabolic markers that represent predefined sample classes. Such classes can be expected according to the choice of experimental design or unexpected and result in class discovery. Examples cover machine learning technologies such as Random Forests, Support Vector Machines (SVM), or discriminant analyses (DA) such as orthogonal partial least squares-DA. 3.14.1

Preprocessing

1. Process data by Log-transformation. This step partially solves typically right-skewed distribution of metabolite abundances into an approximately normally distributed scale of increases and decreases retaining the variance structure [28]. 2. Scale data to null mean and unit standard deviation (Autoscaling): Several scaling and transformations procedures exist [29]; autoscaling performs better to infer the biological context, namely weighing small and large changes equally. By default, unsupervised class discovery methods will grant greater explanatory value to mass features with larger variance in an unscaled matrix (groups based on covariance). Alternatively, in a scaled matrix the grouping will be done according to correlation of mass features (e.g., PCA based on covariance or correlation matrices [30]). The latter represents better molecular systems in which a subtle change in a metabolite can be amplified by a signaling cascade [27]. Metabolite levels are monitored by protein sensors [31] that ultimately trigger responses at the onset of changed metabolite pools (e.g., metabolite level homeostasis) [32]. The cascade of responses and the magnitude of changed metabolite pools are not uniform over biological samples treated differently (e.g., “all-or-nothing” switches) [33]. Therefore, the standard deviation of mean

Multiplexed GC-MS based Metabolite Profiling

227

abundances differs between treated and control samples. These molecular mechanisms define the heteroscedastic nature of metabolite variance structures (i.e., each experimental treatment has a different variance). 3. Impute missing values depending on the scope of the experiment. Use k nearest neighbor method (KNN) [34, 35] to replace missing values by the average abundance of k nearest neighbors that do not have a missing value at that specific Xij cell. Nearest neighbors are defined by Euclidean distance (after scaling, Euclidean distance and correlation are equivalent and hence using one or the other will not alter the results). Similar methods include Bayesian principal component analysis (bPCA), a multiple correspondence analysis (MCA) model, a multiple factor analysis (MFA), and random forests [36]. Imputation of missing data in metabolite matrices has been shown to be biologically more accurate when using KNN [37] than the other mentioned models. Use small values imputing only when the aim is to find metabolic markers of the type presence/absence. Any other imputation of small values has been shown to be detrimental in the analysis of datasets [38]. Depending on the subsequently applied algorithm one may as well skip missing value imputing. 4. Deisotope (i.e., remove all isotopologs except the a0) the data set. This step becomes necessary to avoid groups forming in the following analyses only based on isotopolog correlation. It is best to also remove alternative fragments representing the same analyte manually, guided by statistics (e.g., correlation analysis) (see Subheading 3.11.4). 3.14.2

Class Comparison

1. Analyze the distribution of your data before selecting an appropriate statistical test. A Cullen and Frey graph plots skewness in the x-axis and kurtosis in the y-axis. Skewness measures the asymmetry of the probability distribution around the mean, similarly kurtosis defines the shape of the probability distribution at the tails of it. Moreover, known distributions should be test-plotted before using the data at hand in order to have a better estimate to what distribution your data may fit. The R-package “fitdistrplus” contains these and more features [39]. Due to the preprocessing, the abundances of mass features tend to follow a log-normal distribution. Evaluation of normality can be further done with the Kolmogorov–Smirnov test [40]. 2. Test the equality of variance assumptions (i.e., homoscedastic vs. heteroscedastic variance structures) using the Levene test [41].

228

Alexander Erban et al.

3. Select an appropriate statistical test. If the data is normally distributed and the means have equal variances, ANOVA may be applied. If the data is not normally distributed and the means have equal variances, the nonparametric versions of ANOVA may be applied (e.g., Kruskal–Wallis or Wilcoxon rank) [42]. If the data is normally distributed and the means do not have equal variances, a linear model will suffice to cope with the heteroscedastic nature of mass features. If the data is not normally distributed and the means do not have equal variances, a robust solution is a generalized linear mixed model (GLMM) [43], which parametrizes the mean and the variance using a link function for statistical comparison instead of the actual mean values [28]. 4. Correct for multiple statistical testing. P-values must be adjusted to avoid Type I and II errors. The FDR correction avoids the loss of potentially significant mass features [44]. 3.14.3

Class Discovery

1. Find relevant groups of behavior by clustering the scaled data matrix (in both X and Y axes). Bootstrap [45] the clusters in order to avoid isolated clustering conditions from a single solution and thereby confirm that these groups of mass features confer identity to the clustered biological samples. The resulting groups are visually accessible through a clustered heatmap of scaled abundances (e.g., Clustering algorithm: HCA or Kmeans) that has the mass features on the y-axis and the conditions in the x-axis. The unsupervised grouping map discovers which mass features determine that certain biological samples cluster together, suggesting that these features give identity to each set of clustered biological replicates. 2. After bootstrapping, confirm the separation between experimental conditions relative to the within-replicate variance using PCA. Furthermore, using the conditions as eigenvectors and the masses as projections into the plane, the previously found clusters of mass features can be traced in the biplot to rank their importance. 3. Create a distance matrix on selected masses that belong to significant clusters and subsequent networks in which each node is a metabolite with properties that define its influence in the data set. Consequently, the network approach makes it possible to determine highly influential metabolites acting within the significantly misrepresented clusters [46].

3.14.4

Class Prediction

1. Apply several algorithms that differ in their mathematical procedures to interpret the overlapping metabolic markers found. 2. Methods that provide inherent means of validation are preferred, such as Random Forest and SVM.

Multiplexed GC-MS based Metabolite Profiling

229

Random Forest [47, 48] iterates the classification procedure of the samples across a n-number of repetitions in order to robustly select mass features that are good metabolic markers, namely highly ranked in the classification procedure. By removing one mass feature at a time and reclassifying the dataset, the algorithm is able to assign a value of “importance for classification” to each mass feature. Finally, a fraction of the biological samples can be taken as training or evaluation sets in order to have a degree of external validation for the predictors. In SVMs [49], the training set is mapped into a hyperplane (projection), in which known categories of biological samples become separated by the greatest possible gap or margin between them. The samples that reside in the margin become the support vectors for the boundary between categories. SVMs accomplish high classification rates with a relatively small training set. 3. Other methods do not provide inherent means of validation; for example, orthogonal partial least squares-DA (OPLS-DA) allows for bifactorial class prediction only, while partial least squares-DA (PLS-DA) allows for multifactorial prediction [50]. The use of projection methods based on partial least squares is suitable as a control to confirm the best biological markers after acknowledging the risk of overfitting. Overfitting may be caused by the lack of internal validation procedures such as the splitting of datasets into training and evaluation sets. 3.15 Reconstitution and Matching of Mass Spectra for the Identification of Relevant Metabolites

The preceding discovery process and statistical evaluation provides information on relevant mass features and clusters of relevant mass features. These mass features need to be linked to the metabolite (s) that are represented by those features (Fig. 3). In GC-MS profiling studies this identification process is most efficiently started with a mass spectrum. 1. Reconstitute a mass spectrum from all mass features with overlapping retention index intervals, that is, a reconstituted mass spectrum of a time group. Match reconstituted time group spectra to mass spectral libraries using reverse matching (see Note 27). 2. Reconstitute a mass spectrum of all correlated mass features within a time group, that is, a reconstituted mass spectrum of a cluster of mass features. Match reconstituted cluster spectra to mass spectral libraries using forward matching (see Note 28). 3. Retrieve representative full mass spectra. Apply automated mass spectral deconvolution algorithms to extract full mass spectra from the GC-MS chromatograms of the experimental data set that contain the relevant mass features at highest available

230

Alexander Erban et al.

responses and the least amounts of co-eluting compounds, apply both matching directions (see Note 29). 4. Export and archive all extracted and reconstituted mass spectra in a mass spectral interchange format, such as the MSP format. 5. Perform provisional identification manually (see Notes 30 and 31), consider the best mass spectral match within a RI deviation window 1%, when using identical capillary column types [21]. 6. If no hit is obtained, the respective compound is an unknown metabolite that can be described by mass spectrum and retention index. To further characterize this unknown compound, repeat mass spectral matching without RI constraint and interpret partially matching mass spectra, the fragmentation pattern of the available mass spectra of this unknown compound and attempt to deduce a possible structure or presence of likely substructures [7]. 7. Archive mass spectrum and retention index of the unknown compound, for example by submission to a public database, such as the Golm Metabolome Database or https://www.ebi. ac.uk/metabolights as in [25]. 3.16 Verification of Provisional Metabolite Identifications

Mass spectral and retention index based identification using the respective library compendia needs to be verified by authenticated reference substances, especially for the differentiation of co-eluting isomers. 1. Obtain commercially or synthesize candidate metabolites and alternative isomers. 2. Perform a standard addition experiment. For this purpose, process in parallel, a backup of the biological sample that contains the provisionally identified metabolite, a sample that contains the obtained and authenticated reference compound and an equal mixture. Adjust the amount of the reference compound to be approximately threefold higher than in the biological samples. 3. Check metabolite identity by exact match of mass spectra and chromatographic retention using both GC-EI-MS and GCAPCI-MS technologies. 4. If multiple isomers of the metabolite exist, repeat the standard addition experiment using an alternative type of capillary column with changed polarity of the stationary phase. 5. Check metabolite identity again by exact match of mass spectra and chromatographic retention (see Note 31). 6. Archive the verification status of metabolite identifications or classifications (e.g., identification by standard addition of authenticated reference substance using one chromatographic

Multiplexed GC-MS based Metabolite Profiling

231

system or two systems, preliminary identification by GC-EIMS and GC-APCI-MS mass spectral and retention index matches, or preliminary classification by mass spectral match and exact molecular mass of the tentative molecular ion assignment of GC-APCI-MS analysis). 7. Use the standard addition experiment to obtain a recovery value of the metabolite for the analyzed biological sample type. 8. If identification fails, archive inferred classification, a structure hypothesis, if available, representative mass spectra and retention indices, as well as backup samples, if available, for later identification.

4

Notes 1. The plants can be precultured under optimized control conditions with ambient isotopic material. At the desired developmental stage, precultured plants are shifted to alternative environmental conditions (e.g., heat or cold). The liquid medium may be substituted by the chosen 13C- or 15N-labeled precursor. Environmental and isotopic shifts can be simultaneous or sequential. Harvests may follow during dynamic labeling or at approximate isotopic steady state. 2. Metabolism has a rapid turnover which is stopped by immediate freezing. Samples need to stay frozen until extraction. Gravimetric determination of sample fresh weight by differential weighing, while samples and containers are kept frozen, is influenced by hoarfrost that forms from humid air during the weighing process and by residual droplets of liquid nitrogen. These influences are minimized if the container is controlled for residual liquid nitrogen before weighing and if the air humidity is kept low during differential weighing. Because humidity may change in the course of a day, determine the empty weight of the precooled microvials directly before determining the differential weight with frozen sample. The tolerance of the scales must be 0.1 mg or better. The accuracy of the scales must be tested frequently by gauge weights (Fig. 2). 3. In the case that the single leaf is smaller than the 2.5 mg limit, pooling is required and single leaf analysis in a strict sense is impossible. Note that absolute quantification is delimited by the error of fresh weight determination. In the case that the leaf is heavier than 125 mg or in the case that a complete rosette is analyzed, prepare a frozen homogenate and take a representative aliquot from the frozen homogenate. Surplus frozen homogenate of the sample can be used for analyses of other system components.

232

Alexander Erban et al.

4. The duration and number of bursts required to obtain a fine powder may require adaptation to the physical properties of the plant material. Removal of the steel balls may lead to an uncontrolled loss of sample. Keep the time of storage to a minimum. Do not store at 80  C for longer than 2 months without testing for storage effects. 5. Surface contaminations of all employed vials, flasks, bottles, pipetting devices and steel balls can be avoided by careful washing with extraction solvents. Please consider that autoclaved material, while sterile, may, nevertheless, be chemically contaminated. 6. A recovery mixture of stable isotope labeled standards at concentrations that approximate the expected endogenous metabolite concentrations can be readily obtained by in vivo stable isotope labeling for example of Arabidopsis thaliana, Saccharomyces cerevisiae, or Synechocystis [51–53]. When performing standard addition experiments with nonlabeled endogenous metabolites to determine recovery factors aim for an approximately threefold increased standard addition compared to the endogenous metabolite concentration. 7. Organic solvents can be precooled to 20  C. Water can be kept at 4  C prior to extraction. Precooling is possible when using few internal standards at low concentrations. When using multiple internal standards or high concentrations of single standards, do not precool to avoid precipitation of added internal standards. 8. Leaves of some cold stressed and cold acclimated plants, such as Arabidopsis thaliana, accumulate soluble polar metabolites compared to nonstressed and nonacclimated leaves [8, 9, 25]. The sample fresh weight of cold treated Arabidopsis thaliana leaf samples can be reduced but the ratio of extract volume to sample fresh weight must be kept constant for all samples of an experimental set (see Note 10). 9. To increase metabolite recovery, extract twice with 400 μL bidistilled water. Combine the upper phases and dry a proportionally upscaled aliquot in a vacuum concentrator. Should the bidistilled water for the first extraction contain internal standards, use 400 μL bidistilled water without internal standards for the second extraction step. 10. The fresh weight of single leaves changes during development. To compare small samples (e.g., 2.5–10.0 mg fresh weight) with larger samples, the ratio of extract volume and fresh weight should be kept approximately constant. When using 2.0 mL microvials extract 10 mg with 400 μL solvent, 10–20 mg with 800 μL, 20–30 mg with 1200 μL, 30–40 mg with 1600 μL, and 40–50 mg with 2000 μL. Take a frozen

Multiplexed GC-MS based Metabolite Profiling

233

aliquot from a homogenate to handle larger samples than possible in 2.0 mL microvials. 11. The protocol describes a two-step reaction. A methoxyamine reagent transfers carbonyl moieties into two products, an Eand a Z-methoxyamine. The two products are in most cases separable by gas chromatography. Methoxyamination transforms reducing sugars into open chain products and eliminates acetal and ketal formation prior to trimethylsilylation. Trimethylsilylation substitutes hydrogen atoms that are bound to heteroatoms (e.g., hydroxyl, sulfhydryl, or amino moieties) and thereby transforms most polar primary metabolites into volatile analytes. Note that amino moieties can be nontrimethylsilylated or trimethylsilylated once or twice, depending on the steric hindrance of the structure. The ratio between two trimethylsilylation products of an amine (e.g., an amino acid) can be used to assess the performance of the trimethylsilylation reaction. Nonmethoxyaminated but trimethylsilylated hexoses can be used to monitor the efficiency of the methoxyamination reaction. 12. 4-(Dimethylamino)pyridine is added as catalyst in the presence of high amounts of reducing sugars. Do not change the sequence of dissolving steps. Dissolving methoxyamine hydrochloride first may cause precipitations when adding 4-(dimethylamino)pyridine. The use of 4-(dimethylamino) pyridine is optional. 13. Silylation reagents react with water, lose reactivity, and form polysiloxanes in the process, which contribute to the chemical background of GC-MS profiling. 14. Check that solid residue is completely dissolved. 15. Erban and coauthors [16] describe an automated and miniaturized protocol for inline chemical derivatization. This protocol allows for continuous processing of large sample numbers and is exactly timed between chemical derivatization and injection into the GC-MS system. 16. Perform injections of pure N,O-bis(trimethylsilyl)trifluoroacetamide reagent between samples in regular intervals to control for carryover effects and to counteract buildup of semivolatile chemical deposits within the injector system. Note that complex lipophilic and other nonvolatile compounds accumulate in the GC injection port. An increased frequency of GC maintenance may be required when omitting phase separation using methanol extraction without liquid partitioning into chloroform. Use deactivated glass insert liners [16]. Change liner before each new experimental set. During analysis of each experimental set, change liner depending on the buildup of nonvolatile chemical deposits within the injector system.

234

Alexander Erban et al.

Depending on the type of sample and applied extraction procedure liner may be changed every 12–60 injections. The concentration of chemically derivatized samples is adjusted to monitor a maximal number of peaks. The major metabolites are typically close to or at the upper detection limit when applying splitless injection. To also quantify these major metabolites perform a second analysis of the same samples using split injection. Use the split ratio to adjust the abundance of major metabolites to the linear range of quantification. 17. Store at room temperature. This can be kept as long as spectra are clean enough for calibration purposes. 18. The ionization method determines the complexity of the recorded mass spectrum. Electron impact ionization generates highly repeatable mass spectral fingerprints that can be easily compared between different types of mass spectrometers and between systems of different manufacturers. The sensitivity of profiling analyses declines with aging of mass detectors. Adjustments of detector voltage can be used within manufacturer defined limits. Note that increasing detector voltage may affect signal to noise and thereby the capability to monitor minor components in complex mixtures. Alternatives to electron impact ionization are (atmospheric pressure) chemical ionization approaches which potentially generate a smaller number of fragments or even more complex mass spectra. Such systems are used to analyze the molecular ions of analytes [54] (Fig. 3). 19. The exact type and version of mass spectrometer must be documented. Similar to the influence of the type of capillary column on gas chromatographic separation, the type and build of a mass spectrometer may influence the characteristics and details of recorded mass spectra. Mass spectrometers with higher mass precision can be used to deduce the molecular formula of analytes or of mass fragments from accurate mass recordings. Mass spectrometers set to higher scan rates can be used for fast GC-MS, but a loss of sensitivity and/or separation may occur. Mass spectrometers with lower scan rates should not be employed for fast GC-MS analysis, 10 scans per peak should be recorded. 20. The mentioned artifacts may frequently occur due to system aging. Most artifacts are caused by buildup of nonvolatile deposits. We do not mention rare systems deviations that may also compromise analyses. All quantitative analyses and profiling approaches that are based on chromatography systems hyphenated to mass spectrometric detection are subject to slow drifts in chromatography, mass calibration, and quantitative response. These aspects must be controlled using the

Multiplexed GC-MS based Metabolite Profiling

235

quality control samples mentioned above and by system maintenance. 21. The TagFinder software offers all required options for the processing of GC-MS chromatogram files [23]. The handling of the TagFinder software and recommended parameter settings have been described elsewhere [24]. Here we describe the essential steps of GC-MS data processing for metabolite profiling and quantification experiments. Similar analyses can also be performed by combination of other specialized software. Note that the initial discovery of metabolic markers requires nontargeted peak lists that contain all mass features (RI, m/z, response) observed in the chromatogram files comprising a profiling experiment. The verification of candidate marker metabolites can be performed by a predefined targeted list of selective and metabolite-specific mass features that can be either directly extracted from chromatogram files using vendor software or from the comprehensive peak lists. 22. Mass alignment according to equal nominal mass is a trivial process that is typically applied when using low mass resolution mass spectrometers. Nevertheless, this process needs to be described. With the availability of high-resolution mass spectrometers for GC-MS systems, alternative merging methods are required, for example, the creation of mass bins at decimal intervals that are in agreement with the mass resolution provided by the respective instrument. Minimal, average, and maximal mass and mass range should be documented when using such instruments (see Subheadings 3.11 and 3.12). 23. Chromatographic alignment can be achieved with and without use of internal standards. The alignment results and applied procedures must be reported and archived. Most chromatographic alignment procedures are confounded by random low-response peaks that are caused by electronic and chemical noise. These noise effects can be contained by applying a low-response threshold for analyzed mass features and by coprocessing of a limited number of chromatograms (e.g., 50–200 chromatogram files). However, more confounding are peak broadening effects and retention index shifts which are caused by changes of metabolite concentrations, especially by increases approaching chromatographic overloading. The effects can be contained by accepting wider retention index intervals for mass feature alignment at the cost of a possible loss of isomer resolution. Especially closely eluting minor isomers which provide no differentiating mass fragments may be lost in the presence of large amounts of the abundant isomer. Note that alternative means of chromatographic alignment, such as chromatographic alignment algorithms and software, may be

236

Alexander Erban et al.

applied but need to be tested thoroughly for the mentioned challenges that may confound alignment. 24. Grouping of mass features and “deisotoping” with the aim to remove redundancy can be achieved using diverse approaches. The TagFinder software uses the fact that coeluting redundant mass features are highly correlated in GC-MS analyses, because natural mass isotopomers of elements occur in fixed ratios and because electron impact ionization generates mass fragments in stable ratios. Luedemann and coauthors [23, 24] describe the generation of a correlation network using Pearson’s or Kendall’s correlation coefficients and a core finding algorithm to assign clusters. 25. Mass spectrometry based metabolite profiling experiments may contain large numbers of low-response mass features that are equally likely to be caused either by analytes or by electronic and chemical noise. The majority of irrelevant mass features can be removed before relative or absolute quantification by setting a minimal response threshold. If available through processing software, use signal to noise values of each recorded mass feature to threshold. 26. Note that multiple statistical approaches may lead to the discovery of relevant regulated mass features. Try to combine the results of several tests and focus on those mass features that have high relative changes of normalized responses and that are part of clusters with a high number of correlated mass features. When checking mass features by analysis of absent/present calls avoid mass features with responses that are close to the lower limit of detection. 27. Note that reconstituted mass spectra of time groups are inherently composite mass spectra of coeluting analytes. Therefore, reverse matching of library spectra of pure authenticated reference compounds is the most appropriate matching approach. 28. Note that reconstituted mass spectra of clusters of mass features are inherently partial mass spectra. Mass fragments that are common to coeluting analytes are eliminated because such mass fragments are not highly correlated. Also, mass fragments with low-responses and mass fragments with an overloaded response are eliminated. Therefore forward matching to library spectra of pure authenticated reference compounds is the most appropriate matching approach. 29. Note that automatically deconvoluted mass spectra can still contain systematic errors, such as additional mass fragments of exactly coeluting analytes, missing mass fragments due to false subtraction of fragments from neighboring analytes. If necessary, attempt manual extraction and correction of mass spectra.

Multiplexed GC-MS based Metabolite Profiling

237

30. Note that mass spectral and retention index matches alone represent a preliminary identification. Isomers are frequently present in biological samples (e.g., diverse saccharides). Such isomers can have almost identical mass spectra and highly similar chromatographic retention. 31. An elegant solution for this two column approach to compound identification is provided by two-dimensional GCxGC-MS systems. Note that some isomers cannot even be separated by changed column polarities. A nonseparable set of metabolites should be reported in the sense of summarized information (e.g., the sum of two saccharides or, if nonchiral columns are used, the sum of D- and L-stereoisomers). References 1. Fiehn O, Kopka J, Do¨rmann P et al (2000) Metabolite profiling for plant functional genomics. Nat Biotechnol 18:1157–1161 2. Roessner U, Wagner C, Kopka J et al (2000) Simultaneous analysis of metabolites in potato tuber by gas chromatography–mass spectrometry. Plant J 23:131–142 3. Allwood JW, Erban A, de Koning S et al (2009) Inter-laboratory reproducibility of fast gas chromatography–electron impact–time of flight mass spectrometry (GC–EI–TOF/MS) based plant metabolomics. Metabolomics 5:479–496 4. Wagner C, Sefkow M, Kopka J (2003) Construction and application of a mass spectral and retention time index database generated from plant GC/EI-TOF-MS metabolite profiles. Phytochemistry 62:887–900 5. Schauer N, Steinhauser D, Strelkov S et al (2005) GC–MS libraries for the rapid identification of metabolites in complex biological samples. FEBS Lett 579:1332–1337 6. Kopka J, Schauer N, Krueger S et al (2004) [email protected]: the Golm Metabolome Database. Bioinformatics 21:1635–1638 7. Hummel J, Strehmel N, Selbig J et al (2010) Decision tree supported substructure prediction of metabolites from GC-MS profiles. Metabolomics 6:322–333 8. Kaplan F, Kopka J, Haskell DW et al (2004) Exploring the temperature-stress metabolome of Arabidopsis. Plant Physiol 136:4159–4168 9. Kaplan F, Kopka J, Sung DY et al (2007) Transcript and metabolite profiling during cold acclimation of Arabidopsis reveals an intricate relationship of cold-regulated gene expression with modifications in metabolite content. Plant J 50:967–981

10. Guy C, Kaplan F, Kopka J et al (2008) Metabolomics of temperature stress. Physiol Plant 132:220–235 11. Korn M, G€artner T, Erban A et al (2010) Predicting Arabidopsis freezing tolerance and heterosis in freezing tolerance from metabolite composition. Mol Plant 3:224–235 12. Dunn WB, Erban A, Weber RJM et al (2013) Mass appeal: metabolite identification in mass spectrometry-focused untargeted metabolomics. Metabolomics 9:44–66 13. Sumner LW, Amberg A, Barrett D et al (2007) Proposed minimum reporting standards for chemical analysis. Chemical Analysis Working Group (CAWG) Metabolomics Standards Initiative (MSI). Metabolomics 3:211–221 14. Fernie AR, Aharoni A, Willmitzer L et al (2011) Recommendations for reporting metabolite data. Plant Cell 23:2477–2482 15. Lisec J, Schauer N, Kopka J et al (2006) Gas chromatography mass spectrometry–based metabolite profiling in plants. Nat Protoc 1:387–396 16. Erban A, Schauer N, Fernie AR et al (2007) Nonsupervised construction and application of mass spectral and retention time index libraries from time-of-flight gas chromatography-mass spectrometry metabolite profiles. In: Metabolomics: methods and protocols. Humana Press, Totowa, NJ, pp 19–38 17. Strehmel N, Hummel J, Erban A et al (2008) Retention index thresholds for compound matching in GC–MS metabolite profiling. J Chromatogr B 871:182–190 18. Murashige T, Skoog F (1962) A revised medium for rapid growth and bio assays with tobacco tissue cultures. Physiol Plant 15:473–497

238

Alexander Erban et al.

19. Zhong HH, Painter JE, Salome´ PA et al (1998) Imbibition, but not release from stratification, sets the circadian clock in Arabidopsis seedlings. Plant Cell 10:2005–2017 20. Boyes DC, Zayed AM, Ascenzi R et al (2001) Growth stage–based phenotypic analysis of Arabidopsis. A model for high throughput functional genomics in plants. Plant Cell 13:1499–1510 21. van den Dool H, Kratz P (1963) A generalization of the retention index system including linear temperature programmed gas—liquid partition chromatography. J Chromatogr A 11:463–471 22. Birkemeyer C, Kolasa A, Kopka J (2003) Comprehensive chemical derivatization for gas chromatography–mass spectrometry-based multi-targeted profiling of the major phytohormones. J Chromatogr A 993:89–102 23. Luedemann A, Strassburg K, Erban A et al (2008) TagFinder for the quantitative analysis of gas chromatography—mass spectrometry (GC-MS)-based metabolite profiling experiments. Bioinformatics 24:732–737 24. Luedemann A, von Malotky L, Erban A et al (2012) TagFinder: preprocessing software for the fingerprinting and the profiling of gas chromatography–mass spectrometry based metabolome analyses, in Plant metabolomics: methods and protocols. Humana Press, Totowa, NJ 25. Lawas LMF, Li X, Erban A et al (2019) Metabolic responses of rice cultivars with different tolerance to combined drought and heat stress under field conditions. GigaScience 8:giz050. https://doi.org/10.1093/gigascience/ giz050 26. Wolfender JL, Rudaz S, Choi YH et al (2014) Plant metabolomics: from holistic data to relevant biomarkers. Curr Med Chem 20:1056–1090 27. Zhang Q, Bhattacharya S, Andersen ME (2013) Ultrasensitive response motifs: basic amplifiers in molecular signalling networks. Open Biol 3:130031 28. de Abreu e Lima F, Leifels L, Nikoloski Z (2018) Regression-based modeling of complex plant traits based on metabolomics data. In: Plant metabolomics: methods and protocols. Springer, New York, NY, pp 321–327 29. van den Berg RA, Hoefsloot HCJ et al (2006) Centering, scaling, and transformations: improving the biological information content of metabolomics data. BMC Genomics 7:142 30. Borgognone MAG, Bussi J, Hough G (2001) Principal component analysis in sensory analysis: covariance or correlation matrix? Food Qual Prefer 12:323–326

31. Rogers JK, Guzman CD, Taylor ND et al (2015) Synthetic biosensors for precise gene control and real-time monitoring of metabolites. Nucleic Acids Res 43:7648–7660 32. Kacser H, Burns JA, Kacser H et al (1995) The control of flux: 21 years on. Biochem Soc Trans 23:341 33. Lopez-Fontal E, Milanesi L, Tomas S (2016) Multivalence cooperativity leading to “all-ornothing” assembly: the case of nucleationgrowth in supramolecular polymers. Chem Sci 7:4468–4475 34. Hastie T, Tibshirani R, Narasimhan B et al. (2018) Impute: imputation for microarray data. R package version 1.56.0. 35. Zhang S (2012) Nearest neighbor selection for iteratively kNN imputation. J Syst Softw 85:2541–2552 36. Josse J, Husson F (2016) missMDA: a package for handling missing values in multivariate data analysis. J Stat Softw 70:1–31 37. Do K, Wahl S, Raffler J et al (2018) Characterization of missing values in untargeted MS-based metabolomics data and evaluation of missing data handling strategies. Metabolomics 14:128–128 38. Wehrens R, Hageman JA, van Eeuwijk F et al (2016) Improved batch correction in untargeted MS-based metabolomics. Metabolomics 12:88 39. Delignette-Muller ML, Dutang C (2015) fitdistrplus: an R package for fitting distributions. J Stat Softw 64:1–34 40. Massey FJ (1951) The Kolmogorov-Smirnov test for goodness of fit. J Am Stat Assoc 46:68–78 41. Levene H (1960) Robust tests for equality of variances. In: Contributions to probability and statistics: essays in honor of Harold Hotelling. Stanford University Press, Stanford, CA, pp 278–292 42. Gooch JW (2011) Kruskal-Wallis test. In: Encyclopedic dictionary of polymers. Springer Science & Business Media, New York, NY, pp 984–985 43. Breslow NE (1995) Generalized linear models: checking assumptions and strengthening conclusions. In: Congresso Nazionale Societa’ Italiana di Biometria Centro Convegni S. Agostino, Cortona, Italy 44. Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B Methodol 57:289–300 45. Suzuki R, Shimodaira H (2006) Pvclust: an R package for assessing the uncertainty in

Multiplexed GC-MS based Metabolite Profiling hierarchical clustering. Bioinformatics 22 (12):1540–1542 46. Grapov D, Wanichthanarak K, Fiehn O (2015) MetaMapR: pathway independent metabolomic network analysis incorporating unknowns. Bioinformatics 31(16):2757–27604 47. Breiman L (2001) Random Forests. Mach Learn 45:5–32 48. Touw WG, Bayjanov JR et al (2012) Data mining in the life sciences with Random Forest: a walk in the park or lost in the jungle? Brief Bioinform 14:315–326 49. Cortes C, Vapnik V (1995) Support-vector networks. Mach Learn 20:273–297 50. Westerhuis JA, van Velzen EJJ, Hoefsloot HCJ et al (2010) Multivariate paired data analysis: multilevel PLSDA versus OPLSDA. Metabolomics 6:119–128

239

51. Huege J, Sulpice R, Gibon Y et al (2007) GC-EI-TOF-MS analysis of in vivo carbonpartitioning into soluble metabolite pools of higher plants by monitoring isotope dilution 13 after CO2 labelling. Phytochemistry 68:2258–2272 52. Strassburg K, Walther D, Takahashi H et al (2010) Dynamic transcriptional and metabolic responses in yeast adapting to temperature stress. OMICS 14:249–259 53. Huege J, Goetze J, Schwarz D et al (2011) Modulation of the major paths of carbon in photorespiratory mutants of Synechocystis. PLoS One 6:e16278 54. Strehmel N, Kopka J, Scheel D et al (2014) Annotating unknown components from GC/EI-MS-based metabolite profiling experiments using GC/APCI(+)-QTOFMS. Metabolomics 10:324–336

Chapter 16 Determining the ROS and the Antioxidant Status of Leaves During Cold Acclimation Andras Bittner, Thomas Griebel, Jo¨rn van Buer, Ilona Juszczak-Debosz, and Margarete Baier Abstract Cold slows down Calvin cycle activity stronger than photosynthetic electron transport, which supports production of reactive oxygen species (ROS). Even under extreme temperature conditions, most ROS are detoxified by the combined action of low-molecular weight antioxidants and antioxidant enzymes. Subsequent regeneration of the low-molecular weight antioxidants by NAD(P)H and thioredoxin/thiol–dependent pathways relaxes the electron pressure in the photosynthetic electron transport chain. In general, the chloroplast antioxidant system protects plants from severe damage of enzymes, metabolites, and cellular structures by both ROS detoxification and antioxidant recycling. Various methods have been developed to quantify ROS and antioxidant levels in photosynthetic tissues. Here, we summarize a series of exceptionally fast and easily applicable methods that show local ROS accumulation and provide information on the overall availability of reducing sugars, mainly ascorbate, and of thiols. Key words ROS, H2O2, Superoxide, Thiols, Reducing sugars, Ascorbate, Digital imaging, Quantification

1

Introduction Exposure to unfavorable conditions, such as low temperatures, can impair photosynthetic electron transport and cellular metabolism [1, 2] and lead to enhanced production of reactive oxygen species (ROS), such as superoxide radical anions (O2
), hydrogen peroxide (H2O2) and hydroxyl radicals (OH ), as well as to oxidation of antioxidants [3, 4]. ROS are important signaling molecules but also have a high cell damaging potential [5–8]. Low-molecular weight antioxidants and antioxidant enzymes protect proteins, metabolites, and structures against (uncontrolled) oxidation, epoxidation, and peroxidation [9–12]. The most prominent low-molecular weight antioxidants are ascorbate and glutathione, which accumulate in millimolar concentrations in plant cells [3, 13]. Ascorbate is synthesized mainly from l

l

Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_16, © Springer Science+Business Media, LLC, part of Springer Nature 2020

241

242

Andras Bittner et al. D-mannose and L-galactose [14, 15] and glutathione is formed enzymatically from γ-glutamyl cysteine (γEC) and alanine or glycine [16]. The concentration of both antioxidants can strongly increase in response to stress [17–19]. In the cold, the protection potential is supported by accumulation of osmolytes (e.g., sorbitol and proline) [20–22]. Whereas animal cells mainly detoxify ROS at the expense of glutathione [23], plants evolved highly efficient ascorbate peroxidases to detoxify H2O2 [24, 25]. In plants, glutathione serves mainly as a thiol buffer and for regeneration of ascorbate by dehydroascorbate reductases [26]. Chloroplasts, which are the main site of ROS production in the cold [1, 27], import all their antioxidant enzymes posttranslationally [25]. Superoxide dismutases catalyze oxidation and reduction of superoxide anions at the thylakoid membrane and in the stroma and generate O2 and H2O2 [28]. H2O2 is the most stable type of ROS [29]. Ascorbate peroxidases and peroxiredoxins (including glutathione peroxidases [30]) reduce H2O2 to H2O at the expense of ascorbate or small thiol proteins, respectively [31–33]. Monodehydroascorbate reductase, dehydroascorbate reductase, ferredoxinthioredoxin reductase, and NADPH-thioredoxin reductase regenerate the reductants [26, 34, 35]. The chloroplast antioxidant system widely depends on photosynthetic electron transport and attenuates the electron pressure in the photosynthetic electron transport chain with little effect on ATP biosynthesis [34, 36, 37]. Upon prolonged cold, reorganization of metabolisms, especially accumulation of sugars, promotes biosynthesis of ascorbate [38]. Whereas most genes for chloroplast antioxidant enzymes respond negatively to sugar and ascorbate availability at 20  C [39–41], their transcripts accumulate in prolonged cold [10, 20]. This is not the case for thylakoid ascorbate peroxidase and for CuZn-superoxide dismutase 2, which function close to the thylakoid membrane. The corresponding genes are downregulated in the cold [10]. As soon as the temperature increases, the low-molecular weight antioxidants are metabolized within a few hours [10, 20]. Insufficient antioxidant protection can then lead to elevated ROS levels for several days [10]. Various in situ and in vitro methods have been developed to detect ROS and the antioxidant status in plants. For in situ ROS pattern analysis, preference has been given to histochemical staining [19, 42] and, for quantification, to ROS trapping by fluorophores [43–46]. Here, we present cheap, fast and easy methods for analysis of ROS accumulation in leaves by histochemical ROS staining. Special preparation and illumination of stained plant material enables visualization of ROS production sites not only in the external cell layer, but also in the inner leaf tissues. The method enables monitoring of ROS accumulation in leaves and has been optimized with respect to providing quantitative information. Detection of O2
by NBT (nitro blue tetrazolium) staining and l

ROS and Antioxidant Levels

243

the analysis of H2O2 accumulation by DAB (3,30 -diaminobenzidine) staining follow the same basic setup with minimal variation. Ellman’s reagent (DTNB; 5,50 -dithiobis-2-nitrobenzoic acid) and Tillmann’s reagent (DCPIP; 2,6-dichlorophenolindophenol) can be used to determine quantitative measures for the antioxidant potential of the tissue. Ellman’s reagent reacts selectively with thiols. Tillmann’s reagent reacts with various reductants. Reduction of Tillmann’s reagent is frequently used to estimate the ascorbate level, since ascorbate is the most abundant nonthiol low-molecular weight antioxidant in plant cells [47].

2

Materials All solutions are prepared using distilled water and analytical grade reagents (see Note 1). All reagents and solutions are stored at 4  C (unless indicated otherwise). For disposal of waste material carefully follow your waste disposal regulations (remember that sodium azide is acutely toxic).

2.1

ROS staining

1. 10 phosphate buffered saline (PBS) stock solution: Dissolve 80 g of NaCl, 2 g of KCl, 14.4 g of Na2HPO4, and 2.4 g of KH2PO4 in 800 mL of distilled water and adjust the pH to 7.4. Adjust the volume with distilled water to 1 L. Sterilize by autoclaving and store the solution at 4  C. 2. 1 PBS: To obtain 1 L of 1 PBS, dilute 100 mL of 10 PBS with 900 mL of sterile water. Store at 4  C. 3. 10 mM sodium azide: Dissolve 0.65 g of NaN3 in 950 mL of 1 PBS. Adjust the volume with PBS to 1 L. Caution! Sodium azide is toxic! 4. 1 mg/mL NBT staining solution: Dissolve 1 g of nitro blue tetrazolium chloride (NBT) in 1 L of 1 PBS (see item 1). Store in darkness at 4  C. 5. 1 mg/mL DAB staining solution: Dissolve 1 g of 3,30 -diaminobenzidine (DAB) in 1 L of 1 PBS (see item 1). The solution is stable for up to 1 week in darkness at 4  C. 6. Destaining solution: Mix 100 mL of 100% acetic acid, 100 mL of glycerol, and 300 mL of 96% ethanol. Right before starting the destaining procedure, heat the mixture up to 60–80  C. 7. Sample containers (bottles/beakers): Prepare an appropriate number of sample containers (each sample should be placed in a separate sample container) and fill them with 10 mM NaN3 (for NBT staining) or DAB staining solution shortly before harvesting the plant material.

244

Andras Bittner et al.

8. Desiccators and vacuum pump: The number of desiccators required depends on the number of samples. The same desiccators can be used for the NBT and DAB stainings. 9. Photos: Use either a good camera with a high resolution or a light microscope with an integrated camera. If not using a light microscope, an additional light source might be helpful, which illuminates your samples from the bottom (e.g., White-light plate, INTAS, Germany). For quantification, it is essential that the light is evenly applied and the object is not shaded or shading. 10. Quantification: Download (https://imagej.nih.gov/ij/down load.html) and install on your computer the freeware package ImageJ [48]. For quantification follow the details listed in Subheading 3.1. 2.2 Thiol Determination

1. Extraction buffer: Prepare 100 mL 0.1 M HCl, add EDTA (pH 8.0) to a final concentration of 1 mM and mix the solution, cool the solution on ice prior to use. 2. Assay buffer: Dissolve 120 g NaH2PO4 in 700 mL H2O, adjust the pH to 7.8 with NaOH. Add 12 mL 500 mM EDTA (pH 8.0) to the solution and adjust the volume to 1 L with H2O. Store the buffer at
20  C in 20 mL aliquots until use. 3. DTNB (5,50 -dithiobis-2-nitrobenzoic acid) solution: Dissolve DTNB to a final concentration of 10 mM in 50 mM sodium acetate (pH 5.8). Store 1–2 mL aliquots at
20  C. They can be stored for up to 3 months. DTNB is light sensitive. Protect the solution from direct light exposure. 4. Thiol standards: Dissolve cysteine to a final concentration of 5 mM in 10 mL extraction buffer (stock solution). Store the solution for a maximum of 5 days at
20  C.

2.3 Determination of Ascorbate

1. Extraction buffer: Dissolve 5 g metaphosphoric acid in 90 mL H2O. Add H2O to a final volume of 100 mL. 2. DCPIP indicator solution: Dissolve 1 mg DCPIP in 10 mL 0.2 M sodium citrate (pH 6) (stock solution). Dilute the DCPIP stock solution 1:6 in 0.2 M sodium citrate (pH 6). 3. Ascorbate standard: Dissolve 10 mg sodium ascorbate in 1 mL extraction buffer. Dilute 1:10 in extraction buffer prior to use.

3 3.1

Methods ROS Staining

NBT stains tissue areas blue/purple if O2
ions accumulate and DAB forms an orange-brownish dye with H2O2 (Fig. 1). All staining reactions are light-sensitive. Therefore, the whole staining procedure should be carried out in darkness or at very low light l

ROS and Antioxidant Levels

245

Fig. 1 Arabidopsis thaliana leaves (middle-age and young) stained with NBT (a, b) and DAB (c, d). The samples were harvested before (a, c) and after (b, d) 14 days of cold treatment

intensities. Samples subjected to staining (leaves, seedlings, etc.) should be treated with care, since injuries promote ROS formation and lead to false-positive results. For comparisons, plant material should be harvested in parallel and treated for the same time in the staining solution. 3.1.1 NBT Staining for O2
Detection l

1. Place samples (e.g., detached leaves, whole seedlings, or rosettes) immediately after harvesting in 10 mM NaN3, which inhibits superoxide dismutases and heme-type peroxidases at their active sites. 2. Put the sample containers into the desiccator and apply a mild vacuum 3 times for 30 s (see Note 2) to remove air from the intercellular spaces and to support flooding of the tissue with the azide solution. During this procedure, the azide solution should cover the plant material. Finalize the step by incubating the plant material for 5 min under vacuum. Afterwards, slowly release the vacuum. Take the sample containers out of the desiccator and remove the NaN3 solution carefully. 3. Add an appropriate amount of NBT staining solution to cover the plant material and shake the samples softly for 5 min. Afterwards, place the samples back into the desiccator and apply a vacuum (see Note 3) to remove the oxygen in the desiccator and to protect the NBT from autooxidation. 4. Keep the samples for 6–24 h under vacuum (see Note 4). Then, release the vacuum and subject the samples to the destaining procedure (see Subheading 3.1.3).

246

Andras Bittner et al.

3.1.2 DAB Staining for H2O2 Detection

1. Immediately after harvest, place the samples in containers filled with DAB staining solution. The solution should fully cover the plant material. Place the containers with the samples in a desiccator and apply a mild vacuum 5 times for 30 s to remove the air from the tissue and to flood the intercellular spaces (see Note 3). 2. Incubate the samples in the DAB-solution in darkness. The staining intensity increases with time. Stop the reaction when the contrast between stained and unstained areas is optimal. 3. For clearing of the background from chlorophylls and other pigments, continue with the procedure described in Subheading 3.1.3.

3.1.3 Destaining Procedure

1. Discard the staining solution.

3.1.4 Photographic Documentation of Stained Samples

1. Take the samples out of the destaining solution and place them shortly on a paper towel to remove traces of destaining solution. Alternatively wash the sample once or twice with distilled water.

2. Fill the sample containers (containing the samples) with hot destaining solution, which will remove chlorophylls, carotenoids, and other plant pigments, but will stabilize the dyes. Incubate the samples in the destaining solution until the chlorophyll is completely removed (see Note 5). Nonstained tissue (parts) should be yellowish white. Due to the high glycerol content of the destaining solution, the samples get soft and luminescent. Caution! Acetic acid has a pungent smell (especially when hot). Perform destaining under a fume hood!

2. Place the samples under the microscope or on the external light device (see Note 6). 3. Illuminate the samples with white light from the bottom and take photos (see Note 7). 4. Use your favorite software (e.g., Adobe Photoshop or the freeware GIMP) for digital processing (e.g., background subtraction, contrast improvement, color correction). For taking photos from NBT-stained plants, decrease the yellow background to its minimum. For DAB-stained samples, remove the blue colors to minimize background effects (see Note 8). The photos should be of similar quality after processing to those shown in Fig. 1. 3.1.5 Quantification of Staining Results

1. Open the digital image, which you want to analyze, with ImageJ. 2. Convert the color scale of the photo to grayscale. To do that, choose “Image” from the ImageJ main menu and from the

ROS and Antioxidant Levels

247

Fig. 2 ImageJ main menu (a), “Threshold” window (b) and “Results” window (c) with the most important options marked with red squares

Fig. 3 Quantification workflow: The original picture is converted to grayscale. All areas with stronger staining than the threshold and the total leaf area are determine with the ImageJ area analysis tool

submenus “Type” and “8-bit.” Store the gray-scaled image as a new file. 3. To segment the image into features of interest (area stained stronger than a defined threshold) and background, adjust the threshold for the intensity. Use “Image,” “Adjust,” and “Threshold” in the ImageJ main menu (Fig. 2). In the “Threshold” window set the lower and higher threshold to a level suitable for your corresponding image by moving the slider. Make sure that only the stained area is marked in red (Fig. 3). 4. Choose “Measure” from the ImageJ main menu “Analyze.”

248

Andras Bittner et al.

5. Reopen the picture, repeat until step 4. This time, make sure that the whole leaf area is stained red (Fig. 3). 6. Choose “Measure” from the ImageJ main menu “Analyze.” 7. Calculate the % stained area by dividing the value “Mean” of the stained area by the whole leaf area and repeat for all leaves. 3.2 Determination of the Thiol Status

3.2.1 Extraction and Quantitative Analysis of Thiols

Antioxidants quickly oxidize in contact with air or during storage of the plant material. Therefore, all solutions should be prepared in advance and stored without shaking for a few days or flushed with N2 prior to use to minimize the oxygen content. The plant material should be freshly and quickly extracted in acidic solutions. 1. Quickly homogenize 20–50 mg plant material in 500 μL extraction buffer in a 1.5 mL reaction tube using a micropestle and a tiny bit of quartz sand. 2. Sediment the insoluble material by 3 min centrifugation at 13,000  g. Determine the thiol content immediately in the supernatant. 3. Place 900 μL assay buffer in a 1 mL cuvette and put the cuvette into a photometer. 4. Auto zero the photometer. 5. Add 50 μL DTNB solution and quickly stir the sample (avoiding air bubbles). 6. Add 50 μL sample and quickly stir the sample (avoiding air bubbles). 7. Read the absorption As at 412 nm after 10–20 s. 8. Reaction blank: Mix 900 μL assay buffer and 50 μL extraction buffer in a 1 mL cuvette with 50 μL DTNB solution and read the absorption Arb at 412 nm (see Note 9). 9. Sample blank: Mix 900 μL assay buffer and 50 μL 50 mM sodium acetate in a 1 mL cuvette with 50 μL sample and read the absorption Asb at 412 nm (see Note 10). 10. Calculate the thiol content in the extract using the extinction coefficient of DTNB (13,600 M
1 cm
1) or based on the calibration curve (Fig. 4) (see Note 11). μL rb
A sb 11. Thiol content in the extract As
A  1000 50 μL mM. 13:6

12. Standardize the thiol content on the fresh weight (f.w.) of the sample. Thiol content in micromole per gram fresh weight is calculated using the following formula: Thiol content in the extract ½in mM 

extract volume ½in μL : f :w: ½in mg

ROS and Antioxidant Levels

249

Fig. 4 Determination of the thiol content using DTNB. (a) DTNB forms mixed disulfides with thiols. (b) Examples for calibration curves with reduced glutathione and cysteine. (c) Examples for recovery values determined in samples which were incubated for 15 min at 20  C (slow) and in samples analysed immediately (fast)

3.2.2 Recovery Test and Calculation of the True Thiol Content (See Note 12)

1. Determine the absorption Ac with 5 μL of 5 mM cysteine in 900 μL assay buffer mixed with 45 μL extraction buffer and 50 μL DTNB solution at 412 nm. 5 μL The theoretical absorption is 5 mM  995 μL 
1 13:6 mM ¼ 0:341. Calculate the correction coefficient K: K ¼ Ac/0.341. 2. Quickly homogenize 20–50 mg plant material in 500 μL extraction buffer in a 1.5 mL reaction tube using a micropestle and a tiny bit of quartz sand. 3. Transfer 2 200 μL of extract into fresh reaction tubes. 4. Add 10 μL 5 mM cysteine standard to one tube and 10 μL extraction buffer to the other tube. 5. Sediment the insoluble material in both tubes by 3 min centrifugation at 13,000  g. 6. Determine the thiol contents in both tubes immediately by analyzing the absorptions with 50 μL sample in 900 μL assay buffer supplemented with 50 μL DTNB at 412 nm. Calculate the difference (ΔA) between the two absorptions. Theoretically the difference is

250

Andras Bittner et al.

50 μL 10 μL  5 mM  13, 600 M
1 cm
1  1 cm  K 1000 μL 210 μL ¼ 0:162  K : The recoveryR [in %] is [ΔA/(0.162∗K)]  100% (see Note 13) (Fig. 4). 7. Calculate the recovery-corrected thiol content in the sample. rb
A sb  The true thiol content in the extract is As
A 13:6 1000 μL 100% 50 μL  R ½in% mM. The true thiol content per gram fresh weight is calculated from the uncorrected thiol content per g f.w. using the following formula: Thiol content in μmol per g f :w:  R100% ½in%. 3.3 Determination of Reduced Ascorbate

Reduced ascorbate can be stabilized against oxidation in metaphosphoric acid. Transfer of the sample into the DCPIP solution increases the pH and enables reaction of ascorbate with DCPIP. DCPIP can be reduced by various reducing agents. Since ascorbate is the most abundant and one of the most reducing sugars accumulating in plants, the assay indicates the availability of reduced ascorbate.

3.3.1 Extraction and Quantitative Analysis of Reduced Ascorbate

1. Quickly homogenize approximately 50 mg plant material in 500 μL extraction buffer in a 1.5 mL reaction tube using a micro-pestle and a tiny bit of quartz sand. 2. Sediment the insoluble material by 3 min centrifugation at 13,000  g. Determine the ascorbate content in the supernatant immediately. 3. Autozero the photometer at 600 nm with 0.2 M sodium citrate (pH 6). 4. Give 990 μL DCPIP-solution into a half-micro cuvette and read the absorption ADCPIP at 600 nm. 5. Add 10 μL sample into the cuvette and stir quickly with a small swizzle stick. 6. Read the absorption As at 600 nm. 7. Calculate the content of reducing metabolites in the sample based on the extinction coefficient of DCPIP (16,900 M
1 cm
1) or a calibration curve (Fig. 5) (see Note 11). 8. The content of reducing metabolites in the extract is μL ADCPIP
AS  1000 10 μL mM. 16:9 9. Standardize the result on the fresh weight (see Subheading 3.2.2).

ROS and Antioxidant Levels

251

Fig. 5 Quantification workflow of the ascorbate determination procedure. (a) Sample preparation. (b) Calibration curve for ascorbate 3.3.2 Recovery Test and Correction of the Ascorbate Content (See Note 12)

1. Divide 100 mg plant material in two 50 mg portions. 2. Extract one portion in 500 μL extraction buffer and the other in 250 μL extraction buffer + 250 μL ascorbate standard solution. 3. Determine the absorptions of 10 μL of both samples and of 10 μL of ½ concentrated ascorbate standard solution. Calculate the ascorbate contents of the sample standard mix (AscSample +Standard), the sample (AscSample), and the standard solution (AscStandard). 4. Calculate the ascorbate recovery [in %]: 100%.

AscSampleþStandard
AscSample AscStandard



5. Correct the determined ascorbate contents with the recovery factor.

4

Notes 1. Use of freshly prepared staining solutions improves the quality of staining. PBS can be prepared in advance and stored at 4  C. 2. To enable the comparison between samples (e.g., control and stress treatment), put all of them into the same desiccator and incubate them for the same time. It ensures that the same strength of vacuum is applied to all of them.

252

Andras Bittner et al.

3. To improve penetration of the staining solution into your sample, apply and release the vacuum at least 3 times (each for 5 min) prior to the incubation. 4. Longer staining periods usually give better results, but can also result in overstaining. Therefore, carefully optimize the duration of staining. 5. To speed up destaining, the incubation can be performed in a dry-block incubator (approx. 15–20 min at 60  C). It is recommended to change the destaining solution at least twice for optimal destaining with young tissues and 3 times with older leaves, which are more difficult to destain. 6. To avoid drying of the samples and to flatten the leaves for photographic documentation, place them in a thin water film between two layers of overhead transparencies. 7. Avoid any light from the side. If possible, use a soft ring light for taking the photos. 8. Take care that all digital modifications are identical for all samples treated in parallel. If samples should be compared which are harvested on different days, prepare extra samples for comparison of staining intensities and destaining efficiency and maintain half of them in the dark in the staining solution and half in the destaining solution until the next set of plants is prepared. 9. The reaction blank provides a measure for autooxidation of the dye. 10. The sample blank records staining of the sample which is not caused by the dye reaction (e.g., the yellow background color of flavonoids and flavoproteins). 11. Always control your measurements by analysing at least one sample containing a defined concentration (standard). 12. Part of the antioxidants could be metabolized during the extraction procedure. In the recovery test, a defined amount of the antioxidant is added to quantify which percentage of the metabolite could be detected after extraction. 13. The recovery should be at least 90%. Usually it increases with optimization of the handling. Figure 4 shows the effect of 15 min unnecessary incubation on the recovery and, consequently, on the data quality. References 1. Ensminger I, Busch F, Huner NPA (2006) Photostasis and cold acclimation: sensing low temperature through photosynthesis. Physiol Plant 126:28–44

2. Hurry VM, Huner NPA (1991) Low growth temperature effects a differential inhibition of photosynthesis in spring and winter-wheat. Plant Physiol 96:491–497

ROS and Antioxidant Levels 3. Foyer CH, Lelandais M, Kunert KJ (1994) Photooxidative stress in plants. Physiol Plant 92:696–717 4. Mehler AH (1951) Studies on reactions of illuminated chloroplasts. 1. Mechanism of the reduction of oxygen and other Hill reagents. Arch Biochem Biophys 33:65–77 5. Foyer CH, Noctor G (2016) Stress-triggered redox signalling: what’s in pROSpect? Plant Cell Environ 39:951–964 6. Mittler R, Vanderauwera S, Gollery M et al (2004) Reactive oxygen gene network of plants. Trends Plant Sci 9:490–498 7. Apel K, Hirt H (2004) Reactive oxygen species: metabolism, oxidative stress, and signal transduction. Annu Rev Plant Biol 55:373–399 8. Queval G, Issakidis-Bourguet E, Hoeberichts FA et al (2007) Conditional oxidative stress responses in the Arabidopsis photorespiratory mutant cat2 demonstrate that redox state is a key modulator of daylength-dependent gene expression, and define photoperiod as a crucial factor in the regulation of H2O2-induced cell death. Plant J 52:640–657 9. Huner NPA, Oquist G, Hurry VM et al (1993) Photosynthesis, photoinhibition and low-temperature acclimation in cold tolerant plants. Photosynth Res 37:19–39 10. Juszczak I, Cvetkovic J, Zuther E et al (2016) Natural variation of cold deacclimation correlates with variation of cold-acclimation of the plastid antioxidant system in Arabidopsis thaliana accessions. Front Plant Sci 7:305 11. Kocsy G, von Ballmoos P, Ruegsegger A et al (2001) Increasing the glutathione content in a chilling-sensitive maize genotype using safeners increased protection against chillinginduced injury. Plant Physiol 127:1147–1156 12. Davey MW, Bauw G, van Montagu M (1996) Analysis of ascorbate in plant tissues by highperformance capillary zone electrophoresis. Anal Biochem 239:8–19 13. Foyer CH, Lelandais M, Edwards EA et al (1991) The role of ascorbate in plants, interactions with photosynthesis, and regulatory significance. In: Pell EJ, Steffen KL (eds) Active oxygen/oxidative stress and plant metabolism. ASPP, Rockville, MD, pp 131–144 14. Ishikawa T, Shigeoka S (2008) Recent advances in ascorbate biosynthesis and the physiological significance of ascorbate peroxidase in photosynthesizing organisms. Biosci Biotechnol Biochem 72:1143–1154 15. Smirnoff N, Wheeler GL (2000) Ascorbic acid in plants: biosynthesis and function. Crit Rev Biochem Mol Biol 35:291–314

253

16. Noctor G, Gomez L, Vanacker H et al (2002) Interactions between biosynthesis, compartmentation and transport in the control of glutathione homeostasis and signalling. J Exp Bot 53:1283–1304 17. Noctor G, Arisi ACM, Jouanin L et al (1998) Glutathione: biosynthesis, metabolism and relationship to stress tolerance explored in transformed plants. J Exp Bot 49:623–647 18. Smirnoff N, Pallanca JE (1996) Ascorbate metabolism in relation to oxidative stress. Biochem Soc Trans 24:472–478 19. Vanacker H, Carver TL, Foyer CH (2000) Early H2O2 accumulation in mesophyll cells leads to induction of glutathione during the hyper-sensitive response in the barley-powdery mildew interaction. Plant Physiol 123:1289–1300 20. Zuther E, Juszczak I, Lee YP et al (2015) Time-dependent deacclimation after cold acclimation in Arabidopsis thaliana accessions. Sci Rep 5:12199 21. Thomashow MF (1999) Plant cold acclimation: freezing tolerance genes and regulatory mechanisms. Annu Rev Plant Physiol Plant Mol Biol 50:571–599 22. Heidarvand L, Amiri RM (2010) What happens in plant molecular responses to cold stress? Acta Physiol Plant 32:419–431 23. Flohe´ L, Gu¨nzler WA (1984) Assays of glutathione peroxidase. Methods Enzymol 104:114–121 24. Asada K (1994) Molecular properties of ascorbate peroxidase - a hydrogen peroxidescavenging enzyme in plants. In: Asada K, Yoshikawa T (eds) Frontiers of reactive oxygen species in biology and medicine. Elsevier Science B. V, Amsterdam, pp 103–106 25. Pitsch NT, Witsch B, Baier M (2010) Comparison of the chloroplast peroxidase system in the chlorophyte Chlamydomonas reinhardtii, the bryophyte Physcomitrella patens, the lycophyte Selaginella moellendorffii and the seed plant Arabidopsis thaliana. BMC Plant Biol 10:133 26. Foyer CH, Halliwell B (1976) The presence of glutathione and glutathione reductase in chloroplasts: a proposed role in ascobic acid metabolism. Planta 133:21–25 27. Huner NPA, Bode R, Dahal K et al (2013) Shedding some light on cold acclimation, cold adaptation, and phenotypic plasticity. BotanyBotanique 91:127–136 28. Kliebenstein DJ, Monde RA, Last RL (1998) Superoxide dismutase in Arabidopsis: an eclectic enzyme family with disparate regulation and protein localization. Plant Physiol 118:637–650

254

Andras Bittner et al.

29. Elstner EF (1990) Der Sauerstoff: Biochemie, Biologie, Medizin. BI-Wiss.-Verl, Mannheim, Wien, Zu¨rich 30. Rouhier N (2010) Plant glutaredoxins: pivotal players in redox biology and iron-sulphur centre assembly. New Phytol 186:365–372 31. Asada K (1999) The water-water cycle in chloroplasts: scavenging of active oxygen and dissipation of excess photons. Annu Rev Plant Physiol Plant Mol Biol 50:601–639 32. Ko¨nig J, Baier M, Horling F et al (2002) The plant-specific function of 2-Cys peroxiredoxinmediated detoxification of peroxides in the redox-hierarchy of photosynthetic electron flux. Proc Natl Acad Sci U S A 99:5738–5743 33. Navrot N, Collin V, Gualberto J et al (2006) Plant glutathione peroxidases are functional peroxiredoxins distributed in several subcellular compartments and regulated during biotic and abiotic stresses. Plant Physiol 142:1364–1379 34. Asada K (2000) The water-water cycle as alternative photon and electron sinks. Philos Trans R Soc Lond Ser B Biol Sci 355:1419–1431 35. Spinola MC, Perez-Ruiz JM, Pulido P et al (2008) NTRC new ways of using NADPH in the chloroplast. Physiol Plant 133:516–524 36. Anderson JW, Foyer CH, Walker DA (1983) Light dependent reduction of hydrogen peroxide by intact spinach chloroplasts. Biochim Biophys Acta 724:69–74 37. Foyer CH, Noctor G (2012) Managing the cellular redox hub in photosynthetic organisms. Plant Cell Environ 35:199–201 38. Beck EH, Fettig S, Knake C et al (2007) Specific and unspecific responses of plants to cold and drought stress. J Biosci 32:501–510 39. Horling F, Lamkemeyer P, Konig J et al (2003) Divergent light-, ascorbate-, and oxidative stress-dependent regulation of expression of

the peroxiredoxin gene family in Arabidopsis. Plant Physiol 131:317–325 40. Shaikali J, Baier M (2010) Ascorbate regulation of 2-Cys peroxiredoxin-a promoter activity is light-dependent. J Plant Physiol 167:461–467 41. Heiber I, Cai W, Baier M (2014) Linking chloroplast antioxidant defense to carbohydrate availability: the transcript abundance of stromal ascorbate peroxidase is sugar-controlled via ascorbate biosynthesis. Mol Plant 7:58–70 42. Kawai S, Takeshita S, Okazaki M et al (1994) Cloning and characterization of OSF-3, a new member of the MER5 family, expressed in mouse osteoblastic cells. J Biochem 115:641–643 43. Choi WG, Swanson SJ, Gilroy S (2012) Highresolution imaging of Ca2+, redox status, ROS and pH using GFP biosensors. Plant J 70:118–128 44. Exposito-Rodriguez M, Laissue PP, Littlejohn GR et al (2013) The use of HyPer to examine spatial and temporal changes in H2O2 in high light-exposed plants. Methods Enzymol 527:185–201 45. Deshwal S, Antonucci S, Kaludercic N et al (2018) Measurement of mitochondrial ROS formation. Mitochondrial bioenergetics: methods and protocols. Methods Mol Biol 1782:403–418 46. Hideg E, Schreiber U (2007) Parallel assessment of ROS formation and photosynthesis in leaves by fluorescence imaging. Photosynth Res 92:103–108 47. van der Jagt DJ, Garry PJ, Hunt WC (1986) Ascorbate in plasma as measured by liquid chromatography and by dichlorophenolindophenol colorimetry. Clin Chem 32:1004–1006 48. Sheffield JB (2007) ImageJ, a useful tool for biological image processing and analysis. Microsc Microanal 13:200–201

Chapter 17 Analysis of Changes in Plant Cell Wall Composition and Structure During Cold Acclimation Daisuke Takahashi, Ellen Zuther, and Dirk K. Hincha Abstract The cell wall has a crucial influence on the mechanical properties of plant cells. It therefore has a strong impact on the freezing behavior and very likely also the freezing tolerance of plants. However, not many studies have addressed the question how cell wall composition and structure impact plant freezing tolerance and cold acclimation. In this chapter, we describe a comprehensive workflow to extract total cell wall material from leaves of Arabidopsis thaliana and to separate this material into fractions enriched in crystalline cellulose, pectins, and hemicelluloses by sequential fractionation. We further describe methods for the analysis of chemical structure, monosaccharide composition, and cellulose and uronic acid contents in the total cell wall material and the fractions in response to cold acclimation. Structural properties of cell wall material are analyzed by attenuated total reflectance-Fourier-transform infrared spectrometry (ATR-FTIR) and monosaccharide composition by gas chromatography–mass spectrometry (GC-MS) after isolation of alditol acetate derivatives of the sugars. Key words Cold acclimation, Cell wall, ATR-FTIR, Arabidopsis thaliana, Cellulose, Hemicellulose, Pectin, Uronic acid, Polysaccharide, Monosaccharide

1

Introduction Subzero temperatures can cause extracellular ice formation in plant tissues and expanding ice crystals absorb water from inside of the cell and may directly compress cells physically. Eventually, the growth of ice crystals may, directly or indirectly, lead to the disruption of the plasma membrane and consequentially to cell death [1, 2]. When temperate and boreal plants are exposed to nonfreezing, low temperatures, enhancement of freezing tolerance is induced, which is referred to as cold acclimation (CA) [3]. CA is accompanied by massive changes in gene expression, followed by changes in the abundance of various proteins and metabolites [2, 4–7]. The cell wall is considered to be important as one of the determinants of plant freezing tolerance, because it is the first contact site between ice crystals and plant cells and might regulate

Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_17, © Springer Science+Business Media, LLC, part of Springer Nature 2020

255

256

Daisuke Takahashi et al.

cellular freezing behavior and the extent of freezing injury to the plasma membrane. Previous studies have demonstrated that the cell wall increases in mass, thickness and rigidity during CA in several plant species [8– 14]. Furthermore, structural changes in the cell wall in response to CA, including changes in sugar composition of pectins and hemicelluloses, and in pectin methylation status have also been observed in oilseed rape, wheat and Arabidopsis [12, 15–17]. In comparison to other plant cell components, however, studies of cell walls in association with mechanisms of plant freezing tolerance are limited and further studies in other plant species and under different acclimation condition will be needed. In addition, the study of mutants in genes encoding enzymes active in cell wall biosynthesis and remodeling will help to understand the molecular basis of changes in cell wall structure and composition that have been observed during CA. In this chapter, we describe the experimental workflow and methods to detect CA-induced structural and compositional changes in cell walls (Fig. 1). By following this procedure, total cell wall content, the proportions of the three major cell wall components (crystalline cellulose, pectin, and hemicellulose), content of uronic acid (a major backbone component of pectin), and the composition of neutral sugars in pectins and hemicelluloses can be determined. In addition, a method to determine structural changes in cell walls by attenuated total reflectance-Fourier-transform infrared (ATR-FTIR) spectrometry is described. These procedures are adapted to Arabidopsis but can be equally applied to other plant species, including monocotyledonous plants, with some optimization.

2

Materials Prepare all solutions with analytical grade reagents and ultrapure water with an electrical conductivity lower than 0.055 μS/cm. Follow the regulations determined by local authorities when discarding solutions and experimental materials. Handle all organic solvents under a fume hood.

2.1 Cell Wall Extraction

1. Ball mill. 2. Vortex mixer. 3. Tabletop centrifuge. 4. Ethanol: store at RT or
20  C. 5. 80% (v/v) ethanol: mix 80 mL ethanol and 20 mL water. Store at RT. 6. Acetone: store at RT.

Cell Wall Analysis in Cold-Acclimated Plants

a

257

b Method 3.1.

Total cell wall fraction CDTA fraction

Arabidopsis plants

Na2CO3 fraction Method 3.6.

Method 3.2.

KOH fraction Insoluble fraction

Total cell wall fraction Method 3.3. Steps 1-2

Estimation of structural changes by ATR-FTIR

Method 3.4. Steps 1-6

CDTA-insoluble residues

CDTA extracts

Method 3.3. Step 3

Method 3.3. Steps 7-8

Precipitates

Method 3.5.1.

CDTA fraction

Na2CO3-insoluble residues

Na2CO3 extracts

Method 3.3. Step 4

Method 3.3. Steps 6-8

Crystalline cellulose quantification

Supernatants (200 µL)

Supernatants (100 µL)

Method 3.4. Steps 5-14

Method 3.5.2.

GC-MS samples

Uronic acid quantification

Na2CO3 fraction

KOH-insoluble residues

KOH extracts

Method 3.3. Steps 5, 7-8

Method 3.3. Steps 6-8

Insoluble fraction

KOH fraction

GC-MS analysis Quantification of sugar composition by GC-MS

Fig. 1 Workflow of overall experiments regarding analysis of cell wall described in this chapter. (a) Arabidopsis plant cultivation, total cell wall extraction and preparation of cell wall fractions. (b) quantification of crystalline cellulose and uronic acid, ATR-FTIR analysis for estimation of cell wall structural changes and quantification of neutral sugar composition in matrix polysaccharides

7. Methanol: store at RT. 8. Vacuum desiccator. 9. 10 mM Tris–maleate buffer: add 1.21 g Tris and 1.16 g maleic acid to 50 mL of water. Adjust pH to 6.9 by adding 45–48 mL of 0.2 M NaOH. After diluting to 1 L with water, add 0.11 g CaCl2 (1 mM final concentration) and 0.5 g NaCl (10 mM final concentration). Store at 4  C. 10. Aluminum heating block. 11. α-Amylase solution: α-amylase from porcine pancreas suspended in phosphate-buffered saline solution. Store at 4  C. 2.2 Cell Wall Fractionation

1. 50 mM 1,2-diaminocyclohexane tetraacetic acid (CDTA): adjust pH to 6.5 with NaOH. Prepare freshly. 2. 50 mM Na2CO3: solution is supplemented with 20 mM NaBH4. Prepare freshly (see Note 1). 3. 4 M KOH: solution is supplemented with 20 mM NaBH4. Prepare freshly (see Note 1).

258

Daisuke Takahashi et al.

4. Acetic acid: store at RT. 5. Dialysis tubes: dialysis membrane with a molecular weight cutoff of 3.5 kDa (width: 11.5 mm). Cut into 130 mm length pieces (see Note 2). 6. Dialysis tubing closures: 23 mm width. Prepare both weighted and nonweighted closures. 7. Lyophilizer. 2.3 TFA Hydrolysis and Preparation of Alditol Acetate Derivatives

1. Screw-capped microtubes (1.5 mL). 2. Aluminum heating block for test tubes equipped with needles for nitrogen gas flow. 3. 2 M trifluoroacetic acid (TFA): mix 34 mL water and 6.2 mL TFA. Prepare freshly (see Note 3). 4. 2-Propanol: store at RT. 5. Standard sugars: 1 mg/mL solutions of the following sugars dissolved in water: glucose (Glc), galactose (Gal), arabinose (Ara), xylose (Xyl), rhamnose (Rha), fucose (Fuc), mannose (Man), and myo-inositol (Ino). Prepare freshly. 6. Standard sugar mix for calibration curves: mix Glc, Gal, Ara, Xyl, Rha, Fuc, and Man standards at equal parts and dilute with water to concentrations of 200, 100, 50, 20, 10, 5, 2, 1, 0.5, and 0 mg in 270 μL final volume. Add 30 μL Ino (1 mg/mL) to all sugar mixes as an internal standard. Prepare freshly (see Note 4). 7. Screw capped Teflon-sealing glass tubes. 8. Reduction reagent: dissolve 10 mg/mL NaBH4 in 1 M ammonium hydroxide. Prepare freshly (see Note 1). 9. Glacial acetic acid: store at RT. 10. Acetic acid–methanol (1:9, v/v): store at RT. 11. Methanol: store at RT. 12. Acetic anhydride: store at RT. 13. Pyridine: store at RT. 14. Toluene: store at RT. 15. Methylene chloride: store at RT. 16. Acetone: store at RT.

2.4 Crystalline Cellulose and Uronic Acid Quantification

1. 72% (v/v) H2SO4: store at RT (see Note 5). 2. Standard Glc samples for calibration curves: prepare Glc standards at concentrations of 25, 15, 10, 5, 2, 1, and 0 μg in 100 μL water. Prepare freshly (see Note 6). 3. Anthrone reagent: add 0.1 g anthrone in 50 mL H2SO4. Prepare freshly (see Note 7).

Cell Wall Analysis in Cold-Acclimated Plants

259

4. Aluminum heating block. 5. Spectrophotometer. 6. Standard galacturonic acid (GalA) samples for calibration curves: prepare GalA standards at concentrations of 20, 15, 10, 5, 2, 1, and 0 μg in 100 μL water. Prepare freshly (see Note 8). 7. Sulfamate reagent: add 19.42 g sulfamic acid to 30 mL water. Add 24% (w/v) KOH to dissolve sulfamic acid and adjust pH to 1.6. Fill up with water to 50 mL. Store at RT. 8. Borate buffer: 12.5 mM Na2B4O7 in H2SO4. Store at RT. 9. Biphenyl solution: 0.15% (w/v) m-hydroxy biphenyl in 0.5% (w/v) NaOH. Prepare freshly (see Note 9). 2.5 ATR-FTIR Analysis

3

We use a PerkinElmer Spectrum GX FTIR spectrophotometer (PerkinElmer Inc., Waltham, MA, USA) equipped with Spectrum software (ver. 5.0.1.), and a Golden Gate single reflection diamond attenuated total reflection (ATR) system (Specac, Orpington, Kent, UK). However, there are many other ATR-FTIR systems available that can be used as well. ATR systems allow us to obtain IR spectra directly from cell wall powders that could not be used in conventional transmission FTIR. Spectra are obtained under ambient condition simply by pressing the sample powder onto the diamond stage of the ATR with the in-built sapphire anvil.

Methods

3.1 Plant Growth and Cold Acclimation

1. Sow Arabidopsis thaliana seeds on soil and keep plants at a light intensity of at least 200 μmol/m2/s (8 h day length) and a temperature of 20  C during the day and 18  C during the night for 4 weeks. 2. Transfer plants to a 16 h photoperiod under the same conditions for an additional week (nonacclimated plants, NA). 3. Transfer NA plants to a growth chamber at 4  C with a 16 h day length and a light intensity of 90 μmol/m2/s for up to 7 days (cold acclimated plants, CA).

3.2 Cell Wall Extraction

The following protocol is modified from published methods [18– 20]. Wear gloves, safety glasses and a lab coat throughout the experiments to protect yourself and avoid contamination. 1. Harvest aerial parts of Arabidopsis (1–3 g is suitable for isolation of cell walls) and carefully brush off soil with a soft-bristled brush (see Note 10). 2. Put the plants in a plastic vial. Freeze the sample with liquid nitrogen quickly and store at
80  C (see Note 11).

260

Daisuke Takahashi et al.

3. Grind frozen plant material with a ball mill to obtain a fine powder (see Note 12). Make sure that the material does not melt during grinding. Aliquot sample powder into 2 mL microtubes at 1 g each. Be careful not to thaw the frozen powder during this step. The frozen powder can be stored at
80  C. 4. Add 1 mL 80% ethanol to samples and vortex vigorously. Centrifuge samples at 1500  g for 5 min at RT and discard supernatants by pipetting to remove soluble material. Repeat this step until precipitates turn white or grayish and supernatants are colorless and transparent (see Note 13). 5. Add 1 mL acetone to the final pellets and vortex vigorously. Centrifuge samples at 1500  g for 5 min at RT and discard supernatants by pipetting. 6. Add 1 mL methanol to samples and vortex vigorously. Centrifuge samples at 1500  g for 5 min at RT and discard supernatants by pipetting. 7. Let precipitates dry in a vacuum desiccator for at least 12 h (see Note 14). 8. Add 100 μL Tris/maleate buffer to the samples and incubate at 98  C for 5 min in an aluminum heating block. Let samples cool at RT for 10 min. Add α-amylase solution to the samples at a volume corresponding to 25 U for NA and 40 U for CA material (approx. 1 U for 1 mg powder) and incubate for 1 h at 40  C (see Note 15). 9. Add α-amylase solution to the samples at half the amounts in step 8 and incubate for 30 min at 40  C. 10. Add 400 μL cold ethanol (
20  C) on ice and precipitate polysaccharides at
20  C for 1 h or overnight. Centrifuge samples at 1500  g for 5 min at 4  C and discard supernatants by pipetting. Repeat this step another 3 times. 11. Let precipitates dry in a vacuum desiccator for at least 12 h. Store at RT in well-sealed tubes. 3.3 Cell Wall Fractionation

Wear gloves, safety glasses, and a lab coat throughout the experiments to protect yourself and to avoid contamination. 1. Weigh 30 mg of dry cell wall powder and transfer it into a 2.0 mL screw-capped microtube. 2. Add 1.5 mL 50 mM CDTA solution to the sample on ice and incubate at 4  C for 12 h. Centrifuge sample at 14,000  g for 30 min at 4  C and transfer the supernatant to a fresh 15 mL tube by pipetting. Repeat this step twice and collect the supernatants in the same tube. 3. Add 1.5 mL 50 mM Na2CO3 to the sample on ice and incubate at 4  C for 12 h. Centrifuge sample at 14,000  g for 30 min at

Cell Wall Analysis in Cold-Acclimated Plants

261

4  C and transfer the supernatant to a fresh 15 mL tube by pipetting. Repeat this step twice and collect the supernatants in the same tube. 4. Add 1.5 mL 4 M KOH to the sample on ice and incubate at 4  C for 12 h. Centrifuge sample at 14,000  g for 30 min at 4  C and transfer the supernatant to a fresh 15 mL tube by pipetting. Repeat this step twice and collect the supernatants in same tube. 5. Add 1.5 mL water to the final precipitate (insoluble fraction) on ice and vortex vigorously. Centrifuge sample at 14,000  g for 30 min at 4  C and discard supernatant by pipetting. Repeat this step 3 times. 6. Add acetic acid to Na2CO3- and KOH-soluble fractions (19 μL and 900 μL, respectively) to acidify samples to pH 5.0 (see Note 16). 7. Clip one side of dialysis tubes with weighted closures. Transfer each of the four resultant fractions (Insoluble and CDTA-, Na2CO3-, and KOH-soluble fractions) into dialysis tubes separately. Close the other side of tubes with nonweighted closures. Soak in water at 4  C and keep stirring slowly for 3 days while exchanging water every 12 h. 8. Remove nonweighted closures and transfer dialyzed samples into fresh 15 mL tubes. Freeze them in liquid nitrogen. Close tubes with Parafilm and punch a small hole into the film with a needle. Lyophilize all frozen samples for 3 days. Weigh samples and store them in 2.0 mL screw-capped microtubes at RT. 3.4 TFA Hydrolysis and Preparation of Alditol Acetate Derivatives of Monosaccharides for GC-MS Analysis

All of these procedures must be performed in a fume hood at room temperature. Wear gloves, safety glasses and a lab coat throughout the experiments to protect yourself and avoid contamination. 1. Weigh 1–1.5 mg total cell wall material and cell wall subfractions and transfer into 1.5 mL screw-capped microtubes. Add 30 μL of 1 mg/mL Ino solution to all samples as an internal standard. 2. Add 250 μL of 2 M TFA to all samples and close microtubes tightly. Incubate for 1 h at 121  C and let samples cool at RT for 10 min (see Note 17). 3. Add 300 μL 2-propanol to each sample and let the solvent evaporate at 40  C on an aluminum heating block under a gentle flow of nitrogen gas. Repeat this step twice. 4. Add 300 μL water and vortex vigorously. Centrifuge samples at 14,000  g for 10 min at RT. Separate supernatants carefully from precipitates and transfer into fresh 1.5 mL microtubes. Proceed to crystalline cellulose quantification with precipitates (Subheading 3.5.1, see Note 18).

262

Daisuke Takahashi et al.

5. For the following steps, include all cell wall samples, samples containing standard sugars and the standard sugar mix at the different concentrations for calibration curves in GC-MS analysis. 6. Take all solutions/supernatants and split them into 200 μL in a microtube for uronic acid quantification (proceed to Subheading 3.5.2) and 100 μL in a screw-capped glass tube for alditol acetate derivatization (see Note 18). Dry samples in glass tubes under a gentle flow of nitrogen gas on an aluminum heating block at 40  C. 7. Add 250 μL reduction reagent and incubate at RT for 1 h. 8. Add 20 μL glacial acetic acid for neutralization and incubate until bubbling stops (see Note 16). 9. Add 250 μL acetic acid/methanol (1:9, v/v) to tubes and evaporate at 40  C on an aluminum heating block under a gentle flow of nitrogen gas. Repeat this step 3 times. 10. Add 250 μL methanol to tubes and evaporate at 40  C as above. Repeat this step 4 times. 11. Add 50 μL acetic anhydride and 50 μL pyridine, close the lids tightly and incubate for 20 min at 121  C (see Note 19). 12. Add 200 μL toluene to tubes and evaporate at 40  C as above. Repeat this step twice. 13. Add 500 μL water and 500 μL methylene chloride and vortex. The solution separates into two phases. Transfer the lower phase (methylene chloride) carefully into a 2 mL microtube with a Pasteur pipette (see Note 20). 14. Evaporate the solvent at RT under nitrogen gas flow. Add 100 μL acetone to the tube, transfer it into a GC-vial with a Pasteur pipette and clamp it with a Teflon-sealed lid. Store at
20  C until use. 15. Perform GC-MS analysis as described in Chapter 15. Data analysis can be performed under manual supervision as described in the same chapter. Alternatively, standard gas chromatography or high-performance liquid chromatography methods can be used to quantify the sugars. 3.5 Crystalline Cellulose and Uronic Acid Quantification

All of these procedures must be performed in a fume hood at room temperature. Wear gloves, safety glasses and a lab coat throughout the experiments to protect yourself and avoid contamination.

3.5.1 Crystalline Cellulose Quantification (Seaman Hydrolysis/ Anthrone Assay)

1. Add 175 μL 72% H2SO4 to precipitates after TFA hydrolysis (Subheading 3.4, step 4). Stir for 30 min, sonicate for 15 min and stir for 15 min at RT (see Note 21).

Cell Wall Analysis in Cold-Acclimated Plants

263

2. Add 425 μL water and vortex well. Centrifuge tubes at 1500  g for 5 min at RT. Proceed with 10 μL supernatant and 90 μL water. 3. Add 200 μL anthrone reagent to samples and Glc standards. Heat to 95  C on aluminum block heater for 5 min (see Note 19). Let the samples cool down at RT for 15 min. 4. Read absorbance at 640 nm with a spectrophotometer. Generate a standard curve to quantify total hexose content (Glc) in 100 μL sample volume. To calculate total crystalline cellulose content per milligram cell wall fractions (microgram crystalline cellulose/milligram cell wall fractions), divide results by factor 1.11 to accommodate water addition during conversion from cellulose polymer (162 g/mol) to Glc (180 g/mol). 3.5.2 Uronic Acid Quantification

1. Use 20 μL from supernatant of TFA hydrolysate (Subheading 3.4, step 6). For samples, fill up to 100 μL with water. 2. Add 10 μL sulfamate reagent to samples and GalA standards. Subsequently, add 600 μL borate buffer to all tubes. 3. Vortex well and heat to 121  C on aluminum block heater for 5 min (see Note 19). Let the samples cool down at RT for 15 min. 4. Split all samples and standards into 200 μL and 510 μL volumes for blank and sample measurement, respectively. Add 15 μL biphenyl solution to the samples (510 μL) and incubate for 10 min at RT. Read absorbance in both blank and sample solutions at 540 nm with a spectrophotometer. 5. Before generating the standard curve to quantify total uronic acid content in 20 μL sample volume, subtract the value of blank absorbance from sample absorbance. Calculate total uronic acid content per milligram cell wall fractions (microgram uronic acid/milligram cell wall fractions).

3.6 ATR-FTIR Analysis

Representative spectra obtained from total cell wall material and the four subfractions isolated as described above from leaves of nonacclimated Arabidopsis plants are shown in Fig. 2. Table 1 lists characteristic absorption bands of polysaccharides in the different fractions obtained from ATR-FTIR spectroscopy that can be used for estimation of structural changes in cell walls during cold acclimation. 1. Turn on FTIR spectrometer and install Golden Gate ATR system. Fully dry the sample powder in a vacuum desiccator or in a desiccator over silica gel beads. 2. Collect background spectra from measurements in the absence of any material on the diamond plate of the ATR with the settings described below.

264

Daisuke Takahashi et al.

Total cell wall 1626.03 1562.19

1036.90 1055.94

1413.50 1315.2

1244.36

1017.24 1074.77 1097.15 1049.53

CDTA 1606.62 1622.19

1416.33 1437.93 1370.35 1330.00

1731.80

1234.31

1143.79

917.64

Absorbance (a.u.)

888.54

Na2CO3 1633.51 1720.09

1539.36

1408.23 1343.19

1047.43 1072.87 1017.09 1232.43 1144.18 959.43 915.07 889.94

KOH 1641.32 1037.14

1537.70 1393.51 1446.88

1241.55 1152.70

945.18 897.99 1022.21

Insoluble

1611.06

1369.78 1412.61 1316.09

1155.93 1201.91

1800 1700 1600 1500 1400 1300 1200 1100 1000 Wavenumbers (cm-1)

896.06

900

Fig. 2 Representative FTIR spectra obtained from total cell wall, CDTA, Na2CO3, KOH and insoluble fractions of cell walls in nonacclimated Arabidopsis. The fingerprint regions in the spectra (1800–850 cm
1) are displayed. Representative peaks are marked in each spectrum. See Table 1 for details of peak annotation

3. Place cell wall powder on the diamond plate of the ATR to cover the active sampling area and press the sample gently by turning the torque screw knob. Obtain spectra with the settings shown below.

Cell Wall Analysis in Cold-Acclimated Plants

265

Table 1 A list of ATR-FTIR absorption bands obtained from cell wall components (adopted from SzymanskaChargot and Zdunek [21]) Wavenumber range (cm
1)

Related chemical structure

Origin

1745–1740, 1740–1730

C¼O stretching vibration of alkyl ester

Pectin

1640

H-O-H bending vibration

Absorbed water




1630–1600

COO antisymmetric stretching

Polygalacturonic acid (pectin ester group)

1428, 1426

CH2 symmetric bending

Cellulose

1410, 1400

COO
symmetric stretching

Carboxylate (pectin ester group)

1370, 1362

CH2 bending

Xyloglucan, cellulose

1330, 1320

Ring vibration

Pectin

1317, 1313

CH2 symmetric bending

Cellulose

1268–1230, 1243

C-O stretching

Pectin

1160

O-C-O asymmetric stretching

Cellulose (glycosidic linkage)

1146, 1143

O-C-O asymmetric stretching

Pectin (glycosidic linkage)

1147, 1130

O-C-O asymmetric stretching

Xyloglucan (glycosidic linkage)

1115, 1103

C-O stretching, C-C stretching

Cellulose (C2-O2)

1100, 1093

C-O stretching, C-C stretching

Pectin (ring)

1075, 1071

C-O stretching, C-C stretching

Xyloglucan (ring)

1042

C-O stretching, C-C stretching

Xyloglucan (ring)

1030

C-O stretching, C-C stretching

Cellulose (C6-H2-O6)

1019, 1014

C-O stretching, C-C stretching

Pectin (C2-C3, C2-O2, C1-O1)

1000

C-O stretching, C-C stretching

Cellulose (C6-H2-O6)

960, 954

C-O bending

Pectin

944, 941

Ring vibration

Xyloglucan

899, 895, 893

C1-H bending

Xyloglucan, cellulose (β-anomeric linkage)

833, 832

Ring vibration

Pectin

4. Remove or collect sample and clean ATR diamond plate with absolute ethanol. Place the next sample on the plate and repeat step 3. 3.6.1 Settings of ATR-FTIR Spectrometer

1. Parameters for data acquisition (Spectrum software ver. 5.0.1.): mode, ratio scan; number of scans, 64; scan range,

266

Daisuke Takahashi et al.

4000–400 cm
1; scan resolution, 4.0 cm
1; check H2O/CO2 mode to subtract the influence of H2O and CO2. 2. Steps for post data processing: ATR correction with a contact factor of 0, baseline correction, 1800–850 cm
1; adjust the sum of absorbance to the same value for all samples in the range of 1800–850 cm
1 (see Note 22).

4

Notes 1. Since NaBH4 can be easily degraded by moisture absorption, storage containers must be well-sealed. Na2CO3 or KOH should be completely dissolved prior to NaBH4 addition, because NaBH4 releases hydrogen gas under acidic or neutral pH conditions that may explode upon contact with oxygen. 2. Dialysis tubes contain glycerol as a humectant. In order to remove it, soak tubes in water for at least 30 min and wash thoroughly prior to use. 3. TFA is a highly acidic, corrosive and volatile chemical. Storage containers need to be well-sealed. Preparation of TFA-containing solutions needs to be conducted under a fume hood immediately before use. 4. At first, to minimize weighing errors, prepare sugar mix solution containing 10 mg/mL of each sugar standard. Make dilution series by diluting sequentially with water. Add 30 μL Ino (1 mg/mL) and adjust the volume to 300 μL with water. 5. Since H2SO4 generates heat when mixed with water, gradually mix H2SO4 into the water while stirring. 6. To minimize weighing errors, prepare 1 mg/mL Glc solution. Make dilution series by diluting sequentially with water and adjust the volume to 100 μL. 7. The color should be yellow, not brown. 8. To minimize weighing errors, first prepare 1 mg/mL GalA solution. Make dilution series by diluting sequentially with water and adjust the volume to 100 μL. 9. Prepare in a dark bottle or wrap the bottle in aluminum foil. 10. Stop watering 1 day before harvest to let the soil dry slightly to make soil removal easier. 11. If you need to calculate cell wall content based on dry weight, place harvested plants in an oven at 85  C for 3 days and allow to cool and weigh. Proceed to step 3 with 100 mg dried sample.

Cell Wall Analysis in Cold-Acclimated Plants

267

12. Settings of frequency and duration are 2–3 times at 30 Hz for NA. Since CA plants are harder, increase the number of grinding times to 4 to obtain a sufficiently fine powder. 13. In general, it is sufficient to repeat ethanol-washing 5 times to obtain white precipitates. 14. Stir the powder occasionally to promote complete drying. 15. Cell wall content may change depending on the growth environment. In the first experiment, weigh the cell wall powder at this step to estimate the amount of amylase. 16. In this process, adding acetic acid causes bubbles of hydrogen gas. Perform this step under a well-ventilated fume hood to avoid an explosive reaction. 17. TFA is very volatile and evaporates easily. Make sure that caps are tightened properly without any leakage and complete this step as quickly as possible. 18. TFA precipitates and supernatants can be stored at 4  C for several days. 19. Make sure that caps are tightened properly without any leakage. 20. Make sure that the upper phase (water) is not carried over into the microtube to avoid contamination in further steps. 21. Repeat stirring and sonication if precipitates are not fully dissolved. 22. As an alternative method of normalization, the absorbance can be set to 0 and 1 for the lowest and highest value within the wavenumber range of 1800–850 cm
1.

Acknowledgments This work was supported, in part, by the German Science Foundation (DFG) through Project A3 of the Collaborative Research Center CRC973 (to DKH) and a Grant-in Aid for Scientific Research from the Japan Society for the Promotion of Science (#27328 to DT). In addition, DT gratefully acknowledges a postdoctoral fellowship from the Alexander von Humboldt Foundation. References 1. Guy CL (1990) Cold acclimation and freezing stress tolerance: role of protein metabolism. Annu Rev Plant Physiol Plant Mol Biol 41:187–223 2. Thomashow MF (1999) Plant cold acclimation: freezing tolerance genes and regulatory mechanisms. Annu Rev Plant Biol 50:571–599

3. Levitt J (1980) Responses of plants to environmental stresses, 2nd edn. Academic Press, New York, NY 4. Chinnusamy V, Zhu J, Zhu J-K (2007) Cold stress regulation of gene expression in plants. Trends Plant Sci 12:444–451

268

Daisuke Takahashi et al.

5. Guy C, Kaplan F, Kopka J et al (2008) Metabolomics of temperature stress. Physiol Plant 132:220–235 6. Hincha DK, Espinoza C, Zuther E (2012) Transcriptomic and metabolomic approaches to the analysis of plant freezing tolerance and cold acclimation. In: Tuteja N, Gill SS, Tiburcio AF et al (eds) Improving crop resistance to abiotic stress. Wiley-VCH, Weinheim, pp 255–287 7. Thomashow MF (2010) Molecular basis of plant cold acclimation: insights gained from studying the CBF cold response pathway. Plant Physiol 154:571–577 8. Griffith M, Huner NPA, Espelie KE et al (1985) Lipid polymers accumulate in the epidermis and mestome sheath cell walls during low temperature development of winter rye leaves. Protoplasma 125:53–64 9. Weiser RL, Wallner SJ, Waddell JW (1990) Cell wall and extensin mRNA changes during cold acclimation of pea seedlings. Plant Physiol 93:1021–1026 10. Rajashekar CB, Burke MJ (1996) Freezing characteristics of rigid plant tissues (development of cell tension during extracellular freezing). Plant Physiol 111:597–603 11. Rajashekar CB, Lafta A (1996) Cell wall changes and cell tension in response to cold acclimation and exogenous abscisic acid in leaves and cell cultures. Plant Physiol 111:605–612 12. Kubacka-Ze˛balska M, Kacperska A (1999) Low temperature-induced modifications of cell wall content and polysaccharide composition in leaves of winter oilseed rape (Brassica napus L. var. oleifera L.). Plant Sci 148:59–67 13. Solecka D, Zebrowski J, Kacperska A (2008) Are pectins involved in cold acclimation and

de-acclimation of winter oil-seed rape plants? Ann Bot 101:521–530 14. Domon J-M, Baldwin L, Acket S et al (2013) Cell wall compositional modifications of Miscanthus ecotypes in response to cold acclimation. Phytochemistry 85:51–61 15. Willick IR, Takahashi D, Fowler DB et al (2018) Tissue-specific changes in apoplastic proteins and cell wall structure during cold acclimation of winter wheat crowns. J Exp Bot 69:1221–1234 16. Takahashi D, Gorka M, Erban A et al (2019) Both cold and sub-zero acclimation induce cell wall modification and changes in the extracellular proteome in Arabidopsis thaliana. Sci Rep 9:2289 17. Stefanowska M, Kuras´ M, Kubacka-Ze˛balska M et al (1999) Low temperature affects pattern of leaf growth and structure of cell walls in winter oilseed rape (Brassica napus L., var. oleifera L.). Ann Bot 84:313–319 18. Ruprecht C, Mutwil M, Saxe F et al (2011) Large-scale co-expression approach to dissect secondary cell wall formation across plant species. Front Plant Sci 2:23 19. Neumetzler L, Humphrey T, Lumba S et al (2012) The FRIABLE1 gene product affects cell adhesion in Arabidopsis. PLoS One 7: e42914 20. Pettolino FA, Walsh C, Fincher GB et al (2012) Determining the polysaccharide composition of plant cell walls. Nat Protoc 7:1590–1607 21. Szymanska-Chargot M, Zdunek A (2013) Use of FT-IR spectra and PCA to the bulk characterization of cell wall residues of fruits and vegetables along a fraction process. Food Biophys 8:29–42

Chapter 18 Subcellular Compartmentation of Metabolites Involved in Cold Acclimation Imke I. Hoermiller, Thomas N€agele, and Arnd G. Heyer Abstract Plant cells are heavily compartmentalized, and metabolite concentrations in the various compartments differ significantly. Thus, determination of metabolite abundance in whole-cell extracts may be misleading, when the role of a compound in plant freezing tolerance shall be evaluated. Here, we describe a method for the separation of the largest compartments of plant cells, the vacuole, plastid, and cytosol. With more elaborate analysis, this method can be expanded to also resolve mitochondria and other compartments. Key words Subcellular fractionation, Nonaqueous fractionation, Compartmentation, Cold acclimation

1

Introduction Plant cold acclimation causes a massive reprogramming of metabolism that affects not only metabolite concentrations, but also their subcellular localization [1]. Accumulation of hexoses, which is observed in many plants exposed to low temperature’s and may be important to reduce the amount of freezable water [2], depends on the presence of a vacuolar monosaccharide transporter [3], while the trisaccharide raffinose is actively transported into plastids during cold acclimation [4], where it is required for the protection of photosystem II [5]. Studying subcellular localization of metabolites is complicated by the fact that diffusion as well as enzymatic modification of metabolites must be stalled during separation of the compartments. This is either achieved by rapid filtration of protoplast extracts using filters with different cutoffs [6] or by arresting metabolism and diffusion through freeze drying and separation of compartments in water-free organic solvents [7]. The latter method is capable of analyzing intact tissue (e.g., leaves) and is thus more suited for the study of metabolite redistribution during cold acclimation.

Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_18, © Springer Science+Business Media, LLC, part of Springer Nature 2020

269

270

Imke I. Hoermiller et al.

The method described here uses whole shoots of Arabidopsis thaliana, but can, with some modifications, also be applied to other organs (e.g., potato tubers) [8]. In its standard form, ultracentrifugation is applied for compartment separation, but if this equipment is not available, an alternative approach using repeated ultrasonication and bench-top centrifugation can be employed [9]. As an extension to the method, the use of marker enzyme activities for compartment detection can be replaced by proteomics using peptide analysis via LC-MS/MS. This can improve the resolution and may allow for identification of additional compartments like mitochondria.

2 2.1

Materials Freeze Drying

1. 100 mL round bottom flask with threaded 29/32 joint for freeze drying. 2. Freeze dryer.

2.2 Ultrasonication and Density Gradient Centrifugation

1. Ice bath. 2. Solvent A (see Note 1): mixture of n-heptane and tetrachloroethylene, adjusted at a density of 1.3 g/mL by mixing 131.22 mL of tetrachloroethylene with 68.78 mL of nheptane. 3. Solvent B: pure Tetrachloroethylene (ρ ¼ 1.6 g/mL). 4. Ultra-Sonifier: here, a Branson Sonifier 250 was used with setting: output control 4 (Branson Ultrasonics, Dansbury, CT, USA). 5. Desiccator. 6. Nylon gauze with a pore size of 30 μm. 7. 13.2 mL polyallomer centrifuge tubes for ultracentrifugation, thin walled, 14  89 mm. 8. Gradient pump and mixer for linear gradient formation. 9. Organic solvent-resistant tubing (e.g., Viton/Fluran F-5500). 10. Ultra-centrifuge with swing-out rotor for 121,000  g.

2.3 Marker Enzyme Measurement

1. Extraction buffer for gradient fractions: 50 mM Tris–HCl pH 7.3, 5 mM MgCl2, 1 mM Dithiothreitol (DDT; add immediately before use), store on ice. 2. UV/Vis spectrophotometer. 3. 0.5 mL disposable polystyrene cuvettes. 4. 0.5 mL quart cuvettes. 5. Incubators for 37  C, 95  C.

Subcellular Compartmentation 2.3.1 Acid Phosphatase (Vacuole) [10]

271

1. Incubation buffer AP: 125 mM sodium acetate pH 4.8, 0.125% Triton X-100. 2. Substrate solution: 1 mg/mL p-nitrophenylphosphate (C6H6NO6P) in incubation buffer AP. 3. Stopping solution: 1 M Na2CO3.

2.3.2 UDP-Glucose Pyrophosphorylase (UGPase; Cytosol) [11]

1. Buffer UGP: 100 mM Tris/HCl pH 8.0; 2 mM MgCl2; 2 mM NaF.

2.3.3 Alkaline Pyrophosphatase (PPase, Plastid) [12]

1. Assay buffer: 50 mM Tris–HCl pH 8.0; 10 mM MgCl2; 1.3 mM Na-pyrophosphate. Prepare freshly from stock solutions: 1 M Tris–HCl pH 8.0, 100 mM MgCl2; 100 mM Na-pyrophosphate.

2. Assay solution: to 7 mL of Buffer UGP, add 17.5 μL NADP+ (100 mM), 28 μL UDP-glucose (500 mM), 14 μL glucose1,6-bisphosphate (10 mM), 3 U/mL phosphoglucomutase, 1 U/mL glucose-6-phosphate dehydrogenase.

2. Mixed Reagent Solution (see Note 2): freshly mix 2.5 mL H2SO4 (5 N), 375 μL ammonium molybdate (100 mM), 1.5 mL ascorbic acid (100 mM), 250 μL antimony potassium tartrate (2.743 g/L).

3 3.1

Methods Freeze Drying

3.2 Ultrasonication and Density Gradient Centrifugation

Snap freeze tissue in liquid N2. In a mortar precooled with liquid N2, grind about 1 g of frozen tissue to a fine powder. With a cold spatula, transfer the frozen powder to a round bottom flask and attach it to a freeze dryer. At a pressure of less than 101 mbar, the tissue will dry within about 2–3 days. Tissue should be lyophilized for at least 2 days to ensure complete drying. The dried powder can be stored under vacuum at room temperature for several days. Density gradients are mixtures of two volatile, hazardous organic solvents. Thus, working under a fume hood is required. Resuspend about 80–100 mg of freeze-dried tissue in 10 mL Solvent A. Under constant cooling in an ice bath, the suspension is sonified for 5 s with pauses of 15 s over a time course of 12 min (see Note 3). Subsequently, the suspension is filtered through nylon gauze (pore size 30 μm), and the filtrate is centrifuged at 2350  g, 4  C for 10 min. Remove the supernatant with a pipette, because the pellet is not very stable. The pellet is then resuspended in 1.5 mL Solvent A and stored on ice in a closed vessel. Using a gradient pump, an 8 mL linear gradient from Solvent A to Solvent B is produced in a centrifuge tube, and subsequently, 1 mL of the sonified suspension is added on top of the gradient.

272

Imke I. Hoermiller et al.

Fig. 1 Density gradient centrifugation of leaf homogenate from Arabidopsis thaliana. Freeze-dried tissue was sonicated in a mixture of tetrachloroethylene and heptane and loaded on a gradient of the same solvents with density ranging from 1.3 to 1.6 g/mL. Centrifugation was at 121,000  g for 3 h at 4  C

Use the remaining suspension to tare the tubes for centrifugation. In an ultracentrifuge, gradients are centrifuged at 121,000  g, 4  C, for 3 h. After centrifugation, the gradients are divided into nine fractions of 1 mL by carefully pipetting 1 mL fractions from the top of the gradient. Alternatively, fractionation can be achieved using a peristaltic pump and a glass capillary inserted into the tube. For leaf tissue, the plastidial compartment is usually the least dense and distributed over the first three fractions (Fig. 1). The vacuole is strongly enriched in fraction 9, while the cytosol is distributed over fraction 3–7. Each 1 mL fraction is divided into three aliquots of 300 μL. To each aliquot, 1 mL of n-Heptane is added and mixed. Aliquots are then centrifuged in a benchtop centrifuge at 18,000  g for 5 min. Subsequently, 1 mL of supernatant is removed, and the rest of the solvent is dried under vacuum in a desiccator. 3.3 Marker Enzyme Measurement

Having three sets of aliquots of the gradient fractions, one set is used for marker enzyme measurement. To the dry fraction aliquots, 1 mL Extraction Buffer is added, mixed and stored on ice. For compartments vacuole, cytosol, and plastids, marker enzyme activities are measured as follows.

3.3.1 Acid Phosphatase

Each of the gradient fraction is measured against its own blank, which is generated by adding the Stopping solution prior to Substrate. Therefore, prepare two sets of nine reaction tubes with 450 μL of Incubation buffer B, to which 150 μL of extract from

Subcellular Compartmentation

273

the gradient fractions are given. The tubes (300 μL) are incubated at 37  C for 5 min. To the samples, 100 μL Substrate solution are added and incubated at 37  C for 25 min. Then, reactions are mixed with 400 μL Stopping solution. To the blanks, 400 μL Stopping solution are added, before mixing with 100 μL Substrate solution. A reference is created by mixing 400 μL Stopping solution with 100 μL incubation buffer B. The difference of optical density at 410 nm between sample and blank is measured as proxy for enzyme activity. 3.3.2 UGPase

In quartz crystal cuvettes, 200 μL extract from gradient fractions are mixed with 500 μL Assay solution at room temperature. The reaction is started by adding 14 μL 100 mM Sodium pyrophosphate (Na4P2O7 · 10H2O). The increase of optical density at 334 nm is recorded for 6 min. The slope of a linear regression is proportional to UGPase activity. Although enzymatic activity could be calculated based on the extinction coefficient of NADPH, this is not needed, because only the relative distribution among fractions matters.

3.3.3 PPase

For each gradient fraction, 6 μL of extract are mixed with 800 μL Assay buffer and incubated at room temperature for 10 min. Assay buffer without extract is taken as reference. Reactions are stopped by incubation at 95  C for 5 min. In disposable cuvettes, the samples are mixed with 170 μL mixed reagent solution, and optical density is measured at 882 nm after 2–4 min.

3.4 Calculation of Metabolite Distribution

The method described above is not specific for the detection of certain metabolites, that is, any metabolite, for which a concentration has been determined in the nine gradient fractions, can be allocated to subcellular compartments. Thus, no specific method for metabolite measurements is described here. Since compartments are not completely separated, but show differential distribution among the density gradient, calculation of their allocation to the different compartments follows the rules applied to systems of equations with several variables. Considering three compartments, that is, cytosol, vacuole, and plastids, means that we have three variables, and consequently we need three independent equations to obtain a unique solution. Thus, we would need three gradient fractions to generate three equations. These equations would be: mF1 ¼ P F1 ∗xP þ CF1 ∗xC þ VF1 ∗xV mF2 ¼ P F2 ∗xP þ CF2 ∗xC þ VF2 ∗xV , mF3 ¼ P F3 ∗xP þ CF3 ∗xC þ VF3 ∗xV where mFn is the concentration of metabolite m in Fraction n, PFn is the plastidial marker enzyme activity in fraction n, CFn is the

274

Imke I. Hoermiller et al.

cytosolic marker enzyme activity in fraction n, and VFn is the vacuolar marker enzyme activity in fraction n. The factors xP, xC and xV are the required relative proportions of the metabolite in the compartments plastid, cytosol, and vacuole and should theoretically sum up to 1.0. However, because of experimental errors, it is more robust to divide the gradient into more fractions than required, because this allows solving the equation system to a best fit by the method of least-squares. In the method described above, we have generated nine gradient fractions, and so we have an overdetermined system of equations. The percentage of metabolite m found in gradient fraction i can now be calculated as: m i ½%  ¼

X m j ½%∗Eij ½%, j

where mj is the percentage of the metabolite allocated to compartment j and Eij is the percentage of marker enzyme activity for compartment j found in gradient fraction i. By varying the values of mj, the sum of squared deviations of the calculated mi from the real distribution of the metabolite among the gradient fractions is minimized to determine the optimized mj percentages. This can easily be achieved by using solver functions implemented in spread sheet programs like Microsoft Excel (Microsoft, Redmond, WA).

4

Notes 1. Make sure that all organic solvents are water-free. This can be achieved by adding molecular sieve, pore size 0.4 nm. A density of 1.3 g/mL is obtained by mixing 172 mL n-heptane with 328 mL Tetrachloroethylene. 2. The 100 mM stock solutions of ammonium molybdate and ascorbic acid as well as the antimony potassium tartrate and the DTT stock can be stored as aliquots at 20  C. 3. These parameters apply to a Branson Sonifier 250, output control 4, and may vary for other devices.

Acknowledgments This work was supported by grant HE 3087/10-2 from the German Science Foundation to AGH.

Subcellular Compartmentation

275

References 1. Hoermiller II, Naegele T, Augustin H et al (2017) Subcellular reprogramming of metabolism during cold acclimation in Arabidopsis thaliana. Plant Cell Environ 40:602–610 2. Hoermiller II, Ruschhaupt M, Heyer AG (2018) Mechanisms of frost resistance in Arabidopsis thaliana. Planta 248:827–835 3. Wormit A, Trentmann O, Feifer I et al (2006) Molecular identification and physiological characterization of a novel monosaccharide transporter from Arabidopsis involved in vacuolar sugar transport. Plant Cell 18:3476–3490 4. Schneider T, Keller F (2009) Raffinose in chloroplasts is synthesized in the cytosol and transported across the chloroplast envelope. Plant Cell Physiol 50:2174–2182 5. Knaupp M, Mishra KB, Nedbal L et al (2011) Evidence for a role of raffinose in stabilizing photosystem II during freeze-thaw cycles. Planta 234:477–486 6. Gardestro¨m P, Wigge B (1988) Influence of photorespiration on ATP/ADP ratios in the chloroplasts, mitochondria, and cytosol, studied by rapid fractionation of barley (Hordeum vulgare) protoplasts. Plant Physiol 88:69–76 7. Gerhardt R, Heldt HW (1984) Measurement of subcellular metabolite levels in leaves by

fractionation of freeze-stopped material in nonaqueous media. Plant Physiol 75:542–547 8. Farre´ EM, Tiessen A, Roessner U et al (2001) Analysis of the compartmentation of glycolytic intermediates, nucleotides, sugars, organic acids, amino acids, and sugar alcohols in potato tubers using a nonaqueous fractionation method. Plant Physiol 127:685–700 9. Fu¨rtauer L, Weckwerth W, N€agele T (2016) A benchtop fractionation procedure for subcellular analysis of the plant metabolome. Front Plant Sci 7:1912 10. Boller T, Kende H (1979) Hydrolytic enzymes in the central vacuole of plant cells. Plant Physiol 63:1123–1132 11. Zrenner R, Willmitzer L, Sonnewald U (1993) Analysis of the expression of potato uridinediphosphate-glucose pyrophosphorylase and its inhibition by antisense RNA. Planta 190:247–252 12. Jelitto T, Sonnewald U, Willmitzer L et al (1992) Inorganic pyrophosphate content and metabolites in potato and tobacco plants expressing E. coli pyrophosphatase in their cytosol. Planta 188:238–244

Chapter 19 Mathematical Modeling of Plant Metabolism in a Changing Temperature Regime Lisa Fu¨rtauer and Thomas N€agele Abstract Changes in environmental temperature regimes significantly affect plant growth, development and reproduction. Within a multigenic process termed acclimation, many plant species of the temperate region are able to adjust their metabolism to low and high temperature. Temperature-induced metabolic reprogramming is a nonlinear process affecting numerous enzyme kinetic reactions and pathways. The analysis of metabolic reprogramming during temperature acclimation is essentially supported by mathematical modeling which enables the study of nonlinear enzyme kinetics in context of metabolic networks and pathway regulation. This chapter introduces mathematical modeling of plant metabolism during a dynamic environmental temperature regime. A focus is laid on kinetic modeling and thermodynamic constraints. Key words Temperature acclimation, Climate change, Mathematical modeling, Plant metabolism, Enzyme activity, Kinetic modeling, Metabolic network, Thermodynamics

1

Introduction A detailed study of plant–environment interactions is essential for understanding dynamics and stability of ecosystems in context of global atmospheric changes. Temperature shapes plant growth, development, and physiology, and globally affects ecosystem composition and carbon cycles [1]. Temperature response of photosynthesis has previously been described in a consistent functional relationship across leaf–canopy–ecosystem scales [2]. Hence, quantitative models of photosynthetic metabolism in a changing environment promise to yield predictive output supporting the analysis of ecosystem dynamics in response to climate change. Within the process of temperature acclimation, many plants of the temperate region are able to reversibly adjust their photosynthetic metabolism [3]. While it is well described that plants possess considerable capacity of photosynthetic temperature acclimation, the role of metabolic reprogramming in acclimation is less well understood [4]. Plant cold acclimation is a multigenic process affecting

Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_19, © Springer Science+Business Media, LLC, part of Springer Nature 2020

277

278

€gele Lisa Fu¨rtauer and Thomas Na

expression of hundreds of genes which results in a broad transcriptional and metabolic response [5]. The study of metabolic reprogramming in large scale is focused by omics technologies which enable the simultaneous quantification of hundreds and up to thousands of transcripts, proteins, or metabolites [6]. Combination of omics technologies has unraveled an interplay between transcripts, lipids, and metabolites, resulting in molecular signatures of plant cold response and cold memory [7]. However, due to nonlinear enzyme kinetics and regulatory feedback/feedforward loops mathematical approaches are necessary for efficient data evaluation and predictive modeling [8]. This is particularly true for the analysis of temperature stress response due to additional nonlinear thermodynamic constraints [9]. For example, integration of light and temperature by the central oscillator of the plant circadian clock is preliminary for the precise timing over a range of physiological temperatures and has been recognized as an essential topic for ecosystem management and agricultural innovation [10]. Studies on the plant circadian clock have successfully combined large-scale approaches with mathematical and computational modeling to establish the clock structure [11–13]. Also in the field of biotechnology mathematical modeling of plant metabolism plays an essential role and the availability of plant genome-scale reconstructions for several plant species (e.g., rice, tomato, and maize) enables the functional integration of experimental high-throughput data in context of a genome sequence derived metabolic network [14]. In particular, the need for a combined sink–source model of plant metabolism for crop engineering emphasizes the essential role of mathematical modeling in current plant biology and biotechnology [15]. Based on recent work [8, 16], this chapter intends to introduce to mathematical modeling of plant metabolism in a changing temperature regime. It basically introduces to concepts of (1) model drafting, (2) model programming, (3) parameter optimization and numerical simulation, and (4) thermodynamic constraints of enzyme kinetics (Fig. 1). All steps are applied and explained in context of the central carbohydrate metabolism of photosynthetically active leaf cells which plays a central role in cold acclimation.

2

Development of a Kinetic Model

2.1 Model Construction

Kinetic models generally tend to consider only a relatively low number of reactions because of experimental limitations in recording enzyme kinetics and activities. Nonetheless, kinetic models are still a method of choice for estimating dynamics of a system over time [4]. Typically, mathematical modeling starts with a graphical representation of main enzymatic reactions and metabolites of a pathway (Fig. 1). This graphical representation might be initiated

Mathematics of Plant Temperature Acclimation

279

Fig. 1 Schematic workflow of model drafting, programming and validated simulation

by pen and paper. Yet, also open source software (e.g., CellDesigner [17], Copasi [18], or Virtual cell [19]) support computational model drafting. Furthermore, most of this software helps to translate the model into equations. Cellular metabolism is highly complex and needs to be simplified for kinetic modeling [20]. To provide an example, a simplified scheme of the central carbohydrate metabolism from a leaf mesophyll cell is summarized in Fig. 2. Every kinetic model construction comprises critical steps of metabolic and kinetic simplification which need to be considered very carefully and are exemplarily explained in the following. In this example, simplification leads to a reaction network of the central carbohydrates sucrose, glucose, fructose, and sugar phosphates. Simplified by one arrow are several enzymes, which are necessary for CO2 fixation into carbohydrates. The export function can be interpreted as direct phloem export to roots and export of carbon equivalents to other pathways. In this model subcellular compartmentation is ignored. For example, the reaction of sucrose cleavage catalyzed by invertase would in vivo take place in several compartments catalyzed by various isoforms. Consequently, model simulations will not reveal explicit insight into subcellular partitioning of metabolites which has to be considered when interpreting the output. Frequently, such simplification steps are necessary to allow kinetic modeling because only a fraction of biochemical and kinetic parameters is available. Comprehensive information about metabolic network structure which is necessary

280

€gele Lisa Fu¨rtauer and Thomas Na

Fig. 2 Simplified graphical representation of the central carbohydrate metabolism of leaf cells. Metabolites are described in boxes, and enzymatic reactions are indicated by arrows. Arrows can comprise several biochemical reactions including several metabolites and enzymes. Dashed lines indicate reactions being active under night (dark) conditions. AGPase ADP-glucose pyrophosphorylase, EXP export, Frc fructose, FRCK fructokinase, Glc glucose, GLCK glucokinase, IMP Import, INV invertase, NPS net photosynthesis, RESP respiration, Suc sucrose, SP sugar phosphates, SPP sucrose phosphate phosphatase, SPS sucrose phosphate synthase

for model draftig and construction is available in databases (e.g., KEGG (Kyoto Encyclopedia of Genes and Genomes) [21], plant reactome [22], or Metacyc [23]). The analysis of large omics studies with a kinetic modeling approach is limited by the knowledge of enzymatic parameters; nevertheless, some alternative approaches exist to (partly) overcome these limitations [24–28]. 2.2 Ordinary Differential Equations (ODEs)

Dynamic systems and many natural laws are described by differential equations. A differential equation describes the relationship between functions and related derivative functions, where a derivative function describes a rate of change. If the searched functions only depend on one variable (often on time) only ordinary derivatives appear summarized in ordinary differential equations (ODEs). In biological applications, a dynamic system can be defined by a set of ordinary differential equations describing influx versus efflux (Eq. 1): dx i ðt Þ ¼ influx  efflux ¼ synthesis  degradation dt  chemical interconversion  transport  . . .

ð1Þ

Mathematics of Plant Temperature Acclimation

281

The variable xi might represent a metabolite concentration in the plant metabolome, and its state depends on other state variables, and can be written as xi(t) ¼ {x1(t), x2(t), . . .}. These dependencies result in a coupled system of ODEs. Simultaneous integration of all ODEs will provide a solution of the described system [8]. Giving a concept: for a metabolite concentration xi over a time period the behavior is the result of the sum of all metabolic fluxes and reactions affecting its levels and, therefore, predominantly depends on enzymatic reactions [28]. 2.3 Translation of the Model into Equations

The translation of the graphical model of Fig. 2 into an ODE model results in equations summarized in Table 1. All reaction rates need to be defined by kinetic equations (Table 2). Typically, kinetic models are based on mass action law or Michaelis-Menten kinetics and use ordinary differential equations (see Fig. 1). In the model of Fig. 2, Michaelis-Menten kinetics with competitive inhibition can be applied for reaction rates of sucrose phosphate synthase (SPS, Eq. 2), invertase (INV, Eq. 3), and gluco- and fructokinase (GLCK Eq. 4, FRCK Eq. 5). Inhibitory constants (Ki) quantify the kinetic effect of inhibitors (here: sugar phosphates SP, Eqs. 4 and 5) on enzymatic rates. If insufficient kinetic information is available, mass action laws might be applied to estimate reaction rates based on experimental data on substrate and/or product concentrations (Export, Eq. 6). Gluco- and fructokinase are competitively inhibited by the reaction product SP. Competitive inhibition affects enzymatic substrate affinity, that is, KM values, and inhibitors can be outcompeted by high substrate concentration. A low inhibitory constant Ki

Table 1 Ordinary differential equations for diurnal simulation of the central carbohydrate metabolism Day condition

Night condition

d½CO2  dt

d½CO2  dt

d½SP dt

¼ IMP  NPS

¼ NPS þ GLCK þ FRCK  SPS  AGPase

d½Starch dt

¼ AGPase

d½SP dt

¼ RESP

¼ GLCK þ FRCK  SPS  RESP

d½Starch dt

¼ AMYLASES

d½Suc dt

¼ SPS  INV  EXP

d½Suc dt

¼ SPS  INV  EXP

d½Frc dt

¼ INV  FRCK

d½Frc dt

¼ INV  FRCK

d½Glc dt

¼ INV  GLCK

d½Glc dt

¼ INV  GLCK þ AMYLASES

d½Sink dt

¼ EXP

d½Sink dt

¼ EXP

AGPase ADP glucose pyrophosphorylase, EXP export, Frc fructose, FRCK fructokinase, Glc glucose, GLCK glucokinase, IMP import, INV invertase, NPS net photosynthesis, SP sugar phosphates, SPS sucrose phosphate synthase, Suc sucrose, RESP respiration

282

€gele Lisa Fu¨rtauer and Thomas Na

Table 2 Equations of enzyme kinetics max ∗½SP SPS ¼ v½SP þK M

INV ¼ 

(2) vmax ∗½Suc  ½Frc

½SucþK M ∗ 1þK

i,Frc

GLCK ¼

vmax ∗ ½Glc

½Glc

∗ 1þK



(3)

i,Glc



(4)



(5)

½SP

½GlcþK M ∗ 1þK

i,SP

FRCK ¼

vmax ∗ ½Frc

½SP

½FrcþK M ∗ 1þK

i,SP

EXP ¼ [Suc] ∗ kEXP

(6)

KM substrate affinity, Ki inhibitory constants, k equilibrium constant, EXP export, Frc fructose, FRCK fructokinase, Glc glucose, GLCK glucokinase, INV invertase, SP sugar phosphates, SPS sucrose phosphate synthase, Suc sucrose

indicates efficient inhibition, that is, a low inhibitor concentration is sufficient to significantly reduce enzyme activity. Invertase (Eq. 3) is differentially inhibited by both reaction products glucose and fructose. Fructose is a competitive inhibitor while glucose represents a noncompetitive inhibitor [29]. 2.4 Thermodynamic Constraints of Enzyme Kinetics

Enzyme activity is a function depending on pH and temperature. The velocity of enzyme catalyzed reactions increases by a factor 2–3 per each 10  C according to the van’t Hoff rule [30]. This theory was developed further to the so-called Arrhenius equation (Eq. 7). The influence of temperature on rates of chemical reactions is frequently interpreted in terms of this equation [31]. The rate constant k equals the product of a preexponential factor C and an Ea exponential term e R∗T . The constant C comprises information about geometric molecule position and collision frequency. In a single-step reaction, the exponential term quantifies the proportion of molecules which have sufficient energy to overcome the activation barrier of the reaction. Ea [J/mol] is the activation energy for the reaction, R [J/K mol] is the universal gas constant, and T represents the absolute temperature [K]. The unit of C varies depending on the reaction order, for a first order reaction C has units of [s1]. Ea

k ¼ C∗eR∗T

ð7Þ

The values for the activation energy Ea frequently range between 40 and 50 kJ/mol [30] and can be estimated from the slope of the Arrhenius plot or by measuring v at two different temperatures. Typically, enzyme activity (e.g., vmax) is experimentally recorded under optimum pH and temperature. For mathematical modeling under physiological conditions, recorded experimental

Mathematics of Plant Temperature Acclimation

283

data need to be adjusted to the growth temperature (e.g., to 22  C for nonacclimated plants and to 4  C for cold acclimated plants; exemplarily shown in Fig. 3). The example shows that values measured at optimum temperature (here: 30  C) increase 2.5-fold in cold acclimated plants compared to plants exposed to 22  C (Fig. 3a, b, grey squares). Applying the Arrhenius equation to estimate activity under physiological conditions, that is, at 22 and 4  C, results in a decreased activity at 4  C to 77% of activity at 22  C (Fig. 3a, b, black triangles). This clearly shows that if enzyme activity is experimentally not determined at plant growth temperature (e.g., 4  C) but instead at the temperature optimum of the enzyme, the application of the Arrhenius equation is necessary to adjust reaction rates of computational simulation to plant growth temperature. In conclusion, cold-induced protein accumulation does not necessarily result in a higher reaction rate in vivo. 2.5 Model Programming

To exploit computational capacities for model development and simulation, ODEs are typically programmed within a numerical software environment (e.g., MATLAB; www.mathworks.com; compare Fig. 1). ODEs are frequently referred to as “model states” for which initial conditions, that is, metabolite concentrations at the starting time point, need to be defined. Model parameters comprise kinetic parameters (e.g., KM, Ki or vmax). Following model and ODE programming, a first simulation run can be performed to validate the model syntax. In general, model simulation describes the process of solving the ODE system by numerical integration. Thus, simulation of a metabolic ODE system results in metabolite concentrations as a function of time, M(t). As a consensus platform for model exchange and programming, the Systems Biology Markup Language (SBML) was developed. SBML is an open interchange format for computer models of biological processes, and enables the use of multiple tools without rewriting models for each software tool [32]. In addition to abovementioned open source software, the numerical software MATLAB® is widely used for modeling approaches. Several freely available toolboxes exist which support model programming and simulation (e.g., IntiQuan [https://www.intiquan.com/]) [33]. A limiting step of kinetic modeling is the availability of sufficient kinetic information (e.g., KM, Ki and enzyme activity) [8, 24, 27]. This information is necessary to develop and solve ODEs within biochemically and physiologically meaningful context. Partially, this limitation can be overcome by information provided by enzyme databases such as BRENDA [34], which comprise numerous organisms and pathways, and help to estimate the physiologically relevant range of those parameters within the process of parameter estimation.

Fig. 3 Estimation of vmax values at different temperature regimes by applying the Arrhenius equation. (a) Calculation of adjusted vmax with varied Ea values (b) Differences of measured vmax (gray squares) at a certain measurement temperature and adjusted reaction rate for plants exposed temperature (vmax adj, black triangles)

Mathematics of Plant Temperature Acclimation

285

2.6 Parameter Estimation

After model programming and initial simulation, parameters need to be fitted to experimental data (compare Fig. 1). The best solution for the model is searched within the process of parameter estimation. Solution quality is quantified by a cost function which can be minimized or maximized under predefined constraints. In our model example, the cost function might represent the sum of squared errors between experimentally determined and simulated metabolite concentrations. In an iterative process, parameter estimation algorithms explore a parameter space which is constrained by experimental data on kinetic parameters or enzyme activities. Further, a solution is only biochemically valid if model simulation can reflect experimentally determined metabolite concentrations within their standard deviations. For example, the above calculated adjusted vmax values constrain the parameter space for the estimation process. Within the cell, enzymes are tightly regulated, for example, by product inhibition, and measured vmax values are not representing the “actual” enzyme activity, as the assays are performed under substrate saturation. As analytical solutions for many optimization problems are not available, numerical methods are applied. There exist different optimization strategies, linear and nonlinear as well as local or global optimizers (see, e.g., [35–37]). Nonlinear global optimization is not completely solved in mathematics. Frequently, analogies of nature are employed, for example by particle swarm algorithms. One of the first approaches of particle swarm optimization was developed by Eberhart and Kennedy [38, 39] which was a form of swarm intelligence in which the behavior of a biological social system was simulated by analogy with a school of fish or flock of birds [40]. If a swarm looks for food, each individual will spread into the environment and move around independently, and therefore has a degree of freedom or randomness in its movements. Finally, individuals will find food accumulation and announce this to their neighbors [37]. More abstractly, every particle (variable) is moving within its search space with the information of position and velocity. Further, “memory” is included because each particle movement is influenced by its local best-known position and the knowledge of global optima of all particles. As a result, the swarm will be guided toward the best solution. Parameters in kinetic models are typically vmax values and Ki and KM values, which are estimated in the search space which is constrained by experimental and literature data.

2.7 Mathematical Modeling of Plant Metabolism Is an Iterative Process

Model fitting and model simulation are iterative processes (compare Fig. 1). Model overfitting occurs if the parameters were chosen to closely to a particular set of data and therefore might fail to predict future observations. For example, indications of overfitting are extreme oscillations between two time points of metabolite concentrations. In the case of a time series experiment, the reliability of parameter estimation output and model simulation might be

286

€gele Lisa Fu¨rtauer and Thomas Na

validated by simulation of time points which were not part of model fitting. An identified model allows for the prediction of the system, for example, how metabolite concentrations change during a cold acclimation period or how the system will react if temperature fluctuates. Predictions over time are possible due to the application of a system defined by ODEs. ODEs are powerful in mathematical modeling because it is sufficient to describe the relationship of a systems state and its derivatives within a given time interval (i.e., initial conditions). The solution of ODE systems is able to predict the behavior of the system beyond this time interval. The simulation output has again to be iteratively validated, for example, by further experiments (Fig. 1). Finally, the complete process of kinetic modeling might start anew to prove new hypotheses under a changed temperature regime, and it might also comprise the change of model structure or experiments to finally gain new insights into regulation of plant metabolism.

Acknowledgments We thank Jakob Weiszmann (University of Vienna), Maria Pacheco (University of Luxemburg) and our colleagues from Plant Evolutionary Cell Biology at LMU Munich for their support and valuable discussions. This work was supported by the LMUexcellent Junior Researcher Fund. References 1. Piao S, Liu Q, Chen A et al (2019) Plant phenology and global climate change: current progresses and challenges. Glob Chang Biol 25:1922–1940 2. Ma S, Osuna JL, Verfaillie J et al (2017) Photosynthetic responses to temperature across leaf–canopy–ecosystem scales: a 15-year study in a Californian oak-grass savanna. Photosynth Res 132:277–291 3. Yamori W, Hikosaka K, Way DA (2014) Temperature response of photosynthesis in C3, C4, and CAM plants: temperature acclimation and temperature adaptation. Photosynth Res 119:101–117 4. Herrmann HA, Schwartz J-M, Johnson GN (2019) Metabolic acclimation - a key to enhancing photosynthesis in changing environments? J Exp Bot 70:3043–3056 5. Hannah MA, Wiese D, Freund S et al (2006) Natural genetic variation of freezing tolerance in Arabidopsis. Plant Physiol 142:98–112 6. Weckwerth W (2011) Green systems biology from single genomes, proteomes and

metabolomes to ecosystems research and biotechnology. J Proteome 75:284–305 7. Zuther E, Schaarschmidt S, Fischer A et al (2019) Molecular signatures associated with increased freezing tolerance due to low temperature memory in Arabidopsis. Plant Cell Environ 42:854–873 8. Fu¨rtauer L, Weiszmann J, Weckwerth W et al (2018) Mathematical modeling approaches in plant metabolomics. In: Anto´nio C (ed) Plant metabolomics: methods and protocols. Springer, New York, NY, pp 329–347 9. Ni XY, Drengstig T, Ruoff P (2009) The control of the controller: molecular mechanisms for robust perfect adaptation and temperature compensation. Biophys J 97:1244–1253 10. Gil K-E, Park C-M (2019) Thermal adaptation and plasticity of the plant circadian clock. New Phytol 221:1215–1229 11. Onai K, Okamoto K, Nishimoto H et al (2004) Large‐scale screening of Arabidopsis circadian clock mutants by a high‐throughput real‐time bioluminescence monitoring system. Plant J 40:1–11

Mathematics of Plant Temperature Acclimation 12. Locke JCW, Kozma-Bogna´r L, Gould PD et al (2006) Experimental validation of a predicted feedback loop in the multi-oscillator clock of Arabidopsis thaliana. Mol Syst Biol 2:59 13. Fogelmark K, Troein C (2014) Rethinking transcriptional activation in the Arabidopsis circadian clock. PLoS Comput Biol 10:e1003705 14. Gomes de Oliveira Dal’Molin G, Nielsen LK (2018) Plant genome-scale reconstruction: from single cell to multi-tissue modelling and omics analyses. Curr Opin Biotechnol 49 (Suppl C):42–48 15. Sweetlove LJ, Nielsen J, Fernie AR (2017) Engineering central metabolism – a grand challenge for plant biologists. Plant J 90:749–763 16. Weiszmann J, Fu¨rtauer L, Weckwerth W et al (2018) Vacuolar sucrose cleavage prevents limitation of cytosolic carbohydrate metabolism and stabilizes photosynthesis under abiotic stress. FEBS J 285:4082–4098 17. Funahashi A, Morohashi M, Kitano H et al (2003) CellDesigner: a process diagram editor for gene-regulatory and biochemical networks. Biosilico 1:159–162 18. Hoops S, Sahle S, Gauges R et al (2006) COPASI - a complex pathway simulator. Bioinformatics 22:3067–3074 19. Loew LM, Schaff JC (2001) The Virtual Cell: a software environment for computational cell biology. Trends Biotechnol 19:401–406 20. Klipp E, Liebermeister W, Wierling C et al (2016) Systems biology: a textbook. WileyVCH, Weinheim 21. Kanehisa M, Goto S (2000) KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res 28:27–30 22. Naithani S, Preece J, D’Eustachio P et al (2016) Plant reactome: a resource for plant pathways and comparative analysis. Nucleic Acids Res 45:D1029–D1039 23. Caspi R, Billington R, Fulcher CA et al (2017) The MetaCyc database of metabolic pathways and enzymes. Nucleic Acids Res 46: D633–D639 24. Steuer R, Gross T, Selbig J et al (2006) Structural kinetic modeling of metabolic networks. Proc Natl Acad Sci U S A 103:11868–11873 25. Reznik E, Segre` D (2010) On the stability of metabolic cycles. J Theor Biol 266:536–549 26. Henkel S, N€agele T, Ho¨rmiller I et al (2011) A systems biology approach to analyse leaf

287

carbohydrate metabolism in Arabidopsis thaliana. EURASIP J Bioinform Syst Biol 2011:2 27. Fu¨rtauer L, N€agele T (2016) Approximating the stabilization of cellular metabolism by compartmentalization. Theory Biosci 135:73–87 28. N€agele T (2014) Linking metabolomics data to underlying metabolic regulation. Front Mol Biosci 1:22 29. Sturm A (1999) Invertases. Primary structures, functions, and roles in plant development and sucrose partitioning. Plant Physiol 121:1–8 30. Bisswanger H (2002) Enzyme kinetics: principles and methods. Wiley-VCH, Weinheim 31. Laidler KJ (1984) The development of the Arrhenius equation. J Chem Educ 61:494 32. Hucka M, Finney A, Sauro HM et al (2003) The systems biology markup language (SBML): a medium for representation and exchange of biochemical network models. Bioinformatics 19:524–531 33. IQM-Tools (online) http://www.intiquan. com/iqm-tools/ 34. Schomburg I, Hofmann O, Baensch C et al (2000) Enzyme data and metabolic information: BRENDA, a resource for research in biology, biochemistry, and medicine. Gene Funct Dis 1:109–118 35. Banga JR (2008) Optimization in computational systems biology. BMC Syst Biol 2:47 36. Reali F, Priami C, Marchetti L (2017) Optimization algorithms for computational systems biology. Front Appl Math Stat 3:6 37. Weise T (2009) Global optimization algorithms-theory and application. SelfPublished Thomas Weise 38. Eberhart R, Kennedy J (1995) A new optimizer using particle swarm theory. In: MHS ‘95, proceedings of the Sixth International Symposium on Micro Machine and Human Science. IEEE, Washington, DC, pp 39–43 39. Shi Y, Eberhart R (1998) A modified particle swarm optimizer. In: 1998 IEEE International Conference on Evolutionary Computation Proceedings. IEEE World Congress on Computational Intelligence (Cat. No. 98TH8360). IEEE, Washington, DC, pp 69–73 40. Parrish JK, Hamner WM (1997) Animal groups in three dimensions: how species aggregate. Cambridge University Press, Cambridge

Chapter 20 Characterization of Ice-Binding Proteins from Sea-Ice Microalgae Maddalena Bayer-Giraldi, EonSeon Jin, and Peter W. Wilson Abstract Several species of polar microalgae are able to live and thrive in the extreme environment found within sea ice, where ice crystals may reduce the organisms’ living space and cause mechanical damage to the cells. Among the strategies adopted by these organisms to cope with the harsh conditions in their environment, ice-binding proteins (IBPs) seem to play a key role and possibly contribute to the success of microalgae in sea ice. Indeed, IBPs from microalgae predominantly belong to the so-called “DUF 3494-IBP” family, which today represents the most widespread IBP family. Since IBPs have the ability to control ice crystal growth, their mechanism of function is of interest for many potential applications. Here, we describe methods for a classical determination of the IBP activity (thermal hysteresis, recrystallization inhibition) and further methods for protein activity characterization (ice pitting assay, determination of the nucleating temperature). Key words Sea-ice microalgae, Diatoms, Ice-binding proteins, Antifreeze, Thermal hysteresis (TH), Clifton nanoliter osmometer, Ice recrystallization inhibition (IRI), Recrystallometer, Pitting assay, Nucleation, Supercooling, Lag time

1

Introduction Sea ice is mainly a two-phase system, characterized by a solid phase and liquid brine. During ice formation, solutes in the seawater are excluded from the ice matrix and segregate into brine droplets or brine channels, generally defined as brine inclusions inside sea ice [1]. Time-dependent outflow of high salinity brine and inflow of seawater of lower salinity cause a shift to a higher freezing point in the liquid inclusions, resulting in the narrowing of the inclusions and separation into individual pockets divided by ice bridges. The porous structure of sea ice is determinant for the biological activity that develops inside the ice. Despite the harsh environmental conditions, with temperatures ranging from about
1.8  C on the bottom to
20  C or less on the top of the sea-ice layer [2, 3] and brine salinities up to 200 on the Practical Salinity Scale [4],

Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_20, © Springer Science+Business Media, LLC, part of Springer Nature 2020

289

290

Maddalena Bayer-Giraldi et al.

brine inclusions offer a habitat for a variety of microalgae and other microorganisms. These cells play a crucial role for the ecology of the Polar Oceans, since they represent a concentrated food source in the low-productivity ice-covered sea, and in the months of melting they initiate blooms by seeding the water column [5]. The strategies adopted by ice microorganisms to cope with conditions in sea ice remain to be unraveled. A large variety of cold-tolerant organisms, including fishes, insects, plants, and microorganisms, have ice-binding proteins (IBPs) (for reviews see [6, 7]). These proteins are common in polar species but have never been found in temperate organisms, suggesting that IBPs play a key role in adaptation to subzero conditions. Predominant among seaice microorganisms are IBPs from one family characterized by the “domain of unknown function” (DUF) 3494, as defined in the Pfam database [7]. The DUF4393-IBP family represents today the most widespread of the known IBP families and has been found in bacteria [8–10], diatoms [11–13], and fungal species [14–16]. In the generally accepted adsorption–inhibition model describing the mechanism of action of IBPs, proteins bind to the ice lattice and locally inhibit ice growth by the Gibbs–Thomson effect [17, 18]. The binding mode can be reversible [19, 20] or irreversible [21–23]. One of the most prominent effects of IBPs is thermal hysteresis, which describes the separation of the freezing point of a solution below the melting point. Another effect which defines IBPs is the inhibition of grain growth, the temperature-dependent grain boundary migration resulting in the growth of large crystals at the expenses of smaller ones. In the IBP literature, this grain growth is often referred to as “recrystallization” and we will follow this tradition also in this chapter, although this differs from the terminology used in physical ice research to describe this process. The nomenclature of IBPs varies, depending on their predominant activity, from ice binding to antifreeze (AFPs) or ice structuring proteins. The biological roles of IBPs from sea-ice microorganisms can be diverse. The importance of some IBPs, as observed in fishes or insects, lies in lowering the freezing point below environmental temperature, in order to avoid ice formation in cells or organs. Other IBPs have the function to inhibit recrystallization, as it has been suggested for plant IBPs. In the context of sea ice, it has been observed that most IBPs have a signal peptide and are secreted from the organisms into the surrounding medium. It was suggested that the proteins accumulate within a layer of extracellular polysaccharide substances (EPS) produced by several sea-ice microorganisms [24]. Microalgal IBPs, produced recombinantly or collected from spent growth medium, affect the structure of the ice surface, causing pitting and characteristic microstructural features [11, 13, 25]. This suggests that the proteins shape the frozen environment in order to increase the habitable space within sea ice, potentially

Microalgal Ice Binding Proteins

291

affecting ice porosity and therefore also the biogeochemical imprint within the ice. Other IBPs from sea-ice organisms contain a lipobox signal peptide, which associates the IBP with the cell membrane. In these cases, the IBP may function as a kind of adhesin. The cells can thus be attached to loose ice crystals grown in the ocean, and floating up to the frozen surface, or directly to the sea-ice layer in order to take advantage of better growing conditions under the ice [10, 26]. The characterization of IBPs is of relevance not only to understand their functional role in sea ice, but also in the frame of possible applications of IBPs in the medical field, in the food industry and in other fields related to a control of ice crystals. In the following, we present some standard techniques traditionally applied in IBP research to assess the protein activity in terms of thermal hysteresis (TH) and ice recrystallization inhibition (IRI), which define the proteins as ice binding. Also, we present further methods (ice pitting assay, determination of the nucleating temperature) to characterize the activity of IBPs. However, a large variety of methods has been applied to determine the effects of IBPs from microorganisms from sea ice or the ocean beneath, focusing on ice-growth physics [13], structural studies [8, 23, 26–28], and Molecular Dynamics (MD) simulations [23], among others. However, the description of all these methods would be beyond the scope of this chapter.

2

Materials

2.1 Measurement of Thermal Hysteresis Using a Nanoliter Osmometer

1. Glass micropipettes (prepulled or pulled individually), rubber tubes fitting the micropipettes. 2. Nanoliter osmometer: cooling stage, sample holder plate with sample wells, controlling unit (Fig. 1).

Fig. 1 Clifton nanoliter osmometer and loaded samples. (a) Controlling unit and cooling stage positioned on a stereo microscope; (b) Cooling stage with central sample holder plate, 1-cent coin for size comparison; (c) Sample holder plate as viewed in a stereo microscope, with samples within each of the eight oil-loaded wells

292

Maddalena Bayer-Giraldi et al.

3. Tap water if colder than 18  C or refrigerated circulating cooling fluid, dry air (nitrogen gas), stereomicroscope, cover glass. 4. Immersion oil A (viscosity 150 cSt(lit)), B (viscosity 1250 cSt (lit)), thermal heat sink paste, fine needles, chloroform. 2.2 Measurement of Ice Recrystallization Using an Optical Recrystallometer

1. Optical recrystallometer (Otago Osmometers Ltd, Dunedin, New Zealand). 2. Refrigerated circulating cooling fluid (ethylene glycol), dry air (nitrogen gas). 3. Multimeter. 4. Sample glass tubes (dimensions 8 mm outer diameter, 0.45 mm wall thickness, 8 cm high). 5. A beaker with 100% ethanol cooled overnight to
40  C or
80  C.

2.3 Ice Pitting Activity of IBPs

1. Ice-pitting instrument (Figs. 2 and 3) composed of a sample holder with the IBP solution set in a temperature-controlled bath with refrigerated circulating cooling fluid.

Fig. 2 Schematic of experimental instrument used to observe growth of ice crystals [30]

Fig. 3 Instrument for ice-pitting assay in the laboratory

Microalgal Ice Binding Proteins

293

Coolant Photodiode Al block Sample (in NMR tube) Peltier Brass heat sink Thermocouple Laser

Peltier current

Fig. 4 Schematic of the automatic lag time apparatus (ALTA)

2. Optical microscope. 3. Digital camera or movie-clip recording instruments. 4. Glass petri dishes. 5. Glass slides (Dimension of slides: 0.5 cm (width)  4 cm (height)  0.1 cm (thickness)). 6. Glass sample tubes (Dimension of rectangular part: 1 cm (width)  0.5 cm (width)  2 cm (height)). 2.4 Measuring the Ice Nucleation Properties of IBPs from Sea-Ice Algae

1. Automatic lag time apparatus (purpose-built; Otago Osmometers Ltd, Dunedin, New Zealand, Fig. 4). 2. Dry air (nitrogen gas). 3. Refrigerated circulating cooling fluid (isopropanol–water mix). 4. Data recorder. 5. Purpose built cold stage to sit a differential scanning calorimeter (DSC) pan on or modified commercial cold stage capable of reaching
30  C. 6. Aluminum DSC pans. 7. Refrigerated circulating fluid (isopropanol/water mix).

3

Methods

3.1 Measurement of Thermal Hysteresis Using a Nanoliter Osmometer

One characteristic of IBPs is their ability to lower the freezing point of a solution and separate it from the melting point [29–31]. This effect, known as thermal hysteresis (TH), is noncolligative and differs from equilibrium freezing point depression, where freezing and melting temperatures still coincide. Thermal hysteresis is easily determined with the Nanoliter Osmometer, which measures the melting and the freezing points of small volume samples (1–10 nL). We describe a measurement performed with the Clifton nanoliter osmometer (Clifton Technical Physics, Hartford, NY), but other osmometers follow the same principle (Otago Osmometer Ltd, Dunedin, New Zealand, or LabVIEW operated devices [32]). The

294

Maddalena Bayer-Giraldi et al.

Clifton nanoliter osmometer shown in Fig. 1 consists of a cooling stage, operated by a Peltier element, in contact with a sample holder plate. The cooling stage is regulated by a controller unit, which permits temperature adjustment in the range of approximately 1mOsm ¼ 0.00186  C. Samples are viewed through a stereo microscope. Sample drops are loaded into wells filled with oil to reduce sample evaporation, the melting and freezing points determined visually. 1. Pull the glass micropipettes with a capillary puller in order to have a sharp end. If the fine end of the micropipette is closed, open it by gently scratching it (see Note 1). 2. Fill the glass micropipettes and the rubber tubes by setting them over night in immersion oil A. 3. Assemble the sample loading tube connecting each tube to two micropipettes, one on each end (see Note 2). 4. Connect the cooling fluid device to the osmometer. 5. Set the sample holder plate (cleaned in chloroform and dried) on the cooling stage using thermal paste (see Note 3). Fill the sample wells on the sample holder plate with immersion oil B using a fine needle. 6. Cover the sample with immersion oil A to avoid evaporation. 7. Load the sample into a sample loading tube by gently pressing the rubber part of the tube and then releasing it when submerged in the sample. Clean the loading tip with a tissue to remove remains of sample and obtain a clean tip. 8. Carefully insert the clean tip of the sample loading tube into a sample well. Release a drop of liquid by gently pressing the rubber part of the tube (see Note 4). Fill into all wells. 9. Turn on the dry air in order to avoid condensation during measurement and cover the sample plate with a cover glass. 10. Set the temperature of the osmometer slightly below the melting temperature. Turn on the osmometer, shock-freeze samples at
40  C. A sudden color change of the samples, from clear to dark, indicates their freezing. Release the “Freeze” switch and adjust temperature to the melting point. 11. Melt sample until only one crystal is left (see Note 5). Adjust the volume of the crystal to be as small as possible. Observe the crystal and note melting temperature (shrinking of this grain). 12. Slowly lower temperature and note freezing temperature (growth of the grain). The presence of IBPs becomes evident in a “hysteresis gap,” a temperature range between freezing and melting point. Within this range the crystal will neither melt nor grow. The freezing of the sample can be observed as a “burst,” if protein concentration is not too low (see Note 6).

Microalgal Ice Binding Proteins

3.2 Measurement of Ice Recrystallization Using an Optical Recrystallometer

295

Ice grains, especially small grains as obtained after shock-freezing, are subject to instabilities causing migration of grain boundaries. This effect, called ice recrystallization, leads to the growth of large ice crystals at the expenses of the smaller ones. A characteristic for IBPs is their ice recrystallization inhibition (IRI) activity, inducing a stabilizing effect on the grain dimensions [33]. This inhibiting effect can be measured with an optical recrystallometer (Otago Osmometers Ltd, Dunedin, New Zealand). After shock freezing, a large quantity of ice grains of small dimensions is formed and, as a consequence of light reflection at grain boundaries, the sample appears optically thick. A recrystallized sample with few, large grains appears clear. The recrystallometer detects the light intensity of a beam passing through the sample and gives an estimate of the recrystallization process (Fig. 5). 1. Connect the recrystallometer to the cooling fluid device and to the dry air supply. Connect the multimeter to the recrystallometer. 2. Turn on the recrystallometer and set the temperature to an annealing temperature close to the melting point (e.g.,
4  C) (see Note 7). 3. Cool the sample tubes for at least 10 min in cold ethanol (see Note 8). 4. Load the sample (150 μL) into the sample tubes using a Pasteur pipette. The solution will freeze immediately (see Note 9). Put the tube back into the cold ethanol and incubate for 10 min or more.

Fig. 5 Ice recrystallization measured using the optical recrystallometer. The negative control (red) with a buffer (phosphate-buffered saline) recrystallizes over time, which can be seen as an increase in voltage due to higher light intensities passing through the sample with larger ice grains. In the presence of IBPs, the signal does not change over time, grains maintain their small dimensions

296

Maddalena Bayer-Giraldi et al.

Fig. 6 Images of the deformation on the ice surface as seen with the ice-pitting assay. Scale bars indicate 1 mm [31]

5. Take the sample and quickly dry the tube with a tissue to remove the ethanol (see Note 10). Position the tube into the recrystallometer. 6. Start the multimeter. Measure for 1 h, or the time more appropriate for your needs. 7. Download the multimeter results expressed as tension (millivolt). Recrystallization causes an increase in tension over time, whereas samples with IBPs do not change. 3.3 Ice Pitting Activity of IBPs

Ice-binding proteins have the ability to bind to ice surfaces and to prevent the growth of ice. The proteins usually bind defined planes of an ice crystal, typical for each protein family. Within the hysteresis gap between the melting and the freezing points (see Subheading 3.1) the crystal will only grow in the directions not affected by the proteins, developing characteristic pitting patterns (Fig. 6). For a simple determination of ice-binding activity from samples, the depth and degree of the pitting deformations on small ice plates can be observed and the ice-binding activity can be determined qualitatively [34, 35]. 1. Prior to the experiment, check the osmolality of the samples. The temperature of the cooling liquid in the temperaturecontrolled bath should be slightly below the freezing point. For an osmolality of approximately 1000 mOsm/kg, set the temperature to
1.9  C. 2. Fill half of the glass petri dish with distilled water previously degassed under vacuum, to create the ice plates for the assay. 3. Shock-freeze the water by incubation at
20  C for approximately 40 min, in order to make a flat ice sheet of approximately 0.5 mm thickness (see Note 11). 4. Cool a glass slide at
20  C.

Microalgal Ice Binding Proteins

297

5. Prepare a 0.5 cm  0.5 cm piece of ice from the petri dish and attach it to the chilled glass slide. 6. Put the sample tube filled with approximately 1 mL of the solution to be tested into the temperature-controlled bath. Adjust the position for optimal observation of the sample. 7. Put the glass slide with the ice piece into the cooled sample tube with great care (see Note 12). 8. Adjust the focus and magnification of the optical microscope and take pictures of the deformation on the surface of the ice. 3.4 Measuring the Ice Nucleation Properties of IBPs from Sea-Ice Algae

Determination of the ice nucleation properties of proteins from polar algae requires multiple measurements on the same batch, in the same container, since the temperature of nucleation always differs with successive runs, even when all other factors are kept constant. Polar diatoms synthesize IBPs, but it remains unclear what the actual purpose of these proteins is during the life cycle of the diatoms [10, 24, 26]. One possibility is that the algae use them to bind to the sea ice and remain in the photolayer. In order to determine if diatoms enhance or inhibit ice nucleation we must first know the average nucleation temperature of the culture medium in its container. It is often the container which causes nucleation of supercooled solutions, for example, through a scratch on the wall or a piece of dirt. We describe here two methods for analyzing the nucleation characteristics of diatoms, the first requires a purpose-built device known as an automatic lag time apparatus (ALTA), while the second is simpler, requiring only a cooling stage and differential scanning calorimeter (DSC) pans.

3.4.1 Measurements with an Automatic Lag Time Apparatus (ALTA)

ALTA repeatedly supercools a single liquid volume until it freezes [36]. The sample is cooled linearly, freezing is detected optically, the sample warmed and the process repeated perhaps 200 times, as shown in Fig. 7 (see Notes 13 and 14 for details).

Nucleation temperature

-16 -14 -12 -10 -8 -6 -4 -2 0

1

51

101

151

Run number

Fig. 7 Manhattan for a typical set of runs on ALTA, showing the stochastic nature of nucleation, where each run on the same sample freezes at a different temperature

298

Maddalena Bayer-Giraldi et al.

TT

P

S

CS

TEC HS CF

Fig. 8 Schematic of the DSC pan type measurement arrangement: heat sink (HS), thermoelectric cooling (TEC), cooling fluid (CF), thermal transfer material (TT), DSC pan (P) with liquid sample (S) inside and coverslip (CS). Freezing is detected optically from above

1. A volume of typically 200 μL is placed in a glass tube which resides snugly in a hollowed-out 10 mm thick aluminum block. A thermocouple rests outside the tube to prevent unwanted nucleation sites within the liquid. 2. The metal sample holder is sandwiched between two Peltier modules which are used to heat and cool the sample by computer control. Excess heat is removed by heat sinks cooled by a flowing isopropanol/water mix. 3. Freezing of the sample is monitored by the (interrupted) transmission of a low power diode laser, which causes the computer to switch direction of the current to the Peltier elements. The sample is then heated to 283 K, or more, for some time to ensure melting of all residual ice crystals prior to commencing another run. 3.4.2 DSC Pan Type Measurement

A simpler method is to use a DSC pan sitting on a cold stage, with a typical sample volume of 10 μL of water, which will supercool and freeze at about
23  C [37] (Fig. 8). What temperature will the water/solution freeze at (in that pan) if a sample of diatoms is added? 1. Sample is added to the DSC pan, a coverslip placed on top, and cooling of the pan begun. 2. The freezing event is detected optically and temperature of freezing recorded. 3. The current to the cold stage Peltier elements is reversed and the DSC pan is heated to above 10  C in order to melt the ice. 4. The process is then repeated at least 50 times.

4

Notes 1. You can control the opening size of the micropipette tip by immersing it into water and passing air though the micropipette. The size of the air bubbles will be indicative for the tip opening.

Microalgal Ice Binding Proteins

299

2. Take care to remove air bubbles from the sample loading tubes by gently pressing the rubber part. 3. Use only a small amount of paste. Too much paste will mix with the immersion oil and disturb the measurement. Distribute the paste homogeneously on the sample holder and press it carefully on the cooling stage, in order to ensure maximal heat transfer between the cooling stage and the sample holder plate. 4. The diameter of the liquid sample drop should be around half the diameter of the well. You can control the size by pressing or releasing the rubber part of the sample loading tube. The sample should float in the oil and be positioned centrally within the sample well, without any direct contact with the plate. The sample can be carefully moved with a fine needle. 5. The bias on the freezing point due to the supercooling effect is avoided by observing one individual crystal. 6. Temperature changes should be performed carefully and not too fast, since the temperature response of the osmometer is slow. For time dependence of TH measurements see [32]. 7. The system needs approximately 45 min to adjust to the set temperature. 8. The tubes should be inserted approximately 2/3 into the ethanol. 9. Set the tip of the pipette on the bottom of the sample tube, release the sample and quickly withdraw the pipette before the sample freezes. 10. Dry the glass carefully but do not hesitate to quickly set the tube into the recrystallometer in order to avoid condensation. Water on the glass surface may freeze the tube to the recrystallometer, causing errors in measurement and breaking of the tube. 11. In order to obtain ice with a flat surface, any disturbance or shock during the freezing process should be avoided. 12. When attaching the ice on the chilled glass slide, small drops of water can help to bind the ice to the cooled glass. After the attachment of the ice on the slide, residual water on the glass slide should be removed using a precooled clean tissue. 13. Figure 7 shows these data collected from 200 consecutive heating–cooling cycles on a sample which was cooled at 0.018 K/s. We call this type of plot a “Manhattan” and it neatly demonstrates the stochastic nature of the nucleation process. The lack of slope of the Manhattan indicates that the sample has not changed during these many hours of recording freeze–thaw cycles [38].

300

Maddalena Bayer-Giraldi et al.

Fig. 9 Survival curve for an ALTA sample set showing the spread of nucleation temperatures

14. The natural definition of the supercooling point (SCP) is the temperature at which the survival curve crosses the 50% unfrozen mark, called here a T50. For the data shown in Fig. 9, the proposed SCP, or T50, is 8.17 K below the melting point. Adding substances to the sample tube can shift the S-curve to higher T50s if they enhance nucleation, or to colder regimes if they inhibit nucleation, probably by masking potential nucleation sites inside the existing sample of liquid and/or container. To determine if sea-ice microalgae or their IBPs have any effect on the nucleation temperature, the T50 must first be determined for a sample volume of the culture medium, in the container in which they will be measured [39]. The level of supercooling afforded by the osmolality of the solution is twice the melting point depression [40], that is, if water in a given pan has a T50 of
20  C, then sea water in the same pan will have a T50 that is (2.0 
1.9) ¼
3.8  C lower, that is,
23.8  C. If the T50 of a given pan with fresh culture medium is
24  C and if the T50 of the spent is
19  C then the spent medium contains substances that enhanced nucleation. On the contrary, if it is
28  C it contains substances which are helping to inhibit nucleation. It is also critical that the addition of diatoms does not cause any slope to the Manhattan, which would mean that the sample is changing, almost certainly due to breaking up of the diatoms into smaller pieces with rougher edges or in some way with a better ability to nucleate ice.

Microalgal Ice Binding Proteins

301

References 1. Weeks WF, Ackley SF (1982) The growth, structure, and properties of sea ice. In: CRREL monograph. U.S. Army, Hanover, NH, p 82 2. Maykut GA (1986) The surface heat and mass balance. In: Untersteiner N (ed) NATO ASI series. Plenum Press, New York, NY, pp 396–463 3. Eicken H (1992) The role of sea ice in structuring Antarctic ecosystems. Polar Biol 12:3–13 4. Cox GFN, Weeks WF (1983) Equation for determining the gas and brine volumes in sea-ice samples. J Glaciol 29:306–316 5. Lizotte MP (2001) The contributions of sea ice algae to Antarctic marine primary production. Am Zool 41:57–73 6. Bar-Dolev M, Braslavsky I, Davies PL (2016) Ice-binding proteins and their function. Annu Rev Biochem 85:515–542 7. Vance TDR, Bayer-Giraldi M, Davies PL et al (2019) Ice-binding proteins and the ‘domain of unknown function’ 3494 family. FEBS J 286:855–873 8. Hanada Y, Nishimiya Y, Miura A et al (2014) Hyperactive antifreeze protein from an Antarctic sea ice bacterium Colwellia sp. has a compound ice-binding site without repetitive sequences. FEBS J 281:3576–3590 9. Mangiagalli M, Bar-Dolev M, Tedesco P et al (2017) Cryo-protective effect of an ice-binding protein derived from Antarctic bacteria. FEBS J 284:163–177 10. Vance TDR, Graham LA, Davies PL (2018) An ice-binding and tandem β-sandwich domaincontaining protein in Shewanella frigidimarina is a potential new type of ice adhesin. FEBS J 285:1511–1527 11. Janech MG, Krell A, Mock T et al (2006) Ice-binding proteins from sea ice diatoms (Bacillariophyceae). J Phycol 42:410–416 12. Kim M, Gwak Y, Jung WS et al (2017) Identification and characterization of an isoform antifreeze protein from the Antarctic marine diatom, Chaetoceros neogracile and suggestion of the core region. Mar Drugs 15:318–332 13. Bayer-Giraldi M, Sazaki G, Nagashima K et al (2018) Growth suppression of ice crystal basal face in the presence of a moderate ice-binding protein does not confer hyperactivity. Proc Natl Acad Sci U S A 115:7479–7484 14. Xiao N, Hanada Y, Seki H et al (2014) Annealing condition influences thermal hysteresis of

fungal type ice-binding proteins. Cryobiology 68:159–161 15. Hashim NHF, Sulaiman S, Bakar FDA et al (2014) Molecular cloning, expression and characterisation of Afp4, an antifreeze protein from Glaciozyma antarctica. Polar Biol 37:1495–1505 16. Cheng J, Hanada Y, Miura A et al (2016) Hydrophobic ice-binding sites confer hyperactivity of an antifreeze protein from a snow mold fungus. Biochem J 473:4011–4026 17. Raymond JA, deVries AL (1977) Adsorption inhibition as a mechanism of freezing resistance in polar fishes. Proc Natl Acad Sci U S A 74:2589–2593 18. Kristiansen E, Zachariassen KE (2005) The mechanism by which fish antifreeze proteins cause thermal hysteresis. Cryobiology 51:262–280 19. Zepeda S, Yokoyama E, Uda Y et al (2008) In situ observation of antifreeze glycoprotein kinetics at the ice interface reveals a two-step reversible adsorption mechanism. Cryst Growth Des 8:3666–3672 20. Mochizuki K, Molinero V (2018) Antifreeze glycoproteins bind reversibly to ice via hydrophobic groups. J Am Chem Soc 140:4803–4811 21. Celik Y, Drori R, Pertaya-Braun N et al (2013) Microfluidic experiments reveal that antifreeze proteins bound to ice crystals suffice to prevent their growth. Proc Natl Acad Sci U S A 110:1309–1314 22. Hudait A, Odendahl N, Qiu Y et al (2018) Ice-nucleating and antifreeze proteins recognize ice through a diversity of anchored clathrate and ice-like motifs. J Am Chem Soc 140:4905–4912 23. Kondo H, Mochizuki K, Bayer-Giraldi M (2018) Multiple binding modes of a moderate ice-binding protein from a polar microalga. Phys Chem Chem Phys 20:25295–25303 24. Bayer-Giraldi M, Weikusat I, Besir H et al (2011) Characterization of an antifreeze protein from the polar diatom Fragilariopsis cylindrus and its relevance in sea ice. Cryobiology 63:210–219 25. Raymond JA, Sullivan CW, deVries AL (1994) Release of an ice-active substance by Antarctic sea ice diatoms. Polar Biol 14:71–75 26. Guo S, Stevens CA, Vance TDR et al (2017) Structure of a 1.5-MDa adhesin that binds its

302

Maddalena Bayer-Giraldi et al.

Antarctic bacterium to diatoms and ice. Sci Adv 3:e1701440 27. Do H, Lee JH, Lee SG et al (2012) Crystallization and preliminary X-ray crystallographic analysis of an ice-binding protein (FfIBP) from Flavobacterium frigoris PS1. Acta Crystallogr Sect F: Struct Biol Cryst Commun 68:806–809 28. Mangiagalli M, Sarusi G, Kaleda A et al (2018) Structure of a bacterial ice binding protein with two faces of interaction with ice. FEBS J 285:1653–1666 29. deVries AL, Wohlschlag DE (1969) Freezing resistance in some Antarctic fishes. Science 163:1073–1075 30. Knight C (2000) Adding to the antifreeze agenda. Nature 406:249–250 31. Celik Y, Graham LA, Mok Y-F et al (2010) Superheating of ice crystals in antifreeze protein solutions. Proc Natl Acad Sci U S A 107:5423–5428 32. Braslavsky I, Drori R (2013) LabVIEWoperated novel nanoliter osmometer for ice binding protein investigations. J Vis Exp 72:4189 33. Knight CA, Wierzbicki A (2001) Adsorption of biomolecules to ice and their effects upon ice growth. 2. A discussion of the basic mechanism of “antifreeze” phenomena. Cryst Growth Des 1:439–446

34. Raymond J, Wilson P, de Vries AL (1989) Inhibition of growth of nonbasal planes in ice by fish antifreezes. Proc Natl Acad Sci U S A 86:881–885 35. Raymond JA, Janech MG, Fritsen C (2009) Novel ice-binding proteins from a psychrophilic Antarctic alga (Chlamydomonadaceae, Chlorophyceae). J Phycol 45:130–136 36. Barlow TW, Haymet ADJ (1995) ALTA: an automated lag-time apparatus for studying nucleation of supercooled liquids. Rev Sci Instrum 66:2996–3007 37. Wilson PW, Lu W, Xu H et al (2012) Inhibition of ice nucleation by slippery liquid-infused porous surfaces (SLIPS). Phys Chem Chem Phys 15:581–585 38. Wilson PW, Heneghan AF, Haymet ADJ (2003) Ice nucleation in Nature: supercooling point measurement and the role of heterogeneous nucleation. Cryobiology 46:88–98 39. Wilson PW, Osterday KE, Heneghan AF et al (2010) Type I antifreeze proteins enhance ice nucleation above certain concentrations. J Biol Chem 285:34741–34745 40. Wilson PW, Haymet ADJ (2009) Effect of solutes on the heterogeneous nucleation temperature of supercooled water: an experimental determination. Phys Chem Chem Phys 11:2679–2682

Chapter 21 Isolation and Characterization of Ice-Binding Proteins from Higher Plants Melissa Bredow, Heather E. Tomalty, Laurie A. Graham, Audrey K. Gruneberg, Adam J. Middleton, Barbara Vanderbeld, Peter L. Davies, and Virginia K. Walker Abstract The characterization of ice-binding proteins (IBPs) from plants can involve many techniques, a few of which are presented here. Chief among these methods are tests for ice recrystallization inhibition, an activity characteristic of plant IBPs. Two related procedures are described, both of which can be used to demonstrate and quantify ice-binding activity. First, is the traditional “splat” assay, which can easily be set up using common laboratory equipment, and second, is our modification of this method using superhydrophobic coated sapphire for analysis of multiple samples in tandem. Thermal hysteresis is described as another method for quantifying ice-binding activity, during which ice crystal morphology observations can be used to provide clues about ice-plane binding. Once ice-binding activity has been evaluated, it is necessary to verify IBP identity. We detail two methods for enriching IBPs from complex mixtures using ice-affinity purification, the “ice-finger” and “ice-shell” methods, and we highlight their advantages and limitations for the isolation of plant IBPs. Recombinant IBP expression, necessary for detailed ice-binding analysis, can present challenges. Here, a strategy for recovery of soluble, active protein is described. Lastly, verification of function in planta borrows from standard protocols, but with an additional screen applicable to IBPs. Together, these methods, and a few considerations critical to success, can be used to assist researchers wishing to isolate and characterize IBPs from plants. Key words Ice-binding proteins, Antifreeze proteins, Ice-recrystallization inhibition, Thermal hysteresis, Ice crystals, Ice-affinity purification, Recombinant protein purification, Transgenic IBP expression

1

Introduction Ice binding proteins (IBPs) can be classified into four groups based upon their biological roles [1–3]. The first to be identified were called antifreeze proteins (AFPs) and their role is to inhibit ice growth, thereby allowing the organism to resist freezing by supercooling. In organisms that do freeze, the second group serves to prevent ice recrystallization, which is the growth of large ice crystals

Dirk K. Hincha and Ellen Zuther (eds.), Plant Cold Acclimation: Methods and Protocols, Methods in Molecular Biology, vol. 2156, https://doi.org/10.1007/978-1-0716-0660-5_21, © Springer Science+Business Media, LLC, part of Springer Nature 2020

303

304

Melissa Bredow et al.

at the expense of small ones in the frozen state. These IBPs are more properly called ice recrystallization inhibition (IRI) proteins. The next two functions have been found in single-celled organisms living in lakes and oceans. Secreted IBPs can be used to maintain open pores in ice (ice structuring proteins) and membrane-bound IBPs can be used to tether the organism to ice (ice adhesins). IBPs characterized from higher plants most commonly fall into the IRI category, although a few cases of plants producing AFPs have also been reported (for a recent review of plant IBPs see [4]). As judged by the literature, the isolation and characterization of IBPs from higher plants has lagged behind their analysis in other organisms. Many of the best characterized IBPs are those from freeze-intolerant fish and insects [5]. These, in contrast to many plant IBPs described to date, show higher thermal hysteresis (TH) activity, defined as the depression of the freezing point relative to the melting point. The strategy of most plants is to survive sub-zero temperatures by allowing freezing without substantial supercooling. Freezing is initiated in vascular tissues and the apoplast [6]. Presumably, the advance of ice crystal spicules coming from the extracellular space makes membranes vulnerable to damage, which could be controlled by IBPs. At temperatures close to the melting point, IBPs inhibit the typical growth progression of large ice crystals at the expense of smaller ones. This phenomenon is referred to as Ostwald ripening or ice recrystallization [7]. The observation that grass had comparably higher IRI activity than previously idetified “AFPs” helped spur research into IBPs derived from higher plants [8]. Because of their ability to better control ice recrystallization, IBPs produced by plants are optimal candidates for engineering improved freeze tolerance [9], which has been successfully achieved in Arabidopsis thaliana [10, 11] and Solanum lycopersicum (tomato) [12]. Ice recrystallization is also a major challenge in the transport and storage of frozen foods, and industry has turned to the high IRI found in plants for a solution to this problem [13–16]. As a consequence of commercial interest, it is likely that much of the research on plant IBPs has not been released. Even a quick appreciation of the widespread distribution of IBPs in higher plants can best be obtained by perusing the patent literature. References can be accessed at: tinyurl.com/k5bmebd, tinyurl.com/m2olwo3, tinyurl.com/kkvyrk9, tinyurl.com/jw8omup, tinyurl.com/kza2pky, as well as the academic literature (e.g., [14, 17–32] among others). As these citations indicate, IBPs have been reported in a number of gymnosperms (e.g., spruce and Ginkgo) and angiosperms (e.g., brassicas, carrot, onion, asters, plantain, potato, dandelion, nightshade, maple, legumes, cotton-wood, privet, sea buckthorn, and oak), as well as a variety of monocots, mostly grasses and cereals (e.g., rye grass, fescue, bamboo, brome, bentgrass, daylily, triticale, spring oats, barley, and rye). Our goal in this paper is to present the following methods designed to be useful for researchers moving

Characterization of Ice Binding Proteins from Higher Plants

305

Fig. 1 Assembly showing the equipment for flash freezing a thin layer of sample solution necessary for ice recrystallization inhibition assays (splat method)

from “prospecting” for new plant IBPs to expression of these proteins in planta.

2 2.1

Materials IRI (Splat Assay)

1. Buffer: 25 mM Tris–HCl (pH 7.8) or phosphate buffered saline (PBS), 150 mM NaCl, 0.01 mg/mL bovine serum albumin. 2. Tube with a diameter of at least 5 cm and a minimum length of 1 m (e.g., an empty glass chromatography column or a length of plastic plumbing pipe) (Fig. 1a). 3. Retort stand and carpenter’s level or plumb line. 4. Metal block (e.g., an inverted heating block). 5. Polystyrene container (e.g., an insulated shipping box). 6. Dry ice. 7. Glass microscope cover slides. 8. A low-temperature circulating water bath (e.g., 7 L SpaceSaving, Refrigerated Circulator with MX Controller,
20  C, VWR International) filled with clear ethylene glycol. 9. A double-walled glass chamber attached to the circulating bath (Fig. 2a, b). 10. Polystyrene insulation cut and glued to surround the glass chamber, with hole(s) to allow for illumination from below, plus foam insulation around the tubing.

306

Melissa Bredow et al.

Fig. 2 System for examination of ice crystals to evaulate IRI using the “splat” or superhydrophobic coated sapphire assay. (a) Photograph of the complete setup with microscope, insulated chamber and hoses, and light supply. Imaging is done through a glass plate sealed to the rim of the chamber with vacuum grease. (b) Schematic showing microscope lens (1) with polarizing filter (2) above a custom-made double-walled glass chamber (3) through which the cooling fluid circulates (4). The inside of the chamber (5) contains a second polarizing filter, some desiccant beads, the organic liquid, and a slide or coverslip containing the sample. Illumination must be from below (6). (c) Image of the chamber from above with a sapphire slide containing ten samples in place to be viewed through crossed polarizers. Supports can be glued onto a glass polarizer with epoxy to hold slides in position

11. A 4–5 mm thick glass plate large enough to cover and seal the top of the glass chamber using vacuum grease. 12. Light source (e.g., Fiber-Lite Illumination System 181-1 with gooseneck fiber optic assembly, Dolan-Jenner). 13. Organic liquid such as hexane or isooctane (2,2,4-trimethyl pentane) for use inside the chamber (see Note 1). 14. Desiccation beads (see Note 2). 15. Stereomicroscope with a camera attachment and polarizing filter. 16. Second polarizer (e.g., glass linear polarizer, 20 mm diameter, Edmund Optical, which is robust enough to withstand the solvent used in the bowl). 17. Any camera capable of capturing images in low light with a remote trigger to minimize vibration.

Characterization of Ice Binding Proteins from Higher Plants

2.2 IRI (Superhydrophobic Coated Sapphire)

307

1. Items 1, 4, 8–17 as listed in Subheading 2.1. 2. Oriented sapphire slides, one with traps in a superhydrophobic coating and the other with a slippery coating (see Note 3). 3. Binder clips. 4. Liquid nitrogen. 5. Long tweezers. 6. Scalpel blade.

2.3 TH and Ice Crystal Morphology

1. Nanoliter osmometer, cooling stage, sample holder with wells, and a glass coverslip. A sandwich of two coverslips can help prevent condensation that can interfere with viewing. A ring of glue from a hot-glue gun is applied to a coverslip, a piece of Drierite or a silica desiccant bead is affixed to the inside edge of the glue, and coverslip is added to the top, creating an airtight pocket. 2. Microscope with illumination from below, 400–500 magnification and extra-long working distance objectives above to accommodate the distance to the cooling block. An upright brightfield Nikon Eclipse 80i with a camera port is suitable. 3. Glass micropipettes and a rubber tube attached to a mouthpiece. 4. Immersion oil, three small vials with chloroform, water, and 95% ethanol, respectively. 5. Peristaltic pump to circulate water to lower the temperature of the cooling block. 6. Air pump (aquarium pumps are sufficient) with air passed through a W.A. Hammond Drierite Gas Purifier or a 50 mL conical tube containing Drierite with air coming in to a tube travelling from a hole in the lid to the bottom of the tube and leaving from a similar tube wrapped in a soft paper tissue at the top of the tube. 7. Sample of protein or a plant extract at a suitable concentration to allow for measurement of TH (see Note 4).

2.4 Insoluble Recombinant Plant IBPs

1. A putative open reading frame from an IBP cloned into an expression vector (such as pET24a) and transformed into T7 RNA polymerase–compatible cells. Stop codons must be excluded from the 30 end and the sequence will need to be cloned into the appropriate restriction sites to incorporate the polyhistidine tag (6 His; e.g., NdeI and XhoI for pET24a). 2. Lysogeny broth (LB): 10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl, pH adjusted to 7.5 with NaOH. Autoclave and allow to cool. Add appropriate amount of antibiotic just prior

308

Melissa Bredow et al.

to inoculation (e.g., for pET24a use filter-sterilized 50 μg/mL kanamycin). 3. Filter-sterilized 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG: dissolved in 10 mL of dH2O) prepared fresh before use (see Note 5). 4. 8 M urea stock solution: deionize by incubating with mixed resin ion-exchange beads (5 g/100 mL of sample) for 2 h at room temperature on a shaker (100 rpm). Beads are subsequently removed using a vacuum filter and the solution is degassed by placing the liquid in a sidearm flask while applying a vacuum. Adjust to pH 8.0 (see Note 6). 5. Buffers will vary depending on the properties of the IBP. For purification under denaturing conditions buffers should be prepared fresh on the day of use (see Note 7). Buffer A (Lysis Buffer): 10 mM Tris–HCl, 100 mM NaCl, pH 8.0. Buffer B (Denaturing Buffer): 8 M urea, 100 mM NaH2PO4, 10 mM Tris–HCl, 10 mM β-mercaptoethanol, 2% glycerol (v/v), pH 8.0. Buffer C (Wash Buffer): 8 M urea, 100 mM NaH2PO4, 10 mM Tris–HCl, 10 mM β-mercaptoethanol, pH 6.3. Buffers D (Elution Buffer): 8 M urea, 100 mM NaH2PO4, 10 mM Tris–HCl, 10 mM β-mercaptoethanol, pH 5.9. Buffer E (Elution Buffer): 8 M urea, 100 mM NaH2PO4, 10 mM Tris–HCl, 10 mM β-mercaptoethanol, pH 4.5. Buffers F–I (Dialysis Buffers): Four separate buffers containing either 6 M urea (F), 4 M urea (G), 2 M urea (H) or no urea (I) plus 100 mM NaCl, 15 mM Tris–HCl, 2 mM β-mercaptoethanol, 2% glycerol (v/v), pH 8.0. 6. Mini EDTA-free protease inhibitor cocktail tablets (e.g., cOmplete Protease Inhibitor Cocktail tablets from Sigma-Aldrich). 7. Centrifuge bottles. 8. Ni-NTA beads and gravity flow column. 9. Dialysis tubing. 10. Incubator shakers set at 37 and 24  C. 2.5 Ice Affinity Purification (Ice Finger)

1. Programmable circulating water bath filled with ethylene glycol or antifreeze (plumbing or automotive). Attach the brass cooled probe to the inlet and outlet ports of the bath (Fig. 3). 2. Stir plate and small stir bar (~15 mm  7 mm). 3. Lab jack. 4. Insulated container to fit a 100–500 mL beaker.

Characterization of Ice Binding Proteins from Higher Plants

309

Fig. 3 Apparatus for the purification of IBPs by ice affinity (ice-finger)

Fig. 4 Ice-shell purification apparatus. A round-bottom flask is attached to a rotating motor and partially submerged into a subzero ethylene glycol bath (a). The flask is rotated in the bath until ~50% of the liquid is incorporated into the ice shell (b) 2.6 Ice Affinity Purification (Ice Shell)

1. Programmable circulating water bath filled with ethylene glycol or antifreeze (plumbing or automotive) with an opening large enough to accommodate up to a 1 L round bottom flask (Fig. 4a). 2. Round bottom flasks of various sizes from 100 mL to 1 L.

310

Melissa Bredow et al.

3. Variable speed motor to rotate the flask, such as those that come with a rotary evaporator. 4. Appropriately sized plastic clamps to affix flasks to the rotary shaft. 5. Polystyrene box and ethanol precooled in
80  C freezer. 6. Millipore water precooled in refrigerator. 7. Graduated cylinders. 2.7 Verification in Planta

1. A stationary phase liquid culture of a suitable Agrobacterium strain (e.g., GV3101, LBA4404, EHA105) confirmed to be carrying a binary vector (e.g., pCambia_1305). This must include the desired promoter (e.g., a native promoter sequence, the constitutive promoter CaMV 35S or an inducible promoter such as the plant-specific cold-inducible pOsMYB1R35 [33]) followed by the IBP-encoding DNA sequence of interest. 2. Several 10-cm pots containing 4–5 healthy and well-watered 3to 4-week-old Arabidopsis plants grown under long-day conditions (12–16 h light). Plus additional pots and soil for subsequent steps. 3. A growth chamber or greenhouse set to appropriate conditions (e.g., 150 μE/m2/s light intensity on a 16 h:8 h light–dark cycle at 22  C and 70% relative humidity). 4. LB: 10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl. Autoclave 300 mL in a 1 L flask. Add appropriate antibiotic once solution has cooled (e.g., 50 μg/mL kanamycin for pCambia_1305). 5. Sterile dH2O (autoclave 150-mL aliquots). 6. Inoculation medium: 1.2 g Murashige and Skoog (MS) basal salts, 25 g sucrose in 500 mL dH2O, pH 7.0. Prepare fresh or autoclave in advance. 0.5 μL of a 10 mg/mL stock of benzylamino purine in dimethyl sulfoxide (DMSO) and 90 μL Silwet L-77 (Lehle Seeds, Texas) or other suitable wetting agent are added to the inoculation medium containing Agrobacterium in a shallow container (at least 12.5 cm long and wide and ideally ~5 to 5.5 cm deep) just prior to plant inoculation. 7. Incubator shaker set at 28–30  C. 8. Centrifuge bottles and a centrifuge. 9. Metal sieve with very fine mesh (e.g., as would be used for sifting flour). 10. Small tubes (e.g., microcentrifuge tubes) or bags for seed collection. 11. Sealable glass vessel (e.g., desiccator jar) located in a fume hood. This may require vacuum grease to make a complete seal.

Characterization of Ice Binding Proteins from Higher Plants

311

12. Glass beaker containing 100 mL of household bleach. 13. Concentrated HCl. 14. Transfer pipet. 15. MS-agar plates: 0.5  MS salts (pH 5.8), 8 g/L agar. Autoclave and, in a laminar flow hood, add appropriate antibiotic (e.g., 40 μg/mL hygromycin for pCambia1305) once solution has cooled to ~55  C, just prior to pouring. 16. Laminar flow hood. 17. Micropore surgical tape (3M) and fine forceps. 18. Liquid nitrogen. 19. Mortar and pestle. 20. Extraction buffer: 10 mM Tris–HCl (pH 7.0), 25 mM NaCl. 21. Heating block.

3 3.1

Methods IRI (Splat Assay)

Surveys of plants using a variety of tissues (leaves, stems, roots etc.) can be efficiently accomplished using the “splat” method for the evaluation of IRI. Normally, plant material is collected following a cold acclimation period constituting low temperature exposure (~2 to 10  C), sometimes in combination with shortened photoperiod (6–8 h light). The duration of cold acclimation is species-specific and will need to be tested experimentally. Samples are flash frozen, ground in liquid nitrogen using a mortar and pestle, and suitable buffers are added (e.g., [8, 34], see below). After brief centrifugation (12,000  g) the extract supernatants can be assayed for IRI. This method can also be used for putative IBPs produced recombinantly in Escherichia coli. Rough estimates of IRI activity after the annealing period can be made by determining mean ice crystal diameters or edge lengths, but this must be done within an appropriate dilution range such that crystals are large enough to measure with computer software. Semiquantitative measures of IRI can also be accomplished using a protein dilution series to determine the concentration at which the protein is no longer capable of preventing recrystallization (IR end-point). Importantly, protein samples must be prepared identically with the inclusion of both positive and negative controls in order to ensure reproducibility and to allow for comparisons between samples. 1. Turn on the cooling bath and set to an appropriate subzero temperature such that the temperature within the chamber is between
4 and
6  C. Hexane or isooctane should be added to a depth of ~2 cm and the polarizer should be placed in the bottom of the chamber. Approximately ten desiccation

312

Melissa Bredow et al.

beads should be added around the edge of the polarizer to ensure they do not contact the coverslips. 2. Store a metal block in a
80  C freezer. Place the metal block into a short polystyrene container, surround it with dry ice keeping the top surface free and place on a retort stand. Attach the tube to the retort stand so that it hangs slightly above the metal block (Fig. 1). Ensure that the tube is perpendicular to the ground using a carpenter’s level or a plumb line. 3. Prepare samples for analysis by making 1:10 dilutions of the plant extract or purified protein (100 μL) (see Note 8) and a positive and negative control (see Note 9). 4. Label the corner of a glass coverslip for each dilution and controls. 5. Starting with the negative control, place the coverslip directly underneath the tube and allow the slide temperature to equilibrate. Using an automatic pipette, sample a drop (10 μL) of the diluted extract. Carefully dispense the drop of solution at the top centre of the tube (see Notes 10 and 11). Ensure that the sample has dispersed on the center of the coverslip. 6. Quickly move the slide into the incubation chamber. 7. The polarizer on the microscope should be rotated until the background appears black. If the circulating bath induces vibration (the model listed above does not) it should be turned off briefly while the picture is taken at an appropriate and consistent magnification (e.g., with a field of view spanning ~3 mm). 8. Repeat this procedure for all samples. 9. Once all samples have been photographed, create a tight seal between the glass chamber and a plastic lid with vacuum grease to ensure that the solvent does not evaporate and air moisture does not result in the formation of ice on coverslips which can obscure the visualization of ice crystals. 10. After 16 h photograph slides and compare ice crystal size with earlier time point (time 0) to visually assess the degree of IRI (Fig. 5) (see Note 12). 3.2 IRI (Superhydrophobic Coated Sapphire)

The main advantage of the splat assay is that it requires very little in the way of specialized equipment and it is relatively simple to purchase and set up a basic system. However, its main drawback is that a separate coverslip must be used for each sample. As the ice forms a thin exposed layer on the surface of the glass, coverslips should not overlap. This limits the number of samples that can be safely analyzed at one time to around four, in the apparatus shown in Fig. 2c. To circumvent this, an alternative that employs sapphire slides (chips) with ten wells (traps) per slide has been developed. Each slide is 25 mm  35 mm, which is only slightly larger than a

Characterization of Ice Binding Proteins from Higher Plants

313

Fig. 5 Ice recrystallization inhibition by recombinant B. distachyon IBP (BdIRI). Samples include 50 mM Tris–HCl buffer (pH 7.8) and recombinant BdIRI (brome grass; 1 mg/mL). Images show the ice crystals at the beginning of the experiment (0 h) and after annealing for 18 h at
4  C (see Note 13)

standard square coverslip (22 mm  22 mm). As two of these chips fit easily into the chamber, up to 20 samples can be assayed at once. The chip offers another advantage as less sample volume is used (~1 μL vs. 10 μL). Additionally, as the sample is pipetted directly onto the chip, rather than being dropped from a height, preparation time is reduced and sample wastage is eliminated. It is also easier to keep track of the samples as, for example, successive dilutions can be loaded, in order, on a single chip. 1. Turn on the cooling bath at set to an appropriate subzero temperature such that the temperature within the chamber is
4 to
6  C. Hexane or isooctane should be added to a depth of ~2 cm and the polarizer should be placed in the bottom of the chamber. Around ten desiccation beads should be added around the edge of the polarizer to ensure they do not contact the chips. 2. Prepare samples by tenfold serial dilutions (for range finding) or twofold serial dilutions (for an accurate endpoint) in IRI buffer. 3. Place the chip with the traps on the top surface of the metal block with the narrow side facing up. The block can be placed in ice to keep the chip cold. 4. Pipette 0.8 μL of sample into the center of each trap.

314

Melissa Bredow et al.

5. Carefully slide the chip off the block and hold the long edges between thumb and finger. Place the top chip just above but not touching the bottom chip, again between thumb and finger. Using the other hand, quickly press the two chips together and clamp the long sides using small binder clips. 6. Drop the assembly into freshly decanted liquid nitrogen. Freezing is not optimal if the liquid nitrogen contains small ice crystals formed from moisture absorbed from the air. 7. Use tweezers to remove the chip from the liquid nitrogen, remove the binder clips immediately and drop the sandwiched chips into the precooled (
6  C) chamber with the slippery chip on the top. Carefully pry off the top chip, using a scalpel blade. 8. Use vacuum grease to seal the glass plate to the top of the chamber and then photograph each trap between crossed polarizers at a magnification where the trap fills the field of view. If the depth of field of the system is not sufficient to accommodate the glass plate, it can be applied following photography. The cooling bath mentioned earlier circulates fluid slowly and does not induce vibration in the chamber, but some baths do cause vibration and these must be turned off for a short period (