Receptor Signal Transduction Protocols, 3rd Edition (Methods in Molecular Biology Vol 746)

Methods

in

Molecular Biology™

Series Editor John M. Walker School of Life Sciences University of Hertfordshire Hatfield, Hertfordshire, AL10 9AB, UK



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

Receptor Signal Transduction Protocols Third Edition Edited by

Gary B. Willars and R.A. John Challiss Department of Cell Physiology and Pharmacology, University of Leicester, Leicester, UK

Editors Gary B. Willars Department of Cell Physiology and Pharmacology University of Leicester Leicester, UK [email protected]

R.A. John Challiss Department of Cell Physiology and Pharmacology University of Leicester Leicester, UK [email protected]

ISSN 1064-3745 e-ISSN 1940-6029 ISBN 978-1-61779-125-3 e-ISBN 978-1-61779-126-0 DOI 10.1007/978-1-61779-126-0 Springer New York Dordrecht Heidelberg London Library of Congress Control Number: 2011928685 © Springer Science+Business Media, LLC 2011 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Humana Press, c/o Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or ­dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. While the advice and information in this book are believed to be true and accurate at the date of going to press, neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may be made. The publisher makes no warranty, express or implied, with respect to the material contained herein. Cover image: Confocal image showing membrane localisation of fluorescently labelled neuromedin U-8 (NmU-8; red) bound to HEK293 cells with stable expression of recombinant human neuromedin U type 2 receptors (NMU2) containing a C-terminal, enhanced green fluorescent protein epitope tag (green). Cells were incubated at 37°C for 60 min in the presence of 10 nM fluorescent ligand. Image shows membrane-localised receptor (green) and ligand (red), some internalised receptor and ligand along with some colocalisation (yellow). NmU-8, N-terminally labelled with Cy3B (GE Healthcare, Bucks, UK) was kindly provided by J. Scott and M. Ruediger (GlaxoSmithKline, Harlow, UK). Image: Khaled Al-Hosaini and Gary Willars (University of Leicester, UK). Printed on acid-free paper Humana Press is part of Springer Science+Business Media (www.springer.com)

Preface As our understanding of G protein-coupled receptor (GPCR) signal transduction continues to grow, we cannot help but be struck by the emerging complexity and the ability of this receptor superfamily to continually surprise us as new facets are discovered. At the level of the GPCR itself, we are only now beginning to glimpse how receptors might behave in vivo and many fundamental questions still lack definitive answers. For example, are all GPCRs functionally dimeric and how informative are the available crystal structures with respect to the true conformational repertoires of GPCRs? At the post-receptor level, the diversity of signal transduction mechanisms regulated by GPCRs continues to expand. This, of course, includes G protein-independent signaling, leading some to question the appropriateness of the term “GPCR” and suggesting “7TM receptor” may be a better term, dissociating the receptor from an automatic assumption of G protein association and activation. Indeed, how important is G protein-independent GPCR signaling within the cell? Does such signaling extend beyond the recruitment of b-arrestin and the formation of molecular signaling scaffolds? From a pharmacological perspective, much in the GPCR landscape has changed in recent years. The concepts of inverse agonism, protean agonism, and functional selectivity have all required reappraisal of the simple receptor-G proteineffector hierarchy that we have previously viewed in linear terms. Perhaps the greatest change in GPCR research has been the rapid assimilation of the concept of the receptor as allosterically regulated switches allowing (pharmacological) modulation of ligand affinity and the efficacy of receptor-coupling to signaling pathways. In preparing this third edition of Receptor Signal Transduction Protocols, we have attempted to be mindful of the constant evolution of the GPCR field and to deal with methods that allow researchers to address many of these important issues. This has involved the thorough revision of some core chapters, a complete rewriting of others to encompass new technological developments since the publication of the first and second editions, and the commissioning of chapters to expand on previous coverage. We, therefore, see the current edition very much as a companion to previous editions. We are enormously grateful to all our authors, new and old, for generally living with our deadlines and providing excellent and comprehensive methods, as well as the essential “tricks of the trade” that are often needed to troubleshoot new techniques. Finally, we thank the Series Editor, John Walker, for again giving us this opportunity to complement previous editions with a new volume in the Methods in Molecular Biology series. Leicester, UK Leicester, UK

Gary B. Willars R.A. John Challiss

v

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

v ix

Part I  Developing Models for Studying GPCRs   1 Inducible Expression of G Protein-Coupled Receptors in Transfected Cells . . . . . 3 Beryl Koener and Emmanuel Hermans   2 Using the Flp-In™ T-Rex™ System to Regulate GPCR Expression . . . . . . . . . . . . 21 Richard J. Ward, Elisa Alvarez-Curto, and Graeme Milligan   3 Viral Infection for GPCR Expression in Eukaryotic Cells . . . . . . . . . . . . . . . . . . . 39 Antonio Porcellini, Luisa Iacovelli, and Antonio De Blasi   4 Generation of Epitope-Tagged GPCRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 Yan Huang and Gary B. Willars   5 The Use of Site-Directed Mutagenesis to Study GPCRs . . . . . . . . . . . . . . . . . . . . 85 Alex C. Conner, James Barwell, David R. Poyner, and Mark Wheatley   6 Approaches to Study GPCR Regulation in Native Systems . . . . . . . . . . . . . . . . . . 99 Jonathon M. Willets   7 Heterologous Expression of GPCRs in Fission Yeast . . . . . . . . . . . . . . . . . . . . . . 113 John Davey and Graham Ladds

Part II Examining GPCR Expression and Agonist-Induced Covalent Modifications   8 Radioligand Binding Methods for Membrane Preparations and Intact Cells . . . . . David B. Bylund and Myron L. Toews   9 Quantification of GPCR mRNA Using Real-Time RT-PCR . . . . . . . . . . . . . . . . . Trond Brattelid and Finn Olav Levy 10 Determining Allosteric Modulator Mechanism of Action: Integration of Radioligand Binding and Functional Assay Data . . . . . . . . . . . . . . Christopher J. Langmead 11 Design and Use of Fluorescent Ligands to Study Ligand–Receptor Interactions in Single Living Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stephen J. Briddon, Barrie Kellam, and Stephen J. Hill 12 Examining Site-Specific GPCR Phosphorylation . . . . . . . . . . . . . . . . . . . . . . . . . Adrian J. Butcher, Andrew B. Tobin, and Kok Choi Kong 13 Ubiquitination of GPCRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Adriana Caballero and Adriano Marchese

vii

135 165

195

211 237 251

viii

Contents

Part III Examining Early Events in GPCR Signaling 14 [35S]GTPgS Binding as an Index of Total G-Protein and Ga-Subtype-Specific Activation by GPCRs . . . . . . . . . . . . . . . . . . . . . . . . . . Rajendra Mistry, Mark R. Dowling, and R.A. John Challiss 15 Using Calcium Imaging as a Readout of GPCR Activation . . . . . . . . . . . . . . . . . . Martin D. Bootman and H. Llewelyn Roderick 16 Measuring Spatiotemporal Dynamics of Cyclic AMP Signaling in Real-Time Using FRET-Based Biosensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . Frank Gesellchen, Alessandra Stangherlin, Nicoletta Surdo, Anna Terrin, Anna Zoccarato, and Manuela Zaccolo 17 Determining the Activation of Rho as an Index of Receptor Coupling to G12/13 Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Michio Nakaya, Mina Ohba, Motohiro Nishida, and Hitoshi Kurose 18 The Use of Translocating Fluorescent Biosensors for Real-Time Monitoring of GPCR-Mediated Signaling Events . . . . . . . . . . . . . . . . . . . . . . . . . Carl P. Nelson and R.A. John Challiss

263 277

297

317

329

Part IV Receptor–Receptor and Receptor–Protein Interactions 19 Study of GPCR–Protein Interactions Using Gel Overlay Assays and Glutathione-S-Transferase-Fusion Protein Pull-Downs . . . . . . . . . . . . . . . . . Ashley E. Brady, Yunjia Chen, Lee E. Limbird, and Qin Wang 20 Study of GPCR–Protein Interactions by BRET . . . . . . . . . . . . . . . . . . . . . . . . . . Martina Kocan and Kevin D.G. Pfleger 21 Time Resolved FRET Strategy with Fluorescent Ligands to Analyze Receptor Interactions in Native Tissues: Application to GPCR Oligomerization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Martin Cottet, Laura Albizu, Laetitia Comps-Agrar, Eric Trinquet, Jean-Philippe Pin, Bernard Mouillac, and Thierry Durroux 22 Peptide Affinity Purification for the Isolation and Identification of GPCR-Associated Protein Complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pascal Maurice, Avais M. Daulat, and Ralf Jockers 23 Tandem Affinity Purification and Identification of GPCR-Associated Protein Complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Avais M. Daulat, Pascal Maurice, and Ralf Jockers 24 Identification of GPCR Localization in Detergent Resistant Membranes . . . . . . . Ranju Kumari and Anna Francesconi 25 Analysis of GPCR Localization and Trafficking . . . . . . . . . . . . . . . . . . . . . . . . . . James N. Hislop and Mark von Zastrow

347 357

373

389

399 411 425

Part V Statistical Methods 26 Statistical Methods in Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443 Domenico Spina Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 473

Contributors Laura Albizu  •  Department of Neurology, Mount Sinai School of Medicine, New York, NY, USA Elisa Alvarez-Curto  •  Neuroscience and Molecular Pharmacology, Biomedical & Life Sciences, University of Glasgow, Glasgow, UK James Barwell  •  Clinical Sciences Research Institute, University of Warwick, Coventry, UK Martin D. Bootman  •  Laboratory of Molecular Signalling, Babraham Institute, Cambridge, UK Ashley E. Brady  •  Department of Pharmacology, Vanderbilt University Medical Center, Nashville, TN, USA Trond Brattelid  •  National Institute of Nutrition and Seafood Research, Bergen, Norway Stephen J. Briddon  •  Institute of Cell Signalling, School of Biomedical Sciences, University of Nottingham, Nottingham, UK Adrian J. Butcher  •  Department of Cell Physiology and Pharmacology and the Protein and Nucleic Acid Chemistry Laboratory, University of Leicester, Leicester, UK David B. Bylund  •  Department of Pharmacology and Experimental Neuroscience, University of Nebraska Medical Center, Omaha, NE, USA Adriana Caballero  •  Department of Molecular Pharmacology and Therapeutics, Stritch School of Medicine, Loyola University Chicago, Maywood, IL, USA R.A. John Challiss  •  Department of Cell Physiology and Pharmacology, University of Leicester, Leicester, UK Yunjia Chen  •  Department of Physiology and Biophysics, University of Alabama at Birmingham, Birmingham, AL, USA Laetitia Comps-Agrar  •  Institut de Genomique Fonctionnelle, Centre National de la Recherche Scientifique, Institut National de la Santé et de la Recherche Médicale, University of Montpellier 1 and 2, Montpellier, France Alex C. Conner  •  Warwick Medical School, University of Warwick, Coventry, UK Martin Cottet  •  Institut de Genomique Fonctionnelle, Centre National de la Recherche Scientifique, Institut National de la Santé et de la Recherche Médicale, University of Montpellier 1 and 2, Montpellier, France Avais M. Daulat  •  Institut Cochin, Université Paris Descartes, INSERM, Paris, France John Davey  •  Department of Clinical Sciences, University of Warwick, Coventry, UK Antonio de Blasi  •  Department of Experimental Medicine, University “Sapienza”, Piazzale Aldo Moro, Rome, Italy Mark R. Dowling  •  Novartis Institute for Biomedical Research, Horsham, UK Thierry Durroux  •  Institut de Genomique Fonctionnelle, Centre National de la Recherche Scientifique, Institut National de la Santé et de la Recherche Médicale, University of Montpellier 1 and 2, Montpellier, France Anna Francesconi  •  Dominick P. Purpura Department of Neuroscience, Albert Einstein College of Medicine, New York, NY, USA ix

x

Contributors

Frank Gesellchen  •  Neuroscience and Molecular Pharmacology, Biomedical & Life Sciences, University of Glasgow, Glasgow, UK Emmanuel Hermans  •  Institute of Neurosciences, Université Catholique de Louvain, Brussels, Belgium Stephen J. Hill  •  Institute of Cell Signalling, School of Biomedical Sciences, University of Nottingham, Nottingham, UK James N. Hislop  •  Department of Psychiatry, University of California, San Francisco, CA, USA Yan Huang  •  Department of Cell Physiology and Pharmacology, University of Leicester, Leicester, UK Luisa Iacovelli  •  Department of Structural and Functional Biology, Napoli Ralf Jockers  •  Institut Cochin, Université Paris Descartes, INSERM, Paris, France Barrie Kellam  •  Centre for Biomolecular Sciences, School of Pharmacy, University of Nottingham,University Park, Nottingham, UK Martina Kocan  •  Drug Discovery Biology Laboratory, Monash Institute of Pharmaceutical Sciences and Department of Pharmacology, Monash University, Melbourne, Royal Parade, Parkvile, Victoria, Australia Beryl Koener  •  Institute of Neurosciences, Université Catholique de Louvain, Brussels, Belgium Kok Choi Kong  •  Department of Cell Physiology and Pharmacology and the Protein and Nucleic Acid Chemistry Laboratory, University of Leicester, Leicester, UK Ranju Kumari  •  Dominick P. Purpura Department of Neuroscience, Albert Einstein College of Medicine, New York, NY, USA Hitoshi Kurose  •  Department of Pharmacology and Toxicology, Graduate School of Pharmaceutical Sciences, Kyushu University, Fukuoka, Japan Graham Ladds  •  Department of Clinical Sciences, University of Warwick, Coventry, UK Christopher J. Langmead  •  Heptares Therapeutics Ltd, Welwyn Garden City, UK Finn Olav Levy  •  Department of Pharmacology, Center for Heart Failure Research, University of Oslo and Oslo University Hospital, Oslo, Norway Lee E. Limbird  •  Department of Pharmacology, Vanderbilt University Medical Center, Nashville, TN, USA Adriano Marchese  •  Department of Molecular Pharmacology and Therapeutics, Stritch School of Medicine, Loyola University Chicago, Maywood, IL, USA Pascal Maurice  •  Institut Cochin, Université Paris Descartes, INSERM, Paris, France Graeme Milligan  •  College of Medical, Veterinary and Life Sciences, Institute of Neuroscience and Psychology, Molecular Pharmacology Group, University of Glasgow, Wolfson Link Building, Glasgow, Scotland, UK Rajendra Mistry  •  Department of Cell Physiology and Pharmacology, University of Leicester, Leicester, UK Bernard Mouillac  •  Centre National de la Recherche Scientifique, Institut de Genomique Fonctionnelle, Institut National de la Santé et de la Recherche Médicale, University of Montpellier 1 and 2, Montpellier, France Michio Nakaya  •  Department of Pharmacology and Toxicology, Graduate School of Pharmaceutical Sciences, Kyushu University, Fukuoka, Japan Carl P. Nelson  •  Department of Cell Physiology and Pharmacology, University of Leicester, Leicester, UK

Contributors

xi

Motohiro Nishida  •  Department of Pharmacology and Toxicology, Graduate School of Pharmaceutical Sciences, Kyushu University, Fukuoka, Japan Mina Ohba  •  Department of Pharmacology and Toxicology, Graduate School of Pharmaceutical Sciences, Kyushu University, Fukuoka, Japan Kevin D.G. Pfleger  •  Laboratory for Molecular Endocrinology – GPCRs, Western Australian Institute for Medical Research (WAIMR) and Centre for Medical Research, University of Western Australia, Crawley, Western Australia, Australia Jean-Philippe Pin  •  Institut de Genomique Fonctionnelle, Centre National de la Recherche Scientifique, Institut National de la Santé et de la Recherche Médicale, University of Montpellier 1 and 2, Montpellier, France Antonio Porcellini  •  Department of Physiology and Pharmacology “Vittorio Espèamer”, Piazzale Aldo Moro, Rome, Italy David R. Poyner  •  School of Life and Health Sciences, University of Aston, Birmingham, UK H. Llewelyn Roderick  •  Laboratory of Molecular Signalling, Babraham Institute, Cambridge, UK; Department of Pharmacology, University of Cambridge, Cambridge, UK Domenico Spina  •  The Sackler Institute of Pulmonary Pharmacology, School of Biomedical Science, King’s College London, 5th Floor Franklin Wilkins Building, SEI 9NH, Waterloo Campus, London, UK Alessandra Stangherlin  •  Neuroscience and Molecular Pharmacology, Biomedical & Life Sciences, University of Glasgow, Glasgow, UK Nicoletta Surdo  •  Neuroscience and Molecular Pharmacology, Biomedical & Life Sciences, University of Glasgow, Glasgow, UK Anna Terrin  •  Neuroscience and Molecular Pharmacology, Biomedical & Life Sciences, University of Glasgow, Glasgow, UK Andrew B. Tobin  •  Department of Cell Physiology and Pharmacology and the Protein and Nucleic Acid Chemistry Laboratory, University of Leicester, Leicester, UK Myron L. Toews  •  Department of Pharmacology and Experimental Neuroscience, University of Nebraska Medical Center, Omaha, NE, USA Eric Trinquet  •  CisBio International, Parc Technologique Marcel Boiteux, Bagnols-sur-Cèze, France Qin Wang  •  Department of Physiology and Biophysics, University of Alabama at Birmingham, Birmingham, AL, USA Richard J. Ward  •  Neuroscience and Molecular Pharmacology, Biomedical & Life Sciences, University of Glasgow, Glasgow, UK Mark Wheatley  •  School of Biosciences, University of Birmingham, Birmingham, UK Gary B. Willars  •  Department of Cell Physiology and Pharmacology, University of Leicester, Leicester, UK Jonathon M. Willets  •  Department of Cancer Studies and Molecular Medicine, University of Leicester, Leicester, UK Manuela Zaccolo  •  Institute of Neuroscience and Psychology, College of Medical Veterinary and Life Sciences, Glasgow University, Glasgow, UK Mark von Zastrow  •  Departments of Psychiatry and Cellular & Molecular Pharmacology, University of California, San Francisco, CA, USA Anna Zoccarato  •  Neuroscience and Molecular Pharmacology, University of Glasgow, Glasgow, UK

Part I Developing Models for Studying GPCRs

Chapter 1 Inducible Expression of G Protein-Coupled Receptors in Transfected Cells Beryl Koener and Emmanuel Hermans Abstract Biochemical or pharmacological studies of G protein-coupled receptors (GPCRs) are widely conducted in transfected mammalian cells. A variety of commercially available systems allow the generation of stable cell-lines in which expression of the recombinant receptor can be induced on addition of a defined chemical to the culture medium, which operates as a control switch for the transcription of the cloned sequence. Such systems offer the possibility to induce graded levels of receptor expression in the experimental model, or to induce an abrupt downregulation of receptor expression during the maintenance of the cell-line. This chapter provides an overview of the different systems available and provides methods for the generation and validation of stably transfected cell-lines expressing the GPCR of choice. Key words: Transfection, Inducible expression, Tetracycline, Receptor density

1. Introduction The molecular cloning of the vast majority of known G proteincoupled receptors (GPCRs) now permits the manipulation of the corresponding cDNA sequences in order to express these proteins recombinantly in transfected cells. Hence, transfected cells expressing these cell-surface receptors have probably become the most widely used model for GPCR biochemical and pharmacological studies. Besides transient transfection, which has the advantage that it can rapidly provide expression for a limited series of experiments, generation of stable cell-lines constitutes the preferred approach when a large amount of a validated, functional receptor expression system is required (e.g., for pharmacological

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_1, © Springer Science+Business Media, LLC 2011

3

4

Koener and Hermans

screening). In these models, the use of a robust viral promoter ensures the expression of the receptor, independently of cell cycle, or other environmental considerations. More recently, the development of sophisticated tools for eukaryotic expression of cloned genes has led to the generation of inducible expression systems in which the transcription of the gene of interest in the transfected cells can be artificially controlled (Fig. 1). In these systems, the sequence of the viral promoter controlling the expression of the cloned gene contains additional regulatory elements allowing the exogenous manipulation of the transcriptional activity (1). Using such systems, the expression level of the cloned gene can be tightly controlled through the addition or elimination of a given reagent in the culture medium (chemical inducer). A number of different systems is commercially available and there are now many examples of their use in the generation of stably transfected cell-lines in which the expression of a given GPCR can be induced (2, 3). This allows the study of the influence of receptor density on the pharmacological properties of ligands (4, 5); it also allows alterations of the stoichiometry between the receptor and interacting signaling partners. Other groups have used such inducible systems to downregulate the expression of a receptor during the maintenance of the cell-line, or to

Fig. 1. Typical mechanism of mammalian inducible expression systems. All commercially available systems consist of two plasmids. The first plasmid is used to generate a cell-line that constantly expresses the regulatory protein. The second plasmid contains the cDNA encoding the GPCR of interest which is cloned under the control of a modified viral promoter. The activity of this promoter is controlled by the interaction with the regulatory protein. Depending on the system, this interaction is positively or negatively ­influenced by the chemical inducer, which is therefore used to trigger or repress the expression of the GPCR.

Inducible Expression of G Protein-Coupled Receptors in Transfected Cells

5

switch on the expression while cells are maintained under defined experimental conditions (e.g., in the presence of biochemical activators or inhibitors). This chapter provides an overview, as well as methodological details, regarding the different systems that are commonly used to generate stably transfected cell-lines in which expression of a cloned GPCR can be chemically controlled. The key features that can differ from one system to another are the low level of expression in noninduced conditions (“leakage” of the induction system), and the efficacy/ stability/reproducibility of the induction. 1. Tet-On and Tet-Off systems (Clontech, Mountain View, California). Transcription is controlled by adding (or omitting) tetracycline, or its derivative doxycycline, into the culture medium (suggested references for GPCR expression using this system (5–7)). With the previous versions of Tet-On and Tet-Off vectors (pTRE plasmids), significant leakage of the system could sometimes be detected, depending on the celllines used. In the recent pTRE-Tight plasmids, both TRE and PminCMV have been modified, leading to a decreased basal expression of the gene of interest. This inducible system permits high expression levels of the cloned receptor, often required for pharmacological studies. The main shortcoming of the Tet-On systems relies on the fact that the expression of the gene of interest can only happen in the presence of the regulatory protein, either the tTA or the rtTA. In fact, on the response plasmid, the expression of the gene is under the control of the Tetracycline-responsive element (TRE), made of Tet operator sequences (TetO) upstream the ­minimal CMV promoter. Activation of gene transcription requires the binding of the tTA or the rtTA to the TetO sequences within the TRE, without or with doxycycline, respectively. Therefore, the expression plasmid cannot be used as a conventional mammalian expression vector in cell-lines, where the specific transactivator is not expressed. 2. T-REx system (Invitrogen, Carlsbad, California). Transcription is controlled by adding tetracycline to the culture medium (suggested references for GPCR expression using this system (8, 9)). The advantages of this system hinges on two main characteristics: the absence of viral transactivation domains (present in the Tet-On and Tet-Off systems) in the Tet repressor protein, which avoids potential toxicity issues in some mammalian cell-lines. Secondly, the gene of interest, under the control of a strong CMV promoter, does not depend on another partner to be expressed. The pcDNA4 plasmid can therefore be used in other cell-lines, for other applications, or for validation of the cloned construct.

6

Koener and Hermans

3. LacSwitch II system (Agilent, Santa Clara, California, formerly available from Stratagene). Transcription is controlled by adding isopropyl-b-d-1-thiogalactopyranoside (IPTG) to the culture medium (suggested references for GPCR expression in this system (4, 10, 11)). This system utilizes transcriptional controls derived from prokaryotes and presents minimal toxicity for mammalian cells. Even though it has been used with ­success for several proteins, allowing rapid induction (4–8 h), only a few studies have reported its usage for the expression of GPCRs. 4. GeneSwitch system (Invitrogen). Transcription is controlled by adding mifepristone to the culture medium (suggested reference for GPCR expression in this system (12)). Note that this system shows many similarities to the ecdysoneinducible mammalian expression system (now discontinued, previously available from Invitrogen) in which the expression is controlled by the addition of muristerone or ponasterone (suggested references for GPCR expression in this former ­system (3, 13, 14)). A key advantage of the ecdysone-inducible system is the very low background expression level. However, expression levels obtained after induction remain rather modest. Moreover, its use is restricted to certain cell-types, and the chemical inducers are relatively expensive. All these systems show some mechanistic similarities: The cDNA encoding the protein of interest is cloned into a vector under the control of a viral promoter which contains consensus sequences for the binding of a regulatory protein (repressor or activator). This construct is introduced into a genetically modified cell-line that constitutively expresses this regulatory protein partner. These cell-lines are available from commercial providers, or need to be established through conventional stable transfection with available plasmid constructs. The modulation of transcription operated by the regulatory protein partner on the vector containing the gene of interest is influenced by its interaction with a defined chemical that easily diffuses from the extracellular environment to the cell nucleus (for a more detailed outline of the different systems, see Table  1). It should be noted  that updated versions of these systems are regularly ­proposed and existing versions are often then discontinued. Nevertheless, these versions are accessible in several laboratories and technical details, such as vector maps, remain available on the Web site of the providers. The upgrades generally propose optimized plasmids which contribute to decreasing the expression leakage in noninduced cells, and/or to increase the ­maximal expression obtainable following induction.

7

Inducible Expression of G Protein-Coupled Receptors in Transfected Cells

Table 1 Mechanisms involved in the control of transcription in the commonly used inducible systems T-Rex system. This system uses some native regulatory elements from the Tn10 tetracycline resistance operon from Escherichia coli. The repressor (TetR) encoded by the regulatory plasmid, recognizes an operator (TetO), inserted in the modified CMV promoter controlling the expression of the cloned GPCR gene on the second plasmid. The interaction between TetR and TetO impairs the expression of the cloned cDNA. This repression is suppressed in the presence of tetracycline or its analogs Tet-On and Off systems. These systems use regulatory elements similar to those in the T-Rex system. In the Tet-Off system, the regulatory protein is the tetracycline controlled transactivator (tTA) which is a fusion of the Tet repressor (TetR) and the C-terminus of the activation domain of the herpes simplex virus VP16. Upon binding of doxycycline, a conformational change of the repressor results in an inhibition of gene expression. In the Tet-On system, a four amino acid change in the tTA generates the reverse tetracycline controlled transactivator (rtTA). With this construct, the conformational change induced upon binding of doxycycline to the rtTA switches on the t­ ranscription of the cloned cDNA LacSwitch system. This system uses modified and adapted regulatory elements from the E. coli lactose operon. The regulatory protein is the Lac Repressor, which binds to a Lac operator inserted in the promoter controlling the transcription of the GPCR cDNA. The binding of the repressor to the operator represses gene expression. Synthetic inducers, such as IPTG, interact with the Lac Repressor and the resulting conformational change decreases its affinity for the operator and allows transcription of the cloned gene. GeneSwitch and Ecdysone systems. The GeneSwitch system uses a combination of regulatory elements. The hybrid regulatory protein is composed by a DNA binding domain (originated from a yeast protein), a ligand binding domain (from the human progesterone receptor), and an activation domain (from the human NF-kB protein). When the inducer, mifepristone, binds to the ligand binding domain, it changes the conformation of the hybrid protein, allowing its binding to the promoter site controlling the expression of the gene of interest. In the Ecdysone system, a heterodimer receptor composed by the ecdysone receptor and the retinoid X receptor controls the transcription of the gene of interest. When the synthetic analogs of ecdysone, ponasterone, or muristerone bind to the ecdysone receptor, they allow gene transcription.

2. Materials 2.1. Cloning Vectors

1. Tet-On and Tet-Off systems: pTRE or pTRE-Tight vector (Clontech; Ref. 631059) 2. T-REx system: pcDN4/TO (Invitrogen; Ref. V102020) 3. LacSwitch II system: pOPRSVI/MCS (Agilent; Ref. 217450) 4. GeneSwitch K1060-01)

system:

pGene/V5-His

(Invitrogen;

Ref.

5. The ecdysone-inducible expression system: pIND (previously available from Invitrogen)

8

Koener and Hermans

Maps of all these vectors are available on the Web sites of the providers. Of note, several variants of these vectors are commercially available (e.g., enabling direct PCR product cloning, or generation of tagged fusion proteins) and updated versions of these vectors, which incorporate additional features, are frequently launched. Interested users should refer to provider Web sites. 2.2. Cell-Lines

Expression of GPCRs by transfection has been reported in a large variety of cell-lines. The choice of a cell-line devoid of any endogenously expressed, related GPCR allows the study of a particular receptor type that is well characterized at the molecular level. Obviously, some biochemical properties of the receptor could depend on the cell line used. Fibroblast cell-lines (e.g., CHO, HEK293, COS, BHK-21, NIH/3T3) are the most common hosts for pharmacological and biochemical studies of GPCRs in transfected cells. These cell-lines generally grow fast and are thus easy to maintain. As mentioned above, inducible systems are based on the activity of a repressor or activator protein, which must be constitutively expressed by the cell host. Generation and validation of these cell-lines expressing the appropriate repressor or ­activator protein can be time consuming (see Note 1) and ready-to-use, validated cell-lines are commercially available (see Note 2).

2.3. Transfection Reagents

Methods for cell transfection include calcium-phosphate precipitation, cationic lipid-mediated transfection reagents and electroporation. Transfection (“lipofection”) using cationic lipids has been reported for many cell-types and is commonly recommended for the generation of cell-lines incorporating an inducible expression system. Several companies offer ready-to-use transfection reagents and optimized protocols for the most commonly used cell-lines. The overview of these reagents is beyond the scope of this chapter and the reader should refer to instructions from the various manufacturers’ Web sites.

2.4. Selection Antibiotics

All currently available systems for inducible expression in mammalian cells necessitates the use of dedicated cell-lines expressing the regulatory protein (repressor or activator) which is commercially available, or which must be generated. In both cases, clones stably expressing the regulatory protein should be maintained in the presence of the selection antibiotic. In addition, the system also requires a second transfection using the above-mentioned vectors in which the cDNA encoding the GPCR of interest has been inserted. These vectors harbor a resistance gene and the selection of stable transfectants requires the use of a second selection antibiotic. 1. Tet-On and Tet-Off systems: The pTRE or pTRE-Tight ­vectors do not contain any selection markers and cell ­transfection

Inducible Expression of G Protein-Coupled Receptors in Transfected Cells

9

should include an additional plasmid encoding resistance for hygromycin (Clontech; Ref. 631309, or Invitrogen; Ref. R220-05), or puromycin (Clontech, Ref. 631305, 631306). Vectors provided in previous versions of the Tet-On and TetOff systems contained resistance genes for one of these antibiotics (pTREHyg or pTREPur). The cells-lines dedicated for the Tet-On and Tet-Off systems should be maintained in the presence of G418 (Clontech, Ref. 631307), as the resistance gene is expressed on the plasmid coding for the transactivator (tTA), or reverse transactivator (rtTA). 2. T-REx system: The pcDN4/TO vector supports zeocin resistance. The cell-lines dedicated for the T-REx system should be maintained in the presence of blasticidin (Invitrogen, Ref. R210-01), which is the resistance gene expressed on the ­plasmid coding for the TetR. 3. LacSwitch II system: The pOPRSVI/MCS vector supports G418 resistance. The cell-lines expressing the LacSwitch II system should be maintained in the presence of hygromycin. 4. GeneSwitch system: The pGene/V5-His vector supports zeocin resistance (Invitrogen, Ref. R250-01). The cell-lines expressing the GeneSwitch system should be maintained in the presence of hygromycin, as the corresponding resistance gene is present in the pSwitch plasmid. 5. The ecdysone-inducible expression system: The pIND vector supports zeocin resistance. The cell-lines dedicated for the Ecdysone-inducible expression system should be maintained in the presence of hygromycin as resistance is supported by the regulatory plasmid pVgRXR. 2.5. Chemical Inducers of Gene Expression

All mammalian inducible systems utilize water soluble and ­diffusible chemicals to control the expression of the gene of interest. These chemicals bind to the regulatory proteins and influence their activity, or their binding to the modified viral promoter. 1. In the Tet-Off system, gene expression is turned on when tetracycline (Sigma, Ref. T3383) or doxycycline (Clontech, Ref. 631311) is removed from the culture medium. In contrast, gene expression is turned on in the Tet-On system on addition of doxycycline (note, the Tet-On system is not responsive to tetracycline). Although the Tet-Off system responds to tetracycline and doxycycline, the use of doxycycline is preferred as its half-life is longer (48 h for doxycycline and 24 h for tetracycline) (see Note 3). 2. In the T-REx System, induction is triggered by the addition of tetracycline to the cell culture medium. Doxycycline can be used as an alternative to tetracycline.

10

Koener and Hermans

3. In the LacSwitch II system, the transcription of the cDNA inserted into the pOPRSVI/MCS vector is induced by the addition of IPTG (Agilent Technologies, Ref. 217450) to the culture medium. 4. In the GeneSwitch system, induction is achieved by the addition of the synthetic steroid, mifepristone (Invitrogen, Ref. H110-01). 5. In the Ecdysone-inducible expression system, expression of the cloned gene is initiated by the addition of muristerone (Invitrogen, Ref. H10001) or ponasterone (Invitrogen, Ref. H10101-H10103), which are both analogs of the insect steroid hormone, ecdysone. 2.6. Materials for Validating the Inducible Expression of GPCRs

Several methods are commonly used to demonstrate and quantify GPCR expression in cultured cells: radioligand binding studies, western blot detection of GPCR, functional assays, and quantitative RT-PCR. The description of these assays is beyond the scope of this chapter. Obviously, if specific radioligands are available, binding studies (on intact cells or homogenates) are recommended to validate and quantify the “inducibility” of receptor expression.

3. Methods 3.1. Molecular Cloning of the GPCR cDNA into the Vector

The cDNA sequence encoding the GPCR of interest is generally amplified by PCR and only the coding sequence is inserted into the vector. The vectors provided in all the systems contain a conventional multiple cloning site, where several consensus sites for selected restriction endonucleases are present, facilitating the ­orientated insertion of the GPCR cDNA sequence. Of note, the initial cloning of a PCR amplification product is facilitated using the so-called direct PCR cloning kits, such as the TOPO TA ­cloning system available from Invitrogen. After validation of the cloned sequence, the insert can be easily transferred to the vector designated for inducible expression. In order to enhance the ­efficiency of transcription and translation, it is of critical importance that a TATA box and a Kozak sequence are present between the promoter and the translation initiation site (15).

3.2. Transfection into Appropriate Cells

Transfection of mammalian cell monolayers with cationic lipids is generally rapid and straightforward (see Fig. 2). Detailed protocols are always supplied by the manufacturer. In brief, supercoiled DNA is usually mixed with an appropriate volume of a cationic lipid solution (see Notes 4 and 5). After incubation at room ­temperature for 30–60 min, this mix is added to growing cells, often after the removal of serum from the medium (see Note 6).

Inducible Expression of G Protein-Coupled Receptors in Transfected Cells

11

Fig.  2. Schematic representation of the key steps in the generation and validation of an inducible expression system for GPCRs.

After a few hours, culture medium and serum are added to increase the volume and reach the optimal serum concentration for cell growth (the transfection mix is not removed). Finally, after 24 h the culture medium is removed and replaced by fresh medium. Initially, the cell density, DNA-to-cationic lipid ratio, as well as the incubation time has to be optimized for each cell-type. Recent formulations appear to be more flexible and supplier guidelines are applicable for most cell-types. 3.3. Selection and Isolation of Individual Clones

As indicated above (see Subheading  2.4), isolation of stable t­ ransfectants requires the specific elimination of nontransfected (or transiently transfected) cells. Except for the Tet-On and Tet-Off systems, all dedicated cloning vectors provided in the inducible expression systems mentioned above harbor an ­additional gene which confers permanent resistance to a ­specified antibiotic. For the Tet-On/Tet-Off system, the cell transfection step should include a plasmid encoding resistance for hygromycin or puromycin. Selection is generally started 2 or 3  days after transfection by supplementing the culture medium with the appropriate antibiotic. The concentration of the antibiotic should not be increased progressively, but instead, the appropriate ­concentration should be used immediately. Generally,

12

Koener and Hermans

toxicity and cell death are only noticeable after 3–4 days. At that time, renewal of the culture medium is necessary every day or every other day in order to remove dead cells and debris. It can be useful to set-up sister wells, where DNA was omitted during the transfection protocol, allowing the efficiency of the selection procedure to be evaluated. Interestingly, it has been suggested that the activity of G418 is decreased when the classical antibiotics penicillin and streptomycin are present in the culture medium. As introduced above, the inducible systems presented in this chapter involve the use of cell-lines which permanently express a regulatory protein. These cell-lines are commonly maintained in the presence of a first selection antibiotic, and this antibiotic should be included to guarantee the maintenance of the cell-lines (at the concentration suggested by the provider, or determined by the user for customized cell-lines). Not all cell-lines are equally sensitive to these selection antibiotics, and variations can sometimes be observed between different batches of the same cell-line. The optimal selection concentration (killing nontransfected cells) has to be determined before transfection (see Note 7). Several culture plates are seeded at low density in culture medium supplemented with increasing concentrations of the antibiotic. Guidelines for likely effective selectant concentrations are as follows: hygromycin – 100–400 mg/mL, G418 – 400–1,000  mg/mL, puromycin – 0.5–5  mg/mL, zeocin – 50–1,000 mg/mL, and blasticidin – 1–10 mg/mL. 3.4. Isolation and Propagation of Clones

3.4.1. Picking Colonies

Because of the rather low efficiency of stable transfection, the population of viable cells remaining after selection is highly heterogeneous regarding the level of expression of the transfected gene. Isolation of clonal cell-lines is required in order to obtain constant expression levels and therefore reproducible experimental data. The following techniques describe how to isolate and transfer clonal cells in 24-well plates (see Fig. 2). After isolation and proliferation, the cells are successively transferred to six-well plates and medium-sized flasks before providing enough material to determine the expression of the recombinant GPCR. Growth is rather slow at the beginning and, depending on the cell-type, these proliferation steps take at least 3–6 weeks. 1. Antibiotic-resistant cells generally grow as colonies that are easily detected after the elimination of all nontransfected dead cells and cell debris. On visualizing the colonies to be picked (with the microscope or sometimes with the naked eye), mark their positions by drawing circles on the bottom of the dish. 2. Aspirate the medium and using forceps, place small, sterilized glass cylinders (cloning disks) on to the circled colonies (it is sometimes helpful to glue the disks with a small amount of silicon grease).

Inducible Expression of G Protein-Coupled Receptors in Transfected Cells

13

3. Add 50  mL trypsin solution (0.05% trypsin–EDTA) to the inside of the cylinders and leave (in a laminar-flow hood) for 5 min. 4. Using fresh tips for each colony, collect the contents from each disk and transfer to a single well of a 24-well plate containing 2 mL culture medium. In order to obtain pure clonal cell-lines, it is critical to pick colonies as early as possible once the nontransfected cells have been eliminated. Otherwise, colonies will become too large and may fuse, or cells from large colonies may detach and reseed at a distance, contaminating neighboring colonies. An alternative to the use of cloning disks is to directly pick the colonies using a pipette set to 10 mL and sterile pipette tips. In this case, there is no need to remove the culture medium from the plate, instead, just incline the entire plate in order to scrape the marked colonies which are then directly transferred to the 24-well plate containing culture medium. This technique offers the advantage of being relatively rapid and avoids the possibility of the dryingout of the rest of the cells. If needed, culture medium can be renewed after colony picking and the entire population can then be screened later by limited dilution (see below). 3.4.2. Limited Dilution

1. After elimination of nontransfected dead cells and cell debris, allow the colonies to grow for 2–3 days in order to increase the total number of cells. 2. Trypsinize the entire population of transfected cells and ­collect them in culture medium. After counting, prepare 2  mL of cell suspension adjusted to 1,000  cells/mL. The remaining cells can be frozen or seeded in order to get a “total population.” 3. Transfer 100 mL of the cell suspension into the first row of a 96-well plate (flat bottom) and put 100  mL fresh culture medium in the seven other rows. 4. Using a 12-multichannel pipette, transfer 30 mL from row 1 to row 2. After gentle mixing by repetitive pipetting, transfer 30 mL from row 2 to row 3 and proceed for the other rows until row 8. It is not necessary to change tips between rows. 5. Leave in the incubator for ~7 days. 6. Using a microscope at low magnification, view every well looking for those containing a single round-shaped colony (likely to be found in rows 4–6). Note: the first row rapidly reaches confluence and this might be helpful for setting the focus of the microscope. 7. Trypsinize the cells from the well containing single colonies and transfer to single wells of a 24-well plate containing ­culture medium.

14

Koener and Hermans

3.5. Validation of the Model 3.5.1. Validation of Inducible Expression

A single transfection assay can give rise to many clonal cell-lines potentially expressing the GPCR of interest. However, marked variations in GPCR expression levels between clones are frequently observed. This could be related to differences in the site of insertion within the chromosomes, or from the number of inserted copies of the transgene. Also, despite their resistance to the selection antibiotic (see Note 8), some clones may show low or no expression of the transgene. As early as possible after isolation of clones, the aim is to detect those showing minimal expression before induction and elevated expression after exposure to a maximal concentration of the chemical inducer (see Note 9). Note that the induction of the GPCR may be tested on the entire population of transfected cells, after selection with the antibiotic, but before the isolation of individual clones. Data obtained using this heterogeneous population of cells is solely indicative of the presence of valid clones within the whole population. The most quantitative evaluation of the expression of a given GPCR is to measure the specific binding of an appropriate radioligand. This may require relatively large amounts of cells and therefore frequently delays the elimination of negative clones. In contrast, functional assays or immunodetection can be performed on a relatively small number of cells. However, these approaches do not provide quantitative information that allows the selection of high- or low-receptor-expressing clones. For some GPCRs, radioligand binding can be measured on intact cells seeded in 24-well plates. This requires fewer cells than the preparation of a crude homogenate and is a good compromise between delayed selection and quantitative screening. A binding assay on cell homogenates, using an appropriate radioligand, is thus recommended to validate receptor expression after induction. In order to realize such an assay, cells should be plated in 10 cm Petri dishes (1 × 106 cells per dish) with 8–10 mL of the appropriate culture medium. Before adding the inducer, cells should be adherent and close to 80% confluence. Depending on the cell-type, the amount of time necessary to reach this cell density varies between 24 and 48 h. The culture medium should then be changed to one supplemented with a maximally stimulating concentration of the chemical inducer. It is suggested to prepare a stock solution of medium with the inducer to ensure a uniformity of conditions for the quantitative assay. Sister plates should be maintained in the absence of the inducer (noninduced condition). In addition, nontransfected cells can be grown in ­parallel to constitute a negative control. Radioligand binding assays should be performed on cell homogenates following the instructions provided in Chap. 1 of this volume. In most cases, the binding assay can be limited to the determination of the ­specific and ­nonspecific binding at a relevant concentration of radioligand. The crude homogenate from a single 10  cm Petri

Inducible Expression of G Protein-Coupled Receptors in Transfected Cells

15

dish prepared as indicated below could be sufficient. A  well-­ characterized, unlabeled competitor should be used for determining­ the nonspecific binding. The maximal concentrations of inducer that should be used are the following, according to the User Manuals provided by the respective manufacturers: 1. Tet-On and Tet-Off systems: With these systems, the expression of the cloned receptor is respectively induced or silenced upon the addition of doxycycline or tetracycline for at least 24 h at the concentration of 1–2 mg/mL (suggested concentration to maximize the response). 2. T-REx system: Doxycycline or tetracycline should be used as for the Tet-On/Tet-Off system. 3. LacSwitch II system: IPTG should be added at the final concentration of 1–5 mM, for a period of 4–12 h. Optimal induction time is gene-dependent. 4. GeneSwitch system: Mifepristone must be used at a concentration of 10 nM for a period of 24 h. 5. Ecdysone-inducible expression system: Ponasterone or muristerone are used at a concentration of 5–10  mM for at least 24 h (3). 3.5.2. Optimizing Receptor Expression: Titration of Inducer Concentration

In order to determine the experimental conditions to be used for titrating receptor expression, a range of inducer concentrations need to be screened. Alternatively, one could also decrease the duration of induction, but this approach is less convenient. According to  the respective User Manuals, the following conditions allow a progressive switch from the minimal to maximal expression: 1. In Tet-On and Tet-Off systems: Doxycycline or tetracycline concentrations should vary from 0.01 to 1–2 mg/mL. 2. In the T-REx system: Doxycycline or tetracycline should be used as for the Tet-On and Tet-Off systems. 3. In the LacSwitch II system: IPTG concentrations should vary from 1 to 5 mM, during the 4–12 h induction period. 4. In the GeneSwitch system: In order to modulate the expression of the cell-line, mifepristone concentrations should vary between 0.1 and 100 nM. The duration of exposure to this inducer could vary between 12 and 72 h. 5. In the ecdysone-inducible expression system: Concentrations of muristerone A and ponasterone A in the range of 1–20 mM should be screened (16). Preparation of a crude homogenate from a single 10  cm ­culture dish: 1. Aspirate the culture medium and rinse the cells twice with ice-cold PBS.

16

Koener and Hermans

2. Scrape the cells in 1 mL PBS and collect the suspension in a 2 mL microtube. Rinse the plate with another 1 mL of PBS, collect and transfer to the same tube. Centrifuge at maximal speed in a microfuge for 3  min at 0–4°C and discard the supernatant. 3. Resuspend the pellet in 1 mL 50 mM Tris–HCl pH 7.4 using a 1 mL syringe and a 26 gauge needle. Centrifuge at maximal speed for 15  min at 0–4°C and discard the supernatant. Repeat step 3 twice. 4. The final pellet is resuspended in 200 mL 50 mM Tris–HCl pH 7.4. This crude homogenate can be used for protein and radioligand binding assays after dilution in the appropriate binding buffer.

4. Notes 1. Even though several established cell-lines are available for generating the inducible system, additional cell-lines can be developed using the appropriate plasmids provided by the manufacturers. Thus, a first step of transfection with the ­plasmid encoding the regulatory protein generates cell-lines that permanently express this repressor or activator. The selection of an appropriate clone from this first transfection is of major importance as the stable expression of the regulatory protein contributes to the efficiency of the inducible system. Therefore, before proceeding with the second transfection (with the vector containing the GPCR of interest), the validity of the cell host should be tested using an inducible vector encoding a reporter gene allowing quantification, such as luciferase, or a green fluorescent protein. Appropriate clones are those in which basal expression (no inducer) is minimal, whereas robust and stable expression is obtained after the addition of the chemical inducer. 2. List of commercially available cell-lines for the generation of inducible systems: Tet-On and Tet-Off cell-lines (all from Clontech): (a) HEK293 Tet-On (human embryonic kidney-derived cell-line) (b) HEK-293 Tet-Off (human embryonic kidney-derived cell-line) (c) HepG2 Tet-On (human hepatocellular carcinoma-derived cell-line) (d) HepG2 Tet-Off (human hepatocellular carcinoma-derived cell-line)

Inducible Expression of G Protein-Coupled Receptors in Transfected Cells

17

(e) MCF7 Tet-On (human breast adenocarcinoma) (f ) MCF7 Tet-Off (human breast adenocarcinoma) (g) HeLa Tet-On (human cervical epithelioid carcinoma cell) (h) HeLa Tet-Off (human cervical epithelioid carcinoma cell) (i) CHO-AA8 Tet-Off (Chinese hamster ovary-derived cellline) (j) PC12 Tet-Off (rat adrenal pheochromocytoma-derived cell-line) (k) PC12 Tet-On (rat adrenal pheochromocytoma-derived cell-line) (l) Jurkat Tet-Off (human acute T-cell leukemia-derived cell-line) (m) Jurkat Tet-On (human acute T-cell leukemia-derived cell-line) (n) Saos-2 Tet-Off (human osteosarcoma-derived cell-line) (o) MDCK Tet-Off (Madin-Darby canine kidney type II epithelial-derived cell-line) (p) MEF/3T3 Tet-Off (Swiss 3T3 embryonic mouse fibroblast-derived cell-line) (q) HT1080 Tet-Off (human fibrosarcoma-derived cellline) (r) CHO-K1 Tet-On (Chinese hamster ovary-derived cellline) (s) U2-OS Tet-On (human osteosarcoma-derived cell-line) (t) T-47D Tet-On (human ductal mammary carcinoma (pleural effusion)-derived cell-line) (u) T-47D Tet-Off (human ductal mammary carcinoma (pleural effusion)-derived cell-line T-REx cells (all from Invitrogen) (a) T-Rex-293 (human embryonic kidney-derived cell-line) (b) T-REx-HeLa (human cervical epithelioid carcinoma cell) (c) T-REx-U2OS (human osteosarcoma cell-line) GeneSwitch system: from Invitrogen (a) GeneSwitch-3T3 (Swiss 3T3 embryonic mouse fibroblast-derived cell-line) (b) GeneSwitch-CHO ­cell-line)

(Chinese

hamster

ovary-derived

18

Koener and Hermans

(c) GeneSwitch-293 (human embryonic kidney-derived ­cell-line) For the LacSwitch inducible system, there are no ­commercially available cell-lines and for the ecdysone-inducible system, some lines were previously commercialized which have now been discontinued. 3. The use of tetracycline in animal feeding results in potential contamination in commercially available sera. When using tetracycline-inducible systems in cell cultures requiring fetal bovine serum, this could result in an apparent leakage in the absence of added inducer. The use of batches that are free or very low in tetracycline should be used when minimal basal expression is desired. Tetracycline-free serum is available from Clontech (Tet System approved FBS, Australia-Sourced Ref. 631039-631040 or US-sourced Ref. 631101-631105). 4. Supercoiled plasmid DNA is commonly used for transfection of mammalian cells. Nevertheless, some data from the literature reports on the linearization of the plasmid before stable transfection (avoiding intracellular linearization within the sequence of interest before chromosomal integration). 5. The quantity of plasmid DNA to be used in most transfection protocols depends on the number of cells and, in the case of adherent cells growing as monolayers, is proportional to the area of the culture plates. For cationic lipid-mediated transfection, the final concentration of the reagents in the culture medium also needs to be considered. The following Table indicates the area of plasticware commonly used in mammalian cell transfection and shows the appropriate volume of culture medium to be used during transfection. Culture surface

Area (cm2)

Vol. of medium (mL)

10 cm dish

58.95

10

6 cm dish

19.5

 3

6-well plate

  9.6

 2

12-well plate

  3.8

 1

24-well plate

 2

  0.5

96-well plate

  0.32

  0.1

6. Dividing cells show the highest transfection efficiency. Therefore, transfection is generally performed 24 h after seeding at a density of about 50–70%. Although some protocols allow the presence of serum during transfection, the critical steps are frequently performed under serum-free conditions.

Inducible Expression of G Protein-Coupled Receptors in Transfected Cells

19

Several protocols also recommend that the transfection is ­performed in the absence of any antibiotics. Cell transfection techniques introduce large DNA fragments into the cells and are a potential source of stress for the cells. Therefore, it is advisable to supplement the culture medium with serum as soon as possible after transfection. 7. Antibiotics are not always available as 100% pure powders and the accurate concentration has to be calculated on the basis of purity. Some manufacturers supply “ready to use” standardized solutions of antibiotics. Hygromycin B purity is also ­variable and since its activity is frequently expressed in units, the conversion factor (unit per mg) must be known. 8. Once the transfected cells have been selected and clones isolated, the selection antibiotic is added to the culture medium for maintenance and propagation. Of note, for repeated passaging of selected clones, the final concentration of the selection antibiotic can be decreased by 75%, lowering the costs of the experiments. When cells are seeded for induction, it is generally recommended that the selection antibiotics are removed from the culture medium. 9. Sodium butyrate is known to arrest cell growth and enhances the activity of the cytomegalovirus promoter. It has been sometimes added to the culture media of transfected cells (5  mM for 6–48  h before the experiment) to increase the expression of recombinant GPCRs in HEK and CHO cells (17, 18). Sodium butyrate (n-butyric acid, sodium salt) can be prepared as aqueous solution and is available from Sigma (St Louis, Missouri, USA). References 1. Gossen, M., Bonin, A. L. and Bujard, H. (1993) Control of gene activity in higher eukaryotic cells by prokaryotic regulatory ­elements. Trends Biochem. Sci. 18, 471–475. 2. Van Craenenbroeck, K., Vanhoenacker, P., Leysen, J. E. and Haegeman, G. (2001) Evaluation of the tetracycline- and ecdysoneinducible systems for expression of neurotransmitter receptors in mammalian cells. Eur. J. Neurosci. 14, 968–976. 3. Downey, P. M., Lozza, G., Petro, R., Diodato, E., Foglia, C., Bottazzoli, F., Brusa, R., Asquini, T., Reggiani, A. and Grilli, M. (2005) Ecdysone-based system for controlled inducible expression of metabotropic glutamate receptor subtypes 2, 5, and 8. J. Biomol. Screen. 10, 841–848. 4. Hermans, E., Challiss, R. A.J. and Nahorski, S. R. (1999) Effects of varying the expression

level of recombinant human mGlu1a receptors on the pharmacological properties of agonists and antagonists. Br. J. Pharmacol. 126, 873–882. 5. Theroux, T. L., Esbenshade, T. A., Peavy, R. D. and Minneman, K. P. (1996) Coupling efficiencies of human a1-adrenergic receptor subtypes: titration of receptor density and responsiveness with inducible and repressible expression vectors. Mol. Pharmacol. 50, 1376–1387. 6. Violin, J. D., Dewire, S. M., Barnes, W. G. and Lefkowitz, R. J. (2006) G protein-coupled receptor kinase and b-arrestin-mediated desensitization of the angiotensin II type 1A receptor elucidated by diacylglycerol dynamics. J. Biol. Chem. 281, 36411–36419. 7. Urlinger, S., Baron, U., Thellmann, M., Hasan, M. T., Bujard, H. and Hillen, W. (2000)

20

Koener and Hermans

Exploring the sequence space for tetracyclinedependent transcriptional activators: novel mutations yield expanded range and sensitivity. Proc. Natl. Acad. Sci. USA 97, 7963–7968. 8. Canals, M., Jenkins, L., Kellett, E. and Milligan, G. (2006) Up-regulation of the angiotensin II type 1 receptor by the MAS proto-oncogene is due to constitutive activation of Gq/G11 by MAS. J. Biol. Chem. 281, 16757–16767. 9. Canals, M. and Milligan, G. (2008) Constitutive activity of the cannabinoid CB1 receptor regulates the function of co-expressed m-opioid receptors. J. Biol. Chem. 283, 11424–11434. 10. Williams, N. G., Zhong, H. and Minneman, K. P. (1998) Differential coupling of a1-, a2-, and b-adrenergic receptors to mitogen-­ activated protein kinase pathways and differentiation in transfected PC12 cells. J. Biol. Chem. 273, 24624–24632. 11. Zhong, H., Guerrero, S. W., Esbenshade, T. A. and Minneman, K. P. (1996) Inducible expression of b1- and b2-adrenergic receptors in rat C6 glioma cells: functional interactions between closely related subtypes. Mol. Pharmacol. 50, 175–184. 12. Corti, C., Crepaldi, L., Mion, S., Roth, A. L., Xuereb, J. H. and Ferraguti, F. (2007) Altered dimerization of metabotropic glutamate receptor 3 in schizophrenia. Biol. Psychiatry 62, 747–755. 13. Law, P. Y., Kouhen, O. M., Solberg, J., Wang, W., Erickson, L. J. and Loh, H. H. (2000)

Deltorphin II-induced rapid desensitization of d-opioid receptor requires both phosphorylation and internalization of the receptor. J. Biol. Chem. 275, 32057–32065. 14. Choi, D. S., Wang, D., Tolbert, L. and Sadee, W. (2000) Basal signaling activity of human ­dopamine D2L receptor demonstrated with an  ecdysone-inducible mammalian expression system. J. Neurosci. Methods 94, 217–225. 15. Kozak, M. (1987) At least six nucleotides ­preceding the AUG initiator codon enhance translation in mammalian cells. J. Mol. Biol. 196, 947–950. 16. McDonald, R. L., Balmforth, A. J., Palmer, A. C., Ball, S. G., Peers, C. and Vaughan, P. F. (1995) The effect of the angiotensin II (AT1A) receptor stably transfected into human neuroblastoma SH-SY5Y cells on noradrenaline release and changes in intracellular calcium. Neurosci. Lett. 199, 115–118. 17. Nash, M. S., Selkirk, J. V., Gaymer, C. E., Challiss, R. A. J. and Nahorski, S. R. (2001) Enhanced inducible mGlu1a receptor expression in Chinese hamster ovary cells. J. Neurochem. 77, 1664–1667. 18. Pindon, A., van-Hecke, G., van-Gompel, P., Lesage, A. S., Leysen, J. E. and Jurzak, M. (2002) Differences in signal transduction of  two 5-HT4 receptor splice-variants: ­compound specificity and dual coupling with Gas- and Gai/o-proteins. Mol. Pharmacol. 61, 85–96.

Chapter 2 Using the Flp-In™ T-Rex™ System to Regulate GPCR Expression Richard J. Ward, Elisa Alvarez-Curto, and Graeme Milligan Abstract The development of a cell-based system that allows the integration of a gene of interest (GOI), such as a G protein-coupled receptor (GPCR), into a specific site on the genome, has made the generation of mammalian cell lines able to express such proteins easy and efficient. Flp-In™ stable cell lines are isogenic and hence protein expression is constant across a population of cells. A useful addition to the Flp-In™ system (Flp-In™ T-Rex™) allows this expression to be controlled by the addition of a small molecule inducer to the cell culture medium. Stable cell lines generated as described here can be used to great advantage in the study of receptor pharmacology signalling and oligomerisation. Key words: Flp-In™ T-Rex™, GPCR expression, Inducible locus, Stable expression, Doxycycline induction, Tet repressor

1. Introduction Expression of G protein-coupled receptors (GPCRs) in a ­heterologous system using a variety of mainly mammalian cells has become the norm when studying the pharmacology and function of these crucial signal transduction system components. Stable expression of GPCRs, that is when a gene or cDNA is integrated into the host cell genome, is even more useful to furthering the understanding of the mechanisms involved in cell signalling. For this reason, cell lines that express stably and reliably the proteins of interest, in this case GPCRs are routinely created in many laboratories (1, 2). Transiently transfected cells are commonly used as they are quick to produce but they have many disadvantages, notably large differences in expression level among the population

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_2, © Springer Science+Business Media, LLC 2011

21

22

Ward, Alvarez-Curto, and Milligan

of cells and between experiments. The Flp-In™ T-Rex™ system (Invitrogen Life Technologies; www.invitrogen.com) is designed to create cell lines which stably express proteins of interest in an isogenic and inducible manner. This removes possible variation in expression levels or patterns due to different sites of integration into the chromosome of the gene of interest (GOI) and allows cells to be grown without expression of the protein of interest until such time as this is required. The ability to induce expression by addition to the culture medium of a small molecule has several advantages over either transient transfection, or the generation of constitutively expressing “conventional” stable cell lines. (1) There is no need for repeated, highly variable and expensive transient transfections. (2) All the Flp-In™ T-Rex™ stables derived from the same parental cells are integrated at the same site and hence have the same genetic background allowing comparisons to be made. (3) The Flp-In™ T-Rex™ stable cell lines require only polyclonal selection rather than the isolation of individual colonies. (4) Stable cell lines can be grown to the required density before induction of the GOI, thus avoiding potentially detrimental effects upon cell growth (see Fig. 1). (5) The level of expression can be regulated by varying the concentration of the inducing agent (tetracycline/doxycycline). (6) Since expression of the GOI is only induced when required, there is less likelihood of this expression being lost due to non-expressing cells outgrowing those which are still expressing the GOI. The generation of a Flp-In™ T-Rex™ cell line (that is, a parental cell line which can receive and express a GOI) requires the integration of two plasmids, one containing a F  lp Recombination

Fig. 1. The effect of expression of the human ghrelin receptor upon cell growth. The left hand panel shows a Flp-In™ T-Rex™-293 cell line harbouring the human ghrelin receptor prior to induction. The right hand panel shows cells of the same line after induction with doxycycline. The rounding and poor viability of the cells after induction of the ghrelin receptor is related to the high constitutive activity of this receptor through Gq/G11 family G proteins as it is prevented by both a ghrelin receptor inverse agonist and an inhibitor of Gq/G11 G proteins (K. A. Bennett and G. Milligan, unpublished observations).

Using the Flp-In™ T-Rex™ System to Regulate GPCR Expression

23

Target (FRT) site (3) and the other, a gene expressing the tetracycline repressor (tet repressor) (4). These are maintained within the genome by conferring resistance to the antibiotics zeocin and blasticidin, respectively. It is possible to purchase various cell lines which have been modified in this way from a commercial source (Invitrogen Life Technologies). A list of the available cell lines can be found in Table 1. It can be seen from this that while there are several Flp-In™ cell lines (which have the FRT integration site) and T-Rex™ cell lines (which have the tet repressor system), there is to date only the Flp-In™ T-Rex™-293 (based upon HEK293 cells) which has both. Therefore, in order to work in a background other than HEK293 cells, the plasmids must be obtained and integrated by the user into the desired cell line, for instance, Madin-Darby Canine Kidney (MDCK) cells, which may be used as a polarised cell model (J.H. Robben and G. Milligan, unpublished results). This procedure involves the transfection and integration of the FRT site plasmid (pFRT/lacZeo), the selection of a number of colonies by zeocin resistance and the verification of the b-galactosidase activity that this plasmid confers (Fig. 2). One limitation of the Flp-In™ system is that assays based on b-galactosidase expression and/or complementation are limited in these cells by the high level of b-galactosidase activity derived from the pFRT/lacZeo plasmid. Southern blotting analysis is then used to distinguish a clone with only one copy of the FRT site. The plasmid pcDNA6/TR, which expresses the tet repressor protein, is then transfected into this clone and integrants (independent of the FRT site) selected by blasticidin resistance and checked for b-galactosidase activity (Fig. 2). While this procedure is within the capabilities of many laboratories, the use of the established Flp-In™ T-Rex™-293 cell line clearly makes sense if appropriate to the proposed study. The expression of GPCRs using this system has been found to be an efficient and reliable means of generating cell lines that

Table 1 Commercially available Flp-In™ TRex™ cell lines Flp-In™ cell lines

T-REx™ cell lines

Flp-In™ T-REx™ cell lines

Flp-In™-293

T-REx™-293

Flp-In™ T-REx™-293

Flp-In™-CV-1

T-REx™-HeLa

Flp-In™-CHO

T-REx™-CHO

Flp-In™-BHK

T-REx™-Jurkat

Flp-In™-3T3 Flp-In™-Jurkat

24

Ward, Alvarez-Curto, and Milligan

a FRT-zeo • Flp-InTM T-RExTM parental 293 cells, resistant to zeocin and blasticidin

tetR-blast • co-transfect pcDNA5/FRT/TO/GPCR and pOG44

b

• pcDNA5/FRT/TO/GPCR integrates at FRT site catalysed by Flp recombinase from pOG44. Resistant to blasticidin and hygromycin. TetR protein binds tetOp and represses GPCR expression

FRT-hygro-tetOp -GPCR-FRT

tetR-blast

• add tetracycline / doxycycline to culture c medium

• tetracycline / doxycycline binds tetR protein which is released by tetOp allowing GPCR expression Tet repressor (tetR) = Tetracycline / doxycycline = GPCR =

FRT-hygro-tetOp -GPCR-FRT

tetR-blast

Fig. 2. Schematic representation of the process of making Flp-In™ T-Rex™ cell lines which express a GPCR in a stable and inducible manner. (a) Shows parental Flp-In™ T-Rex™ HEK293 cells, containing the FRT site and expressing the tet repressor. After transfection, (b), the pcDNA5/FRT/TO/GPCR integrates at the FRT site and the tet repressor protein binds to the tetOp region of the integrated pcDNA5/FRT/TO/GPCR which controls the GPCR expression, repressing it. (c) Addition of tetracycline or doxycycline binds the tet repressor, releasing the tet operator and enabling GPCR expression.

are simple to maintain, but which can express GPCRs within a few hours at a level determined by the degree of induction. The system is compatible with the use of epitope tags and fluorescent tags at the N- or C- termini of the GPCR (Figs. 3 and 4). Novel tagging technologies, such as the SNAP/CLIP tagging system (Covalys, www.covalys.com/New England Biolabs Inc; www.neb.com), can also be used.

Using the Flp-In™ T-Rex™ System to Regulate GPCR Expression

25

Fig. 3. Induction and expression of GPCRs using the Flp-In™ T-Rex™ system. Panels (a) and (b) show cells of a stable line which harbour the human cannabinoid CB1 receptor fused at its carboxy-terminal to cyan fluorescent protein (CFP). (a) Non-induced cells (no doxycycline), (b) cells which have been induced with 0.5 mg/mL doxycycline. Panels (c) and (d) are equivalent, but show cells harbouring the human orexin 1 receptor (OX1) fused to enhanced yellow fluorescent protein (eYFP).

2. Materials 2.1. E quipment

1. Tissue culture plastics. 2. “Mr. Frosty”-type Cryo Freezer container (Nalgene). 3. 22 mm thickness 0 glass cover slips.

2.2. R  eagents

1. pcDNA5/FRT/TO (Invitrogen Life Technologies). 2. pOG44 (Invitrogen Life Technologies). 3. Flp-In™ T-Rex™ -293 host cells (Invitrogen Life Tech­no­ logies). 4. Dulbecco’s Modified Eagle Medium 1× (DMEM): + 4.5 g/L glucose, + l-glutamine, – pyruvate. 5. Foetal bovine serum (FBS) tetracycline negative, European Union approved: Added to DMEM to 10% (v/v) of final volume. 6. Penicillin/streptomycin solution: 10,000 units/mL penicillin and 10  mg/mL streptomycin in 0.9% (w/v) NaCl diluted

26

Ward, Alvarez-Curto, and Milligan

Fig. 4. Induction of a double stable cell line expressing c-Myc-tagged human muscarinic M3 receptor fused to cerulean fluorescent protein (c-Myc-hM3-cerulean), inducibly and flag-tagged human muscarinic M3 receptor fused to citrine fluorescent protein (flaghM3-citrine), constitutively. Flag-hM3-citrine is present in the presence and absence of doxycycline (panels a1 and a3, respectively), while c-Myc-hM3-cerulean can only be observed when doxycycline is present (panels a4 and a2, respectively). Western blots of lysates of these cells treated with doxycycline for the times indicated are shown in panel (b). Positive signals are seen in all lanes with the anti-GFP and anti-Flag antisera as these are able to detect the constitutively expressed component (flag-hM3-citrine), but are only present after 6–14 h doxycycline treatment with the anti-c-Myc antiserum as this is specific for the inducible component, c-Myc-hM3-cerulean.

1:100 into DMEM to give 100 mg/mL of streptomycin and 100 units/mL of penicillin. 7. Complete DMEM: DMEM with the addition of 10% (v/v) FBS and penicillin/streptomycin (100 units/mL/100 mg/mL, respectively). 8. Zeocin (Invitrogen Life Technologies): 1 g at 100 mg/mL in water. 9. Complete DMEM + zeocin: As complete DMEM (see item 7), but with the addition of 50–500 mg/mL zeocin.

Using the Flp-In™ T-Rex™ System to Regulate GPCR Expression

27

10. Blasticidin S HCl (Invitrogen Life Technologies): 50  mg ­dissolved in 10 mL water (5 mg/mL), 0.22 mm filtered and stored at −20°C as appropriate aliquots. 11. Complete DMEM + zeocin + blasticidin: As complete DMEM + zeocin (see item 9) but with the addition of blasticidin to a final concentration of 5 mg/mL. 12. Ultrapure sterile water. 13. Complete DMEM + blasticidin: As complete DMEM + zeocin + blasticidin (see item 11), but with the zeocin omitted. 14. Hygromycin B (Roche Diagnostics): 1  g in 20  mL water (50 mg/mL). 15. Complete DMEM + blasticidin + hygromycin: As complete DMEM + blasticidin (see item 13), but with the addition of hygromycin B to a final concentration of 0.5–2 mg/mL. 16. Tetracycline hydrochloride (Sigma-Aldrich Company Ltd): Dissolved in water to 1 mg/mL and passed through a 0.22 mm filter. Stored at −20°C as 0.5 mL aliquots. 17. Doxycycline hyclate (Sigma-Aldrich Company Ltd.): Dissolved in water to 1 mg/mL and passed through a 0.22 mm filter. Stored at −20°C as 0.5 mL aliquots. 18. 2× RIPA buffer: 100 mM HEPES pH 7.5, 300 mM NaCl, 2% (v/v) Triton TX-100, 1% (w/v) sodium deoxycholate and 0.2% (w/v) SDS), stored at 4°C. 1× RIPA consists of 25 mL 2× RIPA with the addition of 10  mM NaF, 5  mM EDTA, 10  mM sodium phosphate buffer pH 7.5, and 5% (v/v) ­ethylene glycol made up to a final volume of 50  mL with water. Protease inhibitors (see item 21) should be added according to the manufacturer’s instructions. This solution is stable for up to 1 week at 4°C. 19. Poly-D-Lysine hydrobromide: Add 50 mL of sterile water to 5 mg in the bottle and store at 4°C. This solution only has to come into contact with cover slips or tissue culture plastic to form an adequate coating and, if kept sterile may be reused repeatedly. 20. HEPES buffer pH 7.4: For 1 L, add 7.6 g NaCl (final 130 mM), 0.37 g KCl (final 5 mM), 4.77 g HEPES (final 20 mM) and 1.8 g glucose (final 9 mM) to 900 mL water. When dissolved, add 1 mL 1 M MgCl2 and pH to 7.2. Correct volume to 1 L with water and add 1 mL 1 M CaCl2. This solution is stable for up to 1 month at 4°C, but should be discarded if it shows signs of microbial growth. 21. Complete protease inhibitors (Roche Diagnostics). 22. Geneticin/G418. 23. Complete DMEM + blasticidin + hygromycin + G418: As complete DMEM + blasticidin + hygromycin (see item 15),

28

Ward, Alvarez-Curto, and Milligan

but with the addition of G418 to a final concentration of 1 mg/mL dissolved in DMEM and added back to the bottle via a 0.22 mm filter. 24. pcDNA3 (Invitrogen Life Technologies, no longer available, replaced with pcDNA 3.1). 25. pcDNA5/FRT/TO and TOPO vectors (Invitrogen Life Tech­nologies). 26. Plasmid Maxi kit (Qiagen). 27. Endofree Plasmid Maxi kit (Qiagen). 28. pFRT/lacZeo (Invitrogen Life Technologies). 29. pcDNA6/TR (Invitrogen Life Technologies). 30. HEK-293 cells: Human embryonic kidney 293 cells (Health Protection Agency Culture Collections). 31. Dimethyl sulphoxide (DMSO). 32. 1× PBS: 120 mM NaCl, 25 mM KCl, 10 mM Na2HPO4 and 3  mM KH2PO4, pH 7.4. This solution can be made up at 10× concentrated, but pH must be checked on dilution. 33. 0.25% Trypsin-EDTA solution. 34. Lipofectamine Reagent (Invitrogen Life Technologies): Used according to manufacturer’s instructions. 35. Lipofectamine 2000 (Invitrogen Life Technologies): Used according to manufacturer’s instructions. 36. Polyethyleneimine (PEI) 25 kDa linear (Polysciences): 1 mg/mL in sterile 150 mM NaCl. Stored as aliquots at −20°C. 37. Non-essential amino acids (NEAA) (Invitrogen Life Techno­ logies).

3. Methods In order to express a GOI, it must first be subcloned into the multiple cloning site of the plasmid pcDNA5/FRT/TO. This is then co-transfected into the Flp-In™ T-Rex™ cell line with a plasmid (pOG44) expressing the Flp recombinase (Fig. 2). Upon integration of the pcDNA5/FRT/TO containing the GOI into the FRT site, the cells are rendered hygromycin resistant and zeocin sensitive, allowing selection of the required integrants. Hygromycin resistance is conferred from the pcDNA5/FRT/ TO and zeocin sensitivity is due to inactivation of the lacZeo fusion gene by the integration process. Expression of the GOI is repressed by the tet repressor protein which binds to the tet operator O 2 sequence upstream of the GOI and prevents

Using the Flp-In™ T-Rex™ System to Regulate GPCR Expression

29

transcription. In  order to induce expression tetracycline or doxycycline must be added to the culture media (Figs. 2 and 3). This then binds to the tet repressor protein, releasing it from the tet operator sequence, which in turn allows transcription and translation of the GOI. 3.1. Generation of Expression Construct

1. Design a cloning strategy for the GOI by assessing the ­multiple cloning site of pcDNA5/FRT/TO. This expression vector contains the FRT site needed for recombinase-mediated integration of the GOI into the genome (see Note 1). 2. Ligation of insert to vector, transformation into Escherichia coli (DH5a or XL-1 Blue strains) and the selection of transformants are carried out by standard molecular biological techniques (see ref. 5 for general techniques). 3. Analyse positive transformants by restriction digest of miniprep DNA and confirm reading frame and orientation by DNA sequencing (see Note 2). 4. Purify plasmid DNA containing the GOI-pcDNA5/FRT/ TO construct. Purify also the Flp recombinase vector (pOG44) ready for co-transfection (see Note 3).

3.2. Growth and Maintenance of Flp-In™ T-Rex™ -293 Host Cells

1. Flp-In™ T-Rex™ -293 host cells (derived from HEK293 parental cells) are available as frozen stocks from Invitrogen Life Technologies, but if desired they (or cell lines of alternative lineage) can be generated in the laboratory (see Note 4). This cell line stably expresses the tetracycline repressor protein (tet repressor) and contains a single integration target site or FRT site (see Note 5). 2. To thaw cells from frozen stock, remove vial from liquid nitrogen and thaw as quickly as possible at 37°C in a water bath. 3. Just before completely thawed, transfer the cell suspension to a sterile 15 mL centrifuge tube containing 10 mL of complete DMEM medium. Centrifuge at 300 × g for 5  min at room temperature. After spinning, remove media, resuspend the cell pellet in 1 mL complete DMEM and transfer to a T-75 flask containing 9  mL of fresh complete DMEM medium (without blasticidin or zeocin). Incubate cells in a humidified incubator at 37°C and 5% CO2 for 16–24 h. 4. Refresh medium using complete DMEM containing 5 mg/mL of blasticidin HCl and 100 mg/mL zeocin (complete DMEM + zeocin + blasticidin) (see Note 6). 5. Incubate and check daily until cells reach about 80% confluency when they are ready to split for further expansion. 6. We recommend freezing cells as soon as possible to keep stocks of the lowest possible passage number (see Note 7).

30

Ward, Alvarez-Curto, and Milligan

3.3. Transfection of Host Cell Lines to Generate Stable, Inducible Flp-In™ T-Rex™ -293 Expression Cell Lines

1. Seed host Flp-In™ T-Rex™ -293 cells into 10  cm dishes and  incubate with complete DMEM + zeocin + blasticidin medium. 2. Integration of the construct containing the GOI into the Flp-In™ T-Rex™ -293 genome is dependent on the Flp recombinase from the pOG44 plasmid. Therefore, an appropriate ratio of pOG44 and GOI-pcDNA5/FRT/TO plasmids may have to be determined for each experimental system. However, in our experience working with Flp-In™ T-Rex™ -293 cells, a ratio of at least 9:1 (w/w) pOG44/ GOI-pcDNA5/FRT/TO gives good results. For instance, 7.2  mg pOG44 and 0.8  mg GOI-pcDNA5/FRT/TO is a good starting point. 3. Prepare DNA mixture in ultrapure sterile water. Prepare also a negative control with empty pcDNA5/FRT/TO but pOG44. 4. Remove medium from cells and replace with complete DMEM supplemented with blasticidin, but without zeocin (complete DMEM + blasticidin). 5. Transfect DNA using the preferred method (see Note 8). 6. 24  h after transfection remove the media and replace with fresh complete DMEM + blasticidin (see Note 9). 7. 48 h after transfection split the cells into several fresh T-75 flasks to a density of less than 25% confluent. Keep cells in complete DMEM with blasticidin but without zeocin for at least another 24 h or until cells attach to flask. 8. Begin the selection of integrants by removing the medium and replacing it with complete DMEM supplemented with the optimal concentration of hygromycin B required for selection of your cell line (complete DMEM + blasticidin + hygromycin) (see Note 10). For most applications involving Flp-In™ T-Rex™ -293 cells, we find that 0.5–2 mg/mL works well. 9. Refresh medium with complete DMEM + blasticidin + hygromycin every 2–3 days until visible foci appear (see Note 11). 10. 2–20 hygromycin B resistant foci should be evident between 10 and 15 days after transfection (see Note 12). The whole polyclonal population of cells should be pooled, expanded and screened for tetracycline-regulated expression (see Note 13). Alternatively, independent clones may be isolated if desired (see Note 14). 11. Freeze a stock of cells of the lowest passage possible as soon as the cell line is established. 12. For subsequent use of these cells, the growing medium should be complete DMEM + blasticidin + hygromycin.

Using the Flp-In™ T-Rex™ System to Regulate GPCR Expression

3.4. Methods for Screening GPCR Expression in Flp-In™ T-Rex™293 Cell Lines 3.4.1. Screening for GPCR Function Using Pharmacological Assays

3.4.2. Screening for Specific DoxycyclineInduced Expression by Western Blotting

31

1. Seed Flp-In™ T-Rex™-293 expressing the GPCR of interest at the desired density with complete DMEM + blasticidin + hygromycin and incubate until attached to plastic. 2. Refresh medium with that containing the appropriate concentration of tetracycline or doxycycline to induce expression (see Note 15). 3. Incubate cells for 24 h (see Note 16). 4. Harvest cells and prepare membranes (as required, see refs. 1, 2 for method), use appropriate assay for the GPCR of interest. This method can be used with cell lines expressing GPCRs fused to an appropriate epitope tag and those for which there is a specific antiserum/antibody against the GPCR. 1. Induce cells with optimal concentration of doxycycline and for the pre-determined length of time. Incubate also a plate or flask of non-induced cells to be used as negative control (see Note 17). 2. Prepare cell lysates using 1× RIPA buffer supplemented with a protease inhibitor cocktail tablet or any other equivalent method. 3. Perform SDS-PAGE gel electrophoresis and transfer of samples onto nitrocellulose membranes according to manufacturer’s instructions. 4. Blot using specific antiserum/antibody and analyse results (6).

3.4.3. Screening by Microscopy

This method may be used when the GPCR has been tagged with a fluorescent protein and you are interested in seeing in  vivo localisation of the receptor (for example, membrane localisation, see Fig. 3). 1. Seed cells onto poly-d-lysine coated glass cover slips and incubate until attached. 2. Induce GPCR expression with the required concentration of doxycycline for 24  h. Leave one coverslip without doxycycline to use as negative control. 3. Remove cover slips off medium and wash twice in warm HEPES buffer pH 7.4. 4. Visualise with microscope. All cells present in the induced sample should show receptor expression and no fluorescence should be detected in those which are non-induced (7).

3.5. Generation of Double Stable Cell Lines in a Flp-In™ T-Rex™-293 Background

This is a very useful approach to study protein–protein interactions and particularly receptor homo- and hetero-dimerisation. It  involves the introduction of a second cDNA that expresses constitutively, into a cell line already harbouring one cDNA at the inducible locus (see Fig. 4).

32

Ward, Alvarez-Curto, and Milligan

1. Seed Flp-In™ T-Rex™-293 cells harbouring the GPCR at the inducible locus into 10 cm dishes and incubate with complete DMEM + blasticidin + hygromycin medium until 60–70% confluent. 2. Prepare DNA construct for the second GPCR, in ultrapure sterile water. 3. Transfect inducible stable cells with 5–10 mg DNA using the chosen method for transfection (see Note 18). 4. 24 h after transfection remove media and replace with fresh complete DMEM + blasticidin + hygromycin. 5. 48 h after transfection, split cells into several 10 cm dishes to various cell densities, always lower than 20% confluent, using complete DMEM + blasticidin + hygromycin (see Note 19). 6. Refresh medium with complete DMEM + blasticidin + hygromycin supplemented with G418 (complete DMEM + blasticidin + hygromycin + G418) every 2–3 days until visible foci appear (see Note 20). Foci comprising at least 20–50 cells should be evident between 10 and 15 days after transfection. 7. Pick individual clones using “cloning rings” and withdraw the cells with 100–500  mL with complete DMEM + blasticidin + hygromycin + G418 medium into 24 well plates (see Note 21). 8. Feed clones and expand until there are sufficient cells for screening (see Note 22). 9. Clones should be checked for constitutive expression of the second cDNA and inducible expression of the gene under the tet-inducible promoter control (see Fig. 4). 10. Freeze a stock of cells of the lowest passage as soon as the cell line is established.

4. Notes 1. We recommend that the GOI insert contains a Kozak sequence (such as GGATCC), immediately prior to the start codon and an appropriate stop codon in order to facilitate correct translation. We routinely include epitope amino or  carboxy terminal tags, such as Flag (DYKDDDDKC), c-Myc (EQKLISEEDL), vsv-G (YTDIEMNRLGK), or HA (YPYDVPDYA), to which well characterised commercial antisera/antibodies are available (see Chapter 4). In addition, fluorescent tags consisting of proteins based upon the many variants of GFP from Aequorea victoria may be added which can be very useful during the characterisation of receptor expression and subsequent use of the newly created cell line.

Using the Flp-In™ T-Rex™ System to Regulate GPCR Expression

33

The addition of such tags during the cloning of the receptor can be done easily using standard PCR-based techniques (see Chapter 4). The pcDNA5/FRT/TO/TOPO vector can be used as alternative system for cloning PCR fragments into pcDNA5/FRT/TO. This system uses the well-established TOPO cloning technology and does not require ligase and many other post-PCR steps making this a good option when rapid and highly efficient cloning is required. It is worth mentioning at this point that it might be helpful to design a cloning strategy that allows simple swapping of the GOI fragment from pcDNA5/FRT/TO to and from the vector pcDNA3, for example, to facilitate transient tranfection or generation of constitutive double stables (see Subheading 3.5). 2. Inserts cloned into pcDNA5/FRT/TO should be sequenced using primers annealing to the human cytomegalovirus promoter region (CMV forward: 5¢-CGCAAATGGGCGGTA GGCGTG-3¢) and to the bovine growth hormone region (BGH reverse 5¢-TAGAAGGCACAGTCGAGG-3¢) present in the backbone plasmid. It is also recommended that specific internal sequencing primers for the GOI are designed to confirm the sequence at the gene-plasmid and gene-tag ligation points. 3. Plasmid DNA must be of highest purity, with no contaminating salts, organic solvents or bacterial debris, for instance, lipopolysaccharide or endotoxin. We routinely purify DNA using Qiagen Plasmid Maxi kit or Endofree Plasmid Maxi kit, but any other equivalent method of obtaining high quality DNA is acceptable. 4. Generation of the Flp-In™ T-Rex™ -293 cell line is achieved by transfecting the pFRT/lacZeo and pcDNA6/TR plasmids into HEK-293 cells. Transfection and subsequent selection of the cells expressing the pFRT/lacZeo plasmid is performed first and the resulting cell line is used as host for pcDNA6/TR transfection. The final resulting stable cell line is resistant to zeocin, resistant to blasticidin and exhibits b-galactosidase activity as conferred by the pFRT/lacZeo plasmid. For further details on how to create such cell lines, see the Invitrogen Web site. 5. The FRT site is achieved by the integration of the pFRT/ lacZeo plasmid into a transcriptionally active region of the genome followed by zeocin selection. The tet repressor is stably expressed after integration of the original pcDNA6/TR plasmid in the genome and expression is maintained under blasticidin selection. 6. The addition of zeocin to the medium selects cells with the FRT site. Concentrations of zeocin between 50 and 1,000 mg/mL should be tested to find the minimum selective concentration

34

Ward, Alvarez-Curto, and Milligan

when generating a new Flp-In™ T-Rex™ host cell line. It is important to note that the effects of zeocin on cells are con­ siderably different from that of other antibiotics, such as hygromycin or geneticin. Cells do not round up and detach as it would be expected, but they suffer dramatic morphological changes. They augment in size and develop an abnormal cell shape, increasing the number of empty cytoplasmatic vesicles. Finally, sensitive cells break down completely only leaving “string-like” cellular debris behind. Cells that are resistant to zeocin do not present any of these characteristics and should look indistinguishable from any other cells. 7. Cells should be in their growing phase and at least 70% confluent at the time of freezing. The cells may be frozen in complete media supplemented with 10% DMSO or in FBS supplemented with 10% DMSO. Cells are washed in 1× PBS, trypsinised and pelleted by centrifugation. The cells are then resuspended in freezing medium and aliquoted into 1 mL cryo-vials that should be rapidly transferred to a −80°C freezer. For ideal cryopreservation, freezing should progress at a rate of 1°C per minute so using a “Mr. Frosty”-type Cryo Freezer container or a homemade equivalent, such as a beaker insulated with a thick layer of cotton wool, is recommended. It is advisable to check cell viability after 24 h of freezing. 8. Transfection can be carried out with standard reagents available for mammalian transfection, such as Lipofectamine or Lipofectamine 2000. A more economical alternative method uses PEI (polyethyleneimine 25 kDa linear (8)). This method does not require serum-free medium, such as Optimen (which is required for the use of Lipofectamine) and the presence of antibiotics does not interfere with the transfection. 9. Integration of the GOI into the FRT site disrupts the lacZeo gene and therefore zeocin resistance is abolished. For this reason after co-transfection of GOI-pcDNA5/FRT/TO and pOG44, the zeocin must be removed from the medium. The integration also places the hygromycin resistance gene of pcDNA5/FRT/TO into frame with the integration site ATG codon and under the control of the SV40 promoter. Expression of this gene then confers hygromycin resistance. 10. It is advisable to generate a hygromycin B kill curve of the host cells, Flp-In™ T-Rex™-293 in this case, to determine the optimal antibiotic concentration that kills any untransfected cells. We recommend testing concentrations ranging from 0.5 to 200 mg/mL. 11. Once hygromycin B selection starts, massive cell death is apparent with only a small number of cells remaining attached to the

Using the Flp-In™ T-Rex™ System to Regulate GPCR Expression

35

plastic. During this period, refresh medium very carefully to avoid losing these cells. 12. In addition to the inherent variation in the efficiency of transfection, it must be taken into account that the recombinasemediated integration of the GOI is a rare event; therefore, you should not expect a large number of integrants. However, anything between 2 and 20 foci should be expected within 10 days of transfection. If no apparent foci are visible after 15–20 days, it is likely that there is a problem with some aspect of the process. Media, reagents, plasmids, and cells should all be rechecked. 13. While maintaining Flp-In™ T-Rex™-293 inducible stable cell lines in culture, special attention must be paid to the type of FBS used to supplement the medium. The serum must be free of any traces of tetracycline as this induces gene expression upon binding to the tet repressor. Only FBS that it is either European Union approved or tested as tetracycline negative must be used. 14. Given that the host cell line is a single integrant of the FRT site, the resulting stable cell line after GOI integration is an isogenic population with each cell carrying only one copy of the GOI at that particular integration site. Therefore, it should show homogeneous levels of expression throughout and is possible to use the pooled population. However, it is still possible to isolate and screen independent clones derived from independent foci, if required. 15. We have replaced the use of tetracycline with doxycycline as it has a longer half-life than tetracycline (at least 24  h) and seems to have the same mechanism of action in this system (see the Invitrogen product literature for more information). The doxycycline concentration used as standard starting point is 1 mg/mL, but we recommend optimising this concentration for your particular system. A good experiment to choose the optimal concentration would be to titrate the amount of doxycycline required to modulate gene expression with a dose curve using concentrations ranging from 1  ng/mL to 1 mg/mL. 16. To find the optimal doxycycline concentration, it is important to find out how long it takes for gene induction to be optimal/ maximal. Therefore, we routinely carry out a time course of doxycycline induction in which we use the concentration that gives the best level of expression. 17. As a negative control, it is a good idea to have a cell lysate of host parental Flp-In™ T-Rex™-293 cells to use as further control for the specificity of the antibody/antiserum.

36

Ward, Alvarez-Curto, and Milligan

18. The second DNA construct can be in pcDNA3 or any other expression vector that is compatible with those already present. It is important to pay attention to the antibiotics used for selection in this case. 19. We recommend using a 1 mg/mL G418 concentration as a starting point. As in the cases of the other antibiotics, it is a good idea to perform a kill curve for G418 with the parental cell line to find the best drug concentration. 20. We advise that cells should be split into 10 cm dishes as this facilitates the process of picking the individual clones. Cells should be seeded at low density to ensure the generation of well separated, individual colonies. We routinely split the transfected cells 1:50 or even 1:100 at this stage. Splitting into several plates for each dilution increases the chances of obtaining positive clones. 21. To pick individual foci, we use “cloning rings” made with either 0.5 cm cut from the wide end of a 1,000 mL pipette tip or the top half of a 0.5  mL tube, with the lid and conical ­bottom removed. These rings are then autoclaved. Mark the position of the foci on the bottom of the plate and remove the media. In order to assist the detachment of the cells from the plate with trypsin, it should be washed carefully with warmed 1× PBS. Spread a very thin layer of vacuum grease on the ring and place over the foci, surrounding it. Make sure that the ring is tightly stuck to the plate and that is not overlapping with any neighbouring foci. Add 100 mL of medium into the ring and carefully pipette up and down to separate the cells from the plate. Transfer medium with cells into a well containing fresh medium in a 12 or 24 well plate. If foci are already quite confluent, 50 mL of trypsin can be used to encourage detachment of the cells. 22. Double stables can be slow to grow, particularly just after selection or while recovering from frozen storage and so it is often beneficial to add NEAA to the medium, until the cells are growing normally. References 1. Ellis, J., Pediani, J.D., Canals, M., Milasta, S., and Milligan, G. (2006) Orexin-1 receptorcannabinoid CB1 receptor heterodimerisation results in both ligand-dependent and -­independent coordinated alterations of receptor localisation and function. J. Biol. Chem. 281, 38812–38824. 2. Smith, N.J., Stoddart, L.A., Devine, N. M., Jenkins, L., and Milligan, G. (2009) The action and mode of binding of thiazolidinedione

ligands at free fatty acid receptor 1. J. Biol. Chem. 284, 17527–17539. 3. Andrews, B.J., Proteau, G.A., Beatty, L.G., and Sadowski, P.D. (1985) The FLP recombinase of the 2 micron circle DNA of yeast: interaction with its target sequences. Cell 40, 795–803. 4. Hillen, W. and Berens, C. (1994) Mechanisms underlying expression of Tn10 encoded tetracycline resistance. Annu. Rev. Microbiol. 48, 345–369.

Using the Flp-In™ T-Rex™ System to Regulate GPCR Expression 5. Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, Second Edition (Plainview, New York: Cold Spring Harbor Laboratory Press). 6. Ward, R.J., Jenkins, L., and Milligan, G. (2009) Selectivity and functional consequences of the interactions of family A G-protein coupled receptors with neurochondrin and periplakin. J. Neurochem. 109, 182–192. 7. Lopez-Gimenez, J.F., Canals, M., Pediani, J.D., and Milligan, G. (2007) The a1b-adrenoceptor

37

exists as a higher-order oligomer: effective ­oligomerization is required for receptor maturation, surface delivery, and function. Mol. Pharmacol. 71, 1015–1029. 8. van Rijn, R.M., van Marle, A., Chazot, P.L., Langemeijer, E., Qin, Y., Shenton, F.C., Lim, H.D., Zuiderveld, O.P., Sansuk, K., Dy, M., Smit, M.J., Tensen, C.P., Bakker, R.A., and Leurs, R. (2008) Cloning and characterisation of dominant negative splice variants of the human histamine H4 receptor. Biochem. J. 414, 121–131.

Chapter 3 Viral Infection for GPCR Expression in Eukaryotic Cells Antonio Porcellini, Luisa Iacovelli, and Antonio De Blasi Abstract This chapter describes the protocol for the preparation of recombinant adenoviruses and infection of target cells to transiently express G protein-coupled receptors (GPCRs) or other proteins of interest. Adenoviruses are non-enveloped viruses containing a linear double-stranded DNA genome. Their life cycle does not normally involve integration into the host genome, rather they replicate as episomal ­elements in the nucleus of the host cell, and consequently there is no risk of insertional mutagenesis. Up to 30 kb out of the 35 kb of the wild-type adenovirus genome can be replaced by foreign DNA. Adenoviral vectors are very efficient in transducing target cells in vitro and in vivo and can be produced at high titers (>1011/mL). The viral infection has a number of useful features: (1) the efficiency of gene transduction is very high (up to 100% in sensitive cells); (2) the infection is easy and does not physically alter the cell membrane for gene transduction; (3) it is possible to infect cells that are resistant to transfection with plasmids (including nondividing cells); and (4) the viral vectors can be used for infection in vivo (including gene therapy) and can potentially be targeted cell-specifically. Key words: G protein-coupled receptors, TSH, Viral vectors

1. Introduction This chapter describes the protocol for the preparation of ­recombinant adenoviruses and infection of target cells to transiently express G protein-coupled receptors (GPCRs) or other proteins of interest. This technique represents a unique tool as it allows the introduction of exogenous DNA into cellular systems, such as primary cultures that are not suitable for transfection using the traditional approaches. Viruses are obligate intracellular parasites, designed through the course of evolution to infect cells, often with great specificity for a particular cell type. They tend to be very efficient at transfecting their own DNA into the host cell, and the DNA is expressed to

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_3, © Springer Science+Business Media, LLC 2011

39

40

Porcellini, Iacovelli, and De Blasi

produce new viral particles. By replacing genes that are needed for the replication phase of their life cycle (the nonessential genes) with foreign genes of interest, the recombinant viral vectors can transduce the cell type it would normally infect. To produce such recombinant viral vectors, the nonessential genes are provided in trans, either integrated into the genome of the packaging cell line or on a plasmid. A number of viruses have been developed, but four types are mostly used: retroviruses (including lentiviruses); adeno-associated viruses; herpes simplex virus type 1 (HSV-1); and adenoviruses. Viral constructs for cell infections can be obtained commercially from different sources (see Note 1). Retroviruses are a class of enveloped viruses containing a singlestranded RNA molecule as the genome. Following infection, the viral genome is reverse transcribed into double-stranded DNA, which integrates into the host genome and is expressed as proteins. Viral infection using retrovirus requires that target cells must be in their proliferative phase. This constraint may be overcome by using lentiviruses, a subclass of the retroviruses that are able to infect both proliferating and nonproliferating cells, although they are far more complicated to use than retroviruses. Adeno-associated viruses are nonpathogenic human parvoviruses, which depend on a helper virus, usually adenovirus, to proliferate. They are capable of infecting both dividing and nondividing cells, and in the absence of a helper virus integrate into a specific point of the host genome (19q 13-qter) at a high frequency (1). HSV-1 is a human neurotropic virus that offers the advantage of allowing gene transfer into the nervous system. After infecting neurones, the wild-type HSV-1 virus can either proceed into a lytic life cycle or persist as an intranuclear episome in a latent state. Latently infected neurones function normally and are not rejected by the immune system. Though the latent virus is transcriptionally almost silent, it does possess neurone-specific promoters that are capable of functioning during latency. This chapter focuses on adenoviruses. Adenoviruses are nonenveloped viruses containing a linear double-stranded DNA genome (Fig. 1). Among over 40 serotype strains of adenovirus, most of which cause benign respiratory tract infections in humans, subgroup C serotypes 2 or 5 are predominantly used as vectors. They are capable of infecting both dividing and nondividing cells. The life cycle does not normally involve integration into the host genome, rather they replicate as episomal elements in the nucleus of the host cell, and consequently there is no risk of insertional mutagenesis. Up to 30 kb out of the 35 kb of the wildtype adenovirus genome can be replaced by foreign DNA (2). There are four early transcriptional units (E1, E2, E3, and E4), having regulatory functions, and a late transcript, which encodes for structural proteins. Progenitor vectors have either the E1 or

Viral Infection for GPCR Expression in Eukaryotic Cells

41

Fig. 1. Schematic representation of the adenoviral genome and of the expression cassette. The sites encoding for the four early transcriptional units (E1, E2, E3, and E4) are indicated. To obtain the recombinant adenovirus, the regions containing the E1a and E1b sites are deleted and the expression cassette is inserted. The expression cassette contains the cDNA of interest (the TSH-R in this case). The deletion of the E1a and E1b cassette prevents the transcription of the major late transcript; this renders the viruses defective for replication and incapable of producing infectious viral particles in target cells. The possible deletion of E3 and/or E4 present in some commercially available viruses is indicated (d E3, d E4). ITR inverted terminal repeats.

E3 gene inactivated, and the missing gene can be supplied in trans either by a helper virus, by a plasmid or it could be integrated into a helper cell genome (HEK 293 cells, (3)). Second generation vectors additionally use an E2a temperature-sensitive mutant or an E4 deletion. However, the first generation remains the most widely used (see Note 1). Viruses with E4 deleted have to be replicated in 911E4 cell lines which can complement E4. The most recent “gutless” vectors contain only the inverted terminal repeats (ITRs) and a packaging sequence around the transgene; all the necessary viral genes being provided in trans by a helper virus (4). In a recent protocol, recombination has been described in prokaryotic cells to improve the yield of recombinant viruses and to facilitate their screening (5, 6). Adenoviral vectors are very efficient in transducing target cells in vitro and in vivo and can be produced at high titers (>1011/mL). With only a few exceptions (7), it is generally reported that transgene expression in vivo from progenitor vectors tends to be transient (2). The essential steps for generating the viral vector are shown in Fig. 2. Viral infection has a number of useful features: 1. The efficiency of gene transduction is very high (up to 100% in sensitive cells). 2. The infection is easy and does not physically alter the cell membrane for gene transduction. 3. Even cells resistant to transfection with plasmids (including nondividing cells) are susceptible to viral infection. 4. The viral vectors can be used for infection in vivo (including gene therapy) and can potentially be targeted cell-specifically.

42

Porcellini, Iacovelli, and De Blasi

Fig.  2. Schematic representation of the essential steps for generating the recombinant adenovirus. The cDNA of interest (the TSH-R) is cloned into the shuttle plasmid. The viral backbone containing plasmid is composed of the sequence of the adenovirus (Ad5) without the region E1 and E3 (see Fig. 1) plus one sequence derived from the pBR322 containing the E. coli ori (origin of replication) and the Amp (ampicillin) resistance. The Bj5183 cells are co-transformed with the linearized shuttle plasmid and the viral backbone containing plasmid for homologous recombination. After recombinant selection, the DNA is digested with PacI, to remove plasmid DNA. The resulting recombinant construct contains the viral backbone plus the expression cassette and is ready for HEK293 transfection. HEK293 provides the transcription factors E1a and E1b for transcription of capsid protein RNAs and viral replication.

Viral Infection for GPCR Expression in Eukaryotic Cells

43

The protocol used is a modification of the method described by He and colleagues (5). The example presented here refers to the thyrotropin receptor (TSH-R) expressed using the pAdTrack-CMV vector.

2. Materials 1. Shuttle vectors: pShuttle-CMV (e.g., Stratagene or Q-Biogene; pAd-TRK-CMV) (see ref. 6). 2. pAdEasy-1 (e.g., Stratagene or Q-biogene). 3. BJ5183 electrocompetent cells. 4. HEK293 cells. 5. Pac I and Pme I restriction enzymes. 6. L Broth: 10  g/L Bacto-triptone, 5  g/L Bacto-yeast, and 5 g/L NaCl. 7. LB-agar: add 15 g/L of bacto-agar to L Broth and autoclave. 8. SOC Medium: 20 g/L Bacto-triptone, 5.5 g/L Bacto-yeast, 10 mM NaCl, and 10 mM KCl; autoclave then add glucose to 20 mM, MgCl2 and MgSO4 to 10 mM each. 9. Electroporation cuvettes, 0.2 cm gap. 10. Lipofectin (Invitrogen). 11. RNAaseA. 12. Neutral Red. 13. Buffer A: 50  mM glucose, 25  mM Tris–HCl pH 8.0, and 10 mM EDTA pH 8.0. 14. Lysis Buffer: 0.2 N NaOH and 1% SDS. 15. Precipitation Solution for 100 mL mix: 5 M potassium acetate 60 mL, glacial acetic acid 11.5 mL, and water 28.5 mL.

3. Methods 3.1. Preparation of Shuttle Plasmid, Adenoviral Backbone Vector, and Competent Cells

1. Subclone the cDNA encoding for the gene of interest (in our case the TSH-R) into the shuttle vector. In this protocol, we refer to the pAdTrack-CMV, which contains the GFP as tracer and the CMV promoter (6) (see Note 2). Prepare the recombinant shuttle vector at high purity (transfection grade) (see Note 3) for the next step. It can be stored at 4°C for up to 6 months. 2. Linearize the recombinant shuttle vector by incubating 1 mg DNA at 37°C for 1  h with the enzyme PmeI (4 units)

44

Porcellini, Iacovelli, and De Blasi

in  100  mL reaction buffer (provided by the manufacturer) (see Note 4). After digestion, dilute to 500 mL with TE and extract with 1 volume of phenol/chloroform/isoamidic 25:24:1 followed by ethanol precipitation. For ethanol precipitation, add 1/10 volume of 4 M LiCl, mix and add 2.5 volumes of ice-cold ethanol. LiCl is preferred as it does not interfere with the ligase or with electroporation transfection efficiency. After extraction and purification, resuspend in ultrapure water (15 mL). It can be stored at −20°C. Run 2 mL on an agarose gel to confirm the digestion (see Note 5). 3. The adenoviral backbone vector can be obtained ready-to-use from the manufacturer. This reagent can also be amplified by transforming competent DH5a (as in Subheading  3.2 but using ampicillin and not kanamycin for LB agar) and purified for further experiments (see Note 3). 4. Prepare competent cells (6) as follows: use a fresh colony or frozen stock of DH5a cells to inoculate 10 mL of LB medium in a 50 mL tube (for BJ5183 cells inoculate 10 mL LB containing 30 mg/mL streptomycin). Grow cells in a shaker overnight at 37°C. The day after, dilute 1 mL of cells into 500 mL of LB medium (for BJ5183, use streptomycin-­containing LB medium) in a 2 L flask. Grow for 2–4 h with vigorous aeration at 37°C, until A550 is ~0.5 for DH5a and A550 is ~0.7 for BJ5183. Stop cell growth by incubating on ice for 10 min to 1 h (the longer the cells are incubated the higher the competency will be). Collect the cells in two 250 mL conical centrifuge tubes. Pellet cells by centrifuging at 2,600 × g at 4°C for 10 min. Wash the pellet by resuspending in 500 mL of sterile ice-cold wash buffer (WB; 10% ultra pure glycerol, 90% distilled water v/v). Centrifuge the cell suspension at 2,500 × g for 30  min. Wash the pellet by resuspending in 250  mL of sterile ice-cold WB. Centrifuge the cell suspension at 2,500 × g for 15  min. Pour the supernatant off gently leaving about 30 mL. Resuspend and transfer the cell suspension to a 50 mL tube. Spin at 2,500 × g for 10 min, and pipette all but 5 mL of the supernatant out (for BJ5183 cells, the final total volume should be limited to 2–3  mL). Resuspend the cell pellet in the  WB remaining in the tube. Aliquot 20–40  mL per tube (the  tubes should be prechilled at −80°C) and store the aliquots at −80°C. You will need at least four aliquots for one recombination. You can store competent cells at −80°C (not in liquid nitrogen). Competent cells are commercially available. 3.2. Generation of Recombinant Adenoviral Plasmids (Fig. 2)

1. To 20 mL competent cells, add 3 mL of linearized shuttle vector and 1 mL (containing 100 ng) of adenoviral backbone vector and mix with the pipette. Transfer into a cuvette (see Note 6) for electroporation. All the reagents, the mix and the cuvette must be on ice.

Viral Infection for GPCR Expression in Eukaryotic Cells

45

2. Electroporate (Gene Pulser, Bio-Rad) at 2,500  V, 200  W, 25 mF for one pulse. 3. Add to the cuvette 500  mL of prewarmed (37°C) SOC medium, mix, transfer into a fresh tube (15 mL), and grow at 37°C for 20 min. 4. Centrifuge (800 × g for 10 min at room temperature), resuspend in 200 mL LB (SOC is fine) and plate in two 10 cm Petri dishes with LB agar with kanamycin 25 mg/mL (see Note 7). Incubate overnight at 37°C. 5. Next day, you should have colonies. Pick up 5–10 colonies (see Note 8) and grow in 3 mL LB plus antibiotic for 15–18 h. Pellet and prepare miniprep as follows: pellet 2 mL overnight culture in 2 mL Eppendorf microfuge tubes, and centrifuge at 15,000 × g for 1 min. Discard the supernatants. Add 100 mL Buffer A and vortex briefly. Add 200  mL Lysis Buffer, and gently mix by inverting the tubes several times. Add 150 mL Precipitation Solution, and mix well by inverting the tubes several times. Spin the tubes at 15,000 × g for 3 min. Pour the supernatant into fresh 1.5 mL tubes. Add 400 mL of 2-phenol/ chloroform (1:1 v/v). Spin at 15,000 × g for 5 min (at room temperature). Transfer the upper phase to a new 1.5 mL tube. Add 1  mL ethanol, leave at room temperature for 10  min. Spin at 15,000 × g for 5 min (at room temperature). Discard the supernatants. Add 500 mL 70% ethanol, vortex and spin at 15,000 × g for 5 min. Discard the supernatants. Briefly spin down (45  s at 15,000 × g) and aspirate the residual liquid in the tubes. Add 30 mL TE/RNAseA (5 mg/mL) to resuspend the DNA. Digest 10 mL of miniprep with PacI (5 units) for 1 h and run on a 0.7% agarose gel. This documents that recombination had occurred (Fig. 3). 6. Take DNA from colonies positive for recombination (as assessed by PacI digestion) and digest for the presence of the insert. For the TSH-R, take 10  mL of recombinant positive colony and digest with BstEII (5 U) for 1 h. Run on a 0.8% gel to assess for the positive band (Fig. 3) (see Note 9). 7. At this stage, you have checked out that the viral backbone has recombined correctly with the plasmid (i.e., it contains the cDNA of interest (the TSH-R) plus the antibiotic resistance), but you do not know whether this construct is able to generate infectious viral particles and to express the protein of interest (the TSH-R) in eukaryotic cells. This is analyzed in Subheading 3.3. For this purpose, transfer (at least) four different insert-positive clones (the 10  mL aliquot remaining after the two digestions) into electro-competent (RecA(−) Escherichia coli strains (such as DH5a) (see Notes 8 and 10). Prepare 100–500  mg of transfection-grade purified plasmid (see Note 2) and save stabs from corresponding colonies.

46

Porcellini, Iacovelli, and De Blasi

Fig.  3. (a) PacI digestion of candidate recombinant clones. This digestion generates two bands: the 33 kb recombinant viral DNA and the 3 kb fragment which contains the kanamycin (Kan) resistance gene. In some clones we found instead a 4.6 kb fragment (likely generated by asymmetric recombination), which turned out to generate infective virus. The presence of a 3–4.6 kb band indicates positive recombination. (b) The same clones digested with BstE II to analyze the presence of the insert (TSH-R cDNA). The digestions shown are with the viral backbone containing plasmid (lane 1), the recombination without insert (lane 2 ) and the positive clone (lane 3 ). The 9.78 kb band containing the TSH-R cDNA is indicated by the arrow. (c) Adenovirus-generated foci in HEK293 cells. 7–10 days after transfection the focus is evident and the cells express the GFP tracer. Note that at this stage you cannot see the focus unless you have a tracer (such as GFP).

3.3. Virus Production in Eukaryotic Cells (see Note 11)

1. 24  h before transfection, plate HEK293 cells (5 × 106/T75 flask). Cell should be 50–70% confluent at the transfection. 2. Digest 50 mg of transfection-grade purified plasmid with PacI (100 U) in 250 mL final volume. Run 5 mL on an agarose gel to verify the digestion. Extract and precipitate DNA. Ethanol must be removed in a sterile hood. Resuspend in 100 mL sterile ultrapure water. 3. Transfect HEK293 cells (see Note 12) with digested DNA using Lipofectin. For each flask, prepare 20  mL of PacIdigested plasmid in 1.25 mL of Optimem and 25 mL Lipofectin in 1.25  mL Optimem and leave for 15–40  min at room ­temperature. Mix these two solutions gently and leave for at least 10  min (is stable for 30–40  min). Wash the HEK293 cells with serum-free medium at least four times, add 3.5 mL Optimem and equilibrate for 15  min at 37°C in a 5% CO2 incubator. Add the mix to the flask containing Optimem and leave in the incubator for 4 h.

Viral Infection for GPCR Expression in Eukaryotic Cells

47

4. Remove the transfection medium and add 10 mL of complete medium with serum. Change the medium every other day. 5. 7–10 days after transfection scrape the cells into the medium and pellet the cells for virus extraction. At this time, unless you have a tracer protein (GFP), you will not usually see clear lysis plaques. The absence of clear plaques at this time does not indicate the absence of recombinant virus (see Note 13) (Fig. 3). 6. Wash the pellet twice in PBS and resuspend in 2 mL PBS and transfer into 1.5 or 2 mL tubes. Freeze and thaw three times in dry ice/ethanol (see Note 14). For thawing, place the ­frozen tube at 37°C until it starts to thaw and then vortex immediately. Avoid complete thawing at 37°C. 7. Spin at 12,000 × g for 30 min and collect the supernatant. 8. Use 1 mL of the supernatant (primary lysate) for the amplification. Save the other 1 mL aliquots at −80°C for the next amplification (stable for up to 6 months). 3.4. Preparation of High Titer Viral Stocks

1. Grow HEK293 cells in flasks to more than complete confluence (better to let them grow 3–4 days after confluence). 2. Wash the cells gently with prewarmed serum-free medium. 3. Add 1mL primary lysate to the cells and gently rock for 2 h at 37°C. Make sure that all the flask surface is in contact with the primary lysate. 4. Remove the primary lysate and add 10  mL medium with serum. Leave cells at 37°C for 72 h. 5. Wash the pellet twice in PBS, resuspend in 2 mL PBS, and transfer into 1.5 or 2 mL tubes. Freeze and thaw three times in dry ice/ethanol (see Note 14). For thawing, place the frozen tube at 37°C until it starts to thaw and then vortex immediately. Avoid complete thawing at 37°C. 6. Spin at 12,000 × g for 30 min and collect the supernatant. 7. Titrate the virus in the lysate (see Note 15). The expected titer is >107PFU/mL. 8. If the viral titer is as expected (see Note 16), infect 5 (or more) T75 flasks of HEK293 cells using this material. Use 2  mL PBS containing 5–10 PFU/cell and proceed as in Sub­ heading 3.4, step 3–6. 9. After centrifugation, you should have 8–10 mL lysate (about 2 mL used for freeze and thaw from five flasks). The expected titer is >109 PFU/mL. This preparation can be used to infect cells. If you need to further purify the virus suspension or if you need a higher titer (for example to infect in vivo experimental animals), proceed to cesium purification (next step).

48

Porcellini, Iacovelli, and De Blasi

10. Add 4.5  g cesium chloride to 8.5  mL of lysate and mix. Transfer this solution to an ultrafuge tube and spin at 150,000 × g (SW 41 rotor) for 18–20 h at 10–12°C. 11. Use a needle to collect the viral fraction (within the cesium chloride gradient). Dilute the collected material 1:2 with storage buffer (expected total volume is 1–2  mL), make 200 mL aliquots and store at −20°C (stable for years). 12. Titrate either by plaque assay, by GFP fluorescence determination or by OD determination of viral DNA. 3.5. Virus Titration

1. Remove all but 2 mL per well of medium from six-well plates containing 80–90% confluent HEK293 cells. Infect with appropriately diluted virus (1 mL) for 2 h. 2. Infect cells with 6 different dilution titers (e.g., 10−3 to 10−8). 3. Prepare the overlay agar as follows: autoclave 100 mL of 2.8% Bacto-Agar (Difco) and keep warm in a 45°C water bath. To 36 mL of 2.8% Bacto-Agar, add 50 mL of prewarmed 2× BME (GIBCO), 10 mL FBS, 1.25 mL of 1 M MgCl2, and 2 mL of 1 M HEPES. Mix well and swirl at 37°C in a water bath. 4. Add 4 mL/well for a six-well plate. Leave plates at room temperature for 30 min to 1 h. 5. Return the plates to a 37°C, 5% CO2 incubator. 6. On days 5–7, overlay 2–3  mL agar containing neutral red (from 100× stock, available from GIBCO-BRL) to each well. Plaques should be visible 16–30  h after the neutral red overlay.

3.6. Infection of Target Cells

1. The efficiency of infection and protein expression depends on the target cell type. You should perform preliminary experiments to determine the optimal Multiplicity of Infection (MOI = ratio between PFU and number of cells) and infection procedure. 2. For NIH3T3 cells, subconfluent cells must be covered with the minimal volume (i.e., 1.8–2 mL for a 100 mm dish) of serum-free medium (see Note 17) containing the virus at the MOI of 100 PFU/cell (see Note 18). This MOI is referred to the minimal volume of virus-containing medium on subconfluent cells. 3. After 2 h incubation at 37°C (possibly with rocking), remove the virus-containing medium and add the medium with serum. If needed, the cells can be harvested and plated by 24 h after infection. 4. The protein (TSH-R) is functionally expressed on the cell surface 36–48  h after infection and cells can be used for experiments.

Viral Infection for GPCR Expression in Eukaryotic Cells

49

4. Notes 1. A number of companies can prepare viral constructs for ­infections on demand. For example: NitAn Biothech LLC, 100 Science Village, 1381 Kinnear Road, Columbus provides custom construction services using adenovirus, lentivirus, and adeno-associated viruses (http://www.nitanbiotech.com/ viralvectors.php); Q-Biogene (Merlin) provides a custom construction service using adenovirus available for Europe (http:// www.qbiogene.com/adenovirus/products/custom/); Vector Biolabs, 3701 Market street, Ste 434, Philadelphia provides custom construction service using adenovirus and adenoassociated viruses (http://www.vectorbiolabs.com/vbs/index. html); Applied Viromics, 4160 Technology Drive, D3, Fremont, CA 94538 provides custom construction service using adenovirus and adeno-associated viruses available for the USA (http://www.appliedviromics.com/Products_1.htm). It is likely that other companies (for example, Invitrogen, Agilent Technologies, Stratagene) could provide a similar service, but it is necessary to contact them and to discuss the specific requests. 2. There are different shuttle vectors commercially available. For example, they may contain the GFP or b-Gal reporter gene. They may also have different promoters or be devoid of promoters to allow the cloning of one promoter of interest to direct the expression of the protein. 3. Use commercially available column purification or CsCl banding. 4. The linearization allows the recombination with the viral backbone and avoids the background of kanamycin-resistant colonies generated by the circular plasmid. 5. If the digestion is incomplete (i.e., <95%), you can purify the linearized DNA from the gel. In this case, extreme care must be taken to avoid any agarose (and perchlorate, if used) residual in the final DNA solution; otherwise, this impairs the electroporation transfection efficiency. 6. Use a 2 mm cuvette and not a 5 mm cuvette. 7. We suggest not to use 50 mg/mL kanamycin, since at this concentration the growth of the recombinant adenoviral plasmid is inhibited. 8. Pick the small colonies as the recombinant containing colonies are usually smaller than those of the shuttle plasmid. 9. Given the size of the recombinant adenoviral plasmid (about 40 kb) and the limited number of diagnostic restriction sites, the digestion may not be sufficient to assess the presence

50

Porcellini, Iacovelli, and De Blasi

(and the orientation) of the insert. This point should then be addressed by Southern blotting the digestion. 10. Do not use the BJ5183 strain since these cells allow further recombination of the recombinant adenoviral plasmid. 11. Safety rules for the use of adenoviruses can be found in laboratory manuals or on Web sites (for example, www-ehs.ucsd. edu/bio/biobk/bioap10.htm, or www.cdc.gov/od/ohs/ biosfty/bmbl4/bmbl4s3.htm). 12. Use care when transfecting HEK293 as these cells tend to detach from the flask, particularly as they are superconfluent. Do not expose cells to Lipofectin for >4 h. 13. At this stage, only a few cells that were infected by the virus have been lysed. After the lysis of these cells, the virus is released and infects the surrounding cells. This is the optimal time to collect the cells to extract the virus (i.e., after infection of surrounding cells and before their lysis, usually 7–10 days after transfection). If you have a GFP tracer, you will see green fluorescent (transfected) cells at day 2 and green infected surrounding cells at days 7–10. If there is no tracer, you can use control plates to see the lysis plaques at day 15–20. This control plate is no longer useful for virus collection (since many infected cells are lysed), but does tell you that the clone is able to generate lytic virus. 14. Always use polypropylene tubes. 15. For titration, you should determine the Plaque Forming Units (PFU) by standard methods. Alternatively, if the virus contains GFP tracer, simply infect superconfluent HEK293 cells with various viral dilutions and count green cells after 24 h. This gives the number of infecting particles, which in our hands corresponds to 5–10 times the PFU value. Determination of PFU is less simple but more quantitative. 16. If the titer is lower, check the transfection efficiency (should be >20%) or start from a different clone. 17. The serum must be carefully removed by washing 2–3 times. For cells that grow in multilayers (such as PC12), it can be difficult to wash out the medium. We suggest that these cells are infected in suspension (in polypropylene tubes) as follows: harvest the cells (using trypsin if needed), wash 2–3 times, and resuspend in a minimal volume (107 cells/mL) of viruscontaining medium. Incubate for 2  h at 37°C, remove the virus, resuspend in medium with serum and plate. 18. For different cell types, the MOI can vary quite substantially. In our hands, to obtain 80–100% of cells expressing the GFP, we have calculated the following MOI for U251 50PFU/cell, PC12 250 PFU/cell, U87MG 5-10 PFU/cell, COS7 50 PFU/ cell, T98G 150 PFU/cell, and for HEK293 5 PFU/cell.

Viral Infection for GPCR Expression in Eukaryotic Cells

51

References 1. Kotin, R.M., Siniscalco, M., Samulski, R.J., Zhu, X.D., Hunter, L., Laughlin, C.A., McLaughlin, S., Muzyczka, N., Rocchi, M. and Berns, K.I. (1990) Site-specific integration by adeno-associated virus. Proc. Natl. Acad. Sci. U.S.A. 87, 2211–2215. 2. Verma, I.M. and Somia, N. (1997) Gene ­therapy - promises, problems and prospects. Nature 389, 239–242. 3. Graham, F.L., Smiley, J., Russell, W.L. and Nairn, R. (1997) Characterization of a human cell line transformation by DNA from adenovirus 5. Gen. Virol. 36, 59–72. 4. Chen, H., Mack, L.M., Kelly, R., Ontell, M., Kochanek, S. and Clemens, P.R. (1997) Persistence in muscle of an adenoviral vector

that lacks all viral genes. Proc. Natl. Acad. Sci. U S A. 94, 1645–1650. 5. He, T.C., Zhou, S., da Costa, L.T., Yu, J., Kinzler, K.W. and Vogelstein, B. (1998) A simplified system for generating recombinant adenoviruses. Proc. Natl. Acad. Sci. U.S.A. 95, 2509–2514. 6. Hanahan, D. (1983) Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol. 166,557–580. 7. Geddes, B.J., Harding, T.C., Lightman, S.L. and Uney, J.B. (1997) Long term gene therapy in the CNS: Reversal of hypothalamic ­diabetes insipidus in the Brattleboro rat by using an adenovirus expressing arginine vasopressin. Nat. Med. 3, 1402–1404.

Chapter 4 Generation of Epitope-Tagged GPCRs Yan Huang and Gary B. Willars Abstract The addition of one or more epitope tags to G-protein-coupled receptors (GPCRs) has facilitated a wide variety of studies on their structure and function. Epitope-tagging is achieved using relatively straightforward molecular techniques but requires careful consideration about the nature of the epitope tag and its location within the receptor. Here, we describe both the strategies and methodologies for the generation of epitope-tagged GPCRs. We highlight a range of possible techniques that depend upon the available starting material, the nature of the epitope to be incorporated, and suggest a strategy to ease the tagging of multiple receptor types. Key words: Cloning, Competent cells, Electrophoresis, Epitope tag, Ligation, PCR, Plasmid, Primer, Restriction digest, Transformation

1. Introduction Epitope-tagging describes the addition of a small antigenic ­peptide (the epitope tag) into a protein molecule so that the resulting product can be recognized by an antibody (generally commercially available) against the epitope tag and/or can be monitored or visualized due to the addition of a tag with fluorescent or enzymatic activity. Typically, epitope-tagging involves inserting a cDNA sequence encoding an epitope into a gene of interest and expressing the gene product as a fusion protein (epitope-tagged protein) in an appropriate host, such as bacteria (e.g., Escherichia coli), yeast (e.g., Saccharomyces cerevisiae), or eukaryotic (e.g., mammalian and insect) cell lines. The ability of antibodies to bind to the epitope tag or for the epitope tag to be visualized by other means (e.g., fluorescence or enzymatic ­activity) allows for protein purification and investigation of, for example, protein expression including subcellular localization (Fig.  1).

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_4, © Springer Science+Business Media, LLC 2011

53

54

Huang and Willars

Fig. 1. Some potential applications of epitope-tagging.

Table 1 Common epitope tags Tag

Residues Sequence/origin

Detection

Purification

FLAG-

8

DYKDDDDK

M1, M2, M5

Immunoaffinity

His-

6

HHHHHH

Anti-His

Metal affinity

HA-

9

YPYDVPDYA from human influenza hemagglutinin

12CA5, 3F10

Immunoaffinity

Myc-

10

EQKLISEEDL from human c-myc protein

9E10

Immunoaffinity

GFP-

238

Green fluorescent protein or variants

Anti-GFP

Immunoaffinity

GST-

220

Glutathione-S-transferase

Anti-GST

Glutathione

WSAPQFEK

Strep-Tactin

Strep-Tactin

Strep-

8

S-

15

KETAAAKFERQHMDS from pancreatic ribonuclease A

Anti-S peptide

S-peptide

Avi-

16

GLNDIFEAQKIEWHE encoding biotin-acceptor peptide 1

Avidin

Avidin

AEEGKLVIW encoding maltose-binding protein

Anti-MBP

Maltose

MBP-

9

As a consequence, epitope-tagging has become a ­standard molecular tool in the study of G-protein-coupled receptor (GPCR) structure, function, and regulation. 1.1. Common Epitope Tags for GPCRs

A wide variety of validated anti-epitope antibodies are c­ ommercially available (see Table 1) along with a range of tags having fluorescent or enzymatic activity. For GPCRs, ideally the epitope tag should not interfere with ligand binding, receptor function, or the processing and trafficking of the receptor.

Generation of Epitope-Tagged GPCRs

55

However, it should provide strong immunoreactivity or allow for maximum ease of use by other analytical techniques. There is no single most appropriate epitope tag for all GPCRs and for all applications. Further, it is difficult to reliably predict how a particular tag will behave in a particular GPCR, or indeed how the behavior of a GPCR will be influenced by the presence of an epitope tag. An obvious starting point is to review the literature for examples of tags added to the GPCR of interest or a related family member as this may provide some clues on how best to tag the receptor for specific uses. If little information is available, it is perhaps worth making several constructs with different epitope tags. Unless there are specific reasons for requiring an N-terminal epitope tag (e.g., use in ELISA assays or other means of specifically monitoring cell-surface expressed receptor), the C terminus of the receptor is a suitable starting point. This strategy would hopefully allow selection of a receptor that, in a range of binding and functional assays, behaved identically to the untagged receptor and produced a tag that behaved as required (e.g., immunoreactivity, fluorescence or enzymatic activity). Among the many epitope tags, which may display different features, the HA-, Myc-, FLAG-, His-, and GFP-tags are perhaps the commonest used to tag GPCRs in eukaryotic expression systems (Table 1). The FLAG-tag was the first commercial epitope tag and was especially designed for protein purification (1). This tag consists of eight amino acids (Table 1), including an enterokinase-­cleavage site (D-D-D-D-K↓-X-), which allows removal of the FLAG-tag from the protein. Several highly specific anti-FLAG antibodies have been developed, including three monoclonal antibodies (denoted as anti-FLAG M1, M2, and M5) and polyclonal antibodies, each with different recognition and binding characteristics. Anti-FLAG M1 binds the epitope in a calcium-dependent manner, which allows FLAG-tagged proteins to be eluted from the antibodies with calcium chelating agents. However, antiFLAG M1 does not recognize either N-terminal FLAG-tagged proteins preceded by a methionine (Met, start codon; MetFLAG-tagged proteins) or C-terminal FLAG-tagged proteins. Anti-FLAG M5 was raised against the sequence of Met-FLAG and hence exhibits a higher affinity for N-terminal ­Met-FLAG-tagged proteins. Anti-FLAG M2 will bind to both N-terminal (including Met-FLAG) and C-terminal FLAG-tagged proteins and is, therefore, more generally applicable. However, M2 and M5 bind to the epitope in a calcium-independent manner, meaning that bound antigens cannot be eluted from affinity columns by calcium chelation (2). The His-tag, composed of six histidine residues (Table  1), is most commonly used for the purification of recombinant proteins by means of metal chelation chromatography, as six consecutive histidine residues form a structure that binds nickel (3).

56

Huang and Willars

The His-tag combines the advantages of small size and ­interaction with a chromatographic matrix (e.g., Ni-NTA resin) that is relatively inexpensive, can withstand multiple regeneration cycles under stringent conditions, and exhibits a high binding capacity. Moreover, elution conditions are mild and flexible. His-tags placed as affinity tags at either the N terminus or C terminus enable purification of the desired protein from the crude extract of the host cells in a single step of immobilized metal affinity chromatography (IMAC). Although this can also be used to detect protein–protein interactions, it is generally considered to be less sensitive. In addition, the use of His-tags and IMAC purification is not recommended for proteins containing metal ions. Similarly, other amino acids including cysteine and naturally occurring histidine-rich regions in host proteins may result in unwanted protein binding during IMAC purification (4). Human influenza hemagglutinin (HA) is a surface glycoprotein required for the infectivity of the human influenza virus. An anti-HA antibody (Clone12CA5, mouse IgG2b/k monoclonal antibody), which recognizes amino acids 98–106 of the HA-molecule (Table 1), was originally used to study how the immune system recognizes the HA virus (5). However, this small peptide sequence has now been extensively used as a general epitope tag (HA-tag) (6). Many HA-tagged proteins have been engineered and expressed, in which the epitope is located at either the N terminus or the C terminus without interference with the bioactivity or subcellular localization of the recombinant protein. The disadvantage of HA-tag is that the 12CA5 often cross-reacts with other mammalian cellular proteins in immunoblotting. Recently, another antiHA-tag antibody (clone 3F10, rat IgG1 monoclonal antibody) has become available from Roche (Basel, Switzerland), which recognizes the same peptide sequence as 12CA5 but with the higher affinity. At the lower concentrations required to identify HAtagged proteins, 3F10 has lower cross-reactivity with other proteins and is well-suited for immunoprecipitation. In addition, the rat origin of 3F10 permits dual localization of a second tag probed using a mouse monoclonal antibody. The Myc-tag consists of ten residues (Table 1) corresponding to amino acids 410–419 of the transcription factor c-myc, a 62-kDa proto-oncoprotein (p62) that plays a role in cell cycle regulation, metabolism, apoptosis, differentiation, cell adhesion, and tumorigenesis. The Myc-tag can be detected when attached to either the C terminus or N terminus of a receptor. The 9E10 antibody (mouse monoclonal IgG1/k antibody) allows the Myctagged protein to be analyzed and visualized using immunochemical methods (7). In addition, Myc-tagged proteins have been purified in a single-step procedure using 9E10 antibodycoupled affinity columns after expression in a variety of recombinant expression systems.

Generation of Epitope-Tagged GPCRs

57

The 238 amino acid (26.9  kDa) green fluorescent protein (GFP) was originally purified from the jellyfish Aequorea victoria (8). For fluorescence, GFP has two excitation peaks, a major one at 395  nm and a minor one at 475  nm, with peak emission at 509  nm in the lower green portion of the visible spectrum. Although much larger than most other epitope tags, GFP has been used extensively to study the trafficking and localization of receptors as it has a number of advantages which includes the following: (a) when illuminated in living cells, GFP is much less phototoxic than most other small fluorescent molecules such as FITC (fluorescein isothiocyanate); (b) GFP-tagged proteins can be directly visualized in live cells using fluorescent microscopy; (c) GFP fluorescence is stable under cell fixation, cell permeabilization, or additional labeling steps and hence suitable for use with immunocytochemistry (ICC) of other epitope tags; and (d) GFPfused GPCRs are more resistant to photo-bleaching with a low background fluorescence than antibodies or ligands, which allows investigators to visualize proteins for a longer duration in an intact cellular environment. Since its discovery, many different GFP mutants have been generated to enable easier detection or, for example, alteration in the fluorescent properties under different conditions (e.g., altered pH). The first major improvement was a single point mutation (S65T), which resulted in increased fluorescence, photo-stability, and a red-shift of the major excitation peak to 488 nm. This excitation range is more compatible with the commonly available optical filters and confocal laser applications (9). Addition of a further mutation to this construct (F64L, S65T) resulted in the generation of enhanced GFP (EGFP; Ex 488 nm, Em 509 nm), which has the dual advantage of increased fluorescence intensity (35× that of GFP) and much higher expression in mammalian cells. EGFP-tags have been used widely in mammalian and yeast expression systems (10, 11). Many color mutants have been made, including blue fluorescent proteins (EBFP, EBFP2, Azurite, and mKalama1), cyan fluorescent proteins (ECFP, Cerulean, and CyPet), and yellow fluorescent protein derivatives (YFP, Citrine, Venus, and YPet). These can also be used as epitope tags providing a range of tags with different properties and allowing coincident imaging of multiple proteins. Further, these have provided tags that can be used for either fluorescence resonance energy transfer (Chapter 21) or bioluminescence resonance energy transfer (Chapter 20). Rabbit polyclonal anti-GFP antibodies are widely available from suppliers (e.g., ab290, Abcam Inc., Cambridge, MA, USA) and react well to all GFP variants. 1.2. Location of the Epitope Tag

In the majority of cases, the epitope tag should be placed at either the N terminus or the C terminus of the protein coding sequence as this choice of location minimizes any potential effects of the tag

58

Huang and Willars

on either the structure or function of most proteins (12). Locating the tag at one end of the protein also maximizes the probability that the tag will be recognized by an antibody, since it is less likely to be buried by the native folding of the target protein. N-terminal epitope-tagging of a GPCR has the potential advantage of allowing monitoring of cell-surface expression by, for example, ICC or ELISA. However, N-terminal tagging may not always be suitable. Although 90% of GPCRs (most Family A members) do not have a signal peptide and use the first transmembrane domain as a uncleaved signal anchor sequence, most Family B and C members are predicted to contain a cleaved signal peptide, in which an epitope tag placed N-terminal to the signal peptide would, therefore, be absent from the mature receptor. Although possible, inserting the epitope tag immediately after the signal peptide has potential problems such as the following: (a) although the cleavage site of the signal peptide can be predicted, this may not be accurate; (b) insertion of an epitope tag sequence at the predicted cleavage site could influence cleavage; and (c) Family B and Family C GPCRs generally have highly structured and large N-terminal domains that contribute to ligand binding such that insertion of a tag may influence ligand binding (and possibly trafficking). For any GPCR, large epitope tags such as GFP inserted at the N terminus may be problematic as they may influence receptor structure, trafficking, and function. Generally, C-terminal tagging with large epitopes is more suited for preserving the native localization and function of GPCRs (13, 14) and indeed other proteins (15). Although the C-terminal domains of GPCRs can be involved in aspects of receptor function including activation, desensitization, and interaction with intracellular proteins, C-terminal tagging has been successfully applied to many receptors with no apparent consequences. An additional advantage of a C-terminal tag is that the detection of the tag will indicate expression of a full-length construct. The various potential advantages and disadvantages of the different tags and locations within a GPCR emphasize the need to very carefully compare the properties of the tagged and untagged receptors wherever possible (16).

2. Materials 1. H2O: ultra pure water, 18 MW quality (e.g., ELGA Labwater, Marlow, UK). For all experiments, ultrapure water should be autoclaved at a temperature of 121°C for 15 min. 2. SOC medium (see Note 1): 0.5% (w/v) yeast extract, 2% (w/v) tryptone, 10 mM NaCl, 2.5 mM KCl, 10 mM MgCl2, 10 mM MgSO4, and 20 mM glucose. Store at 4°C in a fridge in 10-mL aliquots.

Generation of Epitope-Tagged GPCRs

59

3. Luria-Bertani (LB) broth: 1% (w/v) tryptone, 0.5% (w/v) yeast extract, and 1% (w/v) NaCl. 4. LB agar: 1% (w/v) tryptone, 0.5% (w/v) yeast extract, 1% (w/v) NaCl, and 1.5% (w/v) agar. 5. Ampicillin stock (×1,000): 100 mg/mL in H2O, sterilized by passing through a 0.2 mm filter (do not heat), stored in 600-mL aliquots, −20°C. 6. Kanamycin stock (×1,000): 50 mg/mL in H2O, sterilized by passing through a 0.2 mm filter (do not heat) and stored in 600 mL aliquots at −20°C. 7. Annealing buffer: 10 mM Tris-base, pH 8.0, 50 mM NaCl, and 1 mM EDTA, autoclaved and stored at room temperature. 8. TE buffer (pH 8.0): 10 mM Tris-base and 1 mM EDTA, sterilized with 0.2 mm filter, and stored at room temperature. 9. 50% (v/v) glycerol: autoclaved and stored at room temperature. 10. 10× DNA loading buffer: 50% (v/v) glycerol, 50 mM EDTA, and 0.5% (w/v) bromophenol blue, stored at −20°C. 11. Buffer P1: 50  mM Tris–HCl, pH 8.0, 10  mM EDTA, and 100 mg/mL RNase A, stored at 4°C. 12. Buffer P2: 200  mM NaOH and 1% (w/v) SDS, stored at room temperature. 13. Buffer P3: 3  M potassium acetate, pH 5.5, stored at room temperature. 14. Polymerase chain reaction (PCR) tubes (see Note 2): ­available from many suppliers (e.g., Fisher Scientific, Loughborough, UK). 15. Bunsen burner: manipulations should use a Bunsen burner flame to prevent contamination during transfer into or out of containers. 16. Restriction enzyme stocks (10 or 20  U/mL, New England Biolabs, Hitchin, UK), supplied with one of 10× digestion buffers (NEBuffer 1, 2, 3, or 4) and 100× bovine serum albumin (BSA) stock (10 mg/mL) stored at −20°C (see Note 3). 17. T4 Ligase (400,000 U/mL, New England Biolabs, Hitchin, UK) supplied with 10× ligase buffer stored at −20°C (see Note 4). 18. Agarose (Genflow, Fradley, UK), stored at room temperature. 19. TAE (Tris-base/acetate/EDTA) buffer: 0.4 M Tris acetate, 10 mM EDTA, pH 8.3, stored at room temperature. 20. DNA ladder (1 kb plus, Invitrogen, Paisley, UK): DNA size marker for 100 bp to 12 kb range, stored at −20°C.

60

Huang and Willars

21. GelRed™ (Cambridge Bioscience, Cambridge, UK). A ­nontoxic nucleic acid staining solution, stored at room temperature in the dark. 22. pEGFP-N1 (Clontech, Oxford, UK): DNA vector containing an EGFP epitope tag for expression of proteins fused with a C-terminal EGFP epitope tag stored at −20°C. 23. pcDNA3.1(+) (Invitrogen, Paisley, UK): DNA vector for expression in mammalian cells, stored at −20°C. 24. Nescofilm (Bando Chemical Ltd, Kobe, Japan): sealing film. 25. Commercial plasmid DNA mini-prep kit for the preparation of plasmid DNA from E. coli, for example: High Pure plasmid isolation kit (Roche Applied Science, Mannheim, Germany); NucleoSpin Plasmid (Fisher Scientific, Loughborough, UK ) and; GenElute™ HP plasmid Miniprep kit (Sigma-Aldrich, Poole, UK). After adding the RNase, the cell resuspension buffer should be stored at 4°C and the remainder of the contents stored at room temperature unless otherwise directed. 26. QIAquick gel extraction kit (QIAGEN, Crawley, UK): for the extraction of DNA fragments from agarose gel, supplied with QG and PE buffers, stored at room temperature. 27. Vent DNA polymerase: supplied with 10× Thermo Pol buffer and stored at −20°C. 28. 100 mM MgCl2: stored at 4°C after autoclaving. 29. 100 mM CaCl2: stored at 4°C after autoclaving. 30. Competent E. coli resuspension buffer: 100 mM CaCl2, 13.3% v/v glycerol, stored at 4°C after autoclaving. 31. Microbiological wire loop (Sigma-Aldrich, Poole, UK). 32. 70% IMS: 70% industrial methanol in a spray bottle, stored at room temperature. 33. Sterile disposable 10-mL plastic loop (Fisher Scientific, Loughborough, UK).

3. Methods The methods used to generate epitope-tagged GPCRs involve recombinant DNA technology, in which a DNA fragment (linear piece of double-strand DNA) encoding the GPCR sequence of interest (with or without an epitope tag sequence) is introduced into a DNA vector containing (or not containing) an epitope tag sequence. This technique depends upon the ability of restriction endonucleases (REs, see Note 3) to produce single-stranded sticky ends in a double-stranded DNA fragment, which can form hydrogen bonds with a complementary sticky end of another

Generation of Epitope-Tagged GPCRs

61

Fig. 2. Flow chart showing protocols for generating epitope-tagged receptors. Solid lines represent “yes” and dotted lines represent “no.” Once a decision has been made regarding the availability of a commercial or premade vector containing an epitope tag, the protocol should follow lines of one shade throughout (i.e., black, gray, or open gray ).

fragment, allowing the fragments to be ligated (see Note 4). The following protocols can be applied to produce the DNA fragments (both vectors and inserts) with sticky ends that are needed to generate epitope-tagged constructs by a variety of strategies. The strategy required depends upon the starting materials (Fig. 2). For example, (1) if a commercial vector containing an epitope tag sequence is available, follow Subheading  3.1 to insert a DNA fragment encoding the receptor into this vector; (2) if a suitable vector is available without the epitope tag sequence, follow Subheading 3.2 as an example of how to introduce both receptor and epitope tag sequence. Irrespective of the starting strategy, Subheading 3.3 should be followed to ensure the plasmid contains the correct sequence before use in expression studies.

62

Huang and Willars

3.1. Generation of Epitope-Tagged GPCRs Using Commercially Available Vectors

A variety of vectors containing epitope tags are commercially available. For example, the mammalian expression vector pEGFPN1 (see item 22), which by inserting a DNA fragment encoding the GPCR of interest generates a construct containing an EGFPtag fused to the C terminus of the GPCR. The protocols described can be used to prepare both the insert (see Subheadings 3.1.1  and  3.1.2) and the vector with compatible sticky ends (see Subheading 3.1.3), which can then be ligated (see Note 4, Fig. 3).

3.1.1. Subcloning the cDNA Sequence of a GPCR from an Available DNA Construct

If an appropriate construct (plasmid) is available, the following protocol can be employed to isolate the required sequence by RE digest (Fig. 3). The construct should have (a) the cDNA sequence (see Note 5) encoding the GPCR of interest with appropriately positioned start and stop codons (see Note 6) and (b) appropriate restriction sites flanking both ends of the GPCR sequence (see Note 7). The quality of DNA will contribute to the extent of digestion when REs are used to excise the GPCR sequence. If sufficient DNA is available, start the procedure at step 15. However, if more plasmid is required (or the DNA is found to be of poor quality following agarose gel electrophoresis at step 16), then steps 1–14 are required, which describe the preparation of plasmid from either an existing stock of the plasmid (steps 1–14) or a frozen stock of the bacteria containing the relevant plasmid (steps 11–14).

Fig. 3. Generation of an epitope-tagged GPCR using a commercially available vector. A DNA fragment containing sticky ends and the cDNA sequence of the required GPCR can be made by either PCR or subcloning from an existing construct. Using a commercial vector containing an epitope tag, sticky ends that are compatible to those of the insert are made by RE digest in the MCS (multiple cloning site) of the vector. The insert and linear vector are then ligated to form a new construct construct encoding encoding the epitope-tagged GPCR. In this example, the epitope tag sequence is fused to the C terminus of the GPCR.

Generation of Epitope-Tagged GPCRs

63

1. For each plasmid required, warm to room temperature two premade LB agar plates (see Subheading  3.4.1) containing appropriate antibiotics (depending on the antibiotic-­resistance gene carried by the vector). 2. Remove a 50-mL aliquot of DH5a competent cells (generally stored in a 1.5-mL microfuge tube; see Subheading 3.4.7 and Note 8) from −80°C and thaw on ice (~10 min). This is generally best started early afternoon, allowing an overnight incubation following step 9. 3. Pipette 1–10  mL of cDNA construct (~10  ng) directly into the tube containing the competent cells and mix by tapping gently. Do not pipette up and down. 4. Incubate on ice for a further 10 min. 5. Heat shock for exactly 90 s at 42°C (a water bath is ideal for this). Do not mix or shake. 6. Place tube on ice for 2 min. 7. Add 800 mL of prewarmed SOC medium (item 2). 8. Secure the tube in, for example, a rack and leave at 37°C with shaking for 1  h (a standard laboratory orbital shaker set at ~225 rpm is ideal for this). 9. Over two separate LB plates, spread 30 and 100 mL of cells (the aim is to allow the subsequent picking of individual colonies from the plates following growth) and incubate upside down at 37°C overnight. 10. Where a frozen glycerol stock of the bacteria containing the relevant plasmid is available, take a scraping of the frozen stock (using a sterile disposable plastic loop, see item 33) and streak onto the surface of LB plate in such a way as to provide a dilution of the stock, which will at some point allow individual colonies to be picked following growth. Incubate the plate upside down at 37°C overnight. 11. The following morning, plates can either be used or stored short term (~1 month). If storage is required, seal the lid with Nescofilm (see item 24) or equivalent and store the plates inverted at 4°C. 12. Pick a single colony from the LB plate (using a sterile microbiological wire loop, see item 31) and inoculate 5  mL LB broth (see item 3) containing 5 mL of 1,000× stock of appropriate antibiotic (e.g., ampicillin or kanamycin, see items 5 and 6) in a 30-mL sterile universal tube. Grow the bacteria overnight at 37°C with shaking (230 rpm). 13. Prepare the plasmid DNA using a commercially available plasmid DNA mini-prep kit (see item 25) following the manufacturer’s instructions, finally dissolving the DNA pellet with 55 mL of TE buffer (see item 8).

64

Huang and Willars

14. After determining the plasmid DNA concentration (see Subheading 3.4.4), store at −20°C until use. 15. Perform an appropriate RE digestion on 2  mg of plasmid DNA (see Subheading 3.4.5). In this instance, the digestion should be performed for 2 h at 37°C (water bath). 16. Separate and purify the DNA fragment by gel electrophoresis (see Subheading  3.4.2) and extract from the gel using a QIAquick gel extraction kit (see Subheading 3.4.3). 17. Store the purified DNA fragment at −20°C until use. 3.1.2. PCR Cloning of a GPCR cDNA Sequence with Appropriate Restriction Enzyme Sites

If a suitable plasmid (i.e., one containing the GPCR sequence of interest with appropriate restriction sites) is not available, then the GPCR sequence with RE sites may be generated by PCR (see Note 9) using a plasmid, a DNA fragment, or a cDNA library, which contains the cDNA sequence of interest (see Note 5) as the template (provided they contain the GPCR of interest) (Fig. 3). 1. Primer design The following should be considered when designing the 5¢~primer (forward primer) and 3¢~primer (reverse primer). (a) Both primers must contain a portion (at least nine bases) of the relevant coding sequence of the receptor.   C-terminal tagging: the coding sequence of the 5¢~primer should include a Kozak sequence (GCCACC) to enhance the efficiency of ribosomal translation in eukaryotic cells (17) followed by a start codon (ATG) (see Note 6). The 3¢~primer should not contain a stop codon (see Note 6).   N-terminal tagging: the sequence coding for the receptor in the 5¢~primer should not contain a start codon. However, a stop codon should be included in the 3¢~primer (see Note 6). (b) Appropriate RE sites (see Note 7) with flanking bases should be added to the 5¢ end of both the 3¢~primer and the 5¢~primer (see Note 10). (c) Primers should be between 18 and 30 bases in length, with a melting temperature (Tm) of 50–60°C (see Note 11). (d) Optimally, the primers should have a minimum GC content of 40% and should terminate in one or more C or G bases (see Note 12). 2. Perform the PCR and store the reaction mixture at −20°C until agarose gel electrophoresis (see Subheading 3.4.6). 3. Detect and separate the PCR product from the reaction ­mixture by gel electrophoresis (see Subheading 3.4.2).

Generation of Epitope-Tagged GPCRs

65

4. Excise and isolate the PCR product band from the gel (see Subheading  3.4.3). The final step of these protocols allows elution of the purified PCR product with 30 mL of H2O. 5. Apply the entire PCR product to an overnight (16–20 h) RE digestion in a PCR tube (see item 10) using a thermocycler (see Subheading 3.4.5 and Note 10). 6. Purify the RE-digested PCR product by the process described below for purifying a DNA fragment from solution (see Subheading 3.4.3). Remember, the DNA fragment should be finally eluted from the column with 30 mL of TE buffer rather than H2O (see Note 20). 7. Store the RE-digested DNA fragment at −20°C until use. 3.1.3. Prepare a Linear Vector Containing the Required Epitope Tag(s)

The circular plasmid requires linearization and the generation of sticky ends that are compatible with those of the DNA fragment that encodes the GPCR of interest. 1. Digest 2 mg of vector (plasmid DNA) as described below (see Subheading 3.4.5) for 4 h or overnight (see Note 10). 2. Separate and purify the linear vector by gel electrophoresis (see Subheading  3.4.2) and extract from the gel (see Subheading 3.4.3). 3. Elute the linear vector from the column with 30 mL of TE buffer (see step 8) and store at −20°C.

3.1.4. Insertion of a cDNA Fragment Encoding a GPCR Sequence into a Linear Vector by Two-Part Ligation

1. Prepare the ligation mixture in a total volume of 20 mL. This consists of 1 mL of T4 DNA ligase (see step 17), 2 mL of 10× T4 ligase buffer (see Note 4), 4 mL of the GPCR cDNA ­(prepared either from Subheading 3.1.1 or Subheading 3.1.2), 0.5 mL of linear vector (prepared from Subheading  3.1.3), and an appropriate volume of H2O (see Note 14). 2. Conduct the ligation either at room temperature for 3–16 h or at 4°C for approximately 48 h (this can be useful over a weekend). 3. Use ~7.5 mL of the ligation mixture in a bacterial transformation as described above (see step 1–9 in Subheading 3.1.1) with the exception that at the end of the reaction, all the bacteria should be spread onto a single LB plate prepared with the appropriate antibiotic. Before spreading the bacteria, concentrate the mixture by centrifugation (9,000 × g, 2 min in a bench-top microfuge), removing 700 mL of the SOC media and gently resuspending the cells in the remaining media (see Note 8). 4. After incubating the plate at 37°C overnight, check and assess the number of colonies. An excessive number of colonies might suggest an unsuccessful ligation due, for example, to

66

Huang and Willars

the re-ligation of vector that had only a single RE cut. Seal the lid with Nescofilm (see item 24) and store inverted at 4°C. 3.2. Introduction of Synthesized Epitope Tag(s) into Specific Sites Within a GPCR Sequence 3.2.1. Annealing Two Complementary Oligos Encoding an Epitope Tag

This strategy can be used for generating a C-terminal and/or N-terminal epitope tag. The example below describes the generation of a C-terminal HA-tagged glucagon-like peptide 1 (GLP-1) receptor (GLP-1R-HA). This receptor is a Family B GPCR, which contains an N-terminal signal peptide that is cleaved during receptor synthesis (18). 1. Design of complementary oligos for the synthesis of a C-terminal HA-tag (incorporating a XhoI RE site at the N terminus, XbaI RE site at the C terminus and an additional SalI RE site immediately before the XbaI site; see Note 15) (Fig. 4). 2. Dilute the oligos with annealing buffer (see item 7) to, for example, 100 nM. Mix the complementary oligos at equimolar concentrations (e.g., 100 nM, 50 mL of each) in a PCR tube. 3. Using a thermocycler, run the following protocol: (1) heat to 95°C and maintain for 2  min; (2) slowly cool to 25°C for over 45–60 min (e.g., 1.5°C/min) and maintain for 10 min; and (3) hold at 4°C until the next step. 4. Briefly microfuge (1,500 × g, 1  s) to draw all the moisture from the lid. Store the tube on ice or at 4°C until use. 5. Prepare the GPCR insert (e.g., GLP-1R with BglII-XhoI RE sites) using PCR as described in Subheading 3.4.6 with the following details: (a) Template: 0.4  mg of untagged, wild-type (WT) human GLP-1R in pcDNA5-FRT. The linear sequence of the GLP-1R is given in the NCBI database (Accession, NP_002053; Version, NP_002053.3; GI:166795283).

Fig. 4. Oligos for synthesizing the C-terminal HA- tag. Bases with a solid underline represent sticky ends within RE sites (see Note 3), XhoI at the N terminus and XbaI at the C terminus. Bases with a dashed underline encode a Sal I RE site that has been incorporated to allow the plasmid to be used in the future with other constructs (see Note 15). The bases highlighted within the open square boxes have been added such that the ATC sequence of the XbaI RE site (highlighted in gray ) are now in-frame with the HA-tag and form a stop codon (the complementary bases are TAG). The HA-tag sequence is the 27-base pairs between the XhoI and Sal I sites.

Generation of Epitope-Tagged GPCRs

67

Fig. 5. Primers for cloning the GLP-1R (Bgl II-XhoI). Bases with a solid underline represent the RE sites as indicated. The flanking bases (see Note 10) are indicated by a dashed underline. Bases in italic represent the Kozak sequence. Bases with a double underline represent the start codon (see Note 6). The GLP-1R cDNA sequences are shown as indicated. The full length GLP-1R cDNA sequence is 1,389 bp, in which bases 1–3 are the start codon and bases 1,387–1,389 are the stop codon.

(b) 5¢~ and 3¢~primers (Fig. 5): 0.5 mL of each, designed as described previously (see Subheading 3.1.2, step 1). (c) DNA polymerase: 1  mL of Vent DNA polymerase (see item 27). (d) Ta: 59°C and 30 cycles in total. 6. Purify PCR product as described above (see Subheading 3.1.2, steps 3–4). 7. Digest purified PCR product at 37°C for 20 h (see Subheading 3.4.5 and Note 7) in a reaction of 50 mL as below: PCR product

29.0 mL

10× NEBuffer 3

  5.0 mL

100× BSA (10 mg/mL)a

  0.5 mL

BglII (10,000 U/mL)a

  1.2 mL

XhoI (20,000 U/mL)

  0.6 mL

H2O

13.7 mL

a

 Stock concentrations

a

8. Linearise the plasmid pcDNA3.1(+) by a digest at 37°C for 4 h (see Subheading 3.4.5) in a reaction of 50 mL as below: pcDNA3.1(+) (0.56 mg/mL)a

  3.5 mL

10× NEBuffer 3

  5.0 mL

100× BSA (10 mg/mL)

  0.5 mL

BamHI (20,000 U/mL)a

  0.6 mL

XhoI (20,000 U/mL)a

  0.6 mL

H2O

39.8 mL

a

 Stock concentrations

a

68

Huang and Willars

9. Ligate the DNA fragments (see Notes 4 and 14) at room temperature overnight in 20 mL of ligation mixture as given below: Fragment 1

C-terminal HA-tag (XhoI-XbaI)

  2.0 mL

Fragment 2

GLP-1R (BglII-XhoI)

  4.0 mL

Fragment 3

pcDNA3.1(+) (BamHI-XbaI)

  0.5 mL

10× T4 Ligase buffer

  2.0 mL

T4 Ligase (400,000 U/mL) H2O

a

  1.0 mL 10.5 mL

 Stock concentration

a

10. Use ~10 mL of the ligation mixture for the transformation of 100  mL of DH5a competent cells as described previously (Subheading 3.1.4, step 3). 3.2.2. Elongation of a GPCR Sequence to Incorporate an Epitope Tag

This strategy can be used for generating a C-terminal and/or N-terminal epitope tag, although the example below describes the production of an N-terminal HA-tag for generating either an N-terminal HA-tagged GLP-1R (HA-GLP-1R) or an N-terminal HA-tagged and C-terminal EGFP-tagged GLP-1R (HA-GLP1R-EGFP). 1. Primer design (Fig. 6). 2. Cloning using a two-part PCR strategy. Stage 1: Perform the first PCR using the 5¢~primer A and 3¢~ primer A (Fig. 6) as described (see Subheading 3.2.1, step 2)

Fig.  6. Example of primers for generation of N-terminal HA-tags. Bases with a solid underline represent restriction enzyme sites as indicated. Flanking bases are indicated by a dashed underline (see Note 10). Bases in italics are the Kozak sequence. The start codon (ATG) is highlighted by a double underline and the stop codon (TCA) is in a box (see Note 6). The HA-tag sequence and the GLP-1R cDNA sequences are indicated.

Generation of Epitope-Tagged GPCRs

69

to generate the GLP-1R with a truncated N-terminal HA-tag (23 bases). Purify the PCR product as described above (see Subheading 3.1.2, steps 3–4). Stage 2: Perform the second PCR with the purified product (1 mL) obtained from the first PCR as the template, using the 5¢~primer B and either 3¢~primer A (containing a stop codon; for HA-GLP-1R) or 3¢~primer B (no stop codon; for HA-GLP-1R-EGFP) in Fig.  6. Purify the final PCR product as described above (see Subheading  3.1.2, steps 3–4). In this step a DNA fragment is obtained, which contains RE sites, start codon, the full length N-terminal HA-tag, and the GLP-1R sequence either with or without a stop codon. 3. By digestion, prepare HA-GLP-1R (SpeI-XhoI) for insertion into the plasmid. In this step, two sticky ends are made in the purified final PCR product (obtained above in Subheading  3.2.2, step 2). Perform the digest at 37°C for 20 h (see Note 10) in a mixture of 50 mL as given below: PCR product

29.0 mL

10×NBuffer 2

  5.0 mL

100×BSA (10 mg/mL)a

  0.5 mL

SpeI (10,000 U/mL)

  1.2 mL

a

XhoI (20,000 U/mL)a

  0.6 mL

H2O

13.7 mL

 Stock concentrations

a

4. Run a 1% agarose gel (50  mL) at 140  V for 40  min (see Subheading  3.4.2) and extract the DNA fragment at ~1,400  bp from the gel as described previously (see Subheading 3.1.2, steps 3–4). 5. Prepare an NheI-XhoI digested pcDNA3.1(+) (for HA-GLP-1R) or NheI-XhoI digested pEGFP-N1 (for HA-GLP-1R-EGFP) in a digestion mixture of 50  mL as described previously (see Subheading 3.1.3). 6. Ligate the SpeI-XhoI digested PCR products from above into the NheI-XhoI digested vectors (pcDNA3.1(+) (NheI-XhoI) for HA-GLP-1R and pEGFP-N1(NheI-XhoI) for HA-GLP1R-EGFP) as previously described (see Subheading 3.1.4 and Note 16). 3.3. Sequence Confirmation

Following ligation and bacterial transformation, if colonies appear on the plate, the sequence of the inserts (cDNA) should be determined for a number of colonies. A three-step strategy should be

70

Huang and Willars

performed on each DNA clone (plasmid DNA isolated from a colony): step 1 (diagnostic digest; see Subheading  3.3.1) confirms whether a DNA fragment of the expected size has been inserted into the vector at the selected RE sites; step 2 (additional diagnostic digestion; see Subheading 3.3.2) confirms whether the expected RE sites are within the inserted sequence; and step 3 (DNA sequencing; see Subheading 3.3.3) will fully confirm the DNA sequence for the insert. 3.3.1. Preliminary Confirmation by Diagnostic Digestion

1. Pick approximately six individual colonies from the LB plate and inoculate each into a single sterile 15-mL tube containing 3 mL LB broth with appropriate antibiotic(s). Incubate all tubes at 37°C with shaking (230 rpm) overnight. 2. Transfer 1 mL from each culture to a clean 1.5-mL microfuge tube (ensure good labeling). Retain the remaining portion of the cultures and store at 4°C. 3. Spin down the cells in each microfuge tube at 4,500 × g, for 3 min at room temperature. Discharge the supernatant. 4. Resuspend cells in 100 mL buffer P1 (see item 11) containing RNase by pipetting thoroughly. Incubate the cell mixture at room temperature for 5 min. 5. Add 150 mL buffer P2 (see item 12) to each tube and mix by inverting the tubes three times. Incubate the mixture at room temperature for 5 min. 6. Add 150 mL buffer P3 (see item 13) to each tube and invert the tubes four times. Incubate the tubes on ice for 10 min. 7. Spin the tubes at 16,000 × g (maximum speed in a bench-top microfuge), room temperature for 10 min. 8. Carefully transfer 400 mL clear supernatant from each tube to a new microfuge tube. 9. Add 280 mL of isopropanol and mix thoroughly. 10. Spin at 16,000 × g for 30 min at room temperature and discharge the supernatant carefully without disturbing the DNA pellet. 11. Wash the DNA pellet with 500 mL of precooled 70% ethanol. Spin at 16,000 × g for 15 min at room temperature and discharge the supernatant. 12. Dry the DNA pellet in each tube in a 37°C incubator for 20 min. 13. Dissolve each pellet with 40 mL TE (item 8). 14. Label 1.5 mL microfuge tubes. 15. Prepare 120  mL of reaction mixture for six digestions ­containing 12 mL of 10×NEBuffer (type will depend on REs), 1.2 mL of 100× BSA, 15 U of each (of two) REs (see Note 17), and an appropriate volume of H2O.

Generation of Epitope-Tagged GPCRs

71

Fig. 7. Example result of preliminary confirmation of a construct. A DNA fragment of the GLP-1R(NheI-XhoI) was ligated into pEGFP-N1(NheI-XhoI) with the expectation of generating a C-terminal EGFP-tagged GLP-1R (GLP-1R-EGFP) (see plasmid map in Fig. 8). As described above (step 1, Subheading 3.3.1), six individual colonies were picked and cultured, from which plasmid DNA was prepared (step 2, Subheading 3.3.1). After a digest with NheI and XhoI at 37°C for 1.5 h (step 3, Subheading 3.3.1), DNA was separated by 1% agarose gel (lanes 1–6) along with a 1-kb marker (center lane). The digest was expected to generate two bands (4,711 and 1,428 bp) for a correct clone (as shown in lanes 1, 2, and 4). A potentially correct clone is also shown in Lane 6 but with a relatively low DNA concentration. Lane 3 indicates that the clone contained an empty vector (i.e., no insert). Lane 5 suggests a relatively high DNA concentration, in which the lowest band suggests RNA contamination (see Note 18). In this instance, the clonal cultures represented by lanes 1, 2, and 4 were retained.

16. Transfer 19 mL of mixture to each tube and add 1 mL plasmid DNA from each clone. 17. Digest at 37°C for 1.5 h (see Subheading 3.4.5 and Note 18). 18. Separate the products by gel electrophoresis (1% agarose gel, see Subheading 3.4.2) and determine the sizes of DNA bands from each clone under a UV trans-illuminator. Remember to run a lane containing an appropriate marker (see item 20). 19. Retain those clonal cultures that contain the cDNA insert of the correct size (Fig. 7). Store at 4°C. 3.3.2. Further Confirmation by Diagnostic Digestion

From the clones identified above (see Subheading 3.3.1) as containing an insert of the expected size, plasmid DNA should be isolated using a commercially available kit (see item 25) as below. 1. Inoculate 100 mL of each culture of the cDNA clones containing an insert of the expected size (i.e., from above; see Subheading 3.3.1) into a sterile 30-mL universal tube containing 5 mL of LB broth and appropriate antibiotic(s), and incubate overnight at 37°C with shaking (230 rpm). 2. Make a glycerol stock for each clone by adding 600  mL of  overnight culture to 300  mL of sterile 50% glycerol (see item 9). Store glycerol stocks at −80°C. 3. Pellet the remaining bacteria from each culture by centrifugation (4,000 × g, 15 min, 4°C). 4. Miniprep plasmid DNA from each clone using a column following the manufacturer’s instructions.

72

Huang and Willars

5. Measure the DNA concentration of each clone by absorbance (see Subheading 3.4.4). 6. Perform at least two diagnostic digests (20 mL reaction containing 0.5 mg DNA) with different REs. For each digestion, use at least one RE site that should be located within the insert. The following example (Fig. 8) is from a successful ligation of NheI-XhoI digested GLP-1R into a NheI-XhoI digested pEGFP-N1 plasmid to generate a GLP-1R sequence containing a C-terminal EGFP-tag (GLP-1R-EGFP). 7. Store the plasmid DNA of the correct clones at −20°C.

Fig. 8.  Plasmid map and specific restriction enzyme digests of a plasmid containing the GLP-1R-EGFP sequence. The construct was identified by separate RE digests with BamH I, Kpn I, Kas I-EcoR I, and NheI-Xho I. In each digest, 0.5 mg plasmid DNA was incubated with 3 U of each enzyme(s) at 37°C for 2 h in NEBuffers 3, 1, 2, and 2, respectively. The DNA in the reactions was separated by 1% agarose gel, and stained with 0.5% GelRed™ (see Note 21) for visualization (see Subheading 3.4.2). The pattern of bands observed after each digest was as expected for this construct.

Generation of Epitope-Tagged GPCRs 3.3.3. Confirmation by DNA Sequencing

73

In this step, the receptor sequence, including the start and stop codons and epitope tag are confirmed by full DNA sequencing from the both 5¢~ and 3¢~ ends. 1. Design primers for sequencing: (a) Length should be 19–21 bp. (b) Primers should be located approximately 50 nucleotides from the target sequence. (c) The G-C content should be 40–60% with a calculated Tm (see Note 11) of between 44 and 59°C. 2. Send the primers and the positive DNA clones from the diagnostic digests for sequencing following the guidance of the service provider. 3. The receptor sequence can then be checked using Blast [National Center for Biotechnology Information (NCBI)] at http://www.ncbi.nlm.nih.gov/BLAST/. 4. The sequence of a synthesized epitope tag should be confirmed by comparing the sequencing result with the original primers used for synthesis.

3.4. General Protocols for Generating Epitope-Tagged GPCRs 3.4.1. LB Agar Plates

The following protocol is suitable for making approximately 20 plates (see item 4). 1. Add 250 mL of H2O to a 500-mL bottle labeled with autoclave tape. 2. Weigh 5.0 g tryptone, 2.5 g yeast extract, 5.0 g NaCl, and 7.5 g agar. 3. Mix powder well to bring into solution. 4. Add H2O to total volume of 500 mL. 5. Loosen the top and autoclave at 121°C for 15 min. 6. Let the agar cool to ~55°C (you should be able to pick up the bottle without a glove). 7. Add either 500  mL of 1,000× ampicillin stock or 1,000× kanamycin stock to the liquid and mix well but slowly to avoid bubbles. 8. Wipe down the bench top with 70% IMS (see item 32). 9. Remove sterile Petri dishes from plastic bag (retain the bag for storage). 10. Pour a thin layer (10  mm) of LB agar (~20  mL) into each plate, being careful not to lift the cover off excessively (just sufficient to pour). 11. Swirl plate in a circular motion to ensure full coverage of the bottom surface. 12. Cool plates on the bench until agar is solid (~30  min) and store in appropriately labeled plastic bags at 4°C.

74

Huang and Willars

3.4.2. Agarose Gel Electrophoresis

1. Weigh out the required amount of agarose (depending on the gel percentage and volume; for example, 0.5 g for 50 mL of a 1% gel) into a flask that is approximately five times the ­volume of the solution (agarose boils over easily). 2. Add the appropriate volume of TAE buffer and swirl to mix. 3. Boil the mixture in a microwave oven (approximately 1.5 min on medium power) until the agarose melts completely; immediately swirl the flask several times (with care). 4. Cool the solution to 65–70°C and add ethidium bromide or GelRedTM (see item 21 and Note 21) to a final concentration of 0.5 mg/mL. Mix well and pour carefully into a clean casting plate with an appropriate comb. If bubbles appear, push them carefully to the sides with a tip. Allow the agarose to cool until solidified. 5. Immerse the gel in TAE buffer (see item 19) in the gel tank. Mix the DNA samples with 10× loading buffer (see item 10) and then pipette the samples into the wells of the gel. Remember to run an appropriate marker in one lane. 6. Run the gel at a voltage and for a time period that will achieve optimal separation, for example, 3 V/mL gel, 40 min. 7. Detect the DNA bands under UV illumination (see Note 21).

3.4.3. Extraction and Purification of DNA Fragments from Agarose Gel or Solution

Following the protocols below, DNA fragments (70 bp to 10 kb) can be extracted and purified using QIAquick Gel Extraction Kits (see item 26) to remove dNTPs, primers, nucleotides, enzymes and salts in the reaction solution or agarose and GelRed™ in the gel, and other impurities from DNA. If extracting DNA from an agarose gel start at step 1; alternatively, if extracting from solution start at step 5. 1. Record the weight of a 1.5-mL microfuge tube. 2. Excise the band of interest using a clean scalpel under the UV transluminator (low power to avoid nicking the DNA) and place into the preweighed microfuge tube. 3. Reweigh the microfuge tube and calculate the weight of added gel. 4. Add a volume of QG buffer (mL) equivalent to three times the gel weight (mg). Incubate at 50°C with periodic mixing for 10 min or until the gel fragment is dissolved. Proceed to step 6. 5. If purifying a DNA fragment from solution, add three ­volumes of QG buffer and mix. 6. Add isopropanol at a volume (mL) equivalent to that of the gel (mg) or of solution (mL). Mix and apply to a QIAquick column.

Generation of Epitope-Tagged GPCRs

75

7. Centrifuge for 1  min at 16,500 × g and discard the ­flow-through. 8. Add 0.5  mL of buffer QG to the column and microfuge (1 min, 16,500 × g). 9. Wash the column by addition of 0.75 mL of buffer PE and further centrifugation (1 min, 16,500 × g). Discard the flowthrough and centrifuge the QIAquick column (1  min, 16,500 × g) again. This step completely removes residual ethanol in buffer PE. 10. Transfer the column into a clean microfuge tube, then elute DNA (see Note 20) with 30 mL of buffer TE (see item 8) or H2O and collect by centrifugation (1 min, 16,500 × g). 3.4.4. Measuring DNA Concentration by Absorbance

1. Dilute 5 mL of plasmid DNA with 995 mL H2O in a 1.5-mL microfuge tube. 2. With a 1-mL quartz cuvette and using H2O as a blank, measure OD260 and OD280, or where possible use the blanks to zero the spectrophotometer. 3. Transfer the diluted DNA into the cuvette and measure OD260 and OD280. 4. Measure OD260 and OD280 again after a further 1:2 dilution with H2O if the concentration >1.0 mg/mL (see Note 19). 5. Using the relationship that an OD260 of 1.0 = 0.05  mg/mL pure DNA (a solution containing 50  mg/mL of double stranded DNA has an absorbance of 1.0 at a wavelength of 260  nm and path length of 1  cm), DNA concentration (mg/mL) = OD260 × 200 (dilution factor) × 0.05 (mg/mL). 6. The quality of DNA is determined by OD260/OD280, which should be 1.7–2.0. A value outside of this range suggests that the preparation should be performed again (see Note 19). 7. Wash the cuvette with H2O at least three times between ­different samples.

3.4.5. Restriction Endonuclease Digestion

For cloning work (e.g., to prepare inserts or vectors), the resulting DNA fragments may be separated by agarose gel electrophoresis. For this, 2 mg DNA should be used for digestion in a volume of 50 mL and subsequently extracted from the gel. On the contrary, a diagnostic digestion only needs 0.5 mg of DNA in a total volume of 20 mL and there is no need for DNA extraction after running the gel. 1. The reaction should follow the guide provided by the manufacturer of the enzyme used (see item 16). This will normally contain 5 mL (or 2 mL) of 10× RE buffer (generally supplied with the RE), 0.5 mL (or 0.2 mL) of 100× BSA (stock concentration 10 mg/mL), two REs (5 U/mg DNA for each, if they

76

Huang and Willars

are able to work in the same buffer), and H2O to a volume of 50 mL (or 20 mL). 2. Mix the contents gently by tapping the tube and a brief centrifugation (1,500 × g, 1 s) before placing in a thermocycler (or water bath). Incubate for 16–20 h (or 2 h, see Note 13). 3. Where the two enzymes cannot cut in one type of buffer, for example, SalI (100% cut in NEBuffer 3 but 0% in NEBuffer 2) and NheI (100% cut in NEBuffer 2 but 10% in NEBuffer 3), the digestion should be performed sequentially. In the first step, add one enzyme (5 U/mg DNA) into the reaction described above in steps 1–2. 4. Perform a brief purification as described in Subheading 3.4.3, steps 4–10. At the last step, elute DNA from the column with 30 mL of H2O (see Note 20). 5. Apply all the purified single-cut DNA to the digestion with the second enzyme, and repeat steps 1–2. 6. Keep reactions on ice or at −20°C until subjecting them to an agarose gel electrophoresis (see Subheading 3.4.3). 3.4.6. PCR

1. Handling primers: PCR primers are normally shipped in a lyophilized form in 2 mL microfuge tubes at room temperature. On arrival, they should be dissolved in H2O to 1 mg/mL in the original tube (original stock). Take 50  mL of primer from the original stock and add an additional 50 mL H2O for a working stock (0.5 mg/mL), which can be used directly in PCRs. Store both the original and working stocks at −20°C. 2. Set up a reaction mixture in a PCR tube (item 14) containing either 0.4  mg plasmid DNA template or 1.0  mg of cDNA library template (see Note 5), 0.5 mL (0.5 mg/mL) of both the 5¢~primer and 3¢~primer, 1 mL of dNTPs (10 nM stock), 5 mL of 10× DNA polymerase reaction buffer (generally supplied with the DNA polymerase), an appropriate volume of H2O to a total of 49  mL, and finally add 1  mL of DNA polymerase. 3. Mix the contents gently by tapping the tube and a brief microfuge (1,500 × g, 1 s). 4. Place the reaction tubes in a thermocycler with a heated lid (see Note 13) and run the following protocols: (1) 1 cycle of 95°C for 30 s; (2) 24–29 cycles of 95°C for 30 s, annealing temperature (Ta) (see Note 11) for 1 min and 72°C for 3 min; and (3) 1 cycle of 95°C for 30 s, Ta for 1 min and 72°C for 10 min, then hold at 4°C until PCR tubes are removed from the thermocycler (see Note 9).

Generation of Epitope-Tagged GPCRs

77

5. Store the reaction mixture at −20°C until agarose gel electrophoresis. 3.4.7. Preparation of Competent E. coli (DH5a)

 Day 1 1. Autoclave the following items (see Note 22): (a) 500 mL LB broth (b) Two 1-L Erlenmeyer flasks (c) 50 mL of freshly prepared 100 mM MgCl2 (see item 28) (d) 50 mL of freshly prepared 100 mM CaCl2 (see item 29) (e) 20 mL resuspension buffer (see item 30) (f) 200 microfuge tubes (1.5 mL) in a 1-L beaker 2. Inoculate a loop of DH5a cells (from frozen stock) into a 3-mL LB broth and grow cells overnight in a laboratory orbital shaker set at 37°C, ~225 rpm. Day 2 3. Inoculate 200 mL LB with 2 mL of overnight cells and grow the cells under the same condition described in Step 2 of Day 1 until OD600 of 2.5–3.5 (see Note 22). 4. Transfer the cells into six sterile 50-mL centrifuge tubes and incubate on ice for 10  min (all subsequent steps should be performed in a cold room). 5. Pellet the cells using a bench-top centrifuge (6,500 × g for 15 min at 4°C). Pour off and discard the supernatant. 6. Resuspend the cells in each tube with 5  mL of ice-cold 100 mM MgCl2 and combine the cells into two tubes (15 mL per tube). Addition of MgCl2 makes the cells very fragile and they should be handled with great care from this point. 7. Incubate the cells on ice for 20 min. 8. Centrifuge for 15 min, at 6,500 × g, 4°C. Pour off and discard supernatant. 9. Resuspend each cell pellet gently in 20  mL of ice-cold 100 mM CaCl2 and incubate on ice for 20 min. 10. Centrifuge for 15 min, at 6,500 × g, 4°C. Pour off and discard supernatant. 11. Resuspend both pellets gently in a total of 6 mL of resuspension buffer. 12. Prepare 50 mL aliquots in autoclaved 1.5 mL microfuge tubes, and store at −80°C. Once thawed, aliquots should not be refrozen for later use.

78

Huang and Willars

4. Notes 1. SOC is used primarily to aid recovery of competent bacterial cells (see Note 8) following transformation. It aids in the repair of perforations induced by transformation and provides a rich energy source for replication, thereby maximizing transformation efficiency. 2. PCR tubes are not only DNase and RNase free but also have a thin wall, which helps to heat and cool the PCR mixture quickly. 3. Restriction endonucleases (or restriction enzymes; RE) cleave DNA between two nucleotides within specific sites. For example, the common restriction endonuclease BamHI recognizes the sequence GGATCC and cuts between the two G bases on both the top and bottom strands (Fig.  9i). This results in the generation of a single-stranded “sticky end.” This sticky end is used to ligate (see Note 4) to a piece of DNA with a complementary sticky end; in this example, another DNA fragment cut with either BamHI (Fig. 9i) or BglII will also leave a sticky end compatible with a BamHI cut (Fig. 9ii). Although other REs can also generate compatible sticky ends (e.g., SpeI, NheI and XbaI; XhoI and SalI), such products cannot be cleaved by the REs used to initially generate the sticky ends. A very useful source of information can be found on the New England Biolabs website: (http://www. neb.com). 4. The ligation of two pieces of DNA into a single piece is catalyzed by T4 DNA ligase. This enzyme is available through many suppliers (e.g., New England Biolabs, Hitchin, UK)

Fig. 9. Restriction sites of BamH I (i) and Bgl II (ii). Sticky ends are shown in the boxes and the DNA pieces with compatible sticky ends are indicated with double arrows.

Generation of Epitope-Tagged GPCRs

79

and catalyzes the formation of a phosphodiester bond between juxtaposed 5¢-phosphate and 3¢-hydroxyl termini in doublestranded DNA with blunt- or compatible sticky ends. T4 DNA ligase requires ATP as a cofactor, which is normally in the buffer supplied. As this buffer is unstable, it can be degraded by multiple freeze/thaw cycles and it is best to prepare 10–20 mL aliquots from the original stock. 5. cDNA (complementary DNA) is synthesized from a mature mRNA template by reverse transcriptase. A cDNA library is a collection of cloned cDNA fragments inserted into either a plasmid vector or a Lambda phage vector, which together constitute some portion of the transcriptome of the organism. When a cDNA library is used as the template in a PCR, it is recommended to use a little more than when using a plasmid template. 6. For generating an N-terminal epitope-tagged GPCR, a start codon (including a Kozak sequence) should be located before the epitope tag sequence, while the initial start codon of the GPCR sequence should be removed but the initial stop codon should stay at the end of the receptor. On the contrary, for generating a C-terminal epitope-tagged receptor, the initial start codon (including a Kozak sequence) should be located at the beginning of the receptor and the stop codon should be moved from the end of the receptor to the end of the epitope tag (Fig. 10). 7. Appropriate restriction sites should (1) allow both the DNA fragment encoding the GPCR and the vector to be cut by the same or compatible REs (see Note 3); (2) the restriction sites used for isolating the insert from the existing construct must not be present within the GPCR coding sequence and; (3) after ligation to form the new construct, the sequences encoding the epitope tag and the GPCR should be in the same reading frame (i.e., a contiguous and nonoverlapping set of three-nucleotide codons).

Fig. 10. Example cDNA sequence of an epitope-tagged receptor (GLP-1R). WTGLP-1R represents the wild-type untagged GLP-1R (i); HA-GLP-1R represents the N-terminal HA-tagged GLP-1R (ii); and GLP-1R-HA represents the C-terminal HA-tagged GLP-1R (iii). Bases in italic represent the Kozak sequence. The start codon is identified by the single underline; the stop codon by a double underline; the HA-tag sequence by shaded boxes and; and the human GLP-1R cDNA sequence (1,392 bases in total) without a start or stop codon (4-1389) is indicated.

80

Huang and Willars

8. Competent cells (i.e., bacteria treated to allow incorporation of foreign DNA plasmids via transformation) can be purchased commercially or prepared in the laboratory (19) (see Subheading 3.4.7). Although many suppliers claim their competent cells have high transformation efficiency and do not need recovery in SOC medium during the transformation, such recovery may still increase efficiency, which is critical for ligations (see Notes 1, 4, 14). 9. PCR allows amplification of one or a small number of DNA sequences and allows the incorporation of the required restriction sites. The template, primers and dNTPs (single nucleotides of the bases A, T, G, and C) allow the DNA polymerase to initially generate the full-length sequence with restriction sites, which is then amplified in subsequent cycles. Each cycle allows denaturation, annealing and extension. Following the last cycle, samples are incubated at 72°C for an extra 8 min to fill in the protruding ends of newly synthesized PCR products. In some instances, denaturation in the first cycle is carried out for 5  min to increase the probability that long molecules of template DNA are fully denatured. In our experience, this is generally not necessary and for routine PCR we find that an initial denaturation for 90 s at 95°C is adequate. 10. Most REs recognize and cleave DNA poorly when their recognition sites are located at the ends of DNA fragments. It can, therefore, be difficult to achieve complete digestion of PCR products having such restriction sites or to perform double digests at restriction sites that are close to each other in an MCS region of a vector. The incorporation of additional 2–6 bases upstream of an engineered restriction site in a PCR primer will generally greatly increase the efficiency of digestion of the PCR product, although this can be dependent on the RE. Such digestions may also be improved by using relatively long incubation times (e.g., 20 h). Information on RE performance close to the end of DNA fragments, which will help in the design of primers, is available (e.g., New England Biolabs website: http://www.neb.com). 11. The melting temperature (Tm) of a primer (Eq. 1) is a critical determinant for the success of PCR. At the Tm, half of the primers will form duplexes with complementary primers at the same concentration. The temperature used for the annealing step (Ta) in the thermocycler program should be 5°C lower than the calculated Tm of the primers.

Tm (°C) = 61.8°C + [41°C ´ å (G + C)-675°C]/ å (A + T + G + C).  (1)

S(A + T + G + C) is the total number of bases in the primer and S(G + C) is the number of G, C bases in the primer sequence.

Generation of Epitope-Tagged GPCRs

81

12. The presence of G or C bases within the last five bases from the 3¢ end of primers (GC clamp) promotes binding at the 3¢ end as the bonding between G and C bases is stronger than that of A and T bases. 13. When using thermocyclers for molecular biology reactions (e.g., PCR), heated lids can help prevent evaporation of the reaction mixture. If a heated lid is not available, mineral oil can be used to overlay the PCR mixture. Thermocyclers with heated lids are preferential for RE digests with long incubation times. 14. The molar ratio of the insert and vector for ligation can have a significant effect on the outcome of a ligation and the subsequent transformation. However, the optimum ratio can range between a 1:1 and 10:1 molar ratio of insert:vector. The addition of too much DNA (>50 ng total) will decrease the recovery of transformed colonies. It may be necessary to try several ratios and amounts of total DNA in parallel to optimize the reaction. When preparing an insert (500– 2,000  bp) from a plasmid and preparing a vector (3,000– 6,000 bp) for ligation of this insert, we routinely use 2 mg of each for the preparative RE digests. In the final step of purification, the fragments are eluted from the columns with 30 mL of solvent (see Subheading 3.4.3) and 0.5 mL of vector and 4 mL of insert are then used in the ligation reaction. 15. This design allows appropriate RE sites to be inserted (see Note 7) and has several advantages. Firstly, both the GLP-1R (BglII-XhoI) and HA-tag (XhoI-XbaI) fragments are inserted into pcDNA3.1(+) prepared using a BamHI-XbaI digest (see  Note 3). The MCS of this vector includes NheI, PmeI, AflII, HindIII, Asp718I, KpnI, BamHI, BstXI, EcoRI, EcoRV, BstXI, NotI, XhoI, XbaI, ApaI, and PmeI (5¢ to 3¢), and therefore more RE sites are available at the N terminus when the full length GLP-1R is subcloned (or cut) from the resulting construct for ligation into other vectors containing a C-terminal epitope tag (e.g., pEGFP-N1 (NheI-XhoI)). Secondly, the resulting construct offers a vector containing a C-terminal HA-tag, which can be used for tagging other receptors by simply replacing the receptor sequence. Thirdly, the addition of an extra base (G) immediately before the XbaI site at the C terminus of the HA-tag allows the TAG triplet in the XbaI site to function as the stop codon for the GLP1R-HA. Fourthly, an additional SalI site inserted at the C terminus of the HA-tag (immediately before the additional G and XbaI site) allows a GLP-1R-HA fragment either with (when cut with XbaI) or without (when cut with SalI) a stop codon. Finally, a BamHI site is located within the GLP-1R sequence allowing a truncated GLP-1R-HA fragment to be

82

Huang and Willars

isolated. This then can be used to replace the N-terminal region of the receptor with an N-terminal epitope-tagged version to produce a GLP-1R sequence with both an N-terminal epitope tag and C-terminal HA-tag. 16. The strategy here is to generate a version of pcDNA3.1(+) containing an N-terminal HA-tag sequence that can be used for the generation of multiple constructs. Since SpeI and NheI digests result in compatible sticky ends (see Notes 3 and 15), fragments of DNA cut either by SpeI or NheI can be ligated together, but the new site cannot be cut by either SpeI or NheI (see Note 3). In the initial PCR, the 5¢~primer contains a 5¢ truncated HA sequence incorporating an NheI site between the 3¢ end of the HA sequence and the 5¢ end of the GLP-1R. The 3¢~primer contains the GLP-1R with a 3¢ XhoI site. This 3¢~primer will be either with or without a stop codon depending on whether the subsequent construct is to lack or contain a C-terminal tag, respectively. This initial PCR will generate a 5¢-truncated HA sequence at the 5¢ end of the full-length GLP-1R sequence. The second round of PCR will generate a full-length HA-GLP-1R sequence with a 5¢ SpeI site and a 3¢ XhoI site. Ligation of this fragment into an NheIXhoI digested pcDNA3.1(+) will allow the full length GLP-1R sequence to be isolated by NheI-XhoI digestion, but importantly the HA-tag sequence will stay in the vector permanently, allowing other constructs to be inserted for N-terminal tagging. 17. At this preliminary stage, it is normal to use REs that work in the same buffer; ideally the same pair of REs that were used for preparing the vector and insert. 18. In this instance, RNase added to the reaction may not remove all of the RNA and thus an RNA band may be visible in the agarose gel following electrophoresis after the RE digestion. In addition, the RNase and DNase may not be removed during preparation of the DNA and a long RE digest may result in DNA damage. A short digestion (£2  h) is therefore recommended. 19. A spectrophotometer is most accurate when measurements are within the instrument’s linear range (generally 0.1–1.0 OD units). An OD260/OD280 of 1.7–2.0 indicates that the DNA is fairly free of contaminants. Values below this range indicate the presence of protein and membrane fractions. 20. For long-term storage of purified DNA, it is better to elute the DNA fragment with TE (see item 8) rather than H2O. Although the EDTA in TE inhibits enzyme activity, this will have limited effect on a typical RE digestion, particularly as the EDTA will be highly diluted. However, it is recommended

Generation of Epitope-Tagged GPCRs

83

to elute with H2O when all of the eluted material will be used immediately for RE digestion (e.g., PCR product purification and purification of DNA between two different RE digests). 21. Nucleic acid staining (with ethidium bromide or GelRed™) allows visualization of DNA (>50 ng) in the agarose gel under UV light. This exposure to UV light may nick and damage the DNA, so it should be performed as quickly as possible using the minimum intensity of UV available that allows visualization. 22. An OD600 range of 0.3–0.4 has been widely accepted for making DH5a competent cells as it is essential that the number of viable cells does not exceed 108 cells/mL for efficient transformation. For most strains of E. coli, this is equivalent to an OD600 of ~0.4 (19). However, we have found that when cultures are harvested at an OD600 of 0.25–0.30, there is a dramatic increase in transformation efficiency for competent DH5a compared with those prepared at an OD600 of 0.35–0.40. We would, therefore, recommend harvesting at the lower OD600 range of 0.25–0.35 in the current protocol. It is also advisable to prepare a spare flask and more media in case the OD600 is too high. Under these circumstances a small aliquot of the bacterial growth can be used to seed a new flask. References 1. Light, A., Savithri, H.S. and Liepnieks, J.J. (1980) Specificity of bovine enterokinase toward protein substrates. Anal. Biochem. 106, 199–206. 2. Slootstra, J.W., Kuperus, D., Pluckthun, A. and Meloen, R.H. (1997) Identification of new tag sequences with differential and selective recognition properties for the anti-FLAG monoclonal antibodies M1, M2 and M5. Mol. Divers. 2, 156–164. 3. Xu, T. and Rubin, G.M. (1993) Analysis of genetic mosaics in developing and adult Drosophila tissues. Development 117, 1223–1237. 4. Chaga, G.S. (2001) Twenty-five years of immobilized metal ion affinity chromatography: past, present and future. J. Biochem. Biophys. Methods 49, 313–334. 5. Wilson, L.A., Niman, H.L., Houghten, R.A., Cherenson, A.R., Connolly, M.L. and Lerner, R.A. (1984) The structure of an antigenic determinant in a protein. Cell 37, 767–778. 6. Field, J., Nikawa, J., Broek, D., MacDonald, B., Rodgers, L., Wilson, I.A., Lerner, R.A. and

Wigler, M. (1988) Purification of RASresponsive adenylyl cyclise complex from Saccharomyces ceverisiae by use of an epitope addition method. Mol. Cell. Biol. 8, 2159–2165. 7. Evan, G.I., Lewis, G.K., Ramsay, G. and Bishop, J.M. (1985) Isolation of monoclonal antibodies specific for human c-myc protooncogene product. Mol. Cell. Biol. 5, 3610–3616. 8. Shimomura, O., Johnson, F.H. and Saiga, Y. (1962) Extraction, purification and properties of aequorin, a bioluminescent protein from the luminous hydromedusan, Aequorea. J.  Cell. Comp. Physiol. 59, 223–239. 9. Heim, R., Cubitt, A. and Tsien, R. (1995) Improved green fluorescence. Nature 373, 663–664. 10. Chalfie, M., Tu, Y., Euskirchen, G., Ward, W.W. and Prasher, D.C. (1994) Green fluorescent protein as a marker for gene expression. Science 263, 802–805. 11. Cormack, B.P., Bertram, G., Egerton, M., Gow, N.A., Falkow, S. and Brown, A.J. (1997)

84

Huang and Willars

­ east-enhanced green fluorescent protein Y (yEGFP) a reporter of gene expression in Candida albicans. Microbiology 143, 303–311. 12. Jarvik, J.W. and Telmer, C.A. (1998) Epitopetagging. Annu. Rev. Genet. 32, 601–618. 13. Ivic, L., Zhang, C., Zhang, X., Yoon, S.O. and Firestein, S. (2002) Intracellular trafficking of a tagged and functional mammalian olfactory receptor. J. Neurobiol. 50, 56–68. 14. Perret, B.G., Wagner, R., Lecat, S., Brillet, K., Rabut, G., Bucher, B. and Pattus, F. (2003) Expression of EGFP-amino-tagged human mu opioid receptor in Drosophila Schneider 2 cells: a potential expression system for largescale production of G-protein coupled receptors. Protein Expression Purif. 31, 123–132. 15. Palmer, E. and Freeman, T. (2004) Investigation into the use of C- and N-terminal GFP fusion proteins for subcellular localization studies using reverse transfection microarrays. Comp. Funct. Genom. 5, 342–353.

16. Brothers, S.P., Janovick, J., Conn, P.M. (2003) Unexpected effects of epitope and chimeric tags on gonadotropin-releasing ­hormone receptors: implications for understanding the molecular etiology of hypogonadotropic hypogonadism. J. Clin. Endocrinol. Metab. 88, 6107–6112. 17. Kozak, M. (1987) An analysis of 59­noncoding  sequences from 699 vertebrate messenger RNAs. Nucleic Acids Res. 15, 8125–8148. 18. Huang, Y., Wilkinson, G.F. and Willars, G.B. (2010) Role of the signal peptide in the synthesis and processing of the glucagon-like peptide-1 receptor. Br. J. Pharmacol. 159, 237–251. 19. Sambrook, J. and Russell, D.W. (2001) Molecular Cloning: A Laboratory Manual, 3rd ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp 1.112–1.122.

Chapter 5 The Use of Site-Directed Mutagenesis to Study GPCRs Alex C. Conner, James Barwell, David R. Poyner, and Mark Wheatley Abstract G protein coupled receptors (GPCRs) are highly flexible and dynamic proteins, which are able to interact with diverse ligands, effectors, and regulatory proteins. Site-directed mutagenesis (SDM) is a powerful tool for providing insight into how these proteins actually work, both in its own right and when used in conjunction with information provided by other techniques such as crystallography or molecular modelling. Mutagenesis has been used to identify and characterise a myriad of functionally important residues, motifs and domains within the GPCR architecture, and to identify aspects of similarity and differences between the major families of GPCRs. This chapter presents the necessary information for undertaking informative SDM of these proteins. Whilst this is relevant to protein structure/function studies in ­general, specific pitfalls and protocols suited to investigating GPCRs in particular will be highlighted. Key words: Alanine scan, Substituted cysteine accessibility method, Chimeric mutagenesis, Random mutagenesis, Primer design, Plasmid transfection, Plasmid amplification

1. Introduction 1.1. Why Mutagenesis?

G protein coupled receptors (GPCRs) are widely recognised to be archetypal allosteric proteins. They can interact with numerous ligands which in turn promote a range of conformations which can selectively interact with a host of effectors (1). Understanding the fundamental processes underlying ligandbinding and receptor activation is at the forefront of research into the super-family. The crystal structures of rhodopsin and more recently, b1 adrenergic, b2 adrenergic, and adenosine A2a receptors are extremely useful for understanding specific receptor structures and as a basis for homology modelling (2, 3). However, the crystal structure data are a static snap-shot of dynamic proteins which cannot fully define the multiple conformations, and the dynamic

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_5, © Springer Science+Business Media, LLC 2011

85

86

Conner et al.

inter-conversion of these states required for GPCR function. Neither can any single structural technique provide the full range of conformational changes that occur during receptor activation and the precise sites of interaction between a given GPCR, its ligands of differing efficacy and the intracellular machinery. GPCR effectors include not only the different classes of G proteins but also other interacting proteins such as G protein receptor kinases, beta-arrestins, and anchoring proteins (4–6). Finally, the current structural data are restricted to family A GPCRs. Due to lack of sequence homology, this makes it difficult to use these data as a modelling template for the other sub-families (7). Altering GPCRs using site-directed mutagenesis (SDM) has been used to understand many of their functions. This includes the identification of key motifs required for activation such as the DRY site at the bottom of transmembrane helix 3 (TM3) (8) and the NPxxY of TM7. Mutagenic studies have been used to identify many other fundamentals of GPCR function including specific ligand contacts, for example the extracellular loop 2 (ECL2) of the V1a vasopressin receptor (V1aR) (9). This also includes specific intramolecular structural requirements, for example the conserved proline kink in TM6 of the calcitonin gene-related peptide receptor (10) and sites thought to be directly associated with G protein coupling (11). Mutagenic analysis can be split into random or targeted (site-directed). Random mutagenesis was largely superceded as sitedirected technology advanced but has recently been re-visited. Targeted mutagenesis is very widely used for the study of GPCR function and can be broadly split into mutants altering one or two amino acids for a very specific and localised analysis and deletions/chimeras of large regions as a “mining” process to isolate regions of functional importance. The term “site-directed mutagenesis” generally refers to single residue mutation and will be used in that context here. 1.2. Choice of Mutagenesis Strategy

Alanine scans have traditionally been implemented in the study of membrane proteins particularly in alpha helical bundles as it is assumed that the small, non-reactive methyl side chain will not distort the native secondary structure, yet shed light on the function of the native residue. Consequently, alanine scanning is a powerful tool to probe ligand–protein and protein–protein interfaces. For example, it was shown through alanine SDM that two serine residues located in TM5 of the hamster b-adrenergic receptor are essential for catecholamine agonist binding (12). Typically, following mutagenesis, the effects on ligand binding, agonist potency, and receptor expression are measured. A useful framework has been developed by Hulme et al., which interprets the effects of individual mutants on receptor stability, ligand binding, and receptor activation (see Subheading 3.4) (13).

The Use of Site-Directed Mutagenesis to Study GPCRs

87

In spite of the popularity of alanine scans, interpreting their results can be problematic. Typically, only a small fraction of interface residues account for the majority of the binding energy between ligand–protein and protein–protein interfaces (14). Therefore, the relative effect of an alanine substitution is dependent on its impact on the overall total binding energy of the interface. If the alanine mutant substantially changes the total binding energy, it is referred to as a “hot spot.” However, the alanine substitution may be in the binding interface but not make a significant difference to the total binding energy. This could give rise to false-negative results if not interpreted correctly. Holst et al. (15) suggested that steric hindrance mutagenesis could be used to minimise this problem. The addition of larger side chains in the presumed binding region could directly impair ligand interaction and disrupt neighbouring amino acid side chain conformations. Other mutant scanning approaches include lysine scanning. This has been used to investigate the nicotinic acetylcholine receptor to distinguish between core hydrophobic and surface hydrophilic orientations of native side chain residues (16). The substituted cysteine accessibility method is a technique where native residues get consecutively substituted to cysteines and analysing the rate at which sulphydryl specific reagents, such as biocytin maleimide or derivatives of methanethiosulphonate interact, can be used to determine whether the cysteine side chain is exposed to an aqueous environment. Consequently, inferences on the physio-chemical environment of the native residue can be made, which can lead to the identification of residues involved in binding cavities and water channels. For example, Shi and Javitch (17) made ten consecutive cysteine mutants within ECL2 of the dopamine 2 receptor, which lead to the conclusion that the loop was likely to be folded into the ligand-binding crevice of the receptor. Cysteine mutations can also be used to discover distance restraints within and between proteins. Engineering disulphide bonds within a protein, before and after activation can give detailed information on the conformational changes that are essential for signal transduction in receptors. For example, disulphide bonds were successfully engineered in the parathyroid hormone 1 receptor (PTHR1) between the top of TM2 and TM7 suggesting that these regions are in close spatial proximity. The engineered disulphide bond made between F238C/F447C became disrupted after the receptor was activated by the parathyroid hormone, giving insight into the initial stages of the transduction mechanism of this receptor (18). Cysteine mutants have also been extensively used to incorporate biophysical probes. Site-directed spin labelling is an approach that takes advantage of the capability of electron paramagnetic resonance spectroscopy, which can detect structural changes

88

Conner et al.

within a millisecond time frame. Hubbell et al. introduced pairs of sulphydryl-reactive spin labels on the cytoplasmic face of the GPCR rhodopsin. The authors concluded that there was a significant reduction in side chain packing at the cytoplasmic surface during activation (19). This “flowering” effect upon activation was previously hypothesised when a spin label was introduced at intracellular loop 2 (ICL2) and at different locations in TM6 (20). This made it possible to measure the relative changes in distances between TM3 and TM6. Engineering metal ion binding sites within proteins is another strategy to probe the tertiary structure of a protein along with its conformational changes required for function. Sheikh et al. (21) constructed a Zn2+ ion bridge between TM3 and TM6 in retinal rhodopsin and in the b-adrenergic receptor by generating bishistidine metal ion binding sites. It was concluded that the metal ion bridge constrained TM6 locomotion and as a result prevented the activation of the intracellular G protein (i.e. in these cases transducin and Gs, respectively). Moreover, using an evolutionary trace method, Sheikh et al. (22) were able to engineer a Zn2+ ion bridge in the cognate position in the secretin-like PTHR1. The metal ion bridge in the PTHR1 also impaired the ability of the receptor to induce an intracellular cascade. In spite of the obvious lack of sequence homology between the two families, the authors postulated that a shared activation mechanism involving TM6 locomotion is present in both rhodopsin-like and secretin-like GPCRs. Another approach which is possible using mutagenesis is to probe species-selectivity or subtype-selectivity of drugs by mutating residues in one receptor to corresponding residues in the other receptor. For example, this has been used to identify a discriminating residue in the oxytocin receptor contributing to species-specific pharmacology of a non-peptide oxytocin receptor antag­onist developed to prevent pre-term labour (23). It is also possible to use systematic substitution by all other encoded 19 amino acids to probe structural requirements at a specific locus of functional importance (24). Random mutagenesis is a generally untargeted method for determining residues of structural or functional relevance. This technique results in a significant body of data requiring a subsequent selection process. Historically, random mutagenesis involved whole organisms treated with radiation or chemical mutagens to alter phenotypes. The technique has recently undergone an upsurge in experimental popularity, moving towards in  vitro methods more in line with the directed analysis used elsewhere. This involves non-perfect PCR or “gene-shuffling” through recombination strategies for localised random mutagenesis of ­specific gene targets or their regulatory elements; for a review see ref. 25. A recent development in random mutagenesis for the

The Use of Site-Directed Mutagenesis to Study GPCRs

89

analysis of GPCRs involved a comprehensive study of the M3 muscarinic acetylcholine receptor (M3R). This involved a combination of low-frequency random mutagenesis of the entire M3R coding sequence, followed by the application of a yeast genetic screen allowing the recovery of inactivating M3R single point mutations (26). This gives a comprehensive understanding of multiple residues in the ECL2, which is critical for receptor activation. It remains to be seen whether this is more efficient than targeted mutagenesis for the systematic analysis of GPCR structure function. Although only a select number of experimental strategies have been briefly highlighted here, it is evident that there is currently a plethora of experimental techniques that can be successfully employed to investigate membrane proteins. Furthermore, the rapid advance in computer technology has allowed the realm of bioinformatics to have an immediate but sustaining impact on the field of pharmacology. Coupling mutagenesis data with crystallographic information and powerful predictive algorithms can be used to develop molecular models of GPCRs. Such models can provide mechanistic insight, predict receptor phenomena, and direct future mutagenesis in an iterative process (27). Consequently, this has become the dominant strategy for explaining and guiding research in molecular pharmacology. The methods in this chapter deal specifically with targeted (site-directed) mutagenesis.

2. Materials 2.1. Bacterial Cell Handling and Plasmid Recovery

1. Competent cells; e.g. XL10-GOLD cells (Stratagene, UK). 2. Microbiological incubator (many available). 3. LB (Oxoid, UK). 4. LB-agar (Oxoid, UK). 5. Ampicillin (Na-salt; Sigma-Aldrich, UK). 6. Plasmid isolation kit (Wiz-prep; Promega, UK).

2.2. Site-Directed and Chimeric Mutagenesis

1. Pfu DNA polymerase (Promega) and 10× buffer provided with the Pfu polymerase. 2. Oligonucleotide primers (Invitrogen). 3. 10 mM dNTP mixture (Sigma-Aldrich, UK): a mixture of all four dNTPs (dATP, dCTP, dGTP, and dTTP). 4. DNA template plasmid at 100 ng/mL containing the gene of interest. 5. Standard PCR thermo cycler (many available).

90

Conner et al.

6. 0.5 mL sterile PCR tubes (Sarstedt, Germany). 7. 0.8–1.2% (w/v) agarose gel; preferably pure, low melting temperature agarose, e.g. Seaplaque (Cambrex, USA). 8. Agarose gel tank (Bio-Rad Laboratories, USA). 9. 240 V power pack (Bio-Rad Laboratories, USA). 10. Desktop microcentrifuge (many available). 11. DpnI restriction enzyme (New England Biolabs, USA). 12. Competent cells; e.g. XL10-GOLD cells (Stratagene, UK). 13. Microbiological incubator (many available). 14. LB (Oxoid, UK). 15. L-Agar (Oxoid, UK). 16. Ampicillin (Na-salt; Sigma-Aldrich, UK). 17. Plasmid isolation kit (Wiz-prep; Promega, UK). 18. DNA purification (resin-retention kit; Promega, UK, Qiagen, UK and others). 19. T4 DNA ligase (Promega, UK). 20. pCR-Script blunt-cloning vector (Stratagene, UK). 2.3. Cell Culture for Analysis and Transient Transfection

1. Dulbecco’s modified eagle’s medium (DMEM) (Gibco/ BRL, Bethesda, MD) supplemented with 10% foetal bovine serum (FBS; HyClone, Ogden, UT). 2. Solution of 0.25% trypsin (v/v) and 1 mM EDTA (Gibco/ BRL, USA). 3. Teflon cell scrapers (Fisher, UK). 4. “Transfast” transfection reagent (Promega, UK). 5. Serum-free DMEM (Gibco/BRL, Bethesda, MD).

3. Methods 3.1. Site-Directed Mutagenesis Based on Quikchange (Stratagene, UK)

Design of the appropriate oligonucleotide primer pairs is one of the most important design aspects to consider for PCR in general but especially so for SDM. The five golden rules of primer design are as follows:

3.1.1. Oligonucleotide Design

1. To incorporate between 9 and 15 nucleotides either side of the mutation site. 2. To utilise the inherent codon degeneracy in mammalian mRNA translation to reduce the number of changed bases; current thinking suggests that codon usage is not a major issue for expression levels and this has never been noticed as a negative factor in the author’s considerable experience using this technique.

The Use of Site-Directed Mutagenesis to Study GPCRs

91

3. Check for obvious secondary structure and potential mispriming with a BLAST search. Practically, when many primers are being designed it is common to perform these checks only if a problem with individual mutagenesis using certain primers is encountered. 4. Try to avoid multiple runs of individual bases; this has been known to cause mispriming but is often an unavoidable issue. 5. Try to terminate the primer at the 3¢ end with a so-called GC clamp (see Note 1). Terminate the primer at the 3¢ end with 2 Gs, Cs, or a combination of the two. 3.1.2. Mutant Plasmid Amplification

This is the major step in this process and involves the amplification of the template using mutated oligonucleotide primers. This process (and subsequent steps) is outlined in (Fig. 1). 1. The amplification mix contains the following: 100 ng plasmid (see Note 2), 10 pmol each primer, 5–10 U Pfu polymerase

1. Clone in plasmid with target site for mutation

2. Denature plasmid and anneal oligonucleotide primer containing the mutation

3. Using the non-strand displacing action of Pfu DNA polymerase, to extend and incorporate the mutagenic primers

4. Digest the methylated, non-mutated parental DNA template with Dpn 1

5. Transform the circular, nicked ds DNA in appropriate highly competent cells. These repair the nicks in the plasmid

Fig. 1. The Stratagene QuikChange mutagenesis method (based on http://www.stratagene. com/manuals/200518.pdf).

92

Conner et al.

with the associated buffer and excess dNTPs (typically 1 mL at 2.5  mM final concentration). A final volume of 50  mL is advised. Always include a control with one or both oligonucleotides omitted. 2. The melting temperature has to be optimised although a ­general estimate of 60°C is suitable for most primers. Following an initial denaturation step of 95°C for 1  min, cycle as follows, for 12 cycles: 95°C

1 min

55–60°C

1 min

75°C

2 min per kilobase (14 min for a typical GPCR in a pcDNA3 expression plasmid).

To digest the template fragment from the amplification mix, add 2–10  U DpnI (see Note 3), pipette up and down and spin briefly in a microcentrifuge for 1 min. Incubate at 37°C for 1 h (see Note 4). 3. Transform any (see Note 5) commercially available competent cells which are not genetically DAM methylase deficient, using typically 1–5  mL amplification mix; include a DpnI treated-plasmid transformation as a control for DpnI digestion. Transformation is described in more detail in Chapter 4 (see Subheading 3.1.1). 4. Following transformation, it is advisable to incubate in 0.5–1 mL LB (with no selective agent) at 37°C for 1–1.5 h to promote initial antibiotic resistance in free media. This is followed by overnight incubation at 37°C on L-agar plates with the required antibiotic selection agent (50  mg/mL ampicillin in the case of pCR-Script). 5. To identify successful transformants, plasmids obtained from overnight cultures should be run on an agarose gel and sequenced using appropriate sequencing primers as for the chimaera production in Subheading  3.2 (for further details on agarose gel electrophoresis and sequencing, see Chapter 4, Subheadings 3.4.2 and 3.3.3, respectively). If restriction sites have been disrupted by the mutation change, a digestion analysis of transformants can be a useful initial screen to avoid unnecessary sequencing costs. 3.2. Chimeric Mutagenesis

1. Identify the fragments required from each receptor and design PCR primers for their amplification, ensuring that the upstream sense primer and the downstream antisense primer have suitable (different) restriction sites for the final cloning step into the multiple cloning site of the desired expression vector. Oligonucleotide primer design considerations should

The Use of Site-Directed Mutagenesis to Study GPCRs

93

follow that described for routine PCR. Each fragment should be amplified separately using a blunt polymerase (for example Pfu; others are available). 2. Prepare the amplification mix. This should contain the following: 5–10  ng plasmid (as for general PCR); mini-prep DNA is adequately pure and concentrated for the reaction to be successful. 1–10 pmol of each primer (in excess). 1–10 U Pfu polymerase (according to manufacturer’s instructions) with the associated buffer. Excess dNTPs (typically 1  mL at 2.5  mM starting concen­ tration). A final volume of 50 mL is advised, overlayed with an equivalent volume of mineral oil to prevent evaporation/condensation. The melting temperature has to be optimised; however, a general estimate of 60°C is suitable for most primers. Following an initial denaturation step of 95°C for 1  min, cycle as follows, for 30 cycles: 95°C

1 min

55–60°C

1 min

75°C

1 min

Typically two reactions for each desired fragment is sensible. 3. Run all on a 1.2% agarose gel (see Note 6). Remove the gel fragment using a clean, sterile scalpel, retaining as little gel as possible. 4. Purify the sample using a commercially available resin-retention kit. This routinely results in clean, eluted samples of between 30–50 mL and 20–200 ng/mL; however, it is necessary to run an aliquot (1  mL is usually sufficient depending on PCR amplification) on a 1.2% agarose gel to confirm band presence as occasionally gel-clean kits can fail, leading to a loss of DNA. 5. Ligate the two clean bands. Mix approximately 100  ng of each fragment with 1–10  U T4 DNA Ligase (according to manufacturer’s instructions) and add the appropriate ligase buffer as directed. Use a typical final volume of 10–20  mL. Reports differ as to the optimum length and temperature of ligation. Experience suggests 16 h at 14°C is particularly efficient, although 1 h at 37°C is known to be sufficient. 6. Carry out the second PCR step to create the chimeric fragment. Using approximately 1–2 mL of the ligation product as

94

Conner et al.

a template, repeat the PCR as in Subheading 3.2, step 2 using the N-terminal sense primer and the C-terminal antisense primer of the desired chimaera. Include individual primer controls. Run, excise, and clean the PCR product as described above. The product can then be ligated into a blunt-ended cloning vector such as the pCR-Script blunt-cloning vector with blue-white selection. 7. Perform the blunt cloning of the fragment into a plasmid ­vector. Pre-digest the vector with a restriction site producing blunt ends (any single EcoRVI-containing plasmid can be used, but ideally use one with an Srf I site located in the middle of a b-lactamase gene for easier selection; the example here uses pCR-script) and then ligate into the plasmid. Digestion and ligation are according to manufacturer’s instructions but briefly, approximately 100  ng plasmid is digested in appropriate buffer with 5–10  U appropriate restriction enzymes for 1 h at 37°C followed by an enzyme denaturation step for 20 min at 75°C. Ligation can be titrated for efficiency with a starting point of 1:10 plasmid:fragment ratio, at concentrations of approximately 10:100 ng, respectively. To ligate the two clean bands, mix approximately 100 ng each fragment, 1–10 U T4 DNA ligase (according to manufacturer’s instructions) and add the appropriate ligase buffer as directed. Use a typical final volume of 10–20 mL. Reports differ as to the optimum length and temperature of ligation. Experience suggests 16 h at 14°C is particularly efficient, although 1 h at 37°C is known to be sufficient. 8. Transform any (see Note 7) normal E. coli competent cells (unlike the site-directed point mutagenesis described above, few special considerations are necessary). This is followed by overnight incubation at 37°C on L-agar plates with the required antibiotic selection agent (ampicillin in the case of pCR-Script). To identify successfully ligated transformants, spread 100  mL 100  mM IPTG (for b-lactamase promoter) and 100 mL 2% X-Gal (B-lactamase substrate) onto the agar, allowing the components to dry individually to prevent aggregation. White colonies are selected (include one blue colony as an unligated plasmid control; see Note 8). 9. Plasmids obtained from overnight cultures are digested with the restriction enzymes incorporated into the appropriate primers. The digestion will produce the correct chimeric ­construct size with both enzymes but, importantly, not with either restriction enzyme alone. Positive chimeric plasmids identified by restriction enzyme analysis are sequenced using appropriate sequencing primers (see Note 9). Subsequent sub-cloning into an expression vector is routine.

The Use of Site-Directed Mutagenesis to Study GPCRs

3.3. Transfection and Analysis of Mutant Plasmids into Mammalian Cells

95

1. Clearly, the original template choice now needs the appropriate expression machinery as the subsequent analysis will reflect this choice. For example, mutants to be compared in live cell pharmacological assays will require the mammalian expression vectors, whereas those used for structural studies may require bacterial or eukaryotic (often inducible) expression machinery. 2. Generally, transfection involves a confluent flask of cells, and many commercial protocols are available. Expression occurs in typically 24–72 h and transfected cells can be analysed as appropriate.

3.4. Interpretation of Alanine Scans (see Note 10)

1. Measure the expression of the mutated receptor, the Kd for the binding of a given agonist and its efficacy for any given function using the same agonist as well as the basal activity. Compare these parameters against those measured for the wild-type (WT) receptor. 2. If there is no change in any parameter, the side chain of the residue has no specific role. 3. A change in receptor expression in the absence of a change in any other parameter implies that the side chain is involved in receptor stability. 4. A change in agonist Kd in the absence of a change in any other parameter implies that the side chain is involved in ligand binding. 5. A decrease in expression with an increase in agonist affinity, efficacy, and basal activity implies the side chain is important in constraining the receptor in an inactive form. 6. A decrease in basal activity, agonist efficacy, and affinity suggests that the side chain plays a general role in promoting receptor activation. If the efficacy and agonist binding are chiefly impaired, the side chain’s main role is in binding the agonist in the active form of the receptor. On the contrary, if the predominant effect is on basal activity, the side chain is involved in G protein recognition.

4. Notes 1. The A-T interaction has two hydrogen bonds compared with the G-C interaction which has three. The movement of polymerase from the 3¢ end of the annealed primer requires a firm double stranded grip. No such interaction is required at the 5¢ end. As such, ending with a GC clamp can enhance the amplification step significantly.

96

Conner et al.

2. This is not PCR therefore a significant starting mass is required; commercial mini-prep DNA is adequately pure and concentrated for the reaction to be successful. 3. DpnI is a restriction enzyme which requires a methylated DNA substrate. Plasmids will be methylated if recovered from a bacterial host which is DAM methylase-positive. As in vitro amplification has no methylation activity, only the template will be digested. 4. To check that the amplification has been successful, it is useful to remove 5 mL and compare with the pre-DpnI control on a 1.2% agarose gel. This is not essential and occasionally the product is at too low concentrations for electrophoretic analysis. 5. Choice and handling of plasmid and E. coli is essential. Generally, most commercially available cell strains are suitable, provided that the level of competence is guaranteed. This is particularly important for steps where the plasmid insert is nicked/cut removing the super-coiling effect. This makes the plasmid effectively larger and, therefore, more ­difficult to insert through the bacterial cell wall/membrane. Also, where DAM methylation of the template is required for DpnI treatment, it is important that the cell strain is not DAM methylase-negative (most strains are DAM-positive although it is important to check the genome information). 6. Competence can be a limiting factor with transformations and the higher the cell competence, the more chance of success (see Chapter 4, Subheading 3.1.1). 7. Seaplaque agarose is more pure and has a lower melting ­temperature; this is desirable for gel-elution. 8. Ligation of the fragment results in the disruption of the b-­lactamase gene. Plating transformants onto X-gal containing agar will result in blue and white colonies for those without inserts and with a disruptive insert respectively. 9. Refer to http://www.stratagene.com/manuals/211190.pdf for a specific pCR-script protocol (Stratagene, UK). 10. The interpretation of mutagenesis is clearly a huge topic and will depend on the type of experiment carried out. The analysis described here is based on that of Hulme et al. for analysis of the muscarinic receptors (13). It should be noted that mutants may disrupt more than one function. This leads to complicated changes in receptor function; their interpretation is considered in the original article. Other approaches are possible, for example, based on double mutant cycle analysis (28).

The Use of Site-Directed Mutagenesis to Study GPCRs

97

References 1. Kenakin, T.P. (2008) Seven transmembrane receptors as nature’s prototype allosteric protein: de-emphasizing the geography of binding. Mol. Pharmacol. 74, 541–543. 2. Cherezov, V., Rosenbaum, D.M., Hanson, M.A., Rasmussen, S.G., Thian, F.S., Kobilka, T.S., Choi, H.J., Kuhn, P., Weis, W.I., Kobilka, B.K., and Stevens, R.C. (2007) High-resolution crystal structure of an engineered human beta2-adrenergic G protein-coupled receptor. Science 318, 1258–1265. 3. Palczewski, K., Kumasaka, T., Hori, T., Behnke, C.A., Motoshima, H., Fox, B.A., Le Trong, I., Teller, D.C., Okada, T., Stenkamp, R.E., Yamamoto, M., and Miyano, M. (2000) Crystal structure of rhodopsin: A G proteincoupled receptor. Science 289, 739–745. 4. Dromey, J.R. and Pfleger, K.D. (2008) G protein coupled receptors as drug targets: the role of beta-arrestins. Endocr. Metab. Immune Disord. Drug Targets 8, 51–61. 5. Evans, B.N., Rosenblatt, M.I., Mnayer, L.O., Oliver, K.R., and Dickerson, I.M. (2000) CGRP-RCP, a novel protein required for signal transduction at calcitonin gene-related peptide and adrenomedullin receptors. J. Biol. Chem. 275, 31438–31443. 6. Hay, D.L., Poyner, D.R., and Sexton, P.M. (2006) GPCR modulation by RAMPs. Pharmacol. Ther. 109, 173–197. 7. Vohra, S., Chintapalli, S.V., Illingworth, C.J., Reeves, P.J., Mullineaux, P.M., Clark, H.S., Dean, M.K., Upton, G.J., and Reynolds, C.A. (2007) Computational studies of Family A and Family B GPCRs. Biochem. Soc. Trans. 35, 749–754. 8. Rovati, G.E., Capra, V., and Neubig, R.R. (2007) The Highly Conserved DRY Motif of Class A G Protein-Coupled Receptors: Beyond the Ground State. Mol. Pharmacol. 71, 959–964. 9. Conner, M., Hawtin, S.R., Simms, J., Wootten, D.L., Lawson, Z., Conner, A.C., Parslow, R.A., and Wheatley, M. (2007) Systematic analysis of the entire second extracellular loop of the V1a vasopressin receptor: Key residues, conserved throughout a G protein-coupled receptor family, identified. J. Biol. Chem. 282, 17405–17412. 10. Conner, A.C., Hay, D.L., Simms, J., Howitt, S.G., Schindler, M., Smith, D.M., Wheatley, M., and Poyner, D.R. (2005) A key role for transmembrane prolines in calcitonin receptorlike receptor agonist binding and signalling: implications for family B G protein-coupled receptors. Mol. Pharmacol. 67, 20–31.

11. Conner, A.C., Simms, J., Conner, M.T., Wootten, D.L., Wheatley, M., and Poyner, D.R. (2006) Diverse functional motifs within the three intracellular loops of the CGRP1 receptor. Biochemistry 45, 12976–12985. 12. Strader, C.D., Candelore, M.R., Hill, W.S., Sigal, I.S., and Dixon, R.A. (1989) Iden­ tification of two serine residues involved in agonist activation of the beta-adrenergic receptor. J. Biol. Chem. 264, 13572–13578. 13. Hulme, E.C., Lu, Z.L., Ward, S.D., Allman, K., and Curtis, C.A. (1999) The conformational switch in 7-transmembrane receptors: the muscarinic receptor paradigm. Eur. J. Pharmacol. 375, 247–260. 14. Clackson, T. and Wells, J.A. (1995) A hot spot of binding energy in a hormone-receptor interface. Science 267, 383–386. 15. Holst, B., Zoffmann, S., Elling, C.E., Hjorth, S.A., and Schwartz, T.W. (1998) Steric hindrance mutagenesis versus alanine scan in mapping of ligand binding sites in the tachykinin NK1 receptor. Mol. Pharmacol. 53, 166–175. 16. Sine, S.M., Wang, H.L., and Bren, N. (2002) Lysine scanning mutagenesis delineates structural model of the nicotinic receptor ligand binding domain. J. Biol. Chem. 277, 29210–29223. 17. Shi, L. and Javitch, J. A. (2004) The second extracellular loop of the dopamine D2 receptor lines the binding-site crevice. Proc. Natl. Acad. Sci. U.S.A. 101, 440–445. 18. Thomas, B. E., Woznica, I., Mierke, D. F., Wittelsberger, A., and Rosenblatt, M. (2008) Conformational changes in the parathyroid hormone receptor associated with activation by agonist. Mol. Endocrinol. 22, 1154–1162. 19. Hubbell, W.L., Altenbach, C., Hubbell, C.M., and Khorana, H.G. (2003) Rhodopsin structure, dynamics, and activation: a perspective from crystallography, site-directed spin labeling, sulfhydryl reactivity, and disulfide crosslinking. Adv. Protein. Chem. 63, 243–290. 20. Yao, X., Parnot, C., Deupi, X., Ratnala, V.R., Swaminath, G., Farrens, D., and Kobilka, B. (2006) Coupling ligand structure to specific conformational switches in the beta2-adrenoceptor. Nat. Chem. Biol. 2, 417–422. 21. Sheikh, S.P., Zvyaga, T.A., Lichtarge, O., Sakmar, T.P., and Bourne, H.R. (1996) Rhodopsin activation blocked by metal-ionbinding sites linking transmembrane helices C and F. Nature 383, 347–350. 22. Sheikh, S.P., Vilardarga, J.P., Baranski, T.J., Lichtarge, O., Iiri, T., Meng, E. C., Nissenson,

98

Conner et al.

R.A., and Bourne, H.R. (1999) Similar structures and shared switch mechanisms of the beta2-adrenoceptor and the parathyroid hormone receptor. Zn(II) bridges between helices III and VI block activation. J. Biol. Chem. 274, 17033–17041. 23. Hawtin, S.R. (2005) Charged residues of the conserved DRY triplet of the vasopressin V1a receptor provide molecular determinants for cell surface delivery and internalization. Mol. Pharmacol. 68, 1172–1182. 24. Hawtin, S.R., Wesley, V.J., Simms, J., Parslow, R.A., Miles, A., McEwan, K., Keen, M., and Wheatley, M. (2003) An arginyl in the N-terminus of the V1a vasopressin receptor is part of the conformational switch controlling activation by agonist. Eur. J. Biochem. 270, 4681–4688. 25. Beukers, M.W. and Ijzerman, A.P. (2005) Techniques: how to boost GPCR mutagenesis

studies using yeast. Trends Pharmacol. Sci. 26, 533–539. 26. Li, B., Scarselli, M., Knudsen, C.D., Kim, S.K., Jacobson, K.A., McMillin, S. M., and Wess, J. (2007) Rapid identification of functionally critical amino acids in a G protein-coupled receptor. Nat. Methods 4, 169–174. 27. Hawtin, S.R., Simms, J., Conner, M., Lawson, Z., Parslow, R.A., Trim, J., Sheppard, A., and Wheatley, M. (2006) Charged extracellular residues, conserved throughout a G proteincoupled receptor family, are required for ligand binding, receptor activation, and cellsurface expression. J. Biol. Chem. 281, 38478–38488. 28. Naider, F., Becker, J.M., Lee, Y.H., and Horovitz, A. (2007) Double-mutant cycle scanning of the interaction of a peptide ligand and its G protein-coupled receptor. Biochemistry 46, 3476–3481.

Chapter 6 Approaches to Study GPCR Regulation in Native Systems Jonathon M. Willets Abstract The ability to assess whether individual proteins are involved in the signalling or regulation of G ­protein-coupled receptor signalling is highly dependent on the pharmacological tools available. In the absence of appropriate pharmacological agents, alternative molecular approaches have been developed to alter either protein function or expression. This has included the use of mutants, for example catalytically inactive (kinase-dead) enzymes, which when overexpressed function as dominant negatives to inhibit endogenous enzyme function, and more latterly small (21–23 bp) interfering RNA dsRNA oligos, whose antisense strand is designed complementary to the target protein mRNA and which can be used to deplete target protein expression. Critically, the success of these approaches depends on the transfection efficiency, and the chosen experimental assay in the cell type studied. Therefore, three transfection techniques and their merits and drawbacks are described. In addition, one method of examining G proteincoupled receptor (GPCR) regulation, combining siRNA-mediated GRK depletion and imaging of fluorescent GPCR ­signalling reporter biosensors in difficult-to-transfect cells is briefly described. Key words: GPCR, GRK, siRNA, Receptor regulation, Dominant negative, Cell transfection

1. Introduction G protein-coupled receptors (GPCRs) comprise a wide and diverse family of seven-transmembrane proteins which decode a plethora of extracellular chemical signals into physiological outputs, including the mediation of neurotransmission, olfactory detection, and hormonal signals (1, 2). GPCRs represent major therapeutic targets and the continuing identification of orphan GPCRs offers opportunity for future pharmacological and therapeutic development (3). The advent of GPCR cloning and the subsequent use of heterologous expression model cell systems (e.g. HEK293, CHO, see Chapters 1–3, 7) have identified many important molecules involved in signalling, and the regulation of signalling by GPCRs.

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_6, © Springer Science+Business Media, LLC 2011

99

100

Willets

Whilst important, this approach does not always reflect the way that endogenous GPCRs are regulated in their native environments. Indeed, accumulating evidence indicates that identical GCPRs can be regulated differently depending on the cell background and whether they are exogenously or endogenously expressed (4–8). Therefore, to gain understanding into the physiological roles and relevance of GPCR signalling, it is of paramount importance to investigate their signalling and its regulation within their native environment. The task of delineating pathways and identifying individual proteins responsible for endogenous GPCR regulation is often dependent on the pharmacological tools available. For example, the dissection of GPCR-stimulated MAPK signalling pathways is now routine with an extensive choice of specific inhibitory cellpermeable compounds. However, where suitable pharmacological inhibitors are not available, alternative approaches to manipulate individual protein activity/function or expression are required. Whilst useful to identify potential protein–GPCR interactions, resulting data from wild-type protein overexpression studies (designed to elevate protein activity) should be treated with care, since overexpression may lead to “off-target” or non-selective effects (7, 9–11). More importantly, this approach provides no information regarding the potential for endogenous candidate protein–GPCR interactions. Suppression of endogenous protein expression, through RNAi strategies (discussed below), or activity via overexpression of functionally inactive (dominant negative) versions of the protein of interest (12–17) provides attractive alternative methods to explore the function of endogenous proteins. However, due to the requirement for protein overexpression, dominant-negative strategies are also open to the potential criticism of “off-target” effects, or a lack of sufficient specificity towards the protein being targeted. Currently, there are no effective chemical inhibitors available to inhibit G protein-coupled receptor kinase (GRK) function, which makes them ideal candidates for genetic manipulation techniques. Therefore, the relative merits of utilising genetic techniques to manipulate target proteins in native tissues will be discussed in relation to GRK function, in studies designed to characterise their roles in the regulation of endogenous GPCR signalling. As their name suggests, the GRK protein family are kinases that phosphorylate key serine/threonine residues within the third intracellular loop and/or C-terminal domains of agonist occupied GPCRs to initiate the process of receptor desensitisation (4, 18). GRK-mediated phosphorylation enhances GPCR affinity for non-visual arrestin proteins, whose binding physically inhibits GPCR/G-protein interactions to prevent prolonged or inappropriate signalling (19). Although the primary role for GRK proteins is in regulating GPCR signalling (4, 19), many diverse

Approaches to Study GPCR Regulation in Native Systems

101

non-receptor-based functions have been identified ranging from phosphorylation of phosducins, synucleins, tubulin, ribosomal protein P2, ezrin (18) to functioning as a histone deacetylase kinase (20). This chapter, therefore, focuses on two of the commonly used molecular biological techniques employed to manipulate individual GRK protein expression/activities. Whilst the strategies highlighted here are applicable to the majority of cellular proteins, focus will be on their use in the manipulation of GRK activities in native cell systems. GRK function can be altered in two main ways. Firstly, and in common with many kinases, the catalytic domain can be disrupted using site-directed mutagenesis, targeting conserved lysine residues within the ATP-binding domain to create the following mutants: K220RGRK2; K220RGRK3; K216M, K217MGRK4; K215RGRK5; and K215RGRK6 (4). The resultant non-ATP binding (or kinasedead) mutants can then be overexpressed in cells, functioning as dominant-negative mutants (DNMs) by out-competing ­endogenous kinases for GPCR docking sites. Since it is uncertain whether such mutants only compete/sequester specific kinases, or also bind to receptors or other regulatory proteins, other approaches have been used to target specific protein expression. Great advances have been made in this direction recently with the advent of RNA interference technology (21). This approach utilises short unique double-stranded sequences of mRNA (21–23 bp) whose antisense strand is complementary to the target mRNA. When introduced into cells, the dsRNA is unwound and the antisense targeting strand is incorporated into the RNA-induced silencing complex, binding to the target protein mRNA and ­promoting subsequent destruction (21). Overexpression of ­dominant-negative GRK mutants and the degree of siRNA-­ mediated endogenous GRK depletion can be assessed by standard Western blotting techniques. The RNAi technique offers a highly specific method for depleting individual proteins, although care should be taken when designing RNAi constructs to prevent interaction with non-target mRNAs. However, unlike DNMs, which can be used in any cell background, RNAi constructs are usually species-specific, which requires design and validation of new constructs when cell background/species change.

2. Materials 1. Lipofectamine2000 cell culture transfection reagent (see Note 1). 2. Interferin RNAi cell culture transfection reagent (see Note 2). 3. Amaxa nucleofection cell culture transfection kit (see Note 3).

102

Willets

4. Cell culture media, foetal calf serum, antibiotics (penicillin 100  units/mL, streptomycin 100  mg/mL), antimyotics (e.g. Fungizone 2.5 mg/mL), and culture plates. 5. Dominant-negative GRK constructs (e.g. D110A,K220RGRK2, D110A,K220R GRK3, K216M, K217MGRK4, K215RGRK5, K215RGRK6) and appropriate plasmid empty vector control (e.g. pcDNA3) (see Note 4). 6. GRK-protein and anti-GAPDH protein-specific primary ­antibodies (see Note 5). 7. siRNA against the target protein, negative control (non-­ targeting siRNA) positive control anti-GAPDH siRNA (see Note 6). 8. Equipment: Amaxa nucleofection machine (available from Lonza, UK), cell culture incubator, cell culture cabinet, standard Western blotting kit, and access to a fluorescence or laser scanning confocal microscope (although these ­techniques are transferable to virtually any GPCR activity/­signalling assay, see Note 7).

3. Methods 3.1. Transfection Protocol for Plasmid cDNAs (e.g. DNM GRKs)

This protocol is based on cells seeded into a six-well culture plate. 1. Cells should be seeded at a density to produce 40–50% ­confluency 24 h later (see Note 8). 2. After 24 h, change the cell culture medium and replace with 2 mL of serum containing antibiotic/antifungal-free media. 3. To make one transfection reaction solution, add 50  mL of fresh non-supplemented cell culture medium and place in a sterile capped plastic tube. Repeat this into another identical tube. 4. Add cDNA of the required DNM to one tube, replace the lid, and mix by tapping tube gently; do not vortex (see Note 9). 5. Add Lipofectamine2000 to the other tube directly into the cell culture media and not down the side of the tube. Replace cap and mix by tapping tube gently, do not vortex. 6. Incubate tubes at room temperature for 5 min, then transfer the contents of one tube to the other, replace the cap, and mix as described above. Leave for 20  min at room temp­ erature. 7. Add drop by drop to a chosen well of six-well culture plate, mix by gently shaking plate and return to the cell incubator (37°C) for 4 h.

Approaches to Study GPCR Regulation in Native Systems

103

8. Next replace transfection media with fresh pre-warmed (37°C) normal growth media. 9. Finally, return cells to the cell culture incubator and leave it for 48 h prior to experimentation. 3.2. Transfection Protocol for siRNA

1. Cells should be harvested by trypsinisation for 3–5  min. Trypsinisation is stopped with excess culture media and then cells were centrifuged at 200 × g for 5  min. Resuspend cell pellet in 10 mL of supplemented culture media. Determine cell number using a haemocytometer, and seed at 150,000 per well of a six-well culture plate (see Note 6). 2. After 24 h, change the cell culture medium and replace with 2  mL of pre-warmed (37°C) serum containing antibiotic/ antifungal-free media. 3. Resuspend siRNA in appropriate volume of RNase-free water (see Note 10). 4. To make the transfection solution, take 200 mL of fresh nonsupplemented cell culture medium and place in a sterile RNase free-capped plastic tube. Then add the required concentration of siRNA (see Note 6). Finally, add the appropriate ­volume of Interferin transfection reagent (see manufacturer’s instructions). Replace the tube cap and vortex mixture at medium speed for 10  s. Leave at room temperature for 10  min, then add drop by drop to one well of the six-well culture plate. 5. Mix by gently shaking cell culture plate and place in cell incubator for 4 h. 6. After 4 h, replace transfection mix with pre-warmed (37°C) normal growth media. In general, the cells will be ready for experimental use within 48 h of transfection (see Note 11).

3.3. Amaxa Transfection

1. Cells should be passaged approximately for 2–3 days before transfection to ensure that they are in the exponential growth phase. 2. Harvest cell by trypsinisation when 70–85% confluent. Stop trypsinisation with excess culture media then centrifuge (200 × g for 5 min) to pellet and resuspend in 10 mL of fresh media. 3. Count the cells using a haemocytometer and aliquot to give between 1 × 106 and 5 × 106 cells per reaction into a fresh sterile tube (see Note 12). 4. Aliquot 0.5 mL of complete cell culture media into an appropriate number of sterile tubes to match the number of reactions to be performed and pre-warm at 37°C in the cell incubator. Mix together the Amaxa transfection reagent components

104

Willets

(0.5 mL of supplement to 2.25 mL of Nucleofector solution) and warm to room temperature (see Note 13). Set the Amaxa machine to the correct protocol (see Note 3) and prepare enough electroporation cuvettes for the number of transfection reactions. 5. Centrifuge the tube (200 × g for 5 min) to pellet the cells and remove all residual cell culture media. Resuspend the cells in 100 mL of transfection reagent, mix and add the appropriate concentration of cDNA or siRNA (maximum of 5  mg of cDNA or 3 mg of siRNA is recommended, and make sure that the total volume does not exceed 10  mL), and mix again ensuring not to introduce bubbles to the mixture. Do not leave cells suspended in the transfection mix for longer than 15–20  min as this will reduce cell viability and transfection efficiency. 6. Transfer the cell/reaction mix into an Amaxa cuvette taking care not to introduce bubbles as this will decrease the transfection efficiency. Replace the lid and transfer into the Amaxa machine, rotate the cuvette housing, and activate the transfection protocol (see Note 3). 7. Immediately after the transfection process has finished, gently transfer the cells from the cuvette into 0.5  mL of the prewarmed cell culture media using the specially provided micropipette, mix gently, and plate into the appropriate well(s) of a cell culture plate (see Note 14). 8. Repeat above process until all transfection reactions are ­complete and leave cells to grow for 48 h prior to experimentation. To avoid the potential for false-negative data caused by insufficient transfection, it is essential to determine the transfection efficiency in your cells. To achieve this, most commercial suppliers offer both positive (e.g. anti-GAPDH) and negative control (non-targeting) and fluorescently tagged siRNAs (see Notes 6 and 15). 3.4. Assessment of Target Proteins Role in GPCR Signalling

1. Once the transfection technique has been optimised for either maximal overexpression of DNM or depletion of target protein, one is free to undertake experiments of choice to determine the role the target protein plays in GCPR regulation/ signalling. It is, however, worth noting that the choice of experiment will depend heavily upon the transfection efficiency. In our experience, standard cell population biochemical experiments (e.g. MAPK or GPCR phosphorylation assays) only yield discernible results if the DNM transfection efficiency is above 60% or depletion of target protein ³70% by Western blot (see Note 6, Fig. 1b). The potential role that target proteins play in GPCR regulation can be assessed in difficult-to-transfect cells following co-transfection of cells

Approaches to Study GPCR Regulation in Native Systems

105

Fig. 1.  Western blots showing effects of dominant-negative mutants and siRNA-targeted protein depletion on GRK expression in ULTR cells. (a) Overexpression of GRK dominant-negative constructs is shown in cells (grown in a six-well culture plate) transfected (at 50% confluence) with either (0.5 mM) pcDNA3 (vector control, lane 1), D110A,K220RGRK2, D110A,K220RGRK3, K215R GRK5, or K215RGRK6 (dominant-negative GRK construct expression shown in lane 2 ). (b) Specific depletion of individual GRK expression is shown in cells transfected with 100 nM of either, negative-control (lane 1), anti-GRK2 (lane 2 ), anti-GRK3 (lane 3 ), anti-GRK5 (lane 4 ), or anti-GRK6 (lane 5 ) siRNAs. (c) Confocal microscopy image (×20 magnification) of C6 glioma cells 48 h after Amaxa nucleofection transfection with eGFP-tagged K220RGRK2 (1 mg). Transfection efficiency can be determined by comparing the number of eGFP-K220RGRK2 expressing cells (left ) with the total number of cells displayed in the phase contrast image (right ). The image shows that ~80% of cells are transfected with eGFP-K220RGRK2. (d) Representative Western blot showing endogenous GRK2 depletion in C6 glioma cells 48 h after Amaxa nucleofection with anti-GRK2 siRNA. Lower panel shows that GRK3 expression is unaltered by anti-GRK2 siRNA treatment. NT nontransfected cells, NC cells transfected with negative-control (non-targeting) siRNA.

with a fluorescent signalling reporter and either DNM or siRNA against the target protein (13–15, 22). This approach often leads to >90% co-transfection and when combined with confocal imaging enables “real-time” assessment of target protein effects upon GPCR signalling (13–15, 22). An example of this procedure is outlined below. 2. To enable co-transfection, the standard protocols outlined above can be modified slightly. For Lipofectamine2000-based co-transfections, add 0.5  mg of chosen DNM and 0.5  mg of fluorescent reporter (e.g. eGFP-PH\ domain an IP3/PLC biosensor) into the same tube during the initial stage as outlined in Subheading 3.1, step 4. It is important to alter the amount of Lipofectamine2000 added to maintain the ratio of 1 mg of cDNA/siRNA to 3 mL of Lipofectamine2000. Complete the rest of the transfection

Willets

a

R2

R1

IP3 (F/F0 488nm)

3.5

pcDNA3 D110A,K220RGRK2 K215RGRK6

3.0 2.5 2.0 1.5 1.0 0.5 0

b

100 200 300 400 500 600 700 Time (sec) R1

R2

IP3 (F/F0 488nm)

4.0 NC siRNA GRK2 siRNA GRK6 siRNA

3.5 3.0 2.5 2.0 1.5 1.0 0.5 0

c

250

R1

500 Time (sec)

750

Rmax

R2

1000

2.5 IP3 (F/F0 488nm)

106

NC siRNA GRK2 siRNA GRK6 siRNA

2.0 1.5 1.0 0.5 0

250

500 Time (sec)

750

1000

Fig. 2. Manipulations of GRK activity identify specific roles for GRK2 and GRK6 in the regulation of oxytocin and histamine H1 receptor signalling. ULTR cells were co-transfected with eGFP-PH (0.5 mM) and pcDNA3, D110A,K220RGRK2, K215RGRK6 (0.5 mM), anti-GRK2, antiGRK6, or negative-control siRNAs using the Lipofectamine2000 method. After 48 h, cells were stimulated with agonist and translocation of the eGFP-PH IP3 biosensor from plasma membrane to cytoplasm monitored using confocal microscopy as described previously and in Chapter 18. To assess oxytocin receptor desensitisation, cells were challenged twice with oxytocin (100 nM) for 30 s, before (R1) and after (R2) a 5-min washout period. Reduced R2 compared to R1 responses are indicative of GPCR desensitisation (13, 15, 22). (a) The presence of pcDNA3 and D110A,K220RGRK2 failed to prevent the reduced IP3 production to a second oxytocin (100 nM, 30 s) challenge, whilst inhibition of GRK6 following K215RGRK6 expression largely prevents oxytocin receptor desensitisation. (b) siRNA-mediated depletion of GRK6 largely prevents oxytocin receptor desensitisation, whilst GRK2 knockout and negative-control siRNA have no-effect. (c) Due to the existence of a large receptor reserve, an alternative protocol is required to observe histamine H1 receptor desensitisation (13). Cells challenged with an EC50 histamine

Approaches to Study GPCR Regulation in Native Systems

107

as outlined in Subheading 3.1. If siRNA is to be used, then replace DNM with appropriate siRNA at the concentration required to cause maximal protein depletion. For Amaxa transfections, mix the siRNA or DNM GRK cDNA with the fluorescent reporter cDNA during stage in Subheading 3.3, step 6. Complete the rest of the protocol as described in Subheading 3.3. 3. The effects of GRK depletion or DNM inhibition on GPCR signalling can be assessed using standard confocal microscopy techniques 2 days post-transfection (13–15, 22) (see Fig. 2 for experimental data).

4. Notes 1. The suitability of transfection reagents is highly dependent on the cell type to be transfected. Although many good reagents are commercially available, we find that Lipofectamine2000 (available from Invitrogen, Paisley, UK) is particularly good, especially when using difficult-to-transfect cell preparations (13, 14, 22). Achieving maximal cell transfection can be assessed using eGFP or eGFP-tagged constructs with different transfection reagents. cDNA concentrations to transfection reagent volumes should be varied to optimise the procedure. It should be noted that transfection reagents can be cytotoxic especially when used in large volumes. However, cytoxicity can be dramatically reduced by removing antibiotics and antimyotics from the culture medium during the transfection process. As a general rule, when using Lipofectamine2000 a 3:1 ratio of Liopfectamine2000 (mL) to cDNA (mg) is recommended. 2. Many siRNA-compatible transfection reagents are commercially available. We have used Interferin (from Polyplus supplied by Autogen Bioclear, UK) which is relatively non-toxic, depleting >70% of GRK protein expression in many of the cell types we have examined including the ULTR immortalised human myometrial cell line (13, 15) and SH-SY5Y human neuroblastoma cells.

F­ ig. 2. (continued) concentration (10 mM) for 30 s(R1, R2), 5 min either side of a maximal (desensitising) agonist stimulation (100  mM, for 1  min Rmax) show a reduced R2 compared to R1 response, which is indicative of receptor desensitisation. Here GRK2 depletion inhibits H1 receptor desensitisation, whilst GRK6 depletion is ineffective. In all cases, inclusion of dominant-negative or siRNA constructs targeting GRKs 3 and 5 did not affect oxytocin of histamine H1 receptor desensitisation.

108

Willets

3. Amaxa kits are supplied from Lonza, UK, and are individually designed for many diverse cell lines and primary cells, a full list of which can be found on the following website http:// www.lonzabio.com/no_cache/cell-database. Several preset nucleofection protocols may be recommended for one cell type. Therefore, it is important to test all suggested protocols, as although some protocols provide greater transfection efficiency this can sometimes be associated with reduced cell survival. eGFP plasmid cDNA is provided with each kit, which when combined with fluorescent microscopy enables assessment of transfection efficiency and identification of the optimal protocol. 4. DNM GRK proteins (K220RGRK2, K220RGRK3, K215RGRK5, or K215R GRK6) are not commercially available but usually obtainable from most laboratories where they are routinely used. K220R GRK2 and K220RGRK3 proteins directly bind to activated Gaq inhibiting phospholipase C signalling (23). As such, their use is compromised when IP3 or Ca2+ are used as the experimental readout of GPCR regulation. Mutation of the Gaq-binding (e.g. D110A,K220RGRK2 (24) or D110A,K220RGRK3 (14)) domain prevents Gaq-binding, whilst retaining the ability to inhibit endogenous GRK function. 5. The polyclonal GRK antibodies used here are from Santa Cruz (GRK2 cat. no. sc562; GRK3 sc563; GRK5 sc565; GRK6 sc566). GRK2, 3, and 6 antibodies are used at 1:1,000 and GRK5 1:500 dilutions in 50  mM Tris–HCl (pH 7.5), 150  mM NaCl, 5% milk protein (w/v) overnight at 4°C. Secondary anti-rabbit (Sigma-Aldrich, UK, cat. noA6154) is used at 1:1,000 dilution for 1 h at room temperature. 6. Since all assessments of siRNA activity require a suitable reproducible demonstration of protein (or to a lesser extent mRNA) depletion, it is important to optimise firstly the transfection and secondly the detection procedures. Initial cell plating density is important for siRNA transfections. Most transfection reagent manufacturers will provide suggested recommendations for the appropriate number of cells to seed into a certain sized cell culture plate/flask, which should be taken as a rough guide and again can be optimised to suit the cell type. It is worth noting that as with Lipofectaminestyle transfection reagents, siRNA transfection reagents are not suitable for all cell types, particularly difficult-to-transfect cells (e.g. primary vascular smooth muscle). Assessment of transfection efficiency is important to avoid false-negative data. Therefore, it is advisable to conduct experiments using commercially available positive control (e.g. anti-GAPHD siRNA) and negative control non-targeting siRNA, which will assist in determining siRNA and transfection effectiveness.

Approaches to Study GPCR Regulation in Native Systems

109

Fluorescent siRNA constructs are commercially available to assist with determining transfection efficiency, although the resultant data can be difficult to interpret. When planning RNAi-mediated protein depletion experiments, help can be obtained from many websites (both commercial and free) available to assist in the design of RNAi constructs. Since the key to RNAi-based techniques is to specifically deplete only the target protein, a blast search of the sequence is essential to identify any potential cross-reactivity prior to purchase. However, the RNAi sequence will only be made available on receipt of your purchase, so it is advisable to ask the company to undertake the blast search and provide the relevant details. The guarantees provided by commercial suppliers vary so it is sensible to check these details prior to purchase. Many RNAi sequences are guaranteed to deplete >70% of target proteins, and companies will usually supply replacements if it can be proved that their product is unable to achieve this goal. However, commercial suppliers often require proof at the mRNA level whilst most researchers and more importantly reviewers prefer evidence of protein knockout. Crucially, it is important to realise that some RNAi sequences may give false-negative data due to low transfection efficiency; therefore, the cell transfection protocol should be optimised as described above to discount this potential scenario. Furthermore, it is important to note that the observed degree of protein depletion will also depend upon the rate of turnover (synthesis/degradation) of the target protein, which may vary depending upon the cell background. Therefore, it is advisable to conduct a series of experiments to assess the effectiveness of the siRNA over several time points (e.g. 24, 48, 72 h). After optimising the transfection procedure, target protein expression can be assessed via Western blotting (see Fig.  1b) or if no suitable antibody is available by RT-PCR (25). As with all Western blotting, a high quality antibody is required to enable clear assessment of siRNA depletion of endogenous protein. In general, using the above techniques, >70% of ­target proteins (including all GRKs) can be depleted using an siRNA concentration of 10–100  nM after 48  h. Finally, the specificity of siRNA effects should be tested, especially with proteins sharing high degrees of mRNA homology. In the case of the GRK family, this can be achieved by examining the expression of non-targeted GRK proteins. 7. Whilst this chapter focuses on the use of fluorescent reporters to determine the activity/regulation of GPCR signalling pathways, it is important to note that these transfection techniques can be utilised to prepare cells for alternative readouts. Indeed,

110

Willets

the effects of siRNA-mediated depleted GRK expression or dominant-negative-mediated reduced kinase activity have been reported using GPCR phosphorylation (8, 26), and ­biochemical detection of activated MAPK signalling cascades (27, 28) as experimental readouts. However, when undertaking biochemical studies on transfected cells, it is essential to optimise transfection efficiency to prevent false-negative data when not enough (less than 50%) of the cells are transfected (see Note 6). 8. Cell confluency at transfection can significantly alter transfection efficiency and the optimum differs depending upon which cell type is used. As a general rule, cells are seeded at a volume that provides 40–50% confluency 24  h later, although this may need to be reduced for rapidly growing cells. The aim is to have a confluent monolayer for experimentation 48 h posttransfection, although for imaging experiments sub-confluent (70–80%) cultures are generally recommended. This process will need to be optimised with different initial pre-transfection confluencies and fluorescently tagged proteins or eGFP can be used to determine transfection efficiency. 9. The quality and purity of cDNA or siRNA play an important role in any transfection procedure. cDNA should be resuspended in either deionised water or Tris–HCl (10 mM, pH 8.0) EDTA (1 mM) at a concentration of 1–5 mg/mL. Purity can be assessed by determining the absorbance ratio measured at 260 and 280 nm (see Chapter 4), and a value of at least 1.8 is required for efficient transfections. 10. Care must be taken when using siRNA to prevent direct human contact as this will degrade the RNA. It is recommended that when handling siRNA, clean gloves are always worn. To resuspend the siRNA, briefly centrifuge (1  min, 200 × g) and add the required volume of RNase-free water (usually supplied) and vortex. All pipette tips should be sterile and contain filters to prevent contamination and siRNA degradation. 11. The time-course of protein depletion is governed by many variables, of which individual protein half-life and the stability of the siRNA are the most prominent. In our experience, siRNA delivered as a dsRNA oligo causes maximal GRK depletion 48 h after transfection (see Fig. 1b). At later time points, endogenous protein levels begin to rise again due to siRNA degradation. If prolonged protein depletion is required, the RNAi sequence can be cloned into a plasmidbased system, referred to as short-hairpin (sh)RNA. Since this approach is plasmid-based, protein-depletion usually does not reach maximal levels until 72 h after transfection, but has the advantage of continual RNAi production

Approaches to Study GPCR Regulation in Native Systems

111

(i.e. long-term protein depletion). In addition, many shRNA plasmids contain antibiotic selection genes enabling generation of stable cell lines. 12. The number of cells per Amaxa reaction is dependent upon the cell type and the optimum number will be quoted in the kit instructions. However, it is advisable to undertake pilot experiments using the provided eGFP plasmid and fluorescent microscopy to determine the optimum cell number for maximal cell transfection. 13. Once mixed together, the Amaxa transfection reagents have a shelf life of only 3 months; therefore, it may be prudent to mix together only enough to undertake the required number of experiments at any one time. 14. It is essential to pre-warm media (0.5  mL per reaction) to 37°C to allow maximal cell recovery and prevent cell death after the electroporation stage of the Amaxa nucleofection transfection process. 15. Amaxa nucleofection reactions utilise a finite number of cells. Therefore, depending upon the experimental requirements, cells should be divided into the appropriate number of wells to enable confluent monolayers within 2 days. Some optimisation will be required to achieve this aim. References 1. Fredriksson R., Lagerstrom, M.C., Lundin, L.G. and Schioth, H.B. (2003) The G-proteincoupled receptors in the human genome form five main families. Phylogenetic analysis, ­paralogon groups, and fingerprints. Mol. Pharmacol. 63, 1256–1272. 2. Pierce, K.L., Premont, R.T. and Lefkowitz, R.J. (2002) Seven-transmembrane receptors. Nat. Rev. Mol. Cell Biol. 3, 639–650. 3. Wilson, S., Bergsma, D.J., Chambers, J.K., Muir, A.I., Fantom, K.G., Ellis, C., Murdock, P.R., Herrity, N.C. and Stadel, J.M. (1998) Orphan G-protein-coupled receptors: the next generation of drug targets? Br. J. Pharmacol. 125, 1387–1392. 4. Willets, J.M., Challiss, R.A. and Nahorski, S.R. (2003) Non-visual GRKs: are we seeing the whole picture? Trends Pharmacol. Sci. 24, 626–633. 5. Kong, G., Penn, R. and Benovic, J.L. (1994) A b-adrenergic receptor kinase dominant ­negative mutant attenuates desensitization of the b2-adrenergic receptor. J. Biol. Chem. 269, 13084–13087. 6. Simon, V., Robin, M.T., Legrand, C. and Cohen-Tannoudji, J. (2003) Endogenous G

protein-coupled receptor kinase 6 triggers homologous beta-adrenergic receptor desensitization in primary uterine smooth muscle cells. Endocrinology 144, 3058–3066. 7. Debburman, S.K., Kunapuli, P., Benovic, J.L. and Hosey, M.M. (1995) Agonist-dependent phosphorylation of human muscarinic receptors in Spodoptera frugiperda insect cell membranes by G protein-coupled receptor kinases. Mol. Pharmacol. 47, 224–233. 8. Willets, J.M., Challiss, R.A. and Nahorski, S.R. (2002) Endogenous G protein-coupled receptor kinase 6 Regulates M3 muscarinic acetylcholine receptor phosphorylation and desensitization in human SH-SY5Y neuroblastoma cells. J. Biol. Chem. 277, 15523–15529. 9. Willets, J.M., Challiss, R.A., Kelly, E. and Nahorski, S.R. (2001) G protein-coupled receptor kinases 3 and 6 use different pathways to desensitize the endogenous M3 muscarinic acetylcholine receptor in human SH-SY5Y cells. Mol. Pharmacol. 60, 321–330. 10. Diviani, D., Lattion, A.L., Larbi, N., Kunapuli, P., Pronin, A., Benovic, J.L. and Cotecchia, S. (1996) Effect of different G protein-coupled

112

Willets

receptor kinases on phosphorylation and desensitization of the a1B-adrenergic receptor. J. Biol. Chem. 271, 5049–5058. 11. Oppermann, M., Freedman, N.J., Alexander, R.W. and Lefkowitz, R.J. (1996) Phos­ phorylation of the type 1A angiotensin II receptor by G protein-coupled receptor kinases and protein kinase C. J. Biol. Chem. 271, 13266–13272. 12. Schafer, B., Marg, B., Gschwind, A. and Ullrich, A. (2004) Distinct ADAM metalloproteinases regulate G protein-coupled receptor-induced cell proliferation and survival. J. Biol. Chem. 279, 47929–47938. 13. Willets, J.M., Taylor, A.H., Shaw, H., Konje, J.C. and Challiss, R.A. (2008) Selective regulation of H1 histamine receptor signalling by G protein-coupled receptor kinase 2 in uterine smooth muscle cells. Mol. Endocrinol. 22, 1893–1907. 14. Morris, G.E., Nelson, C.P., Standen, N.B., Challiss, R.A. and Willets, J.M. (2010) Endothelin signalling in arterial smooth muscle is tightly regulated by G protein-coupled receptor kinase 2. Cardiovasc. Res. 85, 424–433. 15. Willets, J.M., Brighton, P.J., Mistry, R., Morris, G.E., Konje, J.C. and Challiss, R.A. (2009) Regulation of oxytocin receptor responsiveness by G protein-coupled receptor kinase 6 in human myometrial smooth muscle. Mol. Endocrinol. 23, 1272–1280. 16. Labasque, M., Reiter, E., Becamel, C., Bockaert, J. and Marin, P. (2008) Physical interaction of calmodulin with the 5hydroxytryptamine2C receptor C-terminus is essential for G protein-independent, arrestindependent receptor signalling. Mol. Biol. Cell 19, 4640–4650. 17. Hirotani, S., Otsu, K., Nishida, K., Higuchi, Y., Morita, T., Nakayama, H., Yamaguchi, O., Mano, T., Matsumura, Y., Ueno, H., Tada, M. and Hori, M. (2002) Involvement of nuclear factor-kB and apoptosis signal-­regulating kinase 1 in G-protein-coupled receptor agonist-induced cardiomyocyte hypertrophy. Circulation 105, 509–515. 18. Penela, P., Murga, C., Ribas, C., Tutor, A.S., Peregrin, S. and Mayor, F., Jr. (2006) Mechanisms of regulation of G protein-coupled receptor kinases (GRKs) and cardiovascular ­disease. Cardiovasc. Res. 69, 46–56. 19. Gesty-Palmer, D. and Luttrell, L.M. (2008) Heptahelical terpsichory. Who calls the tune? J. Recept. Signal Transduct. Res. 28, 39–58. 20. Martini, J.S., Raake, P., Vinge, L.E., DeGeorge, B.R., Jr., Chuprun, J.K., Harris, D.M., Gao, E., Eckhart, A.D., Pitcher, J.A. and Koch, W.J.

(2008) Uncovering G ­protein-coupled receptor kinase-5 as a histone deacetylase kinase in the nucleus of cardiomyocytes. Proc. Natl. Acad. Sci. U S A 105, 12457–12462. 21. Rao, M. and Sockanathan, S. (2005) Molecular mechanisms of RNAi: implications for development and disease. Birth Defects Res. C. Embryo Today 75, 28–42. 22. Willets, J.M., Nash, M.S., Challiss, R.A. and Nahorski, S.R. (2004) Imaging of muscarinic acetylcholine receptor signaling in hippocampal neurons: evidence for phosphorylationdependent and -independent regulation by G-protein-coupled receptor kinases. J. Neurosci. 24, 4157–4162. 23. Carman, C.V., Parent, J.L., Day, P.W., Pronin, A.N., Sternweis, P.M., Wedegaertner, P.B., Gilman, A.G., Benovic, J.L. and Kozasa, T. (1999) Selective regulation of Gaq/11 by an RGS domain in the G protein-coupled receptor kinase, GRK2. J. Biol. Chem. 274, 34483–34492. 24. Sterne-Marr, R., Tesmer, J.J., Day, P.W., Stracquatanio, R.P., Cilente, J.A., O’Connor, K.E., Pronin, A.N., Benovic, J.L. and Wedegaertner, P.B. (2003) G protein-coupled receptor kinase 2/Gaq/11 interaction. A novel surface on a regulator of G protein signaling homology domain for binding G alpha ­subunits. J. Biol. Chem. 278, 6050–6058. 25. Wang, X., Dong, L., Xie, J., Tong, T. and Zhan, Q. (2009) Stable knockdown of Aurora-A by vector-based RNA interference in human oesophageal squamous cell carcinoma cell line inhibits tumor cell proliferation, invasion and enhances apoptosis. Cancer Biol. Ther. 8, 1852–1859. 26. Willets, J.M., Mistry, R., Nahorski, S.R. and Challiss, R.A. (2003) Specificity of G proteincoupled receptor kinase 6-mediated phosphorylation and regulation of single-cell M3 muscarinic acetylcholine receptor signalling. Mol. Pharmacol. 64, 1059–1068. 27. Kim, J., Ahn, S., Ren, X.R., Whalen, E.J., Reiter, E., Wei, H. and Lefkowitz, R.J. (2005) Functional antagonism of different G proteincoupled receptor kinases for b-arrestin-mediated angiotensin II receptor signalling. Proc. Natl. Acad. Sci. U S A 102, 1442–1447. 28. Ren, X.R., Reiter, E., Ahn, S., Kim, J., Chen, W. and Lefkowitz, R.J. (2005) Different G protein-coupled receptor kinases govern G protein and b-arrestin-mediated signaling of V2 vasopressin receptor. Proc. Natl. Acad. Sci. U S A, 102, 1448–1453.

Chapter 7 Heterologous Expression of GPCRs in Fission Yeast John Davey and Graham Ladds Abstract In this chapter, we describe methods to heterologously express G protein-coupled receptors (GPCRs) in the fission yeast Schizosaccharomyces (Sz.) pombe. GPCRs regulate a diverse range of biological processes in all eukaryotic cells, including plants, insects, humans, and yeast. The high degree of conservation between GPCRs from different organisms has facilitated the development of a large number of model systems to enable study of this pharmaceutically important family of cell-surface receptors. Of the many model systems available for investigating GPCRs, yeast have proven to be one of the more attractive. Yeasts’ amenability to both genetic and biochemical manipulation, a reduced number of endogenous GPCRs and their relative low culturing costs has facilitated their use in many high-throughput drug screens. Given the high number of detailed methods relating to the expression of GPCRs within budding yeast, we have focused our attention on the use of fission yeast as a model system. We describe the methods used and provide examples from our own experiences of expressing a number of human GPCRs in Sz. pombe cells. Key words: G protein-coupled receptor, Yeast, Schizosaccharomyces pombe, Immunoblotting, b-Galactosidase assays, Cell-surface expression, Green fluorescent protein

1. Introduction Many external signals are detected through a wide variety of cell-surface receptors. One such family is the G protein-coupled receptors (GPCRs) which link to heterotrimeric G proteins consisting of a Ga, Gb, and Gg. In the inactive state, a Ga subunit is bound to a molecule of GDP. On agonist stimulation of a GPCR, nucleotide exchange occurs on the Ga subunit such that GDP dissociates and is replaced by the more cellularly abundant GTP. This promotes dissociation of Ga-GTP from the Gbg dimer. Each moiety is then able to regulate the activity of effector proteins, thereby bringing about changes in cellular behavior. Signaling is

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_7, © Springer Science+Business Media, LLC 2011

113

114

Davey and Ladds

terminated when Ga-bound GTP is hydrolyzed to GDP leading to the re-association of the heterotrimer. This reaction can be catalyzed by regulator of G protein signaling (RGS) proteins that act as GTPase-activating proteins for the Ga-GTP (1). Our understanding of GPCRs and their associated signaling components is not purely restricted to the academic environment. Within humans, GPCRs are responsible for controlling a wide variety of physiological process, including coordination of the immune system and regulation of the autonomic system and, as such, represent one of the most valuable therapeutic targets for the pharmaceutical industry. Indeed ~50% of all prescription drugs currently available on the market today are targeted to GPCRs. The desire for more selective and specific drugs is one of the major drivers for further investigations. However, investigations into GPCR signaling within mammalian systems are often hampered by the large numbers of signaling components expressed within individual cells, e.g., a cardiovascular cell possesses >100 different GPCRs (2). For this reason, many researchers have adopted the use of simple model organisms to complement their studies. Of the many systems developed, yeast are one of the most attractive. The mechanisms of GPCR signaling in yeast are similar to those found within mammalian systems, and in many cases the molecules involved are readily interchangeable. Moreover, yeast have been responsible for a number of key conceptual advances within the GPCR field such as the identification of the first RGS protein (1) and the suggestion that Gbg subunits can activate downstream signaling cascades (reviewed in ref. 3). Through the  use of selected modifications to the yeast cells, it has been possible to use them as hosts for the development and implementation of high-throughput screens. Further, the GPCRs expressed within the yeast cells are readily amenable to both molecular and genetic manipulations. These abilities coupled to the relatively low culturing costs make yeast ideal cells to investigate GPCRs. A great deal of the research in this field has used the pheromoneresponse pathway from the budding yeast Saccharomyces cerevi­ siae, however, in recent times, the distantly related fission yeast Schizosaccharomyces pombe has emerged as an alternative host organism. Although on initial inspection these model systems may appear similar, a number of differences in their abilities to express GPCRs have been identified (for a detailed review see ref. 3). Given the wealth of literature dedicated to the use of Sc. cerevi­ siae, here we have focused specifically on the methods and techniques that enable heterologous expression of GPCRs in Sz. pombe. Sz. pombe is a haploid organism existing in one of two mating types M (Minus) and P (Plus). Under conditions of nutrient limitation, cells of opposite mating type can fuse to form a diploid.

Heterologous Expression of GPCRs in Fission Yeast

115

This process of mating is controlled by the reciprocal exchange of diffusible pheromones, such that an M cell secretes M-factor and expresses the P-factor receptor (Mam2) at its cell surface. Conversely, P-cells secrete P-factor and express the M-factor receptor (Map3). Both GPCRs, irrespective of mating type, on ligand binding activate a Ga subunit (Gpa1) which (unlike in other yeast species) propagates the response by stimulation of a mitogen-activated protein kinase cascade leading to expression of proteins required for mating (reviewed in ref. 4). In their native state, Sz. pombe cells are unable to couple heterologously expressed GPCRs to their endogenous pheromone-responsive signaling components. However, we have created a panel of Sz. pombe cells expressing a number of quantitative reporters that have been specifically engineered heterologously to express GPCRs. For a full listing of the genetic manipulations required to generate these strains, including deletion of the endogenous receptors, integration of ligand-dependent responsive elements within strains, removal of genes responsible for nutrient selection, and the manipulation of the endogenous Ga subunit, Gpa1, readers are directed to refs. 3 and 5. This panel of strains is available on request.

2. Materials 2.1. Sz. pombe Materials

1. Yeast extract (YE).

2.1.1. Media

3. 0.05% Adenine hemisulfate.

2. 20% d-glucose, autoclaved. 4. 0.05% Leucine. 5. 0.05% Uracil. 6. Yeast nitrogen base without amino acids. 7. Ammonium chloride. 8. Di-sodium hydrogen orthophosphate, anhydrous. 9. Potassium hydrogen phthalate. 10. Stock salts (see Subheading 2.1.7). 11. Stock vitamins (see Subheading 2.1.8). 12. Stock minerals (see Subheading 2.1.9). 13. Select agar. 14. 10-cm media plates.

2.1.2. Transformations

1. 0.1 M lithium acetate. 2. 0.1  M lithium acetate containing 50% polyethylene glycol (PEG) 4000.

116

Davey and Ladds

3. 30 and 42°C water baths. 4. Sterile bent glass rods for spreading on plates. 2.1.3. Yeast Extract

1. 5 g/L of yeast extract. 2. 30 g/L of d-glucose. Dissolve the yeast extract and glucose in 1,000 mL of water and autoclave. To supplement with adenine, leucine, and uracil (YEALU), add 250  mg/L adenine hemisulfate, l-leucine, and uracil, respectively, prior to autoclaving. For plates, include 15 g/L of agar before autoclaving.

2.1.4. Synthetic Dropout Medium (AA)

1. 6.7 g/L of yeast nitrogen base without amino acids. 2. 20 g/L of d-glucose. 3. 2 g of appropriate drop out mix/L. Dissolve the yeast nitrogen base without amino acids and the appropriate drop out mix in 900  mL of water and autoclave. When solution has cooled to £55°C, add 100 mL of 20% sterile d-glucose. For plates, dissolve the yeast nitrogen base without amino acids and the appropriate drop out mix in 500 mL of water and autoclave. In a separate bottle, dissolve 20 g/L of d-glucose and 15 g/L of agar in 500 mL of water and autoclave. When the solutions have cooled to £55°C, add the yeast nitrogen base mix to the molten agar. Pour plates immediately.

2.1.5. Dropout Mix

2.1.6. Defined Minimal Medium (Described by Davey et al. (6))

Add 2  g of all the following nutrients into a sterile 100-mL glass  bottle containing two glass marbles: l-alanine, l-arginine, l-asparagine, l-cysteine, l-glutamine, l-glutamate, l-glycine, l-isoleucine, l-lysine, l-phenylalanine, l-proline, l-serine, ­l-threonine, l-tryptophan, l-tyrosine, l-valine, myo-inositol, adenine hemisulfate, l-histidine, l-methionine, uracil, and myo-­inositol. Add 4  g of l-leucine and 0.4  g of para-amino benzoic acid. Premixed dropout powders can be purchased from commercial sources if required. 1. 5 g/L ammonium chloride. 2. 2.2 g/L di-sodium hydrogen orthophosphate anhydrous. 3. 3 g/L potassium hydrogen phthalate. 4. 20% d-glucose. 5. 10 mL of 50× stock salts solution. 6. 1 mL of 1,000× stock vitamins solution. 7. 100 mL of 10,000× stock minerals solution. Dissolve the ammonium chloride, di-sodium hydrogen orthophosphate anhydrous, and potassium hydrogen phthalate in

Heterologous Expression of GPCRs in Fission Yeast

117

880  mL of water and autoclave. When solution has cooled to £55°C, add 100 mL of 20% sterile d-glucose, 10 mL of stock salt solution, 1  mL 1,000× stock vitamins solution, and 100  mL of 10,000× stock minerals solution. To supplement with adenine, leucine, and uracil (DMMALU), add 250 mg/L adenine hemisulfate, l-leucine, and uracil, respectively, prior to autoclaving. For plates include 15 g/L of agar before autoclaving. 2.1.7. 50× Stock Salt Solution

1. 52.5 g/L of magnesium chloride hexahydrate. 2. 735 mg/L of calcium chloride dehydrate. 3. 50 g/L of potassium chloride. 4. 2 g/L anhydrous sodium sulfate. Dissolve all components in 1,000 mL of water and autoclave. Store in 20-mL aliquots at 4°C.

2.1.8 1,000× Stock Vitamins Solution

1. 10 g/L of myo-inositol. 2. 10 g/L of nicotinic acid. 3. 1 g/L of pantothenic acid. 4. 10 mg/L of D-biotin. Dissolve all components in 100  mL of water; sterilize by ­filtration using a 0.22-mm nonpyrogenic sterile filter unit. Store in 20-mL aliquots at 4°C.

2.1.9. 10,000× Stock Minerals Solution

1. 10 g/L of citric acid. 2. 5 g/L of boric acid. 3. 5 g/L of manganese sulfate hydrate. 4. 4 g/L of zinc sulfate heptahydrate. 5. 3.05 g/L of molybdic acid. 6. 2 g/L of iron chloride hexahydrate. 7. 1 g/L of potassium iodide. 8. 0.4 g/L of copper sulfate pentahydrate. Dissolve all components in 100  mL of water; sterilize by ­ ltration using a 0.22-mm nonpyrogenic sterile filter unit. Store in fi 20-mL aliquots at 4°C.

2.1.10. Z Buffer/ONPG

Z buffer 1. 16.1 g/L of Na2HPO4·7H2O. 2. 5.5 g/L of NaH2PO4·H2O. 3. 0.75 g/L of KCl. 4. 0.246 g/L of MgSO4·7H2O. Adjust the pH to 7.0 and sterilize.

118

Davey and Ladds

Z buffer/ONPG Prepare fresh by adding 0.27  mL of b-mercaptoethanol, 500 mL of chloroform, and 67 mg of 2-nitrophenyl b-d-galactopyranoside to 100 mL of Z buffer. 2.2. Confocal Localization of Heterologously Expressed GPCRs

1. Growth media (typically DMMAU – DMM lacking additional leucine).

2.3. Immunoblotting for Heterologously Expressed GPCRs

1. Resuspension buffer (100  mM Tris–HCl, pH 6.8, 10  mM EDTA).

2. Polylysine-coated cover slips. 3. Standard confocal microscope (e.g., Leica TCS SP5 II confocal – Leica Microsystems).

2. 75% Sucrose. 3. TEN buffer (100  mM Tris–HCl, pH 6.8, 10  mM EDTA, 150 mM NaCl). 4. 20× Inhibitor stock. 5. Sucrose cushion buffers of 50, 10, and 5% (see Sub­ headings 2.3.1–2.3.4). 6. Beckman Coulter TL 100 benchtop ultracentrifuge. 7. Thick and thin-walled ultracentrifuge tubes. 8. Antibody-recognizing receptor or epitope tag (e.g., anti-GFP; Santa Cruz Biotechnology Inc.). 9. Apparatus and solutions for standard sodium dodecyl sulfate– polyacrylamide gel electrophoresis (SDS-PAGE) and immun­ oblotting.

2.3.1. 20× Inhibitor Stock

1. 20 mM phenylmethylsulfonyl fluoride (PMSF) (add prior to use from a 100-mM stock). 2. 2 mM N-tosyl-phenylalanine chloromethyl ketone (TPCK). 3. 2 mM pepstatin. 4. 20 mM N-tosyl-l-lysine chloromethyl ketone (TLCK). 5. 20 mM leupeptin. Make fresh each time. Dissolve all components in 250 mL of TEN buffer containing 5% sucrose. Alternatively (for ease) dissolve one tablet of complete mini-protease inhibitor cocktail in 250 mL of TEN buffer containing 5% sucrose supplemented with 2 mM TPCK and 20 mM TLCK.

2.3.2. 5% SucroseCushion Buffer

To make 1 mL 1. 50 mL of 20× inhibitor stock. 2. 67 mL of 75% sucrose.

Heterologous Expression of GPCRs in Fission Yeast

119

3. 883 mL of TEN buffer. Make fresh each time and store at 4°C. 2.3.3. 10% SucroseCushion Buffer

To make 1 mL 1. 133 mL of 75% sucrose. 2. 200 mM PMSF. 3. 2 mM TPCK. 4. 855 mL of TEN buffer. Make fresh each time and store at 4°C.

2.3.4. 50% SucroseCushion Buffer

To make 1 mL 1. 666 mL of 75% sucrose. 2. 200 mM PMSF. 3. 2 mM TPCK. 4. 322 mL of TEN buffer. Make fresh each time and store at 4°C.

2.3.5. Urea Buffer

To make 1 mL 1. 19.22 g urea. 2. 2 g SDS. 3. 12 mL of 0.5 M Tris-HCL, pH 6.8. Make up to 36 mL and dissolve in a 65°C water bath. Confirm that volume is still 36 mL and store at 4°C.

2.3.6. 2× Urea-Containing Sample Buffer for SDS-PAGE

1. 3.6 mL of urea buffer (Subheading 2.3.5). 2. 0.4 mL of b-mercaptoethanol. 3. 10 mg bromophenol blue. 4. 6 mL of sterile water. This urea sample buffer can be stored at 4°C, however, prior to use it is required to be heated to 50°C. Use with an equal ­volume of plasma membrane extracts.

3. Methods 3.1. Sz. pombe Plasmid Transformation Protocols

To simplify the description of the methods, it is assumed that any heterologous GPCR will be expressed within Sz. pombe cells using the thiamine-repressible series of pREP plasmids (7, 8) (See Note 1). These plasmids carry the Leu or Ura nutritional selection. It is advised that whenever possible the pREP3x plasmid (maximal expression and Leu nutritional selection) is used. If co-expression

120

Davey and Ladds

with a second GPCR is required, then the pREP4x (maximal expression and Ura nutritional selection) is advised. 3.1.1. Plasmid Transformations Using Lithium Acetate

The following protocol is based on the highly efficient lithium acetate method which was first developed by Okazaki et al. (9) for high-efficiency transformations within Sz. pombe and was improved by Kanter-Smoler et al. (10). The method can be scaled up or down as required. The desired transformation efficiency is between 102 to 103 transformants/mg of plasmid. If required, more than one plasmid (containing different nutritional selections) can be transformed at one time (e.g., co-expression of two different GPCRs when they are thought to heterodimerize), but be aware that the efficiency will decrease by at least one order of magnitude owing to reduced probability of a yeast cell taking up both plasmids. A small scale version of the transformation protocol: 1. Grow 10  mL of culture in defined minimal medium (DMM) containing low glucose (typically 10% of normal DMM – medium where the glucose concentration is low thereby partially metabolically compromising cells to increase transformation efficiency) to a density of 0.5–1 × 107 cells/mL (typical OD600 = 0.2–0.5). 2. Harvest cells at 1,750 × g, 5  min in a nonrefrigerated ­centrifuge (23–29°C). 3. Resuspend cells in ~1  mL of 0.1  M lithium acetate (LiAc) at 1 × 109 cells/mL and transfer to a sterile 1.5mL Eppendorf tube. Repellet cells by centrifugation (1,750 × g for 1 min at 29°C). 4. Resuspend cells at 1 × 109 cells/mL 0.1  M LiAc and incubate as 100 mL aliquots at 30°C for 60–120 min. Note: Cells will sediment at this stage. 5. Mix ~1 mg of plasmid DNA (in a total volume no more than 10 mL) with 290 mL of 0.1 M LiAc containing 50% (w/v) PEG 4000. Mix thoroughly and prewarm to 30°C. 6. Add one aliquot of cells to each DNA containing tube, being careful not to allow the cells to cool. Mix gently with a pipette and return to 30°C for a further 50–60 min. 7. Heat shock cells, without agitation, at 42°C for 15 min. 8. Collect cells by centrifugation at 800 × g for 3  min at 29°C and resuspend in 200 mL of water. 9. Spread 100 mL aliquots on selective plates. 10. Successful transformants are expected to appear within 4–6 days at 30°C.

Heterologous Expression of GPCRs in Fission Yeast

121

11. Colonies should then be transferred to a fresh selective plate and re-streaked to single colonies. These cells are now ready for the determination of GPCR expression. Note: if a low transformation efficiency is expected, then prior to spreading on selective plates transfer cells to 10 mL of YEALU (see Subheading  2.1.3) and incubate with shaking at 30°C for 60 min. Then collect cells and plate as normal in step 9; alternatively, the electroporation protocol may be attempted. 3.1.2. Plasmid Transformations Using Electroporation

If time is a major issue and a large number of transformations are required to be performed, an alternative to the lithium acetate transformation protocol is the use of electroporation. Precise standard protocols for using this technique in Sz. pombe are rare and generally researchers stick to a method they find works. It is believed that using electroporation increases transformation efficiency, although this is yet to be systematically verified. A major advantage of the electroporation protocol is the ability to premake the cells and to store them at −80°C ready for use on the day (11). The protocol below is the preferred method within our laboratory and generally gives transformation efficiencies of 103–104 transformants/mg of plasmid. It has been suggested that transformation efficiencies of 105–106 transformants/mg of plasmid can be achieved, although this may require the addition of 1,4-dithiothreitol (DTT) (12). This procedure is an amalgamation of a number of different methods, and information contained within the documentation that comes with the electroporators is dependent on instru­ mentation. 1. Grow cells to a density of 0.5–1 × 107/mL (typical OD600 =  0.2–0.5) in DMM. Transformation frequency is not harmed by growth until early stationary phase (OD600 = 1.5). 2. Harvest cells by centrifugation at 1,750 × g for 5  min at 20°C. 3. Cells should be washed once in ice-cold sterile water and prepelleted. 4. Remove ice-cold water and resuspend in ice-cold 1 M sorbitol. Suga and Hatakeyama (12) suggest that the addition of 25  mM DTT for 15  min at this stage enhances electrocompetence. 5. Cells are repelleted again and finally resuspended in ice-cold 1 M sorbitol at a density of 1–5 × 109/mL. 6. Take 40  mL of the cell suspension and add this to chilled Eppendorf tubes containing the DNA for transformation (100 ng) and incubate on ice for 5 min.

122

Davey and Ladds

7. The electroporator is set as follows: (a) 1.5 kV, 200 Ohms, 25 mF (Biorad). (b) Use your manufacturer’s suggested settings if in doubt. 8. Transfer the cells and DNA to a prechilled cuvette and pulse. 9. Immediately add 0.9  mL of ice-cold 1  M sorbitol to the cuvette and then return the cell suspension to the Eppendorf tube and place on ice while other electroporations are being performed. 10. Cells are then plated onto minimal selective medium. It is not required to add sorbitol to the agar and, indeed, it has been suggested that this may even retard growth (11). 11. Successful transformants are expected to appear within 4–6 days at 30°C. 3.2. GPCR Analysis in Sz. pombe Cells 3.2.1. Fluorescence Microscopy to Assess Localization of GPCRs

Following successful yeast transformations and the selection of appropriate colonies, we prefer, before assessing activity, to determine the expression and localization of the heterologously expressed GPCRs (Fig. 1). Experience has shown that a failure of yeast cells to express GPCRs on their plasma membrane directly correlates with an inability to respond to exogenously administered ligand. It is, therefore, advisable to generate a fluorescently labeled GPCR for visualization using standard fluorescence microscopy. Typically, a C-terminal tag can be added to most GPCRs without too much interference with activity. It is, however, recommended that both fluorescent-tagged and nonfluorescent-tagged versions of the same GPCR are expressed in yeast cells to determine activity. To enable sufficient expression of the GPCRs to allow visualization, cells are required to be grown for at least 16 h to enable full expression of the thiamine-repressible nmt1 promoter contained

Fig. 1. Localization of the human CRF-R1A receptor in Schizosaccharomyces pombe. The human CRF-R1A receptor was cloned into pREP3x with GFP fused in frame at the C -terminus, then expressed in an Sz. pombe reporter strain lacking the endogenous mam2 receptor. Cells were cultured to mid-exponential phase in DMM lacking leucine and then analyzed by confocal microscopy. Comparisons were made to (a) GFP expressed alone and (b) Mam2-GFP (the endogenous pheromone receptor) to help determine the localization of (c) hCRF-R1A-GFP. White arrows denote vacuoles and dotted arrows the plasma membrane. The scale for each image is the same as indicated for GFP (a).

Heterologous Expression of GPCRs in Fission Yeast

123

within the pREP series of Sz. pombe expression vectors (7, 8). If a GPCR fails to be localized to the plasma membrane, a number of different signal peptides can be added immediately upstream of the endogenous initiator codon of the receptor. Signal peptide examples include the type II GPCR corticotrophin-releasing factor (CRF) receptor or the signal peptide from the budding yeast a-­ factor precursor, which has been successfully used to enhance ­targeting of GPCRs in yeast (13–18) (see Note 4). 1. Inoculate 20  mL of appropriate growth medium (usually DMMAU – as GPCRs will be expressed using the pREP3x vector) with a single colony of yeast taken from a freshly restreaked plate. 2. Grow the cells for approximately 24  h, and then dilute the growing culture into fresh medium. Typically, this will involve removing 1–2  mL of culture and resuspending in a further 20 mL of appropriate growth medium. 3. Grow these cultures to a density of ~5 × 106 cells/mL (typical OD600 = 0.2). 4. Harvest 1  mL of culture by centrifugation at 1,750 × g for 3 min in a bench microfuge. 5. Wash the cells twice in fresh growth medium, then resuspend in a final volume of 20 mL of Vectashield® (Vector Laboratories, Inc., Burlingame, CA). Vectashield prevents the fluorescent chromophore from rapid fading on excitation. 6. Place 2 mL of cell suspension onto polylysine-coated slides. 7. Cover the suspension with a cover split and seal with nail varnish. 8. Cells can then be viewed using any fluorescence microscope; typically, we use a True Confocal Scanner Leica TCS SP5 II microscope. 3.2.2. Western Blot Analysis to Determine Expression of GPCRs

While fluorescence microscopy can give a good indication as to the cellular location of the heterologous GPCR, it is not a definitive measure of expression. We, therefore, also prepare plasma membrane extracts taken from cells expressing the GPCR of choice to confirm plasma membrane expression. In addition, we use nondenaturing protein gels to indicate the oligomeric state of the GPCR (Fig. 2). The technique below indicates how to produce plasma membrane extracts from our yeast cells. While there are a number of variations of how to produce pure plasma membrane extracts, we have only included our preferred methods.

3.2.2.1. Preparation of Plasma Membrane Extracts

1. Grow 200  mL culture of cells expressing the heterologous expressed GPCR in appropriate medium (usually DMMAU) to a density of between 0.5 and 1 × 107 cells/mL (Typical OD600 = 0.2–0.5).

124

Davey and Ladds

Fig.  2. Immuno-analysis of the yeast pheromone receptors Mam2 and STE2. The endogenous Sz. pombe P-factor receptor Mam2 and heterologously expressed Saccharomyces cerevisiae a-factor receptor STE2 were C-terminally tagged with 6xHis, cloned separately into pREP3x, and expressed in Sz. pombe cells. Plasma membrane cell extracts were separated by nonreducing SDS-PAGE and immunoblotted with anti-6×His antibodies. Lane 1: Mam2-6×His, lane 2: STE2-6×His. The open and filled arrows highlight Mam2-6×His and STE2-6×His homodimers, respectively. Sizes of markers are indicated in kDa.

2. Resuspend the cells in 20  mL of ice-cold TEN buffer and transfer to a 20-mL universal vial. 3. Recentrifuge to harvest the cells (1,750 × g for 5 min, 4°C). 4. Drain the cell pellets completely and add 25 mL of 20× inhibitor stock. 5. Add 10 mL of ice-cold acid-washed glass beads (425–600 mm) and vortex-mix vigorously for at least 1  min, but no more than 3 min – acid-washed glass beads are used to shear the cell wall of the yeast cells. 6. Add 1 mL of 1× inhibitor stock, vortex-mix briefly; pierce the bottom of the universal vial with a small hot hypodermic needle. Insert the tube into a 50-mL Falcon tube and centrifuge at 1,750 × g – this step removes the cell lysate from the glass beads.

Heterologous Expression of GPCRs in Fission Yeast

125

7. Transfer the eluate to a 2-mL Eppendorf tube, and centrifuge for 10 s at 5,500 × g (in a bench microfuge). This will produce a thick “milky” layer away from the cell debris at the bottom of the tube. 8. Transfer the upper fluid layer from the 2-mL tube to a sucrose cushion prepared as follows: Use a 2.3-mL thin-walled ultracentrifuge tube for a TLS55 rotor for a Beckman TL 100 ultracentrifuge. Use 0.5 mL of 50% sucrose and 0.8 mL of 10% sucrose. Layer the cell lysate on top of the 10% sucrose buffer. 9. Centrifuge at 42,000 × g for 30 min at 4°C. It is worth considering using a gentle deceleration from 3,500 × g (typically over 5 min) to minimize disturbance of the membrane layer. 10. Harvest the membranes, a milky layer at the 10–50% sucrose boundary, into a 2-mL Eppendorf tube. 11. To the membrane layer, add ~3 volumes of sterile ice-cold water and centrifuge at 4°C for 1 min at 12,000 × g in a bench microfuge. 12. Resuspend the pellet in ~1.4  mL of ice-cold 10  mM Tris/ HCl, pH 6.8. 13. Transfer to a thick-walled polycarbonate tube in a TLS55 rotor of a Beckman TL 100 ultracentrifuge and centrifuge at 42,000 × g for 15 min at 4°C. 14. Resuspend the membranes in 200  mL of ice-cold 10  mM Tris/HCl, pH 6.8, 20% sucrose. 15. Membranes can be stored at −20°C until they are required for use. 16. To isolate pure plasma membrane fractions, use a discontinuous sucrose gradient (containing inhibitors), but made from three layers of 85, 60, and 50% sucrose. 17. Add membrane extracts from step 14 on to the discontinuous sucrose gradient and centrifuge at 18,000 × g for 14  h in a TLS55 rotor of a Beckman TL 100 ultracentrifuge. 18. The membranes banded at the 85–60% layer are pure plasma membranes. 19. Collect membranes, and repeat steps 12–14. 3.2.2.2. Detection of GPCRs in Plasma Membrane Extracts

1. Thaw the extracts from step 15 or 19 above on ice. 2. Resuspend the extracts in an equal volume of 2× urea sample buffer (Subheading 2.3.6). We find that the addition of urea to the sample buffer enables the GPCRs to be resolved on the SDS-PAGE gel. A failure to add the urea can result in the GPCRs not migrating from the loading wells (see Note 6 and 7).

126

Davey and Ladds

3. Warm the samples to 37°C for at least 10 min – do not boil (see Note 8). 4. Load the samples on to SDS-PAGE gels suitable for resolving the individual GPCRs. Typically, we use the Invitrogen precast 4–20% gradient Bis–Tris gels run following the manufacturer’s instructions (see Note 9). 5. Transfer to polyvinylidene fluoride (PVDF) membranes using standard methods (see Note 10). We also like to determine the total protein content in an extract using coomaise stain (see Note 11). 6. Immunoblot for the presence of the GPCR using either an antibody specific to the GPCR of choice, or one suitable to detect an epitope tag (e.g., GFP or hexa-histidine tags). 3.2.3. Quantitative Characterization of GPCR Activity

Following a successful detection of the expression of the heterologous GPCR, it is necessary to determine whether the receptor couples to the endogenous yeast pheromone-response pathway. We routinely accomplish this through the use of our engineered strains that contain the Escherichia coli LACZ gene that encodes b-galactosidase, under the control of the pheromone-dependent sxa2 genes promoter (19, 20). The sxa2 gene is only expressed in one mating type (M cells) in response to exposure to P-factor (21–23), and encodes for a secreted carboxypeptidase that inactivates the extracellular pheromone. By replacing the sxa2 open reading frame with the LACZ gene, we have generated cells that generate time- and pheromone-dependent production of b-galactosidase. Removal of the endogenous GPCR (Mam2) (see Note 2) and replacement with heterologous GPCRs enable the cells to be stimulated with ligands specific for the GPCR under investigation, and the extent of signaling quantified by measuring b-galactosidase production from the cells, e.g., CRF for the CRF-R1A receptor (Fig. 3). The method for quantifying the level of b-galactosidase produced from cells (5, 18) is a modified protocol adapted from Dohlman et al. (24). 1. Inoculate 20  mL of appropriate growth medium (usually DMMAU – as GPCRs will be expressed using the pREP3x vector) with a single colony of yeast taken from a freshly restreaked selective plate. For GPCRs that have ligands contained within the growth media (see Note 3). 2. Incubate the cells, with shaking, at 29°C for 24 h. 3. Dilute the growing culture one in ten by removing 18 mL of the culture and adding 18 mL of fresh growth medium to the remaining 2 mL of cells. Grow the cells for a further 24 h, with shaking, at 29°C. Note: it takes approximately 16 h for cells, when removed from medium containing thiamine, to maximally express their

Heterologous Expression of GPCRs in Fission Yeast

127

Fig. 3. CRF-R1A signaling in Sz. pombe cells. A series of Sz. pombe yeast reporter strains, each containing a different modified Ga subunit (Gas, Gai2, Gai3, and Ga16) to resemble a human Ga subunit (5), but all lacking the endogenous mam2 receptor, were transformed with pREP3x-CRF-R1A. Cells were exposed to CRF at increasing concentrations (between 0 and 10−4  M for 16  h). Production of b-galactosidase was assayed using 2-nitrophenyl-b-d-galactopyranoside as described in Subheading  3.2.2. Values are means ±S.E.M. of triplicate determinations from three independent isolates. Signaling can be seen to vary dependent on which Ga subunit is expressed within the cell.

GPCRs. This step ensures that all cells are now able to fully express the receptors. 4. Dilute the cultures to ~2 × 105 cells/mL and grow for at least 6 h. 5. During this period, the ligand for the assay can be prepared. If ligands are diluted in an organic solvent (e.g., methanol), this will be required to be removed before the cells are added. For ligands dissolved in an aqueous solution, they can be added directly to cells. 6. Aliquots of the ligand should be added to 2-mL safe-lock Eppendorf tubes in ~5–10  mL volumes at the appropriate required dilution. 7. To the 2-mL safe-lock Eppendorf tubes, add 500 mL aliquots of the cells when the culture has attained a cell density of ~5 × 105 cells/mL (see Note 5). 8. Incubate the tubes at 29°C for 16  h on a slowly rotating wheel (<35 × g). If the rotating wheel is too fast, the cells may settle out and will not respond to the ligand. 9. Following 16  h of incubation, remove the tubes and place them on ice. 10. Transfer a 50-mL aliquot from each tube to a fresh 1.6-mL Eppendorf tube containing 750  mL of Z buffer/ONPG (Subheading  2.1.10) and incubate on a rotating wheel at 29°C for 90 min.

128

Davey and Ladds

11. Reactions are terminated by the addition of 200 mL of 2 M Na2CO3. 12. Transfer each reaction to a 1-mL cuvette and measure the amount of ONPG product formed by analysis at OD420 using a spectrophotometer. 13. The remaining 450  mL of culture from step 9 should be ­gently sonicated to remove any clumps of cells and the cell density is determined. This can be performed using a spectrophotometer (OD600); however, we prefer to use a Z2 Coulter Channelyser (Beckman Coulter, Luton, UK) (see Note 12). 14. The amount of b-galactosidase produced is calculated as the ratio of ONPG product (OD420) formed to assayed cells using the formula OD420/106 cells.

4. Notes 1. A wide range of vectors with different promoters can be used to heterologously express GPCRs in Sz. pombe. Our preferred choice is a range of pREP vectors that express receptors under the control of the thiamine-repressible nmt1 promoter (7). A series of stepwise truncations of the TATA box element just upstream of the transcriptional start site progressively decreases the strength of the nmt1 promoter and forms the basis for a series of vectors that allow a range of expression levels (8). These plasmids carry the Leu or Ura nutritional selection enabling their retention within the yeast cells when plated on to selective medium. Forsburg (25) has previously described a comparison of the strength of different promoters when expressed in Sz. pombe. For a full range of Sz. pombe expression vectors, readers are directed to the Forsburg laboratory website “index of fission yeast plasmids” (http://www-rcf. usc.edu/~forsburg/vectors.html). 2. Successful coupling of heterologous GPCRs in Sz. pombe strains is greatly enhanced by removal of the endogenous Mam2 receptor and modification of the yeast Ga subunit Gpa1. These modifications have been documented in Ladds et al. (5). The panel of strains is available on request direct to the authors. 3. There are occasions where the heterologous GPCR being expressed is a receptor for one of the constituents of the growth medium. If this occurs, care should be taken to ensure that low enough levels of the required nutrient are added to the medium enabling growth, but not inducing signaling. As a rule of thumb, we begin by reducing the concentration in the growth medium by tenfold.

Heterologous Expression of GPCRs in Fission Yeast

129

4. A failure to detect a GPCR on the plasma membrane, or at least expressed within the cell will inevitably lead to a failure to induce a response when the cells are stimulated with the appropriate ligand. We have used a number of signal peptide sequences or pre-pro sequences from secreted proteins to enable expression. Our favored choice is the signal peptide of the human CRF-R1A receptor. 5. It is essential that 2-mL safe-lock Eppendorf tubes are used for the b-galactosidase assay. The yeast cells grow to a high cellular density during the 16 h of the assay and have been known to produce sufficient gas by-products within the tubes, resulting in normal 1.6-mL Eppendorf lids to “pop” open. This results in cells being lost into the incubator and a large mess to clear up the next day. 6. We have identified that the use of urea sample buffer as opposed to standard Laemmli buffer is a great advancement on our abilities to resolve heterologously expressed GPCRs. A simple warming of the extracts contained within the urea sample buffer gives a very good resolution of the GPCRs. 7. If homodimerization is suspected for heterologously expressed GPCRs and it is required to confirm this in the yeast, alternative SDS-PAGE gels can be used that contain ~10% urea. 8. We find that if samples are boiled prior to analysis by SDSPAGE, the GPCRs fail to migrate into the separating gel. This, we assume, is due to the aggregation of the GPCRs. To resolve this, we simply warm the extracts to 37°C prior to loading. 9. Typically, we prefer to use the Invitrogen precast 4–20% gradient Bis–Tris gels. These gels are nondenaturing and enable resolution of potential GPCR homodimers. 10. We typically use PVDF for our immunoblots. We have found a number of problems in the immunoblotting of heterologous GPCRs in yeast extracts when nitrocellulose is used. We do not have an explanation for this, but recommend the sole use of PVDF. 11. Although not quantitative, it is advisable to run a second SDS-PAGE gel that will be stained for total protein content. Typically, we use Coomassie staining for this; however, silver staining can also be utilized. 12. It is well documented that Sz. pombe cells, when challenged with pheromone undergo a unidirectional elongation toward the source of that pheromone. By directly replacing the endogenous GPCR with heterologous GPCRs, we have hijacked the pheromone-response pathway. The direct result of this is that any exogenously administered ligand that induces a response from the GPCR under investigation will result in

130

Davey and Ladds

the cells undergoing the same morphological change. While this is not a problem for the measuring of b-galactosidase, it can result in a discrepancy if the cell number is determined using a spectrophotometer. The increased cell mass results in an increase value for the OD600 which translates to a high cell density. It is for this reason that we recommend using a Coulter Channelyzer to determine cell number. References 1. Ross, E.M. and Wilkie, T.M. (2000) GTPaseactivating proteins for heterotrimeric G proteins: regulators of G protein signaling (RGS) and RGS-like proteins. Annu. Rev. Biochem. 69, 795–827. 2. Tang, C.M. and Insel, P.A. (2004) GPCR expression in the heart; “new” receptors in myocytes and fibroblasts. Trends Cardiovasc. Med. 14, 94–99. 3. Ladds, G., Goddard, A. and Davey, J. (2005) Functional analysis of heterologous GPCR signalling pathways in yeast. Trends Biotechnol. 7, 367–373. 4. Davey, J. (1998) Fusion of a fission yeast. Yeast 14, 1529–1566. 5. Ladds, G. Davis, K., Hillhouse, E. W. and Davey, J. (2003) Modified yeast cells to investigate the coupling of G protein-coupled receptors to specific G proteins. Mol. Microbiol. 47, 781–792. 6. Davey, J., Egel, R. and Nielsen, O. (1995) Pheromone procedures in fission yeast. In: Microbial Gene Techniques: Methods in Molecular Genetics. Adolph KW (Ed), Academic Press, San Diego, CA, USA, pp. 247–263. 7. Maundrell, K. (1990). nmt1 of Fission Yeast: A highly transcribed gene completely repressed by thiamine. J. Biol. Chem. 265, 10857–10864. 8. Maundrell, K. (1993). Thiamine-repressible expression vector pREP and pRIP for fission yeast. Gene 123, 127–130. 9. Okazaki, K., Okazaki, N., Kume, K., Jinno, S., Tanaka, K. and Okayama, H. (1990). Highfrequency transformation method and library transducing vectors for cloning mammalian cDNAs by trans-complementation of Schizosaccharomyces pombe. Nucl. Acid Res. 18, 6485–6489. 10. Kanter-Smoler ,G., Dahlkvist, A. and Sunnerhagen, P. (1994) Improved method for rapid transformation of intact Schizosaccharomyces pombe cells. Biotechniques 16, 798–800.

11. Suga, M. and Hatakeyama, T. (2003) Highefficiency electroporation by freezing intact yeast cells with addition of calcium. Curr. Genet. 43, 206–211. 12. Suga, M. and Hatakeyama, T. (2001) High efficiency transformation of Schizosaccharomyces pombe pretreated with thiol compounds by electroporation. Yeast 18, 1015–1021. 13. Huang, H. J., Liao, C. F., Yang, B. C. and Kuo, T. T. (1992) Functional expression of rat M5 muscarinic acetylcholine receptor in yeast. Biochem. Biophys. Res. Commun. 182, 1180–1186. 14. Hansen, M. K., Tams, J. W., Fahrenkrug, J. and Pedersen, P. A. (1999) Functional expression of rat VPAC1 receptor in Saccharomyces cerevisiae. Receptors Channels 6, 271–281. 15. Weiss, H. M., Haase, W., Michel, H. and Reilander, H. (1998) Comparative biochemical and pharmacological characterisation of the mouse 5HT5A 5-hydroxytryptamine receptor and the human b2-adrenergic receptor produced in the methylotrophic yeast Pichia pas­ toris. Biochem. J. 330, 1137–1147. 16. de Jong, L. A. A., Grunewald, S., Franke, J. P., Uges, D. R. A. and Bischoff, R. (2004) Purification and characterization of the recombinant human dopamine D2S receptor from Pichia pastoris. Protein Express. Purif. 33, 176–184. 17. Grunewald, S., Haase, W., Molsberger, E., Michel, H. and Reilander, H. (2004) Production of the human D2S receptor in the methylotrophic yeast P. pastoris. Receptors Channels 10, 37–50. 18. Kim, T. K., Zhang, R., Feng, W., Cai, J., Pierce, W. and Song, Z. H. (2005) Expression and characterisation of human CB1 cannabinoid receptor in methylotrophic yeast Pichia pastoris. Protein Expr. Purif. 40, 60–70. 19. Didmon, M., Davis, K., Watson, P., Ladds, G., Broad, P. and Davey, J. (2002) Identifying regulators of pheromone signalling in the fission yeast Schizosaccharomyces pombe. Curr. Genet. 41, 241–253.

Heterologous Expression of GPCRs in Fission Yeast 20. Ladds, G. and Davey, J. (2004). Analysis of human GPCRs in fission yeast. Curr. Opin. Drug Discov. Devel. 7, 683–691. 21. Imai, Y., and Yamamoto, M. (1991) Schizosaccharomyces pombe sxa1+ and sxa2+ encode putative proteases involved in the ­mating response. Mol. Cell. Biol. 12, 1827–1834. 22. Ladds, G., Rasmussen, E. M., Young, T., Nielsen, O. and Davey, J. (1996) The sxa2dependent inactivation of the P-factor mating pheromone in the fission yeast Schizosac­ charomyces pombe. Mol. Microbiol. 20, 35–42.

131

23. Ladds, G. and Davey, J. (2000) Sxa2 is a ­serine carboxypeptidase that degrades extracellular P-factor in the fission yeast Schizo­sac­charomyces pombe. Mol. Microbiol. 36, 377–390. 24. Dohlman, H. G., Apaniesk, D., Chen, Y., Song, J. and Nusskern, D. (1995) Inhibition of G protein signalling by dominant gain-offunction mutations in Sst2p, a pheromone desensitisation factor in Saccharomyces cerevi­ siae. Mol. Cell. Biol. 15, 3635–3643. 25. Forsburg, S.L. (1993) Comparison of Schizosaccharomyces pombe expression systems. Nucl. Acid Res. 21, 2955–2956.

wwwwwww

Part II Examining GPCR Expression and Agonist-Induced Covalent Modifications

wwwwwww

Chapter 8 Radioligand Binding Methods for Membrane Preparations and Intact Cells David B. Bylund and Myron L. Toews Abstract The radioligand binding assay is a relatively simple but powerful tool for studying G protein-coupled receptors. There are three basic types of radioligand binding experiments: (1) saturation experiments from which the affinity of the radioligand for the receptor and the binding site density can be determined; (2) inhibition experiments from which the affinity of a competing, unlabeled compound for the receptor can be determined; and (3) kinetic experiments from which the forward and reverse rate constants for radioligand binding can be determined. Detailed methods for typical radioligand binding assays for G proteincoupled receptors in membranes and intact cells are presented for these types of experiments. Detailed procedures for analysis of the data obtained from these experiments are also given. Key words: Affinity, Assay, Binding, Competition, G protein-coupled receptor, Inhibition, Intact cell, Kinetic, Non-specific binding, Radioligand, Rate constant, Receptor, Radioreceptor, Saturation

1. Introduction The radioligand binding assay is a relatively simple but powerful tool for studying G protein-coupled receptors. It can be used to determine the affinity of numerous drugs for these receptors, and to characterize regulatory changes in receptor number and in subcellular localization. As a result, this assay is widely used (and often misused) by investigators in a variety of disciplines. Our focus in this chapter is on radioligand binding assays in membrane preparations from tissues and cell lines, and in intact cells. Similar techniques, however, can be used to study solubilized receptors, receptors in tissue slices (receptor autoradiography), or receptors in intact animals. There are three basic types of radioligand binding experiments: (1) saturation experiments from which the affinity of the radioligand Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_8, © Springer Science+Business Media, LLC 2011

135

136

Bylund and Toews

for the receptor and the binding site density can be determined; (2) inhibition experiments from which the affinity (Ki) of a competing, unlabeled compound for the receptor can be determined; and (3) kinetic experiments from which the forward (k+1) and reverse (k−1) rate constants for radioligand binding can be determined. This chapter presents methods for typical radioligand binding assays for G protein-coupled receptors. 1.1. Saturation Experiments

Saturation experiments are frequently used to determine the change in receptor density (number of receptors) during development or following some experimental intervention, such as treatment with a drug. A saturation curve is generated by holding the amount of receptor constant and using various concentrations of radioligand. From this type of experiment, the receptor density (maximum binding or Bmax) and the dissociation constant (Kd; inverse of affinity) of the receptor for the radioligand can be estimated. The results of the saturation experiment can be plotted with “Bound” (the amount of radioactive ligand that is bound to the receptor) on the Y-axis and “Free” (the free concentration of radioactive ligand) on the X-axis. As shown in Fig. 1, as the concentration of radioligand increases, the amount bound increases until a point is reached where more radioactive ligand does not significantly increase the amount bound. The resulting graph is a rectangular hyperbola and is called a saturation curve. Bmax is the maximal binding which is approached asymptotically as radioligand concentration is increased. Bmax is the density of the receptor in the tissue being studied. Kd is the concentration of ligand that occupies 50% of the binding sites.

Fig. 1. Typical saturation binding experiment. In this simulation, the Bmax (receptor density) is 10 pM and the Kd (the dissociation constant or the free concentration which gives half maximal binding) is 100 pM.

Radioligand Binding Methods for Membrane Preparations and Intact Cells

137

1.2. Inhibition Experiments

The great utility of inhibition experiments is that the affinity of any (soluble) compound for the receptor can be determined. Thus, these assays are heavily used both for determining the pharmacological characteristics of the receptor and for discovering new drugs using high-throughput screening techniques. In an inhibition experiment, the amount of an inhibitor (nonradioactive) drug included in the incubation is the only variable, and the dissociation constant (Ki) of that drug for the receptor identified by the radioligand is determined. A graph of data from a typical inhibition experiment is shown in Fig. 2. The amount of radioligand bound is plotted versus the concentration of the unlabeled ligand (on a logarithmic scale). The bottom of the curve defines the amount of nonspecific binding. The IC50 value is defined as the concentration of an unlabeled drug required to inhibit specific binding of the radioligand by 50%. The Ki is then calculated from the IC50.

1.3. Kinetic Experiments

Kinetic experiments have two main purposes. The first is to establish an incubation time that is sufficient to ensure that steady state (commonly called equilibrium) has been reached. The second is to determine the forward (k+1) and reverse (k−1) rate constants. The ratio of these constants provides an independent estimate of the Kd (k−1/k+1). If the amounts of receptor and radioligand are held constant and the time varied, then kinetic data are obtained from which forward and reverse rate constants can be estimated. A graph of data from a typical association kinetic experiment is shown in Fig. 3. Initially, the rate of the forward reaction exceeds the rate of the reverse reaction. After a period of time (approximately 25 min in this example), the amount of specific binding no longer increases and thus steady state has been reached. From these data, the k+1 can be calculated.

Fig. 2. Typical inhibition experiment. In this simulation, the specific binding is 900 cpm and the IC50 (the concentration of drug which inhibits 50% of the specific binding) is 10 nM.

138

Bylund and Toews

Fig. 3. Typical association binding experiment. In this simulation, steady state is reached after approximately 25 min and lasts until the end of the experiment (42 min).

Fig.  4. Typical dissociation binding experiment. In this simulation, the t½ (the time at which the specific binding has decreased by 50%) is 5 min.

For a dissociation experiment, the radioligand is first allowed to bind to the receptor and then the dissociation of the radioligand from the receptor is monitored as a function of time by the decrease in specific binding (Fig. 4). The rebinding of the radioligand to the receptor is prevented by the addition of a high concentration of a nonradioactive drug that binds to the receptor and thus blocks the receptor binding site, or by “infinite” dilution which reduces the free concentration of the radioligand. Dissociation follows first order kinetics and thus k−1 is equal to the t½ for dissociation divided by 0.693 (natural logarithm of 2). 1.4. Assays in Intact Cells

Although isolated membranes are by far the most common preparation used for radioligand binding assays, for some purposes it is preferable to use intact cells. The most obvious advantage of

Radioligand Binding Methods for Membrane Preparations and Intact Cells

139

assays with intact cells is that the receptor is being studied in its native environment in the cell. A related advantage of intact cell assays is that the binding properties of the receptor can be assessed in the same preparation and under essentially the same conditions as the functional responses mediated by the receptor are measured. This allows a more direct comparison of the receptor binding properties with a wide variety of physiological responses following activation or inhibition of the receptor. Intact cell assays may also be advantageous when many different cell samples need to be studied, because intact cell assays eliminate the need to disrupt cells and isolate membranes prior to assay. For example, intact cell assays have proven very useful for preliminary screening of cell colonies following transfection with cDNA for various G protein-coupled receptors, thus allowing rapid identification of clones for amplification and further analysis. Most of the considerations that make intact cell assays advantageous in certain cases also represent limitations of intact cell assays in other cases. For example, intact cell assays allow studies under physiological conditions, but they make it much more difficult to vary or control the assay conditions to identify factors that modulate receptor binding. Radioligand uptake into cells by various transport processes can occur with intact cells, and care must be taken to ensure that radioligand association with intact cells is due to binding rather than uptake. The occurrence of adaptive regulatory changes in receptor number, localization, and binding properties during the course of binding assays with intact cells can also present a serious complication (1). Finally, intact cells have membrane permeability barriers that are not present in isolated membrane preparations, and therefore the lipid solubility and membrane permeability of both the radioligand and the competing ligands must be considered in assays with intact cells. Lipophilic (“lipid-loving”) ligands generally cross all cell membranes easily and thus have access to both cell-surface receptors and those in intracellular compartments, such as endosomes. On the contrary, hydrophilic (“water-loving”) ligands are relatively impermeable to the plasma membrane, and thus these ligands only label cell-surface receptors. Although these properties can complicate assays with intact cells, they also provide the basis for important radioligand binding-based assays for receptor internalization, as discussed previously (1).

2. Materials The information given in this section is specifically for assays with membrane preparations. The additional information for intact cells assays is given in Subheading 3.4.

140

Bylund and Toews

1. A radioligand appropriate for the receptor being studied (see Note 1). For membrane saturation experiments, add the appropriate volume of radioligand into 550 mL of 5 mM HCl in a glass test tube. Thoroughly mix and add 200 mL of this solution to 300  mL of 5  mM HCl. Prepare successive dilutions in the same manner by adding 200 mL of each dilution to 300 mL of 5 mM HCl to obtain the next lower dilution until six concentrations of radioligand have been prepared. This dilution strategy gives a 100-fold range of radioligand concentrations. Other dilution strategies will give different ranges as indicated in Table 1 (see Note 2). For membrane inhibition and kinetic experiments, only a single concentration of radioligand is needed (see Note 3). 2. A source of receptor, either crude cell lysates or isolate membranes fractions. The standard procedure for a membrane assay is to homogenize the tissue or cells of interest in a hypotonic buffer using either a Polytron (Brinkman) or similar homogenizer. Remarkably, most receptors are stable at room temperature (generally for hours), although it is wise to put the tissue on ice quickly. Homogenize about 500 mg of tissue in approximately 35 mL of wash buffer (50 mM Tris/HCl or similar buffer at pH between 7 and 8) using a Polytron (PT10-35 generator with PT10/TS probe; see Daigger Lab Equipment and Supplies at http://www.daigger.com/) at setting 7 for 20 s (see Note 4). The actual weight of tissue used should be recorded. Centrifuge at 48,000 × g in a Sorvall

Table 1 Dilution of radioligand for saturation experiments Radioligand (mL)

250

200

150

Diluent (mL)

250

300

350

Dilution number

Relative concentration

1

100

100

100

2

50

40

30

3

25

16

4

12.5

6.4

2.7

5

6.2

2.6

0.81

6

3.1

1.0

0.24

7

1.6

0.41

8

0.78

9

0.39

9.0

Radioligand Binding Methods for Membrane Preparations and Intact Cells

141

RC-5B using an SS34 rotor (or similar centrifuge and rotor) for 10 min at 4°C (see Note 5). Decant the supernatant, and repeat the homogenization and centrifugation. The tissue preparation can either be used immediately or stored frozen as a pellet until needed (see Note 6). Generally, protease inhibitors are not needed but could be important in certain tissues (such as liver) or with certain receptors. 3. Membrane assay buffer, 25 mM at pH 7.4 (such as sodium phosphate, or Tris/HCl). For a few receptors the choice of buffer is important, but for most it is not. 4. Wash buffer, such as 25 mM Tris/HCl, pH 7.4. Almost any buffer at neutral pH will often do. 5. 5  mM HCl for diluting labeled and unlabeled ligands. For many ligands, using a slightly acidic diluent will increase stability and decrease binding to dilution/assay tubes. 6. Appropriate unlabeled ligands in solution. 7. 0.1 M NaOH for samples to be used to assay protein. 8. Polypropylene test tubes, 12 × 75 mm (assay tubes). 9. Borosilicate glass test tubes, 12 × 75 mm (dilution tubes). 10. Glass fiber filters (GF/A circles and GF/B strips). 11. Filtration manifold. 12. Scintillation vials if using a 3H-radioligand; or test tubes if using a 125I-radioligand. 13. Scintillation cocktail (if using a 3H-radioligand).

3. Methods 3.1. Saturation Experiment (Membrane Assay)

1. Resuspend washed membrane preparation in distilled water by homogenization. 2. Add three 20 mL aliquots of the tissue suspension to 80 mL of 0.1 M NaOH for estimating protein concentration. 3. Add sufficient ice-cold assay buffer to the membrane suspension to give the appropriate final concentration (see Note 7). 4. Set up a rack of 24 polypropylene incubation tubes, six tubes across and four tubes deep. If using a 125I-ligand, add two additional test tubes to each of the six sets of tubes (for the determination of total added radioactivity). If using a 3 H-ligand, prepare a set of 12 uncapped scintillation vials with GF/A glass fiber filter discs (see Note 8). 5. To the 12 tubes on the last two rows, add 10 mL of a high concentration of an unlabeled ligand to determine nonspecific binding (see Note 9).

142

Bylund and Toews

6. To all 24 tubes, add 970 mL of the membrane preparation. Because this is a particulate suspension, it should be slowly stirred while aliquots are being removed. 7. Starting with the most dilute radioligand solution, add 20 mL to the columns of four tubes, and mix each tube. Also add 20 mL of the radioligand solution to the two filter papers on the scintillation vials (if using a 3H-radioligand) or two test tubes (if using a 125I-ligand) for the determination of total added radioactivity. 8. Mix all the tubes again and incubate (usually at room temperature) for 45 min. Assuming that the system is at steady state, the exact time is not critical. Although generally not necessary, it may be useful to shake the tubes several times during the incubation for some systems. The tubes may need to be rearranged to be compatible with the specific style of filtration manifold used. 9. Filter the contents of the tubes and wash the filters twice with 5 mL of wash buffer. Depending on the rate of dissociation of the radioligand from the receptor, it may be important to use ice-cold wash buffer. 10. Place the filters into scintillation vials, add 5 mL of scintillation cocktail and cap if using a 3H-radioligand, or into test tubes if a using 125I-ligand. 11. Shake the scintillation vials gently for 60 min (or let it stand at room temperature overnight) and then count in a liquid scintillation counter (if using a 3H-radioligand); or count in a g-counter (if using a 125I-ligand) (see Note 10). 3.1.1. Calculation of Results from a Saturation Experiment

Data from a sample saturation experiment are shown in Table 2. 1. Total binding and nonspecific binding can be plotted versus total added as shown in Fig.  5 for the sample experiment. This plot allows one to detect data points that may be problematic. Note that the nonspecific binding is linear (except possibly at the lowest concentrations), and that the specific binding saturates (is relatively constant) at high radioligand concentrations. 2. Specific binding is determined by subtracting nonspecific binding from total binding at each concentration of radioligand (see Table 2). 3. The cpm values are converted to pM values using a conversion factor that accounts for specific activity for the radioligand, the counting efficiency of the particular scintillation counter used, and the conversion factor 2.2 × 1012 dpm/Ci. For this experiment, the counting efficiency was 0.36 and the specific radioactivity of the radioligand was 60 Ci/mmol, and

Radioligand Binding Methods for Membrane Preparations and Intact Cells

143

Table 2 Results of a sample saturation experimenta in cpm Total added (cpm)

Total bound (cpm)

Nonspecific bound (cpm)

Specific bound (cpm)b

2,360

208

18

190

5,601

394

25

369

14,491

597

46

551

32,011

782

88

694

82,520

984

189

795

192,248

1,210

416

794

[ H]RX821002 binding to human a2A-adrenergic receptors in HT-29 cells  The amount of radioligand specifically bound was determined by subtracting the nonspecific bound from the total bound

a  3 b

Fig. 5. Total binding and nonspecific binding versus total added for a sample experiment from Table 2.

the factor for converting cpm to dpm is 0.0210 as shown below: cpm dpm Ci mmol 1, 000 mL 1 mol ´ ´ ´ ´ ´ ml 0.36 cpm 2.2 ´ 1012 dpm 60Ci L 1, 000 mmol ´

1012 pmol 2.10 ´ 10 -02 pmol = mol L

The results of this conversion for the sample experiment are shown in Table 3. 4. Free concentration of radioligand is calculated by subtracting specific bound from total added as shown in Table 3.

144

Bylund and Toews

Table 3 Results of example saturation experiment given in Table 2 converted to pM units Total bound (pM)

Specific bound (pM)

Free (pM)

Bound/free

  4.37

  3.99

46

0.0867

  8.27

  7.74

110

0.0704

12.5

11.5

292

0.0394

16.4

14.5

658

0.0220

20.7

16.7

1,720

0.00971

25.4

16.7

4,020

0.00415

Fig. 6. Saturation curve for data from an example saturation experiment. Specific Bound (pM units) from Table 3 is plotted versus the Free (pM) concentration of the radioligand. The line was drawn using nonlinear regression analysis for one-site binding using the Prism computer program (GrapPad, San Diego, CA).

5. The data are then plotted as Bound versus Free as shown in Fig.  6 for the typical saturation experiment (see Note 11). The Kd and Bmax values are generally calculated by nonlinear regression of the specific binding versus the concentration of radioligand using a computer program such as Prism (GraphPad, San Diego, CA) or a variety of other software packages using the following equation: B=

Bmax ´ F Kd + F

Radioligand Binding Methods for Membrane Preparations and Intact Cells

145

where B is the amount of radioligand specifically bound, F is the free radioligand concentration, Bmax is the radioligand concentration required to saturate all of the binding sites, and Kd is the dissociation constant for the radioligand at these receptors. 6. The Bmax values are dependent on the concentration of protein in the assay. Note that the results are given in pM units. To convert the Bmax values to pmol/mg of protein, the pM values are converted to pmol/mL and divided by the protein concentration (mg/mL) used in the assay. In this example, the protein concentration is 0.072  mg/assay tube and the Bmax value is calculated as shown below:



17.7 pmol 1L 1mL 0.246 pmol ´ ´ = L 1, 000 mL 0.072mg mg 7. In order to better visualize the results and to detect potential problems, the data are frequently transformed (as shown in Table 3) and viewed as a Rosenthal plot (2) in the form of Bound/Free versus Bound as shown in Fig. 7 (see Note 12). The equation for the line is as follows:



B Bound 1 =Bound + max Free Kd Kd In this plot, the intercept with the x-axis (abscissa) is the Bmax and the Kd is the negative reciprocal of the slope (see Note 13). The data points fall close to a straight line indicating a single class of binding sites.

Fig.  7. Rosenthal plot of the saturation binding experiment. The Bound/Free versus Bound data from Table 3 are plotted to obtain the Rosenthal plot.

146

Bylund and Toews

3.2. Inhibition Experiment

1. Prepare a 1-mM solution of the inhibitor(s) in 5 mM HCl (or other diluent as appropriate). Dilute 0.3 mL of this solution with 0.7 mL of 5 mM HCl to give a 0.3-mM solution. Prepare 100 mM, 10 mM, 1 mM, 100 nM, 10 nM, 1 nM, and 0.1 nM solutions by sequentially diluting 100  mL of the previous solution (i.e., tenfold higher concentration) with 900 mL of 5 mM HCl. Similarly, prepare 30 mM, 3 mM, 300 nM, and 30 nM solutions from the 0.3-mM solution (see Note 14). 2. If using a 3H-ligand, prepare two uncapped scintillation vials with GF/A glass fiber filter discs (see Note 8). 3. Set up 24 assay tubes in two rows of 12. Add 10 mL of 5 mM HCl (or other diluent) to the first pair of tubes. Add 10 mL of the appropriate dilution (concentration) of the inhibitor to the other pairs of tubes, starting with the lowest concentration (see Note 15). 4. Add 970  mL of the membrane preparation to each of the 24 tubes. 5. Add 20  mL of radioligand to each of the tubes and mix to start the incubation. Pipette 20 mL of the radioligand solution directly onto duplicate GF/A glass fiber filter discs (if using a 3 H-radioligand) or into two test tubes (if using a 125I-ligand) to determine the total added radioactivity. 6. Mix all of the tubes again and incubate (usually at room temperature) for 45 min. Assuming that the system is at steady state, the exact time is not critical. 7. Filter the contents of the tubes and wash the filters twice with 5 mL of ice-cold wash buffer. The tubes may need to be rearranged to be compatible with the specific style of filtration manifold used. 8. Put the filters into scintillation vials, add 5 mL of scintillation cocktail and cap (if using a 3H-radioligand), or into test tubes (if using a 125I-ligand). 9. Shake the scintillation vials gently for 60 min (or let stand at room temperature overnight) and then count in a liquid scintillation counter (if using a 3H-radioligand), or count in a g-counter (if using a 125I-ligand) (see Note 10).

3.2.1. Calculation of Results from an Inhibition Experiment



The calculation of Ki values from inhibition experiments is relatively straightforward. The inhibition data are simply fit to a sigmoidal curve with the logarithm of concentration of the inhibitor on the abscissa, and the IC50 value (the concentration of the inhibitor that inhibits 50% of the specific binding) are determined using a Hill slope of 1. Y = Bottom +

Top - Bottom 1 + 10(X - Log IC50 )(Hill slope)

Radioligand Binding Methods for Membrane Preparations and Intact Cells

147

Top and bottom refer to the concentration of bound radioligand at the top and bottom of the curve. Y is the amount of radioactive ligand bound at each concentration of inhibitor X. The Ki value is calculated from the IC50 value using the following equation: Ki =



IC50 1 + (F / K d )

where F is the free radioligand concentration and Kd is the affinity of the radioligand. This is often called the Cheng–Prusoff correction (2) (see Note 16). Thus, if the radioligand is present at its Kd concentration, then the Ki is half of the IC50. If more than one binding site is suspected, the Hill slope can be allowed to vary or the equation for two-site binding can be used.

Y =

Span ´ Fraction1 Fraction ´(1 - Fraction1) + 1 + 10X - Log IC501 1 + 10X - Log IC502

In this equation, Span refers to the difference between the top and bottom of the curve, Fraction 1 is the amount of radioligand bound to the high affinity site, and log IC501 and IC502 refer to the inhibition of binding to high and low affinity sites, respectively. An F-test can be used to determine whether the data better fit a one-site or two-site model (see Box 1). These analyses are illustrated with the data from a sample experiment given in Table 4. 1. The data are plotted as Bound (cpm or pM units) versus logarithm of the inhibitor concentration as shown for the sample experiment in Fig. 8. The bottom of the curve plateaus at the same bound value as obtained with norepinephrine, indicating that prazosin is likely binding to the same sites as norepinephrine. 2. The solid curve was obtained using a nonlinear regression analysis of a one-site competition equation using the Prism computer program (GrapPad, San Diego, CA). The data were also fit to a sigmoid curve using nonlinear regression analysis with a variable Hill slope (nH) as is indicated by the dashed line. As is shown in Table 5, the results of the two analyses are essentially identical because the nH is not different from unity. As a rough approximation, nH values need to be less than 0.8 to be significantly different from 1.0 and suggest more complex binding. 3. Similarly, the data can also be fit to a two-site competition equation. The results of this fit are shown in Fig. 8 and Table 5 (see Note 17). 4. An F-test is used to determine whether the data fit a one- or two-site equation better. The F-test for the sample experiment

148

Bylund and Toews

Box 1 F -test for comparison of fit of data from sample competition experiment to a one- versus two-site fit The F-test is used to compare the one-site and the two-site model. The basic steps are 1. Analyze the data for a one- and two-site fit using non-linear regression analysis. 2. Apply the sum of the squares and degrees of freedom to the equation for a F-test: F =

(SS1 - SS2) / (DF1 - DF2) SS2 / DF2

where SS1 = sum of squares for one-site fit, SS2 = sum of squares for two-site fit, DF1 = degrees of freedom for one-site fit, DF2 = degrees of freedom for two-site fit. 3. Determine the P-value from a F-table of statistics. 4. The P value answers the question: If model 1 (one-site fit) is correct, what is the chance that you would randomly obtain data that fits model 2 (two-site fit) much better? 5. If P is low, you conclude that model 2 (two-site fit) is significantly better than model 1. 6. The calculation for the sample saturation experiment is as follows: (25, 467 - 14, 394) / (8 - 6) 14, 394 / 6 F = 2.308 P = 0.1806 F =

Because P > 0.05, the two-site model does not give a significantly better fit and the one-site model is accepted.

is shown in Box 1. The two-site fit was not significantly better than the one-site fit; thus, the curve for the one-site fit was chosen and the IC50 values from the one-site fit were used to calculate the Ki values. 5. The Ki is calculated using the Cheng–Prusoff equation (2). For the sample experiment, the concentration of radioligand was 0.75 nM, the Kd was 0.89 nM, and the log of IC50 was −8.04 (9.08  nM). Putting these numbers into the Cheng– Prusoff equation gives a Ki of 4.9 nM.



Ki =

IC50 9.08nM = = 4.9nM 1 + (F / K d ) 1 + (0.75 / 0.89nM)

Radioligand Binding Methods for Membrane Preparations and Intact Cells

149

Table 4 Data from a sample inhibition experiment a Prazosin concentration (nM)

−Log [prazosin] (M)

Average bound (cpm)

Specific bound b (cpm)

% Specific boundc

(0)d

−12

1,395

1,314

100

0.3

−9.52

1,292

1,211

92.2

1

−9

1,256

1,175

89.4

3

−8.52

958

877

66.7

10

−8

730

649

19.4

30

−7.52

417

336

25.6

100

−7

198

117

8.90

300

−6.52

126

45

3.43

1,000

−6

116

35

2.67

3,000

−5.52

82

1

0.076

10,000

−5

81

0

0

NE

92

e

Prazosin inhibition of [ H]RX821002 binding to a2B-adrenergic receptors transfected into CHO cells with Kd = 0.89 nM. The concentration of radioligand used in all the tubes was 0.75 nM b The amount of radioligand specifically bound was also determined by subtracting the amount bound in the presence of the highest concentration of prazosin c The data were normalized by dividing specific bound by amount of radioligand bound in the absence of prazosin d Although the first inhibitor concentration is zero, in order to run the nonlinear regression program a number needs to be used. Routinely a concentration that is at least one log unit lower than the lowest unlabeled drug concentration is used e Norepinephrine (NE) was used at a concentration of 0.3 mM to determine the extent of nonspecific binding in this experiment a

3

6. Frequently, the total binding for different receptor preparations or different unlabeled ligands is different, so inhibition curves are often normalized so that the percent inhibition can be more easily compared as shown in Fig. 9. 3.3. Kinetic Experiments

1. If using a 3H-ligand, prepare two uncapped scintillation vials with GF/A glass fiber filter discs (see Note 8).

3.3.1. Dissociation Experiment

2. Set up 48 assay tubes in four rows of 12. To the 12 tubes on the last two rows, add 10 mL of a high concentration of an unlabeled ligand to determine nonspecific binding (see Note 9). 3. Add 970 mL of the membrane preparation to each of the 48 tubes. 4. Add 20  mL of radioligand to all tubes and mix to start the incubation. The reaction is allowed to proceed until steadystate conditions are reached (45  min). At appropriate time

150

Bylund and Toews

Fig. 8. Plot of inhibition binding data for an example experiment. The data from Table 4 are plotted as Bound (cpm) versus the prazosin concentration (in log-molar units). The solid curve was obtained using a one-site model. Determination of the IC50 is based on the middle of the curve (725 cpm bound) and not half of the binding in the absence of prazosin (697 cpm bound). The dashed curve is from both a two-site analysis and an analysis of a sigmoid equation with a variable Hill slope (the curves are essentially identical for these data).

Table 5 Results of analysis of the sample inhibition experiment presented in Table 4 Parameter

Single-site fit

Variable Hill slope fit

Two-site fit

Bottom of curve

93 cpm

74 cpm

80 cpm

Top of curve

1,356 cpm

1,390 cpm

1,389 cpm

Log IC50

−8.04

−8.06

−8.62

IC50

9.08 nM

8.72 nM

2.39 nM

Log IC50 second site

−7.76

IC50 second site

17.2 nM

Fraction – second site

0.64

The curves for these data are shown in Fig. 8

intervals, add a high concentration (50 times the IC50) of unlabeled ligand to tubes 2–12 in each row to start the dissociation reaction. The first tube in each row is to determine binding at zero time at the start of the dissociation reaction. All incubations will be terminated at the same time by filtration. Thus, the unlabeled ligand is added at various times, for example, 1, 2, 4, 7, 10, 15, 20, 25, 30, 40, and 60 before

Radioligand Binding Methods for Membrane Preparations and Intact Cells

151

Fig. 9. Plot of normalized data for the example inhibition binding experiment. The normalized data from Table 4 are plotted as a function of the log of the inhibitor concentration. This plot is more useful when multiple datasets are being compared.

filtration. Also pipette 20  mL of the radioligand solution directly onto duplicate GF/A glass fiber filter discs (if using a 3 H-radioligand) or two test tubes (if using a 125I-ligand) to estimate the total added radioactivity (see Note 18). 5. Filter the contents of the tubes and wash the filters twice with 5 mL of ice-cold wash buffer. The tubes may need to be rearranged to be compatible with the specific style of filtration manifold used. 6. Put the filters into scintillation vials, add 5 mL of scintillation cocktail and cap (if using a 3H-radioligand), or into test tubes (if a using 125I-ligand). 7. Shake the scintillation vials gently for 60 min (or let stand at room temperature overnight) and then count in a liquid scintillation counter (if using a 3H-radioligand), or count in a g-counter (if using a 125I-ligand) (see Note 10). 3.3.2. Calculation of Results from a Dissociation Experiment

Data from a sample experiment are given in Table 6. Note that in a dissociation experiment, time zero is the time at which the unlabeled ligand is added to the assay tube. The time course for the dissociation experiment then becomes the time between when the unlabeled ligand is added to the assay tube and the time when the samples are filtered. 1. Nonspecific binding is subtracted from total binding at each time point. Specific binding from a sample experiment is presented in Table 6. 2. The data are plotted as bound versus dissociation time. The data from the sample experiment are plotted in Fig. 10.

152

Bylund and Toews

Table 6 Data from an example dissociation experiment a Time (min)

Specific bound (cpm)

0

577

1

490

2

460

3

400

4

360

6

280

8

215

10

165

12

140

 [ H]DHA was incubated with guinea pig cerebral cortex membranes for 20  min to label b-adrenergic receptors. Isoproterenol (2 mM) was added at time 0. Samples were filtered at the time indicated a 3

Fig. 10. Plot of data from the example dissociation kinetics experiment given in Table 6. The line was drawn using the nonlinear regression equation for one-phase exponential decay using the Prism computer program (GrapPad, San Diego, CA).

3. The data are analyzed with a nonlinear regression program using the equation for exponential decay.

Y = Span ´ e - k ´ X + Nonspecific binding In a dissociation experiment, Span refers to specific binding and k is k−1. Analysis of the data for the sample experiment using a one-site exponential decay analysis and the Prism computer program (GrapPad, San Diego, CA) gives a k−1 of 0.117 min−1.

Radioligand Binding Methods for Membrane Preparations and Intact Cells 3.3.3. Association Experiments

153

1. If using a 3H-ligand, prepare two uncapped scintillation vials with GF/A glass fiber filter discs (see Note 8). 2. Set up 48 assay tubes in four rows of 12. To the 12 tubes on the last two rows, add 10 mL of a high concentration of an unlabeled ligand to determine nonspecific binding (see Note 9). 3. Add 970 mL of the membrane preparation to each of the 48 tubes. 4. At appropriate time intervals, add 20 mL radioligand to all of the tubes and mix. All incubations will be terminated at the same time by filtration. Thus, the radioactivity is added at, for example, 1, 2, 4, 7, 10, 15, 20, 25, 30, 40, 50, and 60 min before filtration. Also pipette 20 mL of the radioligand solution directly onto duplicate GF/A glass fiber filter discs (if using a 3H-radioligand) or two test tubes (if using a 125I-ligand) to estimate the total added radioactivity. 5. Filter the contents of the tubes and wash the filters twice with 5 mL of ice-cold wash buffer. The tubes may need to be rearranged to be compatible with the specific style of filtration manifold used. 6. Put the filters into scintillation vials, add 5 mL of scintillation cocktail and cap (if using a 3H-radioligand), or into test tubes (if using a 125I-ligand). 7. Shake the scintillation vials gently for 60 min (or let stand at room temperature overnight) and then count in a liquid scintillation counter (if using a 3H-radioligand), or count in a g-counter (if using a 125I-ligand) (see Note 10).

3.3.4. Calculation of Results from an Association Experiment

1. Nonspecific binding is subtracted from total binding at each time point to give specific binding. The data from a sample experiment are given in Table 7. 2. Amount bound is plotted versus time as shown in Fig. 11 for the sample association experiment. 3. The data are analyzed using a nonlinear regression analysis of the equation for an exponential association curve.



Y = Ymax (1 - e - kobt ) The rate constant (k observed; kob) obtained is a combination of k1 and k−1 and will vary with the concentration of radioligand (F) added to the assay according to the following equation:



kob = k1F + k-1 k−1 can be determined using a separate dissociation experiment as described above. k1 can then be determined using the following rearrangement of the above equation:



k1 =

kob - k-1 F

154

Bylund and Toews

Table 7 Data from an example association experiment a Time (min)

Specific bound (cpm)

  1.5

   0.0

  0.0

170

  2.5

285

  4.0

380

  6.0

475

  8.0

560

10.0

610

13.0

655

16.0

680

20.0

710

a Radioligand binding of [3H]DHA (0.36 nM) to guinea pig cerebral cortex

Fig. 11. Plot of data from the example association kinetics experiment given in Table 7. The line was drawn using the nonlinear regression equation for exponential association using the Prism computer program (GraphPad, San Diego, CA).

Nonlinear regression analysis of the sample association data (Table  7, Fig.  11) gave a kob = 0.187  min−1 at a radioligand concentration of 0.36  nM. The dissociation rate constant determined from the experiment described above was 0.117  min−1. The association rate constant for the sample experiment thus becomes



k1 =

0.187 - 0.117 min -1 = 0.194 min -1 nM -1 0.36 nM

Radioligand Binding Methods for Membrane Preparations and Intact Cells

155

An alternate way to determine k1 is to do the association experiment at various concentrations of radioactive ligand. The kob determined from these association experiments can then be plotted versus the concentration of radioligand (F). The y-­intercept is the k−1 and the slope of the line is k1. 4. Kd is determined by dividing k−1/k1. For the sample experiment,

3.4. Sample Protocol for a Saturation Experiment with Monolayer Cells

Kd =

0.117 min -1 = 0.60 nM 0.194 min -1 nM -1

Assays with intact cells in suspension are quite similar to assays with membranes. The manipulations required for assays with monolayer cells are more involved, however, and thus a detailed protocol for monolayer cells is presented below (see Note 19). The protocol described is for an eight-point saturation experiment with triplicate determinations for both total and nonspecific binding. The protocol assays the cells in sets of six (three total binding and three nonspecific binding) because this is a convenient number to manipulate within a 1-min time frame (one every 10 s). Accordingly, the cells are plated on six-well plates, and the plates are treated at 5-min intervals. 1. Grow cells to near confluence on eight 6-well plates in 2 mL growth medium per well (see Note 20). 2. Prepare 7.5 mL of HEPES-buffered serum-free growth medium (see Note 21) containing each of the eight concentrations of radioligand, labeled as “A” through “H.” Transfer 3.5 mL of these solutions to each of two polypropylene tubes with a large enough diameter to allow easy use of a 1-mL pipettor tip (see Note 22). To one of each pair of tubes (for total binding), add 35 mL of the vehicle for the agent used to define nonspecific binding and label as “AT” through “HT.” To the other (for nonspecific binding), add 35 mL of 100× concentrated solution of the agent used to define nonspecific binding and label as “AN” through “HN.” Place these tubes in a 37°C water bath to reach physiological temperature (see Note 23). 3. Place a beaker with approximately 300  mL of HEPESbuffered serum-free growth medium in the 37°C water bath as well, to be used as “pre-incubation wash medium.” This same medium can be used as “post-incubation wash buffer,” though in some cases it is beneficial for the post-incubation wash buffer to contain a drug to reduce nonspecific binding (see Note 23). 4. Prepare a Pasteur pipette connected by Tygon tubing to a vacuum flask connected to a vacuum pump or vacuum line to use for aspirating medium from dishes.

156

Bylund and Toews

5. A detailed time-course for a 60-min assay in which a single investigator can conduct the entire assay is presented below, with times presented in minutes. All solution additions are done with a Pipetman or similar hand-held adjustable pipettor. These solutions should be gently added against the inside wall of the dish, not directly onto the monolayers, to avoid loss of cells from the dish due to the multiple medium changes. Initiation of binding: t = 0 min:

At 10-s intervals, aspirate the growth medium and add 2 mL 37°C pre-incubation wash buffer to the 6 wells of the first plate. This step is to wash away serum and bicarbonate and switch the cells to the medium used as assay buffer.

t = 1 min:

At 10-s intervals, aspirate the wash buffer from the plate and add 1 mL AT solution to the top three wells and 1 mL AN solution to the bottom three wells. By starting the T wells before the N wells, it is not necessary to change pipette tips between the total and nonspecific binding solutions. However, the tip must be changed before the next concentration of total binding solution is added to the next set of dishes.

t = 2 min:

Transfer the plate to a 37°C non-CO2 incubator for the 60-min binding time.

t = 5, 6, and 7 min:

Repeat the above steps for the second plate, using the BT and BN solutions.

t = 10, 11, 12, 15, 16, 17 min, etc.:

Repeat the above steps for each of the remaining concentrations of radioligand.

Termination of binding: t = 60 min:

At 10-s intervals, aspirate the binding medium and add 2 mL 37°C post-incubation wash buffer (see Note 24) to the 3 AT wells and then the 3 AN wells. This step is to stop the binding reaction and wash away unbound radioligand.

t = 61 min:

Repeat the step above for a second wash of the first plate.

t = 62 min:

At 10-s intervals, aspirate the wash buffer and add 1 mL 0.2 M NaOH to each of the 6 wells. Set the plate aside (see Note 25).

t = 65, 66, 67 min:

Repeat the above steps for the BT and BN plate.

t = 70, 71, 72, 75, 76, 77 min, etc.:

Repeat the above steps for each of the remaining six plates.

Radioligand Binding Methods for Membrane Preparations and Intact Cells

157

Transfer and quantification of bound radioactivity: after the binding and wash steps are completed for all eight plates, transfer the dissolved cells and the associated radioactivity to vials for scintillation counting or to tubes for g-counting, depending on the radioligand used (see Note 26). 6. Only minimal modifications to this protocol are needed for competition rather than saturation assays. 7. Kinetic assays of association and dissociation become somewhat more complicated with intact cells grown in monolayer culture. Because each well must be started and stopped individually, careful planning of the time points is required. In general, the longer time points are started first and stopped last in order to complete the entire experiment in the shortest possible time. An example of sample timing that allows a single investigator to conduct a time course experiment with time points at 2, 5, 10, 20, 40, and 60 min is presented below. t = 0, 1, and 2 min:

Start the binding reaction for the 60-min plate.

t = 5, 6, and 7 min:

Start the binding reaction for the 40-min plate.

t = 10, 11, and 12 min: Start the binding reaction for the 20-min plate. t = 15, 16, and 17 min: Start the binding reaction for the 10-min plate. t = 25, 26, and 27 min: Stop the reaction for the 10-min plate. t = 30, 31, and 32 min: Stop the reaction for the 20-min plate. t = 35, 36, and 37 min: Start the binding reaction for the 5-min plate. t = 40, 41, and 42 min: Stop the reaction for the 5-min plate. t = 45, 46, and 47 min: Stop the reaction for the 40-min plate. t = 50 and 51 min:

Start the reaction for the 2-min plate; this plate will be stopped immediately

t = 52, 53, and 54 min: Stop the reaction for the 2-min plate. t = 65, 66, and 67 min: Stop the reaction for the 60-min plate.

4. Notes 1. The decision on which radioligand to use is based both on the characteristics of the radioligand and on the specific scientific questions being asked. The important characteristics to be considered include the radioisotope (3H or 125I), the extent of nonspecific binding, the selectivity and affinity of the radioligand for the receptor, and whether the radioligand is an agonist or an antagonist.

158

Bylund and Toews

The advantages of 3H over 125I as a radioisotope include that the radioligand is chemically unaltered and thus biologically indistinguishable from the unlabeled compound, and that it has a longer half-life (12 years versus 60 days). Because of their short half-lives, iodinated radioligands are usually purchased or prepared every 4–6 weeks, whereas 3H-ligands can often be used for several months or even longer. An advantage of the iodinated radioligands is their higher specific activity (up to 2,200 Ci/mmol versus 30–100 Ci/mmol for 3 H-ligands), which makes them particularly useful if the density of receptors is low, or if the amount of tissue is small. It is easier and less expensive to use iodinated ligands, because scintillation cocktail is not needed, thus eliminating purchasing and disposal costs associated with scintillation cocktail. Each radioligand has a unique pharmacological profile. The radioligand should bind selectively to the receptor type or subtypes of interest under the assay conditions used. Although no radioligand is completely selective for any given receptor or receptor subtype, some are better than others. If, for example, several subtypes of a receptor are present in a given tissue, and if the intent is to label all the subtypes, then a subtype nonselective radioligand that has nearly equal affinity for all of the subtypes would be chosen. On the contrary, if only a single subtype is of primary interest, then a radioligand having much higher affinity for that particular subtype, as compared to the other subtypes, would be the preferred radioligand. Usually, the higher the affinity the better, as a lower concentration of the radioligand can be used in the assay, which results in lower levels of nonspecific binding. Furthermore, a higher affinity usually means a slower rate of dissociation, which provides for a more convenient assay. Agonist radioligands may label only a portion of the total receptor population (the high affinity state for G protein-­ coupled receptors), whereas antagonist ligands generally label all receptors. On the other hand, an agonist radioligand may more accurately reflect receptor alterations of biological significance, because it is agonists that activate the receptor. Nonspecific binding is binding to sites other than the receptors of interest. This can include other receptors, nonreceptor binding sites in the tissue (e.g., to the lipid bilayer), and binding to the filter paper. Usually, the radioligand with the lowest nonspecific binding is best. An assay is considered barely adequate if 50% of the total binding is specific, 70% is good, and 90% is excellent. Most radioligands are stored in an aqueous solution, which often contains an organic solvent such as ethanol. These solutions should be stored cold, but not frozen, because

Radioligand Binding Methods for Membrane Preparations and Intact Cells

159

freezing of the solution tends to concentrate locally the radioligand and may increase its radiolytic destruction. 2. At least six concentrations of radioligand should be used with an equal number of concentrations above and below the anticipated Kd value. Thus, the amount of stock radioligand used will depend on the Kd of the radioligand for the receptor type and subtype assayed. The amount of radioligand prepared in this manner is sufficient for three saturation experiments. We routinely use 5 mM HCl to dilute the radioactivity, because we have found that it reduces the nonspecific binding of many radioligands to the dilution tubes and helps to ensure ligand stability. This may not be necessary for all radioligands. The small amount (20 mL) of dilute HCl (5 mM) carried over to the 1.0-mL assay in 25 mM buffer should not alter the pH of the assay (but this should be checked). Because the amount of specific binding approaches the Bmax asymptotically, the specific binding will never actually reach the Bmax and thus true saturation will never be achieved. Furthermore, the use of very high radioligand concentrations is usually limited by the associated high level of nonspecific binding and by prohibitive cost. In practice, for an assay that conforms to a single site, it is sufficient if the highest concentration gives specific binding of ³90% of the Bmax and if the Rosenthal plot describes a straight line. 3. The radioligand concentration used in inhibition experiments should ideally be less than its Kd value. For kinetic experiments to establish steady state, the lowest practical concentration should be used. For experiments to calculate Kd from k+1 and k−1, a concentration near the Kd usually works well. 4. Fibrous tissues, such as the lung, should be filtered through a nylon mesh (about 50 mm). A low-speed centrifugation step (500 × g) is also helpful with some tissues to remove unwanted pieces of tissue. 5. This centrifugation step is generally done at the highest speed possible (without using an ultracentrifuge) for 5–10  min. Centrifugations are routinely carried out at 4°C, although for many receptors this may not be necessary. 6. This tissue preparation is variously called a crude particulate fraction or a membrane fraction. The purpose of the two centrifugation steps is to remove any soluble substances, such as endogenous neurotransmitters and guanine nucleotides, which may interfere with the radioligand binding assay. The choice of buffer for the homogenization is generally not critical; any buffer at neutral pH appears to be sufficient for the majority of receptor preparations. For some receptor assays, it is recommended that EDTA be added to the homogenization

160

Bylund and Toews

buffer and/or an additional incubation (20 min at 37°C) be included after the second homogenization (but before the second centrifugation) in order to more completely remove various endogenous substances. Most receptor preparations are stable to freezing and can be stored at −20°C or −80°C for extended periods of time, either as the original tissue, or as a pellet after the first homogenization/centrifugation. Experience indicates that some receptors and small pieces of tissue (10 mg) do not store well and thus in these particular cases the assay should be run on fresh tissue. 7. A membrane concentration in the range of 2–10 mg original wet weight of tissue per mL (i.e., a dilution of 500–100 volumes) is usually appropriate. This gives a concentration of about 100–500 mg of membrane protein per mL. For transfected cells that over-express the receptor, the protein concentration in the assay will be lower. In general, the higher the membrane (receptor) concentration, the better. Increasing the receptor concentration generally increases the ratio of specific to nonspecific binding, since a large portion of the nonspecific binding is binding to the glass fiber filter. A rule of thumb is that if more than 10% of the added radioligand is bound, then the tissue concentration is too high. 8. Aliquots of the diluted 3H-ligand will be added directly onto the filter paper to determine the total added radioactivity. GF/B filters can be used for this purpose, but GF/A filters are less expensive because they are only half as thick. 9. The choice of the ligand and the concentration used to determine nonspecific binding are critical to the success of the experiment. It is best to use a ligand that is chemically dissimilar to the radioligand to prevent the labeling of “ligand-specific” nonreceptor sites. If at all possible, avoid the use of the unlabeled form of the radioligand. The concentration of the ligand should be sufficiently high to inhibit all specific binding, but none of the nonspecific binding. This can be checked by doing inhibition experiments with several ligands and ensuring that they all give a similar level of nonspecific binding. 10. The time required for shaking and/or waiting before counting will depend on the scintillation cocktail used. The number of cpm in the samples should be identical (within counting error) when recounted 5–10  h later. Many radioligands are lipophilic and for such ligands a nonaqueous scintillation cocktail can be used. For ligands that are less lipophilic, a more expensive aqueous cocktail must be used. If an aqueous cocktail is used, then the 20-mL aliquots of radioligand solution do not need to be spotted onto filter paper, but can be added directly to the cocktail. When used with a nonaqueous

Radioligand Binding Methods for Membrane Preparations and Intact Cells

161

cocktail, the filter paper increases the surface area dramatically so that the radioligand can better partition into the cocktail. 11. In a well-designed experiment there should be an equal number of points above and below the Kd and the highest ligand concentrations should be approximately ten times the Kd. This sample experiment could be improved by using a half to a quarter as much radioligand in order to give a better spread of data points. 12. The term Scatchard analysis is frequently used to describe this linear transformation of saturation data. However, Scatchard’s paper (3) is often not referenced. Even when it is referenced, it seems that the authors either have not read it or do not understand it. The Scatchard derivation assumes a single species of binding macromolecule of known molecular weight and concentration, and the intercept at the abscissa is the number of ligand binding sites per macromolecule. The Bound/Free versus Bound plot was first used by Rosenthal (4), and for most receptor binding studies this is the more appropriate reference. One unique feature of this plot is that radial lines through the origin represent the free radioligand concentration. 13. Ideally the data points should be equally spaced along the line and randomly distributed about the line. In addition, the lowest Bound point should have a Bound/Free ratio of less than 0.1. At ratios higher than 0.1, greater than 10% of the free ligand is depleted, and the equations used for analyzing the data are no longer valid. If a ratio greater than 0.1 is obtained, the tissue concentration should be reduced, or the assay volume increased. 14. An equal number of concentrations above and below the anticipated IC50 value should be used. Typically, nine or ten concentrations of the inhibitor are used. A concentration spacing of half-log units is frequently appropriate. Because the inhibitor will be diluted 100-fold in the assay, the stock solutions are made up at a 100-fold higher concentration. 15. If desired, add 10  mL of the compound normally used to determine nonspecific binding (rather than the inhibitor) to the 12th pair of tubes to confirm that the inhibition caused by the highest concentration of the inhibitor is consistent with that caused by the “standard” compound. 16. The equation published by Cheng and Prusoff was independently derived by Jacobs et al. (5), but is only valid if the Hill slope is unity. For cases when it is not unity, a revised equation has been developed (6). 17. Data from inhibition experiments are always best analyzed using nonlinear regression techniques. If the data are consistent with a single binding site interaction, then the data can

162

Bylund and Toews

be visualized as a simple sigmoidal inhibition curve of bound radioligand (as percent of maximum) versus the logarithm of the inhibitor concentration. If there are multiple sites, however, the data are best visualized using a plot of Bound versus (Bound × Inhibitor) concentration (7). An interesting special case is when the radioligand and the competing ligand have the same affinity for the receptor, such as if the inhibitor is the unlabeled form of the radioligand. In this case, for a plot of Bound versus (Bound × total Free ligand) (i.e., radioligand plus inhibitor), the negative reciprocal of the slope is the Kd, rather than the IC50. 18. Total added is not needed to calculate k−1, but it is good to know that the anticipated amount of radioactivity was actually added to the assay tubes. 19. Several complications of intact cell binding assays are unique to the case of cells in monolayer culture. Whereas the receptor concentration in the assay can be easily varied for suspension cells, for monolayer cells it is difficult to vary the receptor concentration except by varying the number of cells plated per culture dish or by varying the extent of confluence at which the cells are assayed. Furthermore, the monolayer cells are not “in solution” in the assay medium and their concentration is not uniform throughout the medium, and this may complicate some of the theoretical aspects of receptor analysis. 20. The cells must first be grown in the appropriate number and type of vessels for the assays to be performed, with either monolayer cells or cells in suspension. Cells grown in monolayer culture can also be released from the monolayers and assayed in suspension if this is more convenient. An advantage of monolayer culture is that the various medium changes and washes required for the assays can be accomplished by simply aspirating with vacuum and replacing with the next solution. On the contrary, cells in suspension culture require centrifugation for medium changes and washing, which may take longer. Suspension culture cells can be grown in a single vessel, harvested and washed, and then used in assays with individual tubes essentially identical to assays with isolated membrane preparations. The assay tubes can generally be prepared on ice, all started simultaneously by placing the rack of tubes in a water bath, and all terminated simultaneously by filtration with a cell harvester. On the contrary, for monolayer culture, the cells must be grown in as many separate vessels as the number of assays that are to be performed. For a typical saturation experiment with eight concentrations assayed in triplicate for total and nonspecific binding, 48 separate dishes or

Radioligand Binding Methods for Membrane Preparations and Intact Cells

163

wells need to be plated in advance. Thus, details of the ­specific experiment to be conducted need to be known at the time of plating, so that the proper number of dishes or wells are available. Monolayer cells can be plated either on 35-mm dishes and processed individually, or they can be plated on multiwell plates, in which case multiwell plate washers and cell harvesters can be used for washing and collecting the samples. For saturation and competition assays, where the binding time is constant for all samples, multiwell plates are most convenient, since all of the samples can be washed and harvested simultaneously. On the contrary, for kinetic experiments where the time of association or dissociation is varied, each set of dishes must be assayed separately and manually; nonetheless, fouror six-well plates are still more conveniently handled than the same number of individual 35-mm dishes. At the end of the assays, the final collection of samples for counting is by filtration for cells assayed in suspension or by dissolving the cell sheet with its bound radioligand in NaOH or detergent for cells assayed as monolayers. 21. The assay buffer used for intact cells is also critical. To maintain the cells intact and viable, the assay buffer should be isotonic and should contain an adequate energy source. This is most easily accomplished by utilizing serum-free growth medium as the assay buffer, but a balanced salt solution supplemented with glucose as energy source can also be used. Because various portions of the assays are done outside of the culture incubator, a nonvolatile buffer such as Trsi/HCl or HEPES should be used, rather than the CO2/HCO3− buffer system used for growing the cells, and the binding incubations should be done in a cell culture incubator without CO2. 22. Sterile technique is no longer needed. 23. Slightly more solution is prepared at each step than is needed, to ensure that the appropriate number of identical aliquots can be recovered from each tube. 24. Including 100 mM propranolol in the post-incubation wash buffer has been shown to reduce nonspecific binding by unknown mechanisms, not only for the b-adrenergic receptors for which propranolol is a high-affinity antagonist, but also for a1-adrenergic receptors and muscarinic receptors. Phentolamine has also proven useful for some receptors. Thus, testing a variety of drugs for inclusion in the postincubation wash buffer to reduce nonspecific binding may prove beneficial. 25. Pull the NaOH solution into the pipettor tip, and while holding the plate at a 45° angle, rinse the plate from top to bottom twice with the same solution before transferring; we have

164

Bylund and Toews

not found it important to rinse the plate with a second aliquot to obtain reproducible and quantitative transfer. 26. The NaOH may cause problems with chemiluminescence in some scintillation cocktails; this can be avoided by neutralizing the NaOH after transfer or by choosing a different scintillation cocktail. References 1. Toews, M. L. (2000) “Radioligand-binding assays for G protein-coupled receptors” in Regulation of G protein-coupled receptor function and expression (Benovic JL, ed); pp. 199–230;Wiley-Liss, Inc. 2. Cheng, Y. C. and Prusoff, W. H. (1973) Relationship between the inhibition constant (KI) and the concentration of inhibitor which causes 50 per cent inhibition (I50) of an enzymatic reaction. Biochem. Pharmacol. 22, 3099–3108. 3. Scatchard, G. (1949) The attractions of proteins for small molecules and ions. Ann. N. Y. Acad. Sci. 51, 660–672. 4. Rosenthal, H. E. (1967) Graphical method for the determination and presentation of binding

parameters in a complex system. Anal. Biochem. 20, 525–532. 5. Jacobs, S., Chang, K. J. and Cuatrecasas, P. (1975) Estimation of hormone-receptor affinity by competitive displacement of labeled ligand: Effect of concentration of receptor and labeled ligand. Biochem. Biophys. Res. Commun. 66, 687–695. 6. Cheng, H. C. (2002) The power issue: determination of KB or Ki from IC50: A closer look at the Cheng-Prusoff equation, the Schild plot and related power equations. J. Pharmacol. Toxicol. Methods 46, 61–71. 7. Bylund, D. B. (1986) Graphic presentation and analysis of inhibition data from ligand-binding experiments. Anal. Biochem. 159, 50–57.

Chapter 9 Quantification of GPCR mRNA Using Real-Time RT-PCR Trond Brattelid and Finn Olav Levy Abstract Characterisation of G-protein-coupled receptor (GPCR) mRNA expression under normal, different pharmacological and pathological conditions in experimental animal models and human tissue biopsies by quantitative real-time reverse transcription polymerase chain reaction (RT-qPCR) is a valuable approach to understand the regulation of GPCR expression. RT-qPCR is specific and sensitive with a broad dynamic range, which allows precise quantification of mRNA species of interest. In addition to measuring the relative levels of mRNA in a tissue or changes in expression levels between groups of genes of interest, RT-qPCR is also used to identify splice variants and single nucleotide polymorphisms (SNPs) of GPCRs. Even though RT-qPCR has become the standard method for quantification of gene expression, RT-qPCR is sensitive to RNA quality, assay design, normalisation approach and data analysis. This protocol is meant as a guide to RT-qPCR methodology with references to the best standard methods available at present. Key words: Quantitative PCR, RNA isolation, Reverse transcription, Normalisation, Data analysis, RT-qPCR

1. Introduction Quantitative real time reverse transcriptase polymerase chain reaction (RT-qPCR) is a powerful and sensitive method of comparing the expression level of genes encoding GPCRs between different sample populations. It is dependent on: (1) tissue/sample isolation and preparation; (2) RNA isolation; (3) cDNA synthesis; (4) PCR amplification; (5) real time detection and quantification of DNA sequence amplified in the PCR reaction; (6) normalisation strategy and; (7) analysis of results. Each of these steps are vulnerable to the experimental procedures/strategies adopted. The specificity and sensitivity of RT-qPCR make it possible to detect small differences in mRNA expression between samples.

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_9, © Springer Science+Business Media, LLC 2011

165

166

Brattelid and Levy

In biological and medical research, the introduction of hydrolysis probes (e.g. TaqMan®) and intercalating dyes (e.g. SybrGreen) has improved the quantification of nucleic acids by RT-qPCR. Especially in situations where there is a lack of sensitive ligands and/or specific antibodies to determine the protein level of GPCRs, gene expression studies by RT-qPCR have been the best alternative to indirectly detect changes in expression of these receptors. The PCR reaction is based on the amplification of a given DNA sequence restricted by two primers complementary to the two ends of the sequence of interest. Both primers act as starting points for DNA polymerase, which builds a new DNA strand complementary to the template through elongation of the primer by adding single nucleotides present in the PCR reaction. The reaction is driven by temperature cycling as shown in Fig.  1. In theory, the rate of reaction is exponential and the amount/ number of products of the sequence of interest is doubled for each reaction cycle. The standard PCR is an end point method, analyzing the products accumulated at the end of the reaction. RT-qPCR is based upon detection of a fluorescent dye accumulating proportionally to the amount of the PCR product, and measured in each PCR reaction cycle (Fig.  2). The higher the number of templates present in the sample being analysed, the earlier the fluorescent signal will reach a certain threshold value. The number of cycles needed to reach this threshold value is referred to as the quantification cycle (Cq), which can be used as a measure of the initial number of templates. The more templates present, the lower the Cq value. The term quantification cycle (Cq) is now recommended, as compared to the manufacturerplatform terms like “cycle of threshold” (CT) and “crossing point” (CP) value (1). There are several detection chemistries available for RT-qPCR, each with different characteristics. The two most frequently used are the hydrolysis probes, e.g. TaqManÔ, or the intercalating dye SYBR Green, which have both been used to detect GPCRs in different samples (Fig. 3) (2–5). In addition, there are hybridisation probes and different versions of hairpin probes available. The intercalating dye has high affinity for double stranded DNA and this binding enriches fluorescence significantly. Fluorescence therefore increases with the amount of PCR product accumulating. Since the intercalating dye binds to all double stranded PCR products, it does not discriminate between the sequence of interest and any unspecific product generated in the PCR reaction. Although the probes can be of different varieties, they are all based on quenching of a fluorescent signal from a reporter. Loss of quenching increases the fluorescent signal from the reporter. The hydrolysis probe is composed of a single DNA strand, usually less than 30 nucleotides in length, with a fluorophore (reporter dye) linked to the 5¢-end and a quencher to the 3¢-end.

Quantification of GPCR mRNA Using Real-Time RT-PCR

167

Fig. 1. The PCR reaction is driven by temperature cycling. (a) Ordinary three step temperature cycling PCR. (1) The double stranded DNA is melted/denatured (separated) by heating the PCR reaction to 95°C. (2) By temperature cycling of the PCR reaction the sequence of interest is amplified. By cooling down to e.g. 60°C, primers anneal to their complementary sequence on each template. By heating to 72°C, the optimal temperature for the heat-stable DNA polymerase, this enzyme makes a copy of the sequence of interest. By cycling, this temperature regime generates new PCR products in an exponential manner. The emitted fluorescence is recorded at the end of the elongation step. (3) After finishing the temperature cycling the PCR reaction is cooled to 4°C until further processing. (b) A two step temperature cycling PCR with an intercalating dye binding to double stranded DNA. (1) The double stranded DNA is melted/denatured (separated) by heating the PCR reaction to 95°C. (2) By temperature cycling of the PCR reaction the sequence of interest is amplified. Both annealing of primers to their complementary sequence and copy of the sequence of interest are performed at 68°C. Annealing at 68°C increases the specificity of the PCR reaction. By cycling, this temperature regime generates new PCR products in an exponential manner. The emitted fluorescence is recorded at the end of the annealing and elongation step. (3) To control for unspecific priming and amplification a melt curve is run at the end of the PCR cycling. The sample with PCR product is slowly heated from 55°C to 95°C, monitoring the fluorescent signal in the PCR sample. When the sample is heated to the melting point of the amplicon the two DNA strands move apart and the intercalated dye is dissolved into the sample reaction mix resulting in a fading of fluorescence. If several melting points are detected in the same sample there might be splice variants of the gene of interest or the primer pair might not be specific enough. Alternatively, primer-dimer formation may have occurred or genomic DNA might be present.

168

Brattelid and Levy Plateau

Exponential E1

E3

E2

Background Exponential

Plateau

E1

E2

E3

Fluorescence signal (Log)

Background

1

2

Threshold

Cq1

Cq2

Cycle Fig. 2. The fluorescent signal in the RT-qPCR amplification increases with cycle number. Within the exponential phase of the RT-qPCR, where the PCR product theoretically is doubled with each cycle, three phases can be defined. In the “linear” ground phase, E1, the fluorescent signal is below the detection level, whereas in the early exponential phase, E2, and the exponential log linear phase, E3, the fluorescent signal is above background level. The data should be collected where all the samples are in the E3 exponential log linear phase. In the plateau phase the PCR reaction has reached its limit due to depletion of critical reagents, and the amplification process declines dramatically. A sample with high initial template concentration (black curve 1) has a lower Cq value than a sample with a low number of templates (grey curve 2 ).

By Förster resonance energy transfer (FRET), light excites the fluorophore which emits a fluorescent signal that is quenched by the quencher. Breaking up the probe DNA strand connecting the fluorophore and quencher increases the distance of the fluorophore and the quencher and reduces the quenching. If DNA replication occurs in each cycle, the 5¢–3¢ endonuclease activity of the qPCR enzyme cleaves up a hydrolysis probe, which increases the fluorescent signal for each cycle. The PCR reaction is based on the ability of a primer to anneal to its complementary sequence and the DNA polymerase to elongate the primer. If the amount of template is low for the gene of interest, the amplification efficiency might vary between samples due to a reduction in the probability of template amplification. Even when templates are present, differences in amplification in the initial cycles may produce significant differences in the exponential phase of the quantitative RT-PCR reaction. Low levels of

Quantification of GPCR mRNA Using Real-Time RT-PCR

169

Q

Denature

Q

Anneal

Q Pol

Extend

Pol

Fig. 3.  The functional characteristics of hydrolysis probes (left ) and intercalating dyes (right ) in qPCR. Left – The hydrolysis probe with a fluorophore (open circle) attached to the 3¢-end and a quencher (Q) to the 5¢-end does not emit fluorescence upon stimulation by an external light source. Attached to its complementary sequence in close proximity of a primer, the polymerase (Pol) with a 5¢–3¢ endonuclease activity dismantles the probe sequence and allows the fluorophore and quencher to drift apart. The stimulated free fluorophore is then able to emit fluorescence. Right – Exposed to excitation light the DNA intercalating dye is invisible when dispersed in solution. However, when several dyes are gathered close together, bound to a double stranded DNA and exposed to the excitation light a fluorescent light is emitted. This property underlies its use in qPCR.

a GPCR mRNA are therefore vulnerable to inaccurate measurement. However, independent of template concentration and type of application, rigor and accuracy is essential to generate reliable and reproducible data when working with RT-qPCR, especially upstream of the qPCR (Fig. 4). The running conditions (temperature cycling, primer and 2+ Mg concentrations), inhibitors carried over from the tissue or RNA analysis and template conformation, all have an impact on PCR performance. It is therefore important to optimise the running conditions to obtain the best specificity, sensitivity and reproducibility (see Note 1). Good primer design is vital for a robust PCR and verification of the primers is important to confirm their specificity. The ­primers should be located in a region of the gene of interest which is not GC rich and does not contain palindromic sequences, which may form secondary structures. If the gene of interest is encoded by several exons, primers should preferably cover an exon–exon junction, ideally with a 5–7-nucleotide overlap in the 3¢-end of the primer with the following exon. Several GPCRs are encoded by genes with several exons with the opportunity for alternative splicing. This is an advantage when it comes to primer design and makes it possible to design primers of high specificity to the mRNA variant of interest.

170

Brattelid and Levy

f -

-

Fig. 4. Considerations of importance when planning an RT-qPCR experiment. Good design as well as detailed documentation of experimental methods and procedures are prerequisites to enable a thorough interpretation of the RT-qPCR data. Several different methods and applications available at each experimental level demonstrate the complexity of RT-qPCR. Meticulous planning and knowledge of possibilities and limitations of the methods used at all stages are prerequisites of a successful RT-qPCR experiment. Critical procedures in the experimental and analytical steps of RT-qPCR are outlined.

Quantification of GPCR mRNA Using Real-Time RT-PCR

171

Internet based tools including BLAST or mfold (see Note 2) enable in silico evaluation of the specificity of the primer pair and the secondary structure of the template, respectively. Several primer design tools available on the internet have the BLAST search integrated. In addition, the melting temperature (Tm) should be in the range of 58–60°C in the relevant reaction buffer to allow high assay specificity (Fig. 1). However, in silico evaluation does not guarantee a successful PCR. The primer pairs need to be evaluated for specificity by running a PCR. The specificity of the primers should be tested by performing a melting curve analysis on the final PCR product with an intercalating dye such as SYBR Green. Only primers which give a single melting point should be used. Finally, sequencing of the PCR product is optimal to confirm the specificity of the primers. The process from collection of test sample (e.g. tissue or cultured cells) to interpretation of results is highly dependent on experience as well as optimisation of the procedures and techniques involved in RT-qPCR. There is a large palette of kits available for all of the steps involved, each with their particular properties. To increase the reproducibility, each lab must optimise their RT-qPCR method by systematic and careful evaluation of all the procedures and techniques involved. The ability to perform more than one qPCR in each PCR reaction is called multiplex PCR. This may reduce costs, increase throughput and allow evaluation of a set of normalisation standards and one or more genes of interest in the same reaction. However, the presence of more than one primer pair in the PCR reaction requires additional optimisation to avoid accumulation of primer dimers. In addition, the amplification efficiency and detection limit of the sequences analysed should be identical for each individual qPCR. In order to generate a consensus for how to perform and evaluate RT-qPCR experiments, Bustin et  al. (1) generated the “Minimum Information for publication of Quantitative real-time PCR Experiments” (MIQE guidelines), which aim to improve experimental practice, interpretation and communication of RT-qPCR. Indeed, adherence to the MIQE guidelines is now a requirement for publication in several journals. These guidelines should therefore be taken into account at the stage of planning a new RT-qPCR experiment. The aim of this chapter is, therefore, to give an overview of RT-qPCR with its possibilities and limitations, based on our experience gained from RT-qPCR on GPCRs in cardiac tissue. This method is based on manual pipetting and optimised to identify low mRNA levels. In high throughput setups, pipetting robots are usually used to pipet 96/384 well plates, and procedures, techniques, enzymes and reagents indicated here are only suggestive and can be adapted to optimise for other approaches.

172

Brattelid and Levy

2. Materials 1. 10× RNA loading buffer: glycerol 50% (v:v), bromophenol blue 0.4% (w:v) and xylene cyanol 0.4% (v:v) dissolved in nuclease free water. Store at 4°C. 2. Running buffer, 1× TAE (40  mM Tris acetate and 1  mM EDTA) is made by diluting a stock solution of 50× TAE in Milli-Q (deionised) water. To make 1  L of 1× TAE, dilute 20  mL of 50× TAE in 980  mL of Milli-Q water. Store at room temperature. A 50× TAE buffer is made by dissolving 242 g Tris base (FW = 121.14) in 750 mL Milli-Q water. Add 57.1 mL of glacial acetic acid and 100 mL of pH 8.0 adjusted 500 mM EDTA. Adjust the volume to 1 L and store at room temperature. 500 mM EDTA at pH 8.0 is made by dissolving 93.05 g of EDTA disodium salt (FW = 372.2) in 400 mL of Milli-Q water and adjusting the pH with NaOH. Make sure that all the EDTA has been dissolved before adding water to a final volume of 500 mL. Store at room temperature.

3. Methods 3.1. Starting Material

RT-qPCR provides a snapshot of the gene expression in a given tissue or cell type at a given time point. To obtain a clear and focused picture of the gene expression at the time of sample collection, it is important to conserve RNA of good quality and to avoid degradation. Whole tissue biopsies, micro dissected samples or cell cultures can be processed immediately or stored for later use by snap freezing (e.g. liquid nitrogen) or submerged in RNA stabilizing solution. In addition, archived formalin-fixed paraffin embedded samples can be analysed by RT-qPCR.

3.2. Homogenisation, Phase Separation, RNA Precipitation and Washing

In RT-qPCR, the steps prior to PCR are the most vulnerable to the experimental conditions and procedures. The final results should represent the mRNA expression at the time of sample collection. It is therefore important to evaluate the sampling procedures in light of the RNA stability to prevent inconclusive results. The sampling strategy is dependent on the type of sample, sampling time, sampling conditions and technique. Samples can be collected at individual, organ or cellular level. Gross dissection of a whole or part of an organ results in a more heterogeneous sample than e.g. a micro laser dissection collection of a particular cell type. Although the RNA in archived paraffin embedded materials most likely has been fragmented and modified it is possible to recover RNA from such samples. Samples can either be lysed directly in the RNA isolation lysis buffer, conserved in e.g. RNAlater (see Note 3) or flash frozen in liquid nitrogen (see Note 4).

Quantification of GPCR mRNA Using Real-Time RT-PCR 3.2.1. Homogenisation

173

1. A maximum of 50 mg of tissue is homogenised in 1 mL of Trizol (Invitrogen) in single tubes together with ceramic beads in a rock rotating shaker e.g. Precellys (see Note 5). 2. Insoluble material is removed from the homogenates by centrifugation at 12,000 × g for 10 min at 2–8°C. Due to the power of RT-qPCR, the integrity of RNA isolated is critical to produce accurate and reliable results. By all means, a primary goal in RT-qPCR studies is to isolate RNA of good quality. Removed from its cellular environment, RNA is vulnerable and easily degraded if not carefully handled. RNases are abundant and represent a threat to the RNA. Precautions should therefore be taken to avoid contamination by use of RNase free materials, reagents and equipment at all stages. Gloves should be worn during all procedures and changed frequently. Work bench, pipettes and other surfaces which may come in contact with tubes and gloves represent a potential threat to the sample and should be washed with an RNase decontaminating reagent prior to start of the experiment. To avoid problems with contamination several scientific labs have a room or hood dedicated to RNA isolation only. Pipetting at all stages should be performed with filter tips. All tubes should be RNase free and labelled in advance (see Note 6) and reagents (see Note 7) thoroughly mixed (see Note 8).

3.2.2. RNA Isolation

There are several methods available to isolate RNA, including phenol–chloroform extraction and column based methods, which both isolate total RNA or mRNA, or alternatively magnetic bead based methods that isolate mRNA only. The column based isolation of RNA usually gives RNA of good quality that is low in contaminants, whereas phenol–chloroform isolation has a high capacity with a high yield and allows bulk isolation, but is vulnerable to contamination by genomic DNA (gDNA), proteins and/ or organic solvents. The following procedures of the RNA isolation should be carried out in a dedicated room or hood with RNase free surfaces, reagents and equipment. 1. The homogenate is incubated for 5 min at 15–30°C (room temperature). 2. Prepare Phase Lock Gel (5 Prime) tubes (see Note 9) by a quick spin (1,000 × g for 30 s) to collect the gel at the bottom of the tube. 3. Transfer the supernatant from the homogenates to the Phase Lock Gel tubes. 4. Add 200  mL of chloroform per 1  mL of Trizol and shake vigorously by hand for 15 s. 5. Incubate for 2–3 min at 15–30°C.

174

Brattelid and Levy

6. Centrifuge the samples at 12,000 × g for 15 min at 2–8°C. 7. Transfer the clear upper aqueous phase to fresh tubes. 8. Precipitate the RNA by mixing with 500 mL of isopropanol (see Notes 10 and 11). 9. Incubate samples at 15–30°C for 10 min. 10. Centrifuge at 12,000 × g for 10 min at 2–8°C. 11. Remove the supernatant and wash the RNA pellet once with 1 mL of 75% ethanol. 12. Mix the sample by vortexing and centrifuge at 7,500 × g for 5 min at 2–8°C. 13. Remove the ethanol and let the pellet air-dry for 5–10 min. Watch the drying process. If the pellet gets too dry and vitreous it might be difficult to redissolve it in water. 14. Dissolve RNA in RNase-free water (see Note 12) by passing the solution a few times through a pipette tip. 15. Incubate on a heating block for 10 min at 55–60°C to dissolve the RNA completely. 16. Pipette aliquots for spectrophotometric determination and agarose gel electrophoresis. 17. Store RNA at −70°C (see Note 13). To eliminate the problem of contamination of RNA with genomic DNA (gDNA), especially when amplifying genes without exon–exon boundaries, DNase treatment is recommended. In column-based RNA isolation assays this is performed when nucleic acids are bound to the matrix, whereas in phenol–chloroform based assays the DNAse treatment needs to be carried out as an additional operation, which includes an additional precipitation step and loss of some RNA. The phenol–chloroform based Trizol method can result in RNA contaminated by gDNA. To eliminate any traces of gDNA the RNA samples are DNase treated. DNase treatment should also be performed when column-based methods are used for RNA isolation. Usually the DNase treatment can be performed directly on the RNA bound to the column. 3.2.3. DNase Treatment of RNA

1. RNA in RNase-free water: 25 mg. 2. RQ1 RNase-free DNase 10× Reaction Buffer (Promega): 5 mL. 3. RQ1 RNase-free DNase (1 U/mL) (Promega): 5 mL. 4. RNase-free water to a final volume of 50 mL. 5. Incubate at 37°C for 30 min (heat block or water bath). 6. Add 5 mL of DNase Stop Solution to terminate the reaction. 7. Incubate at 65°C for 10 min to inactivate the DNase.

Quantification of GPCR mRNA Using Real-Time RT-PCR 3.2.4. Precipitation of RNA

175

To remove any traces of the DNases prior to cDNA synthesis, the RNA is precipitated and washed. This wash step also dilutes any RT and PCR inhibitors that might be present. 1. Add the following components to the DNase treated RNA samples: 0.1 volume 5 M ammonium acetate (Ambion) 1 mL Glycoblue (Ambion) (see Note 12) 2.5 volumes absolute (100%) ethanol. 2. Mix thoroughly by vortexing. 3. Incubate at −20°C overnight or at −70°C for 30 min. 4. Recover the RNA by centrifugation at 12,000 × g for 30 min at 4°C. 5. Remove the supernatant carefully using a pipette. 6. Centrifuge the tubes briefly and remove the remaining fluid. 7. Add 1  mL of 75% ethanol to wash the pellet. Shake carefully. 8. Centrifuge at 12,000 × g for 10 min at 4°C. 9. Remove the supernatant as in steps 5 and 6. 10. Dry the pellet for 5 min. 11. Add 20 mL of RNase-free water to dissolve the RNA. 12. Pipette aliquots for spectrophotometric determination by the use of NanoDrop as described below and RNA integrity analysis by Agilent Bioanalyzer or BioRad Experion. 13. Add 20 U RNase inhibitor per 20 mL sample volume to the remaining RNA solutions. Mix by careful pipetting. 14. Freeze for storage at −70°C. Since much of the success in RT-qPCR relies on the RNA quality it is important to evaluate the RNA quality. There is a long tradition of evaluating the concentration and purity of the RNA by determining absorbance at 260/280 nm and at 260/230 nm. This is of limited value since it gives no information about the degradation of the RNA. By use of capillary electrophoresis systems, for example, Bioanalyzer (Agilent) or Experion (BioRad), the ratio of the 28S:18S ribosomal RNA levels can be obtained to give a good indication of the RNA quality. Based on experience in the analysis software and characteristics of the ribosomal RNA electropherogram profile it is possible to calculate an RNA integrity number (RIN) (6). On the other hand it is uncertain whether evaluating the ribosomal RNA is representative of the mRNA integrity. However, it has been demonstrated that the detected mRNA expression level of a given gene can correlate with the RIN value, but moderately degraded RNA did not differ from high quality RNA after normalisation of the results (7) (see Note 14).

176

Brattelid and Levy

3.3. Determination of RNA Concentration by the Use of a NanoDrop ND-1000 Spectrophotometer ( see Note 15)

1. Mix 2 mL RNA solution with 8 mL RNase-free water.

3.4. RNA Evaluation by Agarose Gel Electrophoresis

1. Mix 5 mL of the RNA sample with 2 mL of formamide.

2. Run a blank by the use of 2 mL RNase-free water. 3. Use volumes of 2 mL and run each sample twice. Analysis of RNA integrity should be performed by one of the  micro fluid capillary electrophoresis platforms, e.g. Agilent Bioanalyzer or BioRad Experion. Detailed protocols for these instruments can be found elsewhere (see Note 16). However, the cost of the RNA integrity analysis limits the use of this method to evaluate the RNA following DNase digestion prior to cDNA synthesis. At this stage of the RNA isolation, a quick evaluation of the RNA can be performed by a simple agarose gel electrophoresis.

2. Prepare a 1.5% (w:v) agarose gel (MetaPhor Agarose, Lonza) in TAE buffer. Bring the TAE buffer with agarose added to the boiling point in e.g. a microwave oven several times until the agarose is dissolved. Mix thoroughly without shaking after each boiling sequence. 3. Add GelRed (1:10,000 v/v; Life Technologies) (see Note 17) and mix thoroughly. 4. Cast the gel in a horizontal tray. 5. Heat the samples at 70°C for 5–10 min. 6. Cool on wet ice for a couple of minutes. 7. Add 2  mL of RNA loading buffer to each sample and spin quickly to collect the contents of the tubes. 8. Load the samples on to the gel. 9. Use a voltage of 50  V and let the samples run for about 45 min (see Note 18). 10. Picture the gel by the use of UV light. Clear bands of 18S and 28S should be seen with little “smear” in the area below the 18S. This is only a subjective quality assessment and does not give sufficient information about the quantity or degree of degradation. Another approach to evaluate the integrity of the RNA is to analyze the expression level of, for example, a reference gene by RT-qPCR with two independent primer sets, one located in the 5¢-end and the other in the 3¢-end of the mRNA (8). If these two assays give almost identical expression levels, the mRNA can be considered as not degraded.

3.5. cDNA Synthesis

The RT step in the RT-qPCR can be performed either as a singletube or two-tube approach. The single-tube assay is either based on a single enzyme or two independent enzymes for the RT and PCR reactions, performed in the same tube. The single enzyme uses a polymerase, which is able to utilise both RNA and DNA as

Quantification of GPCR mRNA Using Real-Time RT-PCR

177

the template and allows direct PCR on the RNA. This is a time saving approach with reduced probability of contamination but is restricted to the use of gene-specific primers. There is also a reduced possibility for a complete optimisation. The sensitivity of this single-tube approach is low and in multiplex analysis the method is associated with primer-dimer accumulation. The other single-tube approach is based on adding reverse transcriptase and polymerase to the same tube in sequence. Following completion of the cDNA synthesis (where the reverse transcriptase is present in the starting reaction mix primed with oligo-dT), random hexamers or gene specific primers, DNA polymerase, primers for the gene sequence(s) of interest and PCR buffer are added to perform PCR in the same tube. The two-tube approach with RT and PCR in separate tubes allows optimisation of both the RT and PCR steps of the RT-qPCR, which gives a high degree of sensitivity. Following the RT step a portion of the cDNA synthesised is transferred to a new tube for PCR amplification. Although this approach is vulnerable to contamination, the two-tube approach is often the preferred method in research due to its flexibility. To enable comparison of results between different experiments or different labs it is essential to use the same RT strategy without any variation in RNA concentration, type of enzyme, units of enzyme, volume, priming strategy and temperature regime. The reverse transcriptase-mediated synthesis of cDNA from RNA and PCR on the reverse-transcribed cDNA is sensitive to inhibitors found in tissue samples or accumulated during sampling and/or RNA isolation. Inefficient cDNA synthesis and/or PCR lead to incorrect results. Blood in particular is known to contain several inhibitors due to the presence of heme, heparin, immune globulins, bile, urea and lipids. Traces of reagents used during RNA isolation such as organic solvents might also lead to inhibition of cDNA synthesis and PCR efficiency. An assay named SPUD has been developed by the group of Bustin (9) to identify the presence of inhibitors in the reverse transcription or PCR. By comparing PCR on a known concentration of a synthetic sequence added to either the sample RNA or water it is possible to determine the degree of contamination with inhibitors. A higher Cq value in the RNA compared to the water sample is an indication of inhibitors present in the RNA sample. First strand cDNA synthesis (see Note 19) in a final volume of 40 mL: 1. Add the following components to nuclease-free microcentrifuge tubes: Oligo (dT)12–18 (25 mM) (see Note 20)

  2 mL

dNTP Mix, 10 mM (dATP, dGTP, dCTP and dTTP)

  2 mL

Total RNA in RNase-free water

  5 mg

RNase-free water to a total volume of

26 mL

178

Brattelid and Levy

2. Heat at 65°C for 5 min using a heating block and incubate on wet ice for a couple of minutes. 3. Spin quickly to collect the contents of the tubes. 4. Add the following components: 5× First-strand buffer

8 mL

0.1 M DTT

2 mL

RNase inhibitor (40 U/mL)

2 mL

Reverse transcriptase (RT; 200 U/mL)

2 mL

5. Mix gently by pipetting. 6. Incubate at 25°C for 5 min, then at 50°C for 60 min. Finally inactivate the reaction by heating at 70°C for 15 min. Cool to 4°C. 7. Dilute each sample of template (including the standards) with 120 mL of nuclease-free water. 8. Store the cDNA at −20°C. The cDNA is diluted by DNasefree water to avoid pipetting small volumes of template which will increase the variation between samples. The transcription efficiency is dependent on both the priming strategy and gene of interest (10, 11). Comparison of several genes in one study could therefore be sensitive to the choice of priming strategy. However, the nature of the best priming strategy is debated, and some of the main considerations are summarised in the following. 3.5.1. Choice of Primers for First-Strand cDNA Synthesis

Gene-specific mRNA primers are not only specific but also sensitive and efficient and are preferred when the amount of RNA is high (Fig. 5). However, transcription of several sequences in one RT reaction is complicated. Each gene-specific primer represents an independent RT reaction due to different affinities to their complementary sequence. To correct for the possibly different rates of cDNA synthesis between the different mRNA species evaluated, inclusion of a standard curve is required (see Note 21). Oligo dT primers anneal to the poly A tail in the 3¢-end of the mRNA (Fig.  5). It could therefore be more difficult to detect sequences in the 5¢-end of RNA if the transcriptase falls off or is blocked by structural obstacles. RNA forms secondary structures and the cDNA might be truncated in areas with RNA loops not linearised at the temperature used for the reverse transcriptase. Since priming with oligo dT starts the cDNA synthesis from the very 3¢-end of the RNA, oligo dT should be used with caution if analyzing splicing phenomena in the far 5¢-end of the mRNA. Likewise, oligo dT should not be used to reverse transcribe degraded RNA due to the high probability of broken RNA strands, which do not allow synthesis of full length cDNA.

Quantification of GPCR mRNA Using Real-Time RT-PCR

179

Gene specific primer 3’-GATTATACACGAC-5’ 5’-AGUCCAUGCUCGAUCGAUCGGGCUUGCUAAUAUGUGCUGGAAAAAAAAAAAAAAAAAAA-3’ Oligo dT primer 3’-TTTTTTTTTTTT-5’ 5’-AGUCCAUGCUCGAUCGAUCGGGCUUGCUAAUAUGUGCUGGAAAAAAAAAAAAAAAAAAA-3’ Random hexamer 3’-NNNNNN-5’ 5’-AGUCCAUGCUCGAUCGAUCGGGCUUGCUAAUAUGUGCUGGAAAAAAAAAAAAAAAAAAA-3’

Fig. 5. Reverse transcription priming strategies. Gene-specific priming of the RT is specific but restricted to a limited number of genes of interest. The oligo dT primers hybridise to the poly A tail in the 3¢-end of the mRNA whereas random hexamers hybridise throughout the RNA where the sequences are complementary.

Random hexamers are a mix of short primers of six nucleotides in a random sequence that prime both mRNA and ribosomal RNA (Fig. 5). Since random hexamers do not anneal to a conserved sequence and the priming site will vary between each mRNA copy of the gene of interest, the efficiency might also differ from one gene to another. However, the lack of specificity of random hexamers makes them excellent cDNA primers on degraded RNA, e.g. RNA from irreplaceable samples or sources. Longer primers of random nucleotides can improve the transcription efficiency. Increasing the number of nucleotides from 6 to 15 leads to a doubling of the transcription efficiency (12). Resuehr and Spiess (13) tested cDNA synthesis with different priming strategies and found the mix of oligo dT and random hexamers to be the most efficient for PCR targets located 4,000– 12,000 base pairs (bp) away from the poly A tail. If the amplicon of interest is located close to the poly A tail of the mRNA (within 400 bp) the efficiency of oligo dT alone and the mix of oligo dT and random hexamers was identical. Although all reverse transcriptase enzymes have a linear working range, variation might occur from batch to batch and from experiment to experiment due to different running conditions. To ensure a representative transcription of mRNA into cDNA a standard curve with increasing concentrations of RNA should be run in parallel (Fig. 7). If this is not an option, an internal laboratory RNA standard with known expression level of one or more genes should be used to confirm the reverse transcription of mRNA into cDNA.

180

Brattelid and Levy

The transcription efficiency of the mRNA into cDNA is also dependent on the reverse transcriptase used (11). The more heat resistant RT-enzymes allow translation at higher temperatures (50–55°C), which open more RNA loops and linearise the RNA, thereby increasing the cDNA yield compared to RT-enzymes with a lower temperature optimum. 3.6. Design of Primers and Probes

Primers and probes used for RT-qPCR should preferentially be located at exon–exon junctions if possible and designed according to the guidelines of Bustin (14). Prior to synthesis, each primer pair should be tested for specificity in silico by a nucleotide BLAST (blastn) search in the NCBI database (see Note 22) with the forward primer followed by 15n’s and the reverse primer in sequence, and the default blastn settings adjusted to: Word Size = 7; All filters deselected; Expect Value = 1,000. Only primer pairs that detect the relevant sequence should be tested further by ordinary RT-PCR or RT-qPCR with SYBRgreen to verify the specificity by identification of a single product. One of the primers should if possible be located over an exon–exon junction to increase the cDNA specificity (black primers in Fig.  6). If this is not possible, the primer pair should be located in two neighbouring exons to reduce the risk of amplifying contaminating gDNA, which would yield a larger product

E1

I1

E2

I2

E3

I3

E4

pre-mRNA

E1

I1

E2

I2

E3

I3

E4

mRNA processing

E1

Gene

E3

E2

E4

Exon mRNA

E1

E2

E3

E4

Intron Splice site

Fig. 6.  Splicing of exons in precursor RNA to produce an mRNA. The gene with all its exons and introns is copied into precursor RNA (pre-RNA) by transcription. Processing of the pre-RNA skips the introns and fuses the exons to form mature mRNA. The primer pair could be located within one of the exons in the mRNA which would not distinguish between cDNA and genomic DNA (dark grey primer pair), because they would yield products of identical size. Primers located in two different exons will produce two products of different length (light grey primer pair). The shortest amplicon (light grey ) is due to mRNA priming whereas the longest amplicon (light grey and white) is a product of genomic DNA priming. By locating one of the primers over an exon–exon splice site, which exists only in the RNA, only mRNA will be amplified in the PCR (black primer pair ). Due to the intervening intron, the primer located across the exon–exon splice site will not bind properly to the genomic DNA to function as a starting site for the PCR (lack of annealing to the intron indicated in the figure). Although not demonstrated here, the mRNA is reverse transcribed into cDNA prior to the PCR reaction.

Quantification of GPCR mRNA Using Real-Time RT-PCR

_ . 3

181

.

.

_ .

.

.4 2

_ . 4

.

. 2

Fig. 7.  Quantitative RT-PCR standard curves for 5-HT4(b) in rat. The standard curves for 5-HT4(b) are almost identical, independent of the standards used. An absolute standard was constructed by dilutions of a PCR product extending the quantitative RT-qPCR target in both directions (upper panel ). For relative quantification, serial dilutions of RNA were used in the RT step for hippocampus (middle panel ) and left ventricle (lower panel ). The slopes of the hippocampus and left ventricle panels are almost identical, but by evaluating the y-intercept, the levels of 5-HT4(b) mRNA are higher in the hippocampus compared to the left ventricle. The left ventricle 5-HT4(b) standard curve also reveals the presence of inhibitors or an overload of RNA which is observed as a flattening of the Cq values of the standard curve at the highest RNA concentrations.

182

Brattelid and Levy

including the intron (light grey primers in Fig. 6). Such a large product will usually not be amplified with the same efficiency as the short target sequence. The primer pair should ideally define an amplicon of 80–150 nucleotides in length. Primers should be of 15–25 bases, whereas a probe should be no longer than 30 nucleotides. Guidelines for primer design 1. The melting point of the primers should be 58–60°C. 2. Ideally there should be no more than two to three Gs and/or Cs in the last five nucleotides at the 3¢ end of the primers (see Note 23). 3. Primers should be located as close to the probe as possible. Guidelines for probe design 1. The melting temperature of the probe should be 68–70°C. 2. Select the strand that gives the probe more Cs than Gs. Both orientations are possible. 3. The first base at the 5¢ end cannot be a G. The G will then be able to function as a quencher. For the sequenced organisms the different suppliers have primers and probes available for most of the genes (see Note 24) or relevant sequences could be found in primer data bases (see Note 25). However, it might be necessary to design primers and probes by use of shareware available on the internet (see Note 26). 3.7. Real-Time Quantitative RT-PCR

1. Mix the following volumes (mL) of the given components (see Note 27): RNase-free water

5.80 × (n × number of samples + 10%)

Primer 1 (forward) (10 mM)

0.66 × (n × number of samples + 10%)

Primer 2 (reverse) (10 mM)

0.66 × (n × number of samples + 10%)

Hydrolysis probe (5 mM)

0.88 × (n × number of samples + 10%)

2× Master Mix

11.0 × (n × number of samples + 10%)

n – number of replicates of each sample.

2. Pipette 57 mL of this reaction mix and add 9 mL of template dilution. Use 0.2 mL RNase-free microcentrifuge tubes. This volume of 66 mL is now a triplicate master. Keep on ice while working. 3. Use 96/384-well PCR plates and pipette 3 × 20 mL from each triplicate master. 4. Cover the PCR plate with optically clear sealing tape and collect the samples and remove air bubbles by light centrifugation.

Quantification of GPCR mRNA Using Real-Time RT-PCR

183

5. Run real-time quantitative PCR for 40 cycles with the following two step PCR running conditions: Activation (hot start qPCR enzyme)

95°C

5 min

Denaturation

95°C

30 s

Annealing and extension

68°C

1 min

If using an intercalating dye a melt curve from 60°C to 95°C should be run at the end of the experiment to control for amplification of unspecific products. 3.8. Intra and Interassay Variability

It is common to use the mean Cq of two or three (or more) reactions of the same sample to calculate the expression level of each gene analysed. The variation of the reactions of the same sample should be acceptable and a rule of thumb is to exclude Cq values which vary by more than 0.5. The variation between reactions on the same samples is also influenced by the expression level of the particular gene analysed. It might therefore be necessary to accept higher variance when working with low expressed genes, as is the case with many GPCRs. Running at least three reactions for each sample makes it possible to calculate the standard deviation and coefficient of variance (CV) on the concentration or copy number, which enables a more qualified evaluation and interpretation of the results. With at least three reactions, a more general criterion is to allow a maximum CV of 10%. However, the sample size is also of importance regarding statistical handling following data analysis. Extensive analysis of large numbers of samples requires more than one round of PCR to obtain all results. Even if the PCRs are performed in sequence there will frequently be run-to-run variations that need to be accounted for. By including one or preferably more identical calibrator samples in all the runs it is possible to correct for inter-assay variance. The analysis tool in qBase (http: // www.biogazelle.com/) has inter-assay calibration implemented.

3.9. Normalisation Strategy

The sensitivity of RT-qPCR imposes strict requirements on how to compare expression data between samples. Interpretation of RT-qPCR results is highly dependent on the internal control used, especially when measuring minor changes in expression. The most common strategy has been to normalise the expression of a specific gene to a reference gene across all samples. This assumes that the expression of the reference gene is invariant between the compared samples, so that any variation in the reference gene expression between samples would reflect variations in sample preparation and experimental variability and not variation due to e.g. the different physiological or pathological states compared. Selection of an appropriate reference gene is therefore critical (15). However, use of any single reference gene will usually

184

Brattelid and Levy

be vulnerable to changes in expression of the reference gene between experimental groups, and according to the MIQE guidelines (1), use of at least three reference genes is recommended to obtain good data. The process of selecting appropriate reference genes should be conducted each time the experimental condition or procedures are changed. Free, computer-based solutions have been developed to select good reference genes and the most frequently used are NormFinder (16) and geNorm (17) (see Note 28). Evaluation of e.g. 10 candidate reference genes of different functional classes that are selected by literature search, qualified guess or commercially available selection kits and then selecting the three genes with least variability between the experimental conditions or (patho)physiological states to be compared will lay a good foundation for appropriate normalisation of the results. 3.10. Data Analysis

In RT-qPCR the fluorescence increases proportionally with the number of amplicons accumulating in the reaction, and the fluorescence is detected progressively and monitored in real time. During the early cycles the fluorescent signal is below the background level. The Cq value obtained is defined as the cycle when the fluorescent signal is detected above the background at the threshold level set by the operator. In general, a low Cq value reflects a high amount of template, whereas a high Cq value reflects a low amount of template. Calculations of mRNA expression levels based on the Cq values must take into account the amplification efficiency of the PCR reactions. The simplest algorithms assume equal amplification efficiency between all samples analysed, whereas more advanced algorithms have been developed, which aim to compensate for differences in amplification efficiency. There is a wide variety of strategies in use to evaluate the Cq values from qPCR that are based on different mathematical models. The qPCR software supplied with the different PCR machines will perform the calculation, but several laboratories prefer custom-made spreadsheets or one of the analysis tools available, e.g. REST (18), Q-Gene (19) or qBASE (20) for more complex analysis. In the near future, new analysis software tools will be developed with simplified user interfaces. It is therefore important to monitor the progress in development of new RT-qPCR analysis tools. The data analysis software available today has been evaluated by Pfaffl et al. (21). Gene expression can be quantified absolutely or relatively. By absolute quantification the Cq value is transposed to a quantity by fitting to a standard curve based on known copy numbers. Relative quantification, as the name implies, compares the mRNA expression between the samples without absolute quantification, and is typically used to examine differences in mRNA expression between different physiological or pathophysiological conditions.

185

Quantification of GPCR mRNA Using Real-Time RT-PCR

3.11. Relative Quantification

In most RT-qPCR experiments the main goal is to evaluate the changes in gene expression between experimental groups by relative quantification, e.g. changes in GPCR mRNA expression levels in disease or during development.

3.11.1. DCq Method

The DCq method used to calculate relative changes in gene e­ xpression analysis is based on normalisation to a calibrator and assumes that all samples contain the same amount of starting material. One of the experimental samples is used as a calibrator. Usually the calibrator sample is from the experimental control group. The difference in Cq value (DCq) between the calibrator and the samples is used to calculate the difference in ratio between each sample and the calibrator.



Ratio(sample/calibrator) = E

(C q calibrator -C q sample)



(1)

where E is the efficiency of the PCR reaction assumed to be 100% with a doubling for each PCR cycle and therefore E is set to 2.

Ratio(sample/calibrator) = 2

DC q



(2)

DC q = (C q calibrator - C q sample). The DCq method requires the same amount of starting material (cDNA) in all of the samples and assumes the PCR efficiency to be close to 100%. Normalisation of the DCq against a reference gene stably expressed between experimental groups bypasses the need for accurate quantification of starting material. By calculation of the difference in Cq value between the gene of interest and reference gene in all samples, a measure of the expression of the gene of interest relative to the reference gene can be obtained. The relative expression in each sample analysed is then divided by the relative expression in a calibrator sample (chosen as described above) to give the ratio of gene expression. (C q reference gene -C q gene of interest)

Relative expression in control sample = 2

(3)

(C q reference gene -C q gene of interest)

Relative expression in treated sample = 2 Ratio of control expression =

Relative expression in control sample Relative expression in calibrator

Relative expression in treated sample Ratio of treated expression = Relative expression in calibrator 3.11.2. DDCq Method

(4) (5)



(6)

The DDCq method, also called the Livak method, assumes that the gene of interest and reference gene both have a PCR efficiency close to 100% and within 5% of each other. The DDCq method can

186

Brattelid and Levy

be used for relative quantification of the gene of interest relative to the reference gene. In this three step method the Cq value of the gene of interest is first normalised to the Cq value of the reference gene.

DC q(calibrator) = C q(gene of interest) - C q(reference gene)

(7)



DC q(sample) = C q(gene of interest) - C q(reference gene)

(8)

Then the DCq of the sample is normalised to the DCq of the calibrator (usually an experimental control sample) by subtracting the DCq of the calibrator from the DCq of the sample giving the DDCq value.

DDC q = DC q(sample) - DC q(calibrator)

(9)

Equation 8 represents the normalised change in the gene of interest between a sample and the calibrator. Finally the normalised expression ratio is calculated by:

Ratio = 2

- ( DDC q )



(10)

A negative value indicates a reduction in expression level compared to the calibration sample. 3.11.3. Standard Curve Method

By including a standard curve of increasing levels of e.g. RNA translated to cDNA, the Cq can be used to calculate the “RNA concentration” in each sample. Linear regression analysis of a standard curve with known RNA concentrations (Fig. 7) allows calculation of the log10 starting quantity by comparing sample Cq values to the standard curve (Eq. 12). The log quantity value is then converted to real quantities. Interpretation of sample Cq values is limited to the quantities covered by the standard curve.



y = ax + b

(11)



C q = a (log quantity) + b

(12)



((C q -b )/ a )

Quantity = 10



(13)

The slope, Y-intercept and r2 values of the standard curve provide information on how each assay performed as well as the PCR efficiency. The standard curve can also give an indication of the reliability of the PCR compared to previous experiments that were run under identical experimental conditions.

PCR efficency = 10(-1/slope) - 1



(14)

To account for differences in the starting amount in the PCR, the quantity data obtained for the gene of interest should be

187

Quantification of GPCR mRNA Using Real-Time RT-PCR

normalised against one or more reference genes. The ideal single reference gene might be difficult to identify, so at least three reference genes should be used. For example, a normalisation factor can be calculated for each of the samples based on a minimum of three acceptable genes by geNorm or identification of stably expressed genes by NormFinder (16). By dividing the expression level of each gene by the normalisation factor calculated by geNorm, a precise measure of gene expression can be obtained. 3.11.4. REST Method



The DCq and DDCq methods assume that the PCR efficiency is close to 100%, although this is not always the case. Correcting for the differences in PCR efficiency for the genes analysed allows a more precise and reliable calculation model to be achieved. By including a standard curve of known concentrations of RNA transcribed to cDNA in the PCR, the efficiency of the PCR reaction can be included in the calculation. This is the basis of the REST method published by Pfaffl (22). Instead of assuming that the efficiency of the PCR reaction is 100%, the slope of the calibration curve indicating the product accumulation for each PCR cycle can be used to calculate the exact PCR efficiency. Ratio =



(E gene of interest )

DC q gene of interest (calibrator -sample)

(E reference gene )

(E reference gene )

C q (sample)

(E gene of interest )

Ratio =

(15)

DC q reference gene

C q (sample)

Ratio =



(calibrator - sample)



C q (calibrator)

-

(E reference gene )

C q (calibrator)

(E gene of interest )

(E gene of interest )

DC q gene of interest (MEAN control - MEAN sample)

(E reference gene )

DC q reference gene(MEAN control - MEAN sample)

(16)

(17)

The REST application has been updated to enable normalisation to multiple reference genes (23).

3.12. Absolute Quantification

Ratio =

(E gene of interest )

DC q gene of interest (MEAN control - MEAN sample)

(E reference gene index )

DC q reference gene index

(MEAN control - MEAN sample)

(18)

Absolute quantification is performed in order to identify copy numbers of genes of interest in a given amount of sample (e.g. by weight, volume or cell number) following e.g. viral transfection with a GPCR. To enable absolute quantification a standard curve of known copy numbers is needed and used as already described above for the relative quantification. By linear regression analysis of the standard curve with known copy numbers of target gene (Fig.  7) it is possible to calculate quantity of gene of interest comparing sample Cq values to the standard curve (Eq. 12).

188

Brattelid and Levy

Extrapolation from Cq values outside the range of the standard curve is not recommended since the standard curve might not be linear outside the quantities tested. Although absolute quantification is easy to perform it needs a reliable quantification of the standard curve. 3.13. Concluding Remarks

It is also worth remembering that functional conclusions based on RT-qPCR alone are weak since the functional protein expression is dependent not only on mRNA level, but also on translation efficiency, post-translational modifications and rate of protein degradation. However, by optimisation of experimental design and practice to reduce experimental variation, RT-qPCR will provide data of good quality.

4. Notes 1. The PCR performance is dependent on the interactions between template and composition of the reaction buffer. The Mg2+ concentration affects the cDNA folding by modifying its melting temperature. Likewise, the primer concentration will have an impact on the PCR performance. Although time consuming, the PCR can be improved by titration of both the Mg2+ and primer concentrations. Although new quantitative PCR reaction chemistry is less sensitive to Mg2+ concentration it might in some cases be beneficial to optimise the Mg2+ concentration. Titrate the Mg2+ concentration by 0.5 mM increments from 1.5 mM to 5 mM and run the PCR to determine the optimal Mg2+ concentration. Set up of primer pair concentration optimisation assay: Primer

Reverse

Forward Final assay 50 concentration (nM)

100

200

300

50

50/50

50/100

50/200

50/300

100

100/50

100/100

100/200

100/300

200

200/50

200/100

200/200

200/300

300

300/50

300/100

300/200

300/300

Select the lowest primer concentration combination that gives the lowest Cq values. 2. Folding of RNA or DNA sequences can be evaluated by the mfold application found at: http://mfold.burnet.edu.au/ http://mfold.bioinfo.rpi.edu/

Quantification of GPCR mRNA Using Real-Time RT-PCR

189

3. Tissue stored in RNA Later (Ambion) should not be thicker than 2–5 mm. The volume of RNA Later should be five times the tissue sample volume. 4. By clamp-freezing the tissue sample in a zip lock bag it is easy to break loose the amount of sample needed. Tissue samples can be transferred to prelabelled tubes placed on dry ice prior to isolation of total RNA. Use forceps cooled in liquid nitrogen or dry ice for breaking the amount of tissue needed. The clamp-frozen tissue should be placed in between two blocks of dry ice while handled to avoid heating of the sample. 5. Ceramic (zirconium oxide) beads (Precellys) used for homogenisation can be reused. After removal of cell or tissue debris, the beads are washed three times in 1 M NaOH by shaking in e.g. a 50 mL polypropylene tube. The beads are then washed 3 times in RNase free water, dried and stored in e.g. a 50 mL polypropylene tube before adding to a screw cap tube (e.g. #72.694, Sarstedt) for homogenisation. 6. Labelling tubes exposed to temperature variations and solvents can be difficult. However the LABXPERT printer from BRADY (www.brady.com) allows printing of labels that can be laminated and resist freezing in liquid nitrogen and solvents. For 1.5–2 mL Eppendorf tubes use e.g. label XSL86-461 and for 0.2 mL PCR tubes use label XSL-108-461. 7. Always take out aliquots of stock solutions to reduce the risk of contamination. 8. Components (e.g. salts) of reaction mixes might precipitate during freezing. It is therefore important to mix the samples thoroughly before use to ensure accurate experimental/ reaction conditions to all samples. Be aware that some reagents (e.g. the RT-enzyme) are sensitive to mechanical stress and might be damaged by vigorous mixing. 9. Phase Lock Gel (5 Prime) allows a sealed separation of the water and organic phase and simplifies the collection of the RNA-containing water phase. 10. If the RNA amount is low add 1 mL of RNase-free glycogen (AM9510, Ambion) to facilitate the precipitation of RNA. 11. Glycoblue (AM9515, Ambion) is glycogen with a blue dye attached, which simplifies the visualisation and identification of the RNA pellet. 12. Use RNase free water not DEPC treated due to an effect of trace DEPC on RT and PCR enzymes. Bulk volumes of RNase-free water (AccuGENE water mol biolo, 51246 Lonza) can be aliquoted into smaller volumes. Ambion provides RNAse-free water in 1,000-mL bottles (AM9932) or aliquoted in 50-mL bottles (AM9937).

190

Brattelid and Levy

13. If the sample will be stored over a long period of time it is convenient to add 20  U RNase inhibitor (SUPERase In, Ambion) per 20 mL of RNA. 14. RIN values above 5 are considered as good total RNA quality, whereas RIN values above 8 are representative of RNA of perfect quality (7). 15. Quantification of RNA can also be determined by use of RiboGreen (R-11491, Molecular Probes), which stains the RNA and enables the detection of as little as 1  ng/mL RNA. 16. BioRad Experion – http://www.bio-rad.com/webroot/web/ pdf/lsr/literature/Bulletin_10004492.pdf Agilent Bioanalyzer – http://www.chem.agilent.com/Library/ usermanuals/Public/G2938-90034_KitRNA6000Nano_ ebook.pdf 17. Previously, RNA and DNA was visualised by ethidium bromide (EtBr) staining prior to evaluation and picturing under UV light exposure. However, due to the toxicity and extensive handling routines of deposits containing EtBr, new, less hazardous alternatives including GelRed and GelGreen (Life Technologies) are now available. 18. The resolution of the gel is dependent on the voltage. A higher voltage increases the speed of the nucleic acid in the gel but reduces the resolution as well as increases the risk of heating of the gel. As a rule of thumb the voltage should not exceed 5 V/cm gel. 19. For all the samples, parallel reactions should be run in the absence of reverse transcriptase. Replace the reverse transcriptase and the RNase inhibitor volume with RNase-free water. This is the contamination control. 20. The oligo dT12–18 mix can be made by ordering T12, T14, T16 and T18 oligonucleotides and diluting them in RNase free water (Ambion) to 100  mM. Mix equal amounts of each oligo dT and aliquot e.g. 50 mL of the mix into RNase free tubes and store at −20°C pending use. 21. A standard curve for RT-qPCR quantification is made by reverse transcription of increasing known concentration of mRNA to determine the concentration of target mRNA. 22. Gene sequences of interest can be blasted against assembled genomes in the National Center for Biotechnology Information (NCBI) data base to identify similar regions in the target genome: http://www.ncbi.nlm.nih.gov/BLAST/ 23. The rationale to allow no more than 2 Gs or Cs in the last five nucleotides of the primer is to reduce the hybridising strength

Quantification of GPCR mRNA Using Real-Time RT-PCR

191

at the start site of the PCR reaction. This strategy reduces unspecific priming and increases the PCR specificity. 24. Predesigned primers can be purchased from several companies including: Applied Biosystems – http://www3.appliedbiosystems.com/ AB_Home/index.htm Thermo Scientific – http://www.thermo.com SA Biosciences – http://www.sabiosciences.com/ Sigma – home.html

http://www.sigmaaldrich.com/sigma-aldrich/

Invitrogen – http://www.invitrogen.com/site/us/en/home. html Primerdesign – http://primerdesign.gene-quantification.info/ 25. It is also possible to search publicly available databases for primers: Quantitative PCR Primer Database: http://web.ncifcrf.gov/ rtp/gel/primerdb/ RTPrimerDB - http://www.rtprimerdb.org/ qPrimerDepot – http://primerdepot.nci.nih.gov/ 26. Primers for quantitative RT-PCR can be designed by internet-based sources including: Primer3 – http://primer3.sourceforge.net/ Perl Primer – http://perlprimer.sourceforge.net/ http://eu.idtdna.com/analyzer/Applications/OligoAnalyzer/ Default.aspx http://www.ncbi.nlm.nih.gov/tools/primer-blast/index.cgi  27. Due to some loss of the reaction mix during pipetting, 10% extra is added to increase the final volume sufficiently. For intercalating dye assays the volume of probe is replaced with RNase free water. 28. Software ideal to identify good reference genes can be downloaded from: geNorm – http://medgen.ugent.be/~jvdesomp/genorm/ NormFinder – http: //www.mdl.dk/publicationsnormfinder. htm

Acknowledgements Work in the authors’ laboratories has been supported by The Norwegian Council on Cardiovascular Disease, The Research Council of Norway, Anders Jahre’s Foundation for the promotion

192

Brattelid and Levy

of Science, The Novo Nordisk Foundation, The Family Blix Foundation and Shipowner Tom Wilhelmsen’s Foundation. Anne Vatland Krøvel is acknowledged for critical reading of the manuscript. References 1. Bustin, S. A., Benes, V., Garson, J. A., Hellemans, J., Huggett, J., Kubista, M., Mueller, R., Nolan, T., Pfaffl, M. W., Shipley, G.L., Vandesompele, J. and Wittwer, C. T. (2009) The MIQE guidelines: minimum information for publication of quantitative real-time PCR experiments. Clin. Chem. 55, 611–622. 2. Brattelid, T., Tveit, K., Birkeland, J. A., Sjaastad, I., Qvigstad, E., Krobert, K. A., Hussain, R. I., Skomedal, T., Osnes, J. B. and Levy, F. O. (2007) Expression of mRNA encoding G protein-coupled receptors involved in congestive heart failure – a quantitative RT-PCR study and the question of normalisation. Basic Res. Cardiol. 102, 198–208. 3. Amisten, S., Braun, O. O., Bengtsson, A. and Erlinge, D. (2008) Gene expression profiling for the identification of G-protein coupled receptors in human platelets. Thromb. Res. 122, 47–57. 4. Haitina, T., Olsson, F., Stephansson, O., Alsiö, J., Roman, E., Ebendal, T., Schiöth, H. B. and Fredriksson, R. (2008) Expression profile of the entire family of Adhesion G protein-coupled receptors in mouse and rat. BMC Neurosci. 9:43. 5. Moore-Morris, T., Varrault, A., Mangoni, M. E., Le Digarcher, A., Negre, V., Dantec, C., Journot, L., Nargeot, J. and Couette, B. (2009) Identification of potential pharmacological targets by analysis of the comprehensive G protein-coupled receptor repertoire in the four cardiac chambers. Mol. Pharmacol. 75, 1108–1116. 6. Imbeaud, S., Graudens, E., Boulanger, V., Barlet, X., Zaborski, P., Eveno, E., Mueller, O., Schroeder, A. and Auffray, C. (2005) Towards standardisation of RNA quality assessment using user-independent classifiers of microcapillary electrophoresis traces. Nucleic Acids Res. 33(6):e56. 7. Fleige, S. and Pfaffl, M. W. (2006) RNA integrity and the effect on the real-time qRT-PCR performance. Mol. Aspects Med. 27, 126–139. 8. Auer, H., Lyianarachchi, S., Newsom, D., Klisovic, M. I., Marcucci, G. and Kornacker, K. (2003) Chipping away at the chip bias:

RNA degradation in microarray analysis. Nat. Genet. 35, 292–293. 9. Nolan, T., Hands, R. E., Ogunkolade, W. and Bustin, S. A. (2006) SPUD: a quantitative PCR assay for the detection of inhibitors in nucleic acid preparations. Anal. Biochem. 351, 308–310. 10. Ståhlberg, A., Kubista, M. and Pfaffl, M. (2004) Comparison of reverse transcriptases in gene expression analysis. Clin. Chem. 50, 1678–1680. 11. Ståhlberg, A., Håkansson, J., Xian, X., Semb, H. and Kubista, M. (2004) Properties of the reverse transcription reaction in mRNA quantification. Clin. Chem. 50, 509–515. 12. Stangegaard, M., Dufva, I. H. and Dufva, M. (2006) Reverse transcription using random pentadecamer primers increases yield and quality of resulting cDNA. Biotechniques 40, 649–657. 13. Resuehr, D. and Spiess, A. N. (2003) A realtime polymerase chain reaction-based evaluation of cDNA synthesis priming methods. Anal. Biochem. 322, 287–291. 14. Bustin, S. A. (2000) Absolute quantification of mRNA using real-time reverse transcription polymerase chain reaction assays. J. Mol. Endocrinol. 25, 169–193. 15. Brattelid, T., Winer, L.H., Levy, F.O., Liestøl, K., Sejersted, O.M. and Andersson, K.B. (2010) Reference gene alternatives to Gapdh in rodents and human heart failure gene expression studies. BMC Mol. Biol. 11:22. 16. Andersen, C. L., Jensen, J. L. and Ørntoft, T. F. (2004) Normalisation of real-time quantitative reverse transcription-PCR data: a modelbased variance estimation approach to identify genes suited for normalisation, applied to bladder and colon cancer data sets. Cancer Res. 64, 5245–5250. 17. Vandesompele, J., De Preter, K., Pattyn, F., Poppe, B., Van Roy, N., De Paepe, A. and Speleman F. (2002) Accurate normalisation of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 3(7): RESEARCH0034.

Quantification of GPCR mRNA Using Real-Time RT-PCR 18. Pfaffl, M. W., Horgan, G. W. and Dempfle, L. (2002) Relative expression software tool (REST) for group-wise comparison and statistical analysis of relative expression results in real-time PCR. Nucleic Acids Res. 30(9):e36. 19. Muller, P. Y., Janovjak, H., Miserez, A. R. and Dobbie, Z. (2002) Processing of gene expression data generated by quantitative real-time RT-PCR. Biotechniques 32, 1372–1379. 20. Hellemans, J., Mortier, G., De Paepe, A., Speleman, F. and Vandesompele, J. (2007) qBase relative quantification framework and software for management and automated ­analysis

193

of ­real-time quantitative PCR data. Genome Biol. 8(2):R19. 21. Pfaffl, M. W., Vandesompele. J. and Kubista. M. (2009) Data Analysis Software, in RealTime PCR: Current Technology and Applications. (Logan, J., Edwards, K. and Saunders, N., eds.) Caiser Academic Press, pp. 65–83. 22. Pfaffl, M. W. (2001) A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res. 29(9):e45. 23. Pfaffl, M. W. (2004) Quantification strategies in real-time PCR, in A-Z of Quantitative PCR (Bustin, S.A., ed) IUL Biotechnology Series, International University Line, pp. 87–120.

wwwwwww

Chapter 10 Determining Allosteric Modulator Mechanism of Action: Integration of Radioligand Binding and Functional Assay Data Christopher J. Langmead Abstract The drive to produce safer and more receptor subtype selective drugs has sparked a renewed interest in allosteric modulators of G protein-coupled receptors. The increasing use of functional assays has shown that allosteric ligands are capable of modulating both orthosteric agonist affinity and efficacy, as well as mediating receptor activation in their own right. Such a complex range of behaviours can be difficult to discern from single datasets; this chapter seeks to explain how to use radioligand binding and functional assay datasets in concert to elucidate allosteric modulator mechanism of action. Key words: Allosteric, Orthosteric, Binding, Affinity, Efficacy, Potency

1. Introduction G protein-coupled receptors (GPCRs) mediate responses to a wide range of stimuli, from small molecule neurotransmitters and hormones to peptides and photons of light. Their wide expression profile and roles in major physiological functions have made them prime targets for marketed drugs (1). Most drugs targeting GPCRs, whether agonists or antagonists, interact with the same binding site as that for the endogenous ligand for the receptor – the ‘orthosteric’ binding site. The location of this site varies across the receptor family: In Class A monoamine receptors such as muscarinic acetylcholine or dopamine receptors, it is usually located within the transmembrane domain bundle (2). For Class C receptors, such as the gamma amino butyric acid (GABAB) or metabotropic glutamate (mGlu) receptors, the orthostericbinding site is formed by the large N-terminal domain (3).

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_10, © Springer Science+Business Media, LLC 2011

195

196

Langmead

Historically, high-throughput screens for GPCRs have been run using radioligand-binding assays employing a radiolabelled version of the endogenous ligand or a synthetic competitive antagonist. Such an assay format inherently biased the resultant hits towards compounds that interacted with the same site as the radioligand. However, the advent of functional assays as the screening paradigm of choice in the past decade has seen an increase in the detection of compounds that interact with allosteric sites on GPCRs. Allosteric ligands (from the Greek allos, meaning other and stereos, meaning shape) can bind to sites on GPCRs that are topographically distinct from the orthosteric site such that the receptor is able to accommodate two ligands simultaneously. Allostericbinding sites have been identified on many GPCRs, including adenosine (4), muscarinic acetylcholine (5), dopamine (6), chemokine (7), calcium sensing (8), and mGlu (9) receptors. Upon binding, allosteric ligands are able to alter the conformation of the orthosteric site to modify orthosteric agonist activity as described below. Allosteric modulators possess a number of advantages as potential drugs. By virtue of being the binding site for the endogenous agonist or transmitter, orthosteric-binding sites are likely to be subject to evolutionary pressure to remain conserved across receptor families; however, allosteric-binding sites are unlikely to be subject to such pressure and are likely to display divergence within a receptor family. This offers the prospect of selectively targeting individual receptor subtypes, where it has not yet been possible to design orthosteric drugs of sufficient selectivity to avoid unwanted side effects e.g. subtype-selective drugs for muscarinic acetylcholine receptors. In the late 1990s, the pharmaceutical industry switched from using radioligand binding screening assays to functional assays as a result of the availability of high-throughput generic signalling assays. This switch enabled allosteric ligands to be more routinely identified; any compound that perturbed the action of an agonist (or activated the receptor in its own right) could be detected, irrespective of the location of its binding site on the receptor. In the past 5 years, the pharmaceutical industry has seen the approval for market of its first two allosteric modulators of GPCRs: Cinacalcet, a positive allosteric modulator of the calcium-sensing receptor, was approved in 2004 for hyperparathyroidism (10) and in 2007, maraviroc, a negative allosteric modulator of the chemokine receptor CCR5, was approved as an HIV entry inhibitor (11). Many companies are now actively pursuing allosteric modulators of GPCRs as novel therapies for a whole range of disease indications. It is now apparent that allosteric modulators can display a wide variety of molecular behaviours, including positive, negative or neutral modulation of ligand affinity and/or efficacy as well as the ability to activate (or inactivate) a receptor in their own right (12). With such a complex array of potential effects comes the challenge of quantification. This chapter addresses how to understand

Determining Allosteric Modulator Mechanism of Action

197

and apply some of the most recent models of allosterism to everyday functional and radioligand-binding datasets to quantitatively define the mechanism of action of an allosteric ligand. The example cited is that of an allosteric modulator (Compound 1) of a muscarinic receptor in [3H]-N-methyl scopolamine (NMS) radioligand binding and GTPg[35S] functional assays. Data presented were analysed in GraphPad Prism v5 (GraphPad Software, CA, USA). Readers who wish to learn more about analysing data using GraphPad Prism should refer to the excellent handbook and Web site ((13); http://www.graphpad. com). To aid the reader in applying the protocols covered, equations have been given in ‘GraphPad Prism-ready’ format (Fig. 1); a KA=10^LogKA KB=10^LogKB Alpha=10^LogAlpha A=10^LogHot End=Start*((A+KA)/(A+KA/Alpha)) LogBmid=Log(KB*((A+KA)/(Alpha*A+KA))) Y=End+(Start-End)/(1+10^((X-LogBmid)*S))

b I=10^X Part1=(10^LogKA*10^LogKB)/(10^LogAlphaA*B+10^LogKB) Part2=1+(I^S/10^LogKI) Part3=B/10^LogKB Part4=(10^LogAlphaB*(I^S)*B)/(10^LogKI*10^LogKB) KApp=Part1*(Part2+Part3+Part4) Y=(100*(10^LogHot+10^LogKA))/(10^LogHot+KApp)

c

KB=10^LogKB EC50=10^LogEC50 tauB=10^LogtauB A=10^X ab=10^LogAlphaBeta Part1=(A*(KB+ab*B)+tauB*B*EC50)^n Part2=(EC50^n)*((KB+B)^n) Span=Em-Basal Y=Basal+(Span*Part1)/(Part1+Part2)

Fig. 1. Allosteric models are described in this chapter. (a) Allosteric ternary complex model (ATCM). KA and KB represent equilibrium dissociation constants for orthosteric and allosteric ligands, A and B, respectively. a represents the affinity cooperativity factor. (b) An extended ternary complex model accounting for the interaction of a receptor with one allosteric and two orthosteric ligands. Parameters are as for (a); in addition, KI represents the equilibrium dissociation constant for the second orthosteric ligand, I. a represents the affinity cooperativity factor governing the interaction between A and B; a¢ represents the affinity cooperativity factor governing the interaction between I and B. S represents an empirical slope factor. The equation as presented requires specific binding to be normalised as a percentage. (c) Operational ATCM accounting for the ability of an allosteric ligand to modify affinity and/or efficacy and even activate the receptor in its own right. The equation describes a simplified model for positive allosteric modulators, where the orthosteric ligand is a full agonist. KB represents the equilibrium dissociation constants for the allosteric ligand, Basal is the response in the absence of ligand, EC50 is the midpoint of the full agonist concentration-response curve, tB denotes the capacity of the allosteric modulator to exhibit agonism (a function of the intrinsic efficacy and receptor expression), and ab represents a net affinity/efficacy cooperativity parameter which describes the effect of the putative allosteric ligand on agonist function. The terms EM and n denote the maximal possible system response and the slope factor of the transducer function that links occupancy to response, respectively.

198

Langmead

these could be easily adapted to any curve-fitting software capable of global fitting. Datasets for each example are included in the figures to allow the reader to understand how to enter and process data in the software.

2. RadioligandBinding Studies In order to design and analyse experiments to determine allosteric modulator mechanism of action, it is necessary to understand the macromolecular basis of the interaction of allosteric ligands with a GPCR. The simplest model of a three-way inter­ action between a GPCR, an orthosteric ligand and an allosteric modulator is the allosteric ternary complex model (ATCM) proposed by Stockton et al. (14) which is illustrated in Fig. 1a. Both an orthosteric ligand (A) and an allosteric ligand (B) can interact with a receptor (R), forming either binary species (AR, BR) or a ternary complex (ARB). KA and KB denote the equilibrium dissociation constants of AR and BR, respectively. The symbol a denotes the cooperativity factor and is a quantitative measure of the maximal, reciprocal alteration of affinity of A and B for their respective binding sites when both ligands bind simultaneously to form the ternary complex, ARB. In short, the value of a dictates whether the allosteric ligand (B) has a positive or negative effect on the binding of the orthosteric ligand (A) and vice versa. Values of a > 1 denote positive allosteric modulation of affinity, whereas values of 0 < a < 1 denote negative allosteric modulation. A value of a = 1 represents a special situation, whereby the binding of the allosteric modulator does not alter the affinity of the orthosteric probe and vice versa. The effect of allosteric modulators on the binding of a radioligand is shown in Fig. 2. A positive allosteric modulator produces a concentration-dependent, saturable increase in the affinity (Fig. 2a) of the radioligand, which is more readily visualised on a semi-logarithmic plot (Fig.  2c). Conversely, a negative allosteric modulator produces a concentration-dependent, saturable decrease in the affinity of the radioligand, similarly represented in Fig. 2b, d. It is possible to analyse such datasets according to the ATCM to generate estimates of the affinity of the allosteric modulator (KB) and the cooperativity (a) between the modulator and the orthosteric radioligand probe. However, to run what is in effect a Schild analysis using a radioligand-binding assay is unusual and not particularly cost-effective. It is much more common to use a single concentration of a radioligand and run a titration curve of a test compound (often referred to as ‘competition binding’ as historically such assays were used to screen analogues of endogenous ligands). The effects of both positive and negative allosteric modulators in such an assay design are shown in Fig. 2e, f. Note that a

199

Determining Allosteric Modulator Mechanism of Action

a

b

120

80 60 40 20 0

c

100 % specific binding

% specific binding

100

0

2

4 Radioligand (nM)

6

20 0

2

4 Radioligand (nM)

6

8

−10 −9 −8 Log [radioligand] (M)

−7

100 % specific binding

% specific binding

40

120

80 60 40 20 0 −11

−10 −9 −8 Log [radioligand] (M)

f

200 150 100 50

−10

−9

−8 −7 −6 −5 Log [Modulator] (M)

−4

80 60 40 20 0 −11

−7

% control specific binding

% control specific binding

60

d

120

0

80

0

8

100

e

120

200 150 100 50 0

−10

−9

−8 −7 −6 −5 Log [Modulator] (M)

−4

Fig. 2. (a) Effect of a range of concentrations of positive allosteric modulator (KB = 1 mM; a = 3) on the saturation binding of a radioligand (KD = 1 nM) according to the ATCM; (c) transformed into a semi-log plot. (b) and (d) show the same for a negative allosteric modulator (KB = 1  mM; a = 0.3). Titration curves are also shown for the positive (e) and negative (f ) allosteric modulators on the equilibrium binding of a single concentration (1 × KD) of the same radioligand.

positive modulator yields an enhancement of the level of bound radioligand, whereas the negative modulator inhibits the binding of the radioligand. These curves can also be analysed to quantify the affinity and cooperativity of the modulator. It is important to note that the negative modulator is unable to fully inhibit the specific binding of the radioligand. This effect (and the definition of the maximum asymptote for the positive allosteric modulator) is key to determining the degree of cooperativity between the mod-

200

Langmead

ulator and the radioligand. If the negative modulator were to fully inhibit the specific binding of the radioligand, it would be impossible to distinguish such a compound from a simple competitive antagonist (as the degree of negative cooperativity, a → 0). If a compound is truly a negative allosteric modulator, this can be resolved by evaluating the test compound against multiple, higher concentrations of radioligand (e.g. 1 × KD, 10 × KD). Example 1  Equilibrium [3H]-NMS Binding Figure 3 shows an example of the effect of an allosteric modulator (Compound 1) on the equilibrium binding of [3H]-NMS to a muscarinic receptor. The data from the experiment are shown in the table, already normalised to ‘% specific binding’. These data can be analysed according to the ATCM (entered as shown in Fig. 1a); note from the screenshot that the values of Log KA (the logarithm of the equilibrium dissociation constant for [3H]-NMS) and Log Hot (the logarithm of the molar concentration of [3H]-NMS) are set as fixed values. Also note that appropriate estimates of some parameters need to be entered in the ‘Initial values’ tab before the model fits correctly. As can been seen from the graph, Compound 1 very modestly inhibits the specific binding of the radiolabel. Analysis yields an estimate of affinity (KB) and cooperativity (a) with respect to [3H]-NMS as shown in the results table (Fig. 3). As would be expected from the shape of the graph, Compound 1 has moderate affinity (KB = 2.5  µM) with very weak negative cooperativity (a = 0.75); hence, Compound 1 is a weak negative allosteric modulator of [3H]-NMS binding. However, testing allosteric ligand in this fashion only works if the orthosteric ligand that you are studying can be used as a radiolabel. More often than not, researchers want to know the effect of an allosteric modulator on the endogenous agonist and for GPCRs, where the endogenous ligand is of relatively low affinity or labile e.g. acetylcholine at muscarinic receptors, this is impractical. Therefore, it is common for the radioligand of choice to be a high-affinity competitive antagonist. However, one of the hallmarks of allosteric interactions is that although the affinity of an allosteric modulator for a receptor is an intrinsic property of the molecule, its cooperativity is a property unique to the orthosteric/ allosteric ligand pair. In practice, this phenomenon of ‘probedependence’ can result in misleading results if the orthosteric radiolabel is not the endogenous ligand – for example, staurosporine is a positive allosteric modulator of [3H]-NMS binding at the M1 muscarinic receptor but a negative modulator of acetylcholine binding (15).

Determining Allosteric Modulator Mechanism of Action

201

% specific [3H]NMS binding

120 100

LogKA LogKB LogAlpha LogHot Start Slope

80 60 40 20

KB

0

α

−7

−6

−5

= −9.73 −5.63 −0.13 =−9.66 104 2.36 2.3 µM 0.75

−4

Log [Compound 1] (M)

Fig. 3.  Effect of Compound 1 on equilibrium binding of [3H]-NMS and analysis according to the ATCM (Fig. 1a) in GraphPad Prism. Also shown are the binding data as entered in the Prism data table and the parameter constraints. Note that values of KB and a are fitted as logarithms and that values of KA (equilibrium dissociation constant for [3H]-NMS) and Hot (concentration of [3H]-NMS) are set to fixed values.

Example 2  Three-Way Equilibrium [3H]-NMS Binding In cases where the agonist is not available as a radiolabel, it is still possible to determine the effect of an allosteric modulator on an unlabelled orthosteric ligand, provided a radiolabel is available. In this case, as shown in Fig. 4, the effect of multiple concentrations of Compound 1 on the inhibition of [3H]-NMS binding by acetylcholine is studied in a radioligand-binding experiment. This simply requires multiple concentrationresponse curves to the agonist in the absence and presence of varying concentrations of the allosteric modulator. Note that this experiment was performed in the presence of GTP to ensure all binding was to the low agonist affinity state of the receptor. We already know from Example 1 that Compound 1 is a weak negative modulator of [3H]-NMS binding, but from Fig. 4 it can be seen that it produces a saturable, leftward shift in the concentration-response curve to acetylcholine, suggesting that Compound 1 is increasing the affinity of acetylcholine to inhibit [3H]-NMS binding. Therefore, Compound 1 is a positive allosteric modulator of acetylcholine affinity. In order to quantify this effect, an extension to the ATCM is required to account for one allosteric ligand and two orthosteric ligands; this is shown in Fig. 1b. As there is a family of curves, this dataset needs to be fitted globally, whereby all the curves (continued)

202

Langmead

Example 2  (continued) are fitted simultaneously to find the best fit for the parameters (13). In this case the main parameters in question are the affinities of acetylcholine (KI), 3H-NMS (KA), and Compound 1 (KB) and the affinity cooperativity of Compound 1 with respect to [3H]-NMS (a) and acetylcholine (a1). Figure  4 shows the analysis of the dataset according to this model. The data from the experiment are shown in the table, already normalised to ‘% specific binding’. As the analysis is by ‘global ­fitting’ (where the family of curves are fitted together), the molar concentration of Compound 1 is entered as the column heading in the data table (including a value of zero for the control curve). Also shown is the screenshot of the ‘Constrain’ tab from Prism; note that the concentration of Compound 1 (B) is set to be equal to the column heading. The values of Log KA, Log Hot, and slope (S) are fixed, the latter to unity. All other values are set to be shared across all of the datasets, i.e. globally fitted. It is also important to ensure that the estimates provided in the ‘Initial values’ tab are meaningful – complex equations often fail to fit due to inappropriate initial parameter estimates. The results table shows that Compound 1 has relatively low affinity for the receptor with an estimated KB value of 4.1 mM. Compound 1 is a positive allosteric modulator of acetylcholine binding (a = 9.7), despite being a weak negative allosteric modulator of 3H-NMS binding (a = 0.88). This once again reinforces the phenomenon of allosteric probe-dependence. If the analysis fails to fit correctly (either the curves do not go through the data points sufficiently well or the parameter estimates and confidence intervals are not meaningful), then there are a number of steps that can be taken. Firstly, the initial fitting values of the parameters should be checked to make sure that they are appropriate (Is the initial value for the Log KB value physiologically meaningful? e.g. between −10 and −3). If this does not help, then it may be necessary to constrain one or more of the parameters to a value that has been previously determined. In this example, the value of a for Compound 1 and [3H]-NMS could be constrained to 0.75 from the ­original experiment in Fig. 3.

3. Functional Assays As stated in the introduction, the use of functional assays has increased greatly in the past decade, often to the detriment of running ligand-binding assays. The most common method for identifying an allosteric interaction is to determine the effect of an

Determining Allosteric Modulator Mechanism of Action

203

% Specific [3H]NMS binding

120 100

LogKA

Control

80

+ 2 µM Compound 1 + 6.3 µM Compound 1 + 20 µM Compound 1

60 40 20 0 −7

−6

−5 −4 −3 Log [Acetylcholine] (M)

−2

−1

= −9.73

LogKB

−5.39

LogAlphaA

−0.055

B

variable

S

= 1.0

LogKI

−4.68

LogAlphaB

0.99

LogHot

= −9.65

KB α

4.1µM

α’

9.7

0.88

Fig. 4. Effect of Compound 1 on acetylcholine-mediated inhibition of [3H]-NMS binding. Analysis was performed ­according to an extended ATCM (Fig. 1b) in GraphPad Prism; also shown are the data as entered as % specific binding in the Prism data table, the parameter constraints, and initial parameter estimates. Note that values of KB, KI, a (Alpha A), and a´ (AlphaB) are fitted as logarithms and that values of KA (equilibrium dissociation constant for [3H]-NMS), Hot (concentration of [3H]-NMS), and S (slope factor) are set to fixed values.

allosteric modulator on a functional concentration-response curve to an agonist. This is actually the simplest method for qualitatively detecting an allosteric effect. However, due to the additional complexities of functional assays, the quantitative analysis can be rather difficult. Previously, the leftward or rightward shift of an agonist concentration-response curve in a functional assay would have been taken as presumptive evidence of a simple ATCM mechanism

204

Langmead

and analysed accordingly (i.e. changes in agonist potency were taken simply to represent an increase or decrease in agonist affinity). However, it is now commonly recognised that allosteric modulators, in addition to modulation of agonist affinity, can modulate agonist efficacy and even activate receptors in their own right (allosteric agonism; (12)). For example, SIB1893 increases both the potency and the maximal response to L-AP4 in a GTPg[35S] binding assay at mGluR4 (16). Such an effect on maximal agonist response cannot be accommodated by a model which considers only effects of allosteric ligands on orthosteric agonist affinity. To accommodate this range of behaviours, the ATCM has been extended to include the operational model of agonism (17). The operational model effectively dissects agonist potency into its component parameters, affinity (K ) and efficacy (t, tau). The term t is a cell or tissue-dependent expression of the ability of the agonist–receptor complex to produce a stimulus. In this extended ATCM, the ‘efficacy cooperativity’ parameter is neither thermodynamic nor bidirectional, but simply represents the ratio of orthosteric agonist efficacies (t and bt) in the absence and presence of the allosteric modulator, respectively; the ratio (b) represents the efficacy cooperativity factor. The model also accounts for the ability of the allosteric ligand to mediate receptor activation in its own right (tB; (18)). It is a commonly held belief that changes in agonist efficacy result in an increase or decrease in the maximal agonist response. Unfortunately, variations in receptor reserve and/or the maximal assay response mean that this is rarely the case. For an assay with a very low receptor reserve (e.g. native tissue assay), the maximal agonist response may be well below the maximal response of the system; hence, the maximal agonist response could be increased by a positive modulator of efficacy. However, in an assay with high receptor reserve, where the maximal agonist response may well already have reached the ‘ceiling,’ a positive modulator would yield a leftward shift of the agonist concentration-response curve. Example 3  Functional GTPγ[35S]-Binding Assay In order to determine the effect of Compound 1 on acetylcholine function, the effects of varying concentrations of Compound 1 on acetylcholine-stimulated GTPg[35S] binding were measured (Fig. 5). As can be seen from the dataset, Compound 1 causes a concentration-dependent leftward shift in the acetylcholine concentration-response curve. Additionally, Compound 1 yields a small increase in GTPg[35S] binding in its own right, as evidenced by the increased basal response in the absence of acetylcholine. These data were analysed according to a modified version of the operational model of allosterism which accounts for allosteric modulation of affinity, efficacy, and allosteric agonism (Fig. 1c; (18, 19)). In this model, it is (continued)

Determining Allosteric Modulator Mechanism of Action

205

Example 3  (continued) assumed that the maximal agonist response in the assay under test reaches the maximal system response, i.e. it cannot be potentiated. This is usually the case for a full agonist in a recombinant system. The model estimates a ‘net cooperativity’ factor (ab) which encompasses any effects on agonist affinity and efficacy. The data from the experiment are shown in the table, expressed as bound GTPg[35S] in cpm. As before the molar concentration of Compound 1 is entered as the column heading in the data table (including a value of zero for the control curve). Also shown is the screenshot of the ‘Constrain’ tab from Prism; note that the concentration of Compound 1 (B) is set to be equal to the column heading. All other values are set to be shared across all of the datasets, i.e. globally fitted. As in Example 2, it is important to ensure that the estimates provided in the ‘Initial values’ tab are meaningful. The analysis shows that Compound 1 has relatively low affinity for the receptor with an estimated KB value of 3.2 mM, which is in good agreement with the value generated in the radioligand-binding study. Furthermore, Compound 1 is a positive allosteric modulator of acetylcholine function, yielding an estimate of the net affinity/efficacy cooperativity of ab = 4.8. By simple division, it is possible to calculate the efficacy cooperativity factor (b) using the values of ab (4.8) and a (9.7) generated from the functional and radioligand-binding studies, respectively. For Compound 1, this value is calculated as b = 0.5, suggesting that Compound 1 is actually a weak negative modulator of acetylcholine efficacy. Clearly, the overall allosteric effect is positive, as the effects of Compound 1 on affinity outweigh those on efficacy, but in the absence of any radioligand-binding data, it would be impossible to determine whether this positive modulation was due to an increase in affinity or efficacy of acetylcholine. Finally, the analysis also is able to quantify the degree of agonism displayed by Compound 1; in this the value of t is 0.22, indicative of weak partial agonism. Note that due to the cell/tissue dependence of the parameter, it is only possible to compare values of t for the same agonist in different assay systems, or different agonists in the same assay. If the model fails to fit correctly (either the curves do not go through the data points sufficiently well or the parameter estimates and confidence intervals are not meaningful), then a similar set of checks should be applied as previously. Are the initial fitting values of the parameters meaningful? If this does not help, then it may be necessary to constrain one or more of the parameters to a value that has been previously determined. In this example, the value of KB for Compound 1 could be constrained to that determined in the radioligand-binding study.

206

Langmead

Control + 1 µM Compound 1 + 3 µM Compound 1 + 10 µM Compound 1 + 30 µM Compound 1

Bound GTPγ [35S] (cpm)

8000

6000

4000

2000

Basal

−8

−7 −6 −5 Log [Acetylcholine] (M)

LogKB

−5.50

LogEC50

−5.93

LogTauB

−0.66

B

variable

LogAlphaBeta

0.68

n

0.97

EM

7474

Basal

1848

KB

3.2 µM

α.β

4.8

τB

0.22

−4

Fig.  5. Effect of Compound 1 on acetylcholine-stimulated GTPg[35S] binding. Analysis was performed according to a simplified operational allosteric ternary complex model (Fig. 1c) in GraphPad Prism; also shown are the data as entered as bound cpm in the Prism data table, the parameter constraints, and initial parameter estimates. Note that values of KB, EC50, ab, and tB are fitted as logarithms.

It is clear that the analysis in Example 3 can only be applied to net positive modulators in systems where the agonist is full. However, there are more permissive models which account for a fuller range of behaviours (see the full operational model of

Determining Allosteric Modulator Mechanism of Action

207

allosterism; (18)). The problem with using more complex models is that as the number of parameters increases (to describe the greater complexity), so the inherent flexibility in the fitting increases. For these models, even apparently perfect datasets do not fit the model properly because an infinite combination of parameters can describe the dataset. Therefore, in such cases, it is a necessity to constrain the values of some parameters which have been determined independently. Often, the easiest way to do this is to generate estimates of affinity for orthosteric (KA) and allosteric (KB) ligands, affinity cooperativity (a) in radioligandbinding assays, and feed these values back into the analysis of functional datasets. This chapter already assumes that the binding and functional studies have been planned by adhering to good assay design principles (described elsewhere in the book). However, a specific consideration when combining data derived from different assays is that it is important to consider the buffer composition. For instance, most live cell functional assays run in a physiological salt solution, whereas it is not uncommon for radioligand-binding assays to run in Tris or HEPES buffer in the absence of salts. It is well known that ionic composition can have a marked effect on ligand binding, so to ensure that data is as meaningful as possible, radioligand-binding assays should be run in buffer that matches the corresponding functional assay as closely as possible.

4. Discussion This chapter has briefly highlighted the major effects that allosteric ligands can have on GPCR function. Overall, these effects can be complex, involving any combination of positive, neutral or negative modulation of orthosteric agonist affinity and efficacy. Additionally, allosteric ligands are capable of mediating receptor activation (or inactivation) in their own right. The qualitative effects of allosteric modulators are often best determined using functional assays – it is generally easy to see whether a putative modulator is having a positive or negative effect on agonist function. However, to understand how this is mediated at a quantitative level requires further investigation. It may be important to understand whether effects of a modulator can be attributed to changes in agonist affinity or efficacy; the former is a bi-directional thermodynamic event but efficacy is system-dependent. It could be that modulators of efficacy display different effects in recombinant systems compared to a native tissue environment. The example presented herein shows that Compound 1 is a positive modulator of acetylcholine-stimulated GTPg[35S] binding. Analysis according to an extended ATCM which encompasses

208

Langmead

the operational model yielded an estimate of the net affinity and efficacy cooperativity between the two ligands. By using radioligand-binding studies, it was possible to delineate the specific effects of Compound 1 as a strong positive modulator of acetylcholine affinity and, by simple arithmetic, a very weak negative modulator of acetylcholine efficacy. It is also important to reiterate the notion of probe-dependence in allosteric effects; in the example presented here, simply examining the effects of Compound 1 in the muscarinic receptor [3H]-NMS-binding assay is not sufficient; it should be studied with respect to the agonist in question. Compound 1 is a weak negative modulator of [3H]-NMS binding but a strong positive modulator of acetylcholine binding. References 1. Insel, P. A., Tang, C. M., Hahntow, I., and Michel, M. C. (2007) Impact of GPCRs in clinical medicine: monogenic diseases, genetic variants and drug targets, Biochim Biophys Acta 1768, 994–1005. 2. Kobilka, B. K. (2007) G protein coupled receptor structure and activation, Biochim Biophys Acta 1768, 794–807. 3. Pin, J. P., Galvez, T., and Prezeau, L. (2003) Evolution, structure, and activation mechanism of family 3/C G-protein-coupled receptors, Pharmacol Ther 98, 325–354. 4. Gao, Z. G., Kim, S. K., Ijzerman, A. P., and Jacobson, K. A. (2005) Allosteric modulation of the adenosine family of receptors, Mini Rev Med Chem 5, 545–553. 5. Conn, P. J., Jones, C. K., and Lindsley, C. W. (2009) Subtype-selective allosteric modulators of muscarinic receptors for the treatment of CNS disorders, Trends Pharmacol Sci 30, 148–155. 6. Schetz, J. A. (2005) Allosteric modulation of dopamine receptors, Mini Rev Med Chem 5, 555–561. 7. Leach, K., Charlton, S. J., and Strange, P. G. (2007) Analysis of second messenger pathways stimulated by different chemokines acting at the chemokine receptor CCR5, Biochem Pharmacol 74, 881–890. 8. Hu, J. (2008) Allosteric modulators of the human calcium-sensing receptor: structures, sites of action, and therapeutic potentials, Endocr Metab Immune Disord Drug Targets 8, 192–197. 9. Kew, J. N. (2004) Positive and negative allosteric modulation of metabotropic glutamate receptors: emerging therapeutic potential, Pharmacol Ther 104, 233–244.

10. Shoback, D. M., Bilezikian, J. P., Turner, S. A., McCary, L. C., Guo, M. D., and Peacock, M. (2003) The calcimimetic cinacalcet normalizes serum calcium in subjects with primary hyperparathyroidism, J Clin Endocrinol Metab 88, 5644–5649. 11. Fatkenheuer, G., Pozniak, A. L., Johnson, M. A., Plettenberg, A., Staszewski, S., Hoepelman, A. I., Saag, M. S., Goebel, F. D., Rockstroh, J. K., Dezube, B. J., Jenkins, T. M., Medhurst, C., Sullivan, J. F., Ridgway, C., Abel, S., James, I. T., Youle, M., and van der Ryst, E. (2005) Efficacy of short-term monotherapy with maraviroc, a new CCR5 antagonist, in patients infected with HIV-1, Nat Med 11, 1170–1172. 12. Langmead, C. J., and Christopoulos, A. (2006) Allosteric agonists of 7TM receptors: expanding the pharmacological toolbox, Trends Pharmacol Sci 27, 475–481. 13. Motulsky, H. and Christopoulos, A. (2003) Fitting Models to Biological Data using Linear and Nonlinear Regression. A practical guide to curve fitting, GraphPad Software Inc., San Diego, CA. 14. Stockton, J. M., Birdsall, N. J., Burgen, A. S., and Hulme, E. C. (1983) Modification of the binding properties of muscarinic receptors by gallamine, Mol Pharmacol 23, 551–557. 15. Lazareno, S., Popham, A., and Birdsall, N. J. (2000) Allosteric interactions of staurosporine and other indolocarbazoles with N-[methyl-(3) H]scopolamine and acetylcholine at muscarinic receptor subtypes: identification of a second allosteric site, Mol Pharmacol 58, 194–207. 16. Mathiesen, J. M., Svendsen, N., BraunerOsborne, H., Thomsen, C., and Ramirez, M. T. (2003) Positive allosteric modulation of the

Determining Allosteric Modulator Mechanism of Action human metabotropic glutamate receptor 4 (hmGluR4) by SIB-1893 and MPEP, Br J Pharmacol 138, 1026–1030. 17. Black, J. W., and Leff, P. (1983) Operational models of pharmacological agonism, Proc R Soc Lond B Biol Sci 220, 141–162. 18. Leach, K., Sexton, P. M., and Christopoulos, A. (2007) Allosteric GPCR modulators: ­taking advantage of permissive receptor

209

pharmacology, Trends Pharmacol Sci 28, 382–389. 19. Bennett, K. A., Langmead, C. J., Wise, A., and Milligan, G. (2009) Growth hormone secretagogues and growth hormone releasing peptides act as orthosteric super-agonists but not allosteric regulators for activation of the G protein Galpha(o1) by the Ghrelin receptor, Mol Pharmacol 76, 802–811.

wwwwwww

Chapter 11 Design and Use of Fluorescent Ligands to Study Ligand–Receptor Interactions in Single Living Cells Stephen J. Briddon, Barrie Kellam, and Stephen J. Hill Abstract The interaction of ligands with G protein-coupled receptors (GPCRs) has been traditionally studied using radiolabelled variants of receptor ligands. More recently, increased knowledge about the way in which GPCRs exist in a highly organised membrane environment has led to an interest in investigating receptor–ligand interactions in single cells. In addition, substantial improvements in imaging technology and an increase in the expense of radioactive waste disposal have resulted in an expansion in the use of fluorescent technologies. One major requirement for these methods is a suitable fluorescent ligand for the receptor of interest. The design of fluorescent ligands for GPCRs is complex, and has to take into account their pharmacological, photophysical, and also physicochemical properties. Here, we describe some basic considerations in the design of fluorescent GPCR ligands, including choice of pharmacological template, linker, and fluorophore. In addition, we describe basic protocols for determining the photophysical properties of the ligand and determining the cellular localisation of their interaction with the target receptors. Finally, we provide a basic protocol for using fluorescent GPCR ligands to quantify the number and diffusion of receptor–ligand complexes in small areas of the cell membrane. Key words: G protein-coupled receptors, Fluorescent ligand, Fluorophore, Confocal microscopy, Fluorescence correlation spectroscopy, Single cell, Ligand binding

1. Introduction 1.1. Background

Characterising the interaction of G protein-coupled receptors (GPCRs) with their ligands in terms of their affinity and efficacy is a mainstay of pharmacological analysis. Traditionally, the measurement of affinity has involved the use of radiolabelled receptor ligands in either saturation- or competition-binding assays (see Chapter 8). More recently, there has been a substantial increase in the use of fluorescent receptor ligands to study such receptor–ligand interactions (1–3, and references therein).

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_11, © Springer Science+Business Media, LLC 2011

211

212

Briddon, Kellam, and Hill

There are a number of reasons, both scientific and practical, underlying the shift towards fluorescent approaches in this area. It has become increasingly evident, for instance, that GPCRs exist within highly ordered membrane domains, in conjunction with their associated signalling proteins (4, 5). This has resulted in an increased need to study receptor–ligand interactions and functional responses at a single cell or sub-cellular level; something which is particularly amenable to fluorescent approaches. The introduction of a wider range of brighter and more photostable fluorescent labels, at the same time as major advances in detection technology means that fluorescence-based assays can be exquisitely sensitive. Finally, on a practical note, the use of fluorescent ligands circumvents the increasing difficulties and expense in the disposal of radioactive waste. The range of techniques which make use of fluorescent receptor ligands is continually expanding. An obvious primary use is in studying receptor localisation in both cells and tissue slices through fluorescence microscopy (6–9). Ligand–receptor interactions have been quantified using fluorescent ligands in conjunction with fluorescence polarisation (10–12) and Förster resonance energy transfer (13, 14) and the diffusion of receptor–ligand complexes has been determined using fluorescence correlation spectroscopy (FCS) and TIRF (6, 8, 15–18). It is evidently beyond the scope of this chapter to describe detailed methodology for all of the above approaches. Instead, we describe some basic generic protocols and guidelines for choosing and designing a fluorescent ligand, for the initial characterisation of its pharmacological and photophysical properties, as well as protocols for determining its binding properties at the single cell level using confocal microscopy. Finally, we provide a more specialised protocol for subcellular quantification of ligand–receptor interactions using FCS. 1.2. Considerations in Designing a Fluorescent GPCR Ligand

A fluorescent receptor ligand can be thought of as consisting of three distinct, but interdependent parts: the parent pharmacophore, the fluorophore, and the linker region connecting these two components. The requirements for each of these components are different, but should all be considered when designing a new ligand. For each combination considered, it is essential that the novel fluorescent ligand is treated as a new pharmacological entity, and its affinity, efficacy, and subtype selectivity requires careful characterisation. It should never be assumed, even in the case of larger peptide ligands, that adding a fluorescent label leaves the pharmacology of the compound unaltered. To obtain the required pharmacological and photophysical properties from a ligand, several factors need to be taken into account. These include the structure–activity relationship between the pharmacophore and the receptor of interest, the length and chemical composition of

Design and Use of Fluorescent Ligands

213

the linker between fluorophore and pharmacophore, and the chemical structure and photophysical properties of the fluorophore itself. Clearly, the optimal combination for a ligand depends on the target receptor and the nature of the assay in which it is used. The choice of the base pharmacophore is an important one. Careful study of the structure–activity relationship of a compound series at the receptor of interest is necessary in order to choose an appropriate position to add the linker/fluorophore combination. Most commercially available fluorophores have molecular weights in the 400–1,000  Da range, and therefore can substantially increase the size of the pharmacophore, which may well increase the pharmacological effect of such a conjugation. Larger peptides, for instance, may well be less affected than small molecule ligands for aminergic receptors. However, such assumptions should always be tested empirically, as even in such cases positioning of the label on an amino acid residue essential for binding may still cause a decrease in binding affinity (19). The choice of linkage point is therefore influenced by known tolerance for substitution in such a position (with respect to binding affinity and/or efficacy), but may also be affected by the access to medicinal chemistry expertise. Free amine groups, for instance, allow single step coupling to amine-reactive fluorophores, which are generally more widely commercially available. It should also be noted that an increasing number of fluorescent GPCR ligands are now available commercially. The choice of fluorophore is determined primarily by the suitability of its photophysical properties (such as excitation and emission spectra, quantum yield, and fluorescence lifetime) for the required assay. A somewhat bewildering array of fluorophores is now commercially available, covering the full range of the ultraviolet, visible, and infrared spectrum (e.g. the BODIPY™, AlexaFluor™, and cyanine dye series). The optimal excitation– emission combination is influenced both by available light sources and optical set-up (e.g. lasers, filter sets) and also requirements for multi-colour applications (e.g. if using the ligand in conjunction with a GFP-tagged receptor). In addition, for live cell applications, the choice of excitation wavelength can have serious consequences for cell viability and the generation of autofluorescent background when imaging (20, 21). Both of these are generally minimised at longer wavelengths, and are particularly pronounced if ultraviolet wavelengths are used. More specific photophysical properties may also be required, such as spectral overlap for FRET applications, low triplet state population for FCS, or a long fluorescence lifetime for time-resolved applications. For live cell applications in particular, the propensity for the fluorophore to be quenched in aqueous media may also be a consideration (22). This information is often not readily available and

214

Briddon, Kellam, and Hill

may need to be determined experimentally. It should also be noted, however, that there are a number of studies showing that the fluorophore has a substantial effect on the pharmacology of the ligand, e.g. for the adenosine-A1, 5-HT2C, and 5-HT3 receptors (Fig. 1a) (22–24). Finally, the physicochemical properties of the fluorescent label, in combination with those of the linker, should be considered. Many fluorophores, such as the BODIPY dyes, are hydrophobic in nature, and may increase partitioning into cell membranes or increase non-specific binding. They may also increase the amount of intracellular ligand which is seen. Many hydrophobic fluorophores are also often heavily quenched in aqueous environments, which can be advantageous in distinguishing membrane bound from free ligand (22). Other fluorophores,

Fig. 1. Effect of fluorophore and linker length on the potency and efficacy of fluorescent adenosine-A1 receptor agonists. (a) Fluorescent derivatives of the agonist N-ethylcarboxamido adenosine (ABEA-X) were functionally characterised in CHO cells expressing the human adenosine-A1 receptor. Inhibition of forskolin-stimulated cAMP formation was assessed using a secreted alkaline phosphatase reporter gene, driven by a cyclase response element. Fluorophores EvoBlue30, Texas Red, Cy5, and BODIPY630/650 were each attached via the same linker to ABEA-X and functional responses determined. As shown, while ABEA-X-BY630 is a full, potent agonist, changing the fluorophore results in a substantial loss of potency (ABEA-X-Cy5, ABEA-X-Texas Red) and efficacy (ABEA-X-EvoBlue). (b) Structure of ABEA-X-BY630, showing the position of the variable chain alkyl linker (in which “n” refers to the number of carbon atoms in the linker). Inhibition of forskolin-stimulated cAMP production was assessed in response to increasing concentrations of ligands with differing chain lengths with alkyl linkers of n = 3, 4, 5, and 8 carbons. As demonstrated, potency, and efficacy are improved at shorter linker lengths.

Design and Use of Fluorescent Ligands

215

such as those structurally based on rhodamine, are known to increase non-specific binding to cell membranes. The final consideration in ligand design is the linker between the parent pharmacophore and the fluorophore, both in terms of its composition and its length. Linker length has been shown to affect both binding affinity and also efficacy of fluorescent ligands at the adenosine-A1 receptor (Fig. 1b) (15, 22). Again, the optimal linker length varies for each receptor, and more importantly, with each fluorophore (22). For instance, ligands for peptide receptors, where the binding region is on the extracellular N-terminus of the receptor, may not require linkers at all if coupled with a hydrophilic fluorophore. In contrast, a class A aminergic receptor, with a binding pocket buried within the transmembrane regions of the receptor, may require a longer linker to separate the fluorophore and pharmacophore, particularly if the fluorophore of choice is hydrophilic. Conversely, using a hydrophobic fluorophore in such instances may mean a shorter linker can be tolerated. Once the length of the linker has been optimised, further enhancements to affinity and efficacy may be possible by slight changes in the chemical composition of the linker, for instance by changing a saturated alkyl linker to a polyethyleneglycol-based chain of the same length to increase water solubility. Similarly, changes in the flexibility of the linker may be deliberately engineered for applications, such as fluorescence polarisation. 1.3. Synthesis and Purification of Fluorescent Ligands

As noted above, the level of medicinal chemistry input into synthesis and purification of a fluorescent ligand are dependent on the complexity of linkage being performed. With peptides, for instance, N-terminal labelling can be achieved with a simple single step reaction with an amine-reactive fluorophore, such as a succinimidyl ester derivative (11). Similarly, C-terminal labelling with a hydrazide derivative or labelling of unreduced cysteine residues with a thiol-reactive maleimide-conjugated fluorophore can be performed in a single step reaction. Consideration should be given in these cases to the availability of multiple labelling positions on a large peptide (for instance, any lysine residues may also react with the succinimidyl ester), or the importance of the N-terminus or chosen lysine in receptor interaction. Peptide ligands, particularly larger ones, have the advantage of being amenable to purification by size-exclusion chromatography. For small molecule ligands, it may also be possible to complete the synthesis in a single step reaction if a suitable functional group is available at the chosen linkage point (an example of this is given in Subheading 3.1). This can be aided by the availability of reactive fluorophore derivatives with “built-in” linkers (e.g. the aminohexanoyl “X”-linker on some BODIPY fluorophores) (Fig.  2). Where more complex linker strategies are required, it may be

216

Briddon, Kellam, and Hill

Fig. 2. Synthesis and spectral characterisation of XAC-X-BY630. (a) Synthetic scheme for synthesis of XAC-X-BY630 from XAC and the amine reactive fluorophore BODIPY630/650-X succinimydyl ester. (b) Excitation and emission spectra for XAC-X-BY630 (1 mM) in DMSO (solid lines) or HBSS (dashed lines). Left: Spectra normalised to the maximum emission obtained in DMSO, showing the substantial quenching of XAC-X-BY630 fluorescence in aqueous solution. Right : The same data with each curve normalised to its own maximum, showing a small shift of excitation and emission maxima to lower wavelengths in aqueous medium.

necessary to synthesise the pharmacophore from precursors in order to incorporate a suitable functionality in the correct position for linker attachment, or even to incorporate the linker during the synthesis (15). This obviously requires a greater degree of chemistry expertise and multiple reaction steps. In each case, it is generally considered sensible to make fluorophore addition the final step in the synthesis to prevent any chemical, temperature, or light-induced degradation. Integrity and identity of the final product can then be confirmed by using appropriate spectroscopic techniques e.g. nuclear magnetic resonance (NMR) or mass spectrometry (MS) for small molecule ligands produced through organic synthesis. The ligand should be purified to >99% using, for example, reverse phase HPLC. The existence of non-labelled ligand or unattached fluorophore in the final product can lead to misinterpretation of functional and imaging data. As noted above, it may be possible to achieve this purity with peptide ligands using size exclusion chromatography. However, HPLC provides a more

Design and Use of Fluorescent Ligands

217

robust way of ensuring the purity of the final product, via the purification of a single chromatographic peak. 1.4. Pharmacological Characterisation

Accurate pharmacological characterisation of any new fluorescent ligand is essential before interpreting any subsequent localisation or binding studies (see Subheading 4). However, the characterisation should include determining affinity and efficacy relative to the parent pharmacophore at the receptor of interest, preferably using at least two different functional responses. Affinity may be determined using radioligand binding, if a suitable receptor ligand is available. However, it is advisable that this is performed using a whole cell assay rather than purified membranes if subsequent live cell imaging is to be used, as this will most closely replicate the desired experimental conditions. Preferably, the initial characterisation should be performed on as clean a background as possible, such as a cell line expressing only the receptor subtype of interest. Further considerations are interference of the fluorophore in assays which use optical absorbance or luminescence and limited solubility of some ligands containing hydrophobic fluorophores, which can lead to an underestimation of responses at nominally high concentrations.

2. Materials 2.1. Fluorophore Coupling and Purification of XAC-X-BY630

1. Xanthine amine congener (XAC) [N-(2-aminoethyl)-2-(4(2,6-dioxo-1,3-dipropyl-2,3,6,7-tetrahydro-1H-purin-8-yl) phenoxy)acetamide] (Sigma-Aldrich, Poole, Dorset, UK). 2. N,N-Dimethylformamide (DMF) (Sigma-Aldrich). 3. BODIPY 630/650-X-succinimidyl ester (Invitrogen, Paisley, UK). 4. Round-bottomed flask (5 mL). 5. Aluminium foil. 6. High vacuum rotary evaporator and pump. 7. Reversed-phase high performance liquid chromatography (RP-HPLC) equipment suitable for gradient elution. 8. Solvents for RP-HPLC: (1) HPLC grade water (Fischer Scientific, Loughborough, UK) [solvent A]; (2) HPLC grade acetonitrile (Fischer Scientific) [solvent B]. Mobile phases are degassed by helium bubble [solvent A] and sonication [solvent B], respectively. 9. C8-reversed phase HPLC analytical or semi-preparative column (depending on the scale of synthesis).

2.2. Determination of Excitation and Emission Spectra

1. Stock solution of fluorescent ligand (in DMSO, 10  mM). Light sensitive. Store in aliquots in amber microcentrifuge tubes at −20°C. Freeze-thaw sensitivity to be determined on a

218

Briddon, Kellam, and Hill

ligand-by-ligand basis. Vortex thoroughly and ultrasonicate in a sonicating waterbath for 5 min prior to use (see Note 1). 2. Stock solution of unconjugated fluorophore (10  mM). Storage will be fluorophore-dependent, but generally store in aliquots at −20°C. Light sensitive. 3. HEPES-buffered saline solution (HBSS): 147 mM NaCl, 24 mM KCl, 1.3 mM CaCl2, 1 mM MgSO4, 1 mM Sodium pyruvate, 1 mM NaHCO3, 10 mM HEPES, pH 7.4 (see Note 2). 4. DMSO (see Note 3). 5. HPLC grade water, tested for low (e.g. Chromasolv® water, Sigma-Aldrich).

fluorescence

6. Clear flat-bottomed uv-transparent 96-well plate (Costar®, Sigma Aldrich). 7. Fluorescence plate reader capable of spectral scanning (e.g. Flexstation, Molecular Devices, Sunnyvale, CA, USA) (see Note 4). 2.3. Localising Ligand–Receptor Interactions in the Membrane Using Confocal Microscopy

1. Stock solution of fluorescent ligand (in DMSO, 10  mM) (as above). 2. Dulbecco’s Modification of Eagle’s Medium/Ham’s F-12 nutrient mix (Sigma Aldrich) supplemented with 10% foetal calf serum (PAA laboratories) and 2 mM L-glutamine? (see Note 5). 3. Trypsin/EDTA solution. 4. HBSS: 147 mM NaCl, 24 mM KCl, 1.3 mM CaCl2, 1 mM MgSO4, 1 mM sodium pyruvate, 1 mM NaHCO3, 10 mM HEPES, pH 7.4. Autoclave and store at 4°C. Add glucose at 10 mM (90 mg/50 mL) prior to use (see Note 6). 5. Nunc™ Labtek 8-well chambered coverglass (Nalgene Nunc, Thermo Fisher Scientific, Roskilde, Denmark) (see Note 7). 6. Confocal microscope (see Note 8). 7. Appropriate non-fluorescent receptor antagonist (10  mM stock) (see Note 9).

2.4. Quantifying Receptor–Ligand Interactions Using FCS

1. Rhodamine 6G (Sigma Aldrich) 10  mM in ethanol. Light sensitive, store at 4°C. 2. Cy5 (GE Healthcare, Chalfont St. Giles, Bucks, UK) 5 mM in low fluorescence water. Light sensitive. Store in aliquots at −20°C. 3. Water tested for low fluorescence (e.g. Chromasolv® water, Sigma Aldrich). 4. HBSS (as above). 5. Labtek 8-well chambered coverglass (as above).

Design and Use of Fluorescent Ligands

219

6. FCS-capable microscope (see Note 10). 7. Stock solution of fluorescent ligand at 10 mM (as above). 8. Appropriate analysis software (see Note 11).

3. Methods 3.1. Synthesis and Purification of XAC-X-BY630

1. In a 5 mL round bottomed flask (wrapped in aluminium foil to exclude light) dissolve XAC (5.0 mg, 11.7 mmol) in DMF (1 mL). 2. Prepare a solution of BODIPY 630/650-X-succinimidyl ester (5.0 mg, 7.57 mmol) in DMF (1 mL). 3. Add the DMF solution of BODIPY 630/650-X-succinimidyl ester to the aluminium foil-wrapped N,N-DMF solution of XAC, seal the round-bottomed flask and allow to stir in the dark for 2 h (Fig. 2a). 4. Remove the seal and the aluminium foil and transfer the round-bottomed flask to a high vacuum rotary evaporator and remove the DMF. 5. Dissolve the powdery residue in the minimum volume of a mixture of 65% solvent A and 35% solvent B (sonication will facilitate dissolution). 6. Purify the crude reaction material using RP-HPLC on a C8 reversed phase analytical or semi-preparative column. An optimal gradient for good peak separation on an analytical C8 RP-HPLC column (250 × 4.6  mm) is; 35% solvent B/65% solvent A to 100% solvent B/0% solvent A at a flow rate of 1 mL/min over a 30 min period (Approximate XAC-X-BY630 retention time, Rt = 12–13 min). 7. Collect the fractions containing the desired XAC-X-BY630 product and evaporate to dryness on a high-vacuum rotary evaporator.

3.2. Determination of Excitation and Emission Spectra

1. Thaw the stock solution of fluorescent ligand (10 mM), vortex thoroughly and incubate for 5 min in a sonicating waterbath to ensure that any aggregates are disrupted. 2. Prepare a 100 mM solution of ligand by initially adding 2 mL of stock solution to 198  mL of DMSO. Prepare further 10 and 1 mM solutions of ligand by serially diluting 100 mL of this solution to 900 mL of DMSO. Vortex the solutions thoroughly at each stage. 3. Repeat this dilution process using HBSS (with low autofluorescence) as a diluent. 4. Repeat steps 1–3 for the unconjugated fluorophore.

220

Briddon, Kellam, and Hill

5. Dispense 200 mL of each solution into a separate well of a uv-transparent 96-well plate. 6. Read the initial excitation spectra of the DMSO-based solutions from the appropriate wells using the plate reader. Select the emission wavelength to be measured as just above the peak for the parent fluorophore (e.g. for BODIPY 630/650, use 660 nm), and use an appropriate dichroic mirror to minimise the bleed-through of excitation light (e.g. cut off at 610  nm). Subsequently, read the emission intensity over a range of excitation wavelengths, ideally at 2–5 nm intervals (e.g. 400–580  nm). It is likely that you will not be able to capture the most red-shifted section of the excitation curve (580–630 nm) in this first pass due to the bandwidth of the dichroic mirror. This is particularly true if the fluorophore has a small Stoke’s shift (that is a small difference between excitation and emission wavelengths). However, since the emission spectra is not dependent on the excitation wavelength, it is possible to complete the excitation spectra by moving the emission collection to a more red-shifted wavelength (e.g. 720  nm), adjusting the dichroic mirror (e.g. cut off at 680  nm), then repeating the excitation scan to cover the overlap area between excitation and emission spectra (580– 650 nm) (see Note 12). 7. Using these initial data, choose an excitation wavelength just below that giving a peak signal (e.g. 610 nm). Again, using an appropriate dichroic (e.g. 630  nm cut-off), collect the emission intensities over a range of wavelengths around the peak emission (e.g. 660–750  nm). Repeating the scan with the excitation wavelength moved to a blue-shifted wavelength (e.g. 560 nm) should allow the blue-shifted part of the emission spectra to be collected (e.g. 600–660 nm). 8. Data from this initial scan should show that, while the emission intensities will be concentration-dependent, the peak excitation and emission wavelengths should be similar. These data allow judgement as to whether conjugation to the pharmacophore has altered the fluorescent properties of the fluorophore (Fig. 2b). 9. To assess the effect of an aqueous environment on the fluorophore emission, it is necessary to compare the data obtained in steps 6 and 7 with an equivalent concentration of fluorescent ligand/fluorophore in HBSS (to simulate the conditions of live cell experiments). Choose the concentration of fluorescent ligand and fluorophore which give the best signal to noise ratio. Repeat steps 6 and 7 on these two concentrations, but diluted in HBSS, ensuring that the sensitivity settings for detection are unaltered.

Design and Use of Fluorescent Ligands

221

10. A comparison of the peak fluorescence emission intensity of the DMSO sample vs. the HBSS-based sample gives an indication of (a) whether the excitation–emission maxima are altered and (b) whether the ligand is likely to be heavily quenched in an aqueous environment (Fig. 2b). 3.3. Confocal Imaging of Ligand–Receptor Interactions 3.3.1. Initial Characterisation of Fluorescent Ligand Binding

1. Initially, use a cell line which expresses your receptor of interest at a relatively high level on a reasonably clean background (e.g. Chinese Hamster Ovary cells or COS-7 cells) (see Note 13). 2. Functional characterisation of the ligand is essential prior to using confocal imaging to localise receptor–ligand interactions. In the case of a fluorescently labelled agonist, efficacy and potency relative to the parent pharmacophore should be determined in a range of functional assays, depending on the receptor being investigated. In the case of antagonist ligands, functional affinity measurements using Schild analysis (or single point KB estimates, if material is limiting) should be carried out. Radioligand-binding assays can also be used to determine ligand affinity (see Chapter 8). However, consideration should be given to developing whole cell binding assays for these experiments, since these replicate the conditions likely to be used during imaging experiments. If the eventual aim is to use the ligand to study endogenous receptors in tissue or cell lines, then determination of receptor subtype selectivity should be considered (see Note 14). 3. Seed the cells into 8-well chambered coverglasses in culture medium at a density which achieves 70% confluency in 48 h. 4. Incubate the cells in a 37°C incubator in a humidified atmosphere containing 5% CO2 for 48 h. 5. Prepare dilutions of fluorescent ligand in HBSS at tenfold the final concentration required. For initial experiments, consider using final concentrations in the range of 0.5- to 5-fold KB (for antagonists) or 0.5- to 5-fold the functional EC50 for agonists. 6. Wash cells 2–3 times in pre-warmed HBSS by simple removal/ replacement of culture medium. Following the final wash, add 360 mL of HBSS back to each well. 7. If experiments are to be carried out at room temperature, allow cells to equilibrate to RT for 5–10 min. In contrast, if experiments are to be performed at 37°C, then place cells into a temperature controlled environment as quickly as possible and allow to equilibrate (see Note 15). 8. Configure the confocal microscope to collect single confocal images at an excitation and emission wavelength suitable for the fluorescent ligand to be used (e.g. for BODIPY 630/650 use a 633  nm HeNe laser, 633 dichroic and a long pass

222

Briddon, Kellam, and Hill

650 nm emission filter). Preferably, also configure to collect a simultaneous phase/transmitted light image of the cells. 9. Choose an appropriate objective lens to give a field of view of at least 20–30 cells. If available, a water-immersion lens will give the best optical quality when working in a live cell experiment, where the incubation medium is aqueous. In the first well, using white light illumination, focus on the cell monolayer. Under continual scanning, find the appropriate equatorial focal plane to give an equatorial image of the cells within the monolayer, and adjust the field of view to give a suitable selection of cells. Selection of the focal plane can prove troublesome, given that in the first instance no fluorescent label has been added. It is difficult, but possible, to judge the correct focal plane based on the phase image obtained. However, a more consistent method is to put the detector gain to maximum and find the reflection of the laser for the upper surface of the coverslip. Subsequently, focus can be adjusted to a defined distance above the cover slip for each image (which is dependent on cell type). 10. Adjust the pinhole, offset, laser power and detector gain of the fluorescence channel. Initially, choose a pinhole diameter which gives a slice depth of 1–1.5 mm, which gives a suitable image to assess membrane localisation of ligand. Set the detector offset such that the low intensity area (background) is within the lower end of the PMT detector range. Laser power and detector gain are more difficult to set, since in the first instance, only an autofluorescence signal is present and they should be adjusted to minimise this signal (they are readjusted more accurately later). 11. Start a time series of image collection, collecting a single image every minute for 15–20 min. After the first two frames, carefully add 40 mL of 10× the top concentration of fluorescent ligand to be tested to the well, and mix gently once using the addition pipette, being careful not to disturb the cells and focal positioning. 12. Stop the time series when it becomes apparent that the binding of the ligand (membrane or intracellular) has saturated. Using continual scanning, now readjust the laser power and gain to ensure that the whole image is within the dynamic range of the detector. 13. Now proceed to carry out step 11, using the full range of fluorescent ligand concentrations in the remaining seven wells of the plate (see Note 16). Each of these time courses should be collected under the same power/offset/gain settings as determined in step 12, in order that semi-quantitative comparisons between membrane intensities and ligand concentrations can be made (see Note 17).

Design and Use of Fluorescent Ligands

223

14. Subsequent analysis of these data provides information about a number of parameters, including the optimal concentration of ligand to use for subsequent competition experiments. It is important that either (a) all of this analysis is performed on raw image data (i.e. no image adjustment) or that (b) if postcapture adjustment has to be performed it is done equally across all image sets. From one series of images, use appropriate image analysis software to draw regions of interest (ROIs) around the membrane of a selection of 10–15 cells within the field of view (see Note 18). Plotting the mean membrane intensity (IM) of these cells over time gives an indication of the kinetics of binding of the fluorescent ligand to the membrane (Fig. 3). It should be noted, however, that this binding represents total binding of fluorescent ligand (i.e. specific receptor binding plus non-specific membrane binding). Similarly, selection of ROIs from the cytoplasm of 10–15 cells and subsequent calculation of the mean cytoplasmic intensity (IC), gives an indication of the internalisation of the ligand over time, and with changes in ligand concentration. Calculation of the membrane to cytoplasmic ratio (IM:IC) for differing time points and drug concentrations enables the most suitable concentration and time point to be chosen for the competition analysis (see below).

Fig. 3. Confocal imaging of XAC-X-BY630 binding to the adenosine-A3 receptor. CHO cells expressing the human adenosine-A3 receptor were incubated with 50 nM XAC-X-BY630 for 10 min at room temperature. Single confocal slices showing XAC-X-BY630 distribution after 0, 3, 6, and 10 min incubation. Right: Mean ± s.e. mean of fluorescence intensities from the membranes (IM) and Left: the cytoplasm (IC) of 15 cells analysed over the full time course of incubation.

224

Briddon, Kellam, and Hill

Table 1 Example protocol for addition times in a multi-well confocal imaging experiment, using 10 min ligand incubation Time since beginning of experiment (min) Well 1 Well 5 Well 2

Well 6

Well 3 Well 7 Well 4 Well 8

Addition of fluorescent ligand

 0

 4

 8

12

16

20

24

28

Image acquisition

10

14

18

22

26

30

34

38

15. Having chosen a time point which yields substantial membrane binding (if present) over the concentration range, nonspecific binding can be determined. In a second plate, repeat the washing procedure as detailed in step 6, but replace with 320 mL of HBSS at the final stage. 16. Prepare a dilution of non-fluorescent antagonist in HBSS from the 10 mM stock solution at 1,000–5,000 times its KD value for the receptor of interest. To wells 1–4 add 40 mL of HBSS, and to wells 5–8 add 40 mL of this antagonist solution (yielding a final concentration of 100–500 times KD). Allow the antagonist to equilibrate for 30 min at 37°C. 17. Transfer the plate to the microscope and, when equilibrated to the appropriate experimental temperature (see step 7), add fluorescent ligand to well 1 (no antagonist), i.e. 40 mL of 10× required concentration. At the pre-determined time, acquire a single confocal image with settings as determined in step 12. Subsequently, add the same concentration of ligand to well 5 (containing non-fluorescent antagonist) and repeat. 18. Repeat step 17 for three further concentrations of fluorescent ligand. To minimise the amount of time the cells spend on the stage, and to minimise variation in the time exposed to antagonist, stagger the additions to each well as shown in Table 1. 19. Calculate IM, IC and IM:IC in the presence and absence of displacing ligand as described in step 14 (see Note 19). The difference between IM in the presence and absence of nonfluorescent antagonist is an indication of specific binding (Fig.  4). An estimate of the fluorescent ligand KD can be obtained by plotting ligand concentration vs. IM, and fitting to a saturation-binding curve. 3.3.2. Competition-Binding Experiments Using Fluorescent Ligand Binding

1. In similar experiments to those described in Subheading 3.3.1, the fluorescent–ligand interaction can be characterised using competition experiments with a range of concentrations of non-fluorescent ligands (akin to those described using radioligands in Chapter 8). From the experiments carried out

Design and Use of Fluorescent Ligands

225

Fig. 4. Determining the specific binding of XAC-X-BY630 in CHO-A3 cells. (a) Single confocal images of CHO-A3 cells were obtained after incubation with 50 nM XAC-X-BY630 (10 min, 24°C) following incubation (30 min, 37°C) with HBSS (top) or the non-fluorescent adenosine-A3 receptor antagonist MRS1220 (100nM) (bottom). (b) Example ROIs used for the assessment of membrane intensity (IM, solid line) or cytoplasmic intensity (IC, dashed line) are shown. Mean fluorescence intensities calculated from membrane (IM) and cytoplasm (IC) ROIs from 15 cells in each of the images in (a). Specific binding is indicated by the substantial decrease in IM seen following pre-incubation of the cells with MRS1220.

in Subheading 3.3.1, choose a concentration of ligand, close to its KD for the receptor, which gives the best signal to noise ratio for imaging. 2. Using the 10 mM stock solution, prepare a solution of fluorescent ligand in HBSS at ten times the required concentration. 3. Similarly, prepare a range of antagonist concentrations (0.1–100 × KD) in HBSS at ten times the final required concentration. 4. Prepare the cells Subheading 3.3.1.

as

described

in

Steps

3–4

in

5. Wash the cells 2–3 times in HBSS, and replace finally with 320 mL HBSS. 6. Add 40 mL of HBSS to wells 1 and 8, and 40 mL of antagonist solution (10× final concentration) to wells 2–7, and incubate for 30 min at 37°C. 7. Staggering the additions to the plate as described in Subheading  3.3.1, add 40  mL of fluorescent ligand at 10× required concentration, and take single confocal images following the appropriate incubation time. Perform a control measurement (no non-fluorescent ligand) as both the first and the last well to account for changes in cell viability. 8. Analyse the images for membrane intensity as described above.

226

Briddon, Kellam, and Hill

3.4. Quantifying Receptor–Ligand Interactions Using FCS 3.4.1. Introduction

3.4.2. Calibration of the FCS Detection Volume

FCS is an advanced imaging technique, based on analysing fluorescence fluctuations in a small illuminated detection volume, created using the same optical configuration as a standard confocal microscope (16, 20, 21). Despite its specialist nature, the sensitivity, spatial resolution, and quantitative nature of FCS mean that it is becoming more widely used as a method for assessing the number of receptors and ligand–receptor complexes in small microdomains of living cells, as well as quantifying their diffusion. FCS uses a small confocal detection volume (~0.2 fL), which is created by focussing a laser to a diffraction limited spot though a high numerical aperture objective lens. The detection area is then limited by placing a pinhole in the image plane. The volume can be placed accurately on the cell membrane, and as fluorescent species diffuse through this volume, fluorescence emission is detected at the single photon level by highly sensitive avalanche photodiodes. As the volume is small, diffusion of fluorescent species through the volume results in a time-dependent fluctuation in the detected fluorescence intensity. The size and frequency of these fluctuations are dependent on the dwell time of the fluorescent species in the volume, and the number of complexes present during data collection. These two factors (dwell time and particle number) can be determined by autocorrelation analysis of the recorded fluctuations. Thus, when used in conjunction with an appropriate fluorescent ligand, FCS can be used to quantify both the number and mobility of ligand–receptor complexes within small areas (~0.1 mm2) of the membranes of living cells. In addition, the difference in diffusion rates between the fast-moving free fluorescent ligand and the slower receptor–ligand complex allow simultaneous quantification of free and bound ligand in the same measurement (Fig.  5). Alignment of the system (mirrors, pinholes, etc.) is crucial to obtaining accurate and reproducible FCS data, and the alignment and calibration protocol (Subheading 3.4.2) must be completed each time as experiment is performed. In the limited space available here, we provide only a basic protocol for microscope calibration and simple ligandbinding measurements using FCS. For a more detailed practical guide to FCS principles and measurements, we direct the reader to the reviews by Schwille (20) and Kim et al. (21). More detailed explanations of the use of FCS in ligand–receptor interactions are also available (25, 26). 1. Switch on the microscope system, lasers and air conditioning (if present) and allow 1–2 h for the system to equilibrate. 2. Calibration is achieved using an aqueous solution of a fluorophore with appropriate excitation/emission properties. For 458/477/488/514/543 nm excitation use Rhodamine 6G or AlexaFluor546 and for 633/647 nm excitation, use Cy5. Make a 1 mM solution by adding 10 mL of 10 mM fluorophore

Design and Use of Fluorescent Ligands

227

Fig. 5. Determining the number and diffusion characteristics of XAC-X-BY630-adenosine-A3 receptor complexes in CHOA3 cell membranes. (a) CHO-A3 cells were incubated with 5 nM XAC-X-BY630 (10 min, 24°C), and the confocal volume positioned in x and y (white cross) over a cell nucleus using a live confocal image. (b) An intensity scan in z shows distinct peaks for the lower (LM) and upper (UM) cell membranes. The measurement volume was positioned as indicated by the solid line, at 50% the peak membrane intensity. (c) Fluorescence fluctuations were subsequently recorded (upper ), and autocorrelation analysis performed (middle ). Non-linear regression analysis of the autocorrelation function revealed three components with differing dwell times in the confocal volume representing fast-moving free ligand (tD1) and two slower moving components (tD2/3) representing receptor–ligand complexes. The concentration of each component can be calculated from the average particle number in the volume (N ) and the relative percentages of each component.

stock solution to 990 mL of fluorescence free water (0.1 mM final), mix and repeat to give 1  mM. Add 20  mL of this to 980 mL of water to give a 20 nM solution. 3. Add 200 mL of the 1 mM and 20 nM solutions to separate chambers of an 8-well chambered coverglass. 4. Using the appropriate water-immersion objective lens (usually a c-Apochromat 40× or 63×, 1.2NA), add immersion water and place the coverglass on the microscope with the 1 mM solution over the objective. 5. Position the focal volume 200 mm into the fluorophore solution. The procedure varies for this between microscopes, but usually involves first finding the upper coverslip surface using a reflection beampath.

228

Briddon, Kellam, and Hill

6. Select the appropriate beampath combination of laser excitation, dichroic mirror, and emission filter for the fluorescent ligand being used. Turn on the excitation light and adjust the laser power to obtain a count rate at the detector of 250–500 kHz. 7. Using the correction collar on the objective, maximise the count rate (see Note 20). 8. Set the pinhole diameter to 1 Airy unit. Adjust the pinhole position in the x, y, and (if available) the z plane to obtain a maximum count rate in each case. If maxima are not found, then repeat step 7, and possibly adjust the collimation of the excitation light (see Note 21). 9. Once a satisfactory maximum is obtained, move to the 20 nM fluorophore solution, and again, adjust the laser power to obtain a count rate of 20–100 kHz (see Note 22). 10. Collect fluorescence fluctuations for 10 × 10 s reads. 11. Fit the subsequent autocorrelation curve using non-linear regression analysis in appropriate data analysis software. This is often done within the data capture software itself (see Note 23). For calibration data, use a model which assumes free 3D diffusion, with an added pre-exponential to account for fluorophore triplet state: −1

 t   t  G (t ) = 1 + AN 1 +  1 + 2  S tD   tD  

−0.5

−1



(

)



where A = 1 + Tt e −t /t t (1 − Tt ) . N = particle number, tD = diffusion time, S = structure parameter, Tt = triplet fraction and tt = triplet relaxation time (see Note 24). −1

12. If the value obtained for the structure parameter, S, in this fit is between 3 and 8, then the alignment is satisfactory. If it is outside of this range, repeat the alignment process, until a satisfactory value is obtained. 13. In addition, the diffusion time value, tD, from the above fit allows the detection volume to be calculated. The radius of the detection volume, w0, can be obtained using w0 = (4DtD)0.5, where D is the diffusion coefficient of the calibration fluorophore (literature values are D = 2.8 × 10−10 m2/s for R6G and 3.1 × 10−10 m2/s for Cy5 at 22°C). The half height of the volume is then given by w1 = Sw0, and the detection volume, V, is given as V = p1.5w02w1. Knowledge of the volume, allows the concentrations and diffusion coefficients to be calculated from particle number, N, and diffusion time, tD, respectively in subsequent experiments.

Design and Use of Fluorescent Ligands 3.4.3. Determining Diffusion of Receptor– Ligand Complexes Using FCS

229

1. Prepare a 20  nM solution of fluorescent ligand in HBSS (see Note 25). Add 200 mL of this solution to an empty well of a chambered coverglass, and place on the microscope, with this well positioned over the objective. 2. With the confocal volume positioned 200 mm into the solution, determine the diffusion time of the ligand as described for the fluorophore in Subheading 3.4.2, steps 10 and 11. 3. Use cells which have been grown and seeded onto 8-well coverglasses as described in Subheading 3.3 (see Note 26). Wash the cells three times in HBSS, and replace with a final volume of 360 mL of HBSS. Allow the cells to equilibrate to room temperature for 10 min before placing them on the microscope (see Note 27). 4. Add 40 mL of fluorescent ligand at ten times the final required concentration to the first well, and allow incubation with the ligand for a suitable time. 5. Using a live image of the cells, locate a suitable cell for measurement, and position the measurement volume in the x–y plane. 6. To position the volume correctly in the z-plane (i.e. on the upper cell membrane), perform an incremental scan along the z-axis at 0.5 mm intervals using the lowest possible laser power to avoid bleaching. The profile obtained should show two distinct peaks indicating the lower and upper cell membranes. 7. Centre the confocal measurement volume at the point at which the peak intensity for the upper membrane decays to 50% of its value. 8. Collect fluorescence fluctuations using a 15  s pre-bleach exposure to laser (with no data collection) followed by 4 × 15 s of data collection (see Note 28). 9. To confirm that measurement was taken on the upper membrane, repeat the z-scan as per step 7, and then move the measurement volume 3 mm upwards in the z-direction. Repeat the measurement to confirm that only free ligand is present. 10. Data should then be fitted using non-linear regression to a suitable model. For an autocorrelation curve containing free ligand and two binding components, a model assuming one 3D and two 2D components and a pre-exponential (A) for triplet should be used, as follows: −1 −1 −0.5 −1     t   t  t  t   G (t ) = 1 + AN  F1 1 + 1+ + F2 1 + + F3 1 +    t D1   S 2t D1   t D2   t D3   −1

230

Briddon, Kellam, and Hill

where F1,tD1 represent the fractional contribution and dwell time of the free ligand, and F2,tD2 and F3,tD3 represent those for the ligand-bound receptor. Particle numbers (and subsequent concentrations) of each component can be calculated as a simple fraction of the total particle number.

4. Notes 1. The diluents and concentrations at which stock solutions of the fluorescent ligand should be made are obviously dependent on the ligand itself. However, most require an organic solvent, such as DMSO. Ideally, a 10 mM stock solution minimises the final concentration of the solvent at the concentrations used in the experiments. Storage in small frozen aliquots minimises any breakdown caused by freeze-thaw cycles, which should be avoided, particularly for peptides. In addition, the use of translucent amber or opaque storage tubes minimises the effect of light-induced breakdown (which varies depending on the fluorophore used). Since DMSO is hygroscopic, long-term storage of aliquots can lead to a decrease in solubility, and subsequent aggregation of fluorescent ligands which are particularly water-insoluble. Ultrasonication prior to use helps to reverse this process. 2. The aqueous buffer used for determining the photophysical properties of the ligand should be the same as that used for subsequent imaging experiments. 3. Any organic solvent can be used as a comparison. However, DMSO is readily available and in this case is the solvent used for the stock fluorophore solutions. 4. Excitation–emission measurements can be carried out on any equipment capable of varying excitation or emission wavelength in small increments (2–5 nm) over the required range. A plate reader format is more convenient and higher throughput. However, a simple cuvette-based spectrometer, equipped with variable excitation and emission slits, is perfectly adequate, and may actually offer an increased sensitivity. 5. The required growth medium is entirely cell-type dependent. Consideration should be given to using medium without phenol red pH indicator, as this is taken up by cells and can add to cellular autofluorescence. 6. The buffer used for live cell imaging applications may also vary depending on the cell type used and also the incubation conditions used during imaging experiments. A non-bicarbonate based buffer (such as the HEPES used in the recommended buffer here) will be necessary if the imaging platform

Design and Use of Fluorescent Ligands

231

does not have a CO2-fed incubator. If this is the case, then normal serum-free growth medium can be used. The inclusion of a small amount of sodium bicarbonate helps with long-term cell viability and sodium pyruvate minimises any light-induced toxicity resulting from the inclusion of HEPES. 7. Multi-well chambered coverglasses are convenient for the use of multiple treatment conditions, where experiments are performed at room temperature and cells are happy under these conditions for periods of 1–2 h. They are also available with differing numbers of wells per slide, if required. Individual microscopes each have their own stages and incubation chambers, and a carrier or coverslip should be used which is convenient and compatible. Temperature requirements (see below) may determine which carrier is most suitable. For instance, in temperature and humidity controlled chambers, often standard No.1.5 22 or 35 mm diameter coverslips are used, and maintaining even temperature control over multi-well coverglasses is very difficult to achieve. If drug addition via continuous perfusion is required, then the use of a specialist perfusion system, which may have its own coverslip requirements, will be necessary. 8. The protocol as described uses confocal imaging, which has the advantage of being more easily able to distinguish membrane binding from cytosolic due to the optical sectioning obtained. A variety of confocal microscope systems, both “off-the-shelf” and custom built, are available. The specific features and procedures involved differ between microscope manufacturers; here we have described some of the generic methodology involved. Widefield imaging systems can also be used for the screening of fluorescent–ligand binding, and indeed may show a higher degree of sensitivity. However, to obtain accurate determination of membrane vs. intracellular binding, the ability to deconvolve widefield images is a distinct advantage. 9. The choice of competing antagonist is an important one. It should preferentially be of a distinct structure to the parent pharmacophore and be non-fluorescent (even if this is outside the range being tested for the fluorescent ligand). Selectivity may be an issue if the receptor of interest is not expressed uniquely in the cell line used. 10. One of the main disadvantages of using FCS to monitor the diffusion of receptor–ligand complexes is the requirement for expensive and sophisticated imaging hardware. Retro-fitting of FCS detection units to existing confocal microscopes is possible for most major makes of confocal microscope. Alternatively, an FCS microscope can be custom built on a high-end inverted research microscope, but this requires a much higher degree of optical engineering expertise.

232

Briddon, Kellam, and Hill

11. For FCS-modified confocal microscopes, analysis software containing modules for non-linear regression analysis of autocorrelation curves is often provided as part of the system. If not, then other commonly used analysis programmes, such as GraphPad Prism, Matlab, or Origin, are capable of fitting autocorrelation data if the appropriate equations are entered. 12. While the emission spectrum is not dependent on the excitation wavelength, the emission intensity is. Therefore, when obtaining the second part of the spectra by shifting the excitation wavelength, the data should be normalised before combining. 13. The cell line chosen for screening of the ligand should have a reasonable level of expression of the receptor of interest. A cell line transfected to express the receptor on a null background provides a high signal:noise ratio for optimisation of assay conditions, before moving onto a more complex, endogenously expressing system. CHO cells and COS-7 cells are widely used in this regard, as they have low levels of endogenous GPCR expression. A further advantage of using an artificially expressing system is the ability to test the ligand and the parent fluorophore against the native cell line for nonspecific binding. 14. As described in the Introduction, the novel fluorescent ligand should be considered as a new pharmacological entity, and fully functionally characterised. While radioligand binding gives some indication of affinity, for agonists especially, it is important to know that efficacy has been preserved. Ideally, affinity and efficacy should be determined in more than one functional assay. For efficacy determination, one of these should be an assay in which there is little receptor reserve, since this will be more likely to reveal any partial agonism. In each case, the potency of the fluorescent ligand should be compared to the parent ligand to establish the full effect of adding the fluorophore. A further consideration is whether the functional assay itself can be affected by the fluorescent label on the compound being tested. For instance, absorption of light by the fluorophore may interfere with absorption or luminescence readout on a reporter gene assay. Likewise, functional assays should be carried out with due consideration to the solubility and light-sensitivity of the compound. Finally, it should be noted that the receptor subtype selectivity of the pharmacophore is as likely to be altered as the absolute potency, and if the ligand is to be used in endogenously expressing systems, then this should also be determined. 15. The choice of temperature at which to conduct imaging experiments is influenced by a number of factors. Performing experiments at 37°C undoubtedly keeps cells healthy, and gives a more physiologically relevant result. However, it increases the

Design and Use of Fluorescent Ligands

233

binding kinetics of the ligand, increases the amount of non-specific uptake of the ligand into the cells and also increases receptor internalisation in response to agonist stimulation. For these reasons, it may be sensible to perform the initial characterisation of agonist ligands at room temperature. 16. An obvious issue when performing these experiments at room temperature in multi-well carriers is that the cells in the later wells have been on the microscope stage for significantly longer than those in the first wells. This can be tested directly by comparing cell morphology, or repeating control binding in the last well in addition to the first on each plate. It is also good practice to reverse the order of experiments in alternate plates. If the cells used are particularly sensitive, then it may be worth considering moving to 4-well plates or individual coverslips. 17. There are a number of possible outcomes to this experiment, some of which may need further investigation. In an ideal situation, clear membrane binding will be seen, with little background signal between cells from the ligand in solution. It may be that there is no clear membrane binding, even though the functional assays predict that binding should occur. One possible explanation for this is that the fluorophore bleaches heavily (e.g. FITC), or that the fluorophore is present in an aqueous environment (i.e. outside of the receptor protein/membrane) and is heavily quenched by water (e.g. dansyl). Information obtained from the spectroscopy experiments may help to interpret this. At the opposite end of the spectrum, it may be that it is difficult to determine membrane binding, as there is high signal in between cells from the free ligand in solution. In these circumstances, washing the cells in HBSS, then re-imaging will help to determine if any of the signal originates from receptor bound ligand, rather than just from extracellular ligand (21). 18. Where clear membrane labelling is achieved, accurately drawing of the membrane ROI is straightforward. However, often high level of cytoplasmic ligand or low-level membrane binding can make the membrane more difficult to distinguish. In the case of low levels of signal, temporarily increasing the brightness/contrast of the image while marking the ROIs can help. If this still does not improve the definition of the membrane enough, then performing the experiments in the presence of a spectrally separated membrane marker (e.g. DiI, DiO, or fluorescently labelled wheat germ agglutinin) will allow the accuracy required. 19. If displacement is complete in the presence of competing antagonist, then correctly identifying the cell membrane for intensity analysis can be difficult. There are a number of possible solutions to this. Firstly, during the analysis, increase the

234

Briddon, Kellam, and Hill

brightness and contrast temporarily to allow the membrane to be located. If this does not provide a solution, then the experiment can be repeated using a lower concentration of non-fluorescent ligand (20–50 × KD), which will give significant displacement, but allow enough signal to remain that locating the membrane is possible. Finally, as described in Note 16, using a membrane marker or GFP-tagged receptor enables membrane identification under all conditions. 20. An objective lens with a correction collar is a preferred option for FCS instrumentation. This is because a detection volume which is close to a true Gaussian–Lorentzian profile yields higher quality data. Use of a correction collar makes this more achievable in the presence of uneven coverslips and chromatic and spherical aberration. 21. Where adjustment of the pinhole in the z-plane is possible, it may be necessary to adjust the collimator to achieve a signal maximum within the travel of the pinhole. Adjustment of the pinhole in the x and y planes should be repeated following any adjustment in the z-position. 22. The absolute count rate is one indicator of the signal achieved. However, if the software allows, using molecular brightness (h) is a more accurate indicator of signal to noise ratio. This is simply the count rate per molecule in the detection volume. A value of molecular brightness of 40–90 kHz would be preferred for calibration measurements. 23. Autocorrelation curves are usually fitted using a Marquardt non-linear least squares fitting algorithm. This can be performed in any proprietary curve-fitting software if the function is not directly available in the microscope software itself. 24. Triplet state is an excited state of the fluorophore which is longer lived (~1–10  ms) than the standard excited state (1–5 ns). Since the time constant for this “dark state” is within the time resolution of autocorrelation analysis, triplet sate transitions appear as a component of the autocorrelation curve. A term is included in the model equation to account for this. Fluorophores with high triplet state yields (e.g. FITC) are unsuitable for use in FCS as they can cause a significant underestimation of particle number. The structure parameter, S, is a ratio of the height to the diameter of the detection volume, and for an ideal Gaussian–Lorentzian volume is 2.2. As this value is very sensitive to the quality of the data, obtaining a structure parameter of between 3 and 8 during calibration is indicative of a good system alignment. 25. Selection of an appropriate fluorescent tag for use in FCS experiments is more complex than for simple confocal imaging applications. The sensitivity of the detection system means that both autofluorescence and background signal are more

Design and Use of Fluorescent Ligands

235

of an issue. Thus, red-shifted fluorophores are preferable for FCS since both cellular autofluorescence and phototoxicity are reduced at these wavelengths. In addition, low triplet state population is a requirement for accurate quantification of ligand. Again, this tends to be lower in red-excited fluorophores. 26. As described in Note 25, cellular autofluorescence is a major issue with cell-based FCS measurements. Two approaches which minimise this issue are (1) the use of cell culture medium which does not contain phenol red (which accumulates in cells, and has an excitation wavelength in the bluegreen region) and (2) ensuring cells are as healthy as possible for experimentation. 27. Cell-based FCS measurements may be taken at 37°C. However, this increases the frequency of temperature-induced membrane fluctuations, which can produce significant artefacts in the autocorrelation curve. For this reason, experiments are more routinely carried out at room temperature. 28. Optimisation of read times and laser powers should be carried out empirically. The optimal laser power provides a good signal to noise ratio, but does not induce significant bleaching. It should be noted that with FCS, two types of bleaching need to be taken into account. Global bleaching results in a general decrease in the fluorescence signal, and occurs both within and outside the detection volume. This results in significant artefacts in the autocorrelation curve. Spot bleaching occurs within the excitation volume, where the power density is greatest, and fluorescent species are bleached as they diffuse through the excitation volume. This manifests itself as an artificially short dwell time and a leftward shift in the autocorrelation decay curve. Often global bleaching is a result of immobile receptors and is therefore more prevalent at the beginning of exposure to excitation light. To circumvent this problem, a “pre-bleach” period is introduced, where cells are exposed to laser light for a short time prior to the start of data collection. The laser power and bleach time used for this should also be determined empirically. References 1. Middleton, R.J., and Kellam B. (2005) Fluorophore-tagged GPCR ligands. Curr. Op. Chem. Biol. 9, 517–525. 2. Daly, C.J. and McGrath, J.C. (2003) Fluorescent ligands, antibodies, and proteins for the study of receptors. Pharmacol. Ther. 100, 101–118. 3. Leopoldo, M., Lacivita, E., Berardi, F., and Perrone, R. (2009) Developments in fluorescent

probes for receptor research. Drug. Discov. Today 14, 706–712. 4. Ostrom, R.S, and Insel, P.A. (2004) The evolving role of lipid rafts and caveolae in G proteincoupled receptor signalling: implication for molecular pharmacology. Br. J. Pharmacol. 143, 235–245. 5. Insel, P.A., Head, B.P., Ostrom, R.S., Patel, H.H., Swaney, J.S., Tang, C.M., Roth, D.M. (2005)

236

Briddon, Kellam, and Hill

Caveolae and lipid rafts: G-protein-coupled receptor signalling microdomains in cardiac myocytes. Ann. N.Y. Acad. Sci. 1047, 166–172. 6. Hern, J.A., Baig, A.H., Mashanov, G.I., Birdsall, B., Corrie, J.E., Lazareno, S., Molloy, J.E., and Birdsall, N.J. (2010) Formation and dissociation of M1 muscarinic receptor dimers seen by total internal reflection fluorescence imaging of single molecules. Proc. Natl. Acad. Sci. USA 107, 2693–2698. 7. Miquel, M.R., Segura, V., Ali, Z., D’Ocon, M.P., McGrath, J.C., and Daly, C.J. (2005) 3-D image analysis of fluorescent drug binding. Mol. Imaging 4, 40–52. 8. Briddon, S.J., Cordeaux, Y., Middleton, R.J., Weinstein, J, George, M.W., Kellam, B., and Hill, S.J. (2004) Quantitative analysis of the formation and diffusion of A(1)-adenosine receptor-antagonist complexes in single living cells Proc. Natl. Acad. Sci. USA 101, 4673–4678. 9. Daly, C.J., Ross, R., Whyte, J., Henstridge, C., Irving, A., and McGrath, J. (2010) Fluorescent ligand binding reveals heterogeneous distribution of adrenoceptors and ‘cannabinoid-like’ receptors in small arteries. Br. J. Pharmacol. In Press, DOI: 10.1111/j.1476-5381.2009.00608.x 10. Jones, J.W., Greene, T.A., Grygon, C.A., Doranz, B.J., and Brown, M.P. (2008) Cellfree assay of G-protein-coupled receptors using fluorescence polarization. J. Biomol. Screen 13, 424–429. 11. Harikumar, K.G., Pinon, D.L. and Miller, L.J. (2006) Fluorescence characteristics of hydrophobic partial agonist probes of the cholecystokinin receptor. Biosci. Rep. 26, 89–100. 12. Harikumar, K.G., Hosohata, K., Pinon, D.I., and Miller, L.J. (2006) Use of probes with fluorescent indicator distributed throughout the pharmacophore to examine the peptide agonist-binding environment of the family B G protein-coupled secretin receptor. J. Biol. Chem. 281, 2543–2550. 13. Ilien, B., Franchet, C., Bernard, P., Morisset, S., Well, C.O., Bourguignon, J.J., Hibert, M., and Galzi, J.L. (2003) Fluorescence resonance energy transfer to probe human M1 muscarinic receptor structure and drug binding properties. J. Neurochem. 85, 768–778. 14. Tahtaoui, C., Guillier, F., Klotz, P., Galzi, J.L., Hibert, M. and Ilien, B (2005) On the use of nonfluorescent dye labelled ligands in FRETbased receptor binding studies. J. Med. Chem. 48, 7847–7859. 15. Middleton, R.J., Briddon, S.J., Cordeaux, Y., Yates, A.S., Dale, C.L., George, M.W., Baker, J.G., Hill, S.J., and Kellam, B. (2007) New fluorescent adenosine A1-receptor agonists which allow quantification of ligand-receptor

interactions in microdomains of single living cells. J. Med. Chem. 50, 782–793. 16. Briddon, S.J. and Hill, S.J. (2007) Pharmacology under the microscope: the use of fluorescence correlation spectroscopy to determine the properties of ligand receptor complexes. Trends Pharmacol. Sci. 28, 637–645. 17. Cordeaux Y, Briddon SJ, Alexander SPH, Kellam B & Hill SJ (2008) Agonist-occupied A3 adenosine receptors exist within heterogeneous complexes in membrane microdomains of individual living cells. FASEB J. 22, 850–860. 18. Hegener O, Prenner L, Runkel F, Baader SL, Kappler J & Haberlein H (2004) Dynamics of beta2-adrenergic receptor-ligand complexes on living cells. Biochemistry 43, 6190–6199. 19. Ziemek, R., Brennauer, A., Schneider, E., Cabrele, C., Beck-Sickinger, A.G., Bernhardt, G., and Buschauer, A. (2006) Fluorescence- and luminescence -based methods for the determination of affinity and activity of neuropeptide Y2 receptor. Eur. J. Pharmacol. 551, 10–18. 20. Schwille P (2001) Fluorescence correlation spectroscopy and its potential for intracellular applications Cell Biochem. Biophys. 34, 383–408. 21. Kim S, Heinze K & Schwille P (2007) Fluorescence correlation spectroscopy in living cells. Nat. Methods 4, 963–973. 22. Baker, J.G., Middleton, R.J., Adams, L., Briddon, S.J., Kellam, B, and Hill, S.J. (2010) Pharmacological characteristics of different fluorescent ligands at the human adenosine-A1 receptor. Br. J. Pharmacol 159, 772–786. 23. Wohland T, Friedrich K, Hovius R & Vogel H (1999) Study of ligand-receptor interactions by fluorescence correlation spectroscopy with different fluorophores: evidence that the homopentameric 5-hydroxytryptamine type 3As receptor binds only one ligand. Biochemistry 38, 8671–8681. 24. Cornelius, P., Lee, E., Lin, W., Wang, R., Werner, W., Brown, J.A., Stuhmeier, F., Boyd, J.G., and McClure, K. (2009) Design, synthesis and pharmacology of fluorescently labeled analogs of serotonin: application to screening of the 5-HT2C receptor. J. Biomol. Screen. 14, 360–370. 25. Briddon, S.J., Hern, J.A., and Hill, S.J. (2009) Use of fluorescence correlation spectroscopy to study GPCRs. In G-protein coupled receptors – Essential Methods (Poyner, D. and Wheatley, M., ed.), Wiley-Blackwell, 169–191. 26. Pramanik A & Rigler R (2001) FCS-analysis of ligand-receptor interactions in living cells. In Fluorescence Correlation Spectroscopy: Theory and Applications, (Elson, E. & Rigler, R., ed.) Springer, Heidelberg. 101–129.

Chapter 12 Examining Site-Specific GPCR Phosphorylation Adrian J. Butcher, Andrew B. Tobin, and Kok Choi Kong Abstract Phosphorylation of G protein-coupled receptors (GPCRs) is one of the most prominent post-translation modifications mediated by agonist stimulation. This process has been shown to result not only in receptor desensitisation but also, via the recruitment of arrestin adaptor proteins, to promote receptor coupling to numerous signalling pathways. Furthermore, there is now a growing body of evidence suggesting that GPCRs may employ phosphorylation as a mechanism to regulate their cell-type-specific signalling, hence generating tissue-specific functions. These advances have resulted partly from improved methods used in the determination of phospho-acceptor sites on GPCRs and improved analysis of the consequences of phosphorylation. This chapter aims to describe the methods used in our laboratory for the investigation of site-specific phosphorylation of the M3-muscarinic receptor. These methods could easily be applied in the study of other receptors. Key words: Antibodies, G protein-coupled receptor, Immunoprecipitation, Phosphorylation, Proteomics, 2D phosphopeptide maps, SDS–PAGE, Mass spectrometry

1. Introduction It has been well established that most, if not all, G protein-coupled receptors (GPCRs) undergo phosphorylation upon agonist stimulation. Originally associated with receptor desensitisation and internalisation, GPCR phosphorylation has now been implicated in GPCR coupling to G protein-independent pathways (1) and as a mechanism that contributes to tissue-specific signalling (2, 3). Ideally, studies of GPCR phosphorylation would reveal the precise sites of phosphorylation in both transfected and native tissue. The challenge, however, lies in the low expression levels of GPCRs (even in transfected systems) as well as the problems of biochemical extraction and analysis of these highly hydrophobic proteins. Despite this, a number of techniques have been successfully

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_12, © Springer Science+Business Media, LLC 2011

237

238

Butcher, Tobin, and Kong

applied in the determination of the phosphorylation sites of GPCRs. These techniques include phosphopeptide mapping (4, 5), mass spectrometry (6), phospho-specific antibodies (7, 8) and mutagenesis (9–11). In our laboratory, we have adopted a three component strategy to the question of determining the site-specific phosphorylation of the M3-muscarinic receptor. The first is to establish what we refer to as a “phosphorylation signature” of the receptor. This we achieve through a tryptic phosphopeptide map of [32P]-labelled receptor, immunoprecipitated from transfected or native primary cells (usually neurones). The advantage of this approach is that we are able to analyse the full extent of the phosphorylation of the receptor since every phosphopeptide is visualised in the autoradiograph of the phosphopeptide map, an example of this is shown in Fig. 1. Thus, any changes in the phosphorylation status of the receptor are readily detected by a change in the phosphopeptide “spots” in the map. Furthermore, this technique does not rely on the expression levels of the receptor but rather on the incorporation of the [32P]-label, hence this is useful in native primary cells with low receptor expression levels. The disadvantage is that the precise sites of phosphorylation are not determined with this method. The second approach is to determine the sites of phosphorylation of the receptor using mass spectrometry. This technique requires large (1 mg or more) amounts of purified receptor protein

Fig. 1. Flow chart for 2D phosphopeptide map generation and Edman degradation.

Examining Site-Specific GPCR Phosphorylation

239

but has advantages over [32P]-labelling and 2D phosphopeptide mapping in that the peptide identity and specific sites of phosphorylation can be determined. In addition, this technique can be combined with isotopic labelling of peptides using isobaric tag for relative and absolute quantitation (ITRAQ) or stable isotope labelling with amino acids in cell culture (SILAC) such that relative changes at specific sites can be measured in response to stimulation with agonist (12). Receptors can be purified relatively easily from stably expressing cell lines using epitope tags or high affinity receptor-specific antibodies. The major advantage to this approach is that the data can be used to direct the production of phospho-specific antibodies. These antibodies can be used in western blotting, immunocytochemistry or immunohistochemistry studies to establish the extent of receptor phosphorylation in native tissues. In this chapter, the application of these three approaches (i.e. two-dimensional [2D] tryptic phosphopeptide mapping, mass spectrometry, and generation of phosphorylation site-specific antibodies) in our studies of the M3-muscarinic receptor is described. These methods can be readily adapted for similar studies on other GPCRs.

2. Materials 1. [32P]orthophosphate (PerkinElmer) at 10 mCi/mL. 2. Krebs/HEPES: 10 mM HEPES at pH 7.4, 118 mM NaCl, 4.3  mM KCl, 1.17  mM MgSO4, 1.3  mM CaCl2, 25  mM NaHCO3, 1.18 mM KH2PO4, 11.7 mM glucose. 3. RIPA buffer: 10  mM Tris–HCl at pH 7.4, 2  mM EDTA, 20  mM b-glycerophosphate, 160  mM NaCl, 1% (v/v) Nonidet P-40 (NP-40), 0.5% (w/v) deoxycholate. 4. Tris-buffered saline (TBS): 10  mM Tris Base at pH 7.4, 100 mM NaCl. 5. TEG buffer: 10  mM Tris–HCl at pH 7.4, 2  mM EDTA, 20 mM b-glycerophosphate. 6. Protein A-Sepharose (GE-Healthcare), 1.5 g resuspended in 50 mL of TEG buffer. 7. Laemmli buffer: 125 mM Tris–HCl at pH 6.8, 10 mM b-mercaptoethanol, 4% (w/v) sodium dodecyl sulphate (SDS), 20% (v/v) glycerol and 0.05% (w/v) bromophenol blue. 8. 0.5% (w/v) polyvinylpyrrolidone K30 (PVP) in 0.6% (v/v) acetic acid. 9. Trypsin solution: 1 mg of sequencing grade modified trypsin (Promega) dissolved in 50–150 mL freshly prepared 50 mM NH4HCO3 solution.

240

Butcher, Tobin, and Kong

10. pH 1.9 buffer: 88% (v/v) formic acid:acetic acid:water 25:78:897 (v/v/v). 11. pH 3.5 buffer: acetic acid:pyridine:water 100:10:1,794 (v/v/v). 12. Isobutyric acid chromatography buffer: isobutyric acid:nbutanol:pyridine:acetic acid:water 1,250:38:96:58:558 (v/v/v/v/v). 13. MALDI ToF matrix: 10% (w/v) alpha-cyano-4-hydroxycinnamic acid in 50% (v/v) acetonitrile or 10% (v/v) 2,5-dihydroxybenzoic acid (DHB) in 50% (v/v) acetonitrile/1% (v/v) phosphoric acid. 14. CIAP buffer: 50 mM Tris–HCl at pH 8.5, 1 mM EDTA.

3. Methods 3.1. Phosphorylation Site Identification by 2D Phosphopeptide Mapping 3.1.1. [ 32P] Orthophosphate-Labelling and Resolving Phosphorylated Receptors

1. Grow cells expressing the GPCR of interest to about 80–95% confluency in multi-6-well plates and wash twice with 1 mL of phosphate-free Krebs/HEPES. Leave with 1 mL of phosphate-free Krebs/HEPES for 10 min to equilibrate. 2. Remove the buffer and label cells with 1  mL phosphatefree Krebs/HEPES containing 100  mCi/mL of [32P]orthophosphate for 60–90 min (see Note 1) at 37°C. 3. Prepare agonist as a 100× stock in phosphate-free Krebs/ HEPES. Stimulate cells by adding 10 mL per well. 4. Terminate stimulation by removing medium and lysing cells with 1 mL of cold RIPA buffer. 5. Leave cells on ice (or 4°C) for 10 min to solubilise. Transfer all cellular material to a 1.5-mL microcentrifuge tube. If necessary (when receptor expression level is low), material from two or three wells can be combined. Screw-cap microcentrifuge tubes can be used to reduce contamination of the centrifuge. 6. Pellet down particulate material by centrifuging at maximum speed (~16,000 × g) for 5 min at 4°C. 7. Dilute antibody stock with TEG buffer, 1–5 mg of in-house anti-M3-muscarinic receptor antibody per 100 mL TEG buffer per sample. 8. Transfer 900  mL of supernatant into a fresh tube and add 100 mL of the above diluted anti-receptor antibody in TEG buffer. 9. Leave on ice for 60–90 min. 10. Add 180  mL of Protein A-Sepharose slurry to each sample. Rock in 4°C room for 15 min. 11. Pellet down the Protein A-Sepharose beads by centrifuging at 500 × g for 30 s at 4°C.

Examining Site-Specific GPCR Phosphorylation

241

12. Carefully aspirate off the supernatant without disturbing the beads. Leave ~100  mL of the supernatant in the tube as a precaution. 13. Wash beads three times with 1 mL of ice-cold TEG buffer. Centrifuge at 500 × g for 30 s at 4°C each time to pellet down beads and aspirate off supernatant carefully without loosing beads. 14. During the final wash, remove all the supernatant by aspiration with a fine-tipped gel loading pipette. 15. Add 20  mL of Laemmli sample buffer to the Protein A pellet. 16. Mix samples by flicking the tubes and then heating at 50–60°C for 2–3  min. Do not boil the samples as the receptor may aggregate (see Note 2). 17. Centrifuge the samples briefly at high speed to pellet down the beads and load onto a sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS–PAGE, see below) gel (see Note 3). 18. Prepare an 8% polyacrylamide resolving gel (see Note 4) as follow: Distilled water

4.6 mL

30%-acrylamide/0.8%-N, N,-bis-methylacrylamide mix

2.7 mL

1.5 M Tris Base at pH 8.8

2.5 mL

10% (w/v) SDS

0.1 mL

10% (w/v) ammonium persulfate (APS)

0.1 mL

Tetramethylethylenediamine (TEMED)

0.006 mL

Add latter two components last as they polymerize the gel. 19. Layer the top of the gel with some iso-propanol or water so that the gel sets evenly. Leave at room temperature for at least 20 min for gel to set. 20. Rinse off iso-propanol with distilled water. Prepare a 10-well, 4% stacking gel as follows: Distilled water

3.4 mL

30%-acrylamide/0.8%-N, N,-bis-methylacrylamide mix

0.83 mL

0.5 M Tris Base at pH 6.8

0.63 mL

10% (w/v) SDS

0.05 mL

10% (w/v) APS

0.05 mL

TEMED

0.005 mL

242

Butcher, Tobin, and Kong

21. After polymerisation of the separating and stacking gels in the electrophoresis kit, fill the equipment with electrophoresis buffer containing 25  mM Tris Base, 192  mM glycine and 0.1% (w/v) SDS. 22. Load the samples onto the gel and run at constant 200 V (for a mini gel dimensions 8 × 10 cm) until the bromophenol blue front (diaphragm) reaches about 1  cm from the bottom of the gel (this usually take about 40–45  min). Pre-stained molecular weight markers can be run in parallel to facilitate the identification of the receptor (see Note 5). 23. Transfer the proteins on the gel to nitrocellulose membrane (see Note 6) by using wet transfer apparatus. Fill the tank with cold transfer buffer [25 mM Tris Base, 192 mM glycine and 20% (v/v) methanol] and transfer at 100 V for 1 h. 24. Wrap the membrane in cling film and expose to Hyperfilm (GE Healthcare). The membrane must not be dried as this makes the digestion impossible. An overnight film exposure should be sufficient to visualise the bands. The cassette can be stored at −80°C with an intensifying screen to improve the signal. A fluorescent ruler marker can be used as a guide to the position and orientation of the membrane. Alternatively, the membrane can also be exposed to a phosphor-imager. 3.1.2. 2D Phosphopeptide Mapping

1. Superimpose the autoradiogram or a 1:1 printout of the phosphor-imager analysis from Subheading 3.1, step 1 with the nitrocellulose membrane and cut out the bands of interest. Contours of bands may be marked using a needle or a pen. Precision of the cuts should be verified by re-exposing the membrane. 2. Cut bands into small fine pieces and block with 200  mL of 0.5% polyvinylpyrrolidone K30 (PVP) in 0.6% acetic acid at 37°C for 30 min. 3. Aspirate off the blocking solution and wash membrane pieces three times with distilled water. 4. Digest the proteins on membrane pieces by incubating in 50–150 mL of trypsin solution overnight at 37°C. 5. Transfer the supernatant to a new microfuge tube, and wash the membrane once or twice with water for 15–30 min with shaking at 1,500  rpm at room temperature. Pool all the supernatants and dry down completely with a SpeedVac at room temperature (2–6 h). 6. If the membrane pieces still contain a high amount of radioactivity, they can be digested again and the samples pooled. 7. Redissolve the dried pellet with 25–50 mL of pH 1.9 buffer and dry again with a SpeedVac.

Examining Site-Specific GPCR Phosphorylation

243

8. Dissolve the pellet in 5–10  mL of pH 1.9 buffer, vortex vigorously­and centrifuge at maximum speed (24,000 × g) for 1 min. 9. Spot the supernatant in small portions (0.5  mL) onto a 20 × 20  cm cellulose thin layer chromatography (TLC) plate (VWR). Use a hair-drying fan with no heating to dry every portion. Hot air would “bake” the peptides to the plate and it is essential to get the smallest spot possible as this ensures generation of high quality maps. The spot should be 2 cm from one end of an edge of the plate and 4 cm from the adjacent other. 10. Wet the plate with pH 1.9 buffer using a same size (20 × 20 cm) Whatman paper with a circular hole of 1 cm in diameter at the position of the phospho-peptide spot. Wet the areas around the sample first. If done properly, all the buffer converges on the centre of the circle and acts to concentrate the spotted sample. Pat the paper over the rest of the area to make sure the entire surface of the plate is wet. The plate should look dull-grey and not shiny. A shiny appearance indicates too much buffer on the plate. Areas where the buffer has puddled should be blotted carefully with a KimWipe. 11. Electrophorese for 30–40 min at 2,000 V. 12. Air-dry the plate extensively in a fume hood. 13. Scrape off a lane of cellulose 3 cm from top of the plate using a scraper. 14. Run ascending chromatography overnight in isobutyric acid chromatography buffer. 15. Dry the plate extensively in a fume hood, label the corners with radioactive ink and wrap in cling film and expose to film or a phosphor-imager. 3.1.3. Edman Degradation

1. Scrape the spot(s) of interest on the 2D map identified by autoradiography from the TLC plate and transfer to a microfuge tube. 2. Extract two to three times with 200  mL of pH 1.9 buffer. Pool all the supernatant. 3. Before subjecting the samples to Edman degradation, they need to be covalently attached to a Sequelon-AA (Aryl Amine) disk. 4. Dry down the phospho-peptide samples in microfuge tubes using SpeedVac. 5. Redissolve the phospho-peptide in 20  mL of 50% aqueous acetonitrile solution containing 0.1% trifluoroacetic acid (TFA) (caution: very toxic and corrosive!). Heating the samples at 55°C with occasional mixing (vortex or sonication bath) may facilitate dissolving of the sample.

244

Butcher, Tobin, and Kong

6. Rest a sheet of Mylar (or glass slide) on a heat block at 55°C and place a Sequelon-AA disk on top. 7. With a micropipette, apply the samples to the disk, a small drop at a time and allow the sample to dry in between. Then, allow the disk to dry thoroughly (10–15 min). 8. Remove the Mylar (or glass slide) with the disk on top from the heat block, being careful not to touch the disk. 9. Just before use, weigh out approximately 1 mg of EDC (water soluble carbodiimide) in a microfuge tube and add 100 mL of 0.1 M MES pH 5.0 containing 15% acetonitrile. 10. Carefully apply 5 mL of the freshly prepared EDC solution to the disk. Discard remaining EDC solution. 11. Allow the reaction to proceed at room temperature for 20 min. Let the disk dry and then wash in 1 mL of 50% aqueous acetonitrile/0.1% trifluoroacetic acid. The disk is now ready for Edman degradation. 12. Edman degradation is performed in an Applied Biosystems Procise 49× protein sequencer. Place the Sequelon-AA dick in a vertical Blott cartridge of the sequencer. An anilinothiazolinone (ATZ) amino acid collection chemistry cycle is used, which is a modification of the standard pulsed liquid cycle as described by Campbell and Morrice (13). 13. ATZ amino acid is extracted and transferred at each sequencing cycle using 90% methanol: 10% water solution, which is placed in the X3 bottle position on Procise. A transfer line (Applied Biosystems part number 602930) is attached to ATZ/FC port (or port 39) of the Procise and connected to the arm of an external fraction collector. The fraction collector can be interfaced electronically to the Procise and advanced by introducing a Relay 1 pulse function (function 253) into the sequencing programme or operated by time-based collection (setting a time which is equal to the total time of a sequencing cycle). The ATZ amino acid is transferred to the fraction collector with two extractions of 0.2 mL with a 20 s wait between extractions. 14. Setting up the fraction collector in this way ensures that each cycle of degradation resulting in the cleavage on one amino acid is collected in the fraction collector. The aim is to determine if the collected fraction (i.e. the amino acid that has been cleaved) is phosphorylated. This is done by determination of the radioactive content of the fraction. To do this, firstly dry down completely in a SpeedVac at room temperature (2–6 h). 15. Dissolve in 5–10 mL of 50% aqueous acetonitrile/0.1% TFA and apply onto a sheet of 3MM Whatman filter. A square grid (10 × 10 mm) is marked out in pencil on the filter paper, at

Examining Site-Specific GPCR Phosphorylation

245

the centre of each square a circle 2–3  mm in diameter is marked where each ATZ amino acid fraction is applied in a tight spot. 16. Expose for 2–3 days on a phosphor-imager. Erase the screen twice to obtain a really low background on the screen. Alternatively, autoradiograph on Hyperfilm (GE Healthcare) with intensifying screen although a much longer exposure is needed (1–2 weeks) as the sensitivity is decreased. 3.2. Phosphorylation Site Identification by Mass Spectrometry 3.2.1. Receptor Purification and Preparation of Tryptic Peptides

1. Grow cells expressing the GPCR of interest to about 80–95% confluency in 175-cm2 flasks or roller bottles. Grow sufficient cells such that the amount of receptor precipitated is not less than 1 mg. 2. Wash cells twice with Krebs/HEPES and incubate with Krebs/HEPES for 1 h at 37°C. Prepare 100× stock of agonist and incubate cells with agonist for required time. 3. Terminate stimulation by aspirating buffer and resuspend cells by incubating in cold PBS containing 1 mM EDTA for 5 min. Collect cells and centrifuge for 5 min at 1,000 × g at 4°C. 4. Immediately re-suspend cell pellet in 10 mL ice cold TE buffer containing a cocktail of protease and phosphatase inhibitors (Roche, protease inhibitor and phosSTOP tablets) and lyse cells using a Polytron homogeniser. Centrifuge lysate for 5 min at 1,000 × g to remove large debris and non-lysed cells. Collect resulting supernatant and centrifuge for 1  h at 40,000 × g at 4°C to obtain membrane pellet. 5. Resuspend membrane pellet in 5  mL PBS containing 1% NP-40 and protease and phosphatase inhibitors and incubate on ice for 1 h. Remove insoluble material by centrifugation at 20,000 × g for 20 min. 6. Collect resulting supernatant and dilute 1:1 with PBS before adding 200  mL of precipitating antibody (e.g. anti-HA or anti-FLAG) coupled to agarose beads and incubate on rolling platform at 4°C for 4 h. 7. Wash beads with 4 × 10 mL of lysis buffer and elute with 1 vol. of Laemmli sample buffer. Resolve purified receptors by SDS–PAGE on 1.5-mm thick 8% gels as described in Subheading  3.1, step 1 and stain with colloidal Coomassie blue or other mass spectrometry-friendly protein stain.

3.2.2. Preparation of Peptides for Mass Spectrometry

1. Extract receptor band(s) of interest from Subheading 3.2, step 1 using a clean scalpel and cut polyacrylamide into 1–2-mm squares. Wash gel squares 3 × 15  min with freshly prepared 100 mM ammonium bicarbonate and discard the wash. 2. Re-suspend the gel in 10 mM DTT (freshly prepared) dissolved in 50 mM ammonium bicarbonate (enough volume to cover the gel) and incubate for 30 min at 65°C.

246

Butcher, Tobin, and Kong

3. Discard the DTT solution and replace with 100 mM iodoacetamide dissolved in 50  mM ammonium bicarbonate and incubate for 30 min at room temperature in the dark. 4. Discard the iodoacetamide solution and wash the gel 3 × 15 min with 50% acetonitrile in 50 mM ammonium bicarbonate and 1 × 15  min in 100% acetonitrile. Gently remove acetonitrile and evaporate residual acetonitrile in a SpeedVac. 5. Re-swell the gel (add sufficient solution to just cover the gel pieces) with a solution of 50  mM ammonium bicarbonate containing 1  mg of sequencing grade trypsin and incubate overnight at 37°C. 6. Vortex gel pieces for 2  min, remove and keep supernatant. Resuspend the gel pieces in 0.1% TFA containing 50% acetonitrile, vortex and remove supernatant. 7. Supernatants from step 6 can be analysed directly by MALDI mass spectrometry. 1 vol. of sample is mixed with 1 vol. of matrix solution such as 10% DHB in 50% acetonitrile/1% phosphoric acid or 10% alpha-cyano-4-hydroxycinnamic acid in 50% actonitrile and 0.5 mL spotted and dried onto a stainless steel MALDI target plate with the appropriate peptide standards. 8. The resulting MALDI spectra can be compared to in silico tryptic digests of the protein of interest or interrogated using software, such as MASCOT (Matrix Science). 9. More detailed analysis can be carried out using LC–MS/MS using instruments, such as a 4000 Qtrap or LTQ Orbitrap. Here, individual peptides are fragmented and the masses of the resulting fragments are accurately measured. MS/MS spectra are interrogated manually or using MASCOT software and from these, specific sites of phosphorylation may be determined. 3.3. Generation of Phosphorylation Site-Specific Antibodies 3.3.1. Antibody Production

Phosphorylation-site-specific antibodies are highly desirable tools for studying protein phosphorylation. Based upon data obtained from mass spectrometric analysis of receptors overexpressed in cell lines or in  vitro assays, antibodies can be raised to study ­phosphorylation of receptors expressed at very low levels in native tissue by techniques, such as western blotting or immunohistochemistry. Unlike regular polyclonal antibody production, the choice of immunising peptide is limited as it should be centred around the phosphorylation site of interest. Generally, a peptide of 15–16 amino acid residues is sufficient with the phosphoamino acid of interest around the centre. It is better to keep the  sequence as short as possible to restrict the production of ­non-phospho-specific antibodies and a corresponding non-­ phosphorylated version of the peptide is always made to aid

Examining Site-Specific GPCR Phosphorylation

247

purification­of the resulting anti-serum. As with regular anti-peptide polyclonal antibodies, a cysteine residue can be added to the N- or C-terminus to aid conjugation to the carrier protein, such as keyhole limpet haemocyanin (KLH). Design and synthesis of the immunising peptide is often offered as a service and can also be offered as a package which includes immunisation and bleeding­ of the host animals and can in some cases include purification of the resulting anti-serum. These programmes are an attractive and usually cheaper alternative to raising these antibodies in-house and although they may differ slightly from programme to ­programme, the basic principles are similar. For each antibody, 20 mg of each of two peptides are synthesised, one phosphopeptide and one non-phosphopeptide. The phosphopeptide is ­coupled to a carrier protein, mixed with an adjuvant and two ­rabbits are immunised over a 3-month programme with four immunisations and four bleeds. 1. Design and synthesise 20 mg each of two peptides, one phosphopeptide and one non-phosphopeptide. 2. Couple 10 mg of phosphopeptide to carrier protein and carry out immunisation of two rabbits using a 3-month programme consisting of four immunisations and four bleeds. 3. Anti-serum can be screened using ELISA assays to identify the best responding animal. 4. Double affinity purification of final bleed serum to capture phospho-specific antibodies and remove antibodies specific to the non-phosphorylated sequence. A column containing the phosphorylated peptide is used to capture all antibodies, the eluate from this column is then loaded on to a second column containing the non-phosphorylated peptide to remove antibodies specific to the non-phosphorylated sequence. The flow-through from this step should contain only phosphorylation-specific antibodies. 3.3.2. Characterisation of PhosphorylationSpecific Antibodies

Once the phosphorylation-specific antibodies have been isolated, it is worthwhile performing a basic characterisation before the antibodies are used experimentally. This confirms that ­phosphorylation-specific antibodies are not contaminated with non-phosphorylation-specific antibodies and vice versa. In our laboratory, we isolate the target protein by immunoprecipitation and treat with calf intestinal alkaline phosphatase (CIAP) to remove phosphorylation. It is then possible to compare the ­phospho-specific antibody’s ability to detect phosphorylation before and after CIAP treatment. In addition, GST fusion proteins can be generated from bacterial expression systems. These are not phosphorylated and so should not react with the phosphorylationspecific antibodies when tested in western blotting.

248

Butcher, Tobin, and Kong

1. Grow cells expressing the epitope-tagged (e.g. HA or FLAG) GPCR of interest to about 80–95% confluency in multi-6well plates. 2. Wash cells twice with Krebs/HEPES and incubate with Krebs/HEPES for 1 h at 37°C. Prepare 100× stock of agonist and incubate cells with agonist for required time. 3. Aspirate buffer and lyse cells by the addition of 300 mL/well ice cold RIPA buffer containing protease and phosphatase inhibitors. 4. After centrifugation at 24,000 × g for 10  min, immunoprecipitate the receptor of interest by the addition of 40 mL of anti-HA or anti-FLAG coupled to agarose beads followed by incubation at 4°C on a rolling platform for 1–2 h. 5. Wash the beads 3× with RIPA buffer without phosphatase inhibitors and twice with CIAP buffer supplemented with 0.25% n-octylglucoside. Re-suspend the beads in 50 mL CIAP buffer containing 0.25% n-octylglucoside and protease inhibitors and add 40 U CIAP. Incubate at 37°C overnight. 6. Wash the beads 3× with CIAP buffer supplemented with 0.25% n-octylglucoside and re-suspend in 50  mL Laemmli sample buffer. 7. Separate samples by SDS–PAGE and transfer to nitrocellulose as described in Subheading  3.1, step 1. After blocking the membranes, probe with GPCR-specific or phosphorylationspecific antibodies.

4. Notes 1. Although isotopic equilibrium point has not been reached with short-term labelling, longer labelling time was found to have no beneficial effects for the studies of the phosphorylation of the receptor but rather produced detrimental effects to the cells. Thus, we find that 1–2 h labelling time is sufficient for the purpose of the studies of receptor phosphorylation. 2. Due to the hydrophobic nature of GPCRs and most membrane proteins, boiling of samples for denaturation is not recommended as this can cause receptor aggregation. 3. When loading, it is sometime possible that the sample may appear like stringy glue. This would suggest that too many cells were used in the experiment resulting in DNA contamination which precipitates when Laemmli buffer is added. The remedy to this is to use less cells at the beginning of the experiment.

Examining Site-Specific GPCR Phosphorylation

249

4. The percentage of the gel run is dependent on the size (molecular weight) of the receptor. In the case of M3muscarinic receptors, which run as a band at approximately 100 kDa, we routinely use an 8% gel. For most GPCRs, an 8 or 10% gel is usually sufficient. GPCRs often run as broad bands of higher molecular weight due to the hydrophobic nature of the protein, hyper-phosphorylation state and glycosylation of extracellular domains. For example, the M3muscarinic receptor amino acid back bone is 66.2 kDa but on a gel it runs as a broad band at about 100–110 kDa. 5. Pre-stained marker is used as Ponceau S staining may affect the digestion of the proteins in subsequent steps. 6. Polyvinylidene fluoride (PVDF) membrane binds more strongly to the proteins which may affect the yield of digested protein in subsequent steps and therefore should be avoided. References 1. Lefkowitz, R. J. and Shenoy, S. K. (2005) Angiotensin II-stimulated signaling through G proteins and beta-arrestin. Science 308, 512–517. 2. Tobin, A. B. (2008) G-protein-coupled receptor phosphorylation: where, when and by whom. Br. J. Pharmacol. 153 Suppl 1, S167–176. 3. Tobin, A. B., Butcher, A. J., and Kong, K. C. (2008) Location, location, location…site-specific GPCR phosphorylation offers a mechanism for cell-type-specific signalling. Trends Pharmacol. Sci. 29, 413–420. 4. Blaukat, A., Pizard, A., Breit, A., Wernstedt, C., Alhenc-Gelas, F., Muller-Esterl, W., and Dikic, I. (2001) Determination of bradykinin B2 receptor in vivo phosphorylation sites and their role in receptor function. J. Biol. Chem. 276, 40431–40440. 5. Torrecilla, I., Spragg, E. J., Poulin, B., McWilliams, P. J., Mistry, S. C., Blaukat, A., and Tobin, A. B. (2007) Phosphorylation and regulation of a G protein-coupled receptor by protein kinase CK2. J. Cell Biol. 177, 127–137. 6. Trester-Zedlitz, M., Burlingame, A., Kobilka, B., and von Zastrow, M. (2005) Mass spectrometric analysis of agonist effects on posttranslational modifications of the beta-2 adrenoceptor in mammalian cells. Biochemistry 44, 6133–6143. 7. Jones, B. W., Song, G. J., Greuber, E. K., and Hinkle, P. M. (2007) Phosphorylation of the endogenous thyrotropin-releasing hormone receptor in pituitary GH3 cells and pituitary tissue revealed by phosphosite-specific antibodies. J. Biol. Chem. 282, 12893–12906.

8. Tran, T. M., Friedman, J., Qunaibi, E., Baameur, F., Moore, R. H., and Clark, R. B. (2004) Characterization of agonist stimulation of cAMP-dependent protein kinase and G protein-coupled receptor kinase phosphorylation of the beta2-adrenergic receptor using phosphoserine-specific antibodies. Mol. Pharmacol. 65, 196–206. 9. Kara, E., Crepieux, P., Gauthier, C., Martinat, N., Piketty, V., Guillou, F., and Reiter, E. (2006) A phosphorylation cluster of five serine and threonine residues in the C-terminus of the follicle-stimulating hormone receptor is important for desensitization but not for betaarrestin-mediated ERK activation. Mol. Endocrinol. 20, 3014–3026. 10. Mendez, A., Burns, M. E., Roca, A., Lem, J., Wu, L. W., Simon, M. I., Baylor, D. A., and Chen, J. (2000) Rapid and reproducible deactivation of rhodopsin requires multiple phosphorylation sites. Neuron 28, 153–164. 11. Seibold, A., Williams, B., Huang, Z. F., Friedman, J., Moore, R. H., Knoll, B. J., and Clark, R. B. (2000) Localization of the sites mediating desensitization of the beta(2)adrenergic receptor by the GRK pathway. Mol. Pharmacol. 58, 1162–1173. 12. Ong, S.-E. and Mann, M. (2005) Mass spectrometry-based proteomics turns quantitative. Nat. Chem. Biol. 1, 252–262. 13. Campbell, D. G. and Morrice, N. A. (2002) Identification of protein phosphorylation sites by a combination of mass spectrometry and solid phase Edman sequencing. J. Biomol. Tech. 13, 119–130.

wwwwwww

Chapter 13 Ubiquitination of GPCRs Adriana Caballero and Adriano Marchese Abstract In this chapter, we describe a method for detecting the ubiquitination status of G protein-coupled receptors (GPCRs). This involves co-expression of a GPCR with an epitope-tagged ubiquitin construct in a ­heterologous mammalian expression system. Stimulus-dependent modification of the GPCR by ­ubiquitin is detected by immunoprecipitation and subsequent immunoblotting to detect incorporation of the epitope-tagged ubiquitin. We describe here a well-established protocol to detect ubiquitination of the chemokine receptor CXCR4, which can be easily applied to detect the ubiquitination status of other GPCRs. Key words: Ubiquitin, G protein-coupled receptor, Agonist, Immunoblot, Lysosome, Sorting, Degradation, CXCR4, Immunoprecipitation, De-ubiquitination

1. Introduction Many agonist-activated G protein-coupled receptors (GPCRs) are subject to a complex series of events that may lead to their removal from the cell-surface and internalization into an endocytic compartment (1, 2). Once on endosomes, internalized receptors are subject to multiple sorting events that ultimately direct them to a recycling pathway and/or a degradative pathway. If a receptor enters the recycling pathway, it will be brought back to the plasma membrane allowing for a process known as resensitization. Alternatively, if a receptor enters the degradative pathway, it will traffic to lysosomes where it will be degraded, giving rise to a process known as downregulation. Very little is known about the molecular mechanisms mediating the sorting events linked to the postinternalization trafficking of GPCRs. However, we have shown that agonist-promoted ubiquitination of the

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_13, © Springer Science+Business Media, LLC 2011

251

252

Caballero and Marchese

chemokine receptor CXCR4 by the E3 ubiquitin ligase AIP4 is important for targeting the receptor to lysosomes (3, 4). A CXCR4 mutant in which three lysine residues in the C terminus were changed to arginine residues internalized normally, but failed to undergo agonist-promoted degradation (5). Likewise, in cells depleted of AIP4 by RNA interference CXCR4 internalized normally, but was unable to undergo agonist-promoted degradation as compared to control, AIP4-replete cells (4). Both the receptor mutant and AIP4 depletion attenuated CXCR4 ubiquitination, indicating that the ubiquitin moiety is necessary for targeting CXCR4 into a degradative compartment. Other GPCRs may also be targeted to degradative compartments by a similar ubiquitindependent endosomal pathway, although the ubiquitination machinery may be different (6, 7). Of note, direct ubiquitinmodification of some GPCRs may not necessarily constitute the lysosomal targeting signal (8–10). Ubiquitin is a highly conserved 76 amino acid polypeptide that is usually covalently attached to proteins through the formation of an isopeptide bond between the C-terminal glycine residue of ubiquitin and the e-amino group of lysine residues of target ­proteins (11). Through seven internal lysine residues ubiquitin can then be modified by itself to form polyubiquitin chains. However, the most common polyubiquitin chains observed are formed through linkages at K-48 and K-63 (12). Generally, K-48 linkages target ­proteins for proteolysis by the proteasome while K-63 linkages or monoubiquitin subserve nonproteasomal functions. Both K-63 and monoubiquitin have been linked to a role in endosomal trafficking of many integral membrane proteins (13, 14). Ubiquitin is involved in the internalization of cell-surface proteins and the sorting of membrane proteins into the multivesicular body (MVB), a morphologically distinct endosomal compartment containing intraluminal vesicles associated with endosomal degradation (15, 16). The ubiquitin moiety on receptors is recognized by proteins of the endosomal sorting complex required for transport (ESCRT) machinery, of which many components harbor ubiquitin binding domains (UBDs), thus providing the framework for a network of ubiquitin/UBD interactions that ultimately leads to the proper sorting and degradation of ubiquitinated cargo in lysosomes (15). A widely used method to detect the ubiquitination status of a protein is by immunoblotting. The covalent attachment of ubiquitin adds at least 8 kDa to the size of a protein, thereby leading to slower migration when subjected to SDS–PAGE analysis. The ubiquitinated forms can then be detected by immunoblotting using an antibody that recognizes the modified protein, as ubiquitinated forms migrate at ~8 kDa intervals above the unmodified version. However, because the ubiquitinated GPCR may represent only a small fraction of the total membrane pool of the GPCR, antibodies raised against GPCRs are often not sensitive

Ubiquitination of GPCRs

253

enough to detect the small amount of the ubiquitinated GPCR species. A more sensitive method of detecting the ubiquitination status of a GPCR is to concentrate the receptor, typically by immunoprecipitation, followed by SDS–PAGE and immunoblotting with an antibody against ubiquitin. We describe here a procedure that enables the detection of ubiquitinated chemokine receptor CXCR4. This procedure can readily be adapted and applied to detect the ubiquitination status of other members of the GPCR family.

2. Materials 1. Human embryonic kidney cells (HEK293; American Tissue Type Collection). 2. Dulbecco’s Modified Eagle’s medium (DMEM; Mediatech, Herndon, VA) and fetal bovine serum (FBS; Hyclone, Logan, UT). 3. Stromal-cell derived growth factor-1a (SDF-1a) (PeproTech, Rocky Hill, NJ) dissolved in phosphate-buffered saline (PBS) containing 0.1% bovine serum albumin. Note, SDF-1a is also known as CXCL12. 4. TransIT transfection reagent (Mirrus, Madison, WI) and protein-A agarose beads (Roche, Indianapolis, IN). 5. DNA expression constructs: HA-tagged CXCR4 in pcDNA3 (Invitrogen, Carlsbad, CA), FLAG-tagged ubiquitin in pCMV10 (Sigma, St. Louis, MO) (see Note 1). 6. PBS (Hyclone, Logan, UT). 7. Lysis/wash/immunoprecipitation buffer: 50  mM Tris–HCl (pH 8.0), 150 mM NaCl, 5 mM EDTA, 0.5% sodium deoxycholate (w/v), 1% Nonidet P-40 (NP40; v/v), 0.1% sodium dodecyl sulfate (SDS; w/v) (see Note 2). 8. Protease inhibitors are added fresh each time with the following final concentrations: 10 mg/mL aprotinin, 10 mg/mL leupeptin, 0.2  mg/mL benzamidine and 1  mg/mL ­pepstatin-A. Each inhibitor can be obtained from Roche (Indianapolis, IN). 9. N-ethylmaleimide (NEM; Sigma, St. Louis, MO) is added fresh each time to a final concentration of 20 mM (see Note 3). 10. SDS–PAGE gels: We typically use 7% gels cast before use (see Note 4). 11. Denaturation solution for immunoblots: 62.5 mM Tris–HCl (pH 6.7), 100 mM b-mercaptoethanol, 2% SDS (see Note 5).

254

Caballero and Marchese

12. Tris-buffered saline (50  mM Tris–HCl (pH 7.4), 150  mM NaCl) containing 0.05% Tween-20 (TBST). Also, TBST containing 5% (w/v) nonfat dry milk (TBST-5% milk). 13. Antibodies: HRP-labeled or unconjugated anti-FLAG M2 monoclonal antibody (Sigma, St. Louis, MO); anti-mouse IgG conjugated with horseradish peroxidase (if using unconjugated M2 antibody; Vector Laboratories, Burlingame, CA); mouse monoclonal anti-HA and rabbit anti-HA antibodies (Covance, Richmond, CA). 14. 2× sample buffer: 37.5 mM Tris–HCl (pH 6.5), 8% SDS, 10% glycerol, 5% b-mercaptoethanol, 0.003% bromophenol blue. 15. A high sensitivity chemiluminescent substrate, such as SuperSignal® West Dura Extended Duration substrate (Pierce, Rockford, IL).

3. Methods 1. HEK293 cells are maintained in DMEM supplemented with 10% FBS without antibiotics. 2. The day before transfection, passage cells onto 10-cm dishes so that they reach 50–60% confluency on the day of transfection (see Note 6). 3. Transiently transfect cells with 1  mg of DNA encoding HA-CXCR4, 3  mg of FLAG–ubiquitin and 6  mg of empty vector (pCMV or pcDNA, for a total of 10 mg of DNA per 10-cm dish/transfection) using the TransIT-LT1 transfection reagent, according to the manufacturer’s instructions (see Note 7). 4. Passage cells onto 6-cm dishes 24–48 h after FLAG–ubiquitin transfection (typically at 1:4 to 1:5 ratio) (see Note 8). 5. Wash cells by aspirating medium and replacing it with 4 mL of warm DMEM supplemented with 20  mM HEPES, pH 7.4. Repeat this step for a total of two washes. 6. Incubate cells in the same medium in the presence or absence of agonist (10–100 nM of SDF1-a in a volume of 1.5 mL) for 15–30 min at 37°C (see Note 9). 7. Rapidly wash cells once with 2 mL of ice-cold PBS. Aspirate PBS completely and place dishes on ice. 8. Add 1 mL of ice-cold lysis buffer. Scrape cells, and transfer to microcentrifuge tubes. Incubate tubes on ice for ~20 min to allow for complete solubilization (see Note 10).

Ubiquitination of GPCRs

255

9. Sonicate samples at 11% power using a microtip probe (Branson Digital Sonifier; Branson, Danbury, CT) for 10 s on ice (to minimize de-ubiquitination and/or proteolysis). 10. Pellet cellular debris by centrifuging samples at 16.3 × 1000 g for 20 min in a refrigerated table-top microcentrifuge (4°C). 11. Carefully transfer 500 mL of supernatant to a fresh microcentrifuge tube. Add 2.5  mL polyclonal anti-HA antibody and incubate for 1 h while rocking at 4°C (see Note 11). 12. Add 50 mL of protein A-agarose (50% slurry, previously equilibrated with lysis buffer) and continue incubation for an additional 1 h while rocking at 4°C. 13. Collect receptor-protein A-agarose complexes by centrifugation in a microcentrifuge at 16.3 × 1000 g for 5 s. 14. Carefully remove supernatant, resuspend beads in 750 mL of lysis buffer and incubate for 5 min at 4°C while rocking. 15. Repeat steps 13 and 14 two more times. 16. Collect complexes as in step 13 and carefully aspirate the last traces of lysis buffer using a vacuum line. Avoid aspirating pelleted beads. 17. Elute proteins from agarose beads by adding 25 mL of 2× SDSgel sample buffer followed by gently vortex-mixing. Incubate samples for 30 min at room temperature before ­proceeding to the next step (see Note 12). Alternatively, samples can be stored at −20°C or −80°C until further processing. 18. Resolve proteins by SDS–PAGE and transfer to nitrocellulose membranes using a standard Western blot protocol (17) (see Note 4). 19. Block the nitrocellulose membrane for 30 min at room temperature in 15 mL of TBST containing 5% (w/v) nonfat dry milk. 20. Incubate membrane with 10 mL of TBST-5% milk containing a 1:2,000 dilution of monoclonal anti-FLAG antibody (M2–HRP) for 1 h at room temperature (see Note 13). 21. Wash the nitrocellulose membrane five times, each for 10 min, in TBST. 22. Overlay the nitrocellulose with 1–2  mL of chemiluminescence reagent for ~5 min, allow the blot to drip dry, wrap in plastic wrap and visualize on X-ray film. 23. At this point blots can be treated with the denaturation solution to remove bound antibody and reprobed with the monoclonal HA antibody to detect receptor levels.

256

Caballero and Marchese

4. Notes 1. Although commercial anti-ubiquitin antibodies are available, we have had no success at detecting endogenous ubiquitin attached to CXCR4. Thus, in our studies we employ an epitopetagged version of ubiquitin in order to increase the sensitivity of detection. The presence of the tag does not appear to affect conjugation to substrate proteins (18). We subcloned yeast ubiquitin into p3 × FLAG–CMV-10 expression vector (Sigma; Cat# E4401), which carries an in-frame triple FLAG sequence upstream of the cloning site (5). Myc-tagged (18) and hemagglutinin (19) versions have also been used successfully. Several commercially available anti-ubiquitin antibodies from several sources are currently available that recognize specific types of linkages and/or monoubiquitin (20, 21). The anti-ubiquitin antibody clone P4D1 has been used successfully to detect ubiquitination of the b2-adrenergic receptor (22). 2. The recipe for immunoprecipitation buffers varies depending on the receptor being immunoprecipitated. Ideally, high stringency immunoprecipitation and washing conditions are ­preferred in order to reduce the likelihood of immuno­precipitating nonspecific proteins that may confound the ­interpretation of ubiquitin immunoblots. An additional and extremely powerful control would be to identify ubiquitin-deficient mutants to assure that the detection of ubiquitinated receptor species is not confounded by the presence of a ­co-immunoprecipitating protein that may itself be ubiquitinated. 3. As with other types of posttranslational modifications, ubiquitin attachment to proteins is highly dynamic and a tightly regulated process. Once attached to protein, ubiquitin may be removed by a specialized set of proteases known as deubiquitinating enzymes (DUBs), most of which have active site cysteine residues (23). One function of these enzymes is to recycle ubiquitin. Therefore, the ubiquitinated status of a protein is determined by the rate of ubiquitin conjugation and deconjugation. Because the activity of these enzymes can compromise the ability to detect ubiquitinated proteins, it is critical that they are inhibited when a cell lysate is prepared. Many DUBs contain active sulfhydryl groups that can be blocked by alkylating agents, such as iodoacetamide and NEM. Iodoacetamide can be used at a final concentration of 1  mg/mL. We have used NEM at a final concentration as high as 20 mM. These agents should be added fresh to the lysis buffer just before use. 4. We typically use 7% gels because it helps resolve the higher molecular weight species of ubiquitinated receptor.

Ubiquitination of GPCRs

257

5. Even though proteins subjected to SDS–PAGE are significantly denatured, we have found that further treatment of blots in this denaturating solution sometimes enhances the ability to detect ubiquitinated receptor. Thus, nitrocellulose membranes can be incubated with denaturation solution for 30  min at 60°C followed by extensive washing with TBST before blocking and immunoblotting for the detection of ubiquitin. 6. We typically perform experiments 48–72  h after FLAGubiquitin transfection. Therefore, cells should be passaged at an appropriate density such that they are not overgrown the day of the experiment. This variable should be determined empirically. 7. For the yeast a-mating factor receptor, the ability to detect ubiquitinated receptor is enhanced in a yeast strain that is defective in clathrin-mediated endocytosis (24). This is due in part to the ubiquitination reaction occurring at the plasma membrane, thus increasing the proportion of receptors that are modified by ubiquitin, but also because it prevents the receptor from entering a compartment where it is de-ubiquitinated (25). Similarly, co-expression of 1  mg of the dynamin dominant-­ negative K44A mutant, which inhibits CXCR4 internalization, facilitates the detection of ubiquitinated CXCR4 (5). This is potentially useful for the enrichment of ubiquitinated forms of other GPCRs when detection is problematic (26). 8. We have determined that overexpression of epitope-tagged ubiquitin is the most important variable in the successful detection of ubiquitinated receptors. High expression levels can be achieved by allowing an extra day for FLAG-ubiquitin expression (72 h posttransfection) and/or by maximizing the delivery of plasmids with high-performance/cell-specific transfection reagents (e.g., Mirrus TransIT-293 for HEK293 cells). This should be empirically determined for each cell-type used. 9. We initially performed a time-course to determine the optimal length of stimulation that resulted in maximal ubiquitination of CXCR4. Depending on the treatment/agonist concentration, ubiquitination can be detected as early as 10 min. We found that 30 min of stimulation led to the greatest ubiquitination under our conditions; however, this has to be determined for other GPCRs. 10. Membrane solubilization is a particularly important variable in the successful completion of ubiquitination experiments. We have determined that a 70–90% confluent 6-cm dish of HEK293 cells is efficiently solubilized with 1 mL of our lysis buffer. Importantly, increasing cell number or decreasing lysate volume (thus effectively increasing lysate protein concentration) might even be detrimental to ubiquitin detection

258

Caballero and Marchese

by altering CXCR4 solubilization. Because this varies depending on the GPCR and cell-type used, we highly recommend empirically to determine the most efficient solubilization conditions for the cell-type and number of cells to be utilized. In addition, longer incubations while rocking at 4°C can also be performed to allow for complete membrane solubilization. 11. We have been able to successfully immunoprecipitate and detect ubiquitinated CXCR4 from 100 to 250  mg of lysate prepared as described above. For more accurate comparisons between samples, protein amounts in lysates can be quantified using BCA assay (Pierce, Rockford, IL). In addition, incubation times should be kept to a minimum following lysate generation due to the high activity of DUBs present in the lysate. 12. It is critical that samples are not boiled to elute CXCR4 off of the beads. Boiling causes receptors to aggregate and impede their migration on SDS–PAGE. 13. The unconjugated mouse anti-FLAG (M2) can also be used for additional signal amplification and detection of ubiquitin with equivalent results, followed by incubation with a goat antimouse IgG conjugated to horseradish peroxidase (1:10,000).

Acknowledgments Work was supported by NIH grants GM075159 and DA026040 to A.M. References 1. Marchese, A. Paing, M. M. Temple, B. R. and Trejo, J. (2008) G protein-coupled receptor sorting to endosomes and lysosomes Annu. Rev. Pharmacol. Toxicol. 48, 601–629. 2. Hanyaloglu, A. C. and von Zastrow, M. (2008) Regulation of GPCRs by endocytic membrane trafficking and its potential implications. Annu. Rev. Pharmacol. Toxicol. 48, 537–568. 3. Bhandari, D. Robia, S. L. and Marchese, A. (2009) The E3 ubiquitin ligase atrophin interacting protein 4 binds directly to the chemokine receptor CXCR4 via a novel WW domain-mediated interaction. Mol. Biol. Cell. 20, 1324–1339. 4. Marchese, A. Raiborg, C. Santini, F. Keen, J. H. Stenmark, H. and Benovic, J. L. (2003) The E3 ubiquitin ligase AIP4 mediates ubiquit-

ination and sorting of the G protein-coupled receptor CXCR4. Dev. Cell 5, 709–722. 5. Marchese, A. and Benovic, J. L. (2001) Agonistpromoted ubiquitination of the G protein-coupled receptor CXCR4 mediates lysosomal sorting J. Biol. Chem. 276, 45509–45512. 6. Jacob, C. Cottrell, G. S. Gehringer, D. Schmidlin, F. Grady, E. F. and Bunnett, N. W. (2005) c-Cbl mediates ubiquitination, degradation, and down-regulation of human protease-activated receptor 2. J. Biol. Chem. 280, 16076–16087. 7. Shenoy, S. K. Xiao, K. Venkataramanan, V. Snyder, P. M. Freedman, N. J. and Weissman, A. M. (2008) Nedd4 mediates agonist-dependent ubiquitination, lysosomal targeting, and degradation of the b2-adrenergic receptor. J. Biol. Chem. 283, 22166–22176.

Ubiquitination of GPCRs 8. Tanowitz, M. and Von Zastrow, M. (2002) Ubiquitination-independent trafficking of G protein-coupled receptors to lysosomes. J. Biol. Chem. 277, 50219–50222. 9. Baugher, P. J. and Richmond, A. (2008) The carboxyl-terminal PDZ ligand motif of chemokine receptor CXCR2 modulates postendocytic sorting and cellular chemotaxis. J. Biol. Chem. 283, 30868–30878. 10. Meiser, A. Mueller, A. Wise, E. L. McDonagh, E. M. Petit, S. J. Saran, N. Clark, P. C. Williams, T. J. and Pease, J. E. (2008) The chemokine receptor CXCR3 is degraded following internalization and is replenished at the cell surface by de  novo synthesis of receptor. J. Immunol. 180, 6713–6724. 11. Weissman, A. M. (2001) Themes and variations on ubiquitylation. Nat. Rev. Mol. Cell. Biol. 2, 169–178. 12. Ikeda, F. and Dikic, I. (2008) Atypical ubiquitin chains: new molecular signals. EMBO Rep. 9, 536–542. 13. Bonifacino, J. S. and Weissman, A. M. (1998) Ubiquitin and the control of protein fate in the secretory and endocytic pathways. Annu. Rev. Cell. Dev. Biol. 14, 19–57. 14. Hicke, L. (2001) Protein regulation by monoubiquitin. Nat. Rev. Mol. Cell. Biol. 2, 195–201. 15. Raiborg, C. and Stenmark, H. (2009) The ESCRT machinery in endosomal sorting of ubiquitylated membrane proteins. Nature 458, 445–452. 16. Clague, M. J. and Urbe, S. (2008) Multivesicular bodies. Curr. Biol. 18, R402–R404. 17. Mundell, S. J. Orsini, M. J. and Benovic, J. L. (2002) Characterization of arrestin expression and function. Methods Enzymol. 343, 600–601. 18. Ellison, M. J. and Hochstrasser, M. (1991) Epitope-tagged ubiquitin. A new probe for analyzing ubiquitin function. J. Biol. Chem. 266, 21150–21157.

259

19. Courbard, J. R. Fiore, F. Adelaide, J. Borg, J. P. Birnbaum, D. and Ollendorff, V. (2002) Interaction between two ubiquitin-protein ­isopeptide ligases of different classes, CBLC and AIP4/ITCH. J. Biol. Chem. 277, 45267–45275. 20. Wang, H. Matsuzawa, A. Brown, S. A. Zhou, J. Guy, C. S. Tseng, P. H. Forbes, K. Nicholson, T. P. Sheppard, P. W. Hacker, H. Karin, M. and Vignali, D. A. (2008) Analysis of nondegradative protein ubiquitylation with a monoclonal antibody specific for lysine-63-linked polyubiquitin. Proc. Natl. Acad. Sci. USA 105, 20197–20202. 21. Fujimuro, M. Sawada, H. and Yokosawa, H. (1994) Production and characterization of monoclonal antibodies specific to multi-­ ubiquitin chains of polyubiquitinated proteins. FEBS Lett. 349, 173–180. 22. Shenoy, S. K. McDonald, P. H. Kohout, T. A. and Lefkowitz, R. J. (2001) Regulation of receptor fate by ubiquitination of activated b2adrenergic receptor and b-arrestin. Science 294, 1307–1313. 23. Wilkinson, K. D. (2000) Ubiquitination and deubiquitination: targeting of proteins for degradation by the proteasome. Semin. Cell Dev. Biol. 11, 141–148. 24. Hicke, L. Zanolari, B. and Riezman, H. (1998) Cytoplasmic tail phosphorylation of the alphafactor receptor is required for its ubiquitination and internalization. J. Cell Biol. 141, 349–358. 25. Amerik, A. Y. Nowak, J. Swaminathan, S. and Hochstrasser, M. (2000) The Doa4 deubiquitinating enzyme is functionally linked to the vacuolar protein-sorting and endocytic pathways. Mol. Biol. Cell. 11, 3365–3380. 26. Wolfe, B. L. Marchese, A. and Trejo, J. (2007) Ubiquitination differentially regulates clathrindependent internalization of protease-activated receptor-1. J. Cell Biol. 177, 905–916.

wwwwwww

Part III Examining Early Events in GPCR Signaling

wwwwwww

Chapter 14 [35S]GTPgS Binding as an Index of Total G-Protein and Ga-Subtype-Specific Activation by GPCRs Rajendra Mistry, Mark R. Dowling, and R.A. John Challiss Abstract On activation, G-protein-coupled receptors (GPCRs) exert many of their cellular actions through ­promoting guanine nucleotide exchange on Ga subunits of heterotrimeric G proteins to release Ga-GTP and free bg-subunits. In membrane preparations, GTP can be substituted by 35S-labeled guanosine­ 5¢-O-(3-thio)triphosphate ([35S]GTPgS) and on agonist stimulation a quasi-stable [35S] GTPgS–Ga ­complex forms and accumulates. Separation of [35S]GTPgS–Ga complexes from free [35S] GTPgS allows differences between basal and agonist-stimulated rates of [35S]GTPgS–Ga complex formation­to be used to obtain pharmacological information on receptor–G-protein information transfer. Further, by releasing Ga-subunits into solution following the [35S]GTPgS binding step, Ga-subunit-specific antibodies can be used to investigate the Ga-protein subpopulations activated by receptors by immunoprecipitation of [35S]GTPgS–Ga complexes and quantification by scintillation counting. Here, we describe a total [35S]GTPgS binding assay and a modification of this method that incorporates a Gaspecific immunoprecipitation step. Key words: [35S]GTPgS, Guanine nucleotide binding protein, G-protein-coupled receptor, GDP, Receptor–G-protein coupling, Ga-protein-specific immunoprecipitation

1. Introduction The use of agonist-promoted binding of [35S]GTPgS to evaluate the interactions between G-protein-coupled receptors (GPCRs) and G-proteins emerged from experimental attempts to label guanine nucleotide-binding proteins (1). It quickly became apparent that the binding of [35S]GTPgS could be used to assess receptor– G-protein interactions, initially in reconstitution studies (2), and then in membrane preparations (3). Thus, receptor activation stimulates GTP for GDP exchange at the Ga-subunit of G-protein

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_14, © Springer Science+Business Media, LLC 2011

263

264

Mistry, Dowling, and Challiss

heterotrimers. If GTPgS is present, it will compete with GTP for the Ga-GTP binding site and will form a quasi-irreversible Ga– GTPgS complex, as the phosphorothioate group of GTPgS makes this analog highly resistant to Ga-GTPase activity. Therefore, basal and agonist-stimulated rates of GDP/[35S]GTPgS exchange can be compared to gain information about the initial postligand binding step in the GPCR signal transduction process. The [35S]GTPgS binding assay is extremely attractive as it is relatively simple to perform, can be adapted to high-throughput formats, and provides an efficacy readout proximal to the receptor– ligand-binding event. The original Ga protein binding assay has been widely used to profile pharmacologically agonists, antagonists, and inverse agonists at a number of receptors and has been adapted for thin slice/autoradiography, as well as cellular and tissue membrane preparations (3–5). The method is generally satisfactory for studies involving GPCRs that preferentially couple to Gi/o-proteins; however, it has been less used to study GPCRs that couple preferentially to the Gq, Gs, and G12 subfamilies of G-protein. This is because GDP/[35S]GTPgS exchange on these other Ga-proteins can be masked by high basal GDP/[35S]GTPgS exchange rates for the Gi/o-protein subpopulation (6). More recently, elaborations on the original [35S]GTPgS-binding assay have been applied to gain semiquantitative information on the Ga-protein subpopulation(s) activated by the ligand–receptor interaction (7–9). One strategy is to use a photolabile GTP analog that can be covalently attached to its Ga-protein-binding partner(s). Alternatively, an antibody-based method can be used that avoids the need to synthesize and utilize 32P-labeled GTP analogs. This method relies on the availability of Ga-protein subtype-specific antibodies that can be used to immunoprecipitate Ga-subunits to quantify [35S]GTPgS binding to specific Ga-protein species. In addition to providing important information on receptor–Ga-protein-coupling preferences, the method can also substantially improve signal-tonoise in the assay, as it can separate the Ga-specific signal from the (high) basal Ga–[35S]GTPgS binding occurring within the preparation, opening the way to studies of GPCR coupling to Gq, Gs, and G12 subfamilies (9–11). Further, the assay can be adapted to a homogeneous assay format not requiring a separation step (12, 13).

2. Materials 1. [35S]GTPgS [~1,250 Ci (46.25 TBq)/mmol] (see Note 1). 2. Cell preparation: lifting buffer: 10 mM HEPES, 0.9% NaCl, 0.2% EDTA, pH 7.4. 3. Homogenization buffer A – 10 mM HEPES, 10 mM EDTA, pH 7.4.

[35S]GTPgS Binding as an Index of Total G-Protein

265

4. Homogenization buffer B – 10  mM HEPES, 0.1  mM EDTA, pH 7.4 (see Note 2). 5. Assay buffer: 10 mM HEPES, 100 mM NaCl, 10 mM MgCl2, pH 7.4 (see Note 3). 6. GDP, Na salt; GTPgS. 7. Filter-wash buffer: 10 mM HEPES, 100 mM NaCl, 10 mM MgCl2, pH 7.4 at 4°C. 8. Solubilization buffer: 100  mM Tris–HCl, 200  mM NaCl, 1 mM EDTA, 1.25% Igepal CA 630 (see Note 4), pH 7.4; ±0.2% (w/v) SDS. 9. Normal rabbit serum. 10. Protein-A sepharose: 1.5  g diluted in 50  mL 10  mM Tris– HCl, 1 mM EDTA, pH 7.4. The suspension should be stored at 4°C and replaced every 2–3 weeks. 11. Ga-protein-specific antibodies (see Note 5). 12. Scintillant: standard liquid scintillation counting cocktails (e.g., SafeFluor, Lumac-LSC, Groningen, The Netherlands) are adequate and should give 90%+ counting efficiency. 13. Equipment: (1) Total [35S]GTPgS assays: manual filtration ­manifold/vacuum pump or 24/48/96 place cell harvester; scintillation b-counter. (2) [35S]GTPgS/Ga-immunoprecipitation assays: temperature controlled water-bath; high-speed, refrigerated centrifuge; shaker/roller; scintillation b-counter.

3. Methods 3.1. Preparation of Membranes

The [35S]GTPgS-binding method has been applied most commonly to membranes derived from cultured cell-lines expressing either endogenously or recombinantly the receptors and/or G-protein subpopulations being investigated. The membranes generated can also be used for complementary radioligand-binding studies characterizing the ligand–receptor interface. 1. Medium from confluent flask(s) of cells is removed and the cell monolayer briefly washed warm HEPES-buffered saline (10 mM HEPES, 154 mM NaCl, pH 7.4) (2 × 10 mL). 2. Warm lifting buffer (10 mL) is then added for ~5–10 min (or as long as it takes for the cell monolayer to detach from the plastic) and the cells poured into 30-mL tubes on ice. The flask is rinsed with a further 5 mL of lifting buffer and this too is added to each tube. All subsequent steps are conducted at 0–4°C. 3. The cell suspension is centrifuged (~250 × g for 5 min at 4°C) and the supernatant aspirated.

266

Mistry, Dowling, and Challiss

4. The cell pellet is dispersed in 5 mL of Buffer A and homogenized (Polytron, speed £ 20,000 rpm, 5 × 10 s bursts). The homogenate is then transferred to centrifuge tubes diluted to ~50 mL with Buffer A and centrifuged (5,000 × g, 10 min, 4°C). 5. The supernatant is carefully decanted into a fresh centrifuge tube and centrifuged (50,000 × g, 15 min, 4°C). The supernatant is discarded, the pellet resuspended in Buffer B, rehomogenized and recentrifuged as above (50,000 × g, 15 min, 4°C). 6. The final “P2” pellet is resuspended/rehomogenized in Buffer B and its protein concentration determined. The solution is diluted to the required concentration (either 1 or 3 mg protein/mL) and if not used immediately “snap-frozen” in liquid nitrogen and stored at −80°C until required. 3.2. Total [35S] GTP gS-Binding Assay

1. Fresh or frozen membranes should be diluted to an appropriate protein concentration (determined empirically; see Note 6) in ice-cold assay buffer. Typically, 20–100 mg protein per tube is used, depending on the level of receptor expression, the agonist utilized (e.g., partial vs. full) and the predominant Ga protein to which the receptor couples. 2. As the assay is to be terminated by rapid vacuum filtration, relatively high volume tubes (5 mL) are used. GDP is added to assay buffer to give the correct final concentration and [35S]GTPgS is added to give a concentration of ~0.2 nM in a final assay volume of 100 mL (i.e., ~200,000 dpm per assay). To provide an indication of nonspecific binding, unlabeled GTPgS (10 mM) is added to some tubes. 3. Both assay mixture and membranes are brought to 30°C, and membranes added to each assay tube. Reactions are initiated by adding either agonist or vehicle (i.e., the diluent for the agonist) and tubes incubated at 30°C for a predetermined time (usually 30–90 min). 4. Toward the end of the incubation period glass fiber filters (GF/B, Whatman) are wetted with ice-cold filter-wash buffer. Incubations are terminated by the addition of 3 mL ice-cold filter-wash buffer and rapid transfer on to the filters. Tubes/ filters are then washed with 3 × 3 mL filter-wash buffer. 5. Filters containing the membrane-Ga–[35S]GTPgS complex are allowed to dry (either under vacuum or air-dry), transferred to vials and scintillant added. Aliquots of the stock [35S] GTPgS solution are also taken to determine precisely how much radioactivity was added to each assay tube. An example of the data obtained using the above protocol is shown in Fig. 1a, which illustrates the importance of optimizing the GDP concentration in order to maximize signal-to-noise in the assay.

[35S]GTPgS Binding as an Index of Total G-Protein

267

Fig. 1.  Total (a) and Gai-specific (b) [35S]GTPgS–G protein binding in membranes prepared from Chinese hamster ovary cells, recombinantly expressing the human M2 muscarinic acetylcholine receptor (CHO-m2; receptor density = 880  fmol/mg protein). (a) CHO-m2 membranes (50 mg protein) were incubated in the presence of 0.1, 1, or 10 mM GDP for 10 min before incubation with 0.5 nM [35S]GTPgS and methacholine (MCh; 300 mM) or vehicle (basal) at 30°C for 45 min. Basal and agonist-stimulated binding was measured following rapid filtration. (b) CHO-m2 membranes (75 mg protein) were incubated in the presence of 0.1, 1, or 10 mM GDP for 10 min before incubation with 2 nM [35S]GTPgS and methacholine (MCh; 300  mM) or vehicle (basal) at 30°C for 2  min. Basal and agoniststimulated binding was measured following termination by centrifugation, solubilization of membrane pellets, and immunoprecipitation of Gai proteins using a Gai1/2-specific antibody. Representative experiments (performed in triplicate) are shown. For both total and Gai-specific assays the greatest agonist-stimulated increase was observed in the presence of 10 mM GDP: in the total assay, this increase was 52% over-basal, whereas in the Gai-specific assay this increase was 95%. In both cases, preincubation of membranes (for 30 min) with atropine (1 mM) did not affect basal [35S]GTPgS-G protein binding, but completely prevented the agonist-stimulated change (data not shown).

268

Mistry, Dowling, and Challiss

3.3. Ga-Specific [35S] GTP gS-Binding Assay

The assay described above gives a reliable indication of receptor activation that reflects an overall G-protein population activation. Using pertussis toxin (PTx) pretreatment it has been shown that many receptors can couple to both PTx-sensitive and -insensitive Ga-subunits (14). Furthermore, pharmacological observations have provided evidence that different agonists may be able to activate different Ga-subunit complements via the same receptor subtype (15) and has been termed “agonist-directed trafficking of receptor stimulus” (15, 16). Many of these studies have measured distal, functional readouts of receptor activation (e.g., Li+ enhanced [3H]inositol phosphate accumulations; reporter gene assays) where additional inputs, such as pathway cross talk and amplification may affect interpretation. The ability to measure an immediate consequence of receptor activation, Ga-protein GDP/GTP exchange, and further to dissect the Ga-species involved, can provide crucial and definitive pharmacological information. 1. Fresh or frozen membranes (prepared as described in Subheading 3.1) should be diluted to an appropriate protein concentration (see Note 7), typically 50–100 mg per tube is used. 2. As the assay requires several rolling and centrifugation steps, the most convenient tubes to employ are 1.5 mL Eppendorftype tubes. Tubes containing assay buffer, and appropriate concentrations of GDP (Fig.  1b), agonist, and [35S]GTPgS (see Notes 8 and 9) in a volume of 50 mL are maintained at 30°C. Membranes (50 mL) are then added to start the assay. 3. Incubations are continued at 30°C for a predetermined time (see Note 10). 4. Reactions are terminated by the addition of 1  mL ice-cold assay buffer and rapid transfer to a refrigerated centrifuge (prechilled to 4°C). Samples are centrifuged (20,000 × g, 6 min, 4°C). 5. The supernatant is then aspirated to leave a membrane pellet and 50 mL of ice-cold solubilization buffer (containing 0.2% SDS) is added. The pellet is left (on ice) for 15–30 min before attempting to aid the solubilization process (see Note 11). 6. Following complete solubilization of the pellet, the SDS concentration is reduced by the addition of 50 mL ice-cold solubilization buffer (minus SDS). 7. The solution is then precleared by the addition of 100 mL normal rabbit serum (prediluted 1:10 in assay buffer) and 30 mL of the 3% protein-A sepharose suspension. 8. Samples are capped and rolled at 4°C for a minimum of 90 min. 9. Samples are transferred to a refrigerated centrifuge and the protein-A sepharose conjugated material pelleted (20,000 × g, 2–3 min, 4°C).

[35S]GTPgS Binding as an Index of Total G-Protein

269

10. Transfer 100  mL of each supernatant to a fresh tube containing an appropriate dilution of anti-Ga-subunit antibody (see Note 12). 11. Vortex and roll overnight at 4°C. 12. Add 70 mL of protein-A sepharose suspension (see Note 13) and roll as before for 90 min. 13. Centrifuge (20,000 × g, 2–3 min, 4°C). 14. Aspirate the supernatant and add 1 mL ice-cold solubilization buffer (minus SDS). 15. Thoroughly vortex and centrifuge as before. 16. Repeat steps 13–15 so that the beads are washed four times in solubilization buffer (see Note 14). 17. After the final wash, the supernatant is removed and the protein-A beads resuspended in 1.1 mL FloScint IV (or similar scintillant) and vortex-mixed. Radioactivity is detected by liquid scintillation spectrometry. 3.4. Data Analysis

If the specific activity of the [35S]GTPgS is calculated daily, to ensure a consistent final concentration of radioactivity is added, the data are usually comparable to allow a simple plotting of specific cpm/dpm (minus nonspecific binding) versus agonist concentration. Log concentration-response curves can then be analyzed by nonlinear regression using an appropriate commercially available program (e.g., Prism 5, GraphPad Software, San Diego, USA) to produce EC50, Emax, and slope factor values. Alternatively, if the activity of the [35S]GTPgS is routinely counted and the protein added per tube is accurately known, data may be converted into femtomole [35S]GTPgS bound per milligram protein (or femtomole receptor). If there is daily variability between datasets, this may be improved by referencing the results obtained to a maximum concentration of a known full agonist (obtained on the day) and converting the data into a percentage of this maximum. Time-dependency and concentration-response curves for isoproterenol-stimulated Gsa-coupling response in Chinese hamster ovary cell membranes recombinantly expressing the b2 adrenergic receptor are shown in Fig. 2.

4. Notes 1. High specific activity [35S]GTPgS (40–50  TBq/mmol; Cat. No. NEG030H) is supplied by Perkin-Elmer Inc. (Waltham, MA). This product is shipped in a 20  mL volume, which is diluted to 250  mL on arrival with 10  mM Tricine, 10  mM DTT, pH 7.6, to give an ~800 nM stock solution, which is then dispensed as 20  mL aliquots and stored at −80°C.

270

Mistry, Dowling, and Challiss

Fig.  2.  Time course (a) and concentration dependency (b) of isoproterenol (ISO)stimulated [35S]GTPgS binding to Gsa-specific proteins in CHO-b2 membranes. CHO-b2 cell membranes (b2-adrenergic receptor density = 285 fmol/mg protein) were preincubated with GDP (1 mM) for 10 min before the addition of isoproterenol for a further 2 min at 30°C before the addition of 1 nM [35S]GTPgS. In the time-course experiment (a), isoproterenol (1 mM) was added for 2 min and incubations continued for 1, 2, or 5 min after [35S]GTPgS addition. The zero-time point indicates nonspecific binding (assessed in the presence of 10 mM GTPgS). At 5 min a 7.5-fold increase in Gsa–[35S]GTPgS binding was seen in the presence of isoproterenol. Concentration dependency (b) was assessed in the presence of the indicated concentrations of isoproterenol (for 2 min) followed by the addition of [35S]GTPgS for 5 min. The EC50 value for isoproterenol-stimulated Gsa–[35S] GTPgS binding was ~15 nM [pEC50 (M) = 7.83 ± 0.08; n = 3].

[35S]GTPgS Binding as an Index of Total G-Protein

271

American Radiolabeled Chemicals (ARC) Inc. (St. Louis, MO) supply a similar product (ARS-0124). The relatively short half-life of 35S (87.4 days) needs to be accounted for so as to ensure that similar amounts of radioactivity are used in the [35S]GTPgS assay. It is recommended that [35S]GTPgS is used within 2 months of its activity date. 2. This is the basic homogenization buffer – in our experience, superior results can sometimes be obtained by the addition of dithiothreitol (1  mM), MgCl2 (1  mM), and/or a protease inhibitor cocktail. This is more likely to be the case when preparing membranes from tissue (e.g., brain) samples; however, whether any/all of these additions that are necessary should be determined by trial and error. 3. Variations in assay concentrations of Na+, Mg2+, and GDP have been reported by a number of different groups. Preliminary experiments should be performed to assess the effects of ­varying Na+ and Mg2+ concentrations. Most assays use a high salt (Na+) level (routinely 100  mM), although low salt (Na+ = 10 mM) has been proposed to favor some phenomena (e.g., observation of inverse agonist activity). Similarly, a relatively high [Mg2+] (10  mM) is used in most studies. In common with other groups, we have found [GDP] to be a crucial determinant of whether an agonist-stimulated increase in [35S]GTPgS binding is observed (see Fig. 1) and in setting the signal-to-noise of the experimental system. A rule of thumb is that optimal conditions for observing receptor–Gi/o-protein interactions tend to occur at higher concentrations of GDP (10  mM) than for receptor–Gq/11 (0.1–1  mM GDP) or –Gs (1 mM GDP) interactions, however, preliminary experiments should always be performed to ensure that an optimal signalto-noise is achieved. Note also that it may be necessary to use more than one [GDP] to gain a complete experimental picture – e.g., where PTx is used to study receptor coupling to PTxsensitive (Gi/o) and PTx-insensitive G protein subpopulations. 4. To obtain reproducible results from the [35S]GTPgS–Ga/ immunoprecipitation method, it is essential that the membrane pellet is solubilized efficiently. The nonionic detergent Igepal CA-630 [(octylphenoxyl)polyethoxylethanol – SigmaAldrich], which has replaced nonidet P-40 (no longer available), together with sodium dodecylsulfate (SDS) appear to give efficient solubilization without destroying the [35S] GTPgS–Ga association. Optimal solubilization without loss of signal again requires empirical determination. As a rule of thumb, where small amounts of membrane (e.g., <15 mg per assay tube) are used Igepal CA-630 is usually sufficient alone to achieve solubilization and indeed it may be possible to

272

Mistry, Dowling, and Challiss

reduce the concentration of this nonionic detergent in the solubilization buffer. Note also that it might be possible to exclude the termination/centrifugation step and simply to transfer sample tubes at the end of the incubation period to an ice bath and add ice-cold solubilization buffer. This has the potential to avoid the need for harsher solubilization conditions and to remove sometimes the tedious manual labor involved in resolubilizing a tightly packed membrane pellet. 5. There is an ever increasing array of commercially available Ga-protein-specific antibodies. We have achieved satisfactory results with the following Ga-protein antisera from Santa Cruz Biotechnology (Santa Cruz, USA): “pan” Gi1–3a antibody (Gia common; sc-262); Gq/11a (sc-392); Goa (sc-13532); and Gsa (sc-823). If an extensive program of work is to be based on the [35S]GTPgS/immunoprecipitation strategy, then it may be financially prudent to generate an antiserum “in-house.” For example, we have generated our own immunoprecipitating Gq/11a antibody in rabbits against the C-terminal Gq/11a-common sequence (C)LQLNLKEYNLV (where the cysteine has been added for conjugation purposes). This sequence is not only common to human Gqa and G11a, but is also conserved across mammalian species. We are also aware of other groups having good success in raising high-quality antibodies to Gi1–2a, Go, and Gsa. When working with uncharacterized antibodies, experiments must be performed to assess the selectivity/specificity of the antibody for a specific Gasubunit over the others in the cell Ga population. 6. Preliminary experiments need to be performed to establish whether similar [35S]GTPgS-binding data are achieved using fresh or frozen membranes. In addition, there are a number of variables in the [35S]GTPgS-binding assay that need to be investigated to optimize signal-to-noise (i.e., the amount of agonist-stimulated [35S]GTPgS binding relative to the basal level). The most influential assay constituent is GDP. To optimize assay conditions, a sufficient level of GDP needs to be used to suppress basal [35S]GTPgS binding to G proteins and to reveal the maximal agonist-stimulated [35S]GTPgS binding compared to basal [35S]GTPgS-binding activity. The concentration of GDP may also influence the time that the incubation needs to be continued to achieve best results. For a Gi/o-coupled GPCR, a concentration of GDP of 10  mM is commonly used while for Gq/11-coupled GPCRs lower ­concentrations (0.1–1 mM) are generally employed. 7. As this assay ultimately relies on immunoprecipitation (which may not be 100% efficient) and separation of a subfraction of the total Ga-protein population activated, it is often necessary to use a higher protein concentration than may be employed

[35S]GTPgS Binding as an Index of Total G-Protein

273

in the total [35S]GTPgS assay. The optimization of protein concentration should take into account the balance between obtaining a sufficient capture of Ga–[35S]GTPgS complexes with the need to thoroughly solubilize the protein pellet, which becomes increasingly problematic as the amount of material used, and hence the size of the membrane pellet, increases (see also Note 4 above). 8. Depending on Ga-subtype being investigated, the order of addition can vary. While it is usual for membranes to be preincubated with GDP, whether agonist is added prior to, simultaneous with, or after [35S]GTPgS varies. For receptor– Gaq/11 and –Gai/o coupling, these letter two components are often added to GDP-treated membranes together, however, for receptor-Gas coupling much better results are obtained is the GDP-treated membranes are first exposed to agonist and [35S]GTPgS is the final component added to the assay (see Fig. 2). 9. For the reasons given under Note 8, it is necessary to increase the final concentration of [35S]GTPgS. Typically, a final concentration of 2–10 nM is employed (i.e., up to ~2.5 × 106 dpm [35S]GTPgS per assay tube). 10. Because a higher [35S]GTPgS concentration is employed, the rate of [35S]GTPgS–Ga association is more rapid. Careful optimization of the reaction time needs to be established as maximal [35S]GTPgS that may be achieved in as little as 2 min (9). There is also evidence, at least for some Ga-subtypes, that GTP/GTPgS-liganded Ga may also be delipidated in vivo. This would allow membrane dissociation and nonrecovery in the initial membrane pellet. For this reason, agonist incubation periods tend to be minimized. 11. Careful attention needs to be given to the solubilization ­process, as any unsolubilized membrane fragments partition with immunoprecipitated complexes and can therefore ­dramatically affect the quality and reproducibility of data. We have found that 60  min in a vortex-shaker, following by a brief vortexing by hand and a subsequent 60 min in a vortexshaker produces full solubilization without any detrimental effect on the maintenance of [35S]GTPgS–Ga complexes. 12. The final concentration of antibody used needs to be optimized not only for each antibody, but also for each antibody batch if the antibody is supplied commercially. The optimization procedure must balance the need to immunoprecipitate a significant fraction of the relevant Ga-protein against cost considerations. Typically, a good antibody should produce efficient immunoprecipitation of the Ga-protein at a final dilution of 1:100 or greater.

274

Mistry, Dowling, and Challiss

13. Protein-A or protein-G beads may produce higher affinity antibody capture depending on the species of primary antibody. Suppliers (e.g., Amersham Biosciences) provide detailed information allowing optimal protein-A/G-bead selection. 14. The number of washes may well influence the basal-to-­ stimulated signal, depending on the Ga-subunit and initial protein concentration the beads may require two to five washes to produce consistent data. Not carrying out sufficient washes of the bead-antibody-Ga complex is a common cause of low signal-to-noise and variability problems with this assay. References 1. Northup, J.K., Smigel, M.D., and Gilman, A.G. (1982) The guanine nucleotide-activating site of the regulatory component of adenylate cyclase. J. Biol. Chem. 257, 11416–11423. 2. Kurose, H., Katada, T., Haga, T., Haga, K., Ichiyama, A., and Ui, M. (1986) Functional interaction of purified muscarinic receptors with purified inhibitory guanine nucleotide regulatory proteins reconstituted in phospholipid vesicles. J. Biol. Chem. 261, 6423–6428. 3. Hilf, G., Gierschik, P., and Jakobs, K.H. (1989) Muscarinic acetylcholine receptorstimulated binding of guanosine 5¢-O-(3thiophosphate) to guanine-nucleotide-binding proteins in cardiac membranes. Eur. J. Biochem. 186, 725–731. 4. Lorenzen, A., Fuss, M., Vogt, H., and Schwabe, U. (1993) Measurement of guanine nucleotide-binding protein activation by A1 adenosine receptor agonists in bovine brain membranes: stimulation of guanosine-5¢-O(3-[35S]thio)triphosphate binding. Mol. Pharmacol. 44, 115–123. 5. Sim, L.J., Selley, D.E., and Childers, S.R. (1995) In vitro autoradiography of receptor-activated G proteins in rat brain by agonist-stimulated guanylyl 5¢-[g-[35S]thio]-triphosphate binding. Proc. Natl. Acad. Sci. USA 92, 7242–7246. 6. Milligan, G. (2003) Principles: Extending the utility of [35S]-GTPgS binding assays. Trends Pharmacol. Sci. 24, 87–90. 7. Friedman, E., Butkerait, P., and Wang, H.Y. (1993) Analysis of receptor-stimulated and basal guanine nucleotide binding to membrane G proteins by sodium dodecyl sulphate-polyacrylamide gel electrophoresis. Anal. Biochem. 241, 171–178. 8. Burford, N.T., Tolbert, L.M., and Sadee, W. (1998) Specific G protein activation and m-opioid receptor internalization caused by morphine,

DAMGO and endomorphin I. Eur. J. Pharmacol. 342, 123–126. 9. Akam, E.C., Challiss, R.A.J., and Nahorski, S.R. (2001) Gq/11 and Gi/o activation in CHO cells expressing human muscarinic acetylcholine receptors: dependence on agonist as well as receptor-subtype. Br. J. Pharmacol. 132, 950–958. 10. Carruthers, A.M., Warner, A.J., Michel, A.D., Feniuk, W., and Humphrey, P.A. (1999) Activation of adenylate cyclase by human recombinant sst5 receptors expressed in CHOK1 cells and involvement of Gsa proteins. Br. J. Pharmacol. 126, 1221–1229. 11. Young, K.W., Bootman, M.D., Channing, D.R., Lipp, P., Maycox, P.R., Meakin, J., Challiss, R.A.J., and Nahorski, S.R. (2000) Lysophosphatidic acid-induced Ca2+ mobilization requires intracellular sphingosine 1-phosphate production: potential involvement of endogenous Edg-4 receptors. J. Biol. Chem. 275, 38532–38539. 12. DeLapp, N.W., McKinzie, J.H., Sawyer, B.D., Vandergriff, A., Falcone, J., McClure, D., and Felder, C.C. (1999) Determination of [35S] guanosine-5¢-O-(3-thio)triphosphate binding mediated by cholinergic muscarinic receptors in membranes from Chinese hamster ovary cells and rat striatum using an anti-G protein scintillation proximity assay. J. Pharmacol. Exp. Ther. 289, 946–955. 13. Salah-Uddin, H., Thomas, D.R., Davies, C.H., Hagan, J.J., Wood, M.D., Watson, J.M., and Challiss, R.A.J. (2008) Pharmacological assessment of M1 muscarinic acetylcholine receptorGq/11 protein coupling in membranes prepared from postmortem human brain tissue. J. Pharmacol. Exp. Ther. 325, 869–874. 14. Burford, N.T., Tobin, A.B., and Nahorski, S.R (1995) Coupling of muscarinic m1, m2 and m3

[35S]GTPgS Binding as an Index of Total G-Protein acetylcholine receptors, expressed in Chinese hamster ovary cells, to pertussis-toxin sensitive/ insensitive guanine nucleotide-binding proteins. Eur. J. Pharmacol. 289, 343–351 15. Berg, K.A., Maayani, S., Goldfarb, J., Scaramellini, C., Leff, P., and Clarke, W.P. (1998) Effector pathway-dependent relative

275

efficacy at serotonin type 2A and 2 C receptors: evidence for agonist-directed trafficking of receptor stimulus. Mol. Pharmacol. 54, 94–104 16. Kenakin, T. (1995) Agonist-receptor efficacy II: agonist trafficking of receptor signals. Trends Pharmacol Sci. 16, 199–205

wwwwwww

Chapter 15 Using Calcium Imaging as a Readout of GPCR Activation Martin D. Bootman and H. Llewelyn Roderick Abstract Monitoring cellular calcium concentration using fluorescent reporters can provide a rapid, proportional assay of G-protein-coupled receptor activation. Recording calcium changes in single cells, or cell populations, is relatively straightforward, but requires careful deliberation regarding the appropriate calcium reporter and experimental approach. Here, we describe strategies to ensure that calcium changes are recorded with good fidelity and minimal invasiveness. We highlight a range of issues that need to be considered within the design of an experiment to measure cellular calcium, and suggest strategies to avoid common pit-falls. Key words: Calcium, Fluorescence, Pluronic, GPCR, Imaging, Wide-field, Confocal, Photobleaching, Contrast, Brightness, Wavelength

1. Introduction Calcium (Ca2+) is one of the most ubiquitous intracellular ­messengers. It controls processes as diverse as fertilisation, gene transcription and muscle contraction (1, 2). From the perspective of studying G-protein-coupled receptors (GPCRs), monitoring changes in cellular Ca2+ concentration can provide a proximal assay for the kinetics and intensity of cellular stimulation. Ca2+ fluxes occur downstream of the activation of many types of GPCR, although they are not the only means by which Ca2+ signals can be evoked (3). Exogenously expressed GPCRs are often capable of engaging endogenous Ca2+ signalling systems, particularly the pathway involving phospholipase C and inositol 1,4,5-trisphosphate-mediated Ca2+ release. In addition to its generic activation, the popularity of Ca2+ for measuring GPCR activation also stems

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_15, © Springer Science+Business Media, LLC 2011

277

278

Bootman and Roderick

from the availability of relatively cheap Ca2+-binding fluorescent dyes, which can be readily loaded into living cells, and monitor cellular Ca2+ concentration in real-time with excellent signal to noise. Furthermore, there is an abundance of turn-key microscope/cuvette-based technologies available for recording fluorescence from tissues, single cells or populations of cells in suspension. Ca2+ measurements are commonly used in high-content screening assays. A multitude of Ca2+ signals, varying in kinetics, amplitude, frequency and spatial dimension, are utilised by different types of cells. This variability of intracellular Ca2+ signals means that there is no single technique that can be used to study Ca2+ in all situations. Instead, there are many types of apparatus, each with advantages and drawbacks that allow researchers to visualise Ca2+ changes in their particular cellular system. This chapter describes simple steps and considerations for using fluorescent indicators to monitor Ca2+ downstream of GPCR activation. The procedures and approaches described below apply to many different types of imaging modalities, such as wide-field, confocal and total internal reflection microscopy (TIRF). However, the final experimental details can vary greatly depending the cell type, imaging system and characteristics of the Ca2+ signals being studied. It is assumed that the reader understands the process of fluorescence, and how excitation light and emitted light are separated by dichroic mirrors and optical filters. The novice reader is directed to excellent online tutorials provided by commercial fluorescence imaging equipment suppliers (e.g. http://www.olympus.co.uk/microscopy/, https://www.micro-shop.zeiss.com/). The methods described in this chapter are applicable to monitoring Ca2+ using proteinaceous reporters such as pericams (4), cameleons (5), modified yellow cameleons (YCs; (6)) and camgaroos (7), although specific protocols may need to be adjusted. The emphasis of this chapter is on measuring cellular Ca2+ signals using synthetic indicators such as those derived by Tsien and colleagues (8–10).

2. Materials 1. Appropriate cell preparation expressing the receptor of interest. 2. Appropriate receptor ligands. 3. Means of measuring fluorescence (e.g. imaging platform). 4. Appropriate extracellular buffer. This solution will vary depending on the cell type, but generally should be the same as that used during experimentation. 5. DMSO/Pluronic F-127. For preparation of DMSO/Pluronic, dissolve 20% (w/v) Pluronic F-127 in an appropriate volume

Using Calcium Imaging as a Readout of GPCR Activation

279

of dry DMSO. Stir this mixture until the Pluronic is dissolved while simultaneously gently warming the solution. The temperature should not exceed 40°C. This should be done in a ventilated area. This mixture is stable at room temperature for ~6  months. Note that this mixture can be obtained commercially. 6. Sulfinpyrazone. Dissolve 40.4  mg sulfinpyrazone in 1  mL DMSO, thereby making a 100  mM solution. This solution may require considerable agitation for complete dissolution of the sulfinpyrazone. Dilute the sulphinpyrazone solution 1:1,000 (v/v) in the buffer used for indicator loading and experiments. This will give final sulfinpyrazone and DMSO concentrations of 100 mM and 0.1% respectively. 7. Stock solution of a Ca2+ indicator ester (see Note 1). Add an appropriate volume of the DMSO/Pluronic mixture to the indicator powder (see Note 2). A useful final concentration of the stock solution is 1 mM. Make sure that the entire indicator solid has been dissolved in the DMSO/Pluronic mixture by agitating the solution through a pipette tip. Vortex or sonicate, if necessary. The indicator solution can be stored at −20°C for a few weeks. We tend to freeze a volume (~10 microlitres) of an indicator stock sufficient for a single experiment to avoid freeze-thawing cycles. New batches should be used if any deterioration in the responses is observed. 8. Working solution of a Ca2+ indicator ester. The DMSO/ Pluronic/indicator stock solution should be diluted in an appropriate extracellular buffer solution just before incubation with cells. The working concentration of Ca2+ indicator is cell type dependent. Some cells take up and hydrolyse the esterified indicator rapidly, whilst others are more difficult to load. Generally, we use a solution with a 1–2  mM concentration of indicator by adding 4  mL of the 1  mM stock to 2  mL of the extracellular solution. Vortexing or sonicating the diluted indicator solution can sometimes help loading. 9. EGTA stock solution (100 mM). EGTA is relatively insoluble in neutral solutions, but can be rapidly dissolved in alkali. An EGTA stock solution can be made by dissolving powder in weak alkali (e.g. 1 M NaOH), but take usual precautions if doing so. 10. Ca2+ ionophore; 1 mM stock solution in DMSO. Commonly used Ca2+ ionophores are ionomycin and A23817. 4-bromoA23184 may be a better choice if the Ca2+ indicator is excited in the UV range, as it is non-fluorescent and may not affect pH. 11. Extracellular calibration solutions. Prepare the extracellular buffer as used in the experiments but do not add Ca2+. To one half of the solution add Ca2+ (e.g. CaCl2), so that the final Ca2+

280

Bootman and Roderick

concentration is 4 mM. This solution will be used to determine maximum fluorescence. To the other half of the solution, add EGTA so that the final concentration is 2 mM. This solution will be used to determine minimum fluorescence. Check, and if necessary, correct the pH of the two solutions. In particular, the addition of EGTA may significantly alter the pH. To a small volume, say 5 mL, of each solution add Ca2+ ionophore to a final concentration of 10 mM.

3. Methods As there are a multitude of available imaging systems with a variety of imaging modalities, it is not possible to discuss the precise technical set-up for a Ca2+ imaging experiment. However, there are some general considerations that apply to all experiments using fluorescent indicators. Most importantly, the imaging system used needs to be appropriately set to allow detection of Ca2+ changes. This requires some insight into the magnitude, kinetics, and localisation of the cellular Ca2+ signal. Thus, whilst it is not possible to write a prescriptive method that will apply to every imaging device, the following protocols outline typical steps involved in setting up for a Ca2+ imaging experiment. 3.1. Loading Fluorescent Indicators into Living Cells

Cells can be loaded with Ca2+ indicators if they are in free suspension, or adhered to glass/plastic/other material. For microscopic imaging experiments, cells should be either grown on glass coverslips or allowed sufficient settling time to allow them to become attached and not wash away with changes of solution. 1. Replace media overlying the cells with 2  mL of the diluted indicator solution (or whatever volume is required to fill the coverslip/cuvette bearing the cells) and incubate in a dark place for typically 30  min or less (see Note 3), generally at room temperature (see Note 4). Constituents of serum can bind to indicators and reduce their free concentration. If the continual presence of serum is not required, cells should be washed free of culture medium prior to loading. A note of caution here is that incubating cells in the absence of serum may activate autophagy. This could compromise cellular responses. 2. At the end of the loading time, remove the indicator solution. Replace with indicator-free extracellular solution and incubate the cells for an additional 20–30 min at room temperature in the dark to allow complete de-esterification of the indicator (see Note 5). 3. If experiments are being performed at 37°C, the temperature of the solution in which the cells are being maintained should be gradually raised.

Using Calcium Imaging as a Readout of GPCR Activation

3.2. Running an Imaging Experiment

281

1. Switch on imaging system with sufficient time for the electronics to synchronise and light sources to warm up and have stable emission. 2. Load cells with appropriate fluorescent Ca2+ indicator as described above. 3. Mount the coverslip bearing adherent cells in an appropriate chamber on the stage of the microscope attached to the imaging system. 4. To empirically determine appropriate settings, start imaging as per an experiment. Running a dummy experiment is a practical way of checking settings for an imaging session. 5. Depending on the system’s controls, adjust the light intensity, brightness and contrast [gain or off-set] so that background pixels have a positive value and that a substantial proportion of the bit depth is available to record a fluorescence increase (see Note 6). On many systems, a “range check” look-up table (LUT) can be selected to check the intensity of background pixels. 6. Select imaging parameters (e.g. image capture frequency, experiment duration) that suit the expected Ca2+ signal (see Note 6). 7. If possible, select regions of interest to monitor during the experiment. It is very useful to have data readout during the progression of an experiment, as it allows the user to spot potential problems such as cell detachment or focal drift. 8. Run experiment. During the experiment, agonist solution will be added to activate the GPCRs. For strongly adherent cells, we routinely change the solution bathing the cells using a 5 mL pipette. The level of fluid is maintained constant by a suction line to prevent the buffer solution spilling over into the imaging system. Using this simple approach, we can typically exchange the extracellular solution within a couple of seconds. Furthermore, the agonist can be washed off, the cells allowed to recover and another agonist solution subsequently applied. For more accurate timing, gravity-driven superfusion apparatus with rapid solenoid switches to change between solutions are available commercially. The same methods can be used for cells that are weakly adherent, but slower superfusion is required. This needs to be determined empirically. Some cell types are activated by mechanical forces during superfusion. This should be checked using addition of extracellular buffer alone.

3.3. Data Analysis

There are multiple parameters that can be measured to characterise cellular responses to stimulation, as depicted in Fig. 1. Some are not steady-state parameters (e.g. rate of rise, latency), whilst others are (e.g. oscillation frequency). Calculating the proportion of cells

282

Bootman and Roderick

Fig. 1. Parameters for measuring GPCR-evoked Ca2+ signalling. The cartoon depicts an idealised oscillatory Ca2+ signal evoked by application of a GPCR agonist, and the parameters that can be used to monitor the response.

showing a detectable Ca2+ signal is useful, as it gives a measure of the threshold and efficacy of a stimulus. In addition, the peak Ca2+ signal amplitude, latency, area integrated Ca2+ signal and frequency of Ca2+ transients can all be used to provide concentration–response relationships for different stimuli. Which of these parameters is most appropriate depends on the cell type and nature of the experiment. For example, some cells types display all-or-none responses where the amplitude of their Ca2+ signals is not altered by changing the stimulus concentration. Whereas, others have smoothly-graded peak Ca2+ signals (11). For the former type of cell, measurement of peak Ca2+ signal or integrated Ca2+ signal obviously has little use, and the proportion of responding cells or latency could be appropriate parameters to describe the action of a particular stimulus. Cellular Ca2+ signals are also often complex, with spatial and temporal characteristics that make it hard to decide where and what to measure. An example of a GPCR-evoked Ca2+ response in depicted in Fig. 2. The traces show the induction of a Ca2+ signal within a single fluo-3-loaded HeLa cell. The Ca2+ signal was sampled by averaging pixels in three regions; the cytosol, a nuclear area and a Ca2+ puff site (for more information on Ca2+ puffs see refs. 12, 13). It is evident that the Ca2+ signals were not the same in all locations. The nucleus appeared to have a higher resting fluorescence and a larger response than the cytosol. However, this does not neccessarily indicate a real difference of the Ca2+ concentration in these regions. Rather, it can be due to the diverse behaviour of

Using Calcium Imaging as a Readout of GPCR Activation

283

Fig. 2. Spatial and temporal complexity of GPCR-evoked Ca2+ signalling. The figure depicts a Ca2+ signal recorded from a fluo3-loaded HeLa cell evoked by application of histamine. The Ca2+ signal was sampled in three cellular regions, as shown on the cell image. The Ca2+ signals observed in those locations are shown by the correspondingly labelled traces. The response was captured using a Noran Oz confocal imaging system running at 15 frames/s, using the 488 nm excitation light from an argon laser, and collecting emitted light >505 nm.

Ca2+ indicators within the cytosol and nucleus; their affinity for Ca2+ and fluorescence properties are different (14). Since the nucleus occupies a large proportion of many cells’ volume, and as fluorescent indicators are much brighter in that organelle, most cellular Ca2+ recordings will actually be dominated by the response of the nucleus (depending on where the fluorescence signal is sampled). This is particularly true for wide-field imaging, where it is generally not possible to distinguish the cytosolic and nuclear boundary. Confocal microscopy, such as that used to record the response depicted in Fig. 2, will provide spatial resolution so that Ca2+ within the cytosolic and nuclear compartments can be distinguished. The Ca2+ puff site depicted in Fig. 2 produced discrete microscopic Ca2+ signals during the on-set of the response. Ca2+ puffs have been faithfully recorded using confocal microscopy (15, 16) and TIRF imaging (17), but are less readily observed using wide-field imaging. Since Ca2+ puffs are the first Ca2+ signals to be triggered in many cells after GPCR activation, they could represent the primary assay for cellular responses and the threshold activity of a particular agonist. The exact method of data analysis depends on the type of data collected (e.g. single- vs. dual-wavelength Ca2+ indicator) and the required output (e.g. image sequence, plot of fluorescence intensity from a particular region of interest or calibrated Ca2+

284

Bootman and Roderick

measure). We will therefore work from the assumption that the experiment generated a series of images collected over a specified time-course. The first step in data analysis is to remove the background from each image (see Note 7). This can be done in multiple ways. Some systems allow a background image to be captured at the start of an experiment, which can then be subtracted from all the subsequent images. Alternatively, the intensity value of a large portion of pixels outside the fluorescent cells can be obtained for each image. This intensity value can be subtracted from all of the pixels in an image sequence. The public domain Java image processing program “ImageJ” is freely available and able to perform such operations on standard image formats (see http://rsb.info. nih.gov/ij/). Once the background is subtracted, the images can be ratioed if a dual-wavelength indicator was used. From that point it is possible to extract the fluorescence change over time in specific regions of interest (usually the cells) within the images, and use this as a read-out of cellular response. 3.4. Data Presentation

Once the background has been subtracted (see Note 7), the data can be manipulated and presented in a number of different ways. If a movie is required, then the images can be simply left as greyscale or pseudo-coloured using an appropriate LUT. Pseudocoloration helps to enhance visualisation of Ca2+ changes. With dual-wavelength Ca2+ indicators, the data can be represented in the ratio format. With fura-2, for example, data is often presented as “340/380 ratio”. In essence, this is short-hand for “the intensity of emitted light recorded with 340 nm excitation divided by the intensity of emitted light recorded with 380 nm excitation”. For single wavelength Ca2+ indicators, where the cellular concentration of the dye impacts on the recorded intensity, it is usual to normalise the image sequence (or numerical sequence of background-corrected intensity values) by ratioing the fluorescence signal during the experiment against that observed in the first (or first few) images before the cells were stimulated. This procedure is often denoted as “F/F0” or “DF/F0” in published work (with F being the fluorescence recorded during the experiment, F0 representing the pre-stimulation fluorescence intensity, and DF being the change in fluorescence signal between successive data points), and provides a simple way of negating the effect of differences in dye loading between cells.

3.5. Data Storage

Imaging experiments have the potential to generate significant volumes of data. With photometry, or the simple extraction of fluorescence intensity values from groups of cells, the resulting ASCII or text files will not be large. If possible, it is best practice to store the original raw image files, as they can be reanalysed if necessary. However, long-term storage of images can be ­problematic. For example, using a rudimentary set-up with an 8-bit camera

Using Calcium Imaging as a Readout of GPCR Activation

285

capturing­images of 512 × 512 pixels every 5 s for 10 min (i.e. 120 images) will generate a file of ~32 Mb. Obviously, this value would swell dramatically if the image format was larger, the bit-depth was extended or if the number of images captured was increased by a higher capture rate or the need for multiple excitation wavelengths. A long series of experiments will therefore rapidly occupy most standard-sized computer hard drives. Historically, our lab used CD-ROM and DVD to store data, as it was the most cost-effective and provided rapid data access and transportability. More recently, we have utilised portable hard drives, which are available in terabyte sizes and can store many more experiments with rapid access. The advent of BluRay writers opens the possibility of storing ~50  Gb on a small disk. It should be noted that data stored on DVDs and hard drives can degrade over time or due to environmental conditions, and they are not appropriate for long-term data archiving. Plausibly greater security can be obtained by using a central file server that is mirrored and regularly backed-up on to high-content tape drives. 3.6. Obtaining Calibration Data

Calibration of fluorescence data (see Note 8) requires knowledge of the Ca2+ indicator’s Kd, and also the maximum and minimum fluorescence signals. The Kd values of Ca2+ indicators can be obtained from the vendors, the published literature or can be determined in situ (18). Empirical determination of the Kd in the same cell type being studied will generally provide the most accurate calibration. In many cases, the Kd values provided by Ca2+ indicator vendors have been determined in simple solutions that do not resemble the cellular milieu. Many factors, such as ionic strength, pH, and temperature can affect the affinity of an indicator for Ca2+ (19). The maximum and minimum fluorescence signals need to be determined within the cells being studied with the same experimental (contrast, brightness etc.) settings. The following is a simplistic method for obtaining maximum and minimum fl ­ uorescence data. Maximum and minimum fluorescence data can be obtained at the end of an experimental run providing there is no evidence of substantial Ca2+ indicator bleaching or extrusion, and the cells are still viable. If there is clear Ca2+ indicator loss or the cells are looking poor, then calibration data can be obtained using a fresh batch of cells loaded with Ca2+ indicator and examined under the same experimental settings. 1. Superfuse the cells on the coverslip with the ionophore/EGTA solution. Depending on the volume of medium contained by the coverslip holder, this may require several flushes with the solution. The ionophore will make the cells’ membranes rapidly permeable to Ca2+. Ca2+ within the cells will be transported out of the cells via the ionophore (and Ca2+ pumps) and be chelated by the EGTA. The low Ca2+ concentration in the EGTA solution will not allow Ca2+ to enter the cells. Imaging

286

Bootman and Roderick

the cells under this condition will reveal a dramatic decrease in Ca2+ concentration that should reach steady-state within ~10–15 min. This is the minimal fluorescence signal, and is generally denoted Fmin or Rmin for a single-wavelength or dualwavelength Ca2+ indicator respectively. 2. Following determination of the minimal fluorescence signal, the cells should be superfused with the ionophore/4  mM Ca2+ solution. Again, this may take several flushes to ensure complete solution exchange. Imaging the cells will reveal a dramatic increase in Ca2+ concentration that should reach steady-state within ~10 min. This is the maximal fluorescence signal, and is generally denoted Fmax or Rmax for a single-wavelength or dual-wavelength Ca2+ indicator respectively. 3. Using the maximal and minimal calibration data (with background subtracted; see Note 7), the fluorescence recorded during experiments can be converted into Ca2+ concentration using the following equations: For a single-wavelength indicator:



æ F - Fmin ö Ca 2 + = K d ç ÷, è Fmax - F ø where, F is the background-corrected fluorescence intensity recorded from cells during the experiment, Fmin is the ­minimal fluorescence (determined in the ionophore/EGTA solution), Fmax is the maximal fluorescence (determined in the ­ionophore/4 mM Ca2+ solution) and Kd is the affinity of the indicator for Ca2+. For a dual-wavelength indicator:



æ R - Rmin ö æ Fmax,λ 2 ö Ca 2 + = K d ç ÷÷ , ÷ çç è Rmax - R ø è Fmin,λ 2 ø where, R is the ratio value of the background-corrected fluorescence recorded from cells during the experiment, Rmin is the minimal fluorescence ratio, Rmax is the maximal fluorescence ratio and Kd is the affinity of the indicator for Ca2+. The (Fmax,l2/Fmin,l2) factor (often denoted as “b”) is the ratio Fmax/Fmin at the wavelength of the calcium free form of the indicator, l2 (e.g. 380 nm in the case of fura-2).

4. Notes 1. A wide variety of Ca2+ indicators are currently available, with excitation and emission spectra ranging from UV to visible colours (see http: // www.invitrogen.com/site/us/en/home/

Using Calcium Imaging as a Readout of GPCR Activation

287

References/Molecular-Probes-The-Handbook.html) http:// www.bioscience.co.uk/products/ion-indicators. In many situations, however, the choice of indicator is limited by experimental or hardware considerations. For example, confocal and TIRF microscopy generally use lasers to provide illumination. Although the light output can be intense, lasers provide only a limited number of wavelengths of excitation light. In contrast, wide-field imaging tends to utilise arc lamps, which emit a broad spectrum, and thus allow a wider choice of indicator. Ca2+-sensitive fluorescent indicators can be broadly divided into single- and dual-wavelength indicators on the basis of their spectral changes in response to Ca2+ elevation. Singlewavelength Ca2+-sensitive indicators (e.g. fluo-4) change their emission intensity upon binding calcium. However, the intensity of emission is also proportional to the indicator concentration within a cell, so careful calibration is required to allow comparison of results between different cells and experiments. Bleaching of the indicator over the period of the experiment will change the effective indicator fluorescence, and can complicate calibration of Ca2+ concentration. With dual-wavelength indicators, problems of uneven distribution and changes in indicator concentration are generally circumvented. When Ca2+ binds to a dual-wavelength indicator molecule, the optimum excitation or emission wavelength changes. For example, with fura-2, the Ca2+-free molecule has a peak excitation at ~380 nm, whilst the Ca2+bound molecule is optimally excited at 340 nm (with emitted light being collected at >450  nm). By ratioing the fluorescence output of fura-2 when it is alternately excited at 340 and 380 nm, a measure of Ca2+ can be made that is independent of the indicator concentration. Dual-wavelength indicators also display an excitation or emission wavelength where the indicator is insensitive to Ca2+. This is known as the “isosbestic point,” and can be used to measure indicator concentration or explore the interaction of indicators with other metal ions, such as Mn2+, which is commonly used as a Ca2+ surrogate to monitor store-operated Ca2+ entry because it quenches fura-2 fluorescence emission upon binding. The two most commonly employed dual-wavelength indicators are fura-2 and indo-1. The former is dual excitation, whereas the latter is dual emission. The need to alternately excite fura-2 with light at ~340 and 380 nm can slow image acquisition if filters are moved in and out of the excitation light path. Indo-1 only requires excitation at ~340 nm, and the emitted light can be separated by a dichroic mirror into the components that decrease (>450  nm) or increase (375–450  nm) as a Ca2+ signal is evoked. For this reason indo-1 is commonly employed when flow cytometry is used

288

Bootman and Roderick

to measure cytosolic Ca2+. Because no moving parts are required to obtain a dual-wavelength recording using ­indo-1, it can allow rapid monitoring of Ca2+. Indeed, we routinely use indo-1 and twin photomultipliers detectors to monitor Ca2+ in ­cardiac myocytes with millisecond resolution (20). However, in our hands indo-1 has a lesser dynamic range (fold change in fluorescence when going from Ca2+-free to Ca2+-bound) than fura-2. Also, modern filter changers or monochromators can switch excitation wavelength within milliseconds, thus making fura-2 a plausible candidate for rapid dual-wavelength measurements. Dual-wavelength indicators are preferable for long-term recordings because the ratio output is less prone to problems associated with indicator distribution, bleaching and leakage. However, when very rapid image acquisition is required, the time for image pairs to be taken may be unacceptable, in which case a single wavelength indicator may be the most suitable. The two most commonly used Ca2+-sensitive dualwavelength indicators, fura-2 and indo-1, are UV-excitable. The use of UV light can introduce problems such as enhanced cellular auto-fluorescence and cell damage, which are both reduced using indicators that are excited by visible light. Although ratiometric Ca2+ recordings using visible wavelength indictors are less common, there are available reporters­. Notably, ratiometric pericam (21) and fura red (22, 23) have been used for dual-wavelength Ca2+ measurement. In addition, co-introduction of fluo-3 and fura red offers the possibility of dual emission ratiometric confocal measurements (18, 24). In addition to avoiding issues of autofluorescence and cellular damage, the currently-available visible wavelength indicators, such as fluo-4, have a superior dynamic range than their UV-excited counterparts (18). In practical terms, a fluorescent indicator can accurately report changes of Ca2+ concentration approximately one order of magnitude above or below its Kd. Concentrations of Ca2+ in the cell can range from tens of nanomolar to several micromolar, this means that the indicator must be matched to the magnitude of Ca2+ changes of interest. Some insight into the expected Ca2+ response is therefore necessary. Fluorescent Ca2+ dyes have carboxylic acid groups for ­co-ordination of Ca2+. This makes them hydrophilic, and cell impermeant. The free acid forms of the indicators are hydrophilic, and can be micro-injected into cells or introduced via patch pipette. However, the most convenient method for loading the dyes is as hydrophobic compounds in which the carboxylic groups are esterified (commonly acetoxymethyl or acetate esters). The ester versions of the indicators can pass across plasma membranes, and the free acid Ca2+-sensitive dye is liberated following hydrolysis of

Using Calcium Imaging as a Readout of GPCR Activation

289

the ester groups by ­intracellular esterases (25). Alternative methods for Ca2+ ­indicator loading including transient cell permeabilisation, or pinocytic uptake of extracellular dye. These methods allow the introduction of the free acid (Ca2+sensitive) forms of indicators. 2. Ca2+ indicators are light sensitive and labile. They are often purchased as solid powders, and need to be dissolved and stored in the dark at −20°C. The free acid forms of Ca2+ indicators­can be dissolved in water or buffer solution. Typically, Ca2+ indicator ester powders are dissolved in a DMSO/Pluronic acid solution for cell loading. DMSO is used as a general ­solvent and Pluronic acid acts as a detergent to prevent amphipathic Ca2+ indicator molecules forming micelles. Pluronic acid has been demonstrated to enhance Ca2+ indicator loading. DMSO/ Pluronic F-127 mix is available commercially, and we recommend buying this reagent. If needed, the recipe for making DMSO/Pluronic is listed in Subheading 2. However, it should be noted that dissolving Pluronic may require warming the DMSO, which can produce noxious fumes. 3. The duration of incubation for Ca2+ indicator loading is cell dependent, and needs to be worked out empirically. We typically use 30 min or less. Some cell types require longer incubation times. 4. Generally, Ca2+ indicators are trapped more faithfully in the cytosol if the loading is performed at room temperature (i.e. 18–22°C). Higher temperatures often promote the sequestration of indicators in intracellular organelles. 5. Once cells are loaded with a Ca2+ indicator, they should be used immediately. Some cells types have the capacity to ­rapidly extrude indicators, or to compartmentalise them within ­internal organelles. This is particularly evident with experiments performed at 37°C. Such changes in the distribution of indicators make it very difficult to calibrate Ca2+ concentration. If indicator leakage or sequestration is encountered, plausible solutions are to conduct experiments at room temperature (or some empirically determined temperature between 20 and 37°C where the indicator is preserved). Although Ca2+ signalling systems are in part enzymatic and therefore affected by temperature, Ca2+ signals can be reliably evoked at room temperature in many cases. The major alternatives to changing temperature are to use a leakage-resistant Ca2+ indicator, or to pharmacologically block the multi-­ specific organic anion transport that underlies multi-drug resistance (MDR) in cells. The most leakage resistant forms of Ca2+ indicator are those in which the dye is coupled to dextran; a biologically inert carrier molecule. Dextran-conjugated Ca2+ indicators do not leak, or become sequestered, under conditions in which

290

Bootman and Roderick

other dyes are adversely transported. Since they are designed not to be membrane permeant, dextran-conjugated Ca2+ ­indicators need to be microinjected into cells or introduced via patch pipette. However, if ester loading is required, there are a few Ca2+ indicators that have been specifically designed to be leakage resistant. The most popular of these is Fura-PE3, which has similar spectral properties to fura-2, but is much less prone to being exported from cells or compartmentalised within organelles due to its zwitterionic nature (26). If the manoeuvres described above are not appropriate for preventing indicator transport, a final strategy is to block MDR-mediated anion transport. The compound used most often for this purpose is sulfinpyrazone, an uriscosuric medication commonly used in the treatment of gout. We have used sulfinpyrazone in situations where loading of fluorescent Ca2+ indicators or subsequent experiments need to be performed above room temperature (27, 28). To date, we are unaware of any effects of sulfinpyrazone (or the alternative compound; probenecid) on GPCRs or Ca2+ signal transduction. However, as with all drugs, it should be used with caution. Sulfinpyrazone is highly lipophilic and will therefore need to be dissolved in a non-polar solvent such as DMSO or acetone. Since the AM forms of indicator are also reconstituted in DMSO (see above), care must be taken not to exceed deleterious levels of non-polar solvent. 6. With respect to the magnitude of a Ca2+ signal, an imaging system should ideally be set to record the fluorescence change using a wide portion of its “bit depth”. In essence, the bit depth is the number of arbitrary levels that a system can use to monitor changes in the intensity of an image. For example, in a 1-bit recording, the entire grey spectrum would be represented as 0 or 1, and all the pixels that form an image would be black or white. Whereas, an 8-bit recording will have 256 grey levels (28), and an image would therefore have greater intensity resolution. Since the human eye can discriminate less than 50 shades of grey, an 8-bit image gives the appearance of a continuous intensity profile. The camera or photomultiplier used to record the fluorescence emanating from a cell will determine the bit-depth of the resulting image. Most modern digital cameras/photomultipliers are capable of 8-, 12- (4,096 grey levels) or 16-bit (65,536 grey levels) resolution. It is obvious that 12- and 16-bit images allow greater resolution in the detection of small light intensity changes. However, this extra resolution dramatically increases the amount of memory needed to store each image, and is not necessary for many applications. Consideration of bit depth is important when considering how large a Ca2+ signal may be, and how much discrimination is required in the detection of its rising and recovery phases.

Using Calcium Imaging as a Readout of GPCR Activation

A key method for controlling the usage of bit depth is adjusting the “contrast” control on the camera/photomultiplier (sometimes denoted as “gain”), which essentially regulates the amplification of the emitted fluorescent light. This is a particularly important control in Ca2+ imaging experiments, because it needs to be correctly adjusted before a Ca2+ signal is evoked. If the gain setting is not appropriate the detection of the Ca2+ signal will be compromised (Fig.  3). In this context, Ca2+ imaging requires more insight into the nature of the evoked fluorescence signal than in the situation with a fixed, unchanging fluorescent specimen that can be imaged repeatedly with different settings each time. Ideally, the contrast should be moderated so that Ca2+ signals utilise a broad portion of the bit depth of the system. This is critical in experiments where different amplitude Ca2+ signals may occur in the same recording, as shown in Fig. 3. In addition to the genuine fluorescence signal, cameras and photomultipliers will introduce random noise due to electronic fluctuations within their detectors. In many practical situations, this noise normally contributes a small percentage to the signal output from a camera or photomultiplier. However, noise can become an increasing proportion of the signal as the incident fluorescent light from the sample decreases. Increasing the contrast to amplify a weak signal will also enhance the noise making them both more evident. The noise doesn’t actually increase – it just becomes a more significant part of the overall signal. Most imaging systems will have a control for “brightness” (also called “black level” or “off-set”), which also needs to be empirically adjusted. This control essentially sets the intensity at which the camera or photomultiplier starts to detect a signal. As for contrast, this setting needs careful consideration. When performing any kind of fluorescence imaging, it is tempting to adjust the brightness so that the background outside the sample of interest is completely black because it produces images with apparently excellent contrast and no extraneous signal. However, it is incorrect to adjust the brightness to this level, as some of the real fluorescence emission could be negated. This can lead to a miscalculation of Ca2+ changes. In practice, images should be obtained with the brightness setting adjusted so that the areas devoid of cells, or other fluorescent objects, have a positive signal. Although this set-up uses up some of the available bit depth, it is a necessary trade-off to ensure that the entire fluorescence signal is faithfully recorded. The background can be subtracted during data analysis (see below).

291

292

Bootman and Roderick

Fig. 3.  Effects of incorrect adjustment of contrast and brightness settings during Ca2+ imaging. (a) Depicts an idealised cellular Ca2+ response. There are two large Ca2+ transients, and two smaller Ca2+ events. Both types of event are clearly visible as the full bit-depth of the imaging system was used for the recording. In (b), the same idealised Ca2+ signal is presented, but now showing the effect of too high contrast (bi) or too low contrast (bii). In the former situation, the small Ca2+ signals have even better detection, but the large amplitude responses saturate the detector (the grey portions of the trace show parts of the signal that are not recorded). With too low contrast, the large Ca2+ signals are still visible, but the small Ca2+ signals are barely perceptible. (c) illustrates the effects of incorrect brightness setting on the idealised Ca2+ signal. Essentially, the top (ci; brightness too high) or bottom (cii; brightness too low) parts of the recording are missed (the grey portions of the trace show parts of the signal that are not recorded).

Using Calcium Imaging as a Readout of GPCR Activation

293

It is important to differentiate between the effect of adjusting camera/photomultiplier brightness and contrast, and altering the display settings for visualisation of images. With many imaging systems, it is possible to select various look-up tables, adjust the “grey levels to view” or alter the display contrast and brightness settings. These controls can dramatically alter the visualisation of images, but they only affect the way that data is displayed and do not change the recorded information unless a transformation option is available. Final considerations are the intensity of excitation light, image/data acquisition frequency and length of excitation light exposure. These three factors are generally dependent on each other, and there generally has to be a trade-off between excitation intensity, image frequency and excitation duration. For example, acquiring rapid images usually requires the application of brief bursts of intense light in order to excite sufficient fluorescent molecules to generate a sufficient signal to noise. However, in general such experiments will be short-lived because intense illumination of a fluorescent Ca2+ indicator causes it to become photobleached. Such bleaching occurs during excitation of a fluorescent molecule when it reacts with oxygen to form a non-fluorescent molecule, instead of emitting energy as light and returning to a nonexcited state. Photobleaching will compromise the detection and calibration of data obtained during Ca2+ imaging experiments (29). Furthermore, intense illumination of Ca2+ indicators can give rise to the production of reactive oxygen species, which can actually trigger the generation of Ca2+ signals. Indeed, we have found that Ca2+ signals can be easily triggered during confocal imaging of cells loaded with fluo-3 or rhod-2 using moderate laser illumination. This essentially recapitulates events involved in photodynamic therapy, where intense light is delivered to tumour cells containing a photosensitive compound (30). Cell-permeant chemical scavengers of reactive oxygen species (e.g. ascorbate, Trolox) have been successfully used to reduce bleaching and production of reactive oxygen species (29). However, it is important to note that Ca2+ signals and reactive oxygen species are functionally linked, so the introduction of scavengers may alter the biological process being studied. Alternatively, photobleaching of Ca2+ indicators may be mathematically corrected for since it follows a predictable exponential decay (18). In experiments in which Ca2+ signals develop slowly, for example over tens of seconds or minutes, the image capture rate can be slowed to

294

Bootman and Roderick

frequencies ~0.5 Hz or less whilst still allowing good detection of Ca2+ changes. Importantly, with such slow imaging speeds, the illumination generally does not have to be so intense, thereby reducing the occurrence of photobleaching. Advances in illumination and detector technology have significantly helped reduce effects of photoexcitation or photobleaching. However, such advanced technologies may not be available on older rudimentary imaging devices. A further consideration is objective magnification. It is tempting to use higher power objectives to capture cellular images. However, this is generally not necessary. Low power (e.g. 20x) will typically increase the number of cells that can be monitored in one field of view, and will cut the illumination required. A higher power objective (e.g. 40x or 60x) may be required for imaging small cells. 7. Subtraction of background should not be overlooked. It has a significant effect on the apparent change in fluorescence. For example, in the situation where the background pixel intensity was 100 (arbitrary units), the resting pixel intensity within a cell was 200, and after cell stimulation the recorded fluorescence intensity reached 500. Without background subtraction, the response would appear to be a 2.5-fold change in fluorescence (i.e. 500/200 = 2.5). However, with background subtraction, the change is fourfold (i.e. (500 – 100)/(200 – 100) = 4). This may not be critical if the point of an experiment was simply to identify whether a response had occurred. However, without background subtraction it is not possible to accurately convert fluorescence into Ca2+ concentration or compare results between experiments using different imaging system settings. An alternative to subtracting background from each image is to simply measure fluorescence intensity from the required regions of interest, and also from a background region over the time course of the experiment. The background intensity value can then be numerically subtracted from the intensity values obtained from the cells at the corresponding times. This method avoids changes to the images, and is more convenient if a graphical representation of the Ca2+ response in each cell is required. 8. Calibration of fluorescence data is useful because it allows a more realistic comparison between experiments. The intensity of emission from a Ca2+ indicator does not necssarily increase linearly with Ca2+ concentration, but rather increases logarithmically with Ca2+ concentration over certain portions of the calibration curve. This means that a small change in fluorescence intensity could actually report a substantial

Using Calcium Imaging as a Readout of GPCR Activation

295

change in Ca2+ concentration. Many published papers have erroneously presented comparisons between experiments, and corresponding statistics, based on uncalibrated data, where the authors have assumed that the fluorescence is linearly related to Ca2+ concentration. Calibrating data using the methods and equations described provides a transformation of the fluorescence output into a linear Ca2+ concentration. The data can then be averaged and statistically compared. References 1. Berridge, M.J., Lipp, P., and Bootman, M.D. (2000) The versatility and universality of calcium signalling. Nat. Rev. Mol. Cell Biol. 1, 11–21. 2. Berridge, M.J., Bootman, M.D., and Roderick, H.L. (2003) Calcium signalling: dynamics, homeostasis and remodelling. Nat. Rev. Mol. Cell Biol. 4, 517–529. 3. Bootman, M.D., Berridge, M.J., and Roderick, H.L. (2002) Calcium signalling: more messengers, more channels, more complexity. Curr. Biol. 12, R563–565. 4. Nagai, T., Sawano, A., Park, E.S., and Miyawaki, A. (2001) Circularly permuted green fluorescent proteins engineered to sense Ca2+. Proc. Natl. Acad. Sci. U.S.A. 98, 3197–3202. 5. Truong, K., Sawano, A., Miyawaki, A., and Ikura, M. (2007) Calcium indicators based on calmodulin-fluorescent protein fusions. Methods Mol. Biol. 352, 71–82. 6. Nagai, T., Yamada, S., Tominaga, T., Ichikawa, M., and Miyawaki, A. (2004) Expanded dynamic range of fluorescent indicators for Ca2+ by circularly permuted yellow fluorescent proteins. Proc. Natl. Acad. Sci. U.S.A. 101, 10554–10559. 7. Griesbeck, O., Baird, G.S., Campbell, R.E., Zacharias, D.A., and Tsien, R.Y. (2001) Reducing the environmental sensitivity of yellow fluorescent protein. Mechanism and applications. J. Biol. Chem. 276, 29188–29194. 8. Grynkiewicz, G., Poenie, M., and Tsien, R.Y. (1985) A new generation of Ca2+ indicators with greatly improved fluorescence properties. J. Biol. Chem. 260, 3440–3450. 9. Minta, A., Kao, J.P., and Tsien, R.Y. (1989) Fluorescent indicators for cytosolic calcium based on rhodamine and fluorescein chromophores. J. Biol. Chem. 264, 8171–8178. 10. Tsien, R.Y. (1980) New calcium indicators and buffers with high selectivity against magnesium and protons: design, synthesis, and properties of prototype structures. Biochemistry 19, 2396–2404.

11. Bootman, M.D., Cheek, T.R., Moreton, R.B., Bennett, D.L., and Berridge, M.J. (1994) Smoothly graded Ca2+ release from inositol 1,4,5-trisphosphate-sensitive Ca2+ stores. J. Biol. Chem. 269, 24783–24791. 12. Bootman, M.D., Lipp, P., and Berridge, M.J. (2001) The organisation and functions of local Ca2+ signals. J. Cell Sci. 114, 2213–2222. 13. Zeller, S., Rudiger, S., Engel, H., Sneyd, J., Warnecke, G., Parker, I., and Falcke, M. (2009) Modeling of the modulation by buffers of Ca2+ release through clusters of IP3 receptors. Biophys. J. 97, 992–1002. 14. Bootman, M.D., Fearnley, C., Smyrnias, I., MacDonald, F., and Roderick, H.L. (2009) An update on nuclear calcium signalling. J. Cell Sci. 122, 2337–2350. 15. Sun, X.P., Callamaras, N., Marchant, J.S., and Parker, I. (1998) A continuum of InsP3mediated elementary Ca2+ signalling events in Xenopus oocytes. J. Physiol. 509, 67–80. 16. Thomas, D., Lipp, P., Tovey, S.C., Berridge, M.J., Li, W., Tsien, R.Y., and Bootman, M.D. (2000) Microscopic properties of elementary Ca2+ release sites in non-excitable cells. Curr. Biol. 10, 8–15. 17. Smith, I.F., Wiltgen, S.M., and Parker, I. (2009) Localization of puff sites adjacent to the plasma membrane: functional and spatial characterization of Ca2+ signaling in SH-SY5Y cells utilizing mem­brane-permeant caged IP3. Cell Calcium 45, 65–76. 18. Thomas, D., Tovey, S.C., Collins, T.J., Bootman, M.D., Berridge, M.J., and Lipp, P. (2000) A comparison of fluorescent Ca2+ indicator properties and their use in measuring elementary and global Ca2+ signals. Cell Calcium 28, 213–223. 19. Uto, A., Arai, H., and Ogawa, Y. (1991) Reassessment of Fura-2 and the ratio method for determination of intracellular Ca2+ concentrations. Cell Calcium 12, 29–37. 20. Proven, A., Roderick, H.L., Conway, S.J., Berridge, M.J., Horton, J.K., Capper, S.J., and

296

Bootman and Roderick

Bootman, M.D. (2006) Inositol 1,4,5-trisphosphate supports the arrhythmogenic action of endothelin-1 on ventricular cardiac myocytes. J. Cell Sci. 119, 3363–3375. 21. Shimozono, S., Fukano, T., Nagai, T., Kirino, Y., Mizuno, H., and Miyawaki, A. (2002) Confocal imaging of subcellular Ca2+ concentrations using a dual-excitation ratiometric indicator based on green fluorescent protein. Sci. STKE 2002, pl4. 22. Kurebayashi, N., Harkins, A.B., and Baylor, S.M. (1993) Use of fura red as an intracellular calcium indicator in frog skeletal muscle fibers. Biophys. J. 64, 1934–1960. 23. Lohr, C. (2003) Monitoring neuronal calcium signalling using a new method for ratiometric confocal calcium imaging. Cell Calcium 34, 295–303. 24. Lipp, P. and Niggli, E. (1993) Ratiometric confocal Ca2+-measurements with visible ­wavelength indicators in isolated cardiac ­myocytes. Cell Calcium 14, 359–372. 25. Tsien, R.Y. (1981) A non-disruptive technique for loading calcium buffers and indicators into cells. Nature 290, 527–528. 26. Vorndran, C., Minta, A., and Poenie, M. (1995) New fluorescent calcium indicators

designed for cytosolic retention or measuring calcium near membranes. Biophys. J. 69, 2112–2124. 27. Bootman, M.D., Taylor, C.W., and Berridge, M.J. (1992) The thiol reagent, thimerosal, evokes Ca2+ spikes in HeLa cells by sensitizing the inositol 1,4,5-trisphosphate receptor. J. Biol. Chem. 267, 25113–25119. 28. Szado, T., Vanderheyden, V., Parys, J.B., De Smedt, H., Rietdorf, K., Kotelevets, L., Chastre, E., Khan, F., Landegren, U., Soderberg, O., Bootman, M.D., and Roderick, H.L. (2008) Phosphorylation of inositol 1,4,5-trisphosphate receptors by protein kinase B/Akt inhibits Ca2+ release and apoptosis. Proceedings of the National Academy of Sciences of the United States of America 105, 2427–2432. 29. Scheenen, W.J., Makings, L.R., Gross, L.R., Pozzan, T., and Tsien, R.Y. (1996) Photo­ degradation of indo-1 and its effect on apparent Ca2+ concentrations. Chem. Biol. 3, 765–774. 30. Hong, X., Jiang, F., Kalkanis, S.N., Zhang, Z.G., Zhang, X., Zheng, X., Jiang, H., and Chopp, M. (2009) Intracellular free calcium mediates glioma cell detachment and cytotoxicity after photodynamic therapy. Lasers Med. Sci. 24, 777–786.

Chapter 16 Measuring Spatiotemporal Dynamics of Cyclic AMP Signaling in Real-Time Using FRET-Based Biosensors Frank Gesellchen, Alessandra Stangherlin, Nicoletta Surdo, Anna Terrin, Anna Zoccarato, and Manuela Zaccolo Abstract Cyclic AMP governs many fundamental signaling events in eukaryotic cells. Although cAMP signaling has been a major research focus for a long time, recent technological developments are revealing novel aspects of this paradigmatic pathway. In this chapter, we give an overview over current fluorescence resonance energy transfer (FRET)-based sensors for detection of cAMP dynamics, and their application in monitoring local, compartmentalized cAMP signals within living cells. A basic step-by-step protocol is given for conducting a FRET experiment in primary cells with a unimolecular cAMP sensor, which can easily be adapted to a user’s specific requirements. Key words: Imaging, Fluorescence resonance energy transfer, Biosensors, Cyclic AMP, Compartmentalization, Cyclic AMP-dependent protein kinase, Exchange protein directly activated by cAMP

1. Introduction The second messenger molecule, cyclic 3¢,5¢-adenosine monophosphate (cAMP) is generated at the plasma membrane in response to G protein-coupled receptor (GPCR) activation and can elicit a multitude of intracellular responses. The primary target of cAMP is the cyclic AMP-dependent protein kinase (protein kinase A, PKA). In its inactive state, PKA forms a heterotetramer composed of two dimerized regulatory (R-) subunits, each binding one catalytic (C-) subunit. Two major subtypes of regulatory subunits have been described, type I and type II R-subunits (1, 2). Accordingly, PKA holoenzymes are classified as either type I or type II, depending on the subtype of regulatory subunit present. Each R-subunit has two cAMP binding sites, which, when Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_16, © Springer Science+Business Media, LLC 2011

297

298

Gesellchen et al.

occupied, lead to allosteric activation of the enzyme, releasing the active C-subunits which are then free to phosphorylate their substrate proteins (3). The classical textbook representation of this signaling pathway links GPCRs via heterotrimeric G protein activation to stimulation of adenylyl cyclase activity, which in turn raises cAMP levels uniformly inside the cell, leading to PKA activation. This is now generally recognized to be an over-simplification of the pathway’s organization. Indeed, the fact that cAMP is a small, freely diffusible molecule together with the very high number of PKA substrates in the cell raises the question of how the signal can be transmitted with fidelity. In recent years the view has emerged that specificity of response is achieved through mechanisms that confine cAMP signaling spatially as well as temporally. One level of spatial control is implemented by scaffolding proteins that anchor PKA to distinct subcellular compartments. The holoenzyme composition has a major influence on localization, which is mediated by the N-terminal dimerization/docking (D/D) domain of the R-subunits. The property of PKA type II to bind to A-kinase anchoring proteins (AKAPs) has been welldocumented (4), and was fundamental in establishing the concept of compartmentalized cAMP signaling. PKA type I, on the other hand, is generally considered to be localized in the cytoplasm, although a few AKAPs have been reported to be capable of interacting specifically with PKA type I (5, 6) or with both PKA type I and II (7, 8). Indeed, recent evidence suggests that type I holoenzymes can also define distinct cAMP signaling complexes, at least in certain cell-types (9). AKAPs not only serve as scaffolds for PKA, but also for other components of the cAMP signaling pathway; for instance phosphodiesterases (PDEs), protein phosphatases and kinases, or substrates of PKA (4). By hydrolyzing cAMP, PDEs can act as a “sink” for the otherwise freely diffusible cAMP molecule, thus shaping distinct pools of cAMP in the cytoplasm. Together with protein phosphatases, which counteract PKA-mediated phosphorylation, PDEs serve efficiently to terminate cAMP signals, while on the other hand, substrates anchored in the vicinity of PKA allow an efficient relaying of the signal. AKAPs therefore organize highly integrated signaling complexes, which confer stringent spatiotemporal regulation and specificity to the cAMP signals. Important contributions to understanding the spatial and temporal control of this pathway have come from the development of techniques that allow detection of cAMP in real-time in intact cells (10). 1.1. Classical Methods for Measuring cAMP Signaling

The classical approach to measure cAMP still in use today is a radioimmunoassay (RIA) employing anti-cAMP antibodies and radiolabeled cAMP as a tracer, which is displaced by the endogenous cAMP present in the sample (11). Due to its robustness and high sensitivity this method still retains its usefulness today and

Measuring Spatiotemporal Dynamics of Cyclic AMP Signaling in Real-Time

299

has also been adapted for high-throughput screening (12). It is, however, increasingly replaced by non-radioactive methods such as fluorescence polarization (FP) (13), homogeneous timeresolved fluorescence (HTRF), or AlphaScreen (amplified luminescence proximity homogeneous assay) (14). Like the original RIA these are antibody-based detection methods, in which a cAMP tracer (fluorescein-labeled in the case of FP, biotinylated in AlphaScreen) is displaced by cAMP in the sample, resulting in a concentration-dependent signal decrease. A novel high-throughput assay makes use of the PKA holoenzyme itself to assess cAMP levels by exogenously adding PKA to a cell lysate and monitoring ATP consumption of the cAMPactivated enzyme by means of a luciferase/luciferin luminescent reaction (15). Since sample lysis is an inherent step of the described methods, these assays provide no information on the spatial control of cAMP signaling. In addition, they show very limited temporal resolution. Furthermore, antibody-based methods can only detect cAMP accumulation, but not the dynamics inherent to the cAMP signaling events within a cell. While these methods are well-suited to determine total (steady-state) cAMP levels in a cell population, for many applications it is desirable to measure the cAMP response on the single cell level and in real-time. 1.2. Fluorescence Resonance Energy Transfer



Fluorescence resonance energy transfer (FRET) has been established as the method of choice for visualizing cAMP levels in intact living cells and overcoming the limitations of the conventional approaches presented above. FRET takes advantage of a fluorescent molecule’s capability of transferring part of its excited state energy via a nonradiative mechanism (dipole–dipole interaction) to a suitable second nearby fluorophore, provided there is a sufficient overlap of the emission and excitation spectra of donor and acceptor, respectively. This process results in a reduction in the emission of the FRET donor with a concomitant increase in the emission of the FRET acceptor (sensitized acceptor emission). The ratio of acceptor/donor emissions can therefore be used as a measure of resonance energy transfer. The dipole–dipole interactions underlying the FRET phenomenon are highly sensitive to distance and orientation of the fluorophores. Generally, fluorophores must be within 1–10 nm of each other for FRET to occur, while at the same time maintaining an appropriate spatial orientation. According to the formula: -1

é æ r ö6 ù E = ê1 + ç ÷ ú , êë è R0 ø úû the transfer efficiency (E) is inversely proportional to the 6th power of the distance (r) between fluorophores, with E being half-maximal at the so-called Förster radius (R0) characteristic for

300

Gesellchen et al.

a given pair of fluorophores (16, 17) (for calculations of the Förster radii of different green fluorescent protein pairs, see ref. 18). This makes FRET a highly sensitive approach to measure conformational changes and protein–protein interactions. 1.3. FRET-Based Sensors for Real-Time Imaging of cAMP in Living Cells

Generally, a FRET-based cAMP indicator comprises two essential components: (1) a cAMP sensor, which may consist of either two separate interacting protein domains or a single protein domain undergoing a conformational change on cAMP binding, and (2) a donor and an acceptor chromophore fused to the cAMP sensor. Changes in conformation, or in the distance between interacting protein partners, which occurs on cAMP binding, affect the distance or the orientation of the two fluorophores, and thus the efficiency of energy transfer. Variations in energy transfer efficiency, accordingly, correlate with changes in cAMP concentration (see Fig. 1).

1.3.1. Bimolecular (PKA-Based) FRET Sensors

The first FRET-based sensor that allowed real-time imaging of cAMP in living cells was based on fluorescein-labeled PKA-C and rhodamine-labeled PKA-R subunits (FlCRhR, see Table 1) (19) which were microinjected into cells. In a resting cell, under low cAMP conditions, FRET occurs between the two fluorophores,

Fig. 1. Schematic representation of the bimolecular tetrameric R-CFP/C-YFP FRET sensor (a) and of the unimolecular Epac1-camps sensor (b). R, PKA regulatory subunit; C, PKA catalytic subunit; #, cAMP binding domain of Epac1; CFP, cyan fluorescent protein; YFP, yellow fluorescent protein. Excitation of CFP is indicated by a flash, emission of CFP and YFP by short arrows. The respective wavelengths are also indicated. In the cAMP-free state, both sensors exhibit FRET between CFP and YFP on excitation of CFP at 430 nm (left). Binding of cAMP induces the dissociation of PKA C and R (upper right) or a conformational change in the Epac1-camps sensor (lower right) resulting in a decrease in FRET.

Measuring Spatiotemporal Dynamics of Cyclic AMP Signaling in Real-Time

301

while a rise in cAMP concentration correlates with a decrease in FRET due to the dissociation of PKA subunits. While certainly pioneering the field, this sensor had some serious technical limitations, such as the need for microinjection of a relatively large concentration (mM) of a protein complex, the propensity to aggregate and precipitate of the labeled subunits and the non-specific interaction of the labeled subunits with cellular structures (20). The first genetically-encoded PKA-based sensor overcame these limitations (10, 21). In this sensor the regulatory and the

Table 1 Overview of cAMP FRET sensors Sensor

Design

EC50

FlCRhR (19)

Tetrameric PKA

90 nM

R-CFP/C-YFP (21) Tetrameric PKA

0.3 mM

R R230K-CFP/ C-YFP (24)

Mutant tetrameric 31.3 mM PKA

PKA-camps (31)

cAMP-binding domain from PKA

1.9 mM

CFP-Epac-YFP (27), ICUE (28)

Full length Epac1

ND

CFP-Epac (dDEP-CD)YFP (H30) (27)

Mutant Epac1

12.5 mM

mpH30 (30)

Mutant Epac1

20 mM

nlsH30 (30)

Mutant Epac1

17.5 mM

Epac1-camps (31)

cAMP-binding domain from Epac1

2.4 mM

RI_epac (9)

cAMP-binding domain from Epac1

ND

RII_epac (9)

cAMP-binding domain from Epac1

ND

Structure

Sensor: name of cAMP FRET sensor (references in brackets), Design: protein / domain the sensor is based on, EC50: cAMP concentration required to achieve half-maximal FRET response (according to references), Structure: graphical representation of sensor’s structure #: cyclic nucleotide binding domain; RI D/D: dimerization/docking domain of PKA regulatory subunit type I; RII D/D: dimerization/docking domain of PKA regulatory subunit type II; C: PKA catalytic subunit; Rh: rhodamine; Fl: fluorescein; CFP: cyan fluorescent protein, YFP: yellow fluorescent protein; DEP: Dishevelled, Egl-10, Pleckstrin homology; GEF: guanine nucleotide exchange factor; mp: myristoylation/palmitoylation signal; nls: nuclear localization sequence. Mutations in the amino acid sequence are indicated. 2x indicates the presence of two R and C subunits each in the PKA holoenzyme.

302

Gesellchen et al.

catalytic subunits of PKA are fused to cyan and yellow forms of GFP, respectively (see Fig.  1a and Table.  1). Being geneticallyencoded, such sensors can be introduced into cells by transfection of the cDNAs coding for the two chimeric subunits, thereby extending the application to most cell-types. Moreover, as the regulatory subunit of the probe is the type II isoform, the overexpressed R-CFP subunit binds, via its D/D domain, to endogenous AKAPs present in the cell, thereby allowing monitoring of cAMP fluctuations in specific compartments within the cell. Different variations of this PKA-based sensor have been developed. In order to improve the dynamic range of the PKAbased probe a polypeptide linker of 20 amino acids was inserted between the R subunit and the fluorophore (RII-L20-CFP), which doubled FRET efficiency compared to the original sensor (22). Under certain experimental conditions, however, the affinity of the PKA-based sensor for cAMP may be too high, and changes in cAMP concentration may go undetected due to probe saturation. By introducing an R230K mutation in cAMP binding site A of the RII subunit (see Table 1), the affinity for cAMP was lowered by one to two orders of magnitude (23), resulting in a less sensitive FRET cAMP sensor, which is consequently less prone to probe saturation (24). However, PKA-based sensors also have some limitations. First of all, the co-transfection of two separate subunits gives no control over the stoichiometry of the R and C subunits inside the cell. This can lead to artifacts due to a predominance of one fluorophore over the other. Secondly, the cAMP-induced dissociation of R and C subunits occurs through a complex cooperative mechanism (3), and therefore the kinetics of FRET change may be slower than the actual kinetics of cAMP changes. In addition, the PKA-based sensor is catalytically active; this in itself can cause an alteration of cAMP dynamics as a consequence of over-activation of the cAMP/PKA pathway (e.g., via PKA-dependent activation of PDEs). These reasons contributed to the development of unimolecular FRET-based sensors for cAMP. 1.3.2. Unimolecular (Epac-Based) FRET Sensors

Epac is a guanine nucleotide exchange factor (GEF) for the small GTPase proteins Rap1 and Rap2. Epac contains a DEP (Dishevelled, Egl-10, Pleckstrin) domain responsible for membrane anchoring, a regulatory domain (Ras exchange motif, REM), a cyclic nucleotide-binding domain (CNB, homologous to those of PKA) and a C-terminal GEF domain (25). The regulatory domain is auto-inhibitory; when intracellular cAMP levels are low, Epac is folded in an inactive conformation preventing binding of Rap to the GEF domain. When cAMP levels rise, cAMP binding to the Epac nucleotide-binding domain induces unfolding of the protein, thereby releasing inhibition of the GEF domain by the REM domain. This conformational change can be detected by FRET (26, 27).

Measuring Spatiotemporal Dynamics of Cyclic AMP Signaling in Real-Time

303

A number of different unimolecular FRET sensors for cAMP were developed simultaneously in different labs (see Table. 1). A sensor exploiting the cAMP binding properties of Epac is based on sandwiching the full length Epac1 protein between CFP and YFP, as exemplified by the CFP-Epac-YFP sensor (27) and the ICUE (Indicator of cAMP using Epac) sensor (28). The CFPEpac-YFP sensor was further modified by deleting the Epac-DEP domain (29) and by rendering it catalytically-inactive (by introducing a T781A/F782A double mutation in the GEF-domain), leading to the generation of CFP-Epac(dDEP-CD)-YFP, also called H30 (27, 30). This probe displays faster activation kinetics and an extended dynamic range, but a lower affinity for cAMP than PKA. Nikolaev et al. developed sensors based on the fusion of individual cAMP binding domains of PKA or Epac to CFP and YFP (PKA-camps, Epac1-camps and Epac2-camps) (31). Just as with the full length Epac based sensors, FRET between YFP and CFP occurs in the unbound state, while cAMP binding induces a conformational change that is detectable as a FRET decrease (see Fig. 1b). 1.3.3. Targeted Bi- and Unimolecular FRET Sensors and Their Application

The finding that cAMP concentrations can fluctuate rapidly and non-uniformly within the cell led to the development of targeted FRET sensors for cAMP. Depending on the targeting domain these sensors can be localized to specific compartments within the cell, in order to quantify local fluctuations in cAMP.

1.3.4. Monitoring Localized cAMP Pools Using Targeted Sensors

The first targeted indicators for cAMP were generated by fusing specific targeting sequences to the N- or C-terminus of ICUE; thereby targeting ICUE to the plasma membrane, mitochondria or the nucleus, respectively. Simultaneous transfection of these probes revealed different spatio-temporal cAMP dynamics in the cytosol and at the plasma membrane in response to the activation of b-adrenergic or prostanoid receptors (28). A plasma membrane targeted PKA-based sensor (mpPKA) was generated by fusing the myristoylation/palmitoylation signal from the Lyn kinase (32) to the N-terminus of the RII subunit. Another unimolecular sensor targeted to the plasma membrane was generated by fusion of the same myristoylation/palmitoylation signal to the N-terminus of H30. A different modification to H30 was the fusion of a nuclear localization sequence (NLS) to its C-terminus, which induces translocation of the sensor to the nucleus (30). By using these sensors it was possible to demonstrate the existence of contiguous domains with different cAMP levels at the plasma membrane, in the cytoplasm and the nucleus, which was attributed to compartmentalized PDE activity (30). By fusing the dimerization/docking (D/D) domains of PKA-RI or RII subunits to the N-terminus of the Epac1-camps sensor (see Table. 1) the cAMP probe is targeted to those subcellular compartments where PKA type I and PKA type II is normally located in the

304

Gesellchen et al.

Fig. 2. Representative kinetics of FRET changes (R/R0) recorded in cardiac myocytes transfected with either the RI_epac or RII_epac sensor and challenged with isoproterenol (10 nM, a) or PGE1 (1 mM, b) as indicated. Values are normalized to the ratio value at time t = 0s (R0). Localization of RI_epac (c) and RII_epac (d) in neonatal cardiac myocytes (excitation 430 nm, emission 480 nm). Scale bars: 10 mm.

cell. Expression of these probes in cardiac myocytes (9) or skeletal muscle fibers (33) reveals a distinct distribution pattern for the two sensors, with RI_epac tightly associated to the Z and M sarcomeric lines and RII_epac strongly localized to the M line (see Fig. 2c, d). The different localization of the probes is due to the binding to specific endogenous AKAPs, mediated by the respective D/D domains, and allows detection of spatially and functionally distinct cAMP pools (see Fig. 2a, b) (9). It should be noted that the fusion of any targeting domain to a FRET-based probe can affect the FRET characteristics of the sensor. For example, both RI_epac and RII_epac sensors show a 50% reduction in the maximal FRET response compared to the original Epac1-camps sensor (9). 1.3.5. Monitoring Intracellular cAMP Kinetics Using Targeted Sensors

Apart from the spatial control, the magnitude and the duration of cAMP fluctuation play a critical role in cellular signal transduction. Cyclic AMP signals inside the cell can be modulated and terminated by different mechanisms including receptor desensitization, phosphodiesterase (PDE)-mediated hydrolysis of cAMP and Gi coupled receptor activation. Since binding of cAMP to FRET sensors is a reversible process, these probes provide valuable tools for monitoring cAMP dynamics with a temporal resolution that is not possible to achieve with conventional methods. For example, the experiment shown in Fig. 3a shows that isoproterenol stimulation of neonatal rat ventricular myocytes (NRVMs) results in a transient cAMP increase. The response peaks rapidly, within 30 s after the stimulus, and then

Measuring Spatiotemporal Dynamics of Cyclic AMP Signaling in Real-Time

305

Fig. 3. (a) Normalized kinetics of FRET changes (R/R0) recorded in cardiac myocytes expressing the Epac1-camps sensor on application of 10 nM isoproterenol and 100 mM IBMX. (b) Normalized kinetics of FRET changes recorded in cardiac myocytes expressing the RII_epac sensor on application of 10 nM isoproterenol and 30 mM acetylcholine (ACh).

decreases to a stable level after another 30 s. As another example of the ability of the FRET sensors to detect reversible cAMP changes, Fig.  3b shows the termination of the isoproterenolinduced b-adrenergic receptor cAMP response by the activation of the M2 muscarinic receptor. In the following sections we will describe in detail how to perform a FRET experiment using the RI_Epac and RII_epac sensors described above. As these sensors are based on CFP and YFP as donor and acceptor fluorophores, the filter sets used are specific for this scenario. Adjustments will be required when using other types of sensors. The FRET detection method described here is ratiometric and relies on the measurement of donor emission and sensitized acceptor emission. Other methods for FRET measurements have been developed, including acceptor/donor photobleaching (34), or fluorescence life-time measurements (35), which are beyond the scope of the present chapter.

2. Materials 2.1. FRET Imaging System

In principle, any epifluorescence microscope can be adapted for FRET imaging experiments. The basic setup for FRET imaging consists of a wide-field fluorescence microscope, a light source for

306

Gesellchen et al.

Fig. 4. (a) FRET imaging setup, for details see text. (b) Schematic representation of the light paths within the beam splitter. Emission wavelengths are separated by a 505 nm dichroic filter and collected through the band pass filters (480/30 and 545/35). (c) The individual images from the CFP (480 nm) and YFP (545 nm) channel are collected by a CCD camera chips. (d) Image analysis software is used to calculate and display a FRET ratio image.

excitation of CFP, a device to separate YFP and CFP emission signals, a digital camera to collect them, and a computer to store the images and analyze the data (see Fig. 4). Exemplified below is the FRET imaging system we use in our laboratory: 1. Microscope: Olympus IX71 inverted microscope. 2. Illumination source: Lambda LS Stand Alone Xenon Arc Lamp and Power Supply; Sutter Instrument (0661301) (see Note 1). 3. Neutral density filters: filters series 22000, filter set 22000A 25 mm (N.D.: 0.2, 0.4, 0.6, 1.0, 1.5), Chroma Technology (see Note 2). 4. CFP filter setting: excitation filter ET436/20x, dichroic mirror T455LP, emission filter ET480/40m (Chroma Technology). 5. YFP filter setting: excitation filter ET500/30x, dichroic mirror T515LP, emission filter ET535/30m (Chroma Technology). 6. FRET filter setting. In the microscope filter cube: CFP excitation filter ET436/20x, dichroic mirror 455DCLP (all from Chroma Technology). In the beam splitter: dichroic mirror 505DCLP, YFP emission filter 545  nm, CFP emission filter 480  nm (all from Chroma Technology) (see Note 3 and Fig. 4b).

Measuring Spatiotemporal Dynamics of Cyclic AMP Signaling in Real-Time

307

7. Objective: Olympus PlanApoN, 60×, NA 1.42 oil, 0.17/FN 26.5 (see Note 4). 8. Immersion oil: “IMMERSOL” 518F, Carl Zeiss. 9. Beam splitter: Dual-view simultaneous-imaging system (DV2 mag biosystem, Photometrics, ET-04-EM) equipped with the filters indicated in item 6 (see Note 5). 10. Detector: coolSNAP HQ monochrome camera system, Photometrics. 11. Shutter: Lambda 10-3 optical filter changer with smart shutter, Sutter Instrument (see Note 6). 12. Computer: Dell DE6700, 2.66 GHz Intel Core 2 Duo CPU, 3.50 GB RAM, 400 GB hard drive, Windows XP Professional version 2002. 13. Image acquisition and analysis software: Meta imaging series 7.1, MetaFluor, Molecular Devices (http://www.molecular­ devices.com). 2.2. Materials for Transfection

In this section, we are providing sample preparation and transfection protocols for primary cultured rat neonatal cardiomyocytes. For other cell-lines, use appropriate protocols for handling and transfection. 1. Primary cultured rat neonatal cardiomyocytes. 2. Transfection reagent 170–3351, Bio-Rad.

TransFectin™

Lipid

Reagent,

3. Microscope glass cover slips, 24 mm, cat. no 6310161, VWR International. 4. Laminin, mouse, cat. no. 354232, BD Biosciences; dissolved at 0.1 mg/mL in tissue culture water; aliquots kept at −80°C; working solution: dilute in serum-free medium to a final concentration of 0.02 mg/mL. 5. Forceps. 6. Coverslip holders or observation chambers (cat. no. A-7816, Molecular Probes).

3. Methods In this section we describe the workflow of a FRET experiment using the MetaFluor imaging software. A FRET experiment consists basically of three steps: 1. Sample preparation 2. Image acquisition 3. Image analysis

308

Gesellchen et al.

3.1. Sample Preparation 3.1.1. Seeding of Cardiomyocytes

1. Clean and sterilize high quality 24-mm cover glasses with an ether–ethanol solution (1:1), place the coverslips in six-well tissue culture plates and let them dry under a tissue culture hood. 2. Prepare a 1:5 dilution of laminin in serum-free medium (final concentration 0.02  mg/mL). Coat each coverslip with 250 mL of laminin, leave in a tissue culture incubator for at least 90  min. Aspirate the laminin solution before seeding the cells. 3. Seed 600,000 cardiac myocytes per well (see Note 7).

3.1.2. Cardiomyocyte Transfection

24 h after seeding and at least 24 h prior to the experiment, transfect the cardiomyocytes with the FRET sensor(s) of choice. The transfection protocol may vary depending on the cell-type and should be optimized in a pilot experiment. 1. For each well prepare two mixes (scale up quantities as required): (a) 250  mL of serum-free medium + 2.5  mg DNA (coding for FRET-sensor of choice, e.g., RI_Epac or RII_Epac) (b) 250 mL of serum-free medium + 7 mL TransFectin™ 2. Allow to stand for 5 min and mix the two solutions 3. After 20 min of incubation add the mix (500 mL) drop-wise to the well

3.2. Running a FRET Experiment

The example experiment is carried out using the FRET set-up described in Subheading  2.1, with the image acquisition and analysis software Meta imaging series 7.1, MetaFluor.

3.2.1. Software Protocol Set-Up

Before starting a time-course experiment, set up a common protocol for the whole series of experiments. Choose the appropriate objective (we preferably use a 60× objective, but a 40× or a 100× are also fine) and set the following parameters as discussed below: exposure time, frequency of acquisition, number of acquisitions and camera binning.

3.2.2. Exposure Time

Optimal exposure time depends on the expression levels of fluorophores as well as the characteristics of the lamp, the optics and the camera. We usually set an exposure time between 50 and 300 ms depending on the brightness of the cell under investigation (see Note 8). For new cell-lines, or a new FRET sensor a good strategy is to attenuate the light as much as possible and use very short exposure times. In our set-up it is not possible to adjust the lamp intensity; therefore, a good starting point is a neutral density filter with an optical density of 1.0 and an exposure time of 50 ms or less.

Measuring Spatiotemporal Dynamics of Cyclic AMP Signaling in Real-Time

309

3.2.3. Setting the Time Course and Number of Acquisitions

The frequency of acquisition defines the interval between two consecutive acquisitions. It normally varies between 2 and 60 s and depends on the sensor characteristics and the kinetics of cAMP changes under investigation. If following rapid and transient responses, shorter intervals between consecutive acquisitions should be used; however, increasing the frequency of acquisition also increases the probability of photo-damaging of cells and photobleaching of the fluorophores.

3.2.4. Camera Binning

Most CCD cameras have the ability to clock multiple pixel charges into a single larger charge or “super-pixel.” This super-pixel represents the area of all the individual pixels contributing to the charge. This process is referred to as binning. A binning of 1 × 1 means that each individual pixel is used as such; a binning of 2 × 2 means that an area of four adjacent pixels is combined into one larger pixel, and so on. In the case of 2 × 2 binning, there is a fourfold increase in signal (the four single pixel contributions), a twofold loss in resolution and a twofold improvement in signal-to-noise.

3.2.5. Run Protocol

In this phase of the experiment a suitable cell is selected, excited at a wavelength of 430 nm and the resulting CFP (480 nm) and YFP (545 nm) emissions are collected by the camera system on the two adjacent halves of the camera chip (see Fig. 4). 1. Mount the coverslip on the coverslip holder (or chamber) and fill the chamber with 900 mL of cell bathing solution. 2. Apply a small drop of oil to the objective lens and mount the holder on the microscope stage. 3. Visualize the cells in bright field, and establish their overall condition and morphology (see Note 9). 4. Identify fluorescent cells using the YFP filter set. Select cells with a signal to background ratio of at least 3:1 for the experiment; very bright cells as well as very dim ones should be disregarded. 5. Position the cell in the center of the field to avoid artifacts due to spherical and chromatic aberration and field curvature. 6. Switch to the FRET illumination setting (CFP excitation filter ET436/20x, dichroic mirror 455DCLP). Open the Command bar and click on “Focus” to open the Focus window. Select the excitation wavelength for CFP (430 nm) and press “Start Focusing.” This will start the camera acquisition and allow adjusting the focus. Once the image on the computer screen is in focus press “Stop Focusing” and close the window. The software will acquire image stacks in the CFP and the YFP channel and display them in two separate windows. A third window can be used to visualize a ratio image.

310

Gesellchen et al.

7. It is possible to follow online during the experiment the intensity changes of CFP and YFP as well as the ratio changes (see Note 10 for a brief discussion of the importance of this procedure). To follow ratio changes it is important to ensure correct alignment of the two images collected through the beamsplitter. To align the two images that correspond to the YFP and CFP emission wavelength: from the menu bar select “Run experiment” then “Align Wavelengths.” This will open a dialog box with arrows that can be used to move one of the two images with respect to the second one in order to perfectly superimpose them. 8. Draw a region of interest (ROI) on the cell (see Note 14 for some advice on drawing ROIs). From the experiment ­control-panel select “Regions” then choose the wavelength window in which to draw the ROI. The software automatically copies the ROI on CFP, YFP and ratio images and will open two separate windows (Graph 1 and Graph 2). Graph 1 will show the mean fluorescence intensities of CFP and YFP in real time, Graph 2 will display the YFP/CFP ratio changes over time (see Note 10). 9. Start the experiment. On the experiment control-panel tick “Log Data,” “Save Image” and “Save Ratio” (this will save the intensity values averaged over the ROIs as an “.XLS” file, and the image stacks as well as the ratio images as “.INF” files) then press “Start.” During the first few acquisitions the fluorescence intensity traces may drift due to bleaching and photo isomerization processes. This can result in an unstable baseline ratio value. Wait before adding the stimulus until the signal is reasonably stable. If the ratio baseline does not stabilize, select a different cell (see Note 11 for possible pitfalls commonly encountered during image acquisition). 10. Once the ratio signal is stable (normally after about 20 acquisitions), gently add the stimulus between two acquisitions (see Note 12). 11. Wait until the signal stabilizes before adding another stimulus or concluding the experiment. 3.3. Analysis

The MetaFluor software extracts the mean intensity values within the ROIs drawn at the beginning of the experiment from the CFP intensity image, the YFP intensity image and the mean values from the corresponding YFP/CFP ratio image. These values can then be exported in Microsoft Excel format (.XLS) as intensity values over time. However, we highly recommend reanalyzing the data offline to check for alignment, focus drift or changes in the position of the cell that may have occurred during the execution of the experiment. If the cell changes position during the experiment it is necessary to calculate the intensities and ratio

Measuring Spatiotemporal Dynamics of Cyclic AMP Signaling in Real-Time

311

values frame by frame, appropriately repositioning the ROI at every time point. 3.3.1. Data Processing

1. Open experiment: from the control-bar select “Open” and load the “.INF file.” This file contains all the information collected during the experiment recording and, on opening, the CFP and the YFP intensity windows as well as the ratio window will be displayed. 2. Align the CFP and YFP channels in order to optimally superimpose the images (from the menu bar select “Run experiment” then “Align Wavelengths”) (see Note 13). 3. Draw ROIs both on the cell and on the background for both CFP and YFP channels (to do this, from the experiment control-panel select “Regions”) (see Note 14). 4. On the experiment control-panel tick “Log Data”; in the following dialog, save the experiment as an .XLS file. 5. Press “F4 Forward” to rerun the experiment offline. In the .XLS file the software will extract the mean intensity values of CFP, YFP and YFP/CFP from the ROIs drawn. 6. Switch to the Excel sheet to continue the analysis. At each time point, for both CFP and YFP channels subtract the mean intensity of the background ROI from the corresponding mean intensity of the ROI drawn on the cell. From these corrected intensity values the intensity graph shown in Fig.  5a can be generated.

Fig. 5.  (a) CFP and YFP emission from a single cell expressing the cAMP sensor and treated with isoproterenol and forskolin/IBMX. Note the increase in donor emission and the concomitant decrease in acceptor emission. In (b) The CFP and YFP intensities shown in (a) are used to calculate the ratio (R) CFP/YFP.

312

Gesellchen et al.

7. After background correction the intensity values are used to calculate, at each time point, the ratio (R) between CFP and YFP mean intensity, as shown in Fig. 5b (see Note 15). These ratio values correlate with changes in intracellular concentration of cAMP.

4. Notes 1. The most common light sources are mercury or xenon arcdischarge lamp bulbs coupled to appropriate bandpass filters, or to a monochromator. Recent developments in light-emitting diodes (LEDs) have resulted in devices that have significantly more power than halogen and gas discharge lamps. These novel LEDs constitute a practical alternative to traditional microscope light sources. In addition, they offer several advantages, including better stability, longer lifetime (up to 100,000 h), lower cost of replacement, lower risk (no glass bulbs to break), and lower energy consumption than traditional light sources. 2. New lamps generate a more intense light beam than bulbs that have been in use for a longer time. To prevent excessive illumination and damage to the sample it is possible to mount neutral density filters to reduce the intensity of transmitted light. 3. It is essential to use filter sets matching CFP and YFP emission spectra. As a consequence of spectral overlap, the FRET signal will always be affected by direct excitation of the YFP acceptor at the donor excitation wavelength and by CFP emission into the YFP channel. This so-called “bleed-through” can be estimated and it is possible to generate a correction curve for it. CFP bleed-through into the YFP channel can be quantified as the ratio between YFP and CFP channel intensity in cells expressing CFP only, on excitation of CFP. In our system about 50% of CFP emission bleeds through into the YFP channel. If the sensor is efficient and the FRET signal is robust, however, such correction is not necessary. 4. Objective’s parameters such as numerical aperture (NA), magnification, optical tube length, and degree of aberration correction influence the quality of the final image in terms of brightness and resolution. Objectives with high quality lenses, a good degree of aberration correction and high numeric aperture (higher than 1.1) are preferable. 5. Simultaneous imaging of multiple fluorophores can also be achieved by using fast motorized filter wheels. With these devices fluorescence emission is gathered through wheelmounted filters that can switch between adjacent filters in

Measuring Spatiotemporal Dynamics of Cyclic AMP Signaling in Real-Time

313

30–50  ms, and electronic shutters having operating ­specifications in the same range. However, the use of a filter wheel implies that CFP and YFP emission images are collected with a delay that can introduce artifacts if the sample or the focal plane moves during the lag time between the two consecutive acquisitions. A possible alternative to the beamsplitter described above is to equip the microscope with a dichroic mirror to separate CFP and YFP emission signals and use two separate digital cameras for image acquisition. 6. Electronic shutters serve to limit exposure of the sample to damaging radiation during periods in which no images are acquired. Reduction of photodamage, phototoxicity, and photobleaching improves the quality of the images and cell viability over the long periods of time that are often required in time-lapse experiments. 7. Seed the cells at least 2 days prior to imaging to let them attach to the coverlips. Try to avoid high confluence on the day of the experiment, since confluent cells may downregulate cell-surface receptors and may show a reduced response to hormonal stimuli. 8. Increasing the exposure time allows the flux of photons coming from the sample to accumulate in the detector thus enhancing the intensity of the image. At the same time this will increase photodestruction of the fluorophore and photobleaching of the signals. In addition, detectors have a limited capacity to hold electrons. When the maximum capacity is reached the corresponding pixel will be saturated and any further change in the signal will be masked. 9. To perform successful live-cell imaging experiments it is important to carefully monitor the overall status of the cells: temperature, oxygenation, humidity, nutritional supplements, pH and osmolarity are crucial parameters that, if overlooked, may result in high variability among the samples. To limit the free radical damage caused by molecular oxygen, supplement the cell bathing solution with scavengers, such as ascorbate or trolox. In order to reduce the level of background noise and to minimize phototoxicity, the pH indicator phenol red should be avoided in fluorescence live-cell imaging because of its slight autofluorescence. 10. In order to perform a successful FRET experiment, it is important to constantly monitor the intensity of the two fluorescence emissions (CFP and YFP). If bleaching occurs, a decrease in the intensity of the two fluorophores is observed. These unspecific changes in the intensity of the two fluorophores may give rise to a change in the YFP/CFP ratio in the absence of significant cAMP variation. Ratio changes can

314

Gesellchen et al.

r­ eliably be considered as FRET changes only when determined by a change in donor emission which is paralleled by an opposite change in acceptor emission (e.g., an increase in CFP emission must be accompanied by a decrease in YFP emission). 11. Focus drift: this phenomenon is one of the most common problems in long term imaging. It is caused by fluctuations in the axial position of the microscope focal plane during the collection of sequential images. Temperature variations in the imaging chamber or in the microscope room can be responsible for these changes; thermal instability leads to slight bending of the coverslip or temperature gradients in the optical microscope trail give rise to focus drift. Other possible causes include irregularities in the surface of the coverslips or mechanical vibrations. If an oil immersion objective is being used, uneven spreading of the oil and trapped air bubbles may cause focus drift as well. 12. To avoid artifacts the added stimuli should be at the same temperature as the cells. Care must be taken when adding the drugs since this may cause changes in the focal plane and in the position of the cell. Do not touch the cell holder or the microscope stage! 13. If different software is used for acquisition, the images of CFP and YFP emission may not be split automatically. MetaFluor automatically provides the YFP/CFP ratio image in a pseudocolor Look-Up Table (LUT) where typically violet is assigned to black and dark gray pixels, and red colors are assigned to pixels close to absolute white. Changes in fluorescence are thereby represented by change in color. 14. The ROI on the cell should comprise the entire surface of the cell, excluding the nucleus and possible bright aggregates. It should remain within the cell for the entire duration of the experiment. The background ROI should be drawn close to the cell of interest, in an area representative for the background. 15. Most imaging software allows the experimenter to choose between calculating the intensity emission ratio either as 545 nm/480 nm (YFP/CFP) or as 480 nm/545 nm (CFP/ YFP). Since most probes show a decrease in FRET in the presence of cAMP, we usually calculate this ratio as 480 nm/545 nm. This makes the interpretation of the experiment more intuitive because an increase in ratio values corresponds to an increase in cAMP. Ratiometric measurements correct for unequal probe distribution and, within certain limits, for bleaching occurring during the experiment and changes in focus. However, care must be taken when aligning the two channels (to avoid artifacts) and when interpreting ratio values.

Measuring Spatiotemporal Dynamics of Cyclic AMP Signaling in Real-Time

315

If necessary, it is possible to normalize the ratio values to R0, the FRET ratio at time-point t = 0. In this case the FRET response is expressed as DR/R0 %, according to the formula

Rt - Rt1 DR %= 2 ´ 100, R0 R0



with Rt1 and Rt2 being the FRET ratios before and after addition of the stimulus, respectively (averaged over at least five ratio values). References 1. Lee, D. C., Carmichael, D. F., Krebs, E. G. and McKnight, G. S. (1983) Isolation of a cDNA clone for the type I regulatory subunit of bovine cAMP-dependent protein kinase. Proc. Natl. Acad. Sci. USA 80, 3608–3612. 2. Takio, K., Smith, S. B., Krebs, E. G., Walsh, K. A. and Titani, K. (1984) Amino acid sequence of the regulatory subunit of bovine type II adenosine cyclic 3¢,5¢-phosphate dependent protein kinase. Biochemistry 23, 4200–4206. 3. Taylor, S. S., Kim, C., Vigil, D., Haste, N. M., Yang, J., Wu, J. and Anand, G. S. (2005) Dynamics of signaling by PKA. Biochim. Biophys. Acta 15754, 25–37. 4. Wong, W. and Scott, J. D. (2004) AKAP signalling complexes: focal points in space and time. Nat. Rev. Mol. Cell. Biol. 5, 959–970. 5. Gronholm, M., Vossebein, L., Carlson, C. R., Kuja-Panula, J., Teesalu, T., Alfthan, K., Vaheri, A., Rauvala, H., Herberg, F. W., Tasken, K. and Carpen, O. (2003) Merlin links to the cAMP neuronal signaling pathway by anchoring the RIb subunit of protein kinase A. J. Biol. Chem. 278, 41167–41172. 6. Miki, K. and Eddy, E. M. (1998) Identification of tethering domains for protein kinase A type Ia regulatory subunits on sperm fibrous sheath protein FSC1. J. Biol. Chem. 273, 34384–34390. 7. Huang, L. J., Durick, K., Weiner, J. A., Chun, J. and Taylor, S. S. (1997) D-AKAP2, a novel protein kinase A anchoring protein with a putative RGS domain. Proc. Natl. Acad. Sci. USA 94, 11184–11189. 8. Huang, L. J., Durick, K., Weiner, J. A., Chun, J. and Taylor, S. S. (1997) Identification of a novel protein kinase A anchoring protein that binds both type I and type II regulatory subunits. J. Biol. Chem. 272, 8057–8064. 9. Di Benedetto, G., Zoccarato, A., Lissandron, V., Terrin, A., Li, X., Houslay, M. D., Baillie, G. S. and Zaccolo, M. (2008) Protein kinase A type I

and type II define distinct intracellular signaling compartments. Circ. Res. 103, 836–844. 10. Wrighton, K. H. (2009) Sensing second messengers. Nat. Cell Biol. 11, S20–S21. 11. Steiner, A. L., Kipnis, D. M., Utiger, R. and Parker, C. (1969) Radioimmunoassay for the measurement of adenosine 3¢,5¢-cyclic phosphate. Proc. Natl. Acad. Sci. USA 64, 367–373. 12. Kariv, I. I., Stevens, M. E., Behrens, D. L. and Oldenburg, K. R. (1999) High throughput quantitation of cAMP production mediated by activation of seven transmembrane domain receptors. J. Biomol. Screen. 4, 27–32. 13. Prystay, L., Gagne, A., Kasila, P., Yeh, L. A. and Banks, P. (2001) Homogeneous cell-based fluorescence polarization assay for the direct detection of cAMP. J. Biomol. Screen. 6, 75–82. 14. Gabriel, D., Vernier, M., Pfeifer, M. J., Dasen, B., Tenaillon, L. and Bouhelal, R. (2003) High throughput screening technologies for direct cyclic AMP measurement. Assay Drug Dev. Technol. 1, 291–303. 15. Kumar, M., Hsiao, K., Vidugiriene, J. and Goueli, S. A. (2007) A bioluminescent-based, HTS-compatible assay to monitor G-proteincoupled receptor modulation of cellular cyclic AMP. Assay Drug Dev. Technol. 5, 237–245. 16. Förster, T. (1948) Zwischenmolekulare Energiewanderung und Fluoreszenz. Annalen der Physik 437, 55–75. 17. Lakowicz, J. (2006) Energy transfer, in Principles of fluorescence spectroscopy pp 443–471, Springer, New York. 18. Patterson, G. H., Piston, D. W. and Barisas, B. G. (2000) Forster distances between green fluorescent protein pairs. Anal. Biochem. 284, 438–440. 19. Adams, S. R., Harootunian, A. T., Buechler, Y. J., Taylor, S. S. and Tsien, R. Y. (1991) Fluorescence ratio imaging of cyclic AMP in single cells. Nature 349, 694–697.

316

Gesellchen et al.

20. Goaillard, J. M., Vincent, P. V. and Fischmeister, R. (2001) Simultaneous measurements of intracellular cAMP and L-type Ca2+ current in single frog ventricular myocytes. J. Physiol. 530, 79–91. 21. Zaccolo, M., De Giorgi, F., Cho, C. Y., Feng, L., Knapp, T., Negulescu, P. A., Taylor, S. S., Tsien, R. Y. and Pozzan, T. (2000) A genetically encoded, fluorescent indicator for cyclic AMP in living cells. Nat. Cell Biol. 2, 25–29. 22. Lissandron, V., Terrin, A., Collini, M., D’Alfonso, L., Chirico, G., Pantano, S. and Zaccolo, M. (2005) Improvement of a FRETbased indicator for cAMP by linker design and stabilization of donor-acceptor interaction. J. Mol. Biol. 354, 546–555. 23. Zawadzki, K. M. and Taylor, S. S. (2004) cAMPdependent protein kinase regulatory subunit type IIb: active site mutations define an isoformspecific network for allosteric signaling by cAMP. J. Biol. Chem. 279, 7029–7036. 24. Mongillo, M., McSorley, T., Evellin, S., Sood, A., Lissandron, V., Terrin, A., Huston, E., Hannawacker, A., Lohse, M. J., Pozzan, T., Houslay, M. D. and Zaccolo, M. (2004) Fluorescence resonance energy transfer-based analysis of cAMP dynamics in live neonatal rat cardiac myocytes reveals distinct functions of compartmentalized phosphodiesterases. Circ. Res. 95, 67–75. 25. de Rooij, J., Zwartkruis, F. J., Verheijen, M. H., Cool, R. H., Nijman, S. M., Wittinghofer, A. and Bos, J. L. (1998) Epac is a Rap1 guanine-nucleotide-exchange factor directly activated by cyclic AMP. Nature 396, 474–477. 26. Bos, J. L. (2003) Epac: a new cAMP target and new avenues in cAMP research. Nat. Rev. Mol. Cell. Biol. 4, 733–738. 27. Ponsioen, B., Zhao, J., Riedl, J., Zwartkruis, F. J., van der Krogt, G., Zaccolo, M., Moolenaar, W. H., Bos, J. L. and Jalink, K. (2004) Detecting cAMP-induced activation by fluorescence resonance energy transfer: Epac as a novel cAMP indicator. EMBO Rep. 5, 1–5.

28. DiPilato, L. M., Cheng, X. and Zhang, J. (2004) Fluorescent indicators of cAMP and Epac activation reveal differential dynamics of cAMP signalling within discrete subcellular compartments. Proc. Natl. Acad. Sci. USA 101, 16513–16518. 29. De Arcangelis, V., Liu, R., Soto, D. and Xiang, Y. (2009) Differential association of phosphodiesterase 4D isoforms with b2-adrenoceptor in cardiac myocytes. J. Biol. Chem. 284, 33824–33832. 30. Terrin, A., Di Benedetto, G., Pertegato, V., Cheung, Y. F., Baillie, G., Lynch, M. J., Elvassore, N., Prinz, A., Herberg, F. W., Houslay, M. D. and Zaccolo, M. (2006) PGE1 stimulation of HEK293 cells generates multiple contiguous domains with different [cAMP]: role of compartmentalized phosphodiesterases. J. Cell Biol. 175, 441–451. 31. Nikolaev, V. O., Bunemann, M., Hein, L., Hannawacker, A. and Lohse, M. J. (2004) Novel single chain cAMP sensors for receptorinduced signal propagation. J. Biol. Chem. 279, 37215–37218. 32. Resh, M. D. (1999) Fatty acylation of proteins: new insights into membrane targeting of myristoylated and palmitoylated proteins. Biochim. Biophys. Acta 1451, 1–16. 33. Roder, I. V., Lissandron, V., Martin, J., Petersen, Y., Di Benedetto, G., Zaccolo, M. and Rudolf, R. (2009) PKA microdomain organisation and cAMP handling in healthy and dystrophic muscle in vivo. Cell. Signal. 21, 819–826. 34. Kenworthy, A. K. (2005) Photobleaching FRET microscopy, in Molecular Imaging: FRET Microscopy and Spectroscopy (Periasamy, A., and Day, R. N., Eds.), Oxford University Press, New York. 35. Periasamy, A., Elangovan, M., Elliott, E. and Brautigan, D. L. (2002) Fluorescence lifetime imaging (FLIM) of green fluorescent fusion proteins in living cells. Methods Mol. Biol. 183, 89–100.

Chapter 17 Determining the Activation of Rho as an Index of Receptor Coupling to G12/13 Proteins Michio Nakaya, Mina Ohba, Motohiro Nishida, and Hitoshi Kurose Abstract Heterotrimeric G proteins are composed of a, b, and g subunits. G proteins can be activated by a large number of cell-surface hepathelical receptors and can transduce signals from these receptors to various intracellular signaling molecules. When G protein-coupled receptors are bound by their cognate ligand, interaction with specific subtypes of G protein leads to dissociation of the a subunit of the heterotrimeric G protein from the bg dimer, and both Ga-GTP and Gbg are capable of initiating their own signal transduction pathways. G proteins are functionally divided into four groups based on the nature of a subunit into Gs, Gi, Gq, and G12 families. The members of the G12 subfamily are G12 and G13. Increasing evidence indicates that G12/13 proteins play critical roles in various physiological functions. G12 and G13 regulate the small GTPase Rho through modulation of guanine nucleotide exchange factor (RhoGEF) activity to regulate various cellular responses, such as cytoskeletal changes and cell growth. Therefore, Rho activity can often represent a sensitive marker of G12/13 activity. Here, we describe the Rho activation assay to monitor the activity of G12/13 proteins. Key words: G12, Rho, Pull-down, Cardiomyocyte

1. Introduction G proteins cycle between GDP- and GTP-bound forms. The GDP-bound form is the inactive form, and the GTP-bound form constitutes the active state. The active GTP-bound form of the G protein transmits the signal to downstream proteins. Thus far, two methods have been used to evaluate the activation of heterotrimeric G proteins and small GTPases. One is the radioisotope method that estimates the incorporation of [35S]GTPgS into Ga proteins (1). Another method is a nonradioisotopic, biochemical assay using GST-fusion proteins that are unique to each G protein (2, 3). The proteins fused with GST are downstream molecules

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_17, © Springer Science+Business Media, LLC 2011

317

318

Nakaya et al.

and specifically bind the active form of corresponding Ga proteins. Although the original radioisotope method measured all [35S] GTPgS-bound G proteins, combining the [35S]GTPgS-binding step with a pull-down method allows the identification of specific Ga protein-binding partners. Specific-binding substrates of G12/13 proteins have been identified, and it has been reported that the activation status of G12/13 can be measured by pull-down assay using a GST-fused tetratricopeptide repeat (TPR) domain of the type 5 serine/threonine phosphatase (PP5) (4, 5). Through the TPR domain, PP5 strongly interacts with the active form of G12/13. However, the pull-down assay using GST–TPR has a disadvantage in that it lacks sensitivity. In the case of small GTPases, whose activation state is often measured by pull-down assay, the signals received by the GTPases are already amplified by several steps of the upstream signal transduction cascade. In contrast, heterotrimeric G proteins reside immediately downstream of the receptor, allowing less amplification and making the assay rather and insensitive index of receptor–G12/13 protein interactions. To overcome this difficulty, changes in second messenger concentrations have often been used as indices of Gs, Gi, and Gq activation. However, receptor-driven activation of G12/13 proteins does not directly influence the amount of classical second messengers (e.g., cyclic AMP, inositol 1,4,5-trisphosphate/diacylglycerol, Ca2+) significantly. This has led us to focus on Rho activation by G12/13 proteins. It has been reported that in many situations, G12/13 proteins activate the Rho protein and through such activation G12/13 exerts the physiological consequences of various stimuli (6–8). One of the pathophysiological stimuli that activates G12/13 proteins is the stretching of cardiomyocytes as described below (9). In this chapter, we describe the Rho activation assay as an index of G12/13 protein activation.

2. Materials 2.1. For Preparation of GST–Rhotekin-RBD Protein

1. Glutathione S-transferase gene fusion vector (pGEX-6P-1; GE Healthcare, #27-4597-01). 2. Competent Escherichia coli cells (BL21; Novagen, #69449). 3. Isopropyl b-d-1-thiogalactopyranoside (IPTG: Sigma, #I6758). 4. High performance centrifuge (see Note 1). 5. 0.8 × 4  cm poly-prep chromatography column (Bio-Rad Laboratories, #731-1550) (see Note 2). 6. Seamless cellulose tubing (Viskase Companies Inc., #UC18-321002) (see Note 3).

Determining the Activation of Rho as an Index of Receptor

2.2. For Preparation of Rat Neonatal Cardiomyocytes

319

1. Twelve 1-day-old Sprague–Dawley rats. 2. Silicone rubber culture dishes (10 cm2) (STREX). 3. Laminin (Invitrogen, #23017-015) (see Note 4). 4. Fatty acid-free bovine serum albumin (BSA) (SIGMA, #A6003). 5. 4-Hydroxybutyric acid (SIGMA, #H6501). 6. Collagenase A (Roche, #11 088 793 001). 7. 0.20-mm syringe filter (Sartorius Stedim, #16534). 8. Taurine (WAKO, #205-00115) (Stock: 500 mM in H2O). 9. Dulbecco’s modified Eagle medium (DMEM) (low glucose) (GIBCO, #10313-21). 10. Fetal bovine serum (FBS) (Hyclone, Ogden, UT). 11. Insulin–transferrin–selenium #51500-056).

(ITS)

solution

(GIBCO,

12. Penicillin–streptomycin solution (GIBCO, #15140). 13. Sterile surgical blades (ELP, #FKB10 HANDLE3). 14. 70-mm nylon cell strainer (FALCON, #352350). 15. 10-cm plates (Corning, #430591). 2.3. For Pull-Down Assay Using by GST–Rhotekin-RBD Protein

1. Cell scraper (SUMILON, #MS-93170). 2. Glutathione-Sepharose 4B (GE Healthcare, #17-0756-01).

3. Methods 3.1. Preparation of GST–Rhotekin-RBD Protein

1. Subclone cDNA encoding Rho-binding domain (RBD) of mouse Rhotekin (amino acids 7–89) into pGEX-6P-1 vector and confirm the sequence. 2. The plasmid is transformed into E. coli (Strain: BL21) using a standard method. 3. Pick a single colony and inoculate the E. coli into 50 mL LB liquid culture containing 50  mg/mL ampicillin, in a sterile 200-mL flask. 4. Incubate E. coli culture overnight at 37°C with agitation. 5. Transfer 50 mL of the E. coli culture to 500 mL LB medium (containing 50 mg/mL ampicillin) in a sterile 2-L flask. 6. Incubate the culture at 37°C until the OD600 reaches about 0.8 (see Note 5). 7. Add IPTG [final concentration: 500 mM, (see Note 6)] to the culture and incubate for 20–24 h at 20°C with constant shaking.

320

Nakaya et al.

8. E. coli from the culture is collected by centrifugation: ~4,600 × g (equivalent to 5,000 rpm in a Beckman JLA-10.5 rotor) for 30 min at 4°C. 9. Suspend the pellet in phosphate-buffered saline (PBS) containing 10 mM EDTA. 10. The bacterial suspension is centrifuged at ~1,200 × g (2,500  rpm with Beckman prechilled JLA-10.5 rotor) for 30 min at 4°C. 11. The supernatant is removed and the pellet is frozen in a freezer (−80°C). 12. Once the pellet is completely frozen, the pellet is thawed and suspended into 30 mL PBS containing 10 mM EDTA. 13. The E. coli suspension is maintained at 0–4°C and is homogenized using a Dounce homogenizer (ten strokes). 14. Add 3.3 mL of 10% TritonX-100 to the homogenized E. coli suspension (final TritonX-100 concentration = 1%) and mix thoroughly. 15. The E. coli suspension is then subjected to sonication (5 × 1 min sonication periods separated by 5 × 1 min periods on ice) (see Note 7). 16. The lysate is clarified by centrifugation at ~41,000 × g (18,500  rpm with prechilled JA-20 rotor of Beckman) for 30 min at 4°C. 17. Prepare 1 mL of Glutathione Sepharose beads equilibrated by PBS containing 1 mM EDTA (PBS-E) (see Note 8). 18. After centrifugation, carefully pour the supernatant into a 50-mL prechilled tube and add the equilibrated Glutathione Sepharose beads. 19. Rotate the tube gently for 2 h in a cold room (at 4°C). 20. Centrifuge the tube at ~400 × g (1,500  rpm with Beckman prechilled rotor JS7.5) for 5 min at 4°C. 21. Carefully remove the supernatant and suspend the beads gently in 20 mL cold PBS-E. 22. Repeat the above wash procedures (steps 21 and 22) four times. 23. After the final wash, suspend the beads in 5 mL cold PBS-E and move to a cold room with the beads on ice. 24. Fix the 0.8 × 4  cm poly-prep chromatography column to a stand with clamps and pour the well-suspended beads onto the column. Then add a further 3 mL cold PBS-E (see Note 9). 25. Wait until the beads are packed in the column by gravity, and add 3 mL cold PBS-E carefully for an additional wash.

Determining the Activation of Rho as an Index of Receptor

321

26. Elute by adding 1 mL 10 mM glutathione in 50 mM Tris/ HCl, pH 8.0, collecting the eluted fraction (1 mL) in a tube (see Note 10). 27. Repeat step 26 two more times. 28. Pool the three elute fractions together (3 mL) and transfer into the washed dialysis tube (seamless cellulose tubing) (see Note 11). 29. Push air out from the dialysis tube and close the tube tightly using appropriate clean clips. Note that some air can be left within the sac to help float the tube during dialysis. 30. Submerge the tube in the cold 2 L PBS and agitate the solution slowly using a magnetic stirrer for 2 h. 31. Remove the tube and place into fresh, cold PBS (2 L) and stir gently overnight. 32. Transfer the dialyzed solution into a 15-mL tube precooled on ice. 33. 20 mL of the recovered solution is mixed with 2× SDS sample buffer [100 mM Tris/HCl, pH 6.8, 20% glycerol, 2% 2-mercaptoethanol, 2% SDS, 0.04% bromophenol blue] and boiled for 5 min at 95°C. 34. The sample is centrifuged (20,000 × g for 1  min at room temperature) and the supernatant is collected. 35. The recovered sample is subjected to 14.5% SDS– polyacrylamide gel followed by Coomassie Brilliant Blue staining. Check that the band for GST–Rhotekin-RBD protein appears at the appropriate position (~35 kDa). 36. 10 mL of the dialyzed solution is used for the measurement of the protein concentration of the solution. Determine the concentration of the protein solution by standard protein quantification assays, such as Lowry, Bradford, and BCA assays (see Note 12). 37. Dispense 200 mL aliquots into screw-capped tubes and store them at −80°C until required. 3.2. Preparation of Rat Neonatal Cardiomyocytes

1. Add 1 mL of 2 mg/mL laminin solution onto silicone-rubber culture dishes (10  cm2) and incubate overnight at 37°C. Immediately before cell plating, the solution is decanted and washed by 1 mL PBS. 2. Prepare 100 mL PBS solution containing 0.1% 4-hydroxybutyric acid, and 50 mg taurine and adjusted to pH 7.4 at 4°C [referred to as PBS(−) solution below]. 50 mg collagenase A and 50 mg fatty acid-free BSA are dissolved in 50 mL of the PBS(−) solution [referred to as PBS(+) solution below]. 3. Sterilize the solutions by passing through a 0.20-mm syringe filter.

322

Nakaya et al.

4. Twelve 1-day-old Sprague-Dawley rats are deeply anesthetized (see Note 13). 5. Hearts are expeditiously extirpated one after another and kept in ice-cold PBS(−) until tissue sampling has been completed (see Note 14). 6. The atria of the hearts are quickly removed using sterilized forceps and scissors. 7. Ventricles are transferred into fresh 2 mL PBS(−) in a 35-mm plate. 8. The PBS+ is removed and the hearts are washed by the ­addition of another 2  mL PBS(−) to remove the blood completely. 9. Transfer the hearts to a fresh 35-mm plate containing 500 mL PBS(−). 10. Cut each heart into quarters using a sterile surgical blade. Wash the pieces twice with 1 mL PBS(−). Then add 3 mL PBS(−). 11. Cut off the bottom (1–2  mm) of a 1-mL pipette tip at an angle to create a wider tip opening to suck the heart pieces and transfer them to a 50-mL FALCON tube. 12. Add an equal volume (3 mL) of PBS(+) (containing collagenase/BSA; see step 2 above). 13. Wrap the cap of the tube tightly with parafilm and shake the tube vigorously (~90  oscillations/min) in the shaking bath for 2 min at 37°C. 14. Discard the supernatant. Add 8 mL PBS(+). Shake the tube vigorously in the bath at 37°C for 7 min. 15. Collect the supernatant carefully and filter through a 70-mm cell strainer. 16. The filtrate is centrifuged at ~180 × g for 2  min at room temperature. 17. Suspend the cell pellet in 5 mL of 10% FBS/DMEM by gentle trituration. Add 8 mL PBS(+) containing collagenase to the remaining heart pieces and shake the tube vigorously in shaking bath for 7 min at 37°C. 18. Repeat steps 15–17 four times (see Note 15). 19. Combine the cell suspensions together (total volume ~25 mL), and plate the cells onto 10 cm2 plates to attach the fibroblasts present in the cell suspension. 20. Incubate the cells in a CO2:air incubator at 37°C for 1  h. Transfer the nonattached cells to a new tube (see Note 16). Count the number of cells and adjust it to 2 × 106 cells/mL. 21. Plate the cells onto the laminin-coated silicone-rubber culture dishes (~80% confluency).

Determining the Activation of Rho as an Index of Receptor

323

22. On the next day, the medium is replaced by serum-free DMEM with insulin, transferrin, selenium, taurine, penicillin, streptomycin [DMEM (low glucose) 500  mL containing 30 mL ITS solution, 1 mM taurine, 2.5 mL penicillin–streptomycin solution]. 3.3. Pull-Down Assay Using the GST– Rhotekin-RBD Protein

1. Cardiomyocytes in the 10 cm2 laminin-coated silicone rubber culture dishes are stretched by 20% with a stretch system (Fig.  1a) for the desired times and then, the medium is completely removed using an aspirator.

Fig. 1.  Mechanical stretch activates Rho through Ga12/13. (a) A device for the cell stretch experiment. (b) Overall western-blot image of the pulled-down samples immunoblotted by anti-RhoA antibody. An arrow indicates the band for RhoA protein. (c) Time course of RhoA activation by mechanical stretch. Cardiomyocytes on the silicone chamber dishes were stretched for the indicated times and subjected to the RhoA pull-down assay. (d) Ga12/13mediated RhoA activation by mechanical stretch. Cells were transfected with GFP, p115RGS, and inactive form of p115-RGS (p115-mut) by electroporation. RGS is an abbreviation of regulator of G protein signaling. RGS domains consist of ~220 amino acids. They selectively bind to GTP-bound form of a subunit of heterotrimeric G protein, and enhance GTPase activity of a subunit. p115-RGS is the RGS domain of p115RhoGEF. As it selectively binds Ga12/13, it can be used as a Ga12/13-specific inhibitor. p115-mut is a p115-RGS mutant that loses the binding to Ga12/13 (reproduced, with permission, from ref. 9).

324

Nakaya et al.

2. To the cells, 700 mL lysis buffer [50 mM Tris/HCl, pH 7.5, 0.1% TritonX-100, 10% glycerol, 150  mM NaCl, 30  mM MgCl2, 1 mM dithiothreitol, 10 mg/mL leupeptin, 10 mg/ mL aprotinin, 1 mM phenylmethysulfonylfluoride (PMSF)] is added. The cells are scraped using a cell scraper. 3. Lysates are collected into 1.5-mL microcentrifuge tube and frozen in a deep freezer. 4. After freezing and thawing, the lysates are homogenized by passage through a 26G syringe and centrifuged at 20,000 × g for 10 min for 4°C. Supernatants are carefully collected into new tubes placed on ice (see Note 17). 5. 10  mL of the supernatant is sampled, mixed with 2× SDS sample buffer and boiled for 5  min at 95°C. The boiled mixture is centrifuged at 20,000 × g for 1  min at room temperature and the supernatant is collected (referred to below as “samples for normalization”). 6. 10  mg of GST–Rhotekin-RBD premixed with 20  mL bed volume of Glutathione Sepharose beads (40 mL of 50% resin slurry) is added to the remaining cell lysate supernatants (see Note 18). 7. Incubate the mixture with gentle rocking for 120 min at 4°C. After incubation, the mixture is centrifuged at 12,000 × g for 1 min at 4°C and the supernatant removed. 8. The glutathione Sepharose beads are washed two times with 0.5  mL lysis buffer without protease inhibitors. After each buffer addition, beads are sedimented by centrifugation (12,000 × g, 1 min, 4°C) and finally suspended in 60 mL 2× SDS sample buffer. 9. Samples are boiled at 95°C for 5  min, and centrifuged (20,000 × g, 1  min, room temperature). The supernatant is carefully collected into a new tube. The sample is then ready for SDS–polyacrylamide gel electrophoresis (SDS–PAGE). 3.4. SDS–PAGE and Western Blotting

1. 20 mL of the pull-down samples or “samples for normalization” are loaded on to the 14.5% SDS–polyacrylamide gels. After the separation of proteins by SDS–PAGE, the gel is transferred to a PVDF membrane by western blotting [western blotting buffer: 10  mM cyclohexylaminopropanesulfonic acid 10% methanol, pH 11] for 1 h at 50 mA/cm2 using the semidry method. 2. The membrane is blocked in 5% BSA in TBS-T [TBS-T: 20 mM Tris/HCl, pH 7.5, 137 mM NaCl, 0.2% Tween-20] for 1 h at room temperature with gentle shaking in a container. 3. Discard the blocking solution and add a solution containing the anti-RhoA mouse monoclonal antibody (Santa Cruz Biotechnology #sc-418, 1:2,000 dilution) in TBS-T containing 0.5% BSA and 0.1% NaN3.

Determining the Activation of Rho as an Index of Receptor

325

4. Incubate the membrane for 1  h at room temperature with gentle shaking. 5. Remove the antibody solution and wash the membrane three times with TBS-T allowing 10 min for each wash. 6. Incubate the membrane in the antibody solution containing the HRP-conjugated anti-mouse IgG goat antibody (Santa Cruz Biotechnology #sc-2005, 1:10,000 dilution) in TBS-T. 7. Incubate the membrane for 1  h at room temperature with gentle shaking. 8. Remove the secondary antibody solution and wash the membrane three times with TBS-T allowing 10 min for each wash. 9. Incubate the membrane with chemiluminescent substrate (western lightning, Plus-ECL) at room temperature for 1 min. 10. Expose the membrane to X-ray film (typically a few minutes). 11. The bands on the film are scanned and converted to digital images. 12. Band densities are analyzed using Scion Image (for Windows PC) or NIH Image (for Mac) packages (see Note 19). 13. Examples of the results that can be obtained are presented in Fig. 1b–d. 14. We can also detect G12/13-Rho activation in vivo. Fig. 2 presents data for G12/13-mediated Rho activation of mouse hearts after pressure overload.

Fig. 2.  Pressure overload increases the activity of RhoA through Ga12/13 in mouse hearts. We generated mice with a cardiomyocyte-specific overexpression of p115-RGS protein (p115-Tg mice). Pressure overload was induced by surgical transverse aortic constriction (TAC) in WT and p115-Tg mice. One week after the surgery, hearts of WT and p115RGS mice were extirpated. The whole hearts was completely disrupted by a homogenizer in lysis buffer [50  mM Tris (pH 7.5), 0.1% TritonX-100, 10% glycerol, 150  mM NaCl, 30 mM MgCl2, 1 mM DTT, 1 mM PMSF, 2 mg/mL leupeptin, and 10 mM pepstatin] on ice. The homogenized hearts were subjected to freezing and thawing. Then, the protein concentration of the samples was determined by the Bradford protein assay. To the solution containing 2  mg of the total protein, 10 mg GST–Rhotekin-RBD and 30  mL 50% slurry Glutathione Sepharose beads were added and incubated with gentle rocking overnight at 4°C. The wash procedures are the same described in Subheading 3.3, step 3. Representative results are shown (reproduced, with permission, from ref. 9).

326

Nakaya et al.

4. Notes 1. We use an AVANTI HP-25I centrifuge and JLA-10.5, JA-20, JS7.5 rotors (Beckman Coulter) for the purification of GST–Rhotekin-RBD. 2. Cut off the tip of the empty column and wash the column briefly by PBS containing 1 mM EDTA. 3. Seamless cellulose tubing is cut to the appropriate size and washed with double-distilled water. Tubing is boiled for 5  min in PBS and then soaked with new PBS at room temperature. 4. Laminin solution is dispensed as 50 mL aliquots and store at −80°C. For usage, one tube of the solution is thawed and diluted (with 25 mL) PBS to 2 mg/mL. 5. This step usually takes about 1  h. Do not allow OD600 to exceed 0.8. It is important that E. coli is in an exponential growth phase to maximize the expression of fusion protein. 6. IPTG should be dissolved in H2O (stock solution: 500 mM) and stored in aliquots at −20°C. 7. This step depends on the sonicator. Optimal conditions should be experimentally determined. The viscosity of the solution gradually decreases during sonication. Note that oversonication disrupts the structure of the proteins. If the solution is significantly warmed by one sonication, the duration of sonication should be shortened, or the intensity of sonication should be decreased. Try milder conditions first, and gradually increase the intensity of sonication. 8. Shake the bottle of Glutathione Sepharose 4B beads gently to make a homogeneous suspension. Using a pipette, transfer the beads to an appropriate tube. As commercially supplied Glutathione Sepharose 4B is an ~75% slurry. Therefore, calculate the volume of the beads that you need. Add tenfold volume of PBS containing 1 mM EDTA and mix well. Centrifuge the beads at 500 × g for 5 min. Carefully remove the supernatant. Repeat this wash procedure three times. After the final wash, add an equal volume of PBS containing 1 mM EDTA to make an ~50% slurry of beads. 9. Do not allow the beads to dry out during the column preparation. 10. Take care not to disrupt the packed beads. 11. Remember to put on a clean pair of gloves for handling the dialysis tubing to prevent contamination.

Determining the Activation of Rho as an Index of Receptor

327

12. We use the Bio-Rad protein assay kit (#500-0006). We usually obtain ~1  mg/mL protein solution. If the protein solution concentration is lower than this, it can be concentrated by ultrafiltration. Ultrafiltration tubes are sold by several companies (Millipore, etc.). However, remember that some portion of the protein is inevitably lost during the ultrafiltration step. 13. It would be better to use 1-day-old rats delivered on the ­previous day for the experiment. 14. Extirpate a mouse heart as soon as possible. 15. The number of collected cells peaks at third or fourth collection. 16. Wash the plate briefly by the culture medium to collect the cardiomyocytes weakly attached on the plate. 17. Lysis buffer (700  mL) per 10  cm2 silicone dishes produce about 0.5 mg/mL protein solution. 18. Glutathione Sepharose beads are well equilibrated with lysis buffer (50  mM Tris/HCl, pH 7.5, 0.1% TritonX-100, 10% glycerol, 150 mM NaCl, 30 mM MgCl2, 1 mM dithiothreitol, 10 mg/mL leupeptin, 10 mg/mL aprotinin, 1 mM PMSF). 19. Scion Images and NIH Image are available free of charge. References 1. Northup, J.K., Smigel, M.D. and Gilman, A.G. (1982) The guanine nucleotide-activating site of the regulatory component of adenylate cyclase. J. Biol. Chem. 257, 11416–1423. 2. Benard, V., Bohl, B. P. and Bokoch, G. M., (1999) Characterization of Rac1 and Cdc42 activation in chemoattractant-stimulated human neutrophils using a novel assay for active GTPases. J. Biol. Chem. 274, 13198–13204. 3. Ren, X. D., Kiosses, W. B. and Schwartz, M. A. (1999) Regulation of the small GTP-binding protein Rho by cell adhesion and the cytoskeleton. EMBO J. 18, 578–585. 4. Yamaguchi, Y., Katoh, H. and Negishi, M. (2003) N-terminal short sequences of a-subunits of the G12 family determine selective coupling to receptors. J. Biol. Chem. 278, 14936–14939. 5. Nishida, M., Tanabe, S., Maruyama, Y., Mangmool, S., Urayama, K., Nagamatsu, Y., Takagahara, S., Turner, J. H., Kozasa, T., Kobayashi, H., Sato, Y., Kawanishi, T., Inoue, R., Nagao, T. and Kurose, H. (2005) Ga12/13- and

reactive oxygen species-dependent activation of c-jun NH2-terminal kinase and p38 mitogenactivated protein kinase by angiotensin receptor stimulation in rat neonatal cardiomyocytes. J. Biol. Chem. 280, 18434–18441. 6. Kurose, H. (2003) Ga12 and Ga13 as key regulatory mediator in signal transduction. Life Sci. 74, 155–161. 7. Worzfeld, T., Wettschureck, N. and Offermanns, S. (2008) G12/G13-mediated signalling in mammalian physiology and disease. Trends Pharmacol. Sci. 29, 582–589. 8. Suzuki, N., Hajick, N. and Kozasa, T. (2009) Regulation and physiological functions of G12/13mediated signaling pathways. Neurosignals 17, 55–70. 9. Nishida, M., Sato, Y., Uemura, A., Narita, Y., Tozaki-Saitoh, H., Nakaya, M., Ide, T., Suzuki, K., Inoue, K., Nagao, T. and Kurose, H. (2008) P2Y6 receptor-Ga12/13 signalling in cardiomyocytes triggers pressure overload-induced cardiac fibrosis. EMBO J. 27, 3104–3115.

wwwwwww

Chapter 18 The Use of Translocating Fluorescent Biosensors for Real-Time Monitoring of GPCR-Mediated Signaling Events Carl P. Nelson and R.A. John Challiss Abstract The ability to visualize the subcellular localization of proteins by labeling them with fluorescent proteins is a powerful tool in cell biology. In the G protein-coupled receptor signaling field, this technique has been utilized to examine the various aspects of receptor behavior, including activation, internalization and recycling, as well as alterations in the cellular levels of a variety of second messengers and signaling intermediates. Attaching variants of green fluorescent protein on to protein modules, which possess high affinity and selectivity for specific signaling molecules has allowed the visualization of key signaling pathway intermediates in real time, in living cells. This chapter outlines a protocol for the expression and visualization (by confocal microscopy) of such fluorescent “biosensors” and provides guidance on the analysis and interpretation of data obtained from such experiments. Key words: Fluorescent biosensor, Confocal microscopy, Cell signaling, G-protein-coupled receptor, Green fluorescent protein, Phospholipase C

1. Introduction The development of Ca2+-sensitive dyes revolutionized our appreciation of the diverse array of Ca2+ signals and their myriad roles in cellular physiology. In a similar manner, the discovery and exploitation of fluorescent proteins derived from the jellyfish Aequorea victoria has transformed our understanding of many areas of biological science, including a variety of aspects of G-protein-coupled receptor (GPCR)-mediated cell signaling. The combination of advances in imaging technologies, the expansion of the fluorescent protein color palette to span the whole visible spectrum and beyond (1) and the availability of molecular

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_18, © Springer Science+Business Media, LLC 2011

329

330

Nelson and Challiss

biological approaches to introduce these fluorescent proteins within specific cellular protein domains has facilitated the development of an ever-expanding range of fluorescent “biosensors.” By manipulating cells to express fluorescently labeled protein domains, which exhibit high affinity and selectivity for their targets, it has been possible to track the subcellular movements and/or changes in concentration of a variety of components within GPCR-regulated signaling pathways. Since changes in concentration [e.g., increases in the levels of inositol 1,4,5-trisphosphate (IP3) and diacylglycerol (DAG) and decreases in phosphatidylinositol 4,5-bisphosphate (PIP2) concentration on phospholipase C (PLC) activation] and localization [e.g., the translocation of protein kinase C (PKC) isoenzymes to the plasma membrane following PLC activation] of signaling intermediates are largely responsible for encoding the signals transduced through these pathways, the ability to monitor these events in real-time in single living cells has provided information previously unattainable using traditional biochemical methods (2–4). Biochemical assays use large numbers of cells (or whole tissues) to provide quantitative information on the mass changes in a given signaling intermediate, often with a limited temporal resolution. In contrast, fluorescent biosensor assays can provide realtime data in single cells at a subcellular level. They allow specific cell-types within heterogeneous populations to be identified and examined independently and can also facilitate the study of asynchronously responding populations (e.g., oscillatory signals or waves of signaling (2)). In addition, the availability of a range of colors of fluorophore [e.g., yellow and cyan fluorescent proteins (YFP and CFP)] has allowed multiple biosensors to be used within single cells to track two or more signaling intermediates at once (e.g., (5)) and has facilitated the development of fluorescence resonance energy transfer (FRET)-based biosensors for a variety of signaling molecules (see ref. 6 and Chapter 16). Although this methodology has been applied to a number of aspects of GPCR signaling (e.g., b-arrestin recruitment (7) and monomeric G protein translocation (8)), translocating fluorescent biosensor approaches have been particularly effectively applied to the study of the PLC signaling pathway (2–4). PLCb is typically activated by GPCRs coupled to the Gq/11 family of G-proteins and leads to the hydrolysis of the membrane phospholipid PIP2 into two second messengers: DAG, which subsequently recruits and activates PKC isoenzymes and IP3, which causes release of Ca2+ from intracellular stores. Green fluorescent protein (GFP)-based biosensors have been reported for DAG (9) and Ca2+ (10), while GFP-labeled PKC isoenzymes have been used to study directly PKC translocation (e.g., refs. 11, 12). PIP2 biosensors, based on the fluorescently tagged Tubby protein have also recently been described (13–15), but perhaps the most widely used fluorescent

The Use of Translocating Fluorescent Biosensors for Real-Time Monitoring

331

biosensor to date is based on the pleckstrin homology (PH) domain of PLCd1 (eGFP-PH (16, 17)). This biosensor exhibits high affinity and selectivity for PIP2, such that at rest the probe is localized to the plasma membrane, but on PLC stimulation eGFP–PH translocates to the cytosol (16, 17). Since eGFP–PH has a high affinity for both the substrate (PIP2) and one of the products (IP3) of PLC activation, the relative contributions of these two species to its localization has been debated (see refs. 2, 4, 15); as a tool for visualizing PLC regulation in real-time, in a variety of living systems, it has proven to be an invaluable technique. This chapter describes the method for investigating PLC signaling downstream of GPCR activation in a neuroblastoma cell line, using the eGFP–PH fluorescent biosensor; the technique can be easily adapted for use with any fluorescent-labeled biosensor in any cell system that is amenable to transfection. We also include details of how changes in intracellular [Ca2+] can be imaged simultaneously with fluorescent biosensor localization. It is impossible to write a definitive protocol, given the likely availability of distinct imaging hardware and software (as well as numerous potential combinations of cell systems and biosensors) in different laboratories. However, we hope that this protocol, along with the guidance provided in the Notes, will provide a basis for the development of a viable method in your system of choice. In addition, excellent recent articles by Palmer and Tsien (18) and Varnai and Balla (19) also address the use of fluorescent biosensors in real time imaging and may provide further useful resources.

2. Materials 2.1. Preparation and Transient Transfection of Cells

1. SH-SY5Y human neuroblastoma cells. 2. Minimum essential medium, 2  mM l-glutamine, 10% newborn calf serum supplemented (as required) with 100 U/mL penicillin, 100 mg/mL streptomycin, and 2.5 mg/mL amphotericin (all available from Invitrogen, Paisley, UK). 3. 25-mm glass coverslips [thickness 1.5 (160–190 mm), VWR, Lutterworth, UK]. 4. LipofectAMINE2000 (Invitrogen, Paisley, UK). 5. Minimum essential medium (Invitrogen, Paisley, UK), without supplements. 6. Mammalian expression plasmid containing the fluorescent biosensor of choice (in this example, the pleckstrin homology (PH) domain of PLCd1 in peGFP-C1 (16)), in maxiprep quantities.

332

Nelson and Challiss

2.2. Confocal Imaging of Fluorescent Biosensors

1. Krebs–Henseleit buffer (KHB): 118 mM NaCl, 4.7 mM KCl, 1.3 mM CaCl2, 1.2 mM KH2PO4, 1.2 mM MgSO4, 25 mM NaHCO3, 5 mM HEPES, and 10 mM glucose at pH 7.4. 2. Fura Red, AM (Invitrogen, Paisley, UK), 1  mM stock in DMSO. 3. Teflon coverslip dish (V.MSC-TD, Harvard Apparatus, Kent, UK). 4. Peltier unit (PDMI-2 open perfusion micro-incubator, controlled by TC 202A temperature controller, Harvard Apparatus, Kent, UK). 5. Peristaltic pump perfusion system (Gilson Minipuls peristaltic pump, Anachem, Luton, UK). 6. Olympus FV500 laser scanning confocal system (with three photomultiplier tube (PMT) detectors), with Olympus IX70 inverted microscope (Olympus Europa, Hamburg, Germany). 7. Olympus FluoView TIEMPO v.5.0 software. 8. 100  W Mercury lamp (U-ULH) and power supply (both from Olympus Europa, Hamburg, Germany). 9. 60× oil-immersion objective lens (PLAPO60xOI, NA 1.40, Olympus Europa, Hamburg, Germany). 10. 10-mW argon laser (458, 488, 514 nm). 11. 560-nm dichroic beam splitter, BA 505IF long-pass filter (for eGFP) and BA 660IF (for Fura Red). 12. Methacholine (MCh, Sigma, Poole, UK), 10 mM stock solution in KHB or appropriate ligands for receptor under study.

2.3. Analysis of Experimental Data

1. Olympus FluoView v.5.0 software. 2. Prism v.5 (GraphPad Software, San Diego, CA, USA) and Excel (Microsoft, Redmond, Washington, USA) analysis software.

3. Methods The following protocol details the process of transfecting and imaging a typical fluorescent biosensor (based on the pleckstrin homology domain of PLCd1 and used to measure PLC activity) expressed in SH-SY5Y human neuroblastoma cells, but similar approaches may be applied to most model cell-lines (e.g., HEK293, CHO, HeLa) and many primary cell lines. In addition, a wide variety of fluorescently labeled, translocating biosensors may be imaged in a similar manner, with only a minor adaptation of the following methodology (see Note 1).

The Use of Translocating Fluorescent Biosensors for Real-Time Monitoring

3.1. Preparation and Transient Transfection of Cells

333

1. Seed cells, 48–72  h prior to experimentation, onto sterile 25-mm glass coverslips (see Note 2) such that they are 40–60% confluent 24 h later. 2. Transfect cells, after 24 h, with 0.5 mg eGFP-PH cDNA per coverslip using LipofectAMINE2000, according to the manufacturer’s instructions. Typically, we transfect 0.5  mg cDNA with 1.5 mL of LipofectAMINE2000 (i.e., 1:3 ratio), diluted in 100 mL of serum-free minimum essential medium (see Note 3). 3. Remove medium, after 4–6  h, and replace with fresh, fully supplemented medium. 4. 24–48 h post-transfection cells are ready to be imaged (see Note 4).

3.2. Confocal Imaging of Fluorescent Biosensors

1. Carefully remove coverslip containing cells from the medium and wash with prewarmed KHB (1 mL). 2. To co-image changes in [Ca2+]i alongside eGFP–PH translocation, load cells with Fura Red, AM Ca2+-sensitive dye (3 mM, 1 h, room temperature), followed by washing in KHB (1 mL) (see Note 5). 3. Mount coverslip in a Teflon coverslip dish and carefully add prewarmed KHB (1 mL). 4. Maintain cells at 37°C in a Peltier unit and continually perfuse with KHB (5 mL/min) (see Note 6). 5. Use brightfield illumination to focus cells using a 60× oil immersion objective lens on a laser-scanning confocal inverted microscope (see Note 7). 6. Using fluorescence illumination, identify transfected cells (see Note 8) (and assess the success of Fura Red, AM loading where appropriate). Choose an appropriate field of view, ideally containing ³2 transfected cells (see Note 9). 7. Mark a region of interest (ROI) within the cytosol of each cell (see Note 10) and adjust detection settings such that the baseline cytosolic fluorescence is in the range of 200–300 fluorescence units. This should ensure that the system does not become saturated, even by a robust biosensor translocation (see Note 11). 8. Begin recording, typically every 2–3  s (depending on the speed of the response in question) with an intermediate scanning speed (on our system, this corresponds to ~1.7 s/scan) (see Note 12). 9. Where possible, drug additions (to stimulate or inhibit aspects of the PLC signaling cascade) should be made via the continual perfusion system. However, when using lipophilic compounds

334

Nelson and Challiss

(such as wortmannin, thapsigargin, etc.) it is necessary to apply these directly into the bath, so as to avoid contaminating the perfusion system for future experiments (see Note 13). Note the timings of the additions made and how long the cells were exposed to the drug. Stimulation of the endogenous M3 mACh receptor population in SH-SY5Y cells with MCh (100  mM), perfused onto the cells for 60 s, is adequate to observe a maximal, yet transient translocation of the eGFP–PH biosensor from plasma membrane to cytosol (see Fig. 1).

b

(i)

(ii)

Basal

(ii)

(i)

2000 1000 0

0

5

10

1000 0

Distance (µm)

Raw fluorescence data

0

5 10 Distance (µm)

6

1000

5

1000 0

15

0

5

10

15

Distance (µm)

4

750 500

3 2

250 0

2000

Normalised fluorescence

(ii)

1250

F/F0

Fluorescence

(i)

3000

2000

15

Wash

(iii)

3000

Fluorescence

Fluorescence

3000

c

(iii)

+ 100 µM MCh

Fluorescence

a

1 0

100

200

Time (s)

300

400

0

0

100

200

300

400

Time (s)

Fig. 1. Analysis of muscarinic acetylcholine receptor-mediated PLC activation in SH-SY5Y cells by eGFP–PH–PLCd1 translocation. (a) Confocal images of SH-SY5Y neuroblastoma cells transiently expressing eGFP–PH before stimulation (i), during stimulation with MCh (100 mM, 60 s) (ii) and following agonist washout (iii). (b) Fluorescence intensity histograms derived from line scans across a single SH-SY5Y cell [line indicated in (a)(i–iii)], under basal (i), MCh (100 mM) stimulation (ii) and washout (iii) conditions. Data are expressed as fluorescence intensity relative to distance (in micrometers) across the line (from left to right ). (c) Traces illustrating changes in fluorescence intensity within the region of interest (white circle) indicated on the images shown in (a), upon stimulation with MCh (100 mM) for the period indicated by the horizontal bar. (i) Raw mean fluorescence intensity values within the region of interest; (ii) normalized fluorescence values (relative to initial fluorescence) expressed as a self-ratio of F/F0, where F0 is the initial fluorescence intensity and F is the fluorescence intensity at a given time.

The Use of Translocating Fluorescent Biosensors for Real-Time Monitoring

3.3. Analysis of Experimental Data

335

1. Once the experiment is complete, it is important to ensure that the series of acquired images are saved to the hard drive. The experiment can then be revisited at any later point for further analysis. 2. It must then be decided how best to analyze the data. Playing back the experiment allows a visual assessment of any biosensor translocation and where there is a substantial movement of probe, showing still images or a movie of the experiment may provide a clear representation of the experimental result. However, to analyze the data quantitatively, there are at least two options available: (a) perform a line-scan across the image, to assess the fluorescence intensity across the cell; (b) measure the change in fluorescence in a region of interest (ROI) placed within the cytosol (see Note 14). 3. Positioning a line across the width of the image of the cell allows a line-scan to be performed, in which the imaging software measures the fluorescence intensity at each pixel across the image, producing histograms such as those illustrated in Fig. 1b, where the fluorescence intensity is expressed relative to the distance across the image. Obtaining equivalent histograms at various time points in the experiment provides a visual (and to some extent quantifiable) representation of the changes occurring in biosensor localization. For example, the clear localization of the biosensor to the plasma membrane is illustrated by the sharp peaks at either extreme of the histogram from the prestimulated image (Fig. 1b(i)), compared with the more diffuse fluorescence observed on MCh stimulation (Fig. 1b(ii)). 4. If adopting the ROI approach, it is advisable to play the experiment back to determine whether the regions of interest marked out prior to the experiment are suitable for analysis and if necessary redraw/reposition them (see Note 10). Most software programs allow the mean and/or cumulative fluorescence intensity values corresponding to the ROI(s) to be exported into Microsoft Excel, GraphPad Prism, or other suitable data analysis packages. Data can then be plotted as raw fluorescence values (Fig.  1c(i)) or normalized fluorescence values (Fig. 1c(ii)) against time, to provide an indication of the timecourse and extent of biosensor translocation. 5. Further quantitative analysis can be performed when measuring changes in cytosolic fluorescence by ROI. Measuring the peak translocation across a range of cells (where responses are all normalized to initial fluorescence levels to yield a self-ratio) provides an index of the extent of biosensor translocation occurring in response to a given stimulus. Assessment of the concentration dependency of a response may therefore be performed by determining the normalized peak responses to a range of agonist concentrations (see Fig. 2). In this way, fluorescent biosensors

basal

Normalised fluorescence (F/F0)

a

Nelson and Challiss

3 µM

10 µM

30 µM

30 µM

3

100 µM

wash

b

100 µM

10 µM 2

3 µM

1

0 0

100

200

300

400

500

Time (s)

600

700

800

Normalised fluorescence (F/F0)

336

2.5

2.0

1.5

1.0 −8

−7

−6

−5

−4

−3

−2

[MCh] (log M)

Fig. 2. Measuring concentration-dependent eGFP–PH–PLCd1 translocation in SH-SY5Y cells, in response to muscarinic acetylcholine receptor activation. (a) Representative images and trace, illustrating graded changes in cytosolic eGFP–PH–PLCd1 fluorescence (taken from a region of interest within the cytosol) in SH-SY5Y cells following treatment with the indicated concentrations of MCh (3–100 mM; 60 s each). (b) Concentration-response curve for eGFP–PH translocation in response to MCh in SH-SY5Y cells (n ³ 8). Normalized data are expressed as a self-ratio of F/F0, where F is the cytosolic fluorescence intensity and F0 is the initial fluorescence intensity.

can provide quantitative measures, such as agonist EC50 values (e.g., Fig. 2b; (5, 15, 20–22)), as well as kinetic data relating to rates of onset and offset of responses (5, 12, 15, 22, 23) (see Note 15).

4. Notes 1. A variety of fluorescent biosensors, particularly those targeting phosphoinositide signaling pathways, are based on the translocation of fluorescent-labeled modules between cytosolic and membrane compartments on stimulation (3, 4, 24). eGFP-PH is an example of such a probe, which translocates from the plasma membrane to cytosol on PLC activation (16, 17) and this protocol describes how to image and then analyze these translocation events. However, similar techniques may also be applied to study biosensors translocating from cytosol to plasma membrane (e.g., the DAG sensor, eGFP-PKCg-C12 (9)). Experiments may be performed identically to those described in this chapter, including the analysis techniques (line-scan and ROI), with the understanding that an increase in the target molecule (at the membrane) will be represented by a decrease in cytosolic fluorescence signal.

The Use of Translocating Fluorescent Biosensors for Real-Time Monitoring

337

2. The most efficient means of sterilizing commercially available coverslips is by autoclaving (at 121°C for 15 min). For certain cell types (e.g., HEK293) that are relatively poorly adherent on glass surfaces, it may be necessary to coat the coverslips with poly-l-lysine prior to seeding cells. In such cases, in a category II cabinet, incubate the sterilized coverslips in 150–200 mL of poly-l-lysine (50 mg/mL in sterile ddH2O) for 30 min, then remove and air dry before seeding cells onto the coverslips. 3. (a)   It is necessary to optimize the transfection conditions for each cell-type/biosensor combination. We find that 0.5  mg cDNA is optimal in most cases as this yields a good level of expression within transfected cells, while minimizing the amount of cDNA/lipofection reagent used per transfection (an important consideration, since all lipofection reagents are cytotoxic to at least some extent). However, we recommend using this as a ­starting-point from which to perform an optimization in the chosen experimental system. (b) We routinely use LipofectAMINE 2000 (Invitrogen, Paisley, UK) as our lipofection reagent of choice, but have previously used FuGENE (Roche Diagnostics, Burgess Hill, UK) and GeneJuice (Merck Biosciences, Nottingham, UK) to obtain a similar effect. The cDNA:lipofection reagent ratio is another aspect of the transfection, which may require optimization, as ratios of between 1:2 and 1:4 are widely used. Once again, the choice of ratio represents a balance between transfection efficiency and potential cytotoxicity. (c) Performing the transfection (the 4–6 h period in which the cells are exposed to lipofection reagent) in the absence of antibiotics may further improve cell viability, as cells are known to be susceptible to antibiotic damage during the transfection process. (d) The transfection efficiency obtained with standard lipofection reagents varies widely, depending on the cell type being transfected. In immortalized cell-lines, such as HEK293 and CHO cells, efficiencies of up to and exceeding 50% can be achieved, but for primary cell cultures (including vascular smooth muscle cells and hippocampal neurons), the efficiency may be as low as 1–2%. Although greater transfection efficiencies (>70%) can be achieved (even in primary cell cultures) using nucleofection (Lonza, Cologne, Germany) or similar techniques, it is not necessary to go to such lengths for fluorescent biosensor imaging experiments, where imaging one or two cells at a time can be achieved with a relatively low transfection efficiency.

338

Nelson and Challiss

4. In our experience, 24 h is usually sufficient for expression and correct trafficking of the fluorescent biosensors we have used. In addition, we have observed no differences in responses obtained at either 24 or 48 h post transfection. It is therefore possible to image at either time point, but since many of these biosensors have a potential to cause adaptive changes and/or interfere with endogenous signaling cascades within cells, in may be advantageous to perform experiments at the earliest possible time point. 5. Fura Red is a single excitation/single emission Ca2+-sensitive dye, which on excitation emits light at the red end of the spectrum. In its Ca2+-bound form, the red emission is diminished, such that an increase in [Ca2+]i is detected as a decrease in signal. The dynamic range of the dye is generally not as good as for more commonly used Ca2+-sensitive dyes, such as Fura-2 or Fluo-4, but if dye-loading is optimized it can provide a good indication of relative changes in [Ca2+]i occurring alongside alterations in IP3/DAG/PIP2 levels. The choice of Ca2+-sensitive dye is largely determined by the properties of the fluorescent label to be coimaged. We typically use Fura-Red alongside GFP-based biosensors as it ensures good separation between excitation and emission fluorescence spectra. If using a red fluorescent label, it is possible to use more robust Ca2+-sensitive dyes such as Fluo-4, whose excitation/emission spectra overlap those of GFP. Excitation and emission maxima are usually provided by suppliers of fluorescent dyes and fluorophores, while complete excitation and emission spectra are freely available on the Invitrogen website (http://www.invitrogen.com/site/us/en/home/ support/Research-Tools/Fluorescence-SpectraViewer. html). Loading protocols and conditions may vary depending on cell type and dye of choice, so these should be optimized within your system of choice (see also Chapter 15 for a more comprehensive appraisal of Ca2+ imaging techniques). 6. It is possible to perform experiments at lower temperatures, as many of the signaling cascades measured by fluorescent biosensors still occur at room temperature (20–22°C). However, to work at more physiological temperatures, it is advisable to preheat the KHB in a waterbath to ensure a constant flow of warmed buffer over the cells, as perfusion setups often involve lengths of tubing where buffer temperature can fall. It may also be necessary to check the actual bath temperature independently of the reading on the temperature controller, as in our experience this may differ by 2–3°C. A bath temperature of 37°C can therefore be attained by setting the temperature controller to 39–40°C. 7. Although a standard wide-field fluorescence microscope is capable of visualizing fluorescent protein expression within

The Use of Translocating Fluorescent Biosensors for Real-Time Monitoring

339

cells, it is not suitable for imaging changes in the subcellular localization of fluorescent biosensors. The image seen through a conventional microscope represents the sum of the in-focus region (determined by the depth of field of the objective lens, typically around 1 mm) plus the light emitted from the out-offocus regions, such that even in a relatively small cell with a depth of 5 mm, around 80% of the light making up the image may be out of focus (25). Where it is necessary to visualize the localization of a fluorescent protein to (for instance) the plasma membrane, this amount of out-of-focus light yields a very blurry and undefined image. Confocal microscopy overcomes this problem by eliminating the out-of-focus light using a pinhole aperture to restrict the light paths entering the detector to only those coming from the in-focus region. This greatly enhances the contrast and therefore the visibility of finer structures within the specimen and effectively provides an “optical section” through the cell, such as those shown in Fig. 1a. This degree of resolution allows the clear visualization of biosensor translocation between membrane and cytosolic compartments that is required to work with these fluorescent probes. 8. The choice of cells for the experiment is one of the most crucial steps in the process. Under fluorescence illumination, it is possible to scan around the coverslip to ascertain how many cells have been transfected and what range of expression levels are observed within those cells (it may also be useful to look at the cells under a lower magnification (10 or 20×) to observe a wider field). It is advisable to choose cells which are bright enough to image without stretching the system to its limits and certainly without having to increase the laser power beyond 2–3% (in the case of a 10 mW argon laser), as this may lead to significant photobleaching over the time course of experimental observation. However, very highly expressing cells are also to be avoided, as high levels of expression of any recombinant protein may adversely affect cellular function and many fluorescent biosensors bind to their protein/lipid targets at rest (e.g., eGFP-PH binds to PIP2), thereby potentially interfering with their normal functions (15, 17). Ideally, cells showing an intermediate level of expression should therefore be selected for the experiment. It is also important to ensure that the localization of the fluorescent biosensor at rest is consistent with what would be expected. In the case of eGFP-PH, it is important to select cells in which the probe is enriched at the plasma membrane (consistent with its binding to PIP2). If a significant proportion of transfected cells display an abnormal localization, it is advisable to adjust the transfection conditions to optimize cell viability (see Note 3). In general, it is preferable to avoid imaging cells which exhibit an abnormal morphology. This may be verified by comparison (under bright-field illumination) with

340

Nelson and Challiss

­ ntransfected cells elsewhere on the coverslip and with prior u knowledge of the “normal” morphology for the cell type. 9. Obviously, having multiple transfected cells in the field of view increases the amount of data obtained from a single experiment. If coimaging changes in [Ca2+]i alongside biosensor translocation, it may be useful to include a non-transfected (but Ca2+-sensitive dye-loaded) cell in the field of view, as this can provide an internal control for the potentially adverse influence of the recombinantly expressed biosensor on the Ca2+ signaling pathway. These cells can also give an indication of any autofluorescence and/or “bleed-through” between the red and green channels (potentially allowing correction for any artifactual fluorescent signal). 10. Many software programs [including Fluoview (Olympus) and Metafluor (Molecular Devices)] provide an online readout of the average fluorescence intensity within the selected region(s) of interest during the course of an experiment. Even if you ultimately intend to use an alternative analysis method, it is often helpful to get some idea of the response(s) occurring in real time, so it may be worth marking a ROI prior to the experiment to allow this. Larger ROIs will provide a better signal-to-noise ratio, but this must be balanced against the fact that using larger ROIs will increase the chance of a noncytosolic area (plasma membrane, nucleus, or even the extracellular space or another cell) contaminating the region of interest at some point during the experiment. If the cells being imaged remain relatively static during stimulation, larger ROIs may be used, but if the cells are motile (plus or minus stimulus), it is advisable to minimize the size of the ROIs. Other factors, such as the amount of cytosol in view will obviously affect the ease with which a cytosolic ROI may be kept free from artifacts due to cell movement. 11. In our experience, most plasma membrane-to-cytosol biosensor translocations yield a change in cytosolic fluorescence of less than tenfold. To avoid problems with saturation, it is therefore necessary to ensure that the initial cytosolic fluorescence should be greater than background fluorescence levels, but £10% of the maximum detectable fluorescence level. Using a 12-bit system, with a maximum of 4,096 fluorescence intensity units, this requires a basal cytosolic fluorescence of £400 units. 12. The rate of sampling will to some extent be set by the kinetics of the response in question. eGFP-PH translocation in response to receptor-driven PLC activation usually occurs with a time course of tens of seconds (15, 17, 20), so a sampling rate of one image every 2–3 s is adequate. If performing very long experiments, or when working at particularly high ­resolution, it may be necessary to further reduce the sampling

The Use of Translocating Fluorescent Biosensors for Real-Time Monitoring

341

rate, in order to economize on file storage. When observing more rapid translocations, for instance, Ca2+-driven PKC translocations (11, 12, 21), it may be desirable to acquire images at a faster rate (perhaps every 1–1.5 s). The choice of scanning speed (the time of exposure to excitation on each scan) represents a balance between resolution (better resolution obtained with slower scanning speeds) and both the greater potential for photobleaching and the limitation on sampling rate, which occurs with longer scan speeds. For fluorophores that are particularly sensitive to photobleaching (such as YFP), or for experiments requiring faster rates of sampling, some resolution may have to be sacrificed by reducing the scanning speed. 13. For real-time, live cell imaging, it is preferable to constantly perfuse cells to avoid accumulation of waste products/locally released mediators, which may compromise cell viability and the interpretation of the observed responses. Perfusion simplifies drug additions as it only requires the user to change over the input solutions and also allows drugs to be washed out of the system after a predetermined time. When setting up the perfusion system, minimize the dead space as much as possible, practically, and then note the time it takes for solutions to pass through this dead space. This time can then be factored into the analysis of experimental data so that the time at which the drug reaches the bath can be accurately recorded. It is important to keep the perfusion lines clean (thorough washing out with 70% ethanol and H2O at the end of each day should be sufficient); if the lipophilicity of a compound is doubtful, do not run it through the perfusion system. Instead, apply lipophilic compounds (usually as a small volume of concentrated stock) directly into a static bath (of known volume), taking care to minimize addition artifacts resulting from the movement of the stage/coverslip dish. Be sure to clean the coverslip dish and inflow/outflow tubes (with 70% ethanol and H2O) between nonperfusing experiments to ensure that there is no cross-contamination by lipophilic compounds. 14. Analysis of biosensor responses by line scan provides slightly different information from analysis by ROI and in many cases, particularly when first working with a new biosensor, it may be advantageous to apply both forms of analysis. Line-scans provide information about the localization of the probe, since they profile the expression at both membrane and cytosol, while cytosolic ROIs by definition only profile changes in cytosolic fluorescence. However, ROIs are probably more suited to the analysis of time course experiments, since they can profile how the cytosolic fluorescence changes relative to time across the whole experiment. In contrast, line scans only provide a “snapshot” of the experiment. Line scans are less

342

Nelson and Challiss

sensitive to cell movement than ROIs, and are particularly more suitable for motile cells. However, provided that the biosensor does not translocate to a particular subcellular site outside of the confocal plane in view, both techniques report essentially the same information and should therefore in most cases provide corresponding results. 15. In many cases, fluorescent biosensors for particular signaling intermediates (e.g., PIP2, IP3, DAG) have been used to determine quantitative estimates of agonist-potency and kinetic profiles of signaling at the single cell level. Where these estimates have been compared with those determined by traditional biochemical means within the same systems (e.g., refs. 17, 20, 21, 26), good agreement has generally been reported between the two contrasting techniques. However, this may not always hold true and it is certainly important to be aware of the selectivity of a fluorescent biosensor for its supposed cellular target, relative to any other potential binding partners. A particularly good case in point is that of the eGFP–PH itself, which can bind to both PIP2 and IP3, indicating that PLC activitydependent translocations of this biosensor may not necessarily report changes in either PIP2 or IP3, but perhaps a composite of the two (see refs. 4, 15, 27). References 1. Tsien, R. Y. (2009) Constructing and exploiting the fluorescent protein paintbox (Nobel Lecture). Angew. Chem. Int. Ed. Engl. 48, 5612–5626. 2. Nahorski, S. R., Young, K. W., Challiss, R. A. J., and Nash, M. S. (2003) Visualizing phosphoinositide signalling in single neurons gets a green light. Trends Neurosci. 26, 444–452. 3. Halet, G. (2005) Imaging phosphoinositide dynamics using GFP-tagged protein domains. Biol. Cell 97, 501–518. 4. Varnai, P., and Balla, T. (2006) Live cell imaging of phosphoinositide dynamics with fluorescent protein domains. Biochim. Biophys. Acta. 1761, 957–967. 5. Horowitz, L. F., Hirdes, W., Suh, B. C., Hilgemann, D. W., Mackie, K., and Hille, B. (2005) Phospholipase C in living cells: activation, inhibition, Ca2+ requirement, and regulation of M current. J. Gen. Physiol. 126, 243–262. 6. Lohse, M. J., Bunemann, M., Hoffmann, C., Vilardaga, J. P., and Nikolaev, V. O. (2007) Monitoring receptor signaling by intramole­ cular FRET. Curr. Opin. Pharmacol. 7, 547–553. 7. Barak, L. S., Ferguson, S. S., Zhang, J., and Caron, M. G. (1997) A b-arrestin/green

­ uorescent protein biosensor for detecting G fl protein-coupled receptor activation. J. Biol. Chem. 272, 27497–27500. 8. Michaelson, D., Silletti, J., Murphy, G., D’Eustachio, P., Rush, M., and Philips, M. R. (2001) Differential localization of Rho GTPases in live cells: regulation by hypervariable regions and RhoGDI binding. J. Cell Biol. 152, 111–126. 9. Oancea, E., Teruel, M. N., Quest, A. F., and Meyer, T. (1998) Green fluorescent protein (GFP)-tagged cysteine-rich domains from protein kinase C as fluorescent indicators for diacylglycerol signaling in living cells. J. Cell Biol. 140, 485–498. 10. Oancea, E., and Meyer, T. (1998) Protein kinase C as a molecular machine for decoding calcium and diacylglycerol signals. Cell 95, 307–318. 11. Babwah, A. V., Dale, L. B., and Ferguson, S. S. (2003) Protein kinase C isoform-specific differences in the spatial-temporal regulation and decoding of metabotropic glutamate receptor1a-stimulated second messenger responses. J. Biol. Chem. 278, 5419–5426. 12. Nelson, C. P., Willets, J. M., Davies, N. W., Challiss, R. A. J., and Standen, N. B. (2008)

The Use of Translocating Fluorescent Biosensors for Real-Time Monitoring Visualizing the temporal effects of vasoconstrictors on PKC translocation and Ca2+ signaling in single resistance arterial smooth muscle cells. Am. J. Physiol. Cell Physiol. 295, C1590–C1601. 13. Hughes, S., Marsh, S. J., Tinker, A., and Brown, D. A. (2007) PIP2-dependent inhibition of M-type (Kv7.2/7.3) potassium channels: direct on-line assessment of PIP2 depletion by Gq-coupled receptors in single living neurons. Pflugers Arch. 455, 115–124. 14. Quinn, K. V., Behe, P., and Tinker, A. (2008) Monitoring changes in membrane phosphatidylinositol 4,5-bisphosphate in living cells using a domain from the transcription factor tubby. J. Physiol. 586, 2855–2871. 15. Nelson, C. P., Nahorski, S. R., and Challiss, R. A. J. (2008) Temporal profiling of changes in phosphatidylinositol 4,5-bisphosphate, inositol 1,4,5-trisphosphate and diacylglycerol allows comprehensive analysis of phospholipase C-initiated signalling in single neurons. J. Neurochem. 107, 602–615. 16. Stauffer, T. P., Ahn, S., and Meyer, T. (1998) Receptor-induced transient reduction in plasma membrane PtdIns(4,5)P2 concentration monitored in living cells. Curr. Biol. 8, 343–346. 17. Varnai, P., and Balla, T. (1998) Visualization of phosphoinositides that bind pleckstrin homology domains: calcium- and agonistinduced dynamic changes and relationship to myo-[3H)inositol-labeled phosphoinositide pools. J. Cell Biol. 143, 501–510. 18. Palmer, A. E., and Tsien, R. Y. (2006) Measuring calcium signaling using genetically targetable fluorescent indicators. Nat. Protoc. 1, 1057–1065. 19. Varnai, P., and Balla, T. (2008) Live cell imaging of phosphoinositides with expressed inositide binding protein domains. Methods 46, 167–176.

343

20. Nash, M. S., Young, K. W., Willars, G. B., Challiss, R. A. J., and Nahorski, S. R. (2001) Single-cell imaging of graded Ins(1,4,5)P3 production following G-protein-coupled-receptor activation. Biochem. J. 356, 137–142. 21. Bartlett, P. J., Young, K. W., Nahorski, S. R., and Challiss, R. A. J. (2005) Single cell analysis and temporal profiling of agonist-mediated inositol 1,4,5-trisphosphate, Ca2+, diacylglycerol, and protein kinase C signaling using fluorescent biosensors. J. Biol. Chem. 280, 21837–21846. 22. Jensen, J. B., Lyssand, J. S., Hague, C., and Hille, B. (2009) Fluorescence changes reveal kinetic steps of muscarinic receptor-mediated modulation of phosphoinositides and Kv7.2/7.3 K+ channels. J. Gen. Physiol. 133, 347–359. 23. Lenz, J. C., Reusch, H. P., Albrecht, N., Schultz, G., and Schaefer, M. (2002) Ca2+controlled competitive diacylglycerol binding of protein kinase C isoenzymes in living cells. J. Cell Biol. 159, 291–302. 24. Zhang, J., Campbell, R. E., Ting, A. Y., and Tsien, R. Y. (2002) Creating new fluorescent probes for cell biology. Nat. Rev. Mol. Cell Biol. 3, 906–918. 25. Murray, J. M. (2006) Confocal microscopy, deconvolution and structured illumination methods. In Basic Methods in Microscopy (Goldman, D. L. S. a. R. D., ed) pp. 43–81, Cold Spring Harbor Laboratory Press. 26. Balla, A., Kim, Y. J., Varnai, P., Szentpetery, Z., Knight, Z., Shokat, K. M., and Balla, T. (2008) Maintenance of hormone-sensitive phosphoinositide pools in the plasma membrane requires phosphatidylinositol 4-kinase IIIalpha. Mol. Biol. Cell 19, 711–721. 27. Xu, C., Watras, J., and Loew, L. M. (2003) Kinetic analysis of receptor-activated phosphoinositide turnover. J. Cell Biol. 161, 779–791.

wwwwwww

Part IV Receptor–Receptor and Receptor–Protein Interactions

wwwwwww

Chapter 19 Study of GPCR–Protein Interactions Using Gel Overlay Assays and Glutathione-S-Transferase-Fusion Protein Pull-Downs Ashley E. Brady, Yunjia Chen, Lee E. Limbird, and Qin Wang Abstract Numerous recent studies have suggested that the predicted cytosolic domains of G protein-coupled receptors represent a surface for association with proteins that may serve multiple roles in receptor localization, turnover, and signaling beyond the well-characterized interactions of these receptors with heterotrimeric G proteins. This Chapter describes two in vitro methods for ascertaining interactions between G protein-coupled receptors and various binding partners: gel overlay strategies and GST-fusion protein pull-downs. Key words: Gel overlay, G protein-coupled receptor, Glutathione-S-transferase pull-down, Protein– protein interaction

1. Introduction Gel overlays have been used historically to examine proteins whose folding is re-achieved following sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), a property that is determined empirically. For example, a variety of protein–protein interactions have been identified previously using gel overlay strategies, including interacting proteins for calmodulin (1) and for cyclic AMPdependent protein kinase (called A-kinase anchoring proteins, or AKAPs) (2). These approaches have been adapted for studying interactions with G protein-coupled receptors (GPCRs) (3, 4). The overall approach for gel overlay studies is to separate proteins from a mixture (tissue or cell lysates, membrane or cytosolic fractions), using SDS-PAGE, renature (or not) the polyacrylamide gel, transfer to polyvinylidene fluoride (PVDF) or nitrocellulose

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_19, © Springer Science+Business Media, LLC 2011

347

348

Brady et al.

membranes, and probe with a radiolabeled ligand or protein, ­typically created via metabolic labeling in bacteria followed by purification, or by in  vitro translation, as we describe in this Chapter. The term “overlay” is a residual terminology, describing very early protocols in which the SDS-polyacrylamide gel was probed with radiolabeled ligands (including proteins) resuspended in agarose and “overlaid” (and thus immobilized) on the wet polyacrylamide gel. The main variable in current gel overlay protocols that likely needs to be optimized for each protein-interacting pair is the incubation of the SDS-polyacrylamide gel prior to electrophoretic transfer to PVDF or nitrocellulose membranes. Some investigators think that renaturation (e.g. in buffer containing urea or nondenaturing detergents) is essential before transfer; others consider that the renaturation incubation should occur after transfer to the filter, and some include no incubation step to renature proteins following SDS-PAGE. Naturally, optimization of these protocols depends on the proteins involved and the ­chemical and physical properties of their interacting domains. Glutathione-S-transferase (GST)-fusion proteins provide a powerful in  vitro tool for detecting direct interactions between GPCRs and other accessory or regulatory proteins (5–14). The principle of the methodology is that the water-soluble GPCR domain of interest (e.g. the third intracellular loop or C-terminal tail) for examining protein–protein interactions is fused to GST, a dimeric enzyme. Design of “spacer” sequences between the cDNA encoding GST and that encoding the GPCR domain appears routinely to permit unperturbed folding of the GST and the independent fused protein domain. Known quantities of GST fusion protein are incubated with a protein mixture of interest (e.g. cellular extract) or, alternatively, co-expressed in a target cell of interest, and the fused protein is allowed to bind to possible interacting proteins. In the case of testing for an interaction with a specific protein, the target protein can also be labeled via in vitro translation, as described in this Chapter. The incubations (or cells, after lysis) are then exposed to glutathione (GSH) conjugated to agarose. The GST enzyme binds its substrate, GSH, and proteins interacting with its fusion protein are similarly isolated. The appropriate control is to compare eluates of GSH-agarose bound to GST alone versus those bound to the GST-fusion protein.

2. Materials 2.1. Gel Overlay

1. TnT rabbit reticulocyte in vitro transcription and translation kit (Promega). 2. [35S]Methionine ([35S]Met, 1,000 Ci/mmol, at 10 mCi/mL).

Study of GPCR–Protein Interactions Using Gel Overlay Assays

349

3. RNasin ribonuclease inhibitor (Promega). 4. cDNA template. Illustrative example used herein: 3i loop of the a2-adrenergic receptor fused to Gen10, a methionine-rich viral coat protein (15), referred to here as G10-a2-3i loop. 5. Nuclease-free water. 6. 12% SDS-PAGE minigel for analysis of 35S-protein product. 7. Biological preparation of interest to separate on SDS-PAGE. 8. PVDF, such as Immobilon-P (0.45  mm) by Millipore (preferred), or nitrocellulose membranes. 9. Tris-buffered saline (TBS): 20 mM Tris–HCl, 137 mM NaCl, pH 7.6. 10. Blocking buffer: TBS containing 3% (v/v) Tween 20 and 5% (w/v) nonfat powdered milk. 11. Rinsing buffer: TBS containing 0.1% (v/v) Tween 20 and 5% (w/v) nonfat powdered milk. 12. Tris–glycine transfer buffer: 25 mM Tris base, 192 mM glycine, 20% methanol, pH 8.3. 2.2. GST-Fusion Protein Pull-Down

1. pGEX-2T, pGEX-4T or other expression vector encoding GST prepared for in-frame fusion with DNA sequence encoding the protein or domain of interest. 2. E. coli strain DH5a or BL21 (DE3), or any strain that allows for expression of the gene of interest as a GST-fusion protein. 3. Luria broth (LB) (1 L): 10 g of bacto tryptone, 5 g of Bacto yeast extract, 10 g of NaCl, pH 7.0; autoclave for 20 min. 4. 50–100 mg/mL of ampicillin (or other appropriate antibiotic for selection). 5. Isopropyl-b-d-thiogalactopyranoside (IPTG). 6. Tris–Triton (TT) buffer: 100  mM Tris–HCl, pH 8.0, 1% Triton X-100, 100  mM phenylmethylsulfonyl fluoride (PMSF), 1 mg/mL of soybean trypsin inhibitor (STI), 1 mg/ mL of leupeptin, and 10 U/mL of aprotinin. 7. 1 mg/mL of Lysozyme in TT buffer. 8. GSH-agarose (Pierce). 9. 333 mM NaCl in TT buffer. 10. 0.8 × 4 cm Poly-Prep column (BioRad). 11. 10 mM GSH (free acid) in TT buffer. 12. Dulbecco’s phosphate-buffered saline (PBS), sterile: 0.1 g of KCl, 0.1 g of KH2PO4, 4.0 g of NaCl, 1.08 g of Na2HPO4 in 500 mL of double-distilled water (ddH2O), autoclaved. 13. Tris–Triton binding (TTB) buffer (see Note 1): 50 mM Tris– HCl, pH 7.4, 0.05% Triton X-100, 10% glycerol, 0.01% bovine serum albumin (BSA), and 100 mM PMSF.

350

Brady et al.

14. BioRad protein assay reagent (cat. no. 500–0006). 15. 1× Laemmli sample buffer: 62.5  mM Tris–HCl, pH 6.8, 700 mM b-mercapto-ethanol, 2% SDS, and 10% glycerol. 16. EN3HANCE™ autoradiography enhancer (PerkinElmer). 17. Amicon stirred cell with a PM10 (>10,000  Mol  Wt) filter (Amicon).

3. Methods 3.1. Gel Overlay Assays 3.1.1. Generation of 35S-Labeled Protein for Probing Transferred Proteins (Example: 35S-3i Loop of the a2A-Adrenergic Receptor)

1. Mix the following in an autoclaved microfuge tube: 25 mL of TnT lysate, 1 mL of amino acid mix (1 mM, minus methionine), 2 mL of TnT reaction buffer, and 1 mL of TnT T7 RNA polymerase (all reagents provided in TnT kit). 2. Add 4 mL of [35S]Met (1,000 Ci/mmol, at 10 mCi/mL) and 1 mL of RNasin ribonuclease inhibitor (40 U/mL). 3. Add 1  mg of the appropriate cDNA template (presented as the circular plasmid DNA in a vector possessing the T7 RNA polymerase recognition sequence). 4. Adjust the volume to 50 mL with nuclease-free water. 5. Incubate the mixture for 90 min at 30°C. 6. Stop the reaction by placing the tube on ice. 7. Analyze and quantify the reaction product by separation of precisely 1 mL on a 12% SDS-PAGE minigel. 8. Dry the gel and visualize bands by autoradiography. 9. Cut the band corresponding to the estimated probe relative molecular mass (Mr) out of the dried gel and count it in scintillant in a b-counter. Determine c.p.m. of total product per microliter, which corresponds to c.p.m. of 35S-protein of the correct Mr.

3.1.2. Overlay Assay

1. Protein aliquots (2.5–3.0 mg/sample) of biological sample of interest are separated on an SDS-PAGE preparative gel (16-cm long and 1.5-mm thick, see Note 2) using 7.5–20% polyacrylamide gradients. 2. Transfer the resolved proteins to PVDF at 4°C by electrophoresis overnight at 30 V in Tris–glycine transfer buffer. 3. Cut the membrane into vertical strips (2–4 mm wide) for gel overlay and Western blot analysis. 4. Block nonspecific binding to PVDF membranes by incubation for at least 1 h in blocking buffer at room temperature. 5. Wash PVDF membranes for 30 min in rinsing buffer.

Study of GPCR–Protein Interactions Using Gel Overlay Assays

351

6. Incubate PVDF membrane strips with 300,000 c.p.m. of the appropriate amount of [35S]Met-labeled G10-a2-3i loop structure diluted in 1 mL rinsing buffer for 4 h (to overnight) at 4°C with constant rocking (see Note 3). 7. Wash membranes three times for 15  min each with rinsing buffer, twice for 10 min with TBS, and air-dry before autoradiography (see Note 4). 8. Expose the pressed filter to a PhosphorImager and/or X-ray film to permit band identification and quantification. 3.2. GST-Fusion Protein Pull-Downs

1. Subclone the gene of interest into a plasmid designed for GST-fusion protein expression (see Note 5).

3.2.1. Synthesis of GST-Fusion Proteins

2. Transform this plasmid into an E. coli expression strain, such as BL21 (DE3). 3. Use bacteria transformed with this plasmid to inoculate 25 mL of LB medium containing the appropriate antibiotic (e.g. ampicillin at 50–100  mg/mL) to select for the GSTfusion plasmid. 4. Incubate this culture overnight (16–18 h) in a shaking incubator (250 r.p.m., 37°C). 5. Add 25 mL of the overnight culture to 250 mL (1:10 dilution) of LB without antibiotic. 6. Grow bacteria in the shaker (250  r.p.m., 37°C) until the optical density (OD) measured at 600 nm = 0.6 (~1.5–2 h). 7. Add 250 mL of 1 M IPTG (final concentration 1 mM) to the bacterial culture to induce expression of the fusion protein. 8. Continue incubating for 3–6  h (determined empirically for optimal production of each GST-fusion protein) in the shaking incubator (250 r.p.m., 37°C). 9. Centrifuge the samples at 13,500 × g at 4°C for 15  min to pellet bacteria. 10. Discard the supernatant. 11. Resuspend the bacterial pellet in 20 mL of ice-cold TT buffer containing 1 mg/mL of lysozyme and transfer solution to a 50-mL centrifuge tube. 12. Sonicate the sample with a probe sonicator for a 30-s burst. Place the tube of bacterial preparation in an ice bath, and insert the probe into the bacterial suspension two-third from the surface. 13. Allow samples to cool down, and then sonicate for another 30-s burst. 14. Repeat step 13. 15. Centrifuge samples at 39,000 × g at 4°C for 15 min.

352

Brady et al.

16. Collect supernatant and save (GST-fusion protein is in this supernatant). 17. Remove 20 mL of the sample for preliminary analysis of the GST-fusion protein product on SDS-PAGE. 3.2.2. Purification of GST-Fusion Proteins

1. Hydrate 300 mg of GSH-agarose powder in 45 mL of ddH2O for 30 min with rotation at room temperature. 2. Centrifuge the tube at 1,000 × g for 5  min to pellet the hydrated agarose resin. 3. Aspirate the ddH2O, leaving approx. 2 mm above the GSHagarose. 4. Equilibrate agarose by adding 20 mL TT. 5. Centrifuge at 1,000 × g for 5 min to pellet agarose. 6. Aspirate, leaving approx. 2 mm of TT above the GSH-agarose. 7. Add enough TT to make a 1:1 agarose:TT slurry. 8. Add sample (up to 20 mL) to GSH-agarose and mix by inversion at room temperature (30 min) or 4°C (2 h). 9. Centrifuge at 1,000 × g for 5 min. 10. Remove sample, leaving approx. 2  mm of buffer above the GSH-agarose (retain the sample to assay for proteins not adsorbed to GST-fusion protein). 11. Wash the resin twice with 6 mL of TT. 12. Centrifuge and aspirate the sample, leaving approx. 2 mm of buffer above the GSH-agarose. 13. Wash the resin with 3 mL of 333 mM NaCl in TT. 14. Centrifuge and aspirate the wash as described in step 12. 15. Wash the resin with 6 mL of TT. 16. Centrifuge and aspirate the wash as described in step 12. 17. Add 2 mL of TT. 18. Transfer the mixture to a 0.8 × 4  cm Poly-Prep column (see Note 6). 19. Add 2 mL of TT to collect residual resin and transfer this to the column as well. 20. Aspirate the wash buffer from the surface of the settled resin using a 27-guage needle (see Note 7). 21. To elute the fusion protein, add 3 mL of 10 mM GSH in TT buffer to the resin in the Poly-Prep column. Rotate at room temperature for 10  min and collect the flow-through in a sterile tube. 22. Separate the free GSH from the fusion protein by dialysis or by several cycles of concentration-dilution in an Amicon stirred cell using sterile PBS or TTB buffer.

Study of GPCR–Protein Interactions Using Gel Overlay Assays

353

23. Resuspend GST-fusion protein to a final volume of approx. 2  mL and assay protein concentration using the BioRad ­protein assay reagent (according to the manufacturer’s instructions). 3.2.3. GST Pull-Down Assay (Illustrative Example of Saturation Binding of 35S-a2AR-3i Loop to GST-Spinophilin 151–444)

1. Add increasing amounts of GST (control) or GST-fusion ­protein (see Note 8) to 50  mL of GSH-agarose (1:1 slurry equilibrated with TTB buffer) in a total volume of 235 mL. 2. Incubate with rotation for 2 h at 4°C. 3. Add 15  mL of 35S-labeled a2AR-3i loop (31,000  c.p.m.; ­estimated as 2.3 × 10−11 M). 4. Rotate for 2–16 h at 4°C via inversion in a 1.5-mL microcentrifuge tube. 5. Collect the resin by microcentrifugation. 6. Wash resin three times with 1  mL of TTB by resuspension and recollection of the resin by centrifugation. 7. Determine the amount of the 35S-labeled a2AR-3i loop bound to GST versus GST-fusion protein by elution into 1× Laemmli sample buffer and separation of samples by SDS-PAGE (see Note 9). 8. Treat gels with EN3HANCE™ (PerkinElmer) according to the manufacturer’s instructions prior to drying the gel to facilitate detection of 35S-labeled bands by autoradiography. 9. Determine the quantity of bound protein by cutting the appropriate gel bands (e.g. corresponding to the 35S-labeled a2AR-3i loop) from the dried gel and counting in scintillant in a b-counter.

4. Notes 1. NaCl can be added to this buffer up to 200 mM. However, the actual NaCl concentration should be titrated for each ­specific protein interaction studied. Detection of the interaction between spinophilin and a2AR-3i loops was optimal in the absence of NaCl. 2. A preparative gel has two lanes. One narrow lane is at the far edge for running a molecular weight standard. The remainder of the gel consists of a wide lane that spans the entire length of the gel, so the sample is continuous over the entire membrane once transferred. Prestained molecular weight markers should be run in the narrow lane to permit estimation of the approximate molecular weights of the proteins identified by gel overlay analysis.

354

Brady et al.

3. Based on the concentration of methionine contributed to the [35SsMet-labeling reaction by the rabbit reticulocyte lysate (5 mM) and the specific activity of the [35S]Met radiolabel, we estimate that 300,000  c.p.m. of Gen10-a2AR-3i loop represents 5–10 pmol of probe. 4. The number of washes, temperature of wash, and duration of wash can be modified to enhance the signal and minimize background. 5. For example, the pGEX-4T vector from Promega. 6. Use a pipette tip with the end cut off so as not to damage the agarose. 7. The wash buffer can be thoroughly aspirated by briefly touching the surface of the resin with a 27-gauge needle attached to a vacuum source. The resin will not be disturbed, because of the small diameter of the needle. 8. A quantity of 0.00034–34 mg of the fusion protein represents 2.34 × 10−11 to 1.77 × 10−6 M of this particular fusion protein in the incubation. Examining binding over a range of GSTfusion protein concentrations allows examination of the saturability of binding to the GST-fusion protein. Specificity is evaluated by “competition” studies with various “competitors” to determine their ability to bind 35S-a2AR-3i loop, and thus diminish the quantity of radioligand available for interaction with the GST-fusion protein. To assess the relative affinity of these “competitors” with spinophilin for the 3i loop, spinophilin itself is included as one of the competitors. Incubate 5  mg of GST or GST-Sp151-444 with 20  mL of GSH-agarose slurry for 2  h at 4°C. Increasing amounts (0–16 mL) of in vitro-translated Sp151-444 or other “competitors” can be mixed with rabbit reticulocyte lysate to reach a total volume of 16 mL and are then added to each incubation together with 6,000 c.p.m. (estimated as 9.2×10−12 M) of 35S-labeled 3i loop with a total reaction volume at 120 mL. After 2 h of incubation at 4°C, resins are washed and eluted. The 35S-3i loop in the eluate is separated and quantified as described in Subheading  3.2.3, steps 7–9. In these experiments, GST-fusion protein (the “receptor”) is in excess of the 35 S-a2AR-3i loop (ligand) and thus the relative affinity of competitors can be compared, but true thermodynamic ­constants for the interaction cannot be obtained. 9. The percent acrylamide in the gel, with or without a gradient, is determined empirically for optimally resolving proteins in eluates; use 12% SDS-PAGE to identify 35S-labeled a2AR-3i loops.

Study of GPCR–Protein Interactions Using Gel Overlay Assays

355

References 1. Sarkar, D., Erlichman, J., and Rubin, C. S. (1984) Identification of a calmodulin-binding protein that co-purifies with the regulatory subunit of brain protein kinase II. J. Biol. Chem. 259, 9840–9846. 2. Lester, L. B., Coghlan, V. M., Nauert, B., and Scott, J. D. (1996) Cloning and characterization of a novel A-kinase anchoring protein. AKAP 220, association with testicular peroxisomes. J. Biol. Chem. 271, 9460–9465. 3. Prezeau, L., Richman, J. G, Edwards, S. W., and Limbird, L. E. (1999) The z isoform of 14-3-3 proteins interacts with the third intracellular loop of different a2-adrenergic ­receptor subtypes. J. Biol. Chem. 274, 13462–13469. 4. Diviani D., Lattion A., Abuin L., Staub O., and Cotecchia S. (2003) The adaptor complex 2 directly interacts with the a1b-adrenergic receptor and plays a role in receptor endo­cytosis. J Biol. Chem. 278, 19331–19340. 5. Richman, J. G., Brady, A. E., Wang, Q., Hensel, J. L., Colbran, R. J., and Limbird, L. E. (2001) Agonist-regulated Interaction between a 2-adrenergic receptors and spinophilin. J. Biol. Chem. 276, 15003–15008. 6. Becamel, C., Alonso, G., Galeotti, N., et  al. (2002) Synaptic multiprotein complexes associated with 5-HT2C receptors: a proteomic approach. EMBO J. 21, 2332–2342. 7. Hall, R. A., Ostedgaard, L. S., Fremont, R. T., et  al. (1998) A C-terminal motif found in the b2-adrenergic receptor, P2Y1 receptor and cystic fibrosis transmembrane conductance regulator determines binding to the Na+/H+ exchanger regulatory factor family of PDZ proteins. Proc. Natl. Acad. Sci. USA 95, 8496–8501. 8. Smith, F. D., Oxford, G. S., and Milgram, S. L. (1999) Association of the D2 dopamine

receptor third cytoplasmic loop with spinophilin, a protein phosphatase-1-interacting protein. J. Biol. Chem. 274, 19894–19900. 9. DeGraff, J. L., Gurevich, V. V., and Benovic, J. L. (2002) The third intracellular loop of a2adrenergic receptors determines subtype specificity of arrestin interaction. J. Biol. Chem. 277, 43247–43252. 10. Wang, Q. and Limbird, L. E. (2002) Regulated interactions of the a2A adrenergic receptor with spinophilin, 14-3-3z, and arrestin 3. J. Biol. Chem. 277, 50589–50596. 11. Macey, T.A., Liu, Y., Gurevich, V.V., Neve, K.A. (2005) Dopamine D1 receptor interaction with arrestin3 in neostriatal neurons. J. Neurochem. 93, 128–134. 12. Chen, C., Li, J.G., Chen, Y., Huang, P., Wang, Y., Liu-Chen, L.Y. (2006) GEC1 interacts with the kappa opioid receptor and enhances expression of the receptor. J. Biol. Chem. 281, 7983–7993. 13. Wang, X., Zeng, W., Kim, M.S., Allen, P.B., Greengard, P., Muallem, S. (2007) Spinophilin/ neurabin reciprocally regulate signaling intensity by G protein-coupled receptors. EMBO J. 26, 2768–2776. 14. Xu, J., Chen, Y., Lu, R., Cottingham, C., Jiao, K., Wang, Q. (2008) Protein kinase A phosphorylation of spinophilin modulates its interaction with the a2A-adrenergic receptor (AR) and alters temporal properties of a2A-AR internalization. J. Biol. Chem. 283, 14516–14523. 15. Shieh, B. H., Zhu, M. Y., Lee, J. K., Kelly, I. M., and Bahiraei, F. (1997) Association of INAD with NORPA is essential for controlled activation and deactivation of Drosophila phototransduction in  vivo. Proc. Natl. Acad. Sci. USA 94, 12682–12687.

wwwwwww

Chapter 20 Study of GPCR–Protein Interactions by BRET Martina Kocan and Kevin D.G. Pfleger Abstract Bioluminescence resonance energy transfer (BRET) has become an extremely valuable technology for the real-time study of protein–protein interactions in live cells. This technique is highly amenable to the monitoring of G protein-coupled receptor (GPCR)–protein interactions, especially involving scaffolding, regulatory and signaling proteins, such as b-arrestins, which are now known to have significant roles in addition to receptor desensitization. The BRET procedure utilizes heterologous coexpression of fusion proteins linking one protein of interest (e.g. a GPCR) to a bioluminescent donor enzyme, a variant of Renilla luciferase, and a second protein of interest (e.g. b-arrestin) to an acceptor fluorophore. If in close proximity, energy resulting from the rapid oxidation of a cell-permeable coelenterazine substrate by the donor will transfer to the acceptor, which in turn fluoresces at a longer characteristic wavelength. Therefore, the occurrence of such energy transfer implies that the proteins of interest fused to the donor and acceptor interact directly or as part of a complex. BRET detection can be carried out using scanning spectrometry or dual-filter luminometry. The latest improvements in BRET methodology have enabled live cell drug screening as well as monitoring of previously undetectable protein-protein complexes, including constitutive GPCR/b-arrestin interactions. Therefore, BRET is likely to play an increasingly important role in GPCR research and drug discovery over the coming years. Key words: Bioluminescence resonance energy transfer, G protein-coupled receptor, Arrestin, Renilla luciferase8, Rluc8, Fluorophore, Venus

1. Introduction Since the original bioluminescence resonance energy transfer (BRET) technology was described there have been significant improvements in signal sensitivity and stability, enabling this method to be utilized for detection of weak and/or transient G protein-coupled receptor (GPCR)–protein interactions (1–3). Furthermore, the potential applicability of this technology for effective drug discovery high-throughput screening has been demonstrated (3). On oxidation of a coelenterazine substrate,

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_20, © Springer Science+Business Media, LLC 2011

357

358

Kocan and Pfleger

BRET occurs in a nonradiative manner between a bioluminescent donor enzyme, typically a variant of Renilla luciferase (Rluc), and a complementary acceptor fluorophore (1–5). The BRET1 system utilizes the luciferase substrate coelenterazine h (benzylcoelenterazine), with consequent light emission peaking at ~480  nm. Light emission from the corresponding acceptor is then observed at longer wavelengths following energy transfer (e.g. peaking at ~510 nm for enhanced green fluorescent protein (EGFP) or ~530 nm for the yellow fluorescent proteins (YFPs), Venus, or YPet). The extended BRET (eBRET) method was established in order to prolong the detection period. In contrast to BRET1, eBRET uses a long-acting form of coelenterazine h, called EnduRen, with identical spectral properties. However, unlike coelenterazine h, its signal can last for several hours (6). The BRET2 version of the technique utilizes the luciferase ­substrate coelenterazine 400a (also known as DeepBlueC, bisdeoxycoelenterazine or di-dehydro-coelenterazine), with consequent light emission peaking at about 400–420 nm. GFP10 or GFP2 is used as acceptor, with light emission peaking at ~510 nm following energy transfer (1–5). The advantage of BRET2 is the very large separation of donor and acceptor emission spectra, thus providing greater signal resolution. However, signals with BRET2 are weaker than with BRET1 or eBRET, substantially so when using standard Rluc (3). The mutant forms of Rluc, Rluc2 (C124A/M185V) and Rluc8 (A55T/C124A/S130A/K136R/A143M/M185V/ M253L/S287L) (7), greatly improve BRET2 and are strongly recommended for this version of the technique, as indeed they are for all derivations (3). BRET3 has been described very recently, utilizing a combination of Rluc8 and mOrange (a variant of Discosoma spp. red fluorescent protein) (8). This has the advantage of greater spectral resolution than BRET1 or eBRET, while still being able to use coelenterazine h or EnduRen (in contrast to coelenterazine 400a). The mOrange, with light emission peaking at ~565  nm, certainly has advantages for live animal imaging due to the greater tissue penetration of the red-shifted light (8). In terms of live cell applications where this is less of a consideration, a direct comparison between the Rluc8/Venus (3, 9), Rluc8/YPet (10) and Rluc8/mOrange (8) combinations will be required to establish the optimal BRET pairing. Given that the greater spectral resolution with mOrange is likely to be accompanied by less efficient energy transfer, it is not obvious whether a yellow or orange acceptor will be superior in practice, and relative use in the future may well come down to particular applications and personal preferences. BRET enables real-time detection of protein–protein interactions in live cells following genetic fusion of donor or acceptor molecules with proteins of interest (1, 2, 11). This technology has become very popular for monitoring GPCR complexes (12), particularly involving b-arrestins (3, 5, 6, 10, 13, 14). b-arrestins

Study of GPCR–Protein Interactions by BRET

359

regulate a large number of GPCRs and in addition to their role in receptor-G protein uncoupling (desensitization) and receptor internalization, there is increasing evidence for their involvement in the initiation of alternative (G protein-independent) signaling pathways that still remain to be fully elucidated (15, 16). The ability to detect GPCR–protein interactions, such as GPCR/ b-arrestin interactions stimulated by diverse ligands under different conditions, is critically important for better understanding the fundamentals of GPCR signaling, as well as drug discovery.

2. Materials 2.1. Generation and Validation of Fusion Constructs

1. cDNA constructs for complementary BRET donor and acceptor in appropriate expression systems (e.g. pcDNA3.1 or Gateway vectors from Invitrogen). Donor for all BRET derivations, Rluc variants, such as Rluc8 or Rluc2 (from S. S. Gambhir, Stanford University, CA (7)). Acceptor for BRET1 or eBRET, GFP variant such as Venus (from A. Miyawaki, RIKEN, Japan (17)). Acceptor for BRET2, GFP variant such as GFP10 (from Michel Bouvier, Department of Biochemistry, Université de Montréal, Canada (18)) or GFP2 (PerkinElmer). Acceptor for BRET3, mOrange (from R. Y. Tsien, University of California, San Diego, CA (19)). 2. cDNA constructs for proteins of interest (e.g. GPCR and b-arrestin). 3. Instrumentation and reagents for corresponding validation techniques required for determination of proper function and trafficking of fusion proteins (e.g. signaling assays, confocal microscope, fluorimeter).

2.2. Cell Culture

1. Cell line for transfection (e.g. COS7 or HEK293 cells). 2. Cell culture medium (e.g. Dulbecco’s modified Eagle’s medium (DMEM; from Gibco) containing 0.3  mg/mL glutamine (Gibco), 100  IU/mL penicillin, 100  mg/mL streptomycin (Gibco), and 10% fetal calf serum (Gibco)). 3. DMEM without phenol red (Gibco) containing 0.3 mg/mL glutamine (Gibco), 100  IU/mL penicillin, 100  mg/mL streptomycin (Gibco), 10% fetal calf serum (Gibco), and 25 mM HEPES (Sigma). 4. Transfection reagent (e.g. Genejuice Lipofectamine 2000 (Invitrogen)).

(Novagen)

or

5. 0.05% Trypsin; 0.53  mM ethylenediamine tetraacetic acid (EDTA) (Gibco). 6. Cell culture plates, 6-well clear plates (BD Falcon) and 96-well white plates (Nunc).

360

Kocan and Pfleger

2.3. BRET Signal Detection

1. Coelenterazine luciferase substrate: for BRET1 (or BRET3), 5 mM coelenterazine h (benzyl-coelenterazine) dissolved in methanol (Invitrogen); for eBRET (or BRET3), 60  mM EnduRen dissolved in dimethylsulfoxide (Promega); for BRET2, 1  mM coelenterazine 400a (DeepBlueC, bisdeoxycoelenterazine, di-dehydro coelenterazine) dissolved in anhydrous or absolute ethanol (Biotium or Molecular Imaging Products Company). 2. Assay buffer: for coelenterazines h and 400a, Dulbecco’s phosphate-buffered saline (D-PBS) containing 0.1  g/L CaCl2, 0.1 g/L MgCl2∙6H2O, and 1 g/L d-glucose (Gibco); for EnduRen, HEPES-buffered DMEM without phenol red, as described above for cell culture. 3. Microplate luminometer capable of measuring light through two filters (e.g. VICTOR Light (PerkinElmer), Mithras LB 940 (Berthold Technologies), and LUMIstar Omega or POLARstar Omega (BMG Labtech)). 4. Corresponding filter sets for the appropriate donor and acceptor “wavelength windows” respectively: for BRET1 (or eBRET) with a YFP as acceptor, 400–475  nm and 520–540 nm; for BRET1 (or eBRET) with EGFP as ­acceptor, 400–475  nm and 500–550  nm; for BRET2, 370–450  nm and 500–525 nm; for BRET3, 400–520 nm and 550–620 nm. 5. Scanning spectrometry enables visualization of the spectral shift observed with BRET and is a good complementary technique to the dual-filter luminometry. Such spectral analysis is especially useful when detecting constitutive protein–protein interactions (5). Examples of suitable scanning spectrometers include the Cary Eclipse (Varian), the Spex fluorolog or fluoromax (Jobin Yvon), and the FlexStation II (Molecular Devices).

3. Methods Four generations of BRET assay methodology have been developed (BRET1, eBRET, BRET2 and BRET3). BRET1 and eBRET use identical donor and acceptor molecules and only differ in luciferase substrate requirement. The eBRET method uses the EnduRen substrate possessing the same spectral properties, but an extended time period for signal detection in comparison to coelenterazine h, the substrate for BRET1 (3, 6). The BRET3 method uses either coelenterazine h or EnduRen, differing from BRET1 and eBRET as a result of the acceptor used, namely mOrange (8).

Study of GPCR–Protein Interactions by BRET

3.1. Generation and Validation of Fusion Constructs

361

1. Using DNA recombinant techniques, cDNAs for proteins of interest (e.g. GPCR and b-arrestin) are genetically cloned into suitable expression vectors (e.g. pcDNA3.1 or Gateway vectors from Invitrogen) containing the cDNA of the required luciferase donor or acceptor fluorophore in-frame (11). When generating fusion constructs, the stop codon between the cDNA sequences is removed and replaced with a linker region (see Note 1). 2. It is important to validate newly generated fusion proteins to ensure correct functionality of both the luciferase/ fluorophore and the protein of interest. GPCR functionality can be assessed in ligand binding and signaling assays comparing tagged and untagged receptors. Correct GPCR localization and trafficking in the presence and absence of agonist stimulation can be determined using confocal microscopy. This validation method can also be applied to intracellular proteins such as b-arrestins. Luciferase functionality can be assessed by addition of coelenterazine substrate and measurement of luminescence in a luminometer and/or scanning spectrometer. Fluorophore functionality can be assessed, in a fluorimeter with appropriate filters and/or a scanning spectrometer, by excitation with an external light source and measurement of resultant fluorescence (see Note 2).

3.2. Cell Culture

1. Plate cells into 6-well tissue culture plates so that they are ready for transfection with a suitable transfection reagent after 24 h; keep in a humidified incubator at 37°C, 5% CO2. 2. Following 24  h incubation, transfect cells with appropriate amounts of fusion protein cDNA (see Note 3). 3. Following a further 24-h incubation, if BRET detection is to be carried out using adherent cells, the cells are detached (optionally using trypsin-EDTA), resuspended in HEPESbuffered DMEM without phenol red and split (40–100 mL per well) into a 96-well white cell culture plate. The cells are then maintained at 37°C, 5% CO2 in a humidified incubator for a further 24 h to allow attachment (see Note 4). 4. It may be desirable to establish the relative expression of fluorescent and luminescent fusion proteins. In this case, a separate aliquot of each sample is excited by an external light source and the resultant fluorescence is observed. Luminescence from the same sample is then measured following addition of a coelenterazine substrate (see Note 5). 5. A separate aliquot of each sample can also be subjected to scanning spectral analysis if desired.

362

Kocan and Pfleger

3.3. BRET Signal Detection

1. Preparation of the luciferase substrate: for BRET1 (and BRET3), coelenterazine h (Invitrogen) is reconstituted in methanol; for BRET2, coelenterazine 400a (Biotium or Molecular Imaging Products Company) is reconstituted in anhydrous or absolute ethanol; for eBRET (and BRET3), EnduRen (Promega) is reconstituted in tissue culture grade dimethylsulfoxide (DMSO). Coelenterazines h and 400a are gently agitated to resuspend, whereas EnduRen requires extensive vortexing for up to 10 min and warming to 37°C. Aliquots are stored at −20°C protected from light. 2. The luciferase substrate is diluted in appropriate assay buffer immediately before adding to cells. Coelenterazines h and 400a are diluted in D-PBS with CaCl2, MgCl2 and D-glucose (Gibco) to give a final concentration of 5  mM. EnduRen is diluted in HEPES-buffered DMEM without phenol red at 37°C to give a final concentration of 30–60 mM. The substrates continue to be protected from light. 3. Taking extreme care not to detach cells, medium is removed and replaced with luciferase substrate in appropriate assay buffer (see Note 6). 4. BRET with coelenterazines h and 400a is measured immediately, whereas EnduRen is added to cells at least 90 min prior to BRET detection (during which time the cells are incubated at 37°C, 5% CO2 in a humidified incubator). 5. BRET is detected at 37°C using a luminometer measuring light through two appropriate filters. Light from each well is measured for 1–5 s through each filter (either simultaneously or sequentially) before measurement of the next well. 6. Ligand-induced GPCR–protein interactions can be detected in real-time (Figs.  1 and 2). Multiple BRET measurements are taken before and after treatment with ligand (or other modulator), which can be added by injection if time points are required immediately after treatment. Vehicle-treated samples are measured in parallel to provide a control for background signal (see Note 7). 7. In addition to monitoring agonist-induced GPCR–protein interactions, the subsequent effect of antagonist treatment can be observed in real-time using BRET (Fig. 1). 8. Scanning spectrometry, enabling visualization of the spectral shift characteristic of BRET, is a good complementary ­technique to BRET detection by dual-filter luminometry (Fig. 3). A secondary peak or shoulder appears at a ­wavelength characteristic of the acceptor emission (3, 5, 20) (see Note 8). Scanning spectrometry is especially useful as a supplementary technique when detecting agonist-independent/constitutive GPCR–protein interactions (Fig. 3c, d).

Study of GPCR–Protein Interactions by BRET

363

Fig. 1.  Detection of GPCR–protein interactions by real-time BRET1 assays. Kinetics data assessing the effect of an antagonist on agonist (arginine vasopressin; AVP)-induced interactions of vasopressin type II receptor (V2R)/Rluc8 and b-arrestin 2/Venus fusion proteins. The luciferase substrate coelenterazine h was added immediately prior to real-time measurements at 37°C. The transiently co-transfected HEK293 cells were firstly assayed before and after treatment with a submaximal concentration of agonist (AVP, 0.03 mM) or vehicle (PBS). 10 min following addition of agonist/vehicle, cells were treated with antagonist (final concentration of 10 mM) or vehicle (PBS). Data shown are means ± SEM of four independent experiments. Reprinted from ref. 5 with permission. Copyright 2009, The Endocrine Society.

3.4. Calculation and Interpretation of the BRET Signal

1. A ratio is calculated as the emission through the “acceptor ­wavelength window” divided by the emission through the “donor wavelength window” (e.g. for BRET1 (or eBRET) with a YFP as acceptor, 520–540 nm over 400–475 nm; for BRET1 (or eBRET) with EGFP as acceptor, 500–550  nm over 400–475  nm; for BRET2, 500–525  nm over 370–450 nm; for BRET3, 550–620 nm over 400–520 nm). This can be defined as the “fluorescence-over-luminescence” ratio for a particular sample. 2. In order to calculate a “ligand-induced BRET signal”, the above ratio is generated for both a ligand-treated and a vehicle-treated sample. Subtraction of the ratio for the vehicletreated sample from the ratio for the ligand-treated sample is considered to be the “ligand-induced BRET signal” or “ligand-induced BRET ratio” (2, 5, 6) (see Note 9). Using this calculation, it is not necessary to include a “donor-only” control sample, as the vehicle-treated sample represents the background (2, 6) (see Note 10). 3. The BRET signal (“ligand-induced” or otherwise) can be plotted against time in order to generate real-time BRET kinetic profiles (Figs. 1 and 2). Such data can potentially be used to calculate apparent association (or dissociation) rate constants (13).

364

Kocan and Pfleger

Fig. 2.  Detection of GPCR–protein interactions by real-time BRET1 (a, c) and BRET2 (b, d) assays. Kinetics data comparing vasopressin type II receptor (V2R)-wild-type (a, b) and its mutant V2R-R137L (c, d), were generated by monitoring the interaction with b-arrestin 2/Venus (BRET1) or b-arrestin 2/GFP10 (BRET2). The luciferase substrate coelenterazine h for BRET1 (a, c) or coelenterazine 400a for BRET2 (b, d) was added immediately prior to real-time measurements at 37°C. The transiently co-transfected HEK293 cells were assayed before and after treatment with ligand (arginine–vasopressin (AVP) final concentrations of 1, 0.1, and 0.01 mM) or vehicle (PBS). In comparison to wild-type V2R (a, b), mutant V2RR137L (c, d) showed elevated BRET signals for untreated cells indicating agonist-independent interactions with b-arrestin 2. Data shown are means ± SEM of four independent experiments. Reprinted from ref. 5 with permission. Copyright 2009, The Endocrine Society.

4. The BRET signal (“ligand-induced” or otherwise) can be plotted against the logarithm of ligand concentration in order to generate BRET dose-response curves (Fig. 4). Following curve-fitting by nonlinear regression, the concentration eliciting half-maximal response (EC50 value) can be calculated from such data (3, 13). 5. Scanning spectrometry can also be applied to quantify the BRET signal. In this case, within wavelength windows corresponding to the luminometer filters, the area under the curve is obtained so that the BRET signal is calculated in a similar manner to that described above for dual-filter luminometry (see Note 11). 6. The Z¢-factor can be generated to assess the potential suit­ ability of the assay for high-throughput screening. The Z¢-factor

Study of GPCR–Protein Interactions by BRET

365

Fig. 3.  BRET1 (a, c) and BRET2 (b, d) spectral analysis of wild-type and mutant (R137L) vasopressin type II receptors (V2Rs) tagged with Rluc8 in untreated and arginine–vasopressin (AVP)-treated HEK293 cells using coelenterazine h (a, c) or coelenterazine 400a (b, d) as luciferase substrate. Cells were co-transfected with cDNA for b-arrestin 2/Venus (a, c) or b-arrestin 2/GFP10 (b, d) and a V2R construct tagged with Rluc8. Emission spectra were recorded immediately after substrate addition. In V2R wild-type untreated cells, an emission maximum of ~480 nm or ~420 nm corresponds to Rluc oxidizing coelenterazine h (a) or coelenterazine 400a (b), respectively. In V2R wild-type AVP-treated cells, an additional emission peak appears at ~530 nm (a) or ~510 nm (b) that corresponds to emission from Venus or GFP10, respectively, thereby demonstrating BRET due to ligand-induced V2R/b-arrestin 2 interactions. Emission spectra for the V2R mutant, V2R-R137L (c, d), exhibit two emission peaks in untreated as well as treated cells, thereby demonstrating BRET due to agonist-independent as well as agonist-induced V2R/b-arrestin 2 interactions. Data are representative of at least three independent experiments. Reprinted from ref. 5 with permission. Copyright 2009, The Endocrine Society.

is calculated with respect to the mean and standard deviation (SD) of control data using the equation Z¢ = 1 − ((3SD of positive control + 3SD of negative control)/|mean of positive control − mean of negative control|) (21). For example, the “positive control” values are calculated from the “­fluorescenceover-luminescence” ratios for agonist-treated samples and the “negative control” values are calculated from these ratios for vehicle-treated samples (3). In contrast, the Z-factor is used to assess the performance of an actual compound screen. It is calculated in a similar manner to the Z¢-factor, except the mean and SD of the “fluorescence-over-luminescence” ratios for the compounds are used instead of the negative control

366

Kocan and Pfleger

Fig. 4.  Thyrotropin-releasing hormone (TRH) concentration-response curves comparing TRH-induced interactions of different TRH receptor (TRHR)/luminophore fusion proteins with b-arrestin 2/Venus using BRET1 assays. The luciferase substrate coelenterazine h was added to transiently co-transfected HEK293 cells immediately prior to commencing the assay at 37°C. Measurements were taken ~30 min after treatment with various concentrations of TRH or vehicle. Data shown are means ± SEM of three independent experiments. Reprinted from ref. 3 with permission. Copyright 2008, Society for Biomolecular Sciences.

Fig. 5.  BRET Z¢-factor assay performance detecting thyrotropin-releasing hormone (TRH)-induced interactions of TRH receptor (TRHR)/Rluc8 with b-arrestin 2/Venus. Transiently co-transfected HEK293 cells were treated with TRH (positive control) or vehicle (negative control) and luciferase substrate coelenterazine h that were both added immediately prior to real-time measurements at 37°C. Fluorescence/luminescence values are presented against time (a) or well number at 60 min (b) following TRH/phosphate-buffered saline (PBS) addition. In (b), the solid horizontal lines show the means of the positive control (TRH) and negative control (PBS). Broken lines display three standard deviations (SD) from the mean of each data set. Data shown are representative of three independent experiments. Reprinted from ref. 3 with permission. Copyright 2008, Society for Biomolecular Sciences.

values, as it is assumed that the vast majority of the compounds tested will be negative (21). “Fluorescence-over-luminescence” ratios are used rather than one of the forms of BRET ratio, as this enables assessment of the variance of experimental data compared with control data (Fig. 5; see Note 12).

Study of GPCR–Protein Interactions by BRET

367

4. Notes 1. Efficient resonance energy transfer is dependent on a number of factors, including the relative orientation of the donor and acceptor dipoles. Optimal relative orientation becomes more likely with greater freedom of movement (22), which can be achieved by incorporating a linker region between the protein of interest and the donor or acceptor. However, it is important to ensure that the linker region itself does not unintentionally confer altered functional characteristics on the modified receptor, such as inclusion of phosphorylation and/ or b-arrestin binding sites. 2. If newly generated constructs are not functioning properly, and the correct cDNA sequence is confirmed, fusion may be compromising proper function of the protein of interest and/ or of the donor/acceptor. Longer linker regions (see Note 1) or altered donor/acceptor positioning may help to alleviate this problem. When detecting GPCR interactions with intracellular proteins, such as b-arrestins, the GPCR needs to have the donor or acceptor fused to the C-terminus as this is cytosolic. In contrast, b-arrestins have been monitored successfully with donor/acceptor fused to either the N- or C-termini. Some of the possible reasons for observing low relative luminescence or fluorescence counts are: suboptimal transfection of cells (including amount and/or ratio of cDNAs), low cell number, poor substrate viability, the presence of a reducing agent, such as ascorbic acid, and incorrect instrument calibration. 3. The optimal amounts of fusion protein cDNAs need to be established empirically by titration BRET experiments testing various ratios and concentrations, particularly as there is not necessarily a direct relationship between cDNA quantity and final concentration of functional protein expressed. The optimal transfection amounts will depend not only on the expression efficiency of the fusion protein, but also the optimal ratio of donor to acceptor. A greater amount of acceptor than donor is usually advisable to ensure maximal energy transfer, although too much acceptor may inhibit the BRET signal in situations where the acceptor-tagged proteins of interest interact with each other in competition with the donor-tagged protein. 4. If suspended cells are to be used for measuring BRET, such as with BRET cell titration assays (3), they are not detached at this stage. Rather, they are incubated for 48 h after transfection and until immediately prior to BRET detection. Cells are then detached (optionally using trypsin-EDTA), resuspended in BRET assay buffer, and transferred to a white 96-well plate.

368

Kocan and Pfleger

5. It is, for example, important to establish the relative expression of fluorescent and luminescent fusion proteins prior to BRET detection when investigating the effect of GPCR point-mutations on GPCR–protein interactions. In this case, it is particularly advisable to ensure similar expression levels of the mutant and wild-type GPCRs being compared. 6. For eBRET and BRET3 studies using EnduRen, medium removal and replacement is not necessary and may be undesirable if there are concerns regarding cell detachment. EnduRen can be added to HEPES-buffered DMEM without phenol red already present on the cells. 7. The time period over which BRET measurements can be taken following substrate addition is dependent on both the substrate and form of luciferase. As a general guide: BRET with coelenterazine h can be measured for up to 2  h with Rluc2/Rluc8 or up to 1  h with Rluc/hRluc; BRET with EnduRen can be measured for many hours with the limitation usually being cell viability (6); BRET with coelenterazine 400a can be measured for up to 1 h with Rluc2/Rluc8, but for just seconds to minutes with Rluc/hRluc (23). Protected forms of coelenterazine 400a have been developed that extend this time period (24). 8. Cells to be used for BRET detection by dual-filter luminometry and spectral analysis by scanning spectrometry are prepared in a similar manner and thus both assays can be carried out in parallel with duplicate aliquots of samples. For spectral analysis, substrate and ligand/vehicle are added as for BRET detection. However, instead of monitoring the ligand-induced interaction in real-time, the cells are incubated at 37°C, 5% CO2 in a humidified incubator until a predetermined time point at which emission spectral scans are performed using a scanning spectrometer. For example, time-gated luminescence emission spectra are detected in the 350–600 nm wavelength range, with the time gate window set to 300 ms and emission slit to 10 nm (Fig. 3). 9. As the “ligand-induced BRET signal” compares ligand-treated and vehicle-treated samples, a negative value is possible. This may be due to ligand causing the interaction to become weaker, more transient, or occur less frequently. An example of such a scenario would be when an inverse agonist disrupts an interaction resulting from constitutive receptor activity. Importantly, it should also be borne in mind that the BRET signal is very sensitive to distance and relative orientation between donor and acceptor. Therefore, a ligand-induced change in BRET signal (positive or negative) could be due to conformational changes influencing the relative positioning of donor and acceptor, and consequently the efficiency of energy transfer.

Study of GPCR–Protein Interactions by BRET

369

10. Alternatives to the “ligand-induced BRET signal” calculation: (a) A “BRET signal” or “BRET ratio” can be calculated by ­generating “fluorescence-over-luminescence” ratios for the sample containing both donor- and acceptor-linked fusion proteins and a sample containing only the donorlinked fusion protein. The ratio for the donor-only control sample is then subtracted from the sample containing both fusion proteins. This calculation can be used for both ligand-induced, as well as non-ligand-mediated/ constitutive interactions. The assumption is that the ratiometric nature of the calculation accounts for differences in protein expression between samples with and without acceptor-linked fusion protein. However, it is advisable to express similar amounts of donor-linked fusion protein in both the experimental and control samples. (b) The “BRET ratio above wild-type baseline” is calculated by comparing “fluorescence-over-luminescence” ratios for the following samples: the first (experimental) sample contains fusion proteins of a mutant GPCR and its interacting protein of interest (e.g. vasopressin type II receptor (V2R)-R137L/Rluc8 and b-arrestin 2/Venus) treated with ligand; the second (control) sample contains fusion proteins of the corresponding wild-type GPCR and the same interacting protein of interest (e.g. V2Rwild-type/Rluc8 and b-arrestin 2/Venus) treated with vehicle. The ratio for the control sample is subtracted from the ratio for the experimental sample to give the “BRET ratio above wild-type baseline” (Fig.  2). This form of BRET data presentation is advantageous when investigating the effect of GPCR point-mutations on GPCR–protein interactions (5). It is advisable to ensure similar expression levels of mutant and corresponding wild-type GPCRs. 11. An alternative calculation of the BRET signal using scanning spectrometry data requires the Rluc emission peak of each spectrum to be normalized to an intensity of 1. The BRET signal can then be calculated using the area under the curve between 500 and 550  nm for BRET1 (or eBRET), between 480 and 530 nm for BRET2 or between 550 and 600 nm for BRET3. When calculating the “ligand-induced BRET signal” using this approach, the background corresponds to the area under the curve in this wavelength region from a vehicle-treated control sample. Alternatively, it is also possible to use a donor-only control sample (20), or a sample containing a vehicle-treated wild-type receptor to calculate the background, depending on the desired form of data presentation (see Note 10).

370

Kocan and Pfleger

12. Statistical analysis of BRET assays can also be performed using Student’s t-tests or ANOVA with suitable post-tests, again by comparing the variance in the “fluorescence-overluminescence” ratios from experimental and control samples (2).

Acknowledgments KDGP’s work using the BRET methodology is funded by the National Health and Medical Research Council (NHMRC) of Australia (Project Grant #566736). KDGP is an Australian Research Council (ARC) Future Fellow (FT100100271). References 1. Pfleger, K. D. G. and Eidne, K. A. (2006) Illuminating insights into protein–protein interactions using bioluminescence resonance energy transfer (BRET). Nat. Methods 3, 165–174. 2. Pfleger, K. D. G., Seeber, R. M., and Eidne, K. A. (2006) Bioluminescence resonance energy transfer (BRET) for the real-time detection of protein–protein interactions. Nat. Protoc. 1, 337–345. 3. Kocan M., See H. B., Seeber R. M., Eidne K. A., and Pfleger K. D. G. (2008) Demonstration of improvements to the bioluminescence resonance energy transfer (BRET) technology for the monitoring of G protein-coupled receptors in live cells. J. Biomol. Screen. 13, 888–898. 4. Milligan, G. and Bouvier, M. (2005) Methods to monitor the quaternary structure of G-protein-coupled receptors. FEBS J. 272, 2914–2925. 5. Kocan M., See H. B., Sampaio N. G., Eidne K. A., Feldman B. J., and Pfleger K. D. G. (2009) Agonist-independent interactions between b-arrestins and mutant vasopressin type II receptors associated with nephrogenic syndrome of inappropriate antidiuresis. Mol. Endocrinol. 23, 559–571. 6. Pfleger, K. D. G., Dromey, J. R., Dalrymple, M. B., Lim, E. M. L., Thomas, W. G., and Eidne, K. A. (2006) Extended bioluminescence resonance energy transfer (eBRET) for monitoring prolonged protein–protein interactions in live cells. Cell. Signal. 18, 1664–1670. 7. De, A., Loening, A. M., and Gambhir, S. S. (2007) An improved bioluminescence resonance energy transfer strategy for imaging intracellular events in single cells and living subjects. Cancer Res. 67, 7175–7183.

8. De A., Ray P., Loening A. M., and Gambhir S. S. (2009) BRET3: a red-shifted bioluminescence resonance energy transfer (BRET)-based integrated platform for imaging protein–protein interactions from single live cells and living ­animals. FASEB J. 23, 2702–2709. 9. Guo, W., Urizar, E., Kralikova, M., Mobarec, J. C., Shi, L., Filizola, M., and Javitch, J. A. (2008) Dopamine D2 receptors form higher order oligomers at physiological expression levels. EMBO J. 27, 2293–2304. 10. Kamal, M., Marquez, M., Vauthier, V., Leloire, A., Froguel, P., Jockers, R., and Couturier, C. (2009) Improved donor/acceptor BRET couples for monitoring b-arrestin recruitment to G protein-coupled receptors. Biotechnol. J. 4, 1337–1344. 11. Pfleger, K. D. G. and Eidne, K. A. (2003) New technologies: bioluminescence resonance energy transfer (BRET) for the detection of real time interactions involving G-protein ­coupled receptors. Pituitary 6, 141–151. 12. Pfleger, K. D. G. and Eidne, K. A. (2005) Monitoring the formation of dynamic G ­protein-coupled receptor-protein complexes in living cells. Biochem. J. 385, 625–637. 13. Hamdan, F. F., Audet, M., Garneau, P., Pelletier, J., and Bouvier, M. (2005) High-throughput screening of G protein-coupled receptor antagonists using a bioluminescence resonance energy transfer 1-based b-arrestin2 recruitment assay. J. Biomol. Screen. 10, 463–475. 14. Pfleger, K. D. G., Dalrymple, M. B., Dromey, J. R., and Eidne, K. A. (2007) Monitoring interactions between G-protein-coupled receptors and b-arrestins. Biochem. Soc. Trans. 35, 764–766. 15. DeWire, S. M., Ahn, S., Lefkowitz, R. J., and Shenoy, S. K. (2007) b-arrestins and cell ­signaling. Annu. Rev. Physiol. 69, 483–510.

Study of GPCR–Protein Interactions by BRET 16. Dromey, J. R. and Pfleger, K. D. G. (2008) G protein-coupled receptors as drug targets: The role of b-arrestins. Endocr. Metab. Immune Disord. Drug Targets 8, 51–61. 17. Nagai, T., Ibata, K., Park, E. S., Kubota, M., Mikoshiba K., and Miyawaki, A. (2002) A variant of yellow fluorescent protein with fast and efficient maturation for cell-biological applications. Nat. Biotechnol. 20, 87–90. 18. Mercier, J. F., Salahpour, A., Angers, S., Breit, A., and Bouvier, M. (2002) Quantitative assessment of b1- and b2-adrenergic receptor homoand heterodimerization by bioluminescence resonance energy transfer. J. Biol. Chem. 277, 44925–44931. 19. Shaner, N. C., Campbell, R.E., Steinbach, P. A., Giepmans, B. N. G., Palmer, A. E., and Tsien, R. Y. (2004) Improved monomeric red, orange and yellow fluorescent proteins derived from Discosoma sp. red fluorescent protein. Nat. Biotechnol. 22, 1567–1572. 20. McVey, M., Ramsay, D., Kellett, E., Rees, S., Wilson, S., Pope, A. J., and Milligan, G. (2001)

371

Monitoring receptor oligomerization using time-resolved fluorescence resonance energy transfer and bioluminescence resonance energy transfer. J. Biol. Chem. 276, 14092–14099. 21. Zhang, J.-H., Chung, T. D. Y., and Oldenburg, K. R. (1999) A simple statistical parameter for use in evaluation and validation of high throughput screening assays. J. Biomol. Screen. 4, 67–73. 22. Wu, P. and Brand, L. (1994) Resonance energy transfer: methods and applications. Anal. Biochem. 218, 1–13. 23. Jaeger, W. C., Pfleger, K. D. G., and Eidne, K. A. Monitoring GPCR-protein complexes using bioluminescence resonance energy transfer, in G Protein Coupled Receptors: Essential Methods (Poyner, D. and Wheatley, M., ed.), John Wiley & Sons, Hoboken, NJ, 111–132. 24. Levi, J., De, A., Cheng, Z., and Gambhir, S. S. (2007) Bisdeoxycoelenterazine derivatives for improvement of bioluminescence resonance energy transfer assays. J. Am. Chem. Soc. 129, 11900–11901.

wwwwwww

Chapter 21 Time Resolved FRET Strategy with Fluorescent Ligands to Analyze Receptor Interactions in Native Tissues: Application to GPCR Oligomerization Martin Cottet, Laura Albizu, Laetitia Comps-Agrar, Eric Trinquet, Jean-Philippe Pin, Bernard Mouillac, and Thierry Durroux Abstract G protein-coupled receptors (GPCRs) play a key role in the regulation of physiological functions. Deregulation of their activities often results in pathological disorders and therefore these receptors constitute major targets for drug development. The emergence of new concepts such as GPCR oligomerization has modified our understanding of these proteins, and identifying the role of receptor complexes is probably a major challenge for the next decade. Various experimental strategies have been developed to study GPCR oligomers and energy transfer experiments between partners within a complex constitute one of the most convenient approaches. These experimental strategies usually require receptor fusion to tags or fluorescent or luminescent proteins and therefore cannot be easily applied to native tissues. We developed a new experimental approach based on the labeling of receptors with high affinity fluorescent ligands compatible with time-resolved energy transfer measurements. Because of the very high signal-tonoise ratio of the time-resolved fluorescent energy transfer (TR-FRET) signals, this approach constitutes a breakthrough since it allows the direct identification of wild-type GPCR oligomers in native tissues. Key words: Fluorescent ligands, G protein-coupled receptor, GPCR, Time-resolved FRET, Europium, Terbium

1. Introduction Our understanding of the functioning of G protein-coupled receptors (GPCR) has evolved during the last two decades with the emergence of new concepts. Among them, receptor oligomerization (1, 2) is still the subject of intense investigation since the functional consequences of this are far from being well identified. Oligomerization was first reported as the dimerization of two receptors of the same type leading to the Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_21, © Springer Science+Business Media, LLC 2011

373

374

Cottet et al.

formation of receptor homodimers. The concept has evolved in two ways. Firstly, interactions between different types of receptors leading to the formation of heterodimers have been demonstrated. More recently, the existence of larger receptor complexes have been described suggesting the existence of higher-order oligomers. The investigation of receptor oligomerization is crucial since it has been shown to play a role in receptor targeting and internalization and seems to be at the origin of some variations in pharmacological and coupling profiles. Most of the studies focused on receptor oligomerization have consisted of the identification of receptor complexes and various strategies such as co-immunoprecipitation (3), atomic force microscopy (4) and binding or functional assays (5–8) that have been developed. The most convenient experimental methods to demonstrate receptor proximity are those based on resonance energy transfer since the amplitude of the signal is dependent on the distance between the markers. Various versions of resonance energy transfer technologies have been developed (9), including bioluminescent resonance energy transfer (BRET) in which the donor is a luminescent molecule (generally luciferase), and fluorescent resonance energy transfer (FRET) based on the use of green fluorescent protein (GFP) variants, most often yellow fluorescent protein (YFP ) as fluorescent donor and CFP (cyan fluorescent protein) as acceptor fused to the receptor. One of the most important drawbacks of these strategies is that the signal does not only result from receptors targeted to the cell surface but also from those trapped inside cells. The inability of these methods to distinguish between surface receptors participating in cell stimulation and intracellular receptors that cannot be activated by ligands restrains the interest of these strategies, unless sophisticated microscopy approaches are used. Moreover, these strategies are not easily applicable to native tissues since they are based on chimeric receptor expression, which would ultimately require the generation of knock-in mice expressing the desired fluorescent fusion receptors. More recently, time-resolved FRET (TR-FRET) strategies based on the use of lanthanide (essentially terbium or europium) cryptates have opened new perspectives since they exhibit a very high signal-to-noise ratio (10) and allow the labeling of only the receptors targeted to the cell surface. TR-FRET strategies take advantage of the long-lasting fluorescence of the lanthanides. The fluorescence lifetime of terbium and europium is in the millisecond range, and is about 100,000 times longer than those of classic fluorophores (about 10  ns). Therefore, fluorescent measurement after a time delay (50 ms) allows the measurement of acceptor fluorescence only if it is engaged in a FRET since it will exhibit in this case a long-lasting fluorescence. Fluorescent

Time Resolved FRET Strategy with Fluorescent Ligands

375

Table 1 Donor and acceptor fluorophores compatible with time-resolved FRET experiments, with their typical excitation and emission wavelengths Donor

Acceptor Name

lexcitation (nm) lemission (nm) Name

lemission (nm)

Europium-cryptate Eu – PBBP 320–340

620

d2 like fluorophores 665 (d2, d1 Cy5, Alexa 647, DY647…)

Terbium- cryptate

620

-d2 like fluorophores -fluorescein like fluorophores

Lumi4-Tb® 335–340

665 520

acceptors exhibiting fluorescein-like properties (with Lumi4terbium as donor) or emitting in the near infra-red range around 665 nm (d2-like fluorescent properties, with europium-cryptates or Lumi4-terbium as donors), are compatible with TR-FRET experiments (Table 1). Various strategies can be developed to label receptors targeted to the cell surface with TR-FRET compatible fluorophores. For some of them, receptor modifications are mandatory and are achieved by fusing it to tags or to suicide enzymes recognized by specific antibodies (11, 12) or specific substrates (Tag-lite® strategy) (13), respectively. Such techniques cannot be used in native tissues. To overcome this limitation, we developed an alternative strategy allowing the labeling of GPCRs targeted to the cell surface. It is based on the binding of specific fluorescent ligands compatible with time-resolved FRET experiments (14). This strategy of GPCR labeling has been used to study homodimers. In that context, two fluorescent versions of the same ligand (donor-ligand and acceptor-ligand) have to be developed and ­optimal TR-FRET signal is observed when all receptors are labeled, 50% with the donor-ligand and 50% with the acceptor-ligand. Of note, in these conditions, and when considering only dimers and not higher-order oligomer complexes, 25% of the dimers are labeled with two donors, 25% with two acceptors, and 50% with a donor and an acceptor. Only these last complexes contribute to the FRET signal. The optimization of the labeling with donor and acceptor, first to determine the best conditions for receptor labeling and ­second to obtain a donor and acceptor well-balanced labeling, is thus crucial for appropriate homodimer detection. The same strategy could also be used to detect GPCR heterodimers although it has not been done yet. In such a context, the combination of a donor-labeled ligand targeting the first GPCR with a second ligand targeting the second GPCR labeled with an acceptor should allow the labeling of 100% of these heterodimers.

376

Cottet et al.

This receptor labeling strategy is really efficient since it does not require modification of the receptor sequence. However, it is noteworthy that TR-FRET signal amplitude is dependent on ligand binding and therefore on the affinity of the ligands. For the same receptor occupation, low affinity ligands (Ki > 20 nM) have to be used at higher concentrations leading to lower signalto-noise ratio.

2. Materials 2.1. Cells 2.2. Cell Culture

Stable cell lines expressing the receptor of interest or cell lines which can be transiently transfected (see Note 1). 1. Phosphate buffered saline (PBS) prepared from 10× stock solution (Invitrogen, Cergy Pontoise, France). 2. Trypsin–EDTA solution (Invitrogen, Cergy Pontoise, France). 3. Dulbecco’s Modified Eagle’s Medium (DMEM) (Lonza, Verviers, Belgium). 4. Fetal calf serum (FCS) (Lonza, Verviers, Belgium). 5. Penicillin/streptomycin (Invitrogen, Cergy Pontoise, France). 6. Electroporation buffer (EB): 250  mM KH2PO4, 100  mM CH3COOK, and 100 mM KOH prepared from a 5× solution (EB 5×). 7. Solution of MgSO4 (1 M). 8. Cell electroporator (e.g., Gene pulser, Bio-Rad Laboratories, Marnes-La-Coquette, France). 9. Black 96-well plate (e.g., Greiner Cell Star 96-well plate, Dominique Dutscher, Brumath, France). 10. Lipofectamine™ 2000 (Invitrogen, Cergy Pontoise, France).

2.3. Membrane Preparation

1. PBS without calcium and magnesium (Lonza, LevalloisPerret, France). 2. Ice-cold lysis buffer: 15  mM Tris-HCL, 2  mM MgCl2, 0.3 mM EGTA, pH 7.4. 3. Ice-cold 10% sucrose buffer: 10% (w/v) sucrose, 10 mM TrisHCL, 1 mM EDTA, pH 7.4. 4. Ice-cold 35% sucrose buffer: 35% (w/v) sucrose, 10 mM TrisHCL, 1 mM EDTA, pH 7.4. 5. Membrane suspension buffer: 50  mM Tris, 5  mM MgCl2, pH 7.4. 6. An Ultra-Turrax homogenizer Labortechnik, Staufen, Germany).

(Janke

Kunkel

IKA-

Time Resolved FRET Strategy with Fluorescent Ligands

377

7. An ultracentrifuge with a swing SW-28-out bucket rotor (Beckman Coulter, Roissy, France). 8. Bradford protein assay (BioRad protein assay laboratories, Marnes-La-Coquette, France). 2.4. Ligands

1. GPCR ligands derivatized with TR-FRET acceptors (fluorescein, AlexaFluor 488, d2, d1, AlexaFluor 647 or Cy5 can be found in the literature (15) and can therefore be synthesized (see Note 2)). 2. Ligands derivatized with lanthanide cryptates were synthesized by Cisbio Bioassays (Cisbio Bioassays Drug Discovery, Bagnols-sur-Cèze, France). 3. Labeling buffer: 20  mM Tris-HCL, 118  mM NaCl, 1.2  mM KH2PO4, 1.2 mM MgSO4, 4.7 mM KCl, 1.8 mM CaCl2, pH 7.4.

2.5. Reading of the Signal

TR-FRET signals can be read on any microplate reader compatible with HTRF (see www.HTRF.com) (see Note 3).

3. Methods The method should be first optimized on cell lines transiently expressing the receptor of interest and then applied to native tissues. Because experiments can be performed both on intact cells and membrane preparations, the protocol used for membrane preparation is indicated afterwards. 3.1. Expression of the GPCR in a Cell Line

Electroporation was initially used to transiently express Snap-tag receptors. Recently, we have also used Lipofectamine transfection according to the protocol provided by the manufacturer. We indicate below the protocol followed for electroporation (see Subheading 2.2). 1. COS-7 are kept in culture in DMEM medium supplemented with Fetal calf serum (10%), penicillin/streptomycin antibiotics (1%) at 37°C in an atmosphere of 95% air and 5% CO2. Cells are split before they reach confluence. 2. The plasmid mix (152 mL) is prepared in sterile water from plasmid encoding the Snap-tag receptor of interest and supplemented with empty vector to 10 mg (see Note 4). 3. 40 mL EB 5× and 8 mL 1 M MgSO4 are added to the ­plasmid mix. 4. Cell dishes are washed once with PBS (10 mL) and cells are dissociated with prewarmed trypsin–EDTA solution. (4 mL/ dish) for about 3 min at 37°C.

378

Cottet et al.

5. To neutralize trypsin/EDTA solution activity, 6 mL/dish of prewarmed DMEM supplemented with 10% FCS is added and cells are harvested, counted on a Malassez cell and centrifuged at 160 × g for 5 min. 6. Supernatant is removed and cells are resuspended in EB 1× buffer at a density of 100 million cells/mL. Of note, at this density, cell volume represents about 1/3 of the final volume. 7. 100  mL of cells are added to 200  mL of the plasmid mix. Solution is gently mixed by pipetting and incubated at room temperature for 5 min. 8. Each mix preparation is transferred to an electroporation cuvette and placed in the electroporator (see Note 5). 9. After electric shock delivery, cells are resuspended in 10 mL of fresh complete culture medium and are seeded in 96-well plate at a density of 100,000 cells/well. 10. Cells are incubated for 48 h at 37°C, 5% CO2. 3.2. Membrane Preparation 3.2.1. Membrane Preparation from Cultured Cell Line

See Subheading 2.3. 1. Cells cultured in 150-mm culture dishes are rinsed twice with 10 mL of PBS without calcium or magnesium. 2. 4 mL of ice-cold lysis buffer are added and cells are scraped with a rubber policeman. 3. The cells are transferred to a 50-mL tube. 4. Culture dishes are rinsed with 4 mL of ice-cold lysis buffer and the solution is pooled with the previously collected fraction. 5. Cells are homogenized for 30 s with an Ultra-Turrax homogenizer, and centrifuged at 100 × g for 5 min at 4°C. 6. Supernatants are recovered and centrifuged at 44,000 × g for 30 min at 4°C. 7. Pellets are resuspended in suspension buffer and centrifuged at 44,000 × g for 30 min at 4°C. 8. Pellets are resuspended in a small volume of suspension ­buffer. For each membrane preparation, protein concentration is estimated by colorimetric assay (e.g., Bradford assay). 9. Membrane preparation can either be immediately used or stored as aliquots (100 mL containing 100–300 mg membrane protein) in liquid nitrogen.

3.2.2. Membrane Preparation from Native Tissue

Preparation from native tissue is dependent on the tissue collected. Here we indicate the protocol we used to prepare membranes from mammary gland.

Time Resolved FRET Strategy with Fluorescent Ligands

379

1. After sacrifice of 14–18 day lactating rats, mammary glands are collected and kept in ice-cold PBS without calcium or magnesium. 2. Dissected mammary glands are freed of connective tissue and cut into small pieces in ice-cold lysis buffer. 3. Tissues are homogenized for 30  s with an Ultra-Turrax homogenizer (see Note 6). 4. The homogenate is spun at 500 × g for 10 min at 4°C. 5. Supernatants are recovered and centrifuged at 12,000 × g for 30 min at 4°C. 6. Pellets are homogenized in 90  mL of ice-cold 10% sucrose buffer and 15 mL of the homogenate are layered onto 15 mL 35% sucrose buffer in each centrifuge tube. 7. The preparation is centrifuged at 100,000 × g for 2 h at 4°C in a swing SW-28-out bucket rotor. 8. The membranes of mammary gland are collected at the 10–35% sucrose interface. 9. Membranes are dispersed in ice-cold membrane suspension buffer (25–30 mL). 10. Membranes are centrifuged at 40,000 × g for 30 min at 4°C and pellets are resuspended in a small volume (3–5 mL) of the same buffer. For each membrane preparation, protein concentration is estimated by colorimetric assay (e.g., Bradford assay). 11. Membranes are immediately used or aliquots (100  mL containing 100–400 mg of membrane protein) are stored in liquid nitrogen (see Note 7). 3.3. Labeling of Receptor: Optimization of the FRET Signal

Three parameters can vary in order to optimize the FRET signal (see Note 8). See Note 9 and Table  2 for experimental conditions which have to be included in the experiments in order to properly analyze the results.

Table 2 Experimental controls required for the analysis of FRET signals Samples

Negative FRET control

Dynamic FRET control Receptor specificity

Donor-ligand + acceptor-ligand

Donor-ligand

Donor-ligand + acceptor-ligand

Donor-ligand + acceptor-ligand

Expressed receptor (cell or membrane)

Expressed receptor (cell or membrane)

No cell or membrane

Mock cell or membrane or expressed receptor (cell or membrane) + excess unlabeled ligand

380

Cottet et al.

Fig. 1. Optimization of the TR-FRET signal by variation of the ligand concentrations. Different donor-ligand/acceptor-ligand ratios are tested by varying the acceptor concentration while the concentration of the donor remains constant. The optimized ratio corresponds to the peak of the bell-shaped curve. Various curves can be established to determine the optimal concentration of donor. 3.3.1. Optimization of the Ligand Concentrations

Experiments on Cells

As mentioned above, the aim is to optimize the concentration of donor and acceptor labeled ligands to obtain an optimal TR-FRET signal. As a prerequisite, affinities of the ligand and more specifically of the donor-ligand should have been determined. To determine the donor-ligand/acceptor-ligand ratio, we keep the donor concentration constant and vary the acceptor concentration from picomolar to micromolar levels (Fig.  1). Various curves are established with different concentrations of donor-ligand. These experiments can either be performed on cells expressing the receptor of interest or on membrane preparations. 1. Cells are transiently transfected with plasmids encoding the protein of interest in black 96-well plates and incubated for 48 h at 37°C, 5% CO2 (see Subheading 3.1). 2. The different ligands are prepared separately in ice-cold labeling buffer complete with 0.1% BSA and 0.1% glucose, and kept on ice (by working below 15°C, all internalization phenomena are avoided ). At this point, all ligand solutions are prepared at four times the desired final concentration (see Note 10). 3. Premixes of the solutions are performed according to the conditions tested and kept on ice. Cells are placed on melting

Time Resolved FRET Strategy with Fluorescent Ligands

381

ice to cool down. The medium is then removed and replaced with 100 mL of cold mixed ligand solutions. 4. Plates are incubated overnight at 4°C with gentle rocking. 5. The signals are recorded at 620  nm for the donor and at 520  nm (e.g., fluorescein, alexa488) or 665  nm (e.g., d2) depending on the acceptor (see Note 11). 6. Data analyses are performed as indicated in the data analysis section. E xperiments on Membrane Preparations

1. Experiments are performed in a final volume of 200 mL/well. 100 mL of membranes are distributed in each well, in a black 96-well plate, and placed on melting ice. 2. 100  mL of premixed ligand solution (prepared as described previously for experiments on cells, except with 8× initial solutions, due to additional dilution by the membranes) is distributed in each well. 3. Plates are incubated overnight at 4°C with gentle rocking. 4. Fluorescence and TR-FRET signals are recorded as previously described.

3.3.2. Optimization of the Membrane Concentration

After determining the optimal fluorescent ligand concentrations, a second step of optimization of the FRET signal can be performed on the quantity of membrane used in the assay. Indeed the FRET signal is in theory proportional to the receptor labeling. However, different parameters can reduce the FRET signal. For example, if fluorescent ligands are not in excess regarding the number of receptor ÿimmers, the receptor dimer will be labeled with only one fluorescent ligand, either a donor or an acceptor, resulting in the absence of FRET. Therefore, variation of the amount of membrane per assay will result in a biphasic curve, with the ascending portion proportional in the initial phase to the receptor concentration and a descending portion which could be explained by an insufficient occupation of the ÿimmers by the ligands (Fig. 2). 1. The different ligands are prepared separately and ligand premixes are made as described above. 2. Various quantities of membranes (0–300 mg) are distributed in the wells from a stock solution and volumes in the well are completed with the suspension buffer to a final volume of 100 mL. 3. For each membrane concentration, negative controls have to be performed (see Note 9). 4. Plates are incubated overnight at 4°C with gentle rocking. 5. FRET signals are measured as described above.

382

Cottet et al.

Fig. 2. Optimization of the TR-FRET signal by variation of the membrane quantity. The donor-ligand/acceptor-ligand ratio being established, the amount of membrane per assay can be optimized.

3.3.3. Optimization of the Fluorescent Ligands

Various FRET signal intensities can be measured depending on the nature of the ligand. Indeed, crosstalk between binding sites of protomer within a dimer can lead to positive or negative cooperativity of binding. Positive cooperativity favors the binding of two ligands per dimer leading to a greater FRET signal. Testing various pairs of fluorescent ligands can therefore be relevant to optimize the FRET signal. Using various GPCRs, we repeatedly observed a larger signal with fluorescent antagonists than with fluorescent agonists (see Note 12).

3.4. Data Analysis

Two parameters are classically used to measure the FRET signal. 1. D665 (for d2-like acceptor) or D520 (for fluorescein-like acceptor) corresponding to variation of the fluorescence emission intensity of the acceptor. This parameter is the simplest one but intensity variations can also reflect variations in the experimental conditions. 2. DF % (see the formula below): this parameter takes into account the variation of the donor concentration and corresponds to a percentage of the FRET increase compared to a negative control. However, the DF % parameter does not allow the comparison of experiments that are not performed with the same donor-ligand concentrations.

Time Resolved FRET Strategy with Fluorescent Ligands 3.4.1. Determination of the D665 Specific FRET Signal (or D520, Depending on the Acceptor Fluorophore Used)

383

The calculation of the D665 (for example) allows the representation of the specific FRET signal from which has been subtracted nonspecific FRET (bleed-through from the donor, FRET resulting from random donor-acceptor collision: dynamic FRET). It corresponds to D665 = signal at 665  nm of the sample – signal at 665 nm of the negative control. The negative control can include: 1. Sample with only the donor-ligand. 2. Mock cells or membranes with donor and acceptor labeled ligands (gives an estimation of the dynamic FRET, especially if labeled ligands are used at concentrations higher than 10 nM). 3. Sample with donor and acceptor labeled ligands and an excess of unlabeled ligand (e.g.: 1 mM of unlabeled ligand).

3.4.2. Calculation of the DF%



The FRET ratio is simply calculated for each well as the ratio between the acceptor FRET signal (at 665 or 520 nm, depending on the acceptor used) and the donor emission at 620 nm: Ratio: (signal at 665  nm/signal at 620  nm) × 10,000 (see Note 13). The resulting data represents the FRET efficiency. The calculation of the DF% is based on the different FRET ratios: Ratio sample − Ratio neg ∆F % = × 100 Ratio neg As mentioned above, it is important to be very careful when c­ omparing DF % values obtained with different donor-ligand ­concentrations. If variations in the Rationeg are observed, it may not be relevant to compare DF % values. In such a case, it is preferable to compare between D665 values. FRET values are plotted against the acceptor-ligand concentration (for the FRET optimization assay) or against receptor expression/membrane quantity (for the dimerization assay).

4. Notes 1. Two cell lines have been used: CHO and COS-7 cell lines. 2. A large collection of these ligands are now commercially available from Cisbio Bioassays (Cisbio Bioassays Drug Discovery, Bagnols-sur-Cèze, France) since they are used in Tag-lite® binding assays (see www.HTRF.com). 3. We read TR-FRET signal on a Rubystar reader (BMG Labtechnologies) and on a Tecan Infinite F500 microplate reader (Tecan).

384

Cottet et al.

4. Because receptor expression is extremely variable from one receptor to another, the quantity of plasmid encoding for one receptor has to be optimized. For example, we generally use between 200 ng and 1 mg of vectors coding vasopressin V2 or dopamine D2 receptors and between 9 and 9.8 mg of empty vector. 5. Parameters of electroporation differ slightly from one cell line to another. For COS-7 cell lines, we use 270  V and 1,000 mF. 6. We recommend homogenizing the tissue little-by-little to avoid the formation of aggregates and to keep the preparation on ice. 7. The dissection of mammary gland of one lactating rat routinely gives between 10 and 40 mg of membrane protein. 8. Three parameters can vary in order to optimize the FRET signal: (a) The concentrations of the donor-ligand and the acceptorligand: the concentration of donor-ligand and ­acceptor-ligand should be high enough to label receptors. The donor/acceptor ratio has to be determined in order to get 50% of the receptors labeled with donor-ligand and 50% with acceptor-ligand. Finally the lower the concentration of donor-ligand and acceptor-ligand, the lower the nonspecific FRET signal resulting from nonspecific dynamic FRET signal (FRET resulting form random collision of donors and acceptors in suspension). (b) The amount of membrane: the FRET signal intensity varies with the number of receptors labeled with ligands. However, increasing the amount of membrane per well increases the signal to a maximum and then the signal decreases. This signal decrease can be explained by an excess of receptor amount compared to the available ligand. (c) The nature of the labeled ligand (e.g., agonist or antagonist): we observe that the TR-FRET signal intensity can vary depending on the nature of the ligand and the interaction between the receptors. 9. The following conditions have to be included in the experiments in order to analyze the results: (a) A negative control obtained in the presence of donorligand and the absence of acceptor-ligand. (b) A positive control based on the combination of donor and acceptor labeled ligands. (c) A nonspecific control obtained with an excess of ­unlabeled ligand, in the presence of donor and acceptor labeled ligands.

Time Resolved FRET Strategy with Fluorescent Ligands

385

Table 3 Example of ligand dilutions for triplicates performed in 96-well plate formats

Mock

Cells/membranes

Donor-ligand Acceptor-ligand Cold ligand 4× (e.g., 4× 4× (e.g., 4× 4× (e.g., 4× 1 nM) 1 pM to 1 mM) 1 mM)

Labeling buffer qsp 400 mL

Mock

100 mL

100 mL

–

200 mL

100 mL

–

–

300 mL

100 mL

100 mL

100 mL

100 mL

100 mL

100 mL

–

200 mL

Donor only Nonspecific

Expressing receptor

Experiment

(d) Another nonspecific control can be performed using cells transfected with an empty vector or membrane preparation from cells or tissues that do not express the receptor of interest. Such controls can replace the control performed with an excess of unlabeled ligand. We strongly recommend this control for each donor and acceptor ligand concentration since mixtures of donor and acceptor can lead to nonspecific dynamic TR-FRET signal. 10. For the calculation of volumes, the assay is performed in triplicates, with a final concentration in 100  mL/well. For an example of ligand dilutions see Table 3). 11. Since the plate readers are generally not thermostated, plates should be keep on ice as much as possible. 12. Although a general rule cannot be drawn from experiments performed on only a few types of GPCRs, we think that the design and the synthesis of fluorescent antagonists are potentially more relevant (Fig. 3). 13. The factor 10,000 just allows an acceptable scale on the Y axis.

Acknowledgements Thanks are due to Dr. L. Prezeau for his critical reading of the manuscript. This work was supported by research grants from CNRS, INSERM, ACI Molécules Cibles et Thérapeutiques (no. 240 and 355), ANR (06-Blanc-0087-03). Thanks to Plateforme de Pharmacologie-Criblage Interactome of Montpellier and the Region Languedoc-Roussillon for making this work possible.

386

Cottet et al.

Fig. 3. Optimization of the TR-FRET signal by testing various ligands. TR-FRET signal amplitude can vary depending on the nature of the ligand. As illustrated, we observed on significant variation of the FRET signal between fluorescent agonists on the one hand, and fluorescent antagonists on the other.

References 1. Milligan, G. (2004) G protein-coupled ­receptor dimerization: function and ligand pharmacology. Mol. Pharmacol. 66, 1–7. 2. Terrillon, S. and Bouvier, M. (2004) Roles of G-protein-coupled receptor dimerization. EMBO Rep 5, 30–34. 3. Gomes, I., Gupta, A., Filipovska, J., Szeto, H.H., Pintar, J.E., and Devi, L.A. (2004) A role for heterodimerization of mu and delta opiate receptors in enhancing morphine analgesia. Proc. Natl. Acad. Sci. U. S. A. 101, 5135–5139. 4. Fotiadis, D., Jastrzebska, B., Philippsen, A., Muller, D.J., Palczewski, K., and Engel, A. (2006) Structure of the rhodopsin dimer: a working model for G-protein-coupled receptors. Curr. Opin. Struct. Biol. 16, 252–259. 5. Roess, D.A., Horvat, R.D., Munnelly, H., and Barisas, B.G. (2000) Luteinizing hormone receptors are self-associated in the plasma membrane. Endocrinology 141, 4518–4523. 6. Urizar, E., Montanelli, L., Loy, T., Bonomi, M., Swillens, S., Gales, C., Bouvier, M., Smits, G., Vassart, G., and Costagliola, S. (2005) Glycoprotein hormone receptors: link between receptor homodimerization and negative cooperativity. EMBO J. 24, 1954–1964.

7. Waldhoer, M., Fong, J., Jones, R.M., Lunzer, M.M., Sharma, S.K., Kostenis, E., Portoghese, P.S., and Whistler, J.L. (2005) A heterodimer-selective agonist shows in  vivo relevance of G protein-coupled receptor dimers. Proc. Natl. Acad. Sci. U. S. A. 102, 9050–9055. 8. Wreggett, K.A. and Wells, J.W. (1995) Cooperativity manifest in the binding properties of purified cardiac muscarinic receptors. J. Biol. Chem. 270, 22488–22499. 9. Angers, S., Salahpour, A., and Bouvier, M. (2002) Dimerization: an emerging concept for G protein-coupled receptor ontogeny and function. Annu. Rev. Pharmacol. Toxicol. 42, 409–435. 10. Bazin, H., Trinquet, E., and Mathis, G. (2002) Time resolved amplification of cryptate emission: a versatile technology to trace biomolecular interactions. J. Biotechnol. 82, 233–250. 11. Albizu, L., Balestre, M.N., Breton, C., Pin, J-P., Manning, M., Mouillac, B., Barberis, C., and Durroux, T. (2006) Probing the existence of G protein-coupled receptor dimers by positive and negative ligand-dependent cooperative binding. Mol. Pharmacol. 70, 1783–1791. 12. Maurel, D., Kniazeff, J., Mathis, G., Trinquet, E., Pin, J.P., and Ansanay, H. (2004) Cell ­surface

Time Resolved FRET Strategy with Fluorescent Ligands detection of membrane protein interaction with homogeneous time-resolved fluorescence resonance energy transfer technology. Anal. Biochem. 329, 253–262. 13. Maurel, D., Comps-Agrar, L., Brock, C., Rives, M-L., Bourrier, E., Ayoub, M.A., Bazin, H., Tinel, N., Durroux, T., Prézeau, L., Trinquet, E., and Pin, J-P. (2008) Cell-surface protein-protein interaction analysis with timeresolved FRET and snap-tag technologies: application to GPCR oligomerization. Nat. Methods 5, 561–567.

387

14. Albizu, L., Cottet, M., Kralikova, M., Stoev, S., Seyer, R., Brabet, I., Roux, T., Bazin, H., Bourrier, E., Lamarque, L., Breton, C., Rives, M.L., Newman, A., Javitch, J., Trinquet, E., Manning, M., Pin, J. P., Mouillac, B. and Durroux, T. (2010) Timeresolved FRET between GPCR ligands reveals oligomers in native tissues. Nat. Chem. Biol. 6, 587–594. 15. Middleton, R.J. and Kellam, B. (2005) Fluorophore-tagged GPCR ligands. Curr. Opin. Chem. Biol. 9, 517–525.

wwwwwww

Chapter 22 Peptide Affinity Purification for the Isolation and Identification of GPCR-Associated Protein Complexes Pascal Maurice, Avais M. Daulat, and Ralf Jockers Abstract Protein networks and their dynamic regulation play a fundamental role in biological systems. Seven transmembrane-spanning G protein-coupled receptors (GPCRs) constitute the largest family of membrane receptors controlling the flow of information from the extracellular environment into cells by inducing intracellular signaling pathways. Several GPCR-associated protein complexes (GAPCs), particularly those binding to the intracellular carboxyl-terminus (C-terminus), have been identified over the last 20 years. Recent optimizations in purification protocols and advances in mass spectrometry-based protein identification techniques have considerably accelerated the identification of GAPCs. We will concentrate here on a description of the latest version of the peptide affinity purification approach dedicated to the purification of GAPCs interacting with GPCR C-termini or any other soluble receptor subdomain. Key words: G protein-coupled receptor (GPCR), Immobilized metal affinity chromatography (IMAC), Protein complexes, Proteomics

1. Introduction The superfamily of seven transmembrane G protein-coupled receptors (GPCRs) constitutes the largest family of membrane proteins with up to 800 members in humans (1). GPCRs are the most common therapeutic targets as they mediate the majority of cellular responses to hormones and neurotransmitters. They are also involved in vision, olfaction, and taste. A common theme of GPCRs is their capacity to activate heterotrimeric G proteins. More recently, an increasing number of G protein-independent signaling pathways have been discovered for GPCRs and a wide range of previously unappreciated GPCR-associated protein complexes (GAPCs) identified (2–7). The composition of these

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_22, © Springer Science+Business Media, LLC 2011

389

390

Maurice, Daulat, and Jockers

GAPCs can determine the targeting of GPCRs to a specific cellular compartment, its association with signaling or structural proteins and the fine-tuning of signal transduction through receptor desensitization and resensitization. In addition, depending on the extra- and intracellular environment and the physiological state of the cell, the composition and dynamics of GAPCs vary, thus participating in the dynamic regulation of GPCR function. Several genomic and proteomic approaches have been used to isolate and identify GAPCs (8). However, the analysis of membrane proteins, such as GPCRs and their associated protein complexes, has always been a challenge because these receptors are typically expressed at low levels, are poorly immunogenic and are difficult to solubilize into a functionally-active form. Soluble GPCR subdomains, typically the C-terminus, have been extensively used in the past to screen for GAPCs to circumvent problems associated with the hydrophobic nature of full-length GPCRs. Moreover, the GPCR C-terminus is one of the most attractive targets for identification of GAPCs, because the sequence and binding motifs within the C-terminus are specific for each GPCR. Many splice variants of GPCRs also show sequence variations within their C-terminus, and important posttranslational modifications, such as palmitoylation and phosphorylation take place within this domain. The main strategies for isolating GAPCs from receptor subdomains are based on yeast two-hybrid screening or peptide affinity purification. The yeast twohybrid approach was applied to a large number of GPCRs. However, the main limitation of this genomic approach is that only direct interactions occurring in the yeast nucleus are identified. Peptide affinity purification relies on bead-immobilized baits to isolate protein complexes from mammalian tissue lysates. Tagged-baits produced in bacteria, or chemically synthesized can be used. However, tagged-baits produced in bacteria often lead to high backgrounds. The reader is referred to the review of Daulat et  al. for the main features and limitations of currently used approaches for the identification of GAPCs (8). Here, we focus on the description of a proteomic approach that uses isolated GPCR subdomains and improved peptide affinity chromatography for the isolation of GAPCs from mouse brain lysates (see Fig.  1). This peptide affinity purification combines chemically synthesized 6×His-tagged baits with metal affinity immobilization on a Ni–NTA matrix. Advantages of this approach are low nonspecific binding due to the use of chemically synthesized baits of high purity, optimal pH conditions that preserve protein interactions and the use of a 6×His-tag (840 Da), which is much smaller than the commonly used GST tag (26  kDa). Whereas the His-tag is compatible with protein separation by 1D electrophoresis, the GST tag is not. In addition, this approach,

designed for the identification of protein complexes formed in a specific tissue, can be applied to comparisons of interactomes from different tissues and to the identification of interactomes under pathophysiological conditions, or after in  vivo pharmacological treatments. Potential limitations of this peptide affinity approach

Peptide Affinity Purification for the Isolation

391

Fig. 1. Summary scheme of the peptide affinity purification. An additional aspect of this approach is the use of ­quantitative mass spectrometry if comparing GAPCs of the same receptor subdomain from different experimental conditions (different tissues, pathophysiological states or pharmacological treatments).

392

Maurice, Daulat, and Jockers

are the non-native structure of the bait, the solubility and size of synthetic peptides that might become limiting in some cases and the loss of GAPCs whose binding depends on more than one GPCR subdomain (e.g. heterotrimeric G proteins).

2. Materials 1. Biological material to prepare fresh tissue lysates (see Note 1). 2. Lyophilized synthetic 6×His-tagged baits (see Note 2). 3. Ni–NTA-agarose beads (Qiagen). 4. Crushing buffer: 20 mM NaH2PO4, 2 mM Na3VO4, 10 mM NaF, protease inhibitors EDTA-free cocktail (Roche), pH 8.0 (see Note 3). All products, except the protease inhibitors, were from Sigma. 5. Homogenization buffer: 20 mM NaH2PO4, 2 mM Na3VO4, 10 mM NaF, protease inhibitors, EDTA-free cocktail, 10 mM CHAPS (Sigma), 150 mM NaCl (Sigma), pH 8.0 (see Note 3). 6. Washing buffer: 20 mM NaH2PO4, 2 mM Na3VO4, 10 mM NaF, protease inhibitors EDTA-free cocktail, 10 mM CHAPS, 150  mM NaCl, and 20  mM imidazole (Sigma), pH 8.0 (see Notes 3 and 4). 7. Elution buffer for quantification of eluted material: 2% SDS in phosphate-buffered saline (PBS) (see Note 5). 8. Elution buffer for 1D electrophoresis: SDS-PAGE loading buffer (62.5  mM Tris-HCL, 2% SDS, 10% glycerol, and 0.5% bromophenol blue, pH 6.8). 9. Elution buffer for 2D electrophoresis: 8 M urea, 2 M thiourea, and 4% CHAPS. All products were from Sigma. 10. Antibodies against known interacting proteins for functional validation and specificity of the peptide columns (see Note 6). 11. Equipment: Ultra-Turrax T25; spectrophotometer; highspeed, refrigerated centrifuge; shaker/roller; protein quantification kit (BCA™ Protein Assay Kit, Thermo Scientific).

3. Methods 3.1. Preparation of Brain Lysate

1. Mice are sacrificed by cervical dislocation. A dorsal cut is made in the skin to expose the skull. Using a pair of pointed scissors, dorsal cuts are made along the midline of the skull to expose the brain. The brain is then removed from cranial cavity using a spatula (see Note 7).

Peptide Affinity Purification for the Isolation

393

2. After rinsing in PBS, the mouse brain is crushed in a tube containing 15 mL of crushing buffer and using an Ultra-Turrax T25 at maximum speed at 4°C for 30 s. CHAPS (10 mM final concentration) and NaCl (150  mM final concentration) are then added and the volume is adjusted to 20 mL with a solution of 20 mM NaH2PO4, pH 8.0 (see Note 8). 3. The homogenate is solubilized for 3–5 h at 4°C under gentle end-over-end mixing. 4. After solubilization, the homogenate is centrifuged at 10,000 × g for 1 h at 4°C, and the supernatant (= brain lysate) collected. 5. Protein concentration in the supernatant is then measured. 3.2. Preparation of Peptide Columns

1. The lyophilized synthetic 6×His-peptide is dissolved at 1 mg/mL ideally in 20 mM NaH2PO4, pH 8.0 (see Note 9). 2. An aliquot is taken and the absorbance of the peptide solution is measured at 280  nm using a spectrophotometer (see Note 10). 3. After three washes with 1 mL of the same buffer used to dissolve the peptide (see Note 9) to eliminate any preservatives present in the bead stock solution, Ni–NTA agarose beads (20 mL/peptide) are equilibrated with 1 mL of the same solution for 15 min at 4°C under gentle end-over-end mixing. A control condition must be added (uncoated beads) to determine nonspecific binding of brain proteins to the beads. 4. The dissolved 6×His-peptide is then incubated with the beads for 30–90  min at 4°C under gentle end-over-end mixing (see Note 11). 5. After centrifugation (6,500 × g, 30 s, 4°C), the supernatant is collected and the absorbance is measured at 280 nm to determine the amount of 6×His-peptide immobilized on the beads (see Notes 10 and 12). 6. After three washes with 1 mL of peptide reconstitution buffer, beads are ready for use (see Note 13).

3.3. Peptide Affinity Purification

3.3.1. Optimization of the Peptide Affinity Chromatography Conditions

Before performing a large-scale purification, it is important to optimize the peptide affinity chromatography conditions (amount of protein lysate and imidazole concentrations to be added) and to verify, if possible, in a small-scale purification the recruitment of known binding proteins to validate the functionality and specificity of the peptide column. 1. Increasing concentrations (2–10  mg) of protein lysate at 0.5–1 mg/mL are incubated with uncoated beads (20 mL/ condition) in the presence of increasing concentrations

394

Maurice, Daulat, and Jockers

(0–30  mM) of imidazole overnight at 4°C under gentle end-over-end mixing. 2. Beads are then washed five times with 500  mL of washing buffer containing the same imidazole concentration. 3. Nonspecifically retained proteins are eluted in 50 mL of 2% SDS in PBS (see item 7, Subheading  2) and quantified (see Note 14). 3.3.2. Small Scale Purification: Quantification and Western Blot Detection of Known Binding Proteins

1. The optimized amount of protein lysate is incubated with uncoated (control) or bait-coated beads (20 mL/condition) in the presence of the optimized imidazole concentration overnight at 4°C under gentle end-over-end mixing (see Note 15). 2. The beads are then washed five times with 500 mL of washing buffer containing the optimized imidazole concentration. 3. Retained proteins are eluted with 50 mL of 2% SDS in PBS (95°C for 10  min) and a fraction of the sample is used for protein quantification (see Note 15). 4. The functionality and specificity of the peptide column is then validated by checking the recruitment of known binding proteins in the eluate by conventional western blotting protocols after protein separation by 1D electrophoresis (see Note 6).

3.3.3. Large Scale Purification and Mass Spectrometry Analysis

1. Retained proteins are eluted with 50 mL of elution buffer for 1D (10 min, 95°C) and/or 2D electrophoresis (5  h, room temperature). Purifications performed in parallel can be pooled together if a larger amount of protein is needed. Eluted proteins are resolved by 1D or 2D electrophoresis (see Note 16). 2. The gels are silver-stained by conventional mass spectrometry (MS)-compatible protocols. 3. Protein spots or bands are then excised from the gels for trypsin digestion and mass spectrometry analysis (see Note 17).

3.4. Discussion

Detailed knowledge about the advantages and limitations of a technique is a prerequisite for its optimal application. As previously mentioned, main limitations of the optimized peptide affinity purification approach described here are the potential loss of the native secondary structure of the bait, the solubility and size of the bait that might become limiting in some cases and the loss of GAPCs whose binding depends on more than one GPCR subdomain. Accordingly, identified interacting partners need to be confirmed in intact cells using entire GPCR to eliminate false positives. It is important to note that the approach described here constitutes one approach among others dedicated to the identification of GAPCs. These approaches are often complementary rather

Peptide Affinity Purification for the Isolation

395

than mutually exclusive (8). The following Chapter will describe another approach, the tandem affinity purification (TAP).

4. Notes 1. We typically use whole brains of C57/Bl6 mice, but any other tissue can be considered. 2. Typical GPCR-derived baits are intracellular loops, C-terminus, or specific protein motifs (i.e., PDZ-binding motifs). Note that a peptide, encompassing the C-terminal tail of the MT1 melatonin receptor (61 amino acids) has been successfully synthesized. The 6×His-tag can be introduced anywhere in the peptide sequence allowing oriented immobilization of the bait and guaranteeing homogenous coupling. We introduced the 6×His-tag at the N-terminus of the C-terminal tail of the receptor to mimic the natural situation (mobilized N-terminal part by the transmembrane domain 7 and free cytosolic C-terminus). 3. It is important to adjust all the buffers to pH 8.0. The histidine residues in the 6×His-tag have a pKa of approx. 6.0 and will become protonated if the pH is reduced (pH 4.5–5.3). Under these conditions, the 6×His-tagged bait can no longer bind to the nickel ions and will dissociate from the Ni–NTA resin during the incubation with the tissue lysate. In addition, keeping the pH at 8 is also important for preserving the integrity of protein complexes. 4. Since there is a higher potential for binding of background contaminants from tissue lysates under native conditions than under denaturing conditions, addition of low concentrations of imidazole (20–30  mM) in the washing buffer, but also during the incubation of the beads with the brain lysate is recommended. The imidazole ring is part of the structure of histidine. Imidazole binds to the nickel ions and disrupts the binding of dispersed histidine residues in non-tagged background proteins. At low imidazole concentrations, nonspecific, low affinity binding of background proteins will be prevented, while 6×His-tagged bait remains bound to the Ni–NTA matrix. 5. 2% SDS was used because of its compatibility with the protein quantification kit. Other elution buffers can be considered. According to the manufacturer, 100–250  mM imidazole or acidic pH (4.5–5.9) can be used to elute the 6×His-tagged bait and their interacting partners. Reagents, such as EDTA or EGTA, that chelate Ni2+ ions and remove them

396

Maurice, Daulat, and Jockers

from the NTA groups can also be used. Elution ­conditions are highly reproducible, but must be determined for each 6×His-tagged bait. 6. It is recommended to validate the functionality and specificity of the peptide columns by verifying the recruitment of known interacting proteins. We monitored the presence of  known interacting proteins, such as heterotrimeric G ­proteins and G protein-coupled receptor kinases (GRKs), when using the GPCR C-terminus as bait. If the 6×Histagged bait contains a PSD-95/Disc-large/ZO-1 (PDZ) domain binding motif, antibodies against PDZ domain-containing proteins should be used. 7. Ideally, fresh tissues are used, but tissues stored at −80°C can also be considered as we did not observe any apparent difference in terms of identified GAPCs. 8. One mouse brain contains 25–35 mg of soluble protein. This amount is sufficient to perform two to three purifications if using 10 mg of protein per condition. CHAPS is added after crushing to avoid detergent foaming. As cell lysates are typically acidic, it is recommended to verify that the pH is maintained at pH 8 after the addition of the homogenization buffer (see Note 3). 9. If the peptide is insoluble in phosphate buffer, denaturing agents such as urea or guanidine hydrochloride (up to 8 and 6 M, respectively) can be used according to the manufacturer’s instructions. We successfully used 6  M urea in 20  mM NaH2PO4, pH 8.0 without any interference with the binding of 6×His-tagged baits to Ni–NTA agarose beads. Once immobilized, bait-coated beads can be washed several times in buffer containing decreasing concentrations of urea to eliminate the denaturing agent before incubating with the protein lysates. Note that the immobilization of a peptide on beads tends to increase its solubility. 10. Measuring the absorbance of the 6×His-tagged bait solution at 280 nm (if aromatic residues are present) before and after immobilization on the Ni–NTA agarose beads allows quantification of the amount of bait immobilized on the beads. If the peptide does not contain aromatic residues, absorbance can be measured at 214 nm, a wavelength that detects peptide bonds. 11. According to the manufacturer, the binding capacity of the Ni–NTA agarose beads is up to 50 mg His-tagged peptide/mL of resin. In our experiments, we used bait-saturated beads and incubated 500 mg of 6×His-tagged peptide at 1 mg/mL with 20 mL of beads. After 90 min incubation, 300–350 mg was immobilized. The amount of immobilized 6×His-tagged

Peptide Affinity Purification for the Isolation

397

bait can be reduced, but this reduction will lead to an increase in the binding of nonspecific proteins from the tissue lysate. 12. If the 6×His tagged peptide does not bind to the Ni–NTA agarose beads, verify that the binding conditions are correct by checking the pH of the peptide reconstitution buffer and ensuring that there are no chelating or reducing agents present in the buffer. Another possibility is to render the 6×His-tag accessible by using denaturing conditions. If the problem persists, you should consider changing the position of the tag within the peptide. 13. We recommend using Ni–NTA agarose beads only once and not to recycle 6×His-tagged peptides to guarantee reproducible results. 14. Addition of imidazole is very important to reduce nonspecific binding of proteins on the beads (see Note 4). In the absence of imidazole, the nonspecific binding on uncoated beads is high and increases concurrently with the amount of protein lysate added. At 20 mM imidazole, the nonspecific binding is dramatically decreased irrespective of the amount of protein lysate used (between 12 and 15  mg when using 10  mg of brain protein lysate). Note that when using the coated beads, the nonspecific binding of proteins will be less due to the steric hindrance of the bait. 15. It is important to add another control condition using baitcoated beads incubated with the buffer alone. This control, processed in parallel, will be helpful in estimating the amount of bait that will be eluted from 2% SDS in PBS, and will be subtracted from the amount of GAPCs measured. 16. Several purifications can be pooled together depending on the size of the gels and the amount of sample that will be used for 1D or 2D electrophoresis. We typically use 50 mg of eluted proteins in 50 mL 1D elution buffer and 150 mg in 250 mL 2D elution buffer for MS analysis. 17. Trypsin digestion can be performed in solution from the eluate (9). However, it is important to note that peptides derived from proteins present in large quantities in the eluate could mask the MS detection of the less abundant peptides. Trypsin digestion directly on the beads is not recommended especially if the bait contains trypsin digestion sites.

Acknowledgements We thank Patty Chen (Institut Cochin, Paris) for comments on the manuscript.

398

Maurice, Daulat, and Jockers

References 1. Lagerström, M. C., and Schiöth, H. B. (2008) Structural diversity of G protein-coupled receptors and significance for drug discovery. Nat. Rev. Drug. Discov. 7, 339–357. 2. Maurice, P., Daulat, A. M., Broussard, C., Mozo, J., Clary, G., Hotellier, F., Chafey, P., Guillaume, J. L., Ferry, G., Boutin, J. A., Delagrange, P., Camoin, L., and Jockers, R. (2008) A generic approach for the purification of signaling complexes that specifically interact with the carboxyterminal domain of G ­protein-coupled receptors. Mol. Cell. Proteomics 7, 1556–1569. 3. Daulat, A. M., Maurice, P., Froment, C., Guillaume, J. L., Broussard, C., Monsarrat, B., Delagrange, P., and Jockers, R. (2007) Purification and identification of G proteincoupled receptor protein complexes under native conditions. Mol. Cell. Proteomics 6, 835–844. 4. Becamel, C., Alonso, G., Galeotti, N., Demey, E., Jouin, P., Ullmer, C., Dumuis, A., Bockaert, J., and Marin, P. (2002) Synaptic multiprotein complexes associated with 5-HT2C receptors: a proteomic approach. EMBO J. 21, 2332–2342.

5. Becamel, C., Gavarini, S., Chanrion, B., Alonso, G., Galeotti, N., Dumuis, A., Bockaert, J., and Marin, P. (2004) The serotonin 5-HT2A and 5-HT2C receptors interact with specific sets of PDZ proteins. J. Biol. Chem. 279, 20257–20266. 6. Joubert, L., Hanson, B., Barthet, G., Sebben, M., Claeysen, S., Hong, W., Marin, P., Dumuis, A., and Bockaert, J. (2004) New sorting nexin (SNX27) and NHERF specifically interact with the 5-HT4a receptor splice variant: roles in receptor targeting. J. Cell Sci. 117, 5367–5379. 7. Enz, R. (2007) The trick of the tail: proteinprotein interactions of metabotropic glutamate receptors. Bioessays 29, 60–73. 8. Daulat, A. M., Maurice, P., and Jockers, R. (2009) Recent methodological advances in the discovery of GPCR-associated protein complexes (GAPCs). Trends Pharmacol. Sci. 30, 72–78. 9. Wisniewski, J. R., Zougman, A., Nagaraj, N., and Mann, M. (2009) Universal sample preparation method for proteome analysis. Nat Methods 6, 359–62.

Chapter 23 Tandem Affinity Purification and Identification of GPCR-Associated Protein Complexes Avais M. Daulat, Pascal Maurice, and Ralf Jockers Abstract The first tandem affinity purification (TAP) protocol was described in 1999. Originally designed for the purification of protein complexes in yeast RNA splicing, its application rapidly expanded towards whole proteome analysis in yeast and mammalian cells. More recently, TAP has been applied to the purification of G protein-coupled receptor (GPCR)-associated protein complexes (GAPCs). This approach is particularly attractive for GPCRs, as the native, seven transmembrane structure is used as bait to purify GAPCs from mammalian cells expressing receptors at physiological levels. Here, a detailed protocol of the TAP method applied to GPCRs is presented. Key words: G protein-coupled receptor, Tandem affinity purification, Protein complexes, Proteomics

1. Introduction Members of the G protein-coupled receptor (GPCR) superfamily share a common and complex topology consisting of an extracellular amino-terminal domain, a hydrophobic core of seven transmembrane (7TM) a-helices that interact together to form a three-dimensional barrel within the plasma membrane, and a cytosolic carboxyl terminus (C-terminus) (1). Whereas extracellular loops and the hydrophobic 7TM core are involved in ligand binding, the intracellular domain of the receptor, composed of three loops and the C-terminus, are important for signal transmission, receptor trafficking and desensitization. All these functions are accompanied by the dynamic recruitment of different protein complexes to intracellular receptor subdomains. Some interactions rely on linear molecular determinants of one single receptor subdomain,

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_23, © Springer Science+Business Media, LLC 2011

399

400

Daulat, Maurice, and Jockers

and many others on the interaction with several different receptor subdomains. The recently described Tandem Affinity Purification (TAP) method for GPCRs appears to be appropriate for the identification of proteins interacting with multiple intracellular subdomains as native GPCRs and not isolated subdomains are used as bait (2). Further interesting ­features of the TAP approach are the preservation of the subcellular localization, post-translational modifications, quaternary structure of receptors, and the possibility of recovering protein complexes on agonist activation of receptors. Potential limitations of the TAP approach are the potential interference of the TAP-tag with receptor function, the potential loss of low-affinity interactions due to the two-step purification protocol, and the establishment of stable cell lines or transgenic mice models that might be considered to be too time consuming in some cases. The TAP method relies on the expression of the TAP-tagged target protein in the host cell. The original TAP tag consists of two immunoglobulin (IgG) binding units of protein A from Staphylococcus aureus, a cleavage site for the tobacco etch virus (TEV) protease and a calmodulin binding peptide (CBP). The protein of interest and its associated protein complexes are ­recovered in a two-step purification protocol that first involves binding to IgG-coated agarose beads and subsequent cleavage by the TEV protease at the Glu–X–X–Tyr–X–Gln/Ser consensus sequence. The eluted complex is then immobilized on calmodulincoated beads in the presence of calcium and finally eluted with EGTA. Other TAP tags have been described mostly based on ­different combinations of HA, Flag, and His tags (2). The major limitation of these tags is the use of anti-tag antibodies during the purification process that could interfere with mass spectrometry analysis. More recently, a tandem tag consisting of the streptavidin binding peptide (SBP) and the CBP gave excellent results in ­deciphering protein complexes in the Wnt/b-catenin pathway (3). This SBP tag has high affinity for streptavidin beads and the elution can be quickly performed in the presence of biotin. The TAP method, initially described in 1999, was used for large-scale purification of soluble protein complexes (4, 5). Recent advances include our adaptation of the TAP method to the specific needs of GPCRs, namely MT1 and MT2 melatonin receptors, to isolate GPCR-associated protein complexes (GAPCs) from HEK 293 cells stably expressing C-terminally TAP-tagged receptors (6). Further, successful examples are the identification of GAPCs of the N-terminally TAP-tagged a1D-adrenergic receptor (7) and the comparison of GAPCs of the wild-type and L148S mutant forms of the orphan GPR54 receptor, which is associated with idiopathic hypogonadotropic hypogonadism (8).

Tandem Affinity Purification and Identification of GPCR-Associated

401

2. Materials 1. HEK293 cells from ATCC® Number: CRL-1573™. 2. Dulbecco’s modified Eagle’s medium (DMEM), 4.5  g/L glucose, 100  U/mL penicillin, 0.1  mg/mL streptomycin, and 1 mM glutamine (Invitrogen). 3. Geneticin® antibiotic (Invitrogen). 4. FuGENE® 6 Transfection Reagent (Roche). 5. Rabbit IgG-agarose beads (Sigma-Aldrich). 6. TEV protease (Invitrogen). 7. Calmodulin-sepharose beads (Stratagene). 8. Costar® Spin-X® (Sigma-Aldrich). 9. Protease inhibitor EDTA-free cocktail (Roche). 10. Lysis buffer: 5 mM Tris/HCl, and 2 mM EDTA, pH 8.0. 11. Solubilization buffer: 10% (v/v) glycerol, 75 mM Tris/HCl, 2 mM EDTA, 5 mM MgCl2, protease inhibitor EDTA-free cocktail (Roche), 1  mM orthovanadate, 2  mM NaF, and detergent of choice, pH 8.0. All products, except the protease inhibitors, were from Sigma (see Note 1). 12. Calmodulin binding buffer: 75 mM Tris/HCl, 5 mM MgCl2, 2 mM CaCl2, and detergent of choice, pH 8.0. 13. Calmodulin rinsing buffer: 50 mM ammonium bicarbonate, 2 mM CaCl2, and detergent of choice, pH 8.0. 14. Calmodulin elution buffer: 50 mM ammonium bicarbonate, 100 mM EGTA, and detergent of choice, pH 8.0. 15. NuPAGE® Pre-Cast Gel System 4–12% Bis/Tris Gel, NuPAGE® LDS Sample Buffer (Invitrogen). 16. Modified porcine trypsin (Promega). 17. Specific antibodies against the GPCR of interest and known interacting proteins such as heterotrimeric G proteins (i.e., anti-Gai or anti-Gb antibodies, Santa Cruz Biotechnology) should be used to validate the TAP protocol. GPCR–TAP can be also detected with a goat anti-rabbit HRP conjugated antibody (Santa Cruz Biotechnology). 18. Equipment: Ultra-Turrax T25; spectrophotometer; highspeed, refrigerated centrifuge; shaker/roller; protein quantification kit (BCA™ Protein Assay Kit, Thermo Scientific), luminometer.

402

Daulat, Maurice, and Jockers

3. Methods 3.1. Generation of GPCR–TAP Constructs

We used the originally described dual affinity tag composed of two IgG binding domains from protein A and a calmodulin binding peptide (CBP) separated by the TEV protease site (9) (see Note 2). This double tag can be fused at the C-terminus or the N-terminus of the GPCR. Fusion to the N-terminus might require the addition of a signal peptide for proper surface expression (see Fig. 1). We chose to clone the GPCR–TAP construct into the pcDNA3.1 mammalian expression vector under the control of the CMV ­promoter, but other expression vectors can be used.

3.2. Verifying the Functional Expression of GPCR– TAP Constructs

To confirm the innocuousness of the tag, functional expression of the GPCR–TAP construct should be verified in transient transfection experiments. 1. Grow HEK293 cells in complete DMEM to 40–50% confluence in 10-cm Petri dishes and transfect 2 mg of vector using FuGENE® 6 Transfection Reagent according to supplier instructions (see Note 3). 2. 48 h after transfection, check the expression of the GPCR– TAP construct by lysing transiently transfected cells in solubilization buffer supplemented with your detergent of choice, or if the detergent screen has not been yet performed use as a

Fig.  1. Organization of N-terminally and C-terminally TAP-tagged GPCRs. CBD, ­calmodulin-binding domain; IgGBD, ­immunoglobulin (IgG) binding domains of protein A from Staphylococcus aureus; SS, signal peptide sequence; TEV site, cleavage site for the tobacco etch virus protease.

Tandem Affinity Purification and Identification of GPCR-Associated

403

“default” 1% NP-40. Separate the lysate by SDS–PAGE and perform western blot analysis using anti-receptor antibodies or goat anti-rabbit HRP-conjugated antibodies. 3. Monitor the correct subcellular localization of the GPCR– TAP construct by fixing transiently transfected cells with 4% paraformaldehyde and detect receptors by immunofluorescence microscopy using anti-receptor antibodies or goat antirabbit FITC-conjugated antibodies. 4. Perform functional assays (see Notes 4 and 5). 3.3. Establishing and Validating Stable GPCR–TAP Cell Lines

For those GPCR–TAP constructs whose functional properties are unchanged by the TAP-tag, stable cell clones can be established. We used HEK293 cells to establish stable cell clones (see Note 6). 1. Grow HEK293 cells in complete DMEM to 40–50% confluence in 10-cm Petri dishes and transfect 2 mg of vector using FuGENE® 6 Transfection Reagent according to supplier instructions (see Note 3). 2. Add Geneticin (0.4 mg/mL) 24 h post-transfection to select for cells harboring the pcDNA3.1 expression plasmid that contains the neomycin resistance gene. Resistant clones can be isolated by serial dilutions and the presence of the GPCR– TAP construct in the clones verified by radioligand binding or western blot. Established clones should be functionally characterized as described above.

3.4. Amplification of Cells for LargeScale Experiments

Before starting large-scale experiments, the TAP protocol should be validated in small-scale experiments (approx. 5 × 10-cm Petri dishes) to estimate the overall purification yield (see Note 7). As GPCRs are expressed at low levels compared to most other cellular proteins, a relatively large amount of cells is needed in order to perform mass spectrometric analysis. The quantity of cells required depends on the expression level of the GPCR–TAP and the purification yield. We start the expansion from one ­confluent 10-cm Petri dish and expand into two 15  cm Petri dishes followed by ten and forty 15-cm Petri dishes if necessary.

3.5. Preparation of Crude Membranes

We prepare crude membranes to enrich the GPCR-containing fraction and to remove most of the nuclear and cytosolic proteins (see Note 8). 1. HEK293 cells expressing the GPCR–TAP of interest in 15-cm Petri dishes are washed once with PBS and detached with 10  mL of PBS supplemented with 2  mM EDTA. Cells are gently transferred into 50 mL conical tubes and collected by centrifugation at 500 × g for 5 min at 4°C. 2. The pellet is resuspended and lysed in 7  mL lysis buffer ­supplemented with protease and phosphatase inhibitors.

404

Daulat, Maurice, and Jockers

Lysates are homogenized with a polytron (Ultra-Turrax T25) at maximum speed at 4°C for 2 × 15 s and the homogenate centrifuged at 48,000 × g for 45 min at 4°C. 3. The crude membrane pellet is resuspended in solubilization buffer and the protein concentration is determined. We typically adjust the protein concentration to 2  mg/mL for the solubilization step. 3.6. Solubilization of Membrane Proteins

1. This step is crucial as the choice of the detergent will ­determine the amount and quality of the GPCR–TAP and associated protein complexes that will be subsequently purified. Detergent concentration and solubilization time should also be optimized (see Note 9). We typically solubilized melatonin receptors using 2  mg/mL of membrane proteins in the ­presence of 0.5% digitonin or 0.25% Brij96V in solubilization buffer for 16 h at 4°C with gentle rocking. 2. Centrifuge the solubilizate at 20,000 × g for 45  min at 4°C and collect the supernatant (solubilized fraction).

3.7. The TAP Protocol 3.7.1. Protein A Affinity Chromatography

All manipulations must be performed at 4°C and the buffers stored on ice. 1. Equilibrate 100 mL of packed rabbit IgG-coated beads with three washes of 1 mL of solubilization buffer. Sediment beads by centrifugation at 800 × g for 1 min at each step. 2. Incubate IgG-coated beads with solubilized fraction for 6–8 h with gentle rocking. 3. Perform three washes with solubilization buffer followed by three washes with calmodulin binding buffer (CBB) (see Note 10). Spin down beads by centrifugation at 800 × g for 1 min at each step. 4. Elute the protein complexes from IgG-coated beads in 400 mL of CBB supplemented with 100 U of TEV protease overnight with gentle rocking (see Note 11). Recover the supernatant by centrifugation and wash the beads twice with 400  mL of CBB. Pool the supernatant and the two washes (TEV eluate) and filter them using a Costar® Spin-X® column by centrifugation to eliminate any remaining IgG-coated beads.

3.7.2. Calmodulin Affinity Chromatography

All manipulations must be performed at 4°C and the buffers stored on ice. 1. Equilibrate 20–50 mL of packed calmodulin–sepharose beads with three washes of 1  mL of CBB. Spin down beads by centrifugation at 800 × g for 1 min at each step. 2. Incubate the TEV eluate with calmodulin–sepharose beads in the presence of 50 mM CaCl2 for 4 h with gentle rocking.

Tandem Affinity Purification and Identification of GPCR-Associated

405

3. Perform three washes with 1  mL of CBB in presence of 50 mM of CaCl2. Spin down the beads by centrifugation at 800 × g for 1 min at each step. 3.8. Sample Preparation for Mass Spectrometric Analysis

As the presence of detergents is not compatible with mass ­spectrometry analysis, the detergent has to be eliminated from the sample at this step of the protocol. Depending on the experimental settings, different procedures, described below, can be used to separate the retained protein complexes from calmodulin– sepharose beads (see Note 12). 1. Option A:  the retained proteins can be eluted by heating the beads in NuPAGE® LDS sample buffer for 10 min at 95°C and resolved by 1D electrophoresis using NuPAGE® pre-cast gel system 4–12% (see Note 13). Gels are silver stained using conventional mass spectrometry-compatible protocols (10). Protein bands are then excised from the gel for trypsin digestion and mass spectrometry analysis. 2. Option B:  the calmodulin–sepharose beads can be washed twice with 1 mL of calmodulin rinsing buffer and the retained proteins are eluted with three times 200  mL calmodulin ­elution buffer. Then, perform a buffer exchange between detergent and urea followed by trypsin digestion as described (11). 3. Option C:  the calmodulin–sepharose beads can be washed twice with 1 mL calmodulin rinsing buffer without detergent and resuspended into 100 mL of 50 mM ammonium bicarbonate, pH 8.0. Trypsin digestion can be performed directly on beads by adding 1 mg of trypsin overnight at 37°C with agitation. The TAP approach is only one of several different approaches that can be applied to the identification of GAPCs. The reader is directed to a recent review article comparing the respective advantages and disadvantages of currently used techniques (12) and also to the previous chapter describing an improved peptide affinity purification approach recently described by Maurice et al. (13).

4. Notes 1. All buffers should be filtered with 0.44-mm filters to remove any trace of dust and keratins. 2. Other affinity tags have been described in the literature (3). Among these tags, we obtained good results for a TAP-tag composed of a streptavidin binding domain and a CBP. We first performed a streptavidin affinity chromatography, followed by elution with 50 mM biotin (pH 8.0) into calmodulin

406

Daulat, Maurice, and Jockers

binding buffer. The second purification step (calmodulin affinity chromatography) is performed as described above. 3. We typically use HEK293 cells, which are particularly well characterized for GPCR expression and signaling, but any other suitable cell-type can be used. We recommend using FuGENE® 6 Transfection Reagent, which provides moderate and homogeneous expression levels, but any other transfection reagent can be used. Expression levels should preferably be kept close to physiological levels of your protein of interest. 4. Protein expression can also be assessed by SDS–PAGE ­followed by western blotting using receptor-specific antibodies. If a radiolabeled ligand is available for the GPCR, it is recommended to determine the amount of ligand binding-competent receptor and the affinity of the receptor for the radioligand. Expression of the GPCR–TAP construct at the plasma ­membrane can be confirmed by immunofluorescence microscopy. Appropriate signal transduction assays should also be performed. Considering that most GPCR can activate the ERK pathway, we recommend testing ERK activation by the GPCR–TAP construct. Wild-type receptors should be run in parallel for each of these assays. 5. Functional expression of GPCR–TAP constructs is strongly recommended to obtain physiologically meaningful results. Mislocalization or misfolding, for example, is expected to completely change the repertoire of receptor-associated ­protein complexes. If the efficient transport of the GPCR– TAP construct to the plasma membrane is disrupted, addition of a signal sequence should be considered. If fusion of the TAP-tag to the C-terminus interferes with the recruitment of intracellular signaling proteins, transfer of the tag to the N-terminus should be considered. 6. We used HEK293 cells for the proof-of-concept, but other cell-types can be used, as long as sufficient cell quantities can be produced. We started with 1 × 109 cells expressing approximately 200–1,000 fmol of receptor/mg of protein for mass spectrometry analysis. 7. We validated the TAP procedure by monitoring the amount of GPCR–TAP and of copurified known interacting proteins during the solubilization and purification process. We followed the presence of heterotrimeric G protein by western blots using anti-Ga and Gb antibodies. If ligand binding is maintained on receptor solubilization, as in the case of ­melatonin receptors, the amount of ligand binding-competent receptors can be monitored by labeling receptors with radioligands in crude membranes followed by solubilization and receptor purification.

Tandem Affinity Purification and Identification of GPCR-Associated

407

8. Direct solubilization of cells without the preparation of crude membranes is also possible. We used crude membrane preparations as a first purification step to eliminate most nuclear and cytosolic proteins. Any kind of subcellular ­fractionation can be performed prior to purification. This might be particularly interesting for the isolation of GAPCs present in specific subcellular compartments. 9. Extraction of GPCRs from the membrane requires detergents that ideally preserve the structure of the receptor itself and the integrity of the receptor-associated protein complexes. Establishing optimal solubilization conditions should have highest priority if satisfactory results are to be obtained. Three parameters should be considered, the nature, concentration, and incubation time of the detergent. As successful solubilization depends on the protein–detergent ratio, we recommend using a maximal membrane protein concentration of 2–3 mg/mL and detergent concentrations above the critical micelle concentration (CMC). For melatonin receptors, best results were obtained with 0.5% of digitonin and 0.25% of Brij96V. We encourage the reader to extend the screen to other mild detergents, including dodecylmaltoside, CHAPS, Brij35, Brij98, lauryl dimethylamine n-oxide. Two different approaches were used to determine the solubilization efficiency of the different detergents (6). First, to quantify the amount of solubilized receptor, a GPCR-Rluc (Renilla luciferase) fusion construct was used. Cells expressing the fusion protein were incubated during different times with different detergent concentrations and solubilization yields were determined by measuring luciferase activity in the soluble ­versus nonsoluble fraction. Second, to evaluate the integrity of solubilized GPCRs, we labeled the receptor with the radioligand in crude membranes. Knowing that ligand binding to solubilized melatonin receptors requires the native receptor conformation, we determined the quantity of intact (radiolabeled) receptors at each step of the purification protocol by measuring the receptor-associated radioactivity. 10. Avoid drying of the beads in order to avoid protein precipitation on the beads. 11. TEV protease is a highly site-specific cysteine protease recognizing the Glu–X–X–Tyr–X–Gln/Ser consensus sequence, which is not present in the human proteome. This recognition specificity leads to an exclusive elution of the tagged GPCR. The elution with TEV protease is performed at 4°C, even though the optimal temperature for TEV protease activity is around 30°C. 12. The purpose of these three options is to remove the detergent for optimal mass spectrometry analysis. The major

408

Daulat, Maurice, and Jockers

l­imitation of this step might be a low protein recovery yield. We propose three options used in routine in our laboratory. Option A takes advantage of protein separation by gel ­electrophoresis to increase the mass spectrometry identification efficacy by decreasing the number of proteins analyzed simultaneously by the mass spectrometer. Option B is based on mild EGTA elution to complex Ca2+ ions that is necessary for binding to calmodulin. This procedure reduces the background, but has the lowest protein recovery rates compared to the other options. Option C also known as “shot gun approach” is based on direct analysis of the protein complex without any sample preparation. This approach yields typically high recovery rates, but the presence of high amounts of calmodulin can hinder the detection of some peptides during mass spectrometry analysis. 13. We recommend the use of a precast gel system, because of their greater resolving power and to avoid keratin contamination. It is of course also possible to use any other adequately prepared gel migration system.

Acknowledgments We thank Dr. Luc Camoin for help in establishing proteomics procedures in our laboratory and Patty Chen for help in preparing the manuscript. AMD held an EGID fellowship. References 1. Rosenbaum, D. M., Rasmussen, S. G., and Kobilka, B. K. (2009) The structure and function of G-protein-coupled receptors. Nature 459, 356–363 2. Nakatani, Y., and Ogryzko, V. (2003) Immunoaffinity purification of mammalian protein complexes. Methods Enzymol 370, 430–444 3. Angers, S., Thorpe, C. J., Biechele, T. L., Goldenberg, S. J., Zheng, N., MacCoss, M. J., and Moon, R. T. (2006) The KLHL12-Cullin-3 ubiquitin ligase negatively regulates the Wntbeta-catenin pathway by targeting Dishevelled for degradation. Nat Cell Biol 8, 348–357 4. Bouwmeester, T., Bauch, A., Ruffner, H., Angrand, P. O., Bergamini, G., Croughton, K., Cruciat, C., Eberhard, D., Gagneur, J., Ghidelli, S., Hopf, C., Huhse, B., Mangano, R., Michon, A. M., Schirle, M., Schlegl, J., Schwab, M., Stein, M. A., Bauer, A., Casari, G., Drewes, G., Gavin, A. C., Jackson, D. B.,

Joberty, G., Neubauer, G., Rick, J., Kuster, B., and Superti-Furga, G. (2004) A physical and functional map of the human TNF-alpha/ NF-kappa B signal transduction pathway. Nat Cell Biol 6, 97–105 5. Gavin, A. C., Bosche, M., Krause, R., Grandi, P., Marzioch, M., Bauer, A., Schultz, J., Rick, J. M., Michon, A. M., Cruciat, C. M., Remor, M., Hofert, C., Schelder, M., Brajenovic, M., Ruffner, H., Merino, A., Klein, K., Hudak, M., Dickson, D., Rudi, T., Gnau, V., Bauch, A., Bastuck, S., Huhse, B., Leutwein, C., Heurtier, M. A., Copley, R. R., Edelmann, A., Querfurth, E., Rybin, V., Drewes, G., Raida, M., Bouwmeester, T., Bork, P., Seraphin, B., Kuster, B., Neubauer, G., and Superti-Furga, G. (2002) Functional organization of the yeast proteome by systematic analysis of protein complexes. Nature 415, 141–147 6. Daulat, A. M., Maurice, P., Froment, C., Guillaume, J. L., Broussard, C., Monsarrat, B.,

Tandem Affinity Purification and Identification of GPCR-Associated Delagrange, P., and Jockers, R. (2007) Purification and identification of G protein-­ coupled receptor protein complexes under native conditions. Mol Cell Proteomics 6, 835–844 7. Lyssand, J. S., DeFino, M. C., Tang, X. B., Hertz, A. L., Feller, D. B., Wacker, J. L., Adams, M. E., and Hague, C. (2008) Blood pressure is regulated by an alpha1D-adrenergic receptor/dystrophin signalosome. J Biol Chem 283, 18792–18800 8. Wacker, J. L., Feller, D. B., Tang, X. B., Defino, M. C., Namkung, Y., Lyssand, J. S., Mhyre, A. J., Tan, X., Jensen, J. B., and Hague, C. (2008) Disease-causing mutation in GPR54 reveals the importance of the second intracellular loop for class A G-proteincoupled receptor function. J Biol Chem 283, 31068–31078 9. Rigaut, G., Shevchenko, A., Rutz, B., Wilm, M., Mann, M., and Seraphin, B. (1999) A generic protein purification method for protein

409

c­ omplex characterization and proteome exploration. Nat Biotechnol 17, 1030–1032 10. Rabilloud, T. (1999) Silver staining of 2-D electrophoresis gels. Methods Mol Biol 112, 297–305 11. Wisniewski, J. R., Zougman, A., Nagaraj, N., and Mann, M. (2009) Universal sample preparation method for proteome analysis. Nat Methods 6, 359–362 12. Daulat, A. M., Maurice, P., and Jockers, R. (2009) Recent methodological advances in the discovery of GPCR-associated protein complexes. Trends Pharmacol Sci 30, 72–78 13. Maurice, P., Daulat, A. M., Broussard, C., Mozo, J., Clary, G., Hotellier, F., Chafey, P., Guillaume, J. L., Ferry, G., Boutin, J. A., Delagrange, P., Camoin, L., and Jockers, R. (2008) A generic approach for the purification of signaling complexes that specifically interact with the carboxyl-terminal domain of G protein-coupled receptors. Mol Cell Proteomics 7, 1556–1569

wwwwwww

Chapter 24 Identification of GPCR Localization in Detergent Resistant Membranes Ranju Kumari and Anna Francesconi Abstract Lipid domains of the plasma membrane were originally described as a cell matrix insoluble in cold ­nonionic detergents and enriched in glycosphingolipids. Because of these biochemical properties, these membrane domains were termed detergent-resistant membranes (DRMs) or detergent-insoluble ­glycolipid-enriched (DIG) membranes. Membrane rafts and caveolae are two types of lipid domains that share these properties, as well as structural/functional dependence on membrane cholesterol. Membrane rafts and caveolae are believed to act as signaling platforms for ligand-activated receptors, thereby contributing to the regulation of receptor function. Here we describe a simple method to assess the association of GPCRs with detergent resistant membranes in native brain tissue and cultured cells. Key words: GPCR, mGluR, Membrane rafts, Caveolae, Lipid domains, Detergent resistant membranes, DRM, DIG, Density gradients, Flotation

1. Introduction Membrane rafts and caveolae are two types of membrane lipid domains that share some common biochemical and functional properties. Membrane rafts – previously termed lipid rafts – were originally defined in biochemical terms as membrane microdomains that are enriched in cholesterol, sphingomyelin, and glycolipids and are insoluble in cold nonionic detergents. These domains can be separated from other complexes formed by detergent extraction, by virtue of their buoyancy in density gradients caused by the high lipid content (1). It was proposed early that membrane rafts formed specialized domains that could include or exclude proteins depending on intrinsic affinity for the raft lipid environment, thus contributing to lateral heterogeneity in the plasma membrane. Studies prevalently based on resistance to detergent extraction, flotation properties, and sensitivity to cholesterol depletion, ­determined Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_24, © Springer Science+Business Media, LLC 2011

411

412

Kumari and Francesconi

that proteins post-­translationally ­modified by glycosylphosphatidylinositol, cholesterol, acyl or palmitoyl groups and some integral membrane proteins could associate with rafts. Caveolae (or caveolar rafts) are a subset of detergent insoluble lipid domains that can form W-shaped invaginations (2). Caveolae are believed to originate from lipid domains by cholesterol-dependent polymerization of caveolins, a family of three palmitoylated integral membrane proteins that bind cholesterol. A concept that soon emerged from the analysis of the compositions of DRMs was that membrane rafts, like caveolae, could form “signaling platforms” for ligand-activated receptors; this view was buttressed by the observed enrichment of these membrane fractions with effector proteins that participate in signal transduction (3, 4). Membrane rafts have also been shown to be involved in vesicular traffic, including endocytosis, sorting and transport of proteins and lipids to the plasma membrane and intracellular organelles (5). In neurons, they have been implicated in cell polarity, axon guidance, receptor clustering, and maintenance of dendritic spine density and morphology (6–8). Nevertheless, in recent years, the nature and even the existence of membrane rafts have undergone intense scrutiny (9, 10). Because rafts lack morphologically distinguishable features and coat proteins, they have been difficult to visualize in living cells without chemical perturbations. This discrepancy has raised legitimate doubts regarding the physiological significance of membrane rafts as originally defined. Thus, membrane rafts are now defined as “small (10–200  nm), heterogeneous, highly dynamic, sterol- and sphingolipid-enriched domains that compartmentalize cellular processes” (11). Studies employing advanced imaging techniques have recently provided evidence for the existence of such small domains in the membrane of living cells (12). Although detergent resistant membranes (DRMs) do not equate to lipid microdomains present in the membranes of living cells, they remain a useful biochemical indicator of the presumptive affinity of a protein for different membrane environments. Here, we described a simple protocol to determine the co-­fractionation of GPCRs with detergent resistant membranes enriched with known markers of membrane rafts.

2. Materials 2.1. Preparation of Samples from Brain Tissue

These procedures require dissecting tools, such as sharp scissors, (long and short blade), forceps and small spatulas. 1. Normal saline solution: 0.9% (w/v) NaCl. Saline solution is used to rinse off blood from dissected tissue. It can be sterilized by autoclaving and stored at 4°C.

Identification of GPCR Localization in Detergent Resistant Membranes

413

2. Homogenization buffer: 10 mM Tris–HCl, pH 7.4, 5 mM ethylenediamine tetraacetic acid (EDTA), 320 mM sucrose. Prepare the buffer on the day of the experiment. 3. Extraction buffer: 50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, 0.5% Triton X-100 (v/v). It can be prepared in advance and stored at 4°C for 1 week: add Triton X-100 on the day of the experiment. 4. 80% sucrose (w/v) stock solution prepared by dissolving sucrose in extraction buffer. 30% sucrose (v/v) and 5% (v/v) solutions: prepared from 80% sucrose stock diluted in extraction buffer. These solutions are prone to contamination: sterilize by filtration and, if needed, store at 4°C for £2 days. 5. Complete cocktail of eukaryotic protease inhibitors (Sigma or other vendor). 6. Conical glass tissue homogenizer (approx. clearance 0.004″). 2.2. Preparation of Samples from Cultured (HEK293) Cells

1. Reagents for culturing human embryonic kidney (HEK) 293 cells (ATCC): growth medium (Dulbecco’s Modified Eagle’s Medium, 10% fetal bovine serum, 10 U/mL penicillin/10 mg/ mL streptomycin), phosphate-buffered saline (PBS), 0.25% trypsin/1 mM EDTA solution. 2. Reagents for transfection by calcium phosphate precipitation: (a) 2.5 M CaCl2 solution: filter sterilize and store in aliquots at 4°C. (b) 2× HEPES-buffered saline (2× HeBS): 50 mM HEPES, 280  mM NaCl, and 1.5  mM Na2HPO4. For 100  mL, dissolve powders in 90  mL of H2O, adjust pH to 7.05 ± 0.5 (room temperature) with NaOH and bring to volume. Filter sterilize and store in aliquots at −20°C. 3. TNEX buffer: 50  mM Tris–HCl, pH 7.4, 150  mM NaCl, 5 mM EDTA, and 1% Triton X-100 (v/v). 4. TNE buffer: 50  mM Tris–HCl, pH 7.4, 150  mM NaCl, 5 mM EDTA. 5. 80% sucrose: 80% (w/v) stock solution prepared in TNE buffer. 30% sucrose (v/v) and 5% (v/v) solutions are prepared from 80% sucrose stock diluted in TNE. 6. Complete cocktail of eukaryotic protease inhibitors (Sigma or other vendor). 7. Teflon cell scrapers.

414

Kumari and Francesconi

2.3. Preparation of Gradients

The instructions assume the use of a Beckman Coulter ­ultracentrifuge but can be adapted to equivalent equipment. 1. Thinwall, ultra-clear, 14-mL, 14 × 95 mm, centrifuge tubes, (Beckman Coulter). 2. SW40 Ti rotor, swinging bucket, 6 × 14  mL (Beckman Coulter).

2.4. Detection of Membrane Raft Markers, Non-raft Proteins and GPCRs

Instructions assume the use of mini-gel electrophoresis and transfer systems (Mini-Protean System, BioRad Laboratories, or equivalent model). 1. Reagents for electroblot:

protein

electrophoresis

and

wet

tank

(a) Gel Stacking buffer (4×): 0.5 M Tris–HCl, pH 6.8, 0.4% (w/v) sodium dodecyl sulfate (SDS). Store at room temperature. (b) Gel Resolving buffer (4×): 1.5  M Tris–HCl, pH 8.8, 0.4% (w/v) SDS. Store at room temperature. (c) 40% (w/v) acrylamide stock solution: 37.5:1 acrylamide: bis-acrylamide (National Diagnostics). (d) 10% (w/v) ammonium persulfate, dissolved in water. Store single use aliquots at −20°C. (e) N,N,N,N ¢-Tetramethyl-ethylenediamine Store at room temperature.

(TEMED).

(f) Running buffer (10×): 0.25 M Tris base, 1.92 M glycine, 1% (w/v) SDS. Do not adjust pH. (g) Transfer buffer (10×): 0.25 M Tris base, 1.92 M glycine. pH should be ~8.3: do not adjust. Transfer buffer (1×): for 1  L, use 100  mL of 10× Transfer buffer, 200  mL methanol (20% final concentration), 700  mL of deionized water. (h) Laemmli buffer (1×): 50  mM Tris–HCl, pH 6.8, 10% (v/v) glycerol, 2% (w/v) SDS, 100  mM dithiothreitol (DTT), 0.03% (w/v) bromophenol blue. Store aliquots at −20°C. 2. TBS-T buffer: 100 mM Tris–HCl, pH 7.4, 150 mM NaCl, and 0.1% (v/v) Tween-20. 3. Blocking buffer: 5% (w/v) non-fat dry milk dissolved in TBS-T. Store aliquots at −20°C. 4. Nitrocellulose membrane (0.45 mm; Fisher Scientific or other vendor). 5. Cholera toxin B subunit-horseradish peroxidase (HRP) conjugate (Invitrogen).

Identification of GPCR Localization in Detergent Resistant Membranes

415

6. Primary antibodies: anti-transferrin receptor 1 (Invitrogen), anti-flotillin-1 (BD Biosciences), anti-caveolin-1 (BD Biosciences), anti-Thy-1 (BioLegend). Secondary antibodies: goat anti-mouse and/or anti-rabbit HRP-conjugated IgG (Jackson ImmunoResearch Laboratories or other vendor). 7. ECL Plus (GE Healthcare or equivalent from other vendor). 8. Restore Plus Western Blot Stripping Buffer (Pierce). 9. Protein Assay Kit (BioRad Laboratories). 10. (Optional). Amplex Red Cholesterol Kit (Invitrogen). The kit includes all the reagents and standards needed for quantification of free cholesterol and cholesteryl esters. The procedure also requires black, untreated 96F-microwell plates (NUNC), and fluorescence microplate reader.

3. Methods Several procedures have been described for the biochemical, large-scale isolation of membrane domains enriched with membrane rafts markers, including extraction with different nonionic detergents and detergent-free methods (13, 14). The method described here relies on the criteria of insolubility in cold Triton X-100 and flotation in density gradients to isolate membrane fractions enriched in raft markers, such as the ganglioside GM1, flotillin-1, and caveolin-1. The procedure is suitable for the analysis of total cell receptors but not for the selective analysis of receptors present in the plasma membrane. However, the protocol offers the advantage of requiring a relatively small amount of starting material, thus affording rapid analysis of samples prepared from transiently transfected cells. It is important to stress that cofractionation with DRMs does not equate to localization to lipid microdomains as they exist in living cells, but merely it indicates the presumptive affinity of a given protein for different membrane environments. 3.1. Preparation of Samples from Rodent Brain Tissue

1. Brain dissection. Animals (see Note 1) are euthanized according to approved institutional guidelines. The brain is quickly removed, transferred to a beaker containing ice-cold saline solution and rinsed three times with cold saline (~30 mL per wash) to remove excess blood. The brain is transferred to a Petri dish lined with filter paper soaked in cold homogenization buffer. Meninges are removed (this may be more difficult in younger animals due to the softness of the tissue) and the brain dissected (see Note 2). The cerebellum is removed first and the cortex is separated into two halves. (Optional: the

416

Kumari and Francesconi

hippocampal formation can be removed using small spatulas). The dissected tissue is quickly rinsed with cold homogenization buffer, drained of excess liquid, and weighed. The tissue can be flash-frozen and stored at −80°C or used immediately (see Note 3). 2. All procedures are carried out at 4°C and/or on ice with icecold reagents and tools. The dissected tissue is suspended in homogenization buffer (10% w/v), supplemented with protease inhibitors, and homogenized by 15 strokes: homogenization is performed slowly to avoid foaming. The homogenate is centrifuged at 800 × g for 10 min to pellet nuclei. The supernatant is transferred to fresh tubes and centrifuged at 10,000 × g for 15 min. 3. The pellet is dissolved in 2.2  mL of extraction buffer with protease inhibitors, sonicated on ice for 5 s (microtip, single pulse, 10% output amplitude) and incubated on ice for 10  min. An aliquot of tissue extract (~0.2  mL) should be reserved and stored at −80°C for analysis of total proteins and lipids. Protein content in the extract can be estimated by a colorimetric assay (BioRad Protein Assay kit). 4. The extract (2 mL) is adjusted to 40% sucrose (final concentration) by adding 2 mL of 80% sucrose stock solution (see Note 4). 3.2. Preparation of Samples from Cultured HEK293 Cells

Each sample is prepared from three confluent dishes of cells expressing the receptor of interest. For transfected recombinant receptors, it is important to optimize cell plating and transfection conditions to achieve the desired final cell density (see Note 5). 1. Cell plating and transfection. Day 1: cells (3–4 × 106 per plate) are seeded into three 100 mm, tissue culture-treated plates. Day 2: transfection by the calcium phosphate method (see Note 6). Day 3: transfection medium is replaced with fresh growth medium. Day 4: sample preparation. 2. All procedures are carried out on ice with ice-cold reagents and tools. Plates are transferred onto an ice bath and rinsed twice with ~5 mL PBS. Cells are harvested with PBS (~4 mL per plate), by lifting them off with a cell scraper, and transferred to a centrifuge tube. 3. Cells are centrifuged at 500 × g for 5  min: the pellet is dissolved in 2.2  mL TNEX buffer containing eukaryotic protease inhibitors.

Identification of GPCR Localization in Detergent Resistant Membranes

417

4. Cells are sonicated on ice for 5 s (microtip, single pulse, 10% output amplitude) and incubated on ice for 10  min (see Note  7). An aliquot of cell extract (~0.2  mL) should be reserved and stored at −80°C. Protein content in the extract can be estimated by colorimetric assay. 5. The cell extract (2 mL) is adjusted to 40% sucrose (final concentration) by adding 2  mL of 80% sucrose stock solution made in TNE (see Note 4). 3.3. Preparation of Gradients

1. Samples are transferred to ultracentrifuge tubes and overlaid with 6 mL of 30% sucrose solution followed by 3 mL of 5% sucrose solution (see Note 8). Samples are centrifuged at 36,000 rpm (230,000 RCF) for 16 h (4°C) with a SW40 Ti rotor. 2. After centrifugation, the tubes are removed from the rotor and placed on a stand: it is important to proceed carefully at this stage to prevent disturbing the gradients. Detergent resistant membranes appear as cloudy, opaque material floating in the upper third of the gradient (Fig.  1a). Thirteen, 1 mL fractions are collected starting from the top of the gradient (see Note 9). Each fraction is thoroughly mixed by pipetting three to four times (see Note 10).

Fig. 1. Isolation of detergent resistant membranes enriched in membrane raft markers. (a) Illustration of the expected distribution of buoyant, detergent resistant membranes (DRMs) in discontinuous sucrose density gradients: 13 fractions are collected starting from the top of the gradients. (b, c) DRMs are enriched in lipids and proteins that associate with membrane rafts: shown are preparations from (b) mouse brain and (c) HEK293 cells. Transferrin receptor 1 (TFR1), which does not associate with membrane rafts, is excluded from DRMs. The distribution of GM1 in gradient fractions is monitored by ­cholera toxin B (CTB) binding. FLOT1 flotillin-1, CAV1 caveolin-1.

418

Kumari and Francesconi

3.4. Detection of Membrane Raft Markers, Non-raft Proteins, and GPCRs

All procedures are carried out at room temperature unless ­ therwise noted. o

3.4.1. Detection of GM1

1. A strip of nitrocellulose membrane (approximately 2.5 × 9 cm) is equilibrated in TBS-T, placed on clean Parafilm, and allowed to dry for ~5 min. 2. One to 2 mL of each gradient fraction are spotted in triplicate onto the nitrocellulose membrane strip. The membrane is allowed to dry (5–10 min) to facilitate sample adsorption and binding. 3. The membrane is incubated with blocking buffer for 30 min with constant agitation. After pouring out the blocking ­buffer, the membrane is incubated for 1  h with cholera toxin B ­subunit-HRP conjugate diluted in blocking buffer (see Note 11). 4. The membrane is washed by three changes (15 min each) of TBS-T and incubated with ECL Plus (or other ­chemilumiscent/ fluorescent substrate) as per manufacturer’s instructions. Luminescent signal is detected by exposing membranes to autoradiography film or with an imaging station (Kodak 2000R or equivalent).

3.4.2. Detection of Proteins Associated with DRMs

1. 25 mL of each gradient fraction and 10 mL of total extract is loaded on acrylamide gels prepared in advance, and resolved by SDS–PAGE (see Note 12). 2. Proteins are transferred onto nitrocellulose membranes by wet tank electroblotting (see Note 13). 3. After electroblot, the membranes are incubated with blocking solution for 30 min, followed by incubation with primary antibodies (diluted in blocking buffer) at 4°C for 16 h or at room temperature for 1 h. Antibodies raised against flotillin-1, caveolin-1, or Thy-1 are used to identify DRMs (Fig. 1b, c); antibodies raised against transferrin receptor 1 are used to identify gradient fractions containing proteins that are excluded from DRMs (Fig. 1b, c; see Note 14). 4. Membranes are washed with four changes of TBS-T (15 min per wash) and incubated with secondary antibody conjugated to HRP (diluted in blocking buffer) for 1 h. After washing four times with TBS-T, membranes are incubated with ECL Plus substrate and exposed for signal detection (see above). 5. Expected results and interpretation. DRMs are identified by enrichment with GM1, flotillin-1, caveolin-1, Thy-1 and absence of transferrin receptor 1 (Fig. 1b, c). GM1 is normally enriched in fractions 3–5 of the gradient; Thy-1 is enriched in

Identification of GPCR Localization in Detergent Resistant Membranes

419

fractions 3–4 (lighter membranes) whereas flotillin-1 and caveolin-1 are enriched in fractions 3–5 (see Note  15). Transferrin receptor 1, which is excluded from DRMs, is enriched in fractions 10–13. Co-fractionation of GPCRs with membrane raft markers is indicative of the ­propensity of a given receptor to associate with detergent resistant membranes and potentially membrane rafts (see Note 16). In the case of group I mGluRs, a subset of receptors co-fractionates with buoyant detergent resistant membranes in both transfected HEK293 cells and brain tissue, indicating potential association with membrane rafts in vivo (15). 6. (Optional: Detection of cholesterol by fluorometric enzymatic assay). All the necessary reagents, provided with the Cholesterol Amplex Red kit, are reconstituted at working concentration as directed. The reaction mix contains 300 mM Amplex Red, 2 U/mL HRP, 2 U/mL cholesterol oxidase, 0.2 U/mL cholesterol esterase. Serial dilutions of cholesterol (0–0.8 ng/mL) and resorufin (0–20 mM) standards are prepared in sufficient amount for triplicate determinations (150 mL). Samples are prepared in triplicate: for each measurement, 5  mL of each gradient fraction is added to 45  mL of reaction buffer. Standards and samples (50  mL) are transferred into a 96F-microwell plate. The reaction mix (50 mL) is added to each sample and incubated at 37°C for 30  min in the dark (reagents are light sensitive). Fluorescence is measured with a microplate reader at 560–590 nm Ex-Em.

4. Notes 1. Animals. We routinely use male mice; the age chosen for experimental animals will clearly depend on the specific research objectives. However, we have observed consistent age-dependent differences in the relative enrichment of mGluRs (and mGluR-associated proteins) in DRMs (Kumari and Francesconi, unpublished). Age-dependent changes in the composition of DRMs have been previously reported (16). The potential developmental regulation of receptor localization in DRMs should be taken into consideration when interpreting results from native tissue. 2. Brain tissue. We routinely use brain cortex: in general, after microdissection, ~0.15–0.2  g (wet weight) of tissue is obtained per adult mouse. Detergent extraction and gradient conditions are optimized for cortical tissue from one to two adult mice (2–8 month-old). For neonatal animals, or for

420

Kumari and Francesconi

­ ifferent brain regions (i.e., hippocampus), the tissue from d several animals can be pooled. 3. Frozen tissue. Many studies have reported analysis of DRM composition from frozen brain, including post-mortem human tissue. We found that, for the raft markers and receptors examined (15), the relative enrichment in DRM does not significantly differ between fresh vs. frozen tissue. However, we have observed variability in the recovery in DRMs of one mGluR-binding protein (Kumari and Francesconi, unpublished observation). We therefore suggest the comparison of samples prepared from fresh vs. frozen tissue, when first examining the presence of novel proteins in DRMs. 4. Sucrose should be thoroughly mixed by pipetting up and down several times while avoiding the formation of foam. Do not vortex. 5. Instructions are optimized for ~6–8 mg total proteins in the extract. Under these conditions, the approximate detergent: protein ratio is ~3–2.5 to 1 (17). 6. Transfection. Cells are fed with 4.5  mL of fresh growth medium (without antibiotics) and returned to 37°C for ~2 h. Allow the CaCl2 and HeBS solutions to equilibrate at room temperature for at least 30  min. For each plate, 12.5  mg of plasmid DNA is diluted in water (0.2 mL final volume): CaCl2 solution (50 mL) is added to the DNA and mixed by gentle vortexing. The DNA/CaCl2 mix is added dropwise to 0.25 mL of 2× HeBS while vortexing (~800  rpm). The DNA mix is incubated at 37°C (in a prewarmed water bath) for 2 min and added dropwise to the cells while gently swirling the plate. 7. If desired, cells can be homogenized before sonication, as described for brain tissue. However, we have not observed a significant increase in the relative amount of material recovered in DRMs using homogenization. 8. Sucrose solutions are added by slowly pouring them along the side of the centrifuge tube. 9. Collect fractions with a slow circular movement to ensure recovery of material that may adhere to the walls of the tubes. 10. Gradient fractions can be analyzed immediately or stored frozen. To avoid frequent freeze/thaw cycles, gradient fractions can be dispensed into aliquots and stored at −80°C for several months. 11. GM1 is detected by bound cholera toxin B subunit. The sensitivity of this method depends on the relative abundance of endogenous GM1, which can vary between cell types. For brain tissue samples, in which GM1 is most abundant, cholera toxin B subunit-HRP is used at 10  ng/mL; for HEK293 cells, use at 40 ng/mL.

Identification of GPCR Localization in Detergent Resistant Membranes

421

12. For two 10% acrylamide gels, prepare 15  mL of acrylamide solution (7.5 mL per gel) by adding 3.75 mL of acrylamide stock solution, 3.75 mL of Resolving buffer (4×) to 7.5 mL of water: mix, add 0.05 mL of ammonium persulfate and 0.01 mL TEMED. Pour immediately in casted gel glasses and allow to polymerize at room temperature (³30  min): overlay with water-saturated butanol. Prepare 5 mL of stacking acrylamide solution by adding 0.49 mL of 40% acrylamide stock solution, 1.25 mL of Stacking buffer (4×) to 3.26 mL of water: mix, add 0.025  mL of ammonium persulfate and 0.005  mL TEMED. Rinse off butanol with water, pour the stacking gel, insert combs and let polymerize. Run gels in 1× Running buffer at 100 V for approximately 1 h. Using a 1.5 mm-thick gel with a 15-well comb affords simultaneous examination of all the fractions. A 10% acrylamide concentration is suitable for fractionation of flotillin-1 (~48 kDa), ­caveolin-1 (~24 kDa), Thy-1 (~18  kDa) and transferrin receptor 1 (~85  kDa). Depending on the molecular mass of the receptor under study, the gel concentration may need to be adjusted. 13. For electroblot, equilibrate gels in Transfer buffer (1×) for ~30 min at room temperature. For each gel, submerge two pieces (7.5 × 8.5 cm), of 3MM paper, one pre-cut nitrocellulose membrane (7.5 × 8.5 cm) and two fiber pads in Transfer buffer. Assemble a gel sandwich in the gel holder cassette by overlaying, in this order, first fiber pad, one piece of filter paper, gel, membrane, filter paper, and fiber pad. Close the gel holder tightly, transfer to electroblot apparatus and submerge with transfer buffer. Transfer at 400 mA for 1–2 h at 4°C with constant buffer mixing with a stir bar. 14. Antibodies to flotillin-1 and caveolin-1 are suitable to identify DRMs in both brain tissue and HEK293 cell samples, whereas antibodies to Thy-1 are used for brain samples. Antibodies against transferrin receptor 1 work equally well for both brain tissue samples and HEK293 cells. Membranes can be sequentially probed (up to three times) with different antibodies after removing bound IgGs. For this, membranes are incubated with Restore Plus Stripping Buffer for 5 min at room temperature and rinsed three times with TBS-T before applying the blocking buffer again. 15. Flotillin-1 and caveolin-1, although enriched in DRMs, are also present to some extent in the heavy fraction of the gradient. 16. Controls and general considerations. The cholesterol dependence (one characteristic of membrane rafts) of co-­fractionation with DRMs, can be tested by examining the composition of gradient fractions after acute cholesterol depletion. This is achieved by incubating cultured cells with the ­cholesterol-binding

422

Kumari and Francesconi

agent methyl-b-cyclodextrin (mbCD; 5 mM, 60 min at 37°C), which sequesters cholesterol from the plasma membrane. Cells incubated with either mbCD alone or with mbCD followed by cholesterol replenishing medium (DMEM containing 2.5 mM mbCD, 0.25 mM cholesterol; 60 min, 37°C) are used to comparatively examine the composition of DRMs. Another consideration is that the association of receptors with lipid domains can be affected by stimulation with agonist (18). While the association with DRMs can only be examined at steady state in native tissue, in cultured cells it is possible to test the impact of agonists on receptor recruitment to DRMs. In some cases, the cell culture medium itself may contain endogenous agonists, for example glutamate. In this situation, cells can be incubated with reagents that effectively lower the concentration of free glutamate in the medium (19).

Acknowledgments Supported by the U.S. National Institute of Health/NIMH (MH082870 to A.F.) References 1. Simons, K. and Ikonen, E. (1997) Functional rafts in cell membranes. Nature 387, 569–572. 2. Parton, R.G. and Simons, K. (2007) The multiple faces of caveolae. Nat. Rev. Mol. Cell Biol. 8, 185–194. 3. Simons, K. and Toomre, D. (2000) Lipid rafts and signal transduction. Nat. Rev. Mol. Cell Biol. 1, 31–39. 4. Allen, J.A., Halverson-Tamboli, R.A., and Rasenick, M.M. (2007) Lipid raft microdomains and neurotransmitter signalling. Nat. Rev. Neurosci. 8, 128–140. 5. Hanzal-Bayer, M.F. and Hancock, J.F. (2007) Lipid rafts and membrane traffic. FEBS Lett. 581, 2098–2104. 6. Tsui-Pierchala, B.A., Encinas, M., Milbrandt, J., and Johnson, E.M., Jr. (2002) Lipid rafts in neuronal signaling and function. Trends Neurosci. 25, 412–417. 7. Hering, H., Lin, C.C., and Sheng, M. (2003) Lipid rafts in the maintenance of synapses, dendritic spines, and surface AMPA receptor stability. J. Neurosci. 23, 3262–3271.

8. Kamiguchi, H. (2006) The region-specific activities of lipid rafts during axon growth and guidance. J. Neurochem. 98, 330–335. 9. Munro, S. (2003) Lipid rafts: elusive or illusive? Cell 115, 377–388. 10. Nichols, B. (2005) Cell biology: without a raft. Nature 436, 638–639. 11. Hancock, J.F. (2006) Lipid rafts: contentious only from simplistic standpoints. Nat. Rev. Mol. Cell Biol. 7, 456–462. 12. Eggeling, C., Ringemann, C., Medda, R., Schwarzmann, G., Sandhoff, K., Polyakova, S., Belov, V.N., Hein, B., von Middendorff, C., Schonle, A., and Hell, S.W. (2009) Direct observation of the nanoscale dynamics of membrane lipids in a living cell. Nature 457, 1159–1162. 13. Ostrom, R.S., and Liu, X. (2007) Detergent and detergent-free methods to define lipid rafts and caveolae. Methods Mol. Biol. 400, 459–468. 14. Macdonald, J.L., and Pike, L.J. (2005) A simplified method for the preparation of ­detergent-free lipid rafts. J. Lipid Res. 46, 1061–1067. 15. Francesconi, A., Kumari, R., and Zukin, R.S. (2009) Regulation of group I metabotropic

Identification of GPCR Localization in Detergent Resistant Membranes glutamate receptor trafficking and signaling by the caveolar/lipid raft pathway. J. Neurosci. 29, 3590–3602. 16. Jiang, L., Fang, J., Moore, D.S., Gogichaeva, N.V., Galeva, N.A., Michaelis, M.L., and Zaidi, A. (2008) Age-associated changes in synaptic lipid raft proteins revealed by two-dimensional fluorescence difference gel electrophoresis. Neurobiol. Aging, doi: 10.1016/j. neurobiolaging.2008.11.005 17. Ostermeyer, A.G., Beckrich, B.T., Ivarson, K.A., Grove, K.E., and Brown, D.A. (1999) Glycosphingolipids are not essential for formation of detergent-resistant membrane rafts in

423

melanoma cells. methyl-beta-cyclodextrin does not affect cell surface transport of a ­GPI-anchored protein. J. Biol. Chem. 274, 34459–34466. 18. Kong, M.M., Hasbi, A., Mattocks, M., Fan, T., O’Dowd, B.F., and George, S.R. (2007) Regulation of D1 dopamine receptor ­trafficking and signaling by caveolin-1. Mol. Pharmacol. 72, 1157–1170. 19. Francesconi, A., and Duvoisin, R.M. (1998) Role of the second and third intracellular loops of metabotropic glutamate receptors in mediating dual signal transduction activation. J. Biol. Chem. 273, 5615–5624.

wwwwwww

Chapter 25 Analysis of GPCR Localization and Trafficking James N. Hislop and Mark von Zastrow Abstract Localization and trafficking of G protein-coupled receptors (GPCRs) is increasingly recognized to play a fundamental role in receptor-mediated signaling and its regulation. Individual receptors, including closely homologous subtypes with otherwise similar functional properties, can differ considerably in their membrane trafficking properties. In this chapter, we describe several approaches for experimentally assessing the subcellular localization and trafficking of selected GPCRs. Firstly, we describe a flexible method for receptor localization using fluorescence microscopy. We then describe two complementary approaches, using fluorescence flow cytometry and surface biotinylation, for examining receptor internalization and trafficking in the endocytic pathway. Key words: Microscopy, Receptor endocytosis, Flow cytometry, Biotinylation, Receptor trafficking

1. Introduction The subcellular localization of receptors can impact fundamentally on the strength and specificity of cellular signaling. Individual GPCRs can differ considerably in their localization between the plasma membrane and various intracellular compartments, as well as between different domains of the plasma membrane. In many cases, this represents a steady state, reflecting dynamic trafficking pathways that are subject to physiological regulation (1, 2). Clinically relevant drugs influence the number or subcellular distribution of receptors in target tissues and, for some drugs, effects on receptor localization are thought to represent the primary therapeutic mechanism (3). Accordingly, it is of considerable interest to have experimental methods suitable for assessing receptor localization and trafficking.

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_25, © Springer Science+Business Media, LLC 2011

425

426

Hislop and von Zastrow

In this chapter, we discuss several approaches for assessing the subcellular distribution and trafficking of receptors, based on our experience with several members of the GPCR family. We focus on immunochemical methods that require receptor-specific antibodies or the expression of mutant receptors fused to an epitope-tag. These methods complement pharmacological and cell fractionation assays used traditionally to assess receptor distribution, and can be applied both to cell culture and intact tissue preparations. Each of the methods described is highly flexible, and can be modified to accommodate particular experimental questions and constraints. They are also flexible with regard to the deployment of newer experimental technologies, such as engineered fluorescent proteins and covalent protein tagging. 1.1. Visualization of Subcellular Localization of GPCRs

How does one begin to investigate the localization and trafficking fate of a GPCR? An obvious starting point is to visualize the subcellular distribution of receptors using immunocytochemical staining and microscopy. A major limitation in many cases has been the relative dearth of antibodies that are capable of specifically detecting receptors of interest when expressed at their (typically low) native levels. The widespread availability of recombinant receptors, as well as application of methodologies such as epitope and fluorescent protein tagging (discussed elsewhere in this volume), has greatly facilitated progress in this area. There is increasing application of enzyme-directed covalent modification (e.g., SNAP-tag) as an alternate means of receptor labeling. Of course one must consider, and control for, the possibility that recombinant receptors may not faithfully mimic native receptor properties. Tagged, recombinant receptors are typically amenable to conventional methods of chemical fixation, immunocytochemical staining, and microscopic imaging. Generally receptor localization is carried out using fluorescence microscopy, after staining tagged receptors with a fluorescent antibody conjugate. Receptors tagged with an intrinsically fluorescent protein [such as green fluorescent protein (GFP)] can be visualized directly, but may lose fluorescence intensity after fixation. Anti-GFP antibodies and conventional fluorescent conjugates are often used to augment detection if this is a problem. Fluorescent staining methods are generally rapid and amenable to imaging the localization of multiple receptors and/or other proteins in the same preparation, taking advantage of the availability of a wide range of spectrally resolved fluorescent conjugates and suitable optics for multicolor acquisition. Conventional fluorescence imaging methods are intrinsically limited in their ability to resolve small structures by Abbe’s law, which specifies that the ability of light to resolve objects is directly proportional to the illuminating wavelength (4). Even with the highest quality microscope and objective lens, the resolution limit of epifluorescence or confocal fluorescence microscopy is on the order of ~300  nm, a usable

Analysis of GPCR Localization and Trafficking

427

range for many (but not all) questions of receptor localization. Much higher spatial resolution (on the order of a few nanometer) can be achieved using electron microscopy. This is because the effective wavelength of an electron beam is considerably smaller than that of light. Immunoelectron microscopy is similar in principle but involves additional sample preparation and reagents, description of which is beyond the scope of this chapter. There are several patterns of subcellular localization typically described for GPCRs, and recognizing these can be a useful first step in inferring the basic membrane trafficking properties of receptors. Many GPCRs are localized primarily in the plasma membrane, particularly in cells not exposed to an agonist. Other GPCRs may be localized primarily in the endoplasmic reticulum (5) or subcompartments of the Golgi apparatus (6), or may be found in endosomes at steady state (7). The addition of agonist for between 5 and 120  min often causes a redistribution of receptors from the plasma membrane to a more punctate distribution within the interior of the cell, reflecting the occurrence of ligand-induced endocytosis. Co-staining with various markers can help determine these localization patterns. For receptors that undergo ligand-induced internalization, subcellular localization at various times after inducing receptor endocytosis can provide information about the subsequent trafficking fate of receptors in the endocytic pathway. Receptor localization in the plasma membrane is often obvious in cultured cells, based on a smooth peripheral pattern of receptor immunoreactivity. This can be verified by establishing surface accessibility of receptors, or by colocalization with fluorescent concanavalin A (a lectin that labels surface glycoproteins), or the Na+, K+ ATPase that is localized primarily in the plasma membrane of most cell types. A number of markers have been established that help the investigator to distinguish different types of internal membrane compartments relevant to GPCR trafficking. These include fluorescent transferrin and transferrin receptor antibodies (these mark early endosomes and the major recycling pathway), EEA1 (concentrated on early endosomes), LAMP1 and 2 (concentrated on late endosomes and lysosomes), BiP and calnexin (localized to endoplasmic reticulum), galactosyl transferase (Golgi apparatus), and TGN38 (trans-Golgi network). Specific antibodies or fluorescent conjugates recognizing these proteins are available commercially, and colocalization relative to GPCRs is practical using the multilabeling flexibility of fluorescence microscopy. 1.2. Use of Flow Cytometry to Determine CellSurface Expression Levels of GPCR

Although microscopic visualization offers a quick and generally easy method to estimate the predominant subcellular localization of GPCRs, and to infer basic aspects of receptor trafficking, this method is not easily quantifiable. As such, it is difficult to determine absolute receptor numbers or precisely assess redistribution

428

Hislop and von Zastrow

of receptors. We routinely use fluorescence flow cytometry as a more quantifiable tool, which is particularly useful for assessing receptor internalization and recycling. In the simplest application, a fluorescent antibody conjugate is used to detect surface-accessible receptors before and after the exposure of cells to agonist, and then again after agonist removal. Changes in surface-accessible receptor immunoreactivity are then measured as a function of time. This allows one to use kinetic models to estimate rates of receptor internalization and recycling (8). Flow cytometric assays are similar in principle to cell surface ELISA, but we generally find that flow cytometry has a wider linear range of detection. It is possible to calibrate flow cytometry methods to allow determination of absolute receptor number, although this method is typically used to assess relative changes. 1.3. Use of BiotinProtection, Degradation Assay to Measure Receptor Stability

Changes in receptor localization and surface receptor immunoreactivity are useful in assessing rapid trafficking processes, but such changes do not provide direct information about generally slower processes such as receptor biosynthesis or proteolytic destruction. A traditional method for inferring receptor biosynthesis and proteolysis is via radioligand binding assay (discussed elsewhere in this volume). Radioligand binding remains an extremely useful technique but is limited to receptors for which suitable ligands (usually high-affinity antagonists) are available with appropriate specificity and in radiolabeled form. Also, as radioligand binding reports changes in net receptor number (e.g., changes in Bmax estimated from saturation isotherms), it can be challenging to distinguish the effects of receptor biosynthesis from degradation using this technique. This is a particular concern in transfected cell systems, and for those receptors whose down-regulation is slow, as significant changes in receptor destruction can be effectively masked by the high level of biosynthesis driven by nonendogenous promoters typically used in these systems. Pulse-chase metabolic labeling and immunoprecipitation methods, described in detail elsewhere, are useful for investigating receptor biosynthesis. We have found nonradioactive methods, based on chemical modification of receptors by surface biotinylation, advantageous for selectively assessing endocytic trafficking, and proteolysis. Using a variation in which existing surface receptors are rendered nonreactive, the method can also be used to assess delivery of receptors to the plasma membrane from internal pools (6). Chemically “cleavable” biotinylation reagents can be used to allow selective labeling of internalized receptors, even if this pool is minor quantitatively (9, 10). The particular method described below (in Subheading  3.3) incorporates this principle.

Analysis of GPCR Localization and Trafficking

429

2. Materials 2.1. Cell Fixation Materials 2.1.1. Complete Cell Culture Medium

1. Dulbecco’s modified Eagle medium (DMEM). 2. 10% fetal bovine serum. 3. 100 U/mL penicillin and 100 mg/mL streptomycin. 4. 4 mM l-glutamine. 5. 25 mM glucose.

2.1.2. Poly-l-Lysine solution

1. Sterile H2O. 2. Poly-l-lysine solution (0.1% w/v in H2O). Make a 1/100 dilution of the poly-l-lysine solution in sterile H2O.

2.1.3. Phosphate-Buffered Saline

1. 135 mM NaCl. 2. 15 mM Na2HPO4·7H2O. 3. 2.68 mM KCl. 4. 1.47 mM KH2PO4. 5. 0.68 mM CaCl2. 6. 0.49 mM MgCl2·6H2O.

2.1.4. Phosphate-Buffered Saline/EDTA

1. 135 mM NaCl. 2. 15 mM Na2HPO4·7H2O. 3. 2.68 mM KCl. 4. 1.47 mM KH2PO4. 5. 1 mM EDTA.

2.1.5. Fixation Buffer

1. Phosphate-buffered saline (PBS). 2. 4% Formaldehyde (made from paraformaldehyde). Dissolve 4  g of paraformaldehyde in 100  mL PBS, heating gently (to about 65°C with stirring) in a fume-hood. Be careful because formaldehyde fumes can be dangerous, particularly if inhaled or exposed to the eyes. Add 2–4 drops of 1 M NaOH to help the paraformaldehyde dissolve. Chill on ice, aliquot and store at −20°C. In some cases, the PBS can be replaced with a PIPES-buffered solution, or supplemented with an osmotic substance such as 300  mM sucrose. These modifications can help in maintaining the morphology of some subcellular compartments.

430

Hislop and von Zastrow

2.1.6. Permeabilization Buffer

1. PBS. 2. 1% bovine serum albumin. 3. 0.5% Triton X-100 (or Nonidet P40).

2.1.7. Quench/Wash Buffer Tris-Buffered Saline

1. 50 mM Tris base. 2. 275 mM NaCl. 3. 6 mM KCl. 4. 2 mM g/L CaCl2. Adjust pH to 7.4 with HCl.

2.2. Flow Cytometry Materials 2.2.1. Complete Cell Culture Medium 2.2.2. PBS 2.2.3. Primary/Secondary Antibody Solutions

    1. Antibody (appropriate concentration; we typically use M1 anti-Flag at 1 mg/mL). 2. PBS. 3. 0.1% BSA.

2.2.4. Fixation Solution

1. PBS. 2. 1% Formaldehyde.

2.3. Biotinylation Materials

  

2.3.1. Poly-Lysine Solution 2.3.2. PBS 2.3.3. 100× Biotinylation Solution

1. 30 mg/mL cleavable biotin (EZ-link NHS-SS-biotin, Thermo Scientific). 2. DMSO.

2.3.4. 1× Biotinylation Solution

1. PBS. 2. 100× Biotinylation Solution. Dilute the 100× biotinylation solution in PBS just prior to addition to cells.

2.3.5. Tris-Buffered Saline

1. 50 mM Tris base. 2. 275 mM NaCl. 3. 6 mM KCl. 4. 2 mM g/L CaCl2.

Analysis of GPCR Localization and Trafficking

431

Adjust pH to 7.4 with HCl. 2.3.6. Complete Cell Culture Medium

1. DMEM. 2. 10% fetal bovine serum. 3. Penicillin and streptomycin.

2.3.7. Stripping Buffer-1

1. 50 mM Tris–HCl, pH 8.8. 2. 100 mM NaCl. 3. 1 mM EDTA. 4. 1% BSA. 5. 100 mM MESNA.

2.3.8. Stripping Buffer-2

1. 5 mM glutathione. 2. 75 mM NaCl. 3. 10% fetal bovine serum. 4. pH with 7.5 mL of 1 M NaOH.

2.3.9. Quench Buffer

1. 100 mL PBS. 2. 1 g iodoacetamide. 3. 1 g BSA.

2.3.10. Lysis Buffer: 10 mL

1. 250 mM NaCl. 2. 50 mM Tris/HCl, pH 7.4. 3. 5 mM EDTA. 4. 0.5% Triton X-100. 5. 10 g/L iodoacetamide. 6. 1× Complete mini protease inhibitor cocktail tablet (Roche). Add the iodoacetamide and inhibitor cocktail tablet fresh.

3. Methods 3.1. Cell Fixation and Staining

For the purposes of this chapter, it is assumed that your GPCR of interest is expressed in HEK293 cells; however, the general principles remain the same for whichever cell line/tissue you are using. Stable cell-lines expressing your GPCR should be maintained in complete cell culture medium until the time of experiment. Coverslips should be sterilized and placed into appropriate cell culture dishes; typically we use 12-well plates and 18-mm diameter glass coverslips. The use of poly-l-lysine provides a charged surface on the glass, and allows significantly greater attachment of cells. This becomes important with the large

432

Hislop and von Zastrow

number of washes required and ensures that there will still be cells attached to the coverslips at the end of the protocol. 1. Add 1  mL of poly-l-Lysine solution to the wells to completely cover the coverslip. 2. Leave for 10 min. 3. Aspirate off the poly-l-lysine solution and allow coverslips to air-dry for 1 h. 4. Remove the medium from the cells and lift with PBS/ EDTA. 5. Return to complete cell culture medium and replate on to coverslips so they will be ~80% confluent on the day of experimentation, typically 24–48 h after plating. The conditions of the experiment itself will vary depending on the specific question being asked, but the principles for fixation and staining remain the same as for a simple endocytosis experiment. Cells should be left untreated, others should have agonist added to the culture medium for 10, 30, and 90 min to measure endocytosis. Following endocytosis, the cells may be washed and returned to complete medium in the presence of an antagonist to allow the receptors to recycle. Following agonist treatments it is important to rapidly cool the cells to prevent further trafficking (which is essentially zero at 4°C) and then fix the cells. A number of different fixatives can be used, depending on the particular antigenic properties of the protein you wish to stain, but the most commonly used is formaldehyde. Formaldehyde is a small membrane permeable fixative that forms methylene crosslinks with various parts of the cellular proteins and maintains a lot of the cellular architecture. Organic solvents, such as ice-cold methanol are sometimes used, particularly if aldehyde fixation negatively impacts antigenicity. Following fixation it is important to quench the fixation reaction to prevent the formaldehyde interfering with the staining steps; this can be achieved by incubation with a primary amine, such as glycine, or by carrying out multiple washes in Tris-buffered saline (TBS). 6. Wash cells twice with ice-cold PBS. 7. Fix cells for 10 min with 1 mL/well fixation buffer. 8. Wash cells three times in TBS. 9. Quench fixation buffer for 20 min with 1 mL/well TBS. Once fixed, the cells need to be permeabilized to allow antibody access to intracellular compartments. A number of different detergents can be used, commonly Triton or NP-40. Milder reagents, such as saponin and digitonin may also be used. When used at low concentration these selectively extract a subset of lipids, effectively creating holes in membranes with less perturbation of overall membrane structure. Follow-

Analysis of GPCR Localization and Trafficking

433

ing permeabilization it is important to block nonspecific binding of antibodies by incubating the cells in PBS containing 1% bovine serum albumin (alternative blocking solutions are 1% gelatin or 10% serum from the species that the secondary antibody was raised in), typically, the blocking solution also contains detergent (as do all the antibody solutions). 10. Permeabilize cells for 10 min with 1 mL/well permeabilization buffer. 11. Incubate the cells for 30 min in blocking buffer. 12. Dilute primary antibodies to appropriate concentration in blocking buffer (see Note 1). Any number of antibodies can be added at one time, provided that there is enough specificity for secondary antibodies. Typically, two are used at any time, normally raised in mouse and rabbit; however, two mouse antibodies can be used if they are monoclonal subtypes (e.g., IgG1 and IgG2). 13. Incubate cells for 1 h in primary antibody solution. 14. Wash cells three times in TBS (see Note 2). 15. Leave in TBS for 30 min in TBS. 16. Dilute fluorophore-conjugated secondary antibodies to appropriate concentration in permeabilization buffer. The secondary antibodies should have the required immunological specificity (e.g., anti-mouse vs. anti-rabbit IgG, or anti-IgG1 vs. anti-IgG2). There are several excellent commercial sources for these reagents, and specific advice regarding their selection and use, such as at Jackson ImmunoResearch (Malvern PA, USA). For double-labeling, one would typically choose a green fluorescent conjugate (such as fluorescein or Alexa488) for one secondary antibody and a red conjugate (such as TexasRed or Alexa594) for the other. The specific fluorochromes used should be chosen according to the ability of the available imaging equipment to spectrally resolve them (see Note 3). 17. Stain cells for 1 h in secondary antibody solution. 18. Wash cells twice in TBS. 19. Leave cells for 10 min in TBS. 20. Remove TBS. 21. Repeat steps 19 and 20, three times (see Note 2). Lastly the samples need to be mounted on to microscope slides before imaging. 22. Place two drops of fluoromount onto a microscope slide. 23. Lift coverslips out of the wells using a needle and forceps. 24. Wash in ddH2O. 25. Place coverslip cell side down onto the drop of fluoromount.

434

Hislop and von Zastrow

26. Seal coverslip with nail polish to prevent drying and movement under microscope. 27. Store in the dark at 4°C. It is generally advisable to visualize specimens soon (within a few hours) after preparation. Depending on the dissociation rate of the antibodies used in staining, however, it is sometimes possible to store specimens in the cold for several days. Imaging is typically carried out using a high-quality ­fluorescence microscope. Resolution of subcellular structures typically requires use of water or oil immersion objective lenses that provide a relatively high numerical aperture (NA). Optical resolution is inversely proportional to NA, and objectives used for examining subcellular localization of receptors generally have NA ³ 1. Oil immersion lenses offer the highest possible NA, and thus attainable spatial resolution. Water immersion objectives are often preferred for thicker specimens because of spherical aberration that becomes problematic when imaging deeper into aqueous medium. Standard epifluorescence microscopy can be useful for cultured cells that are relatively flat. Confocal fluorescence microscopy is often preferred, however, especially for thicker specimens. Confocal microscopes reject a fraction of the emission or scattered light originating from out of the focal plane, allowing one to localize receptors in “optical sections” with reduced background. Many excellent instruments are available commercially or in multiuser facilities. By way of example, we typically use a shared Zeiss LSM510 confocal microscope fitted with a 63 × 1.4 NA oil-immersion objective lens for imaging relatively flat tissue culture cells. However, there are many other excellent instruments and options. There are also a variety of newer developments in fluorescence imaging methodology, which may be preferred for particular questions or applications (see Note 4). 3.2. Flow Cytometry

1. Plate cells into 12-well plates so that each time point/drug treatment can be measured in triplicate. 2. Treat cells. 3. Wash cells twice with ice-cold PBS. 4. Add 0.5 mL of PBS to each well and lift cells by mechanical trituration and place in flow cytometry tubes. 5. Dilute primary antibodies to appropriate concentration in staining buffer (see Note 1). 6. Stain cells for 1 h in primary antibody solution. 7. Centrifuge tubes at 500 × g for 5 min. 8. Resuspend cells in 3 mL of PBS. 9. Repeat steps 7 and 8, two times (see Note 2).

Analysis of GPCR Localization and Trafficking

435

If the primary antibody is directly conjugated to a fluorophore then skip ahead to step 15. 10. Leave in PBS for 30 min. 11. Dilute fluorophore-conjugated secondary antibodies to appropriate concentration in permeabilization buffer. 12. Stain cells for 1 h in secondary antibody solution. 13. Wash cells twice by centrifugation. 14. Leave cells for 10 min in PBS. 15. Add 0.5 mL of fixation buffer. 16. Analyze surface fluorescence using a flow cytometer. The details of flow cytometric analysis methods are beyond the present scope, but many institutions have multiuser facilities with suitable instrumentation and expertise. We routinely use a BD FACSCalibur instrument (Becton Dickinson Biosciences, San Jose, CA), and collect data from >10,000 cells per sample. Changes in cell-surface receptor immunoreactivity are typically estimated by determining changes in the mean or geometric mean calculated from the cell population. 3.3. Surface Biotinylation Assay

This assay is a very simple method for analyzing the stability of specific GPCRs in the presence of prolonged agonist treatments; however, it contains a large number of washing steps. Although tissue culture dishes are already treated to allow cell-lines to grow, we treat the dishes with poly-l-lysine to encourage attachment of the cells to the plates and prevent cell loss during the many washes and manipulations required in this assay. The size of the tissue culture plate required will depend on the receptor expression level and the amount of endocytosis the receptor undergoes. For a cell-line expressing receptors that undergo substantial endocytosis, 60-mm dishes will be sufficient, but can be scaled up to 100-mm dishes if required. 1. Add enough poly-l-lysine solution to cover the bottom of 6 × 60-mm dishes (see Note 5). 2. Leave for 10 min at room temperature. 3. Aspirate off the poly-l-lysine solution and leave to air dry for 1 h. 4. Plate cells so they are 80% confluent on the day of the experiment (see Note 6). On the day of the experiment label the plates 100%, “strip,” NT, agonist 30 min, agonist 90 min and agonist 180 min (lengths of agonist treatments can be altered depending on the stability of any particular GPCR), and place the plates on ice (all manipulations should be carried out on ice to prevent endocytosis or trafficking taking place when not wanted). The first step is to label all

436

Hislop and von Zastrow

of the cell-surface receptors with biotin. The biotinylation reagent used includes a reactive N-hydroxysuccinimide (NHS) ester group, which forms a stable amide bond with free amine groups in their nonprotonated form (see Note 7). The biotin used here is the so-called cleavable sulfo-NHS-SS-biotin, indicating that it contains a disulphide bond that can be broken by thiol reducing agents, thereby removing it from the labeled protein. 1. Wash all plates 2× with ice-cold PBS. 2. Aspirate off the PBS. 3. Carefully add 3 mL of 1× biotinylation solution to each plate (made fresh from 100× biotinylation solution, see Note 8). 4. Leave plates at 4°C for 30 min. 5. Quench the biotinylation reaction by washing 2× in ice-cold TBS (the amine groups in the Tris–HCl will “quench” unreacted biotin). 6. Leave the 100% strip plates in TBS and place at 4°C. 7. Add 5 mL of prewarmed complete medium to the other four dishes and return to the incubator. 8. Let cells equilibrate in the incubator for 30 min before addition of agonist to the three agonist plates (leave the NT plate in the incubator without agonist). At this point the agonist plates will need to be removed at different times; however, the protocol is the same for of all the plates. The NT plate should be left in the incubator for the same length of time as the longest agonist-treated plate. All plates (with the exception of the 100% plate) need to be “stripped” – treated with a reducing agent to remove any noninternalized biotin (see Note 9). 1. Wash the plates with 2× ice-cold PBS. 2. Add 3 mL of stripping buffer to each plate (except the 100% plate, which remains in PBS). 3. Leave for 10 min at 4°C, rocking gently. 4. Aspirate the first strip and repeat with another 3 mL of stripping buffer. 5. Leave for 10 min at 4°C, rocking gently. 6. Aspirate the stripping buffer. 7. Add 5 mL of quench buffer to each plate (including the 100% plate). 8. Leave at 4°C for 20 min. The last step in the protocol is the recovery of biotinylated proteins and the detection of receptors. This is achieved by a modified version of an immunoprecipitation assay. Firstly, the cells are lysed and the nuclei and insoluble cytoskeletal components

Analysis of GPCR Localization and Trafficking

437

are removed by centrifugation. Samples are then mixed with streptavidin-conjugated agarose beads to recover the biotinylated proteins. Recovered proteins are then separated by size and charge by polyacrylamide gel electrophoresis, and the amount of receptor detected by an anti-epitope (typically Flag or HA) antibody and enhanced chemiluminescence. 1. Aspirate off the quench buffer. 2. Add 1 mL of lysis buffer (ensuring no reducing agents). 3. Scrape cells off of the plate and place in a 1.5-mL microcentrifuge tube. 4. Clear the supernatant by centrifugation at 20,000 × g for 10 min at 4°C. 5. Transfer supernatant to a new set of microcentrifuge tubes. 6. Add 35 mL of streptavidin-conjugated agarose to each tube. 7. Incubate on a rocker at 4°C for 3–16 h. 8. Centrifuge at 6,000 × g for 1  min (streptavidin-conjugated agarose beads should sediment rapidly), discard the supernatant, and resuspend in 1  mL of lysis buffer. Repeat three times, rotating for 5 min between each wash and removing as much liquid as possible using a vacuum aspirator attached to a 22- to 25-gauge needle (to prevent aspiration of beads). 9. Repeat step 28, but resuspend in 1 mL of 20 mM Tris/HCl, pH 7.4. 10. Add 10 mL of PNGase-F solution. 11. Incubate at 37°C for 1 h. 12. Add 10 mL of sample buffer. 13. Incubate at room temperature for 30  min. We do not boil samples, because we find this promotes the formation of aberrant high molecular weight receptor aggregates. 14. Centrifuge 14,000 × g for 10  min at room temperature and load supernatant on SDS-polyacrylamide gels suitable for resolving the candidate GPCR. 15. Transfer to nitrocellulose or PVDF membranes using standard methods and immunoblot for GPCR using anti-epitope antibody (typically HA-11 or M2 Flag). The different plates should give different amounts of immunoreactivity. The 100% plate is the total amount of receptor present at the cell-surface, the “strip” is a control plate to ­determine the efficiency of the protocol, and this lane should be essentially empty. The NT lane shows the agonist-independent endo-cytosis occurring during the time-course of the experiments. The agonist for 30 min is the pool of endocytosed receptor­. The other agonist plates will demonstrate the stability of the ­receptor. If these plates are the same or more than the 30-min

438

Hislop and von Zastrow

time point then the receptor is fairly stable and likely to recycle­, if they are less than the 30-min point then the endocytosed pool has undergone rapid proteolysis.

4. Notes 1. In our experience the most critical determinant of success or failure in immunocytochemical staining is the primary antibody. In all cases, one must find the optimal antibody ­concentration and incubation time. For monoclonal antibodies and epitopetagged receptors in dissociated cells, the staining is typically highly robust over a range of times and concentrations. For antisera, as well as for all types of staining applied to tissue sections, these conditions are more critical. An important control for nonspecific staining for tagged receptors is to apply the same conditions to cells not expressing the tagged receptor, or expressing the same receptor tagged with an irrelevant epitope. For endogenous proteins the nonspecific control is often limited to blocking peptide or immunogen. This is an important control, but is less definitive because cross-reactive, but irrelevant antigens can also be blocked. In this case the ideal control is to stain cells/ tissue from the relevant knockout animals, if available. 2. The second most important variable for successful staining is adequate washing after antibody incubations. This must be determined for each situation, but, in general, one must wash extensively and make sure that complete solution changes occur in each wash. The same principle applies to flow cytometric assays, although this method can be more forgiving, because the passage of cells through the saline solution used in the internal plumbing of the instrument provides additional washing before analysis. 3. In any multilabel experiment, one should verify specific detection of each antigen. In our experience, the two most common causes of trouble are immunochemical cross-reactivity between primary and secondary reagents (i.e., one of the secondary antibodies has significant affinity for the “wrong” primary antibody) and spectral “bleed-through” (i.e., the imaging system does not adequately separate the fluorochromes used to label the secondary antibodies). One can test for immunochemical cross-reactivity in control specimens using single labeling for the “wrong” combination of primary and secondary antibodies. One can test for spectral bleed-through in control specimens using single labeling for the “right” individual combination of primary and secondary antibodies, and imaging in the “wrong” channel on the microscope. To be valid, these controls must be carried out in parallel specimens and using the same microscope acquisition settings.

Analysis of GPCR Localization and Trafficking

439

4. We typically use conventional epifluorescence or laser-scanning confocal microscopy for fixed specimens. One major area of rapid advance in the analysis of GPCR trafficking is the imaging of receptor localization and redistribution in living specimens. Such approaches can provide important insights into regulatory events occurring acutely in response to physiological or pharmacological stimuli. Fluorescent antibody conjugates can be useful for live-cell imaging, but the popularity of this approach has exploded with the availability of intrinsically fluorescent protein tags, such as GFP and its many variants. Live imaging involves many of the same principles and techniques described above, with some adaptations. Additional fluorescence imaging methods, such as spinning disk confocal microscopy and twophoton microscopy are useful for determining three-dimensional localization of receptors in cells and tissue preparations. Total internal reflection fluorescence (TIRF) microscopy is gaining popularity for imaging receptor redistributions among microdomains of the plasma membrane (11, 12). Another area of current advance is toward achieving higher spatial resolution than the theoretical diffraction “limit” of optical microscopy. This has the exciting potential to reveal, potentially even in live preparations, fine details of receptor localization that were previously accessible only in fixed specimens using electron microscopy. These so-called super-resolution methods of fluorescence imaging involve the merging of conventional optical techniques with additional physical principles. One approach, called photoactivated localization microscopy (PALM), is based on the principle of single-particle tracking and adapted to the availability of improved photo-convertible fluorescent dyes (13). Another promising approach, called stimulated emission decay (STED) microscopy, is a variation of laser-scanning confocal microscopy that increases the effective resolving power of the laser excitation (14). A number of recent studies have carried out liveimaging of GPCR trafficking in both cultured cells and tissues. Single-particle tracking has been applied to GPCRs for some time (15), and it is likely that GPCR localization using newer super-resolution methodologies will be forthcoming soon. 5. It is important to include controls to make sure that the biotinylated signal detected represents the labeled receptor and not a nonspecific signal, at least until the experiment is standardized and the characteristic protein mobilities established. Typically this involves analyzing, in parallel, cells not expressing the tagged receptor or expressing an alternately tagged version. 6. Optimal plating density and cell health is essential to keep cells from lifting off the dish through the extensive series of incubations and washes. This is particularly true for less adherent lines, such as HEK293 cells. 7. The major reactive groups are the N-terminal a-amine (if not blocked) and the e-amine present in lysine side-chains, both of

440

Hislop and von Zastrow

which are basic, and thus the efficiency of cell-surface biotinylation can be quite low when conducted in PBS (pH ~7.4). If weak signals are obtained, alkalinize the biotinylation reagent slightly (note, it is advisable not to exceed pH ~8.5 due to the tendency of alkaline conditions to produce ­cytotoxicity). Owing to the generally lower pKa of the a-amine compared to e-amine group (~9 vs. ~10.5), manipulating the reaction pH can also be used to adjust the degree to which lysine side chains are modified relative to the N-terminal a-amine. 8. The biotinylation reagent can react also with water. This reagent must be prepared fresh in DMSO, and the powder should be kept in a desiccator to avoid inactivation during storage. 9. The reducing agent breaks the disulphide bridge of any biotin still on the surface of the cell. We provide recipes for two different stripping buffers, both of which are effective, but some cells may tolerate one buffer better than the other. References 1. Hanyaloglu, A. C. and von Zastrow, M. (2008) Regulation of GPCRs by endocytic membrane trafficking and its potential implications. Annu. Rev. Pharmacol. Toxicol. 48, 537–568. 2. Marchese, A., Paing, M. M., Temple, B. R. and Trejo, J. (2008) G protein-coupled receptor sorting to endosomes and lysosomes. Annu. Rev. Pharmacol. Toxicol. 48, 601–629. 3. Bernier, V., Bichet, D. G. and Bouvier, M. (2004) Pharmacological chaperone action on G-protein-coupled receptors. Curr. Opin. Pharmacol. 4, 528–533. 4. Lippincott-Schwartz, J., and Patterson, G. H. (2003) Development and use of fluorescent protein markers in living cells. Science 300, 87–91. 5. Petaja-Repo, U. E., Hogue, M., Laperriere, A., Walker, P. and Bouvier, M. (2000) Export from the endoplasmic reticulum represents the limiting step in the maturation and cell surface expression of the human d-opioid receptor. J. Biol. Chem. 275, 13727–13736. 6. Kim, K. A. and von Zastrow, M. (2003) Neurotrophin-regulated sorting of opioid receptors in the biosynthetic pathway of ­neurosecretory cells. J. Neurosci. 23, 2075–2085. 7. Vickery, R. G. and von Zastrow, M. (1999) Distinct dynamin-dependent and -independent mechanisms target structurally homologous dopamine receptors to different endocytic membranes.J. Cell Biol. 144, 31–43. 8. Gage, R. M., Kim, K. A., Cao, T. T. and von Zastrow, M. (2001) A transplantable sorting signal that is sufficient to mediate rapid ­recycling of G protein-coupled receptors. J. Biol. Chem. 276, 44712–44720.

9. Tsao, P., Cao, T. and von Zastrow, M. (2001) Role of endocytosis in mediating downregulation of G-protein-coupled receptors. Trends Pharmacol. Sci. 22, 91–96. 10. Bartlett, S. E., Enquist, J., Hopf, F. W., Lee, J. H., Gladher, F., Kharazia, V., Waldhoer, M., Mailliard, W. S., Armstrong, R., Bonci, A. and Whistler, J. L. (2005) Dopamine responsiveness is regulated by targeted sorting of D2 receptors. Proc. Natl. Acad. Sci. USA 102, 11521–11526. 11. Puthenveedu, M. A. and von Zastrow, M. (2006) Cargo regulates clathrin-coated pit dynamics. Cell 127, 113–124. 12. Yudowski, G. A., Puthenveedu, M. A., Henry, A. G. and von Zastrow, M. (2009) Cargo-mediated regulation of a rapid Rab4dependent ­recycling pathway. Mol. Biol. Cell. 20, 2774–2784. 13. Manley, S., Gillette, J. M., Patterson, G. H., Shroff, H., Hess, H. F., Betzig, E. and Lippincott-Schwartz, J. (2008) High-density mapping of single-molecule trajectories with photoactivated localization microscopy. Nat. Methods 5, 155–157. 14. Willig, K. I., Harke, B., Medda, R. and Hell, S. W. (2007) STED microscopy with continuous wave beams. Nat. Methods 4, 915–918. 15. Hegener, O., Prenner, L., Runkel, F., Baader, S. L., Kappler, J. and Haberlein, H. (2004) Dynamics of b2-adrenergic receptorligand complexes on living cells. Biochemistry 43, 6190–6199.

Part V Statistical Methods

wwwwwww

Chapter 26 Statistical Methods in Research Domenico Spina Abstract Statistical methods appropriate in research are described with examples. Topics covered include the choice of appropriate averages and measures of dispersion to summarize data sets, and the choice of tests of significance, including t-tests and a one- and a two-way ANOVA plus post-tests for normally distributed (Gaussian) data and their non-parametric equivalents. Techniques for transforming non-normally distributed data to more Gaussian distributions are discussed. Concepts of statistical power, errors and the use of these in determining the optimal size of experiments are considered. Statistical aspects of linear and non-linear regression are discussed, including tests for goodness-of-fit to the chosen model and methods for comparing fitted lines and curves. Key words: Statistical analysis, Descriptive and comparative statistics, Parametric, Non-parametric, Normal distribution, Statistical power

1. Introduction Experimental results of all kinds are prone to uncertainty or error, particularly in biology. This error can be minimized by applying the accepted criteria for good experimental design – such things as making sure that data values are independent, assigning treatments at random to experimental units, making sure that sources of bias are eliminated, controlling for known sources of variability and using an experimental approach with a wide range of applicability – but it is not possible to eliminate error completely. When performing an experiment to investigate a phenomenon or to measure some quantity, it becomes important to have some measures of this error, which is where statistical methods come in. For the most part, biologists use statistics to describe their data by calculating summary statistics (averages) and quoting quantities which describe variability in these data (measures of

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0_26, © Springer Science+Business Media, LLC 2011

443

444

Spina

dispersion), to look for relationships between data values using regression and correlation techniques and also to evaluate the effects of experimental manipulations, using tests of significance. It should not be forgotten, however, that these are tests of statistical not scientific significance; they cannot conceal poor results and should not be used as a substitute for scientific argument (1). They can help to interpret results of experiments, but it is vital to choose the appropriate method for statistical analysis of results at the time of designing the experiment, rather than (as happens all too often) at the end of the project. Then, the experimental protocol can be designed to make sure that the appropriate data are collected to allow the chosen analysis. It should be remembered that all statistical methods are based on the calculation of probabilities, rather than certainties. Thus, statistical results cannot prove that any scientific hypothesis is true, but only provide some justification for believing that it might not be false. However, it is expected that most biologists seeking to publish their work include statistical results to satisfy journal editors and reviewers. The purpose of this chapter is to discuss how to choose appropriate methods to analyze data, and how to use statistical techniques to improve experimental design.

2. Summary Statistics Implicit in the quoting of an average or summary value for a quantity measured in an experiment is the idea that this represents the true value for this average – the population average. Since experiments can never measure the entire population, they have to rely on measuring the average for a sample. Statistical methods assume that this sample is randomly chosen from the population, so when designing your experimental protocol you should consider whether or not you violate this assumption and try to avoid bias in both sample selection and measurement. In reality, this assumption is rarely met, and at best we obtain samples of convenience that are randomly allocated to different treatments (2, 3). Most biological data fit a Gaussian or Normal (as in “pattern”, not “usual”) distribution, which has defined parameters with particular properties (Fig. 1). Statistical techniques which assume that data fits a particular distribution are known as parametric techniques and are the most powerful and widely used. Nonparametric statistics, on the other hand, do not make any assumption about the shape of data distribution and are thus universally applicable.

Statistical Methods in Research

−3

−2

−1

0 mean

+1

+2

445

+3

Fig. 1. The Normal (Gaussian) distribution calibrated in terms of standard deviation. 68% of data values in such a distribution are expected to fall within ±1 SD of the mean, 95% within ±1.96 SD, and 99.7% within ±3 SD.

Table 1 Some useful transformations Type of data

Distribution

Count (e.g. dpm, cpm)

Poisson distribution

Proportion or %

Binomial distribution

Normalizing transform

C Arcsine

P Measurement (e.g. EC50)

2.1. Averages: Measures of Central Tendency or Position

Log Normal

Log (M )

Given the choice, most scientists calculate the arithmetic mean as an appropriate average value for any data set. This is normally appropriate as a summary as long as the data fit a normal distribution. Fortunately most biological data do not deviate too far from normality and the mean is usually a good average. If data are clearly not normally distributed, as has been shown for IC50 and Ki values from ligand-binding experiments or EC50 values from agonist concentration-effect curves (4, 5), it is appropriate to transform them (in these cases using a logarithmic transformation) so that they fit the normal curve and do the statistical analysis on transformed data. Table  1 lists this and some other transformations which have been found to be useful in normalizing the distribution of different types of data before undertaking statistical analysis. The summary statistics can then be transformed back to the original units to quote final results. Other data types commonly used in pharmacology which present problems in analysis may include data expressed as percentages, fractions, or multiples of control values. These are paired data values, but the differences between pairs are larger as

446

Spina

the control value increases, hence the usefulness of expressing the difference as a ratio. But ratios are asymmetric, so it makes more sense to analyze a logarithmic transformation of this type of data (see ref. 6 and Subheading 3.1.1 below). Alternatively, percentages between 0 and 100% follow a binomial distribution and for data in this range the arcsine P transformation may be used to convert to a more Gaussian distribution. Transformation is not possible for certain other data types that are clearly not normally distributed – discontinuous data, category data, or heavily skewed continuous data – and here too the arithmetic mean is not a good summary statistic. It is better in these cases to use the median (the middle value) of the data set, since this is less affected by outliers, or even the mode or modes (the most frequent value(s)) as an appropriate summary. Box plots are a convenient graphical format for non-parametric data sets. 2.2. Measures of Dispersion

Having selected an average, it remains to choose a measure of dispersion to describe the data distribution. For normally distributed data, it is possible to use an estimate of the population variance (the sum of squares of deviations from the mean, divided by the number of data values minus one) to describe the distribution. This is not very convenient, since the units of variance are the square of the data units. More conventionally we use the standard deviation (SD, the square root of the variance) to indicate the spread of values in a normally distributed data set. In fact, SD has a precise meaning defined by the Gaussian curve; 68% of data values in such a distribution are expected to fall within ±1 SD of the mean, 95% within ±1.96 SD, and 99.7% within ±3 SD (Fig. 1). Thus, the bigger the SD, the more scattered the data so that the sample SD describes the variability in a particular set of data. As the size of the data set increases, the SD approaches the population SD (s). To estimate the true value of a measure of central tendency, it would be appropriate to make repeated determinations in separate experiments and calculate a mean. The distribution of sample means can then be used to describe a measure of dispersion, the standard error of the mean (SEM, calculated as the SD divided by the square root of n). As the size of the data set increases, the SEM approaches zero as the sample mean approaches the population mean. Thus, a small SEM may not represent tight data, just a large sample size. So, should you quote SD or SEM? After all, you can calculate both for any data set. Many experimenters prefer SEM since it is clearly smaller than the SD, but it should be remembered that it is describing a different property of the data set – the variability in the mean, not the spread of data values. As a general rule, it is

Statistical Methods in Research

447

more appropriate to use mean ± SD when summarizing a single experiment so that the reader knows how variable the values were, and mean ± SEM when presenting results from a series of repeated experiments because you are then estimating how close the sample mean is to the population mean. Some (mainly clinical) journals prefer the use of confidence intervals (CI) as summary statistics rather than SDs and SEMs and p values (7). These give a better idea of the (im)precision of the study estimates as generalized to the wider population. A 95% confidence interval for a mean (the most frequently chosen value for CI) expresses the range of values between which you can be 95% certain that the population mean lies (see Example 1). Certainly, it is easy for the reader to grasp an idea of variation in the data from the 95% confidence intervals, which are expressed in the same units as the original data. For a population mean, the 95% CI is calculated as the mean ± 1.96 SEM but for experimental (i.e. real) data, it is necessary to invoke the t distribution to correct for small sample size, as below:

95%CI = sample mean ± t (0.05, n )´ SEM.

Example 1  Calculation of 95% Confidence intervals Blood pressure measurements for a group of 41 randomly selected students yielded a mean of 123.4 mmHg with SD of 14.93. The t value for this group size is 2.021; thus, the 95% CI are 123.4 - (2.021 ´ 14.93 / 41) to 123.4 + (2.021 ´ 14.93 / 41) or 118.7–128.1 mmHg. This would be quoted as 123.4 (118.7–128.1) mmHg. A random sample of ten students was drawn from this group and their blood pressures used to perform a similar calculation. The mean value was 110.5  mmHg with SD of 16.76. The t value for n = 10 (9 df ) is 2.262 and so the mean and 95% CI in this case are calculated as 110.5 (98.5–122.4) mmHg, wider than that found from the larger group.

If you are using the median or mode as an average (i.e. implying that you do not consider the data to be Gaussian), you should not quote SD or SEM as measures of dispersion. The simplest such measure in these cases is the range of data values. If using the median, which is actually the location of the second quartile, you can calculate the first and third quartiles for your data and present

448

Spina

your data as median (Q2) and inter-quartile range (Q1–Q3). Either of these approaches has the advantage of making the direction of any skew in the data values obvious to the reader. 2.3. Are Your Data Normally Distributed?

If there is no published consensus on the matter, deciding if your data are normally distributed can be a problem. It may be helpful to inspect the distribution by eye using a dot-plot or box and whisker diagram (there is rarely enough data to draw a meaningful histogram) and come to a conclusion. Also one can have a good idea of whether or not particular data types are expected to be normally distributed from their nature – continuous measurements are usually normally distributed whereas category data, such as scores, are not. The experimental protocol may lead to skewed data if there are prohibited values or cut-off points imposed by the experimental procedure. In Example 2, a condition of the Home Office Project Licence was that all mice should be removed from the hot plate at 60 s so that no value in excess of this was possible, even when some mice had not reacted by this time. In some experiments, the range of possible values is dictated by physiological factors. If the dependent variable is known to deviate from a Gaussian distribution (e.g. EC50), then it is expedient to apply an appropriate transformation (e.g. logarithmic). Alternatively, one has to make an assumption that the sample distribution conforms to a Gaussian distribution or alternatively, use non-parametric tests where there is no assumption of the underlying distribution. The statistical software may offer one or more tests for normality, such as the Kolmogorov–Smirnov, Levene, or Wilk– Shapiro tests. However, a problem with these is that they give reliable results only when applied to large data sets (typically n > 30) and the small group sizes most commonly used in real experiments violate their assumptions (5).

Example 2  Calculation of the median, range, and inter-quartile range for skewed data In an experiment, measuring the time taken for control mice to respond to a standard noxious stimulus (being placed on a hot plate at 49°C) the following response times (s) were obtained: 29.4, 14.8, 60.0, 29.2 17.5, 19.5, 29.0. The median response time is 29.0  s, the range is 14.8–60.0 s and the inter-quartile range is 17.5–29.4 s. In this case, the inter-quartile range is more indicative of the direction of skew, but the range clearly shows the extreme outlier. An alternative to quoting these numbers would be to draw a box-plot of the data where the median, Q1, Q3, and the range could be shown.

449

Statistical Methods in Research

3. Tests of Significance It is conventional to use tests of statistical significance to help assess the scientific significance of differences seen between treatment groups in experiments. The general procedure for such a test is to start by defining a null hypothesis – that there is no difference between treatment groups – and then perform the test to calculate the probability that a result as extreme as that observed experimentally could have arisen by chance. If that probability is small – conventionally one takes p < 0.05 as the critical probability (alpha) – then it can be assumed that the probability of the null hypothesis being true is also small (i.e. less than 5%) and that by implication there is a difference between groups. This is an entirely arbitrary level at which to decide to reject the null hypothesis, due originally to Fisher, and any other level may be chosen. Obviously, the smaller the value of p obtained in any case, the less likely it is that the null hypothesis is true, but it should be remembered that extreme values may occur in any experiment by chance and so the null hypothesis is falsely rejected. Statistical reasoning is all about probabilities, not certainties (see discussion of statistical power, and Type I and Type II errors below). Choosing a statistical test is fairly straightforward. If data ­values are normally distributed then tests assuming this may be used. In other cases it is safer to rely on results of tests which make no assumptions about the shape of the data’s distribution; these are referred to as non-parametric tests or distribution-free methods. Several statistics texts provide flow-charts or tables that illustrate the analysis of data through selection of an appropriate test (8–10). Several statistics software packages also provide some guidance in a user-friendly manner (e.g. GraphPad Prism™, Instat™). An indispensible guide for the use of SPSS™ is highly recommended (11). 3.1. Tests Based on the Normal Distribution 3.1.1. Tests for Two Groups



For the comparison of two groups of normally distributed data values, t-tests are appropriate. These were devised to be applicable to small group sizes and different procedures are available for applying them to single sample data (comparing observed data with known values), to two independent treatment groups, or to treatment groups where the values are paired (either because you have repeated measures data, or observations on paired individuals, or matched pairing like cells from the same source plated into 96-well plates but receiving different treatments). The t statistic has the following mathematical form Difference between group means t statistic = . Pooled standard error of the difference



450

Spina

It assumes that data fit the normal distribution and that variances in the two groups are the same. A variant of the independent samples t-test (Welch’s test) has been developed for cases where the variances are not the same and is offered by several statistics packages. In cases where you have paired data expressed as a % or ratio of control values (see Subheading  2.1), Motulsky (6) suggests performing a t-test of the ratios by converting the data to logarithms and using a paired t-test to find the probability of obtaining a ratio as far from 1.0 (i.e. no difference) as observed in your experiment. Earlier, the notion of samples of convenience was described (see Subheading 2) and it has been argued that all data sets should be analyzed using a randomization or permutation test which does not rely on the assumption that samples were randomly derived from a normal population (2, 3). However, most statistical packages do not offer this approach, but it appears that parametric tests are a good approximation of the permutation test (2). The probability of the calculated t-value arising by chance (p) depends on the degrees of freedom (df, best explained as the number of observations which are free to vary; [n − 1] for a paired test and [n1 + n2 − 2] for an unpaired test) and can be represented by the area under the tail of the curve for the appropriate value of t. Because the distribution is symmetrical about 0, there are both positive and negative values of t and hence 2 tails and 2 areas to be added together to obtain a value for p. The sign of t obtained by calculation depends on the direction of the difference between the means and is disregarded in assessing p. It is conventional in biology not to specify in advance the direction of this difference i.e. to perform a two-tailed test. This is regarded as the robust way to proceed. However, it is sometimes appropriate to predict in advance the direction of difference and perform a one-tailed test, in which case the area, and hence the p value, is halved. 3.1.2. Comparing More Than Two Groups: One-Way Analysis of Variance

Experiments comparing only 2 groups are not generally considered to be examples of good experimental design because they yield limited information. Designing your experiment to compare more than 2 groups at once allows an examination of the effect of several treatments (e.g. different drugs, doses of a single drug, time courses, etc.) under the same conditions using a single control group and is thus economical of time, effort, and materials. However, the analysis of such experiments becomes more complex, requiring the use of the powerful technique of analysis of variance (ANOVA) rather than performing repeated t-tests. ANOVA avoids the problem of multiple comparisons, where the chance of making a Type I error (erroneously rejecting the null hypothesis, see below) increases with the number of pairwise comparisons being made. ANOVA considers the overall

Statistical Methods in Research

451

variability in the data and compares between-group variation with within-group variation producing a ratio of variances, the F ratio. If there is no difference between the groups, the F ratio will be close to 1. Where the between-group variability exceeds the withingroup variability, F is large and the p value for this arising by chance can be calculated, taking into account the df for the betweengroup and within-group variance. As with all tests of significance, conventionally p < 0.05 is taken as indicating a statistically significant result, allowing rejection of the null hypothesis. The example below shows a typical one-way ANOVA results table. In Example 3, animals were randomized to receive only one treatment and therefore, each providing one independent variable (i.e. eosinophil number). The results of the ANOVA indicate that the null hypothesis of no statistical difference between sample means must be rejected and hence, we conclude that there is an overall treatment effect. An important assumption of ANOVA which should be met is homogeneity of sample variance and most statistical software runs this type of analysis (e.g. Levene’s test of equality of error variances or Bartlett’s test). In Example 3, a test for homogeneity of variance yielded a p value of 0.979 and so we can conclude that sample variance was similar between the 4 groups and we can accept the results of the ANOVA. In the circumstance where sample variances are different, it might be prudent to try a normalizing transform (see Table. 1) and then repeat the test. If this does not help, then one could use a non-parametric test to analyze the data. Since the null hypothesis that the sample means are similar has been rejected, it still remains to be determined where this difference occurs (i.e. which pair(s) of sample means when compared gives rise to a rejection of the null hypothesis?). There are two approaches to ascertain which treatment groups are responsible for the rejection of this null hypothesis. Statisticians distinguish between comparisons planned before doing the experiment i.e. based on predictions from the scientific hypothesis (a priori tests) and tests to investigate differences shown up after the initial ANOVA (post hoc tests). 3.1.3. A Priori Tests

A priori tests enable the experimenter to make specific comparisons (contrasts) between treatment groups and involves partitioning the between group variance. This choice must be made before the experiment has been performed and cannot be decided retrospectively after the results have been inspected as this can lead to bias. In this case, a post hoc test must be performed. The number of planned comparisons possible is one less than the number of treatment groups (k − 1). However, it is conventional that these contrasts must be orthogonal (i.e. independent comparisons). In Example 3, there are three orthogonal contrasts

452

Spina

as follows: saline vs average of the three treatments (ln SPSS, this would be defined as Level 1 vs later); antigen vs average of the two monoclonal treatments (Level 2 vs later); and finally MoAb A vs MoAb B (Level 3 vs Level 4) (see Example 4). In some circumstances, it might be appropriate to undertake a non-orthogonal contrast although there is debate concerning the validity of these comparisons since by definition they are not independent comparisons. One pragmatic approach is that orthogonal contrasts are preferable but if the contrasts do not answer a particular experimental question, then it is legitimate to run a non-orthogonal contrast (12) (see Example 5). Example 3  Output from a one-way ANOVA to analyze the effect of two monoclonal antibodies (MoAb) on antigeninduced eosinophil accumulation in rabbit lung. The table below shows the number of eosinophils measured in lavage fluid in different rabbits randomized to the four treatments. The null hypothesis is that the four groups come from a population with the same mean and variance Saline

Antigen

Antigen + MoAb A

Antigen +  MoAb B

33

142

131

30

36

143

132

27

36

139

127

23

22

125

116

9

51

160

139

34

Testing these data with a one-way ANOVA gives the following standard output ANOVA table: ANOVA SS

df

MS

F

Sig.

175.4

<0.001

Between groups

5,6152.5

3

1,8717.5

Within groups

1,707.2

16

106.7

Total

5,7859.7

19

SS sum of squares, df degrees of freedom, MS mean squares

The value for F, the ratio between the between group and within group variances, is 175.422. The significance value is the probability of this value arising by chance if the null hypothesis is true (i.e. F is actually 1.0) and is <0.001. In this case, the F ratio suggests that the means of each group differ (i.e. there is a treatment effect).

Statistical Methods in Research

453

Example 4  Planned orthogonal contrast of the data in Example 3 There are three planned contrasts that are orthogonal (Helmert contrast in SPSSTM): Description

Contrast

Saline vs average of antigen + MoAb A + MoAb B

Level 1 vs Later

Antigen vs average of MoAb A + MoAb B

Level 2 vs Later

MoAb A vs MoAb B

Level 3 vs Later

Contrast Results (K matrix) Treatment Helmert Contrast Level 1 vs Later

Contrast estimate Hypothesized value Difference (estimate-hypothesized) Std error Significance

−62.867 0 −62.867 5.334 <0.001

Level 2 vs Later

Contrast estimate Hypothesized value Difference (estimate-hypothesized) Std error Significance

65.000 0 65.000 5.658 <0.001

Level 3 vs Level 4

Contrast estimate Hypothesized value Difference (estimate-hypothesized) Std error Significance

104.400 0 104.400 6.533 <0.001

The contrast estimate is the difference in means between the two groups (e.g. mean of MoAb A vs mean Mo Ab B for the level 3 vs level 4 contrast). The analysis reveals a significant difference between saline and the antigen groups; between antigen and the two drug treatment groups and finally between the two MoAb treated groups. We would conclude that while antigen significantly increased eosinophil recruitment to the lung and was blocked by antibody treatment, this effect was specific to MoAb B only.

Other forms of planned comparisons include linear, quadratic, or cubic contrasts based on whether the treatment groups follow a “dose-relationship” of the form described by the contrast. Planned comparisons are not a feature of GraphPad Prism™, however, statistical software like SPSS™ can run this analysis.

454

Spina

Example 5  Planned non-orthogonal contrast of the data in Example 3 An example of 3 non-orthogonal planned contrasts (repeated contrast in SPSSTM): Description

Contrast

Saline vs antigen

Level 1 vs Level 2

Antigen vs MoAb A

Level 2 vs Level 3

MoAb A vs MoAb B

Level 3 vs Level 4

Contrast Results (K matrix) Treatment Repeated Contrast Level 1 vs Level 2

Contrast estimate Hypothesized value Difference (estimate-hypothesized) Std error Significance

−106.200 0 −106.200 6.533 <0.001

Level 2 vs Level 3

Contrast estimate Hypothesized value Difference (estimate-hypothesized) Std error Significance

12.800 0 12.800 6.5335 0.068

Level 3 vs Level 4

Contrast estimate Hypothesized value Difference (estimate-hypothesized) Std error Significance

104.400 0 104.400 6.533 <0.001

The analysis reveals a significant difference between saline and the antigen group; and between the two MoAb treated groups but not between the antigen and MoAb A group. We would conclude that while antigen significantly increased eosinophil recruitment to the lung this was blocked by MoAb B only.

3.1.4. Post hoc Tests

There is much discussion in the literature as to the most appropriate test procedure, based on the need to avoid type I errors by restricting the pair-wise error rate, as opposed to detecting real differences between groups i.e. avoiding type II errors (5, 8, 10, 13–15). In order to avoid this particular problem, one must use a post hoc analysis. It is inappropriate to simply apply multiple t-tests between different groups because it invalidates the assumption of independence and results in the all too frequent erroneous rejection of the null hypothesis. In general, choose a comparison

Statistical Methods in Research

455

procedure based on the design of your experiment. There are a bewildering number of post hoc tests and a description of each test has been described elsewhere (13). Statistical software offers a variety of such tests; some of more use than others. Least significant difference (LSD) is a procedure to calculate the smallest difference between means which could be considered significantly different. This procedure was popular because the calculation could be readily carried out on paper; however, it is not now considered to be a reliable post-test. The Bonferroni modification (see Example 6) of the t-test is widely accepted as a conservative method of restricting the family wise (experiment) type I error rate to an acceptable level, reducing the probability of making a “false discovery”. The Bonferroni correction is not advisable to use if you have more than five groups, since it is a conservative test and thus has low power to detect differences. Widely used alternatives for performing multiple comparisons are the Tukey or Sheffé tests (5). Most texts recommend Tukey’s or Student Newman–Keuls tests as being more powerful than Bonferroni for large numbers of comparisons, or the Ryan-EinotGabriel-Welsch procedure (14). In the case of Example 7, a Bonferroni correction was used to reject the null hypothesis that Example 6  Post-test results for eosinophil data Since there are only four groups and the experimenter was interested in relatively few specific comparisons i.e. between Saline and Antigen, Saline and MoAb A, Saline and MoAb B and MoAb A vs MoAb B, a Bonferroni test was used. The relevant parts of the output (from SPSSTM) are shown below: Multiple Comparisons Bonferroni Mean diff. Std. (I − J) error

Sig.

I

J

Saline

Antigen

106.2*

6.533

<0.001

Antigen

MoAb A

11.0

6.533

0.670

Antigen

MoAb B

93.4*

6.533

<0.001

Antigen/MoAb A

Antigen/MoAb B

104.4*

6.533

<0.001

SPSSTM helpfully identifies (by *) mean differences, where p < 0.05 to help you make sense of the output. So we can see that antigen clearly increased eosinophil recruitment and that MoAb A was not as effective as MoAb B in inhibiting this response.

456

Spina

glucose levels were similar between each time point and the postglucose injection time point (mean difference ± sem; e.g. time 1 vs time 2; 11 ± 1.3, p < 0.001; time 1 vs time 3; 12 ± 1.5, p < 0.001; and time 1 vs time 4; 9.8 ± 1.9, p = 0.002). Similarly, one could undertake a Bonferroni post hoc test to ascertain the location of the difference between apparatus in Example 8. For example, the measurement of contractile agonist potency is significantly different between apparatus 1 vs 2 (p < 0.001), 1 vs 3 (p < 0.001), 1 vs 4 (p < 0.001), 1 vs 5 (p < 0.001), and 1 vs 6 (p < 0.001). In this case, we are undertaking a comparison of the main effect. The “day” factor is ignored and the agonist pD2 values are averaged over the 4 day for each apparatus to allow comparisons of agonist mean pD2 values between each apparatus. If you want to restrict the number of comparisons by only comparing each group with a specified control group, as is sometimes appropriate in pharmacology, then Dunnett’s post-test is designed to do this specifically. Although widely used in the literature, Duncan’s Multiple Range test is not recommended for post hoc testing in statistics texts, since it has no means of controlling the pair-wise error rate (14). If you have a very large number of complex comparisons or contrasts to make, then Sheffé’s method is probably the best to use although rather conservative. 3.1.5. More Complex Designs of ANOVA

One of the advantages of ANOVA as a statistical technique is its flexibility – it is possible to analyze complex experiments which make use of pairing, blocking, and repeated measures designs to reduce variability as well as designs examining the effects and possible interactions of two or more treatments (usually called factors in statistical texts), using two-way or multi-way analyses that can also accommodate repeated measures designs. One problem with these procedures is the complexity of the output from statistics packages. The examples below illustrate the sections of the analysis output table that are used to interpret the result of a one-way repeated measures design (see Example 7), a two-way independent ANOVA (see Example 8) and a two-way repeated ANOVA (see Example 9). In order to compare the effect of more than one factor in a single experiment then the ANOVA can become more complex but can also reveal if one factor influences (or interacts with) another. These factorial designs are economical in terms of experimental units and can also feature repeated measures if your experimental design controls for inter-subject variability, either by matching subjects or repeating measurements on the same subjects. If the results of a two-way ANOVA indicate significant differences between groups, post-tests can identify these as for a one-way ANOVA. Most statisticians regard t-tests and ANOVA as robust to slight deviations from normality, particularly for large group sizes, so in most cases one would tend to ignore the results of a normality

457

Statistical Methods in Research

Example 7  A one-way ANOVA with repeated measures Mice were injected with glucose (2 g/kg) and plasma glucose levels (mM) measured at various timepoints after injection (post time 0). The null hypothesis is that there is no difference in plasma glucose levels over time. Time (min) Animal

0

15

30

60

1

17.1

30.2

33.3

32.4

2

15

24.4

26.9

26.1

3

15.3

22.8

26.9

26.2

4

12.4

20.8

26.7

30.1

5

12.4

21.6

28.7

28.8

6

10.5

24.8

28.8

29.2

7

9.5

22.4

23.9

15.3

8

9

23.5

19.7

11.5

9

8.5

17.1

14.9

9.4

16

17.1

14.4

10

10

11

9.4

14.6

10.5

10.9

12

9

30.4

26.1

21.6

Using SPSSTM and assuming sphericity, the following output is obtained. Source

SS

df

MS

F

Sig.

1,191

11

108.3

7.8

<0.01

Time

1,108.194

3

369.4

26.7

<0.001

Error

455.879

33

13.8

Between subjects Animal With subjects

Total

2,559

19

The results show that there is a significant difference in glucose levels over time. We are not interested in the between group variance but the analysis reveals that the time course profiles differ between mice.

458

Spina

Example 8  A two-way independent ANOVA The –log pD2 values for agonist contractile potency on ­isolated perfused hearts was investigated on four separate occasions in six different apparatus. The null hypothesis that pD2 values are not affected by day or apparatus can be tested. Day D1

D2

D3

D4

A1

8.9

9.2

9.4

9.5

A2

7.0

7.3

7.5

7.8

A3

7.0

7.2

7.5

7.8

A4

5.5

5.9

6.0

6.1

A5

6.0

6.3

6.5

6.6

A6

6.0

6.4

6.5

6.6

In this design, a different heart was used under each ­condition, hence, there are a total of 24 independent measurements. The ANOVA table tabulates a number of sources of variance including; between group variation (i.e. day, apparatus) and within group variation (i.e. error). Since there are no replicate values in each cell, no interaction term is necessary: Between-subjects effects SS

df

MS

F

Sig.

Apparatus

29.5

5

5.901

1,187

<0.001

Day

1.468

3

0.489

98.4

<0.001

Error

0.075

15

0.005

Total

31.05

23

The ANOVA summary table shows that pD2 values differ depending on the day the experiment was performed (p < 0.001) and the apparatus that was used (p < 0.001). Example 9  A two-way ANOVA with repeating factors The −log pD2 values of a range of compounds at receptor X was determined using a 96-well plate format (see table). pD2 values were determined in triplicate and the mean values tabulated under each treatment condition. Experimental compounds were dissolved in different solvents (S1, S2) and tested on two different days (day 1, 2). Since cells from the same cell line were used for a compound under different treatment ­conditions, the data derived for each compound can be considered a (continued)

Statistical Methods in Research

459

Example 9  (continued) matched-pair (i.e. n = 1 not 4 for each compound) or alternatively, blocked design. The null hypothesis that pD2 values are not affected by solvent or day of assay can be tested. Day 1 Day 2 S1

S2

S1

S2

Compound 1

8.9

9.2

9.4

9.5

Compound 2

7.0

7.3

7.5

7.8

Compound 3

7.0

7.2

7.5

7.8

Compound 4

5.5

5.9

6.0

6.1

Compound 5

6.0

6.3

6.5

6.6

Compound 6

6.0

6.4

6.5

6.6

In this example, cells from a cell line are randomized to receive the same compound but under two treatment conditions (solvent, day). In this case, a mixed design has been employed to test the fixed factors, solvent and day on agonist potency. Do not confuse with a two-way independent ANOVA. The ANOVA table has to account for a number of sources of variation; within-subject for each factor (solvent and day in this example) separately, interaction between the factors, total variation and error variation (not shown). Because this is a mixed design, the main interest is the within-subjects effects. The important variance terms are shown: Within-subjects effects SS

df

MS

F

Sig.

Solvent

0.350

1

0.350

247

<0.001

Day

1.084

1

1.084

161

<0.001

Solvent*day

0.034

1

0.034

5

0.073

The ANOVA summary table shows that pD2 values differ depending on the solvent used (p < 0.001) and the day of the experiment (p < 0.001). test. However, if following an ANOVA an omnibus F value gave a p value close to 0.05 but with small group sizes then one might treat the results with caution and you have to make an individual decision whether to reject the null hypothesis based on your experience of similar experiments. A particular problem concerning repeated measures ANOVA is a violation of sphericity (homogeneity of covariance) and this leads to bias and inflation of the type I error rate, and this issue

460

Spina

has been discussed in detail elsewhere (11, 13, 16, 17). Statistical software packages like SPSS™ (but not GraphPad Prism™) test for violation of sphericity and automatically run a more conservative analysis using various correction factors (e.g. Greenhouse– Geisser adjustment). Most statistics packages test for homogeneity of covariance (SPSSTM uses Mauchly’s test of sphericity; see Example 7). In this example, the Mauchly’s statistic yielded a p value of 0.014, and hence an appropriate adjustment is required. In this case, a Greenhouse–Geisser correction was chosen and a significant time effect was still observed (F [1.885, 20.7] = 26.74, p < 0.001). Note the df values are smaller than those presented in the table. The test is now more conservative. In Example 9, both solvent and day can be considered repeating factors and from the analysis we note that there is a significant difference in agonist pD2 depending on the solvent employed and day of the experiment. Since there is no significant interaction (solvent*day), one can rightly conclude that the solvent effect is consistent over time. Some statistics packages plot the group means as a visual display of the results. In this case, you would get two parallel lines. Where there is evidence of interaction, the lines intersect or approach each other. It might be tempting to undertake a post hoc test to compare differences between mean groups, but this is redundant for a 2 x 2 analysis particularly as there is no significant interaction. It has been argued that for time-response relationships, a suitable summary statistic (e.g. area under the curve, peak height) should be calculated for each time-response curve and the data then analyzed using t-tests for two groups or ANOVA when there is more than two groups (16), although there is the potential for a loss of sensitivity and a failure to detect a treatment effect when using this approach (17). Both of these approaches for analysis of serial data are valid and the choice depends on the experimental question and the judgement of the experimenter. An alternative approach not discussed here is the use of linear mixed models for repeated measures (11) If you expect differences between groups to follow a specific order (e.g. in a dose-response or a time-course), it is possible to apply a post-test for trend, rather like the procedure for linear regression, rather than testing pairs of groups. This tests whether there is a trend for mean values to increase or decrease as you move through the groups.

4. Non-parametric Methods Where data clearly do not fit a Gaussian distribution there are several possibilities. You can use parametric tests on the basis that they are robust to deviations from normality. Alternatively, you can transform the data to fit a normal distribution so that parametric

Statistical Methods in Research

461

tests may be used (Table 1) or you can use distribution-free methods, the so-called non-parametric statistics. Non-parametric tests make no assumptions about data distributions. They are less powerful than the corresponding tests based on the normal distribution so significant results obtained using them might in one sense be regarded as more reliable. The usual non-parametric equivalents to t-tests are the Mann–Whitney U test (for independent samples; see Example 10) and the Wilcoxon test (for paired data). For multiple comparisons of independent samples the Kruskal– Wallis test (a non-parametric equivalent of a one-way ANOVA, which is a generalization of the Mann–Whitney U test for >2 groups) may be used (see Example 11), with Dunn’s test as a posttest. Post-tests for non-parametric ANOVA are not available in many statistical packages but the calculations are not too involved. Methods may be found in the classic text of Siegel and Castellan (18) or a more recent book by Neave and Worthington (19). These tests rely on ranking data values and comparing ranks within groups to assess whether or not they differ. If you have category data or proportions you may find Chi square tests applicable. The usual applications for Chi square, tests for proportions, or for goodness of fit. Where the assumptions of a Chi square test for proportions are violated (usually because the group sizes are too small making expected values <5) an alternative is Fisher’s exact test, which can, as its name suggests, calculate exact probabilities. Example 10  Mann–Whitney U test results for response times in mice In the experiment which yielded the control data presented in Example 2, a further group of mice treated with 9 mg/kg of a k opioid agonist had the following response times (s) to the same stimulus: 60, 60, 48.8, 46.0, 60, 60, 60. Here, median (range) = 60 (46.0–60) s, showing a very skewed data set, hence the decision to use a non-parametric method for analysis. Comparison of the 2 groups using a Mann–Whitney test gave the following results (taken from SPSSTM output): Response

N

Mean rank

Sum of ranks

Control

7

5.29

37.0

k Agonist

7

9.71

68.0

Total

14

Mann–Whitney U 9.000 2 Tailed Sig. 0.034 This test uses the ranks of data values to compute the test statistic, not the actual values. The final p value suggests that response times in the two groups were significantly different.

462

Spina

Example 11  Kruskal–Wallis ANOVA applied to haematology data The following T lymphocyte cell counts were obtained in an experiment comparing the responses of groups of mice to recombinant a thymosin. Since cell counts are regarded as not being normally distributed, a non-parametric ANOVA was used. Control

1 mg thymosin

5 mg thymosin

10 mg thymosin

609

700

678

800

584

569

987

578

573

790

945

593

711

987

934

493

456

800

888

579

Kruskal–Wallis results (from SPSSTM): Test statistic

df

Significance

8.730

3

0.0331

This indicates that the groups differ. GraphPad Prism™ gives identical results for the ANOVA and offers Dunn’s post-test, with the following results (comparison of selected groups), Control vs 1 mg thymosin p > 0.05 Control vs 5 mg thymosin p < 0.05 Control vs 10 mg thymosin p > 0.05 Thus, the 5  mg thymosin dose appeared to be the only dose which altered T cell counts in these mice.

5. What Is the Optimal Size of an Experiment?

If attempts have been made to improve sensitivity and reduce variability, then sample size and, by default, df available to estimate experimental error will determine the precision of the experiment. To illustrate this in graphical terms, Fig. 2 represents the critical values of t at the 5% probability level as a function of df. At 10 df, the t value is 2.23 which falls to 2.09 with 20 df. As we can see from this figure, df greater than 10 does not yield a substantial reduction in the critical value of t. Thus, for comparisons between two samples, a minimum of six experimental units (independent data values) per group should be employed. If one can achieve a df for error (denominator) between 10 and 20, then one will obtain high power; any number above this value

Statistical Methods in Research

463

Crtical values of t 0.05

15

10

5

0

0

10

30 40 50 20 degrees of freedom (n−1)

60

70

Fig. 2.  Relationship between critical values of t and degrees of freedom. Shaded area highlights t values for degrees of freedom between 10 and 20. No substantial changes in the magnitude of the t value greater than 20  df. A t value of 1.96 is achieved at infinity.

will lead to only a slight increase in precision at the expense of wasting experimental units, which is particularly unethical if the units are animals. 5.1. Power of Experiments and Type I and I Errors

Hypothesis testing involves making a decision as to whether two samples are often from the same population. We normally set the significance level at which the null hypothesis is rejected at the small probability of 5%, and in this way avoid making wild claims that a treatment has had a significant effect. Rejecting the null hypothesis when it is in fact true is known as a Type I error. Type I errors are usually committed when multiple t-tests are applied between a control group and various treatment groups. If five “independent” comparisons are made, the probability of rejecting the null hypothesis is no longer 5% but increases to 23%. Thus, it is highly likely that a rejection of the null hypothesis will be made. In order to reduce the potential of a Type I error in this example, the threshold to keep the overall risk of a Type I error to 5% is now 0.0102. Thus, the null hypothesis could only be rejected if a probability of 1.02% or less was obtained. This is an example of a Bonferroni correction. On the other hand, by accepting the null hypothesis, you are stating that you have not found a big enough difference in your experiment to reject the possibility that the difference arose simply by chance. So, if in your experiment you fail to find a significant difference, it could be that there is no difference between sample means, or there is a difference, but you have missed it. It is this latter case that is commonly referred to as a Type II error (i.e. failure to detect a difference when it in fact exists). This is denoted by b. In a perfect world, sample distributions would not overlap, and we could avoid making both Type I and Type II errors, but in the real world we can only try to balance the two.

464

Spina

5.2. Power of a Test



In order to observe a statistical difference between treatments, your test requires sufficient power: Power of a test = 1 - b. Statistical power is the complement to a Type II error – the probability of detecting a given effect in a sample if it actually exists in the population. Techniques of power analysis are important in planning investigations, particularly to help determine necessary sample sizes to detect particular effect magnitudes. By convention, an acceptable level of statistical power is set to 0.80 (80%). Power is dependent on several factors, including the significance level (alpha: conventionally set to 0.05); sample size (n); the population standard deviation (s); and the effect size one wishes to detect and depends on the nature of the experiment i.e. how small a difference between groups you would consider to be scientifically significant, or how small a difference your experimental system can actually detect. How these parameters are precisely linked to power is a function of the statistical test to be used. Hence, you must have decided on the method of analysis before you start. Because of the complexity of the calculations, it is usual to use proprietary software, such as nQuery Advisor™ or the freeware statistical programme, G*Power3 (20) to calculate n, the minimum number of subjects or replicates for a particular effect size. This is mandatory when planning clinical trials, to avoid subjecting patients to needless risk by undertaking a study which is either too small to detect anything of interest or too big and thus a waste of resources. A similar position could become standard in laboratory experiments, given pressure on time and resources. For most laboratory investigations, the population variance and hence SD are not known making this calculation difficult. Possible solutions might be to estimate the population SD based on pilot experiments and an example of this approach has been described (21). Alternatively, the population variance might be obtained from other published studies in the appropriate area (i.e. prior knowledge). Examples of this approach and comprehensive ­discussion of this matter can be found in the public domain (14). Alternatively, Mead (22) suggested a rule of thumb resource equation, which is somewhat easier to apply, viz:



E = N - T - B,



where E = error degrees of freedom, N = total degrees of freedom (number of experimental units minus 1), T = treatment degrees of freedom (number of treatment groups minus 1), and B = block degrees of freedom (number of blocks minus 1).

Statistical Methods in Research

465

This approach is valid only for data conforming to a Gaussian distribution capable of being analyzed by parametric methods, such as t-tests or ANOVA. The desirable value for E is between 10 and 20, based on the distribution of critical values for t as shown in Fig. 2, where it is seen that increasing the df above 10 has very little effect on critical values of t. This approach does not take into account any consideration of effect size, unlike the full power analysis (see Example 12). In most cases in research, it may be of more use to use power analysis retrospectively after a preliminary experiment to answer the question: “Given the observed variability in the experiment, what is the smallest change that we could be expected to identify?” You could substitute the observed effect size and SD in the appropriate power equation to work out the power of your experiment. It is then possible to choose group sizes for follow-up studies to have a good chance of performing an experiment that has 75–80% power. This “reversed” power analysis may be used to justify publication of negative results if it shows that the sample size used was sufficient to demonstrate a specified effect and yet none was detected.

Example 12  Using the resource equation to estimate an ­appropriate n number for different experimental protocols In a comparison of cellular responses to drug treatment using one control and one treatment group (suitable for analysis by independent samples t-test), we have B = 0 and T = 1. Therefore, assuming the error degrees of freedom, E = 10 and solving Mead’s resource equation, yields N = 10 + 1 = 11. Therefore, the total degree of freedom is 11 and hence, one would require six experimental units (animals, 96-well plates, etc.) per group. Any fewer and E would be too small, whereas using more might be a waste of resources without providing any extra information. If one wanted to test two other drugs, then two additional control groups would be required and would be inefficient and wasteful (i.e. a total of 36 experimental units). If the experimental protocol was changed to one comparing three different treatments with a control (requiring a oneway ANOVA analysis), T = 3 and, in order to obtain E ³ 10, N needs to be at least 13. You would probably elect to have a balanced design (equal numbers in each group) and so have N = 15 (i.e. four experimental units per group). Using a factorial design, a total of 16 rather than 36 experimental units would be a more efficient experimental design and provide sufficient degrees of freedom to detect a difference.

466

Spina

6. Statistical Aspects of Line and Curve fitting 6.1. Choosing a model

6.2. Telling if Data Fit Your Model and Comparing Data Sets 6.2.1. Linear Regression

Before the advent of personal computers, most models for experimental data used transformations so that the data could be fitted to straight lines, the easiest manipulation to carry out, using the least squares method. These methods are now regarded as inappropriate since the transformed data often violate the assumptions of linear regression and hence leads to inaccurate values for parameters, such as Kd, Bmax, etc., in radioligand-binding experiments (5, 23). Software packages make it possible to use non-linear regression to fit any data to a bewildering variety of equations with ease – the problem is not how to perform the analysis but rather, how to choose an appropriate model and then to tell if the results indicate a good fit. Choosing a model for any particular data set must depend on custom and practice in that area and accepted methods. If your curve-fitting package does not contain the equation you require as a built-in function, you will have to add it as a user-defined option. To start the curvefitting process, which is carried out by a process of iteration, you need to supply initial values for parameters of the curve. Most packages automatically generate such values for their built-in equations based on the range of data provided, but you may need to inspect these and decide if you could improve on them before starting the curve-fitting process. If you use a user-defined equation, you may have to provide the initial values. The iteration process is automatic to converge on the best fit based on the model that you specify, but the software has no knowledge of the biology of the system you are modelling and does not identify any clearly impossible values. Consequently, if there are values that you want to fix (e.g. if modelling exponential decay, the curve must plateau at zero, or you have a set of results where you wish to fix the maximum response to 100%) you should do so at the outset. For further discussion of appropriate methods for receptor studies, see Kenakin (23) or Motulsky (5) and earlier chapters in this book. The least squares method for linear regression minimizes the squares of the vertical distances of each data point from the regression line and finds the slope and intercept of the straight line which best fits the data, assuming a linear relationship. Thus, any data entered into regression software generate values for these parameters (see Example 13). To tell if the fit is good, one needs to examine other statistics from the regression. The Pearson correlation coefficient, r, is often quoted, but a better parameter is the coefficient of determination, the square of this quantity, r2 and measures the strength of the relationship between the independent and dependent variable. This is a measure of the fraction of variation in y which may be explained by its dependence on x.

Statistical Methods in Research

467

It is computed by comparing the ratio of the sums of squares of distances of y values from the regression line with the sums of squares of deviations from the null hypothesis line (y = 0):

r2 =

SS (regression ) . SS (total )

Absorbance at 540 nm

The better the data fit a straight line, the closer r2 is to 1.0 An F test of the regression (i.e. MSregression/MSresidual) is used to determine whether the regression model is a better prediction of the values of the independent variable than if there was no such relationship. Alternatively, the F ratio tests the null hypothesis that the overall slope of the line is zero. A p value <0.05 for this test is additional evidence for a linear relationship between x and y. Most statistics packages give this output as an ANOVA table in the form we have already seen for a one-way ANOVA. An alternative approach is to use t-tests to test the null hypothesis that the slope and the intercept are equal to zero (this would imply no linear relationship in the data). Again, this is often standard output from a statistics package and you have to look for p values <0.05 in these tests to confirm that the calculated coefficients are not zero. The p value for the slope, of course, duplicates the information given by the F test. A further confirmation is to inspect the 95% CIs for the coefficients, if given, since these will not span zero if the relationship is linear. The 95% confidence intervals for the regression line give limits within which the true line may lie. From a plot, it may be seen that these are curved (Fig.  3), demonstrating how this technique places most emphasis on data points nearest the mean values. A plot of residuals (vertical distances for each data point from the regression line, which are easier to see from a graph than from

expt1 expt2

1.0 0.8 0.6 0.4 0.2 0.0 0.0

0.1

0.2

0.3

0.4

0.5

Salicylate (mg/ml)

Fig. 3. The two regression lines (with 95% confidence intervals) from Example 13, showing the difference in slopes in each experiment (expt).

468

Spina

a table of the figures) should show points randomly scattered about y = 0, not varying with x. There should be no clusters of adjacent points all above or below the line. A runs test ­(commonly available as an option for output) highlights such clusters, by ­calculating the probability of observing as few or fewer runs or ­clusters of points as appear in the data. If the p value for the runs test is <0.05, one may conclude that the data does not fit a straight line. Alternatively, one could calculate the variance due to the deviations from linearity to ascertain whether the regression deviates from linearity (see Example 17.8 in Zar (15)). 6.2.2. Comparison of Several Regression Lines

Zar (15) gives a method for comparing the slopes and intercepts of regression lines using analysis of covariance (ANCOVA). This analysis is available as an option for linear regression in GraphPad Prism™. The starting point is to compare the slopes using an F test. If the resulting p value is <0.05, the lines have different slopes and are not identical. Applying a suitable post-test (such as Tukey’s test to compare pairs of lines, or Dunnett’s test if one of the lines is a control and the other lines are to be compared with it) will allow you to identify the location of the difference(s). If, on the other hand, the slopes are not significantly different (i.e. p > 0.05), then it is possible to use a further F test to compare the intercepts to decide whether the lines are identical (i.e. the intercepts are not significantly different). If p for this comparison is <0.05, then one can conclude that the lines are parallel but separate and a post-test as above will identify the differences. ANCOVA is used to test the null hypothesis that several means are the same but controls for the effect of a covariate (in the case of Example 13, the covariate is salicylate concentration). The two lines are shown in Fig. 3 with 95 CIs. The data were described by a linear relationship and intercept and slope values were calculated (see Example 13). We can test the null hypothesis of whether the two slope values for these lines are similar using ANCOVA (in SPSS). A test of homogeneity of regression slopes is performed (not shown), and we find an F value of 32.79 (df 1,32) and for this, p < 0.001 i.e. the slopes (and therefore the lines) are significantly different, and there is no reason to compare the intercepts. Thus, the results of this assay vary from day to day, and it is always necessary to include standards.

6.2.3. Goodness of Fit for Non-linear Regression

Curve-fitting software fits any data to a non-linear equation which is selected based on an understanding of the relationship between the independent and dependent variable (e.g. hyperbolic function to estimate Kd and Bmax values from a binding experiment). The programme generates best fit values for curve parameters, and these should be within the expected range(s). Parameter estimates (e.g. EC50) which are obviously erroneous (e.g. negative values or values off the scale of measurement) suggest that the data are not a good fit. The reason may be that there are too few points in the

469

Statistical Methods in Research

Example 13  Linear regression and comparison of two lines An assay for salicylate gave the following results for absorbance at 540 nm: Absorbance at 540 nm Salicylate conc. (mg/mL)

1

2

3

0

0

0

0

0.1

0.175

0.189

0.201

0.2

0.360

0.390

0.400

0.3

0.500

0.580

0.596

0.4

0.650

0.750

0.800

0.5

0.990

0.968

0.941

The triplicate samples had been prepared individually rather than being repeated measurements on one sample so were treated as independent determinations in the analysis. The linear regression gave the following results for coefficients (±SEM): Slope = 1.898 ± 0.0512 Intercept = −0.002762 ± 0.01568 R2 = 0.9982

The slope significantly deviated from zero (F  [df 1, 16] = 1,343, p < 0.0001). The runs test for deviations from linearity gave p = 0.783 Thus, the data fit a good straight line, as we would hope for assay standards. A second experiment carried out later under the same conditions gave the following results: Absorbance at 540 nm Salicylate conc. (mg/mL)

1

2

3

0

0

0

0

0.1

0.158

0.162

0.146

0.2

0.324

0.323

0.291

0.3

0.465

0.483

0.461

0.4

0.640

0.641

0.606

0.5

0.760

0.831

0.748

Results as before: Slope = 1.564 ± 0.027 Intercept = 0.002698 ± 0.008114 R2 = 0.9953

The slope significantly deviated from zero (F  [df 1, 16] = 3,405, p < 0.0001 The runs test for deviations from linearity gave p = 0.782.

470

Spina

region of the curve closest to the EC50 and hence the need to repeat the experiment to obtain better data. Assuming the estimated parameters and the generated graph look reasonable, then values for r2, calculated from the sums of squares of deviations from the curve as for linear regression and plots of residuals and runs test results can tell you if the fit was close to the chosen curve. 6.2.4. Comparison of Several Curves

Although it is possible to use a two-way ANOVA with appropriate post-tests to compare dose-response or binding curves, the results may not be easy to interpret because they ignore trends in the data and also depend on sample size. Instead, Motulsky (6) recommends using non-linear regression to fit curves and comparing the derived parameters. Just as with linear regression, you can use F tests to determine whether slopes and derived parameters are the same. If any of these tests give a significant result (p < 0.05), a suitable post-test will identify the differences. It is also possible to use an F test to compare the goodness of fit of two different models (e.g. one-site vs two-site binding).

7. Choosing a Statistics Package Before embarking on any statistical analysis, it is good practice to inspect your data. This usually means plotting it in some way and looking for trends, obvious deviations from expectation or anomalies. Thus, using a graphics package to try out different forms of presentation is a good first step. Modern graphics software is mostly user-friendly, often includes statistical routines and may be all that is needed for your data analysis. GraphPad Prism™ has the advantage that it was specifically designed for presentation and analysis of pharmacological results. It also has particularly good World Wide Web support for users (see Web Resources at the end of this chapter). Many scientists make widespread use of spreadsheets for data manipulation and presentation. The graphics capabilities of these packages are generally of sufficient quality for scientific presentation. They also have some statistical routines built in, although these tend to be limited. One of the most widely used spreadsheets, Microsoft Excel™, comes with a data analysis add-in option which does limited statistical analyses, including t-tests, correlation and linear regression, and one-way and two-way ANOVA (for balanced designs only) but does not include routines for post-tests or nonparametric analyses. Various commercial packages developed as Excel add-ins extend these capabilities. Spreadsheets and graphics packages cannot usually cope with more complex analyses or in some cases with unbalanced

Statistical Methods in Research

471

designs (unequal group sizes) and missing data values. On the other hand, dedicated statistics packages perform more tests than most users ever need and allow graphical representation of data, although this is rarely in a suitable form for publication. Most traditional statistics packages have evolved over time from mainframe versions (e.g. SPSS™; Minitab™), although it is advisable for the novice, to obtain some training on how to use these programmes. Packages written originally for PCs (e.g. Systat™, Statgraphics™, Origin™) can take some time getting used to. The best advice is to get hold of one of the specialized texts dealing with your particular software, choosing one relevant to your discipline. These are often easier to understand than the official manual. You may well be constrained to use whatever package is supplied on your in-house network. If you need to do particular analyses routinely, there may be specialized software available (e.g. nQuery Advisor™ for power calculations, Arcus ProStat Biomedical™ for analysis of clinical trial data) or you may choose a package that allows you to customize analyses. In the end, the choice of software is a personal one, balancing your needs with the availability of user support, budgetary constraints, and user-friendliness.

8. Web Resources Software houses tend to have useful Web sites, with the support for users. In particular, the GraphPad™ site (www.graphpad.com) offers advice, downloadable manuals and links to other sites of interest. Modern textbooks often have supplementary information available on the relevant publishers’ Web sites. A useful site for freeware for Power analysis is G*Power3 (www.psycho. uni-duesseldorf.de/abteilungen/aap/gpower3/). References 1. Huff, D. (1955) How to lie with statistics Penguin Books, London 2. Ludbrook, J. (2008) Statistics in biomedical laboratory and clinical science: applications, issues and pitfalls. Med. Princ. Pract. 17, 1–13. 3. Lew, M.J. (2008) On contemporaneous controls, unlikely outcomes, boxes and replacing the ‘Student’: good statistical practice in pharmacology, problem 3. Br. J. Pharmacol. 155, 797–803. 4. Hancock, A.A., Bush, E.N., Stanisic, D., Kyncl, J.J., and Lin, C.T. (1988) Data normalization before statistical analysis: keeping

5. 6. 7. 8. 9.

the horse before the cart. Trends Pharmacol. Sci. 9, 29–32. Motulsky, H.J. (1995) Intuitive biostatistics. Oxford University Press, New York Motulsky, H.J. (1999) Analysing data with Graphpad Prism. Graphpad Software Inc, San Diego Altman, D.G., Machin, D., Bryant, T.N., and Gardner, M.J. (2000) Statistics with confidence. BMJ Books, London Sokhal, R.R. and Rolf, F.J. (1995) Biometry. W.H. Freeman & Co, New York Petrie, A. and Sabin, C. (2000) Medical statistics at a glance. Blackwell Science, Oxford

472

Spina

10. Dyson, C. (2003) Choosing and using statistics. Blackwell Science, Oxford 11. Field, A. (2009) Discovering statistics using SPSS, 3rd ed., SAGE Publications Ltd, London 12. Klugh, H.E. (1986) Statistics: The essentials for research. 3rd ed., Lawrence Erlbaum Associates, London 13. Wallenstein, S., Zucker, C.L., and Fleiss, J.L. (1980) Some statistical methods useful in ­circulation research. Circ. Res. 47, 1–9. 14. Quinn, G.P. and Kenough, M.J. (2002) Experimental design and data analysis for biologists. Cambridge University Press, Cambridge 15. Zar, J. (1999) Biostatistical analysis. 4th ed., Prentice Hall International, New Jersey 16. Matthews, J.N., Altman, D.G., Campbell, M.J., and Royston, P. (1990) Analysis of serial measurements in medical research. Brit. Med. J. 300, 230–235.

17. Ludbrook, J. (1994) Repeated measurements and multiple comparisons in cardiovascular research. Cardiovasc. Res. 28, 303–311. 18. Siegel, S. and Castellan, N.J. (1988) Nonpara­ metric statistics for the behavioural sciences. 2nd ed., McGraw-Hill International, New York 19. Neave, H.R. and Worthington, P.L. (1988) Distribution-free tests. Unwin Hyman Ltd, London 20. Faul, F., Erdfelder, E., Lang, A.G., and Buchner, A. (2007) G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav. Res. Methods 39, 175–191. 21. Festing, M.F. (2003) Principles: the need for better experimental design. Trends Pharmacol. Sci. 24, 341–345. 22. Mead, R. (1988) The design of experiments. University Press, Cambridge, New York 23. Kenakin, T. (1996) Molecular pharmacology: a short course. Blackwell Science, Oxford

Index A Abbe’s law....................................................................... 426 Acceptor....................................54, 299, 300, 305, 311, 312, 314, 358–363, 367–369, 374, 375, 377, 380–385 Acceptor-ligand............................... 375, 379, 380, 382–385 Adenovirus................................................39–42, 46, 49, 50 Adenylyl cyclase............................................................. 298 a2-Adrenergic receptor................................................... 349 Affinity..............................7, 54–56, 95, 100, 135–137, 147, 157, 158, 162, 163, 166, 196–202, 204, 205, 207, 208, 211–213, 215, 217, 221, 232, 239, 247, 274, 283, 285, 286, 302, 303, 330, 331, 354, 376, 390, 391, 393–395, 399–408, 411, 412, 415, 428, 438 Affinity cooperativity.......................................197, 202, 207 Agarose gel..............................44–46, 60, 69, 71, 72, 74–75, 82, 83, 90, 92, 93, 96, 176 Agarose gel electrophoresis.............................62, 64, 74–77, 92, 174, 176 Agar plates.......................................................63, 73, 92, 94 Agonist-directed trafficking........................................... 268 A-kinase anchoring proteins (AKAPs)...................298, 302, 304, 347 Alanine scan..........................................................86, 87, 95 Allosteric...................................................85, 195–208, 298 Allosteric agonism.......................................................... 204 Alternative splicing......................................................... 169 Amplicon......................................... 167, 179, 180, 182, 184 Amplification..................................... 10, 47, 80, 91–93, 95, 96, 139, 165–168, 171, 177, 183, 184, 258, 268, 291, 318, 403 Analysis of covariance (ANCOVA)................................ 468 Annealing..................................... 33, 59, 66–68, 76, 80, 95, 167, 168, 178–180, 183 Antibiotic....................................8–9, 11, 12, 14, 19, 23, 34, 36, 45, 63, 65, 70, 71, 92, 94, 102, 103, 107, 111, 254, 337, 349, 351, 377, 401, 420 Antibiotic resistance........................................12, 45, 63, 92 Antibody(ies)...............................31, 53, 102, 118, 166, 238, 252, 264, 298, 323, 375, 392, 400, 415, 426, 452 Antibody production.............................................. 246–247 A priori tests........................................................... 451–454 Arrestin.............. 86, 100, 330, 358, 359, 361, 363–367, 369

Association constant............................................... 154, 363 Autocorrelation............................... 226–229, 232, 234, 235 Autofluorescence......219, 222, 230, 234, 235, 288, 313, 340 Autoradiogram............................................................... 242 Autoradiography...............135, 243, 264, 350, 351, 353, 418

B Background subtraction.................................................. 294 Bait.................................................. 390, 392, 394–397, 400 Bartlett’s test................................................................... 451 b-galactosidase............................................23, 33, 126–130 Biocytin maleimide........................................................... 87 Bioluminescent resonance energy transfer (BRET)...............................................357–370, 374 Biosensors................................ 105, 106, 297–315, 329–342 Biotin..........................54, 117, 400, 405, 428, 430, 436, 440 Biotinylation............................ 428, 430–431, 435–438, 440 BiP......................................................................................427 Bit depth.................................................281, 285, 290–292 BLAST.............................................................91, 171, 180 Blasticidin........................................ 9, 12, 23, 24, 27, 29–33 Bleaching...................................57, 229, 235, 285, 287, 288, 293, 310, 313, 314 Bleed-through................................. 220, 312, 340, 383, 438 Blunt cloning.............................................................. 90, 94 Bmax.................................... 136, 144, 145, 159, 428, 466, 468 Bonferroni.......................................................455, 456, 463 BRET2............................................358–360, 362–365, 369 BRET3.................................... 358–360, 362, 363, 368, 369 BRET ratio.....................................................363, 366, 369 Butyrate............................................................................ 19

C Ca2+ indicator...........279–281, 283–287, 289, 290, 293, 294 Ca2+ ionophore....................................................... 279, 280 Calcium.............. 55, 117, 196, 277–295, 376, 378, 379, 400 Calcium phosphate..............................................8, 413, 416 Calf intestinal alkaline phosphatase (CIAP)................................................240, 247, 248 Calibration of fluorescence..................................... 285, 294 Calmodulin affinity chromatography..................... 404–406 Calnexin......................................................................... 427

Gary B. Willars and R.A. John Challiss (eds.), Receptor Signal Transduction Protocols: Third Edition, Methods in Molecular Biology, vol. 746, DOI 10.1007/978-1-61779-126-0, © Springer Science+Business Media, LLC 2011

473

Receptor Signal Transduction Protocols 474  Index

  

Cameleons...................................................................... 278 Camgaroos..................................................................... 278 Ca2+ puff................................................................. 282, 283 Cardiomyocytes........307, 308, 318, 319, 321–323, 325, 327 Ca2+-sensitive dye............................ 288, 329, 333, 338, 340 Caveolae................................................................. 411, 412 Caveolin.......................................... 412, 415, 417–419, 421 cDNA..................................3, 4, 6–8, 10, 21, 31, 32, 41–43, 45, 46, 53, 62–69, 71, 76, 79, 102–105, 107, 108, 110, 139, 165, 175–180, 185–188, 302, 319, 333, 337, 348–350, 359, 361, 365, 367 Cell fixation...................................................... 57, 429–434 Cell lysate........................................... 31, 35, 124, 125, 140, 256, 299, 324, 347, 396 Cesium purification.......................................................... 47 Chemiluminescence........................................164, 255, 437 Cheng–Prusoff................................................147, 148, 161 Chimeric mutagenesis.....................................89–90, 92–94 Chi square test................................................................ 461 Cholera toxin B (CTB)...........................414, 417, 418, 420 Cholesterol.............................. 411, 412, 415, 419, 421, 422 Chromosomal integration................................................ 18 Clonal isolation.......................................................... 12–14 Cloning rings.............................................................. 32, 36 CMV......................................................5, 7, 33, 43, 256, 402 Codon......................................32, 34, 55, 62, 64, 66–69, 73, 79, 81, 82, 90, 123, 361 Coelenterazine................................. 357, 358, 360–366, 368 Competence............................................................. 96, 121 Competent cells..................43–44, 63, 68, 80, 83, 89–92, 94 Competition binding...............................198, 211, 224–225 Competitive antagonist.......................................... 196, 200 Concanavalin A.............................................................. 427 Confidence interval (CI).........................202, 205, 447, 467 Confocal fluorescence microscopy.......................... 426, 434 Confocal microscopy..............................105–107, 122, 212, 218, 283, 339, 361, 439 Coomassie.......................................................129, 245, 321 Cooperativity factor.................................197, 198, 204, 205 Co-transfection................................... 29, 34, 104, 105, 302 Cq........................................................................... 184–186 Cryopreservation.............................................................. 34 Curve fitting.................................... 198, 234, 364, 466–470 CXCR4...................................................252–254, 256–258 Cyan fluorescent protein (CFP).......................25, 300–303, 305, 306, 309–314, 330, 374 Cyclic 3′,5′-adenosine monophosphate (cAMP)....................... 214, 297–305, 309, 311–314

D Data transformation........................ 145, 445, 446, 460, 466 DeepBlueC............................................................. 358, 360 De-esterification............................................................. 280 Degrees of freedom......................... 148, 450, 452, 463–465

Desensitization.................................. 58, 304, 359, 390, 399 Detergent extraction............................................... 411, 419 Detergent-insoluble glycolipid-enriched membranes......................................................... 412 Detergent-resistant-membranes (DRMs).............. 411–422 Diacylglycerol (DAG)..................... 318, 330, 336, 338, 342 Diagnostic digestion............................................. 70–72, 75 Dialysis............................................................321, 326, 352 Dichroic mirror..............................................220, 228, 278, 287, 306, 309, 313 Diffusion time........................................................ 228, 229 Dimerisation..................................................................... 31 Dispersion.......................................................444, 446–448 Dissociation constant............................. 136, 137, 145, 197, 198, 200, 201, 203 DNA concentration.........................................64, 71, 72, 75 DNA extraction................................................................ 75 DNA polymerase..........60, 67, 76, 80, 89, 91, 166–168, 177 DNase.................................................. 78, 82, 174–176, 178 Dominant negative.................. 100–102, 105, 107, 110, 257 Donor........................................299, 300, 305, 311, 312, 314, 358–361, 363, 367–369, 374–376, 380–385 Donor-ligand................................... 375, 379, 380, 382–385 Down-regulation............................................................ 428 Doxycycline........................5, 7, 9, 15, 22, 24–27, 29, 31, 35 Dropout mix................................................................... 116 Dual affinity tag............................................................. 402 Duncan’s multiple range test.......................................... 456 Dunnett’s test......................................................... 456, 468 Dunn’s test.............................................................. 461, 462 Dwell time...............................................226, 227, 230, 235 Dynamin........................................................................ 257

E Ecdysone................................................... 6, 7, 9, 10, 15, 18 Edman degradation.........................................238, 243–245 EEA1�����������������������������������������������������������������������������427 Efficacy............................ 5, 86, 95, 196, 197, 204, 205, 207, 208, 211–215, 217, 221, 232, 264, 282, 408 Efficacy cooperativity..............................197, 204, 205, 208 Electrophoresis............................. 31, 62, 64, 65, 71, 74–77, 82, 92, 118, 174–176, 241, 242, 324, 347, 350, 390, 392, 394, 397, 405, 408, 414, 437 Electroporation................................. 8, 43, 44, 49, 104, 111, 121–122, 323, 376–378, 384 ELISA...............................................................58, 247, 428 Emission spectra............................. 213, 216–221, 286, 312, 338, 358, 365, 368 Endocytosis..................................... 257, 412, 427, 432, 435 Endonuclease............................ 10, 60, 75–76, 78, 168, 169 Endosomal sorting complex required for transport (ESCRT)............................................................ 252 Endosome.......................................................139, 251, 427 EnduRen.................................................358, 360, 362, 368

Receptor Signal Transduction Protocols 475 Index     

Enhanced GFP (eGFP)........................... 57, 60, 62, 68, 69, 71, 72, 105–108, 110, 111, 331–334, 336, 339, 340, 342, 358, 360, 363 Epac.................................................................300–305, 308 Epifluorescence.......................................305, 426, 434, 439 Epitope tag..................................... 24, 31, 53–83, 118, 126, 239, 248, 257, 426, 438 Error......................................... 160, 271, 443, 446, 449–451, 453–459, 462–465 Ethidium bromide...............................................74, 83, 190 Europium............................................................... 374, 375 Excitation spectra................................................... 220, 299 Experimental design................ 188, 443, 444, 450, 456, 465 Experimental power....................................................... 463 Expression...........................3–19, 21–36, 39–50, 53, 86, 99, 114, 165, 195, 237, 253, 266, 304, 326, 331, 349, 359, 374, 400 Extended BRET (eBRET)................................... 358–360, 362, 363, 368, 369

F Fisher’s exact test............................................................ 461 Fixation.....................................................57, 426, 429–435 FLAG-tag...........................................................26, 55, 253 Flotation................................................................. 411, 415 Flotillin–1................................................415, 417–419, 421 Flow cytometry....................... 287, 427–428, 430, 434–435 Flp-In™. ..................................................................... 21–36 Flp recombination target (FRT)................................ 22–25, 28–30, 33–35, 66 Fluo–3���������������������������������������������������� 282, 283, 288, 293 Fluorescence correlation spectroscopy (FCS).........212, 213, 218–219, 226–232, 234, 235 life-time.....................................................213, 305, 374 microscopy.................................. 57, 108, 111, 122–123, 212, 403, 406, 426, 427 polarisation....................................................... 212, 215 Fluorescent biosensor............................................. 329–342 Fluorescent ligand...................................211–235, 373–386 Fluorophore.....................166, 168, 169, 212–220, 226–230, 232–235, 299, 300, 302, 305, 308, 309, 312, 313, 330, 338, 341, 358, 361, 374, 375, 383, 433, 435 Fluoview................................................................. 332, 340 Focal plane...............................................222, 313, 314, 434 Focus drift.............................................................. 310, 314 Förster radius.................................................................. 299 Förster resonance energy transfer (FRET).............168, 213, 297–315, 330, 373–386 FRET ratio......................................................306, 315, 383 F-test...................................................................... 147, 148 Fura–2............................................. 284, 286–288, 290, 338 Fura-PE3........................................................................ 290 Fura red...................................................288, 332, 333, 338

G G12..............................................................................317–327 G418...................................................... 9, 12, 27, 28, 32, 36 Galactosyl transferase..................................................... 427 Ganglioside GM1.......................................................... 415 Ga-protein antisera........................................................ 272 Gaussian distribution...................... 445, 446, 448, 460, 465 GC clamp..............................................................81, 91, 95 GDP��������������������������������113, 114, 263–268, 270–273, 317 GelGreen........................................................................ 190 Gel overlay............................................................. 347–354 GelRed™������������������������������������������60, 72, 74, 83, 176, 190 GeneSwitch......................................... 6, 7, 9, 10, 15, 17, 18 Geneticin.....................................................27, 34, 401, 403 Genomic DNA (gDNA).........................167, 173, 174, 180 Glutathione sepharose.............................319, 320, 324–327 Glutathione-S-transferase (GST)............54, 247, 317–321, 323–326, 347–354, 390 Glycerol stock............................................................. 63, 71 Glycolipids..................................................................... 411 Gpa1....................................................................... 115, 128 GPCR-associated protein complexes.................... 389–397, 399–408 G protein.................................22, 86, 88, 95, 100, 113, 114, 237, 263–274, 298, 317–327, 330, 359, 389, 392, 396, 401, 406 G protein-coupled receptor kinase (GRK)............ 100–103, 105–110, 396 Gradient centrifugation.................................................. 125 GraphPad Prism™. ......................... 197, 201, 203, 206, 232, 335, 449, 453, 460, 462, 468, 470 Green fluorescent protein (GFP).................................. 26, 32, 43, 46–50, 54, 55, 57, 58, 122, 126, 213, 234, 302, 323, 330, 338, 359, 374, 426, 439 GSH-agarose...........................................348, 349, 352–354 GST-fusion protein.........................................317, 348–354 GTP.................. 113, 114, 201, 263, 264, 268, 273, 317, 323 GTPase............................................ 264, 302, 317, 318, 323 GTPase-activating proteins............................................ 114 [35S]GTPgS.............................................263–274, 317, 318 GTPgS-binding...............................................263–274, 318

H HA-tag...................................56, 66, 68, 69, 79, 81, 82, 253 Helmert contrast............................................................ 453 Heterodimer........................................................7, 374, 375 Heterologous expression................................... 99, 113–130 Hill slope.................................................146, 147, 150, 161 His-tag................................................................55, 56, 396 [3H]-N-methyl scopolamine (NMS).......197, 200–203, 208 Homodimer.............................................124, 129, 374, 375 HPLC........................................................................216–219 hRluc..................................................................................368

Receptor Signal Transduction Protocols 476  Index

  

Hydrophilic ligands........................................................ 139 Hygromycin.................... 9, 11, 12, 19, 24, 27, 28, 30–32, 34

I IC50...........................................................................146, 147 ICUE...................................................................... 301, 303 Imaging.....................................57, 105, 110, 213, 216, 217, 221–226, 230–234, 277–295, 300–308, 312–314, 329, 331–335, 337–341, 358, 360, 362, 412, 418, 426, 433, 434, 438, 439 Immobilized metal affinity chromatography (IMAC)...... 56 Immunisation................................................................. 247 Immunoblotting.................56, 118–119, 129, 252, 253, 257 Immunocytochemistry (ICC)..............................57, 58, 239 Immunoelectron microscopy.......................................... 427 Immunoprecipitate.................. 238, 248, 256, 258, 264, 273 Immunoprecipitation................................ 56, 247, 253, 256, 265, 267, 271–273, 374, 428, 436 Immunoreactivity.............................. 55, 427, 428, 435, 437 Indo–1........................................................................287–288 Inducible expression..........................................3–19, 32, 95 Inducible locus............................................................ 31, 32 Induction.............. 5, 6, 9, 10, 14, 15, 19, 22, 24–26, 35, 282 Infection..................................................................... 39–50 Infinite dilution.............................................................. 138 Inhibition...........................................................7, 106, 107, 136, 137, 139, 140, 146–151, 159–162, 177, 201, 203, 214, 302 Inositol 1,4,5-trisphosphate............................277, 318, 330 Intercalating dye...................... 166, 167, 169, 171, 183, 191 Internalization........................................ 139, 251, 252, 257, 359, 374, 380, 427, 428 Iodoacetamide.................................................246, 256, 431 IPTG...................... 6, 7, 10, 15, 94, 318, 319, 326, 349, 351 Isosbestic point............................................................... 287 ITRAQ........................................................................... 239

K K+1....................................................................136, 137, 159 K–1...........................................................136–138, 152, 153, 155, 159, 162, 451 Kd..................................................................................... 95 Ki............................................................................ 148, 376 Kill curve.................................................................... 34, 36 Kolmogorov–Smirnov test.............................................. 448 Kozak................................................... 10, 32, 64, 67, 68, 79 Kruskal–Wallis test................................................. 461, 462

L LacSwitch II....................................................6, 7, 9, 10, 15 LAMP...................................... 287, 306, 308, 312, 332, 427 Lanthanides............................................................ 374, 377 Least significant difference test...................................... 455 Levene test............................................................. 448, 451

Ligand binding..................................... 7, 10, 14, 16, 54, 58, 86, 87, 95, 115, 135–164, 195–208, 217, 221–225, 231, 232, 264, 265, 361, 376, 399, 403, 406, 407, 428, 445, 466 Ligase.....................33, 44, 59, 65, 68, 78, 79, 90, 93, 94, 252 Ligation.............................................. 29, 33, 65–66, 68, 69, 72, 78–82, 93, 94, 96 Linearization.........................................................18, 49, 65 Linear regression............................. 186, 187, 460, 466–470 Line fitting............................................................. 466–470 Line-scan.........................................................334–336, 341 Linker.............................................. 212–216, 302, 361, 367 Lipid rafts....................................................................... 411 Lipofection................................................................. 8, 337 Lipophilic ligands................................................... 139, 160 Localisation....................................... 31, 212, 217, 222, 280 Luciferase...................16, 299, 358, 360–366, 368, 374, 407 Lysine scanning................................................................ 87 Lysis................................................ 43, 45, 47, 50, 172, 245, 253–257, 299, 324, 325, 327, 348, 376, 378, 379, 401, 403, 431, 437 Lysosome.........................................................251, 252, 427

M Mam2.............................................. 115, 122, 124, 126–128 Mann–Whitney U test................................................... 461 Map3.............................................................................. 115 Mass spectrometry.......................... 216, 238, 239, 245–246, 319, 391, 394, 397, 400, 405–408 Melting temperature................................. 64, 80, 90, 92, 93, 96, 171, 182, 188 Membrane....................... 31, 41, 58, 82, 86, 89, 96, 99, 106, 119, 122–126, 129, 135–164, 195, 212, 214, 215, 217, 218, 222–227, 229, 231, 233–235, 242, 245, 248, 249, 251, 252, 255, 257, 258, 263–273, 288, 290, 297, 302, 303, 324, 325, 330, 331, 334–336, 339–341, 347–351, 353, 376–385, 389, 390, 395, 399, 403–404, 406, 407, 411–422, 425, 427, 428, 432, 439 Membrane preparation...........................135–164, 263, 264, 376–381, 385, 407 Membrane raft markers...........................414–415, 417–419 Membrane rafts............... 411, 412, 414–415, 417–419, 421 Metabolic labeling.................................................. 348, 428 MetaFluor....................................... 307, 308, 310, 314, 340 Metal ion bridge............................................................... 88 Methanethiosulphonate.................................................... 87 Methylated DNA............................................................. 96 Methyl-b-cyclodextrin (mbCD).................................... 422 M-factor......................................................................... 115 mfold..........................................................................171, 188 mGluR............................................................204, 419, 420 Microdomain................................... 226, 411, 412, 415, 439 Mifepristone........................................................6, 7, 10, 15 MIQE guidelines................................................... 171, 184

Models...............................3, 4, 14–16, 23, 89, 99, 114, 147, 148, 150, 184, 187, 197, 198, 200, 202, 204–208, 228, 229, 234, 332, 400, 414, 428, 460, 466–470 mOrange................................................................ 358–360 Multiple cloning site................................. 10, 28, 29, 62, 92 Multiplex PCR............................................................... 171 Multiplicity of infection................................................... 48 Multivesicular body (MVB)........................................... 252 Muristerone A.................................................................. 15 Muscarinic.................................... 26, 89, 96, 163, 195–197, 200, 208, 238–240, 249, 267, 305, 334, 336 Mutagenesis..........................................40, 85–96, 101, 238 Myc-tag...............................................................26, 56, 256

N N-ethylmaleimide (NEM)..................................... 253, 256 Nitrocellulose............................31, 129, 242, 248, 255, 257, 347–349, 414, 418, 421, 437 Non-linear regression.....................................148, 227–229, 232, 466, 468–470 Non-orthogonal..................................................... 452, 454 Non-parametric....................... 446, 448, 449, 451, 460–462 Non-raft proteins............................................414–415, 418 Nonspecific binding........................... 14, 15, 137, 141–143, 149, 151–153, 155–163, 224, 232, 266, 269, 270, 350, 390, 393, 397, 432 Normal distribution.........................................445, 449–461 Nucleofection.......................... 101, 102, 105, 108, 111, 337 Null hypothesis.......................................449–452, 454, 455, 457–459, 463, 467, 468 Numerical aperture (NA)................................226, 312, 434

O Oil immersion.........................................314, 332, 333, 434 Oligo dT.........................................................177–179, 190 Oligomerization..................................................... 373–386 Oligonucleotide.................................................. 89–92, 190 One-site binding............................................................ 144 One-way analysis of variance.................................. 450–451 Operational model...........................................204, 206, 208 Optical resolution........................................................... 434 Orthogonal............................................................. 451–454 Orthosteric.............................. 195–198, 200, 201, 204, 207 Oscillation.............................................................. 281, 322

P Parametric................444, 446, 448–451, 460–462, 465, 470 Pearson correlation coefficient........................................ 466 Peptide affinity purification.................................... 389–408 Peptide columns..............................................392–394, 396 Pericams................................................................. 278, 288 Permeabilization................................ 57, 289, 430, 433, 435 Pertussis toxin (PTx).............................................. 268, 271 P-factor...........................................................115, 124, 126

Receptor Signal Transduction Protocols 477 Index      Pharmacophore........212, 213, 215–217, 220, 221, 231, 232 Pheromone...................................... 115, 122, 124, 126, 129 Phosphatidylinositol 4,5-bisphosphate (PIP2)........330, 331, 338, 339, 342 Phosphodiesterase.................................................. 298, 304 Phospholipase C (PLC).................................105, 108, 277, 330–334, 336, 340, 342 Phospho-peptide map.................................................... 243 Phosphor-imager.............................................242, 243, 245 Phosphorylation..................... 100, 101, 104, 110, 237–249, 298, 367, 390 Phospho-specific antibodies....................238, 239, 246, 247 Photoactivated localization microscopy (PALM)........... 439 Photobleaching...........57, 293, 294, 305, 309, 313, 339, 341 Phototoxicity.......................................................... 235, 313 Plaque forming units (PFU)................ 47, 48, 50, 89, 91, 93 Plasmid..........................................4–7, 9, 11, 16, 18, 22, 23, 28–30, 33, 35, 40–46, 49, 50, 60–67, 69–72, 75, 76, 79–81, 89–96, 102–103, 108, 110, 111, 119–122, 128, 257, 319, 331, 350, 351, 377, 378, 380, 384, 403, 420 Plasmid linearization.............................................18, 49, 65 Pluronic F–127....................................................... 278, 289 Poly A tail............................................................... 178, 179 Polymerase...............................59, 60, 67, 76, 80, 89, 91, 93, 95, 165–169, 176, 177, 350 Polymerase chain reaction (PCR).....................8, 10, 33, 59, 62, 64–69, 76–83, 88–90, 92–94, 96, 109, 165–192 Polyvinylidene fluoride (PVDF)............. 126, 129, 249, 324, 347–351, 437 Ponasterone A.................................................................. 15 [32P]orthophosphate............................................... 239–242 Post-hoc tests..........................................451, 454–456, 460 Potency............... 86, 204, 214, 221, 232, 342, 456, 458, 459 pREP................................................................119, 123, 128 Preparative gel........................................................ 350, 353 Primer................................33, 64, 67–69, 73, 74, 76, 80–82, 89–95, 166–169, 171, 176–182, 188, 190, 191 Primer-dimers.................................................167, 171, 177 Probability...................................58, 80, 120, 168, 177, 178, 309, 449, 450, 452, 455, 462–464, 468 Probe-dependence.................................................. 202, 208 Promoter..................................4–7, 9, 10, 19, 32–34, 40, 43, 49, 94, 122, 126, 128, 402, 428 Proteasome..................................................................... 252 Protein A affinity chromatography................................. 404 Protein A-sepharose........................ 239, 240, 265, 268, 269 Protein G........................................................................ 274 Protein kinase A (PKA)......................................... 297–303 Protein kinase C (PKC)...................................330, 336, 341 Protein–protein interactions....................... 31, 56, 300, 347, 348, 358, 360 Proteomics...................................................................... 390 Pull-down................................ 318, 319, 323–324, 347–354 Puromycin...............................................................9, 11, 12

Receptor Signal Transduction Protocols 478  Index

  

Q Quantification cycle....................................................... 166 Quantitative PCR...........................................183, 188, 191 Quencher.................................................166, 168, 169, 182

R Rabbit reticulocyte lysate................................................ 354 Radioligand.......................... 10, 14, 16, 135–164, 195–208, 217, 224, 232, 354, 403, 406, 428 Radioligand binding............................. 10, 14, 16, 135–164, 195–208, 217, 221, 232, 265, 403, 428, 466 Random hexamers........................................................... 177, 179 mutagenesis.....................................................86, 88, 89 Range........................... 15, 54, 55, 57, 59, 75, 81–83, 85, 86, 128, 140, 160, 171, 179, 188, 195, 196, 199, 204, 206, 212, 213, 220–222, 224, 225, 228, 230, 231, 279, 281, 288, 302, 303, 313, 330, 333, 335, 338, 339, 354, 368, 374, 375, 389, 426–428, 438, 443, 446–448, 456, 458, 461, 466, 468 Rate constants................................. 136, 137, 153, 154, 363 Reading frame.....................................................29, 79, 126 Real-time imaging...........................................300–305, 331 Receptor number......................................................135, 139, 428 purification....................................................... 245, 406 reserve........................................................106, 204, 232 Receptor-specific antibodies............................239, 406, 426 Recombinase...................................................24, 28–30, 35 Recombination.............................. 22, 42, 44–46, 49, 50, 88 Recycling.........................................................251, 427, 428 Reference gene........................................176, 183–187, 191 Region of interest (ROI)........................ 223, 233, 283, 284, 294, 310, 311, 314, 333–336, 340, 341 Regulator of G protein signaling (RGS).........114, 323, 325 Renaturation................................................................... 348 Renilla luciferase (Rluc).................... 358, 359, 365, 369, 407 Resonance energy transfer.......................212, 299, 367, 374 Resource equation.................................................. 464, 465 REST method................................................................ 187 Restriction digest.......................................................................... 29 endonuclease........................................10, 60, 75–76, 78 enzyme................... 43, 59, 64–65, 68, 72, 78, 90, 94, 96 site..............................................49, 62, 64, 78–80, 92, 94 Retrovirus......................................................................... 40 Reverse transactivator (rtTA)............................................. 9 Reverse transcription.......................................177, 179, 190 Rho......................................................................... 317–327 Rhod–2........................................................................... 293 Rhotekin.......................................... 318, 319, 321, 323–326 Ribonuclease......................................................54, 349, 350 Ribosomal RNA............................................................. 175 RIPA buffer......................................... 27, 31, 239, 240, 248

Rluc2..................................................................358, 359, 368 Rluc8................................. 358, 359, 363, 365, 366, 368, 369 RNAi.......................................................100, 101, 109, 110 RNA integrity number....................................173, 175, 176 RNA interference................................................... 101, 252 RNA isolation................................. 165, 172–174, 176, 177 RNase..................................................59, 60, 70, 78, 82, 103, 110, 173–178, 182, 189–191 Rosenthal plot........................................................ 145, 159 RT-qPCR............................................................... 165–192 Ryan-Einot-Gabriel-Welsch procedure......................... 455

S Saccharomyces cerevisiae...............................................53, 124 Saturation binding........................... 136, 145, 199, 224, 353 Scaffolding proteins........................................................ 298 Scanning spectrometry.................... 360, 362, 364, 368, 369 Scatchard........................................................................ 161 Schild...................................................................... 198, 221 Schizosaccharomyces pombe.........................................114–129 SDM. See Site-directed mutagenesis (SDM) Selection..........................8–9, 11–12, 14, 19, 22, 23, 28–30, 33, 34, 36, 42, 55, 88, 92, 94, 111, 115, 119, 120, 122, 128, 183, 184, 222, 223, 234, 274, 349, 433, 444, 449 Sequencing........... 29, 33, 70, 73, 92, 94, 171, 239, 244, 246 Sheffé test....................................................................... 455 Short-hairpin (sh)RNA.................................................. 110 SH-SY5Y........................................ 107, 331, 332, 334, 336 Shuttle plasmid..................................................... 42–44, 49 Signaling platforms........................................................ 412 Signal peptide...................................... 58, 66, 123, 129, 402 Significance..................... 158, 412, 444, 449–460, 462–464 siRNA..................................................................... 101–110 Site-directed mutagenesis (SDM)...................... 85–96, 101 SNAP.................................................................24, 307, 426 Snap-tag................................................................. 377, 426 Solubilization......................... 254, 257, 258, 265, 267–269, 271–273, 401, 404, 406, 407 Southern blotting....................................................... 23, 50 Span................................................. 147, 152, 329, 353, 467 Spatial resolution............................. 226, 283, 427, 434, 439 Specific activity........................................142, 158, 269, 354 Specific binding................................ 14, 137, 138, 142, 144, 146, 151–153, 159, 160, 197, 200, 202, 203, 224, 225, 232, 318 Sphericity........................................................457, 459, 460 Sphingomyelin............................................................... 411 Spin labelling.................................................................... 87 Spinophilin............................................................. 353, 354 Stable expression........................................................ 16, 21 Stable isotope labelling with amino acids in cell culture (SILAC)....................................... 239 Standard deviation (SD).................................183, 365, 366, 445–447, 464, 465

Receptor Signal Transduction Protocols 479 Index     

Standard error................................................................. 446 Sticky-ends...............................60–62, 65, 66, 69, 78, 79, 82 Stimulated emission decay (STED) microscopy............ 439 Stoke’s shift.................................................................... 220 Streptavidin affinity chromatography............................. 405 Stripping......................................... 415, 421, 430, 431, 440 Student Newman–Keuls test.......................................... 455 Substituted cysteine accessibility method......................... 87 Subtype-selective............................................................ 196 Sucrose............. 118, 119, 125, 376, 379, 413, 416, 417, 429 Suicide enzymes............................................................. 375 Sulfinpyrazone........................................................ 279, 290 Summary statistics.................................................. 443–448 Supercoiled DNA............................................................. 10 SybrGreen.............................................................. 166, 180

T t1/2.......................................................................................138 Tandem affinity purification (TAP)................395, 399–408 Tandem tag..................................................................... 400 TAP. See Tandem affinity purification (TAP) TAP tag........................................... 400, 402, 403, 405, 406 TaqMan®........................................................................ 166 TATA box................................................................. 10, 128 Temperature cycling........................................166, 167, 169 Template........................................64, 66, 69, 76, 79, 80, 86, 89, 91, 92, 94–96, 166–169, 171, 177, 178, 182, 184, 188, 349, 350 Terbium.................................................................. 374, 375 Ternary complex............................................................. 198 Tet-Off......................................................5, 7–9, 11, 15–17 Tet-On......................................................5, 7–9, 11, 15–17 Tetracycline.................. 5, 7, 9, 15, 18, 22–25, 27, 29–31, 35 Tetracycline repressor................................................. 23, 29 Tetracycline-responsive element (TRE)............................. 5 Tetratricopeptide repeat (TPR)...................................... 318 Thin layer chromatography (TLC)................................ 243 Threshold value.............................................................. 166 Time-resolved fluorescent energy transfer (TR-FRET)............................................... 373–386 TIRF. See Total internal reflection microscopy (TIRF) TLC. See Thin layer chromatography (TLC) TnT........................................................................ 348, 350 Tobacco etch virus (TEV) protease.........400–402, 404, 407 Total binding.................................... 87, 142, 143, 149, 151, 153, 155, 156, 158, 223 Total internal reflection microscopy (TIRF)..........212, 278, 283, 287, 439 TPR. See Tetratricopeptide repeat (TPR) Trafficking.................................... 54, 57, 58, 251, 252, 268, 338, 359, 361, 399, 425–440 Transactivator............................................................. 5, 7, 9 Transfection.............................3, 6, 8, 10–12, 14, 16, 18, 19, 22–24, 29, 30, 32–35, 39, 41–47, 49, 50, 90, 95, 101–105, 107–111, 139, 187, 253, 254, 257, 302,

303, 307, 308, 331, 333, 337–339, 359, 361, 367, 377, 401–403, 406, 413, 416, 420 Transfection efficiency............................ 18, 44, 49, 50, 104, 105, 108–110, 337 Transferrin............................... 319, 323, 417–419, 421, 427 Transformation..............................29, 65, 68, 69, 78, 80, 81, 83, 92, 96, 115–116, 119–122, 161, 293, 295, 445, 446, 448, 466 Transgene................................................................... 14, 41 Trans-Golgi network (TGN38)..................................... 427 Transient expression............................................... 334, 377 Translocation................... 106, 303, 330, 334–336, 339–342 TRE. See Tetracycline-responsive element (TRE) T-Rex................................................... 5, 7, 9, 15, 17, 21–36 Triplet state.............................................213, 228, 234, 235 Trypsin digestion.............................................394, 397, 405 Tryptic digest.................................................................. 246 Tryptic peptides.............................................................. 245 t-test...................................................................450, 455, 465 Tubby.............................................................................. 330 Tukey test............................................................... 455, 468 2D electrophoresis...........................................392, 394, 397 Two-site binding.................................................... 147, 470 Two-way analysis of variance...................456, 458, 459, 470 Type I error............................. 449, 450, 454, 455, 459, 463 Type II error............................................449, 454, 463, 464

U Ubiquitin.................................................252–254, 256–258

V Variance.................................. 183, 366, 370, 446, 450–452, 457–459, 464, 468 Venus......................................... 57, 358, 359, 363–366, 369 Viral titre.......................................................................... 47 Virus...................................... 7, 39–41, 46–50, 56, 400, 402

W Water immersion.............................................222, 227, 434 Welch’s test..................................................................... 450 Wilcoxon test................................................................. 461 Wilk–Shapiro test.......................................................... 448

Y Yeast................................. 7, 43, 53, 57–59, 73, 89, 113–130, 256, 257, 349, 390 Yellow fluorescent protein (YFP).......................25, 57, 300, 301, 303, 305, 306, 309–314, 330, 341, 358, 360, 363, 374

Z Zeocin........................................9, 12, 23, 24, 26–30, 33, 34 Z’-factor................................................................. 364–366 Z-factor.......................................................................... 365