Transcription Factors: A Practical Approach, 2nd edition

Transcription Factors

The Practical Approach Series SERIES EDITOR B. D. HAMES Department of Biochemistry and Molecular Biology University of Leeds, Leeds LS2 9JT, UK

See also the Practical Approach web site at http://www.oup.co.uk/PAS * indicates new and forthcoming titles

Affinity Chromatography Affinity Separations Anaerobic Microbiology Animal Cell Culture (2nd edition) Animal Virus Pathogenesis Antibodies I and II Antibody Engineering * Antisense Technology Applied Microbial Physiology Basic Cell Culture Behavioural Neuroscience Bioenergetics Biological Data Analysis Biomechanics - Materials Biomechanics - Structures and Systems Biosensors Carbohydrate Analysis (2nd edition) Cell-Cell Interactions The Cell Cycle Cell Growth and Apoptosis if Cell Separation

Cellular Calcium Cellular Interactions in Development Cellular Neurobiology * Chromatin if Chromosome Structural Analysis Clinical Immunology Complement * Crystallization of Nucleic Acids and Proteins (2nd edition) Cytokines (2nd edition) The Cytoskeleton Diagnostic Molecular Pathology I and II DNA and Protein Sequence Analysis DNA Cloning 1: Core Techniques (2nd edition) DNA Cloning 2: Expression Systems (2nd edition) DNA Cloning 3: Complex Genomes (2nd edition) DNA Cloning 4: Mammalian Systems (2nd edition)

* Drosophila (2nd edition) Electron Microscopy in Biology Electron Microscopy in Molecular Biology Electrophysiology Enzyme Assays Epithelial Cell Culture Essential Developmental Biology Essential Molecular Biology I and I * Eukaryotic DNA Replication Experimental Neuroanatomy Extracellular Matrix Flow Cytometry (2nd edition) Free Radicals Gas Chromatography Gel Electrophoresis of Nucleic Acids (2nd edition) if Gel Electrophoresis of Proteins (3rd edition) Gene Probes 1 and 2 Gene Targeting Gene Transcription if Genome Mapping Glycobiology if Growth Factors and Receptors Haemopoiesis * High Resolution Chromotography Histocompatibility Testing HIV Volumes land 2 * HPLC of Macromolecules (2nd edition) Human Cytogenetics I and II (2nd edition)

Human Genetic Disease Analysis * Immobilized Biomolecules in Analysis Immunochemistry 1 Immunochemistry 2 Immunocytochemistry if In Situ Hybridization (2nd edition) lodinated Density Gradient Media Ion Channels if Light Microscopy (2nd edition) Lipid Modification of Proteins Lipoprotein Analysis Liposomes Mammalian Cell Biotechnology Medical Parasitology Medical Virology MHC Volumes 1 and 2 if Molecular Genetic Analysis of Populations (2nd edition) Molecular Genetics of Yeast Molecular Imaging in Neuroscience Molecular Neurobiology Molecular Plant Pathology I and II Molecular Virology Monitoring Neuronal Activity Mutagenicity Testing * Mutation Detection Neural Cell Culture Neural Transplantation Neurochemistry (2nd edition) Neuronal Cell Lines

NMR of Biological Macromolecules Non-isotopic Methods in Molecular Biology Nucleic Acid Hybridisation Oligonucleotides and Analogues Oligonucleotide Synthesis PCR 1 PCR 2 *PCR3:PCR In Situ Hybridization Peptide Antigens Photosynthesis: Energy Trans duction Plant Cell Biology Plant Cell Culture (2nd edition) Plant Molecular Biology Plasmids (2nd edition) Platelets Postimplantation Mammalian Embryos if Post-Translational Modification Preparative Centrifugation

Protein Blotting if Protein Expression Protein Engineering Protein Function (2nd edition) Protein Phosphorylation Protein Purification Applications Protein Purification Methods Protein Sequencing Protein Structure (2nd edition) Protein Structure Prediction Protein Targeting Proteolytic Enzymes Pulsed Field Gel Electrophoresis RNA Processing I and II if RNA-Protein Interactions Signalling by Inositides Subcellular Fractionation Signal Transduction * Transcription Factors (2nd edition) Tumour Immunobiology

Transcription Factors A Practical Approach Second Edition Edited by

DAVID S. LATCHMAN Professor of Molecular Pathology, and Director, Windeyer Institute of Medical Sciences, University College London Medical School, London, UK

OXPORD UNIVERSITY PRESS

1999

OXFORD UNIVERSITY PRESS

Great Clarendon Street, Oxford OX2 6DP Oxford University Press is a department of the University of Oxford and furthers the University's aim of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Athens Auckland Bangkok Bogota Buenos Aires Calcutta Cape Town Chennai Dar es Salaam Delhi Florence Hong Kong Istanbul Karachi Kuala Lumpur Madrid Melbourne Mexico City Mumbai Nairobi Paris Sao Paulo Singapore Taipei Tokyo Toronto Warsaw and associated companies in Berlin Ibadan Oxford is a registered trade mark of Oxford University Press Published in the United States by Oxford University Press Inc., New York © Oxford University Press 1999 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press. Within the UK, exceptions are allowed in respect of any fair dealing for the purpose of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act, 1988, or in the case of reprographic reproduction in accordance with the terms of licenses issued by the Copyright Licensing Agency. Enquiries concerning reproduction outside those terms and in other countries should be sent to the Rights Department, Oxford University Press, at the address above. This book is sold subject to the condition that it shall not, by way of trade or otherwise, be lent, re-sold, hired out, or otherwise circulated without the publisher's prior consent in any form of binding or cover other than that in which it is published and without a similar condition including this condition being imposed on the subsequent purchaser Users of books in the Practical Approach Series are advised that prudent laboratory safety procedures should be followed at all times. Oxford University Press makes no representation, express or implied, in respect of the accuracy of the material set forth in books in this series and cannot accept any legal responsibility or liability for any errors or omissions that may be made. A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data (Data available) ISBN 0-19-963697-4 (Hbk) 0-19-963696-6 (Pbk) Typeset by Footnote Graphics, Warminster, Wilts Printed in Great Britain by Information Press, Ltd, Eynsham, Oxon.

Preface In the five years since the first edition of this work was published, the number of groups studying transcription factors has dramatically increased. Thus, those studying the regulation of individual genes are increasingly occupied with the study of the transcription factors which regulate such genes, whilst numerous sequences isolated as expressed sequence tags or as part of genome projects have been shown to encode transcription factors. Similarly, a wide range of new methodologies have been developed to facilitate the study of these factors whether isolated as part of studies on the regulation of an individual gene or as part of genome screening procedures. A new edition of this book is therefore necessary to allow it to continue to provide a full description of the methods required to fully characterize the function and activity of an individual transcription factor. The opportunity has been taken of updating the material in the original edition, all the original chapters having either been comprehensively updated by the original authors or completely rewritten by new authors. Most importantly, however, four new chapters have been added, both widening the scope of the original edition and adding specific methods that have become of particular importance since it was first published. As before, the book begins with an initial series of chapters aimed at characterizing the proteins binding to a specific DNA sequence. In particular, the initial chapter on the DNA mobility shift assay (Chapter 1) is now followed by a completely rewritten chapter by new authors dealing with a wide range of methods for characterizing the DNA-protein interaction in more detail (Chapter 2). As before, methods for characterizing the transcription factor protein in more detail are described in the next chapter (Chapter 3). Such a characterizing of the factor paves the way for the isolationg of cDNA clones encoding it, either by the use of oligonucleotides predicted from its protein sequence (Chapter 4) or by screening of a cDNA expression library with its DNA binding site (Chapter 5). However, an increasing number of factors are now identified on the basis of their homologies to other known factors as determined either by experimental screening procedures or by database searches (Chapter 6). In this situation, little or no information will be available about the targets for these transcription factors and for this reason an additional chapter has been added which describes the methods available for identifying the target genes of a previously uncharacterized factor (Chapter 7). Evidently, once a transcription factor has been cloned and its target genes characterized, a range of methods are available to analyse its function further. As in the previous edition, a specific chapter thus deals with many of the methods available for doing this (Chapter 8). This chapter is now supplemented, however, by additional chapters which deal with specific specialized

Preface aspects such as analysing transcription factor function by carrying out in vitro transcription assays using transcriptionally active nuclear extracts derived from rat brain (Chapter 9) and for similarly analysing the effects of these factors on chromatin structure (Chapter 10). In addition, since the activity of many transcription factors in these and other assays is affected by their posttranslational modification, a final chapter describes the methods for analysing the phosphorylation or glycosylation state of transcription factors (Chapter 11). It is hoped that the general updating of existing chapters and the provision of additional chapters will allow this work to build on the success of its predecessor and to continue to provide a comprehensive guide for those who wish to apply the appropriate methods for studying these critical factors. As before, I would like to thank all the contributors for the efforts they have made which have rendered their methods accessible for use by others and to thank the staff of Oxford University Press for deciding to commission this new edition and for their skills in producing it. London April 1998

D.S.L.

VIII

Preface to the First Edition As well as being the essential first step in the conversion of the genetic information in the DNA into protein, the process of transcription is also the major point at which gene expression is regulated. Thus, whilst some cases of post-transcriptional control do exist, in most cases gene regulation is achieved by activating (or repressing) the transcription of particular genes in specific cell types or in response to a specific signal. Once this has occurred, all the other stages of gene expression (RNA processing, translation etc.) follow and the appropriate protein is produced in a cell-type specific or inducible manner. Both the basal process of transcription itself and its regulation are controlled by specific short DNA sequences in the gene promoters or enhancers. These sequences act by binding specific proteins known as transcription factors, which then influence the rate of transcription of the gene. The study of these transcription factors is, therefore, a critical aspect of gene regulation. In general, however, the characterization of these factors involves a distinct set of methods for studying the proteins themselves and their interaction with DNA, which may not be available even in a laboratory skilled in the standard molecular biology techniques for studying DNA and RNA. The aim of this book is to provide such a set of methods which will allow the user who has identified specific regulatory regions in the gene of interest to completely characterize the protein(s) which bind to them. Initially, this study will involve identifying the proteins binding to a specific DNA sequence within the regulatory region by the DNA mobility shift assay (Chapter 1) as well as characterizing the DNA-protein interaction in more detail using DNasel footprinting and methylation interference techniques (Chapter 2). Subsequently, the biochemical characteristics of the protein can be studied, allowing determination of its size and its ability to form complexes as well as to stimulate transcription in vitro (Chapter 3). Although many such studies emphasize the ability of the factor to stimulate transcription, it should not be forgotten that DNA-binding transcription factors are members of a larger class of proteins which have the ability to bind to DNA. Many of the methods developed for these proteins are, therefore, applicable to transcription factors also (Appendix 1). This characterization of the factor paves the way for its eventual purification, which in turn allows the partial protein sequence to be determined, thereby allowing the isolation of cDNA clones for the factor by screening cDNA libraries with appropriate oligonucleotides (Chapter 4). Other methods of isolating cDNA clones for the factor which may be more convenient in some situations also exist, for example direct screening of cDNA expression libraries with the DNA-binding site for the factor or antibodies to it (Chapter 5) or cloning by homology to other known factors (Chapter 6).

Preface to the First Edition Once the cDNA clones have been isolated, all the standard techniques of molcular biology can be applied to studying the gene structure, expression pattern, and DNA sequence of the transcription factor. It will also be important, however, to characterize the regions of the protein which are responsible for its various properties such as DNA binding or transcriptional activation, and this can readily be achieved using the cDNA clones (Chapter 7). Ultimately, therefore, an appropriate combination of the methods described here will allow the experimenter to move forward from the characterization of a DNA sequence involved in the basal or regulated transcription of a specific gene and to obtain a detailed understanding of the protein(s) binding to this sequence and the manner in which it controls transcription. Finally, I would like to thank all the contributors for the efforts they have made to render their methods accessible to others, as well as the staff of Oxford University Press for their continuous assistance. London May 1992

D.S.L.

Contents List of contributors Abbreviations 1. The DNA mobility shift assay

xix xxi 1

C. L. Dent, M. D. Smith, and D. S. Latchman 1. Introduction

1

2. Detection of DNA binding proteins Applications of the DNA mobility shift assay Selection of DNA probe Preparation of labelled oligonucleotide probes for retardation assay Labelling of fragment probes Preparation of protein extracts The binding reaction Preparation of mini-extracts

2 2 6 7 8 9 10 13

3. Other sources of protein for use in the binding assay Expression of DNA binding proteins in bacteria Expression of proteins in mammalian cells Expression by in vitro transcription and translation Expression of cloned transcription factor in baculovirus Purification of transcription factor protein

14 14 15 16 17 17

4. Investigation of DNA binding specificity

17

5. Characterization of DNA binding proteins Addition of antibodies Addition of potential ligands Proteolytic clipping band-shift assay

19 19 21 21

6. Study of protein-protein interactions

23

7. Concluding comments

25

References

25

2. Footprint analysis of DNA-protein complexes in vitro and in vivo

21

Craig Spiro and Cynthia T. McMurray 1. Overview

27

2. In vitro footprinting

29

Contents Analysis at base resolution of binding sites on 32P-end-labelled fragments Analysis of binding on closed, circular plasmid

29 41

3. In vivo footprinting Modification of DNA in vivo Nested gene-specific primers Linker for LMPCR Visualization to nucleotide resolution by LMPCR

47 48 55 55 56

References

61

3. In vitro transcription and characterization of transcription

63

Austin J. Cooney, Sophia Y. Tsai, and Ming-Jer Tsai 1. Introduction

63

2. In vitro transcription assays

63

3. Determination of the molecular weight of native transcription factors Molecular weight determination using gel filtration chromatography Molecular weight determination using glycerol gradient centrifugation Determination of molecular weight of native factors using nondenaturing gradient gel electrophoresis 4. Determination of the molecular weight of denatured transcription factors UV cross-linking of transcription factors to DNA UV cross-linking with a bromodeoxyuridine-substituted DNA UV cross-linking of transcription factors to non-substituted probes Renaturation of transcription factors from SDS-polyacrylamide gels 5. Analysis of the monomer/dimer structure of the DNA binding forms of a transcription factor Analysis of transcription factors by DNA mobility shift assay Analysis of transcription factor subunits by chemical cross-linking Analysis of cooperative binding between dimers bound to adjacent response elements

69 69 73 75 76 77 77 80 82 84 84 85 86

6. Identification and initial characterization of non-DNA binding transcription factors Assay of direct interactions between DNA binding and non-DNA binding transcription factors 'Supershift' gel mobility shift analysis Analysis of the dissociation of transcription factor-DNA complexes in the presence and absence of the non-DNA binding transcription factor Analysis of direct interactions by co-immunoprecipitation References

88 90 90 91 92 95

XII

Contents

4. Purification and cloning of DNA binding transcription factors

97

R. H. Nicolas, G. Hynes, and G. H. Goodwin 1. Introduction

97

2. Buffers and solutions

99

3. Preparation of nuclear extract

99

4. DNA-cellulose chromatography

103

5. DNA affinity chromatography

104

6. Reverse-phase chromatography

111

7. Production and isolation of peptides

114

8. Design of oligonucleotides for cDNA isolation

118

Acknowledgements

121

References

121

5. Cloning transcription factors from a cDNA expression library

123

Ian G. Cowell and Helen C. Hurst 1. Introduction

123

2. Handling bacteriophage expression libraries Library selection Library plating for screening Plaque purification

124 124 125 126

3. Screening methods Introduction Screening with DNA binding-site probes Immunological screening Other approaches to library screening and their relative merits

126 126 127 134 137

4. Proving the identity of the factor

139

References

142

6. Cloning transcription factors by sequence similarity

145

Alan Ashworth 1. Introduction

145

xiii

Contents 2. Methods for cloning related transcription factors Low-stringency hybridization Polymerase chain reaction approaches In silica approaches to identifying novel transcription factors

145 145 148 160

3. Some examples of the cloning of transcription factors by homology

163

Acknowledgements

163

References

163

7. Identification of target genes for a transcription factor by genomic bindingsite cloning

ies

Satoshi Inoue, Shigeru Kondo, and Masami Muramatsu

1. Introduction

165

2. Method for genomic binding-site (GBS) cloning Preparation of a transcription factor protein Source of DNA Selection procedure GBS cloning procedure GBS cloning from CpG islands

165 167 169 170 172 173

3. Molecular cloning of target genes using the DNA fragments obtained by GBS cloning Zoo blotting Northern blot analysis Sequencing and DNA database searches Binding activity of the genomic fragment Enhancer activity of the genomic fragment Molecular cloning of target genes using the DNA fragment obtained by cloning

174 174 175 175 176 176 177

4. Concluding remarks

177

References

178

8. Analysis of cloned factors

181

Roger White and Malcolm Parker 1. Introduction Identification of conserved domains

181 182

2. Mapping and analysis of domains by deletion and point mutagenesis Preparation of deletion mutants using PCR Point mutagenesis

182 182 185

xiv

Contents 3. Expression systems The use of in vitro translation systems and overexpression systems to analyse transcription factor function Expression in bacteria Expression in mammalian cells Overexpression in mammalian cells Overexpression in yeast and insect cells Comparison of expression systems

187

4. Analysis of the properties of cloned factors Transient transfection in mammalian cells Reporter genes and control plasmids Methods of transfection Assays for reporter genes in transfected cells Identification of transactivation domains using chimeric proteins

196 196 196 198 200 202

5. Analysis of protein-DNA interactions Analysis of DNA binding activity using a gel retardation assay

202 202

6. Analysis of protein-protein interactions in vitro

206

Analysis of interactions between proteins using a GST pull-down assay Analysis of protein-protein interactions using immunoprecipitation Analysis of protein-protein interactions using the gel retardation assay Analysis of protein-protein interactions on DNA using an ABCD assay Analysis of protein-protein interactions in cells using two-hybrid analysis

187 189 192 193 194 195

206 208 209 209 211

4. Summary

212

References

213

9. Neuronal promoter analysis by in vitro transcription using nuclear extracts from rat brain

215

M. L. Schwartz and W. W. Schlaepfer 1. Introduction

215

2. Background

216

3. Applications Comparison to transfection and transgenic mouse studies Studies of the effect of chromatin structure on transcription

216 216 217

4. Preparation of nuclear extracts Overview

217 217

xv

Contents Purification of nuclei from rat brain Extraction of nuclear proteins 5. In vitro transcription Overview of the reaction Typical results of in vitro transcription Notes on in vitro transcription Promoter templates 6. Concluding remarks References

10. Preparation of chromatin templates for transcription studies Joan Boyes 1. Introduction 2. Experimental approaches using chromatin templates Introduction to chromatin structure Initiation and elongation of transcription on chromatin templates Mononucleosome templates versus nucleosome arrays 3. Preparation of chromatin templates Technical considerations Xenopus nuclear extracts Drosphila embryo nuclear extracts Minichromosome preparation Mononucleosome preparation 4. Preparation of histones Principle of hydroxyapatite purification 5. Reconstitution of mononucleosomes by salt-urea dialysis Considerations for the DNA fragment for reconstitution Amount of DNA needed in the reconstitution reaction Calculation of the amount of histones in the reconstitution reaction Purification of mononucleosome reconstitutes Characterization of the reconstitutes 6. Transcription factor binding to nucleosome templates Acknowledgements References

11. Analysis of transcription factor modifications

217 220 222 222 224 224 226 227 227

229 229 229 229 230 230 232 232 233 233 234 234 236 236 243 243 246 248 249 252 255 257 258

261

N. Shaun B. Thomas 1. Introduction

261

xvi

Contents 2. Phosphorylation Gel electrophoresis Dephosphorylation in vitro Radioactive labelling Mapping phosphorylation sites Phosphorylation in vitro Generation and use of phosphorylation-site-specific antibodies Functional analysis Other methods for analysing phosphorylated proteins

261 263 266 270 274 281 283 283 284

3. Other post-translational modifications O-Linked glycosylation

285 285

Other modifications

290

4. Conclusions

290

Acknowledgements

291

References

291

Appendix 1 List of suppliers

295

Index

297

XVll

This page intentionally left blank

Contributors ALAN ASHWORTH

Chester Beatty Laboratories, The Institute of Cancer Research, Fulham Road, London SW3 6JB UK. JOAN BOYES

Chester Beatty Laboratories, The Institute of Cancer Research, Fulham Road, London SW3 6JB UK. AUSTIN J. COONEY

Department of Cell Biology, Baylor College of Medicine, 1 Baylor Plaza, Houston, TX 77030 USA. IAN G. COWELL

Department of Biochemistry and Genetics, The Medical School, Newcastle University, Newcastle Upon Tyne, NE2 4HH UK. C. L. DENT Glaxo Wellcome Research and Development, Medicines Research Centre, Gunnels Wood Road, Stevenage, Hertfordshire, SGI 2NY UK. G. H. GOODWIN

Haddow Laboratories, Institute of Cancer Research, 15 Cotswold Road, Sutton, Surrey, SM2 5NJ UK. HELEN C. HURST

ICRF Molecular Oncology Unit, Hammersmith Hospital, Du Cane Road, London W12 ONN UK. G. HYNES

Chester Beatty Laboratories, Institute of Cancer Research, Fulham Road, London, SW3 6JB UK. SATOSHIINOUE

Department of Biochemistry, Saitama Medical School, 38 Morohongo, Moroyama Iruma-Gun, Saitama 350-04, JAPAN. SHIGERU KONDO

Department of Biochemistry, Saitama Medical School, 38 Morohongo, Moroyama Iruma-Gun, Saitama 350-04, JAPAN. D. S. LATCHMAN

The Windeyer Institute for Medical Sciences, University College London Medical School, The Windeyer Building, 46 Cleveland Street, London W1P 6DB, UK.

Contributors CYNTHIA T. McMURRAY

Department of Pharmacology, Mayo Foundation, 200 1st Street, SW Rochester, MN 55905 USA. MASAMIMURAMATSU

Department of Biochemistry, Saitama Medical School, 38 Morohongo, Moroyama Iruma-Gun, Saitama 350-04, JAPAN. R. H. NICOLAS

Imperial Cancer Research Fund, PO Box 123, Lincoln's Inn Fields, London WC2A3PXUK. MALCOLM G. PARKER

Molecular Endocrinology Laboratory, Imperial Cancer Research Fund, PO Box 123, Lincoln's Inn Fields, London WC2A 3PX UK. W. W. SCHLAEPFER

Division of Neuropathology, 435 Johnson Pavillion, University of Pennsylvania Medical School, Philadelphia, PA 19104-6079 USA. M. L. SCHWARTZ

Division of Neuropathology, 435 Johnson Pavillion, University of Pennsylvania Medical School, Philadelphia, PA 19104-6079 USA. M. D. SMITH

The Windeyer Institute for Medical Sciences, University College London Medical School, The Windeyer Building, 46 Cleveland Street, London W1P 6DB, UK. CRAIG SPIRO

Department of Pharmacology, Mayo Foundation, 200 1st Street, SW Rochester, MN 55905 USA. N. SHAUN B. THOMAS

Department of Haematology, University College London Medical School, 98 Chenies Mews, London WC1E 6HX UK. MING-JER TSAI

Department of Cell Biology, Baylor College of Medicine, 1 Baylor Plaza, Houston, TX 77030 USA. SOPHIA Y. TSAI

Department of Cell Biology, Baylor College of Medicine, 1 Baylor Plaza, Houston, TX 77030 USA. ROGER WHITE

Molecular Endocrinology Laboratory, Imperial Cancer Research Fund, PO Box 123, Lincoln's Inn Fields, London WC2A 3PX UK xx

Abbreviations A x a b s o r b a n c e a t wavelength x BLAST basic local alignment search tool BPV bovine papilloma virus BSA bovine serum albumin bZIP basic zipper CAT chloramphenicol acetyl transferase cdk cyclin-dependent kinase CIP calf intestinal phosphatase CNBr cyanogen bromide CoA coenzyme A COX7RP cytochrome oxydase subunit VH-related protein CRE cAMP-responsive enhancer DATP diallyl tartardiamide DCC dextran-coated charcoal dd dideoxy dATP 2'deoxyadenosine 5'triphosphate dCTP 2'deoxycytidine 5'triphosphate dNTP deoxynucleoside triphosphate dGTP 2'deoxyguanosine 5'triphosphate dTTP 2'deoxythymidine 5'triphosphate DBS diethylstilboestrol DMEM Dulbecco's modified Eagle's medium DMP dimethylpimelidate DMS dimethyl sulfate DNAse deoxyribonuclease DSP dithiobis (syccinimyl propionate) DTT dithiothreitol EBAG9 oestrogen receptor-binding fragment-associated gene 9 EDTA ethylenediamine tetraacetic acid EFP oestrogen-responsive finger protein EGTA ethylene glycol-bis (p-aminoethyl ether) N,N,N',N' tetraacetic acid EMSA electromobility shift assay ENU ethylnitrosourea ER oestrogen receptor ER-DBD DNA binding domain of oestrogen receptor ERE oestrogen-responsive element ESMS electrospray mass spectrometry FACT facilitates chromatin transcription GBS cloning genomic binding-site cloning

Abbreviations GST HMG HSV IPTG LB broth LDAO MALD-MS MBD NGF NP40 NURF OD PAGE PBS PCR PCV p.f.u. PMSF PR pRb PRE PVDF RSC SDS SRF SSC SV40 TCA TE TEMED TF TFA tk Tris vitERE

glutathione-5-transferase high mobility group herpes simplex virus isopropyl 3-D-thiogalactopyranoside Luria-Bertani broth lauryl dimethylamine oxide matrix-assisted laser desorption mass spectrometry methyl-CpG binding domain nerve growth factor Nonidet P-40 (octylphenoxypolyethoxyethanol) nwcleosome remodelling/actor optical density polyacrylamide gel electrophoresis phosphate-buffered saline polymerase chain reaction packed cell volumes plaque forming units phenyl methyl sulfonyl fluoride progesterone receptor retinoblastoma protein progesterone-response element polyvinylidene fluoride remodel the structure of chromatin sodium dodecyle sulfate serum response factor saline-sodium citrate Simian virus 40 trichloroacetic acid Tris-EDTA (buffer) N,N,N' ,N' ,-tetramethylethylenediamine transcription factor trifluoroacetic acid thymidine kinase Tris (hydroxymethyl)-aminomethane vitellogenin ERE

xxii

1 The DNA mobility shift assay C. L. DENT, M. D. SMITH, and D. S. LATCHMAN

1. Introduction A knowledge of the DNA binding protein content of a cell or tissue type can be an important aid towards a more thorough understanding of the functional role of that tissue. Important processes, such as the development of an adult organism from the single-celled zygote and the maintenance of the developed structure and biochemical characteristics of tissues, are increasingly being demonstrated to be dependent on DNA binding transcription factors. The spectrum of transcription factors present determines what genes may be transcribed in a cell type, including genes that encode further transcription factors. The expression of a single gene may be determined by a number of different transcription factors, and the ratio of their concentrations may be vital, particularly if competing positive and negative factors are involved. Factors other than those that bind the DNA may be involved, such as protein cofactors, metal ions, and ligand molecules, and a further important factor is the chromatin structure of the DNA. The gene must be available for binding of factors and subsequent transcription. The DNA mobility shift assay attempts to determine the potential for a gene to be transcribed in a particular cell type by providing an assay for the presence of DNA binding proteins capable of binding to a promoter, providing a very effective point of first entry into a more detailed understanding. This chapter will detail the techniques and considerations involved in analysing DNA-protein interactions using the DNA mobility shift assay, and the example of the herpes simplex virus 1 (HSV-1) immediate-early gene promoter TAATGARAT sequence that has been studied in our laboratory (1). Protein extracts from cell types that are permissive and non-permissive to lytic infection by HSV-1 were assayed for proteins binding to the TAATGARAT sequence (the TAATGARAT sequence is a promoter element that has been demonstrated to be important in immediate-early gene transcription via the binding of a complex of a cellular transcription factor protein, Oct-1, and a viral component, Vmw65—see refs 2 and 3). Both permissive and non-permissive cell types contain this cellular factor, but non-permissive

C. L. Dent et al. cell types were found also to contain a second factor believed to be a neuronal form of the B-cell-specific protein Oct-2, which is closely related, in DNA binding specificity and structurally, to Oct-1. However, Oct-2 cannot interact with the viral component Vmw65 (4) and cannot act as a transcriptional activator from the viral TAATGARAT sequence. It is, however, able to bind TAATGARAT, thus preventing the access of Oct-l/Vmw65 complexes, and subsequent transcription of the immediate-early genes of HSV-1. Figure 2 shows a DNA mobility shift assay demonstrating the presence of DNA binding proteins in permissive (BHK) and non-permissive (neuronally derived ND7) cells showing the presence of large amounts of the second protein (lower band) in the non-permissive cell line. Further experiments have supported the fact that this protein is a neuronal form of Oct-2 and does indeed have a repressive capacity (5), thus demonstrating how a single band from a mobility shift may eventually explain an important cellular phenomenon.

2. Detection of DNA binding proteins 2.1 Applications of the DNA mobility shift assay Sequence-specific DNA binding proteins are involved in the regulation of transcription and DNA replication. The DNA mobility shift assay provides a powerful tool for the detection of factors binding to specific sequences. The method relies on the ability of a protein to bind to a radiolabelled DNA fragment (probe) in vitro, followed by electrophoretic separation of DNAprotein complexes from the unbound DNA on non-denaturing polyacrylamide gels (6,7). One or more proteins binding to the DNA fragment may be identified. In general, the larger the DNA-protein complex that is formed, the greater the extent of retardation of mobility within the gel. The principle of the DNA mobility shift assay is illustrated in Figure 1. Figure 2 illustrates a typical example of such an assay. The unbound probe containing the octamer motif is represented by the very heavy unresolved band at the bottom of the photograph in all four tracks. Two of the tracks show the complexes formed by two DNA binding proteins of different mobilities binding to the same DNA sequence in extracts prepared from different cell types. The second track shows only one slowly migrating DNA-protein complex, whereas the fourth track shows two. The protein present in both tracks is Oct-1 and has a molecular weight of 100 kDa. The protein present only in the fourth track is Oct-2 and has previously been demonstrated to be smaller than Oct-1 (molecular weight 60 kDa). The complex formed by the smaller protein migrates further within the gel, as would be expected. An accurate determination of the molecular weight of a DNA binding protein is not possible on this type of gel. The protein must be somehow purified (for example by cutting out from a gel after UV cross-linking to the labelled DNA probe) and then 2

1: The DNA mobility shift assay

Figure 1. DNA mobility shift assay- Binding of a cellular protein (B) to the radioactively labelled DNA causes it to move more slowly upon gel electrophoresis and hence results in the appearance of a retarded band upon autoradiography to detect the radioactive label.

run on a denaturing sodium dodecyl sulfate (SDS)-polyacrylamide gel (see Chapter 3). The DNA sequence specificity of the protein within the complex can be tested by competing for binding with non-radiolabelled DNA fragments. If a molar excess of a DNA fragment capable of binding the same protein is introduced into the binding reaction, much of the protein will bind to the unlabelled DNA, leaving less protein available for binding to the probe. This will lead to a reduction in, or elimination of, the band corresponding to the complex formed by that protein (Figure 3). In Figure 2, the first track represents 3

C. L. Dent et al.

Figure 2. DNA mobility shift assay using nuclear extracts prepared from BHK cells (a) or ND cells (b) and a labelled octamer oligonucleotide. The tracks show the results in the absence (-) or presence (i ) of a 100-fold excess of unlabelled octamer oligonueleotide competitor. For further details see ref, 1.

the same binding reaction as the second track, but in the presence of a 100foJd excess of unlabelled oligonucleotide. The band formed by the interaction between the probe and Oct-1 is competed away. Similarly, the third track represents the same reaction as the fourth track in the presence of excess unlabelled oligonucleotide. The band formed by the interaction between the probe and Oct-1 is competed away. Both specific and non-specific DNA sequences should be used because the failure of a non-specific sequence to compete provides proof that the complex is indeed DNA-scquence specific. Identification of proteins within the complex may be made by including antibodies against known proteins in the binding reaction. These antibodies may bind to the complex, causing further electrophoretic retardation and 'supershifting' of the complex, or completely inhibit complex formation by binding to a vital site within the binding protein. These types of experiment will be discussed fully later in this chapter (Section 5.1). Applications of the DNA mobility shift assay include the identification of both known and novel factors binding to a candidate DNA fragment, usually the DNA sequences 5' to a transcription unit. It may also be used to identify fluctuations in the levels of known transcription factors in response to stimuli. 4

1: The DNA mobility shift assay

Figure 3. Use of unlabelled competitor DMAs in the DNA mobility shift assay. If the unlabelled competitor is capable of binding the same protein as the labelled probe, it will do so (B) and the retarded band will not be observed.

e.g. growth factors. In our laboratory, we have used the assay both for the study of known factors and identification of new factors. In one series of experiments (8), we have taken a sequence, identified within the human papillomavirus 16 enhancer as being related to the octamer motif, and used this 5

C. L. Dent et al.

Figure 4, DNA mobility shift assay using an overlapping octamer/TAATGARAT oligonucleotide (ATGCTAATGAGAT) from the HSV-1 IE1 gene promoter and extracts from 3T3 cells (track 1J, BHK-21 cells (track 2), Jurkat T cells (track 3), S115 mammary epithelial cells (track 4), 310 primary cervical cells with no evidence of papillomavirus infection (track 5), 310 A cells (310 cells transformed with HPV-16 DNA, track 6), and SiHa cervical carcinoma cells (track 7). The arrow indicates the cervical-specific band. For further details see ref. 8.

sequence to identify a protein, contained in cervical cell extracts, that binds to this sequence. Figure 4 shows a band-shift assay carried out using the papillomavirus octamer sequence, showing binding to Oct-1 in all seven cell lines studied, and to a second protein with a higher mobility only in those extracts made from cervical cell lines. It is important to realize, however, that the DNA mobility shift assay does have limitations. In particular, this assay does not reveal which nucleotides within the sequence are recognized by, and interact with, the protein. As with many techniques, it is at its most valuable when used in conjunction with the other binding and functional assays discussed in the accompanying chapters of this book.

2,2 Selection of DNA probe A restriction fragment or synthetic oligonucleotide probe may be used, but the size of the fragment is normally kept below about 250 base pairs (bp) to 6

1: The DNA mobility shift assay enable clear distinction of the probe from any complexes. The type and size of probe used depends on the nature of the investigation. If a previously identified factor is to be studied, then an oligonucleotide probe should be used. An oligonucleotide probe simplifies the interpretation of results as a site is isolated from other unidentified possible sites present within adjacent regulatory sequences. However, interactions of a protein on an isolated site may not mirror the situation in vivo as competition and cooperation between proteins binding to adjacent sites may be observed. This is exemplified in the U2 small nuclear RNA enhancer where Oct-1 and Sp-1 bind cooperatively to adjacent binding sites (9). In this case, oligonucleotides covering both sites were constructed, and their binding of the two proteins compared with oligonucleotides in which one or the other site was rendered non-functional by mutation. Protocols 1-3 below give methods for labelling both oligonucleotide and fragment probes.

2.3 Preparation of labelled oligonucleotide probes for retardation assay Synthetic binding sites are made as two complementary single-stranded oligonucleotides that are subsequently annealed to generate a double-stranded oligonucleotide. Annealing is achieved by mixing equimolar amounts of the two oligonucleotides, heating to 80°C for 5 min, and allowing the oligonucleotides to cool slowly down to room temperature. If so desired, the oligonucleotides may be designed to possess the overhanging ends of a restriction enzyme site when annealed. This permits them to be cloned into promoter constructs, facilitating the assay of their activity in cells. The most common method for the labelling of oligonucleotides is to add a 32 P-labelled phosphate to the 5' end using T4 DNA kinase (see Protocol 1): the enzyme catalysing the transfer of the gamma-phosphate of ATP to the 5' hydroxyl group of the terminal nucleotide. Fragment probes may also be labelled using this method but the terminal phosphate must be removed from the DNA by phosphatase treatment before the addition of labelled phosphate (Protocol 2). It is usually simpler to label fragment probes by filling in the recessed ends created by restriction enzymes. The methods for labelling restriction enzyme fragments are described in Section 2.4. All the methods for labelling DNA are described by Sambrook et al (10). Protocol 1. End labelling DNA with T4 kinase Equipment and reagents •

32

STE: 10 mM Tris-HCI pH 8.0, 100 mM NaCI, 1 mM EDTA Sephadex G-25 column Water bath at 37°C

P-labelled ATP, at 110 TBq/mmol (370 MBq/ml . 50 mM Tris-HCI pH 7.6, 10 mM MgCI2, 5 mM dithiothreitol (DTT), 0.1 mM EDTA . T4 DNA kinase

7

C. L. Dent et al. Protocol 1.

Continued

Method 1. Mix 2 pmol of annealed oligonucleotide with 20 uCi (0.74 MBq) [32P]ATP in the presence of 50 mM Tris-HCI pH 7.6, 10 mM MgCI2, 5 mM DTT, 0.1 mM EDTA, and 0.5 ul (5 units) of T4 DNA kinase. 2. Incubate at 37°C for 30 min. 3. Add 200 ul of STE to the reaction and separate the labelled oligonucleotide from unincorporated label by centrifugation at 2000 x g for 2 min at room temperature through a 1 ml Sephadex G-25 column. The probe passes through in approximately 200 ul, of which 1 ul (10fmol of DNA) is used per binding reaction.

Protocol 2.

Dephosphorylation of DNA

Equipment and reagents • Calf intestinal phosphatase (CIP) . 10 x CIP buffer: 0.5 M Tris-HCI pH 9.0, 10 mM MgCI2,1 mM ZnCI2, 10 mM spermidine . STE: 100 mM Tris-HCI pH 8.0, 1 M NaCI, 10 mM EDTA

PhenolArichloromethane (phenol:trichloromethane:/so-pentanol, 25:24:1) 10 mM Tris-HCI pH 8.0 Water bath at 37°C

Method 1. Dissolve DNA in a minimum volume of 10 mM Tris-HCI pH 8.0. 2. Add 5 ul 10 x CIP buffer and 0.01 units of CIP per picomole of DNA ends, and make up to 50 ul with distilled water. 3. Incubate at 37°C for 30 min. 4. Add 50 ul H2O,10 ul STE, and 5 ul 10% SDS. 5. Heat to 68°C for 15 min. 6. Extract the DNA with an equal volume of phenol/trichloromethane. 7. Take the upper aqueous layer and repeat step 6. 8. Transfer the upper aqueous layer to a fresh tube and extract with an equal volume of trichloromethane. 9. Take the aqueous layer and precipitate the DNA with 2 volumes of ethanol overnight at -20°C.

2.4 Labelling of fragment probes There are three classes of fragment probes, depending on the nature of DNA ends generated by the enzymes used for their isolation. The enzymology of labelling these fragments varies, depending on the type of ends generated. (a) Fragments with 5' overhanging ends, generated by enzymes such as BamHI and EcoRI. These fragments are the most straightforward to 8

1: The DNA mobility shift assay label using the Klenow fragment of DNA polymerase I; therefore, usually enzymes that generate these ends are chosen preferentially when deciding upon a strategy for the isolation of DNA fragments for labelling. (b) Fragments with 3' overhanging ends, generated by enzymes such as KpnI and SstI. These are labelled using T4 DNA polymerase which possesses a 3'-5' polymerase activity as well as a potent 5'-3' exonuclease activity (which must be kept inactive to prevent unwanted degradation of the DNA fragment). (c) Blunt-ended fragments, generated by enzymes such as Smal and Rsal. These are also labelled by T4 DNA polymerase after limited digestion by the exonuclease activity of the enzyme. Protocol 3. Filling in 5' overhangs using the Klenowfragment of E. coli DMA polymerase I Equipment and reagents 32

P-labelled dCTP (20 vC\, 0.74 MBq); 1 mM unlabelled dATP, dTTP, and dGTP 50 mM Tris-HCI pH 7.5, 10 mM MgSO4, 0.1 mM DTT

Klenow enzyme STE (Protocol 1) Sephadex G-50 column Water bath at 37°C

Method 1. Mix DNA, [32P]dCTP, unlabelled dATP, dTTP, and dGTP, 50 mM Tris-HCI pH 7.5, 10 mM MgS04, 0.1 mM DTT, and 1 ul (1 unit) Klenow enzyme. 2. Incubate at 37°C for 30 min. 3. Make volume up to 200 ul with STE and centrifuge through a Sephadex G-50 column (as described in Protocol 1).

Protocol 4. Filling in 3' overhangs using T4 DNA polymerase Equipment and reagents 32

P-labelled dCTP (20 uCi, 0.74 MBq); 1 mM unlabelled dATP, TTP, dGTP Acetate buffers: Tris acetate 33 mM pH 7.9, 66 mM K acetate, 10 mM Mg acetate Bovine serum albumin (BSA, 0.1 mg/ml)

STE (Protocol 1) T4 DNA polymerase Sephadex G-50 column Water bath at 37 °C

Method 1. Mix DNA, [32P]dCTP, unlabelled dATP, TTP, dGTP, acetate buffers, 0.5 mM DTT, 0.1 mg/ml BSA, and 1 ul (2.5 units) T4 DNA polymerase. 2. Incubate at 37°C for 30 min. 3. Make volume up to 200 ul with STE and centrifuge through a 1 ml Sephadex G-50 column (as described in Protocol 1).

9

C. L. Dent et al. Blunt-ended fragments are labelled by a modification of Protocol 4. The DNA is incubated with T4 DNA polymerase in the absence of dNTPs for about 1 min, allowing the exonuclease activity of the polymerase enzyme to cut back. If a single dNTP is included during this incubation, the exonuclease will only cut back until it reaches that base in the fragment, thus controlling the digestion. The label and remaining dNTPs are then added, and the digested DNA is resynthesized, incorporating the labelled dNTP.

2.5 Preparation of protein extracts Protein extracts may be prepared from whole cells or isolated nuclei. There are advantages to using both types of extract, or even a combination, and comparing results obtained with both types. The preparation of nuclear extracts results in the isolation of only those binding factors with access to the DNA. Factors isolated from the nucleus will thus have the potential to bind to sites on the chromosomal DNA. However, it must be considered that much of the chromatin is masked by histones and other DNA binding proteins, thus rendering it inaccessible to transcription factors, hence a binding site in vitro may not necessarily represent a binding site in vivo. The preparation of extracts from whole cells enables the entire DNA binding protein content of the cell to be examined. Some proteins may be present in the cytoplasm rather than the nucleus, and can be identified by the comparison of the binding profiles of nuclear and cytoplasmic extracts. Proteins present in the cytoplasm cannot be involved in transcriptional regulation at that time, due to their lacking access to the chromatin, but are likely to be on 'standby', ready for a quick transport into the nucleus following a certain stimulus. This transport of transcription factors from the cytoplasm to the nucleus may be studied by comparing whole-cell and nuclear extracts before and after treatment of the cells. Whole-cell extracts are easier to prepare, requiring fewer steps in their preparation, which makes their use favourable when a tissue sample is limiting because fewer manipulations present fewer stages at which protein will be lost or damaged. It is also necessary to prepare whole-cell extracts when a sample has been frozen, as freezing damages the nuclear membrane, preventing the preparation of intact nuclei, and thus the preparation of nuclear extracts. The method for the preparation of nuclear extracts is given in Protocol 5, and is a modification of that described by Dignam et al. (11). If fewer than 5 X 107 cells are to be used, the method in Protocol 5 may be scaled down accordingly. For very small numbers of cells, the mini-preparation method given in Protocol 8 should be used. Whole-cell extracts are made by a modification of the method (12) given in Protocol 5. Harvest the cells and wash with PBS. Resuspend in 1 ml of buffer C and homogenize with 20 strokes of a tight-fitting homogenizer. Then add 10

1: The DNA mobility shift assay NaCl to a final concentration of 300 mM and continue from step 9 of the nuclear extract preparation method.

Tissue samples can be more difficult to homogenize than cultured cells. Small pieces of soft tissue may be treated as cultured cells, and we have successfully used this method for the isolation of binding proteins from rat dorsal root ganglion and brain tissue. Larger pieces of tougher tissue or frozen samples may have to be treated more drastically to release the protein. It should prove sufficient for most tissues to use a tissue macerator instead of a 11

C. L. Dent et al. Dounce homogenizer, but particularly tough tissues or frozen samples may have to be frozen in liquid nitrogen and ground to a fine powder before homogenization. Samples that have at any stage been frozen may only be used for the preparation of whole-cell extracts. Fresh tissue samples should be dealt with as quickly as possible to prevent degradation by proteases.

2.6 The binding reaction Conditions for the binding of protein to DNA vary between research groups. We use a 45 min incubation on ice in order to maintain an extent of uniformity between different studies. The binding conditions routinely used in our laboratory are given in Protocol 6.

Protocols. Conditions for binding Equipment and reagents Binding solution: 20 mM Hepes pH 7.9, 1 mM MgCI2, 4% ficoll, 0.5 mM DTT Poly dldC (Pharmacia)

0.25 X TBE (1 X TBE = 100 mM Tris-HCI, 100 (um boric acid, 2 mM EDTA) Gel drying apparatus

Method 1. Make a 20 ul binding reaction by mixing binding solution with KCI to a final salt concentration of 50 mM,a 2 ug poly dldC, and 1 ul (10 fmol) 32 P-labelled probe and protein extract. 2. Incubate on ice for 40 min. 3. Run samplesb on a 4% polyacrylamide gel (0.25 x TBE). The gel should be pre-run at 150 V for about 2 h before electrophoresis (current will drop from 20-30 mA to approximately 10 mA during this time). Then run samples for 2.5 h (or until the bromophenol blue marker dye has run about two-thirds of the way down the gel). 4. Dry gel on to filter paper (1 h, 80°C, with vacuum) and autoradiograph overnight. 'Remember that any extract added is in 300 mM NaCI. 'Do not add any loading dye to samples; the ficoll in the binding buffer provides the density required for loading. Bromophenol blue in glycerol may be added to a spare track as marker. Poly dldC is added to the binding reaction as a non-specific competitor for the binding of any general DNA binding proteins, leaving the probe free to bind sequence-specific binding proteins. The amount of extract required will vary depending on the protein concentration of the extract and the abundance and affinity of the factor to be studied. It is advisable to determine the amount giving the best results experimentally. The concentration of protein in extracts should be determined by the method of Bradford (ref. 13, Protocol 7), allowing the comparison of equal amounts of protein from different extracts.

12

1: The DNA mobility shift assay Protocol 7. Bradford assay for protein concentration Equipment and reagents Dye reagent: 100 mg Coomassie brilliant blue G, 30 mg SDS, 50 mg 95% (v/v ethanol, 100 ml 85% (v/v) phosphoric acid dilute the mixture to a final volume of 1 litre, using distilled water

PBS 0.25 M Tris-HCI pH7.5 Liquid nitrogen and 37°C water bath

Method 1. Following transfection, wash the cells with PBS, harvest them, transfer to a 1.5 ml microcentrifuge tube and centrifuge at 1000 x g for 5 min at 4°C. 2. Add 100 ul of 0.25 M Tris-HCI to the cell pellet. 3. Disrupt the cells by freezing and thawing. To freeze-thaw, immerse the tubes in liquid nitrogen for 2 min and then transfer them to a 37°C water bath. Repeat the cycle three times. 4. Pellet the cell debris by centrifugation at 3000 x grfor 3 min at 4°C and save the supernatant to test for protein. Samples may be saved at this point by storage at -20°C. 5. Add 1 ml of dye reagent to 10 uJ of each sample. 6. Measure the absorbance of the sample at 595 nm (A595) after 15 min. 7. If the absolute concentration of protein is needed, construct a standard curve using BSA as standard (draw A5S5 vs. [BSA] mg/ml).

2.7 Preparation of mini-extracts Several methods have been published for the preparation of mini-extracts, when cells or tissue are limiting. Mini-extracts are particularly quick and easy to prepare, and all steps can usually be carried out in a microcentrifuge tube without the need for steps such as homogenization, eliminating the loss of sample that may occur when it is continually being transferred between tubes. The method given in Protocol 8 is for the preparation of mini-whole-cell extracts (14). Protocol 8. The preparation of mini-whole-cell extracts Equipment and reagents Microcentrifuge and microcentrifuge tubes Freezing bath (dry ice/ethanol) and 37°C water bath

PBS

Extraction buffer: 20 mM Hepes pH 7.8, 450 mM NaCI, 0.4 mM EDTA, 0.5 mM DTT, 25% glycerol, 0.5 mM PMSF

Method 1. Harvest the cells into 1 ml PBS in a microcentrifuge tube, centrifuge for 1 min at 1000 x grforS min at 4°Cto pellet. 13

C. L. Dent et al. Protocol 8. Continued 2. Wash with 1 ml PBS, centrifuge for 1 min at 1000 x g for 5 min at 4°C, remove all traces of PBS. 3. Resuspend cell pellet in 100 ul (or less) extraction buffer. 4. Freeze (dry ice/ethanol bath) and thaw (37°C water bath) three times. 5. Spin at 10000 x g for 10 min in a microcentrifuge at 4°C. 6. Use 10 (ul of supernatant for each binding reaction.

3. Other sources of protein for use in the binding assay Purified transcription factor expressed from the cloned gene may be used as a substrate for a DNA binding assay. The expression of a cloned transcription factor can allow the study of the binding of that factor in the absence of other proteins with which it may interact in vivo and, when expressed in bacterial systems, in the absence of some post-translational modifications. Expression can serve to confirm the identity of a clone by comparing the band formed by the expressed cloned protein with that observed in cellular extracts. It must be remembered, however, that a cloned protein may not necessarily bind in the same manner as the endogenous protein does in a cellular extract, or even bind at all. Several factors may contribute to differences in binding. Other proteins may be required for, or contribute towards, the formation of a complex, resulting in a different binding pattern, or inability to bind. Proteins synthesized in bacterial systems will not have undergone certain posttranslational modification that occur in a mammalian cell; the most common of these being phosphorylation (see, for example, ref. 15). One of the greatest advantages of using expressed cloned transcription factors for DNA binding and expression studies is that it is possible to manipulate specific regions of the transcription factor protein by mutating the cloned gene. Truncations, deletions, and point mutations of the protein can be made. Binding assays and expression assays can then be used to determine which mutations are defective in DNA binding or transcriptional activation (see Chapter 8). There are several methods of expressing cloned proteins, these include: (a) expression in bacteria (b) expression in mammalian cells (c) in vitro synthesis in reticulocyte lysate from an in vitro translated mRNA (d) baculovirus expression systems

3.1 Expression of DNA binding proteins in bacteria Expression in bacterial systems is one of the less commonly used systems for the synthesis of transcription factor protein. All the other systems described use eukaryotic cells or cell extracts for the synthesis of protein, and it makes good sense to synthesize a eukaryotic protein in a eukaryotic system. 14

1: The DMA mobility shift assay This system has been used in a number of cases. There are several factors that must be considered when using bacterial expression vectors for the expression of transcription factors. First, as with any expression vector, the coding sequence must be inserted into the vector in the correct orientation and reading frame, otherwise the RNA made will not be able to be translated into protein. It is also important to use a cDNA clone rather than one which is interrupted by introns, as bacteria are incapable of editing these from the RNA to give a mature translatable product. Proteins that have been expressed in bacteria will not have undergone certain post-translational modifications that would occur in the eukaryotic cell. This may be a problem when experiments are being carried out to determine, for example, whether the transcription factor encoded by a certain cloned piece of DNA encodes the same factor as that giving a certain band in a retardation assay, because proteins lacking certain modifications may not bind properly, may interact correctly with other proteins, or may even possess an incorrect apparent molecular weight. However, the production of unmodified protein may be an advantage when studying the modification process itself. If the protein expressed in a bacterial system behaves differently from that expressed in a cellular extract or eukaryotic expression system, it would suggest that that protein required post-translational modification for full functional activity. The converse experiment is also possible, i.e. a phosphorylated protein from a cell extract or eukaryotic expression system may also be dephosphorylated by treatment with phosphatases, and the effect of this treatment on binding activity studied. As a final comment, it is important to make sure that all experiments involving the expression of transcription factor protein are properly controlled. First, binding assays carried out using extracts expressing a cloned protein must be compared with assays using extracts made from cells expressing the intact vector only, in order to identify any binding of bacterial rather than expressed eukaryotic proteins. However, even when this precaution is taken it is often impossible to rule out the possibility that the expressed protein is binding with the aid of a bacterial cofactor protein (which may or may not be an analogue of a protein that is required in vivo), and results must be interpreted with this in mind.

3.2 Expression of proteins in mammalian cells The expression of proteins in mammalian cells is one of the most commonly used means of producing recombinant transcription factor for gel retardation or expression assay analysis (see Chapter 8, Section 3). It is often possible to express the cloned transcription factor in a cell line in which it is usually active, therefore ensuring that all cofactors and modification proteins required for the expression of that factor will be present. If this is done, it will be necessary to devise some means of distinguishing the cloned transcription factor from the endogenous factor present in the cell. This may be relatively 15

C. L. Dent et al. easy if the cloned factor is expressed as a fusion protein, or has been deleted or truncated, and is therefore a different size from the endogenous factor, but it is possible that this may constitute a major problem. If so, or if the nature of the investigation makes such an experiment of interest, it is also possible to express a transcription factor in a mammalian cell line in which it is not usually expressed. The transcription factor is cloned into an expression vector as already described in the discussion of expression of proteins in bacterial systems. The vector systems are similar to those found for bacterial systems, in that the coding DNA must be inserted in the correct orientation for expression, in the correct reading frame, and that the resulting expression product may be synthesized as a fusion protein. In addition, the vector must contain sequences to specify capping and polyadenylation of the transcribed RNA sequence. Protein expression is under the control of a mammalian promoter (often a mammalian viral promoter such as the cytomegalovirus gene promoter) which may be inducible. The methods for introducing DNA into mammalian cells by transfection are given in Chapter 8. Extracts can be made from the cell line expressing the transcription factor in order to detect the DNA binding of the transfected factor by use of the DNA mobility shift assay. Mini-preparation methods should initially be used to isolate protein, but it is possible that transiently transfected cells will contain insufficient protein for detection by a mobility shift assay. If this is the case, it is possible to prepare a protein extract from more plates of transfected cells, or to prepare stably transfected cell lines such that a higher proportion of the cells present are expressing the transcription factor protein.

3.3 Expression by in vitro transcription and translation In vitro transcription and translation is another method that is used very widely for the production of recombinant or cloned transcription factor protein (see Chapter 8, Section 3.1 and ref. 16). The major advantage of this method is that it gives a clean protein product which is likely to be free of contaminating transcription factors. cDNA encoding the transcription factor is cloned into a vector which possess the promoter sequence for a viral RNA polymerase; usually SP6, T3, or T7 polymerases are used (17). The bluescript vector is particularly suitable for this purpose as it contains both T3 and T7 RNA polymerase promoters either side of a multiple cloning site polylinker. This enables the in vitro transcription of either sense or antisense RNA from the cloned sequence, depending in which direction transcription is inititated. After checking an aliquot of the RNA by electrophoresis (there should be sufficient RNA to visualize on an agarose slab gel without the necessity for Northern blotting) it can be translated in a cell-free translation system. A good system, particularly for the translation of transcription factor proteins, is the rabbit reticulocyte lysate. Reticulocytes do not contain a nucleus and, 16

1: The DNA mobility shift assay therefore, have no requirement for transcription factors. This makes the interpretation of mobility shift assay results from the cloned transcription factors much simpler in the absence of additional endogenous factor or factors. However, it is still necessary to include an aliquot of translation extract (after incubation without RNA or with antisense RNA in a mobility shift) to make absolutely certain that it does not contain endogenous proteins that are capable of binding to the probe and thus confusing the interpretation of results. It is also possible to label any translation product radioactively with [35S]methionine during translation, allowing the protein to be sized accurately on an SDS-polyacrylamide gel. Commercially available kits now make simultaneous transcription and translation reactions possible.

3.4 Expression of cloned transcription factor in baculovirus Baculovirus vectors are a very efficient means of generating recombinant protein (see Chapter 8, Section 3.5). Baculovirus is a virus that infects insect cells and the cloned protein is, therefore, produced in an insect cell line. Vmw65 has been expressed in a baculovirus system (18), with the result that enough protein was expressed to be clearly visible when infected (with recombinant baculovirus) and mock-infected cell extracts were compared by electrophoresis on an SDS-polyacrylamide gel. In this case, the recombinant protein represented about 3% of the soluble protein extracted from the infected cells, equivalent to 20 mg of recombinant protein from 109 infected cells. This protein was used directly for DNA mobility shift studies.

3.5 Purification of transcription factor protein The purification of transcription factor proteins can be applied both to endogenous transcription factors and to an artificially expressed factor, and can be used to purify (or partially purify) a factor from a cell extract (see Chapter 4). The purified factor may then be used for DNA binding assays to further study the relationships involved in binding to the probe, or can be used as a substrate for protein sequence analysis as part of a strategy for the isolation of a clone encoding the factor. Purification methods may also be used to prepare an extract depleted in a factor. The depleted extract can be used to dissect the activity of the protein, for example by adding back a cloned transcription factor.

4. Investigation of DNA binding specificity Sequence-specific DNA binding proteins will often be capable of binding to a series of variations on a basic consensus sequence. A binding site that is further removed from the consensus may bind the factor less strongly and, therefore, be a weaker site for transcriptional activation. Differences in the sequence of flanking DNA outside of the core consensus binding site may also 17

C. L. Dent et al. affect binding. These differences in affinity of different sites for the same factor will contribute to the differential effects of a single transcription factor on different genes, and can be studied by competition analysis. The approach used for the measurement of the affinity of a transcription factor for different binding sites is discussed below. If a binding reaction is set up between a radiolabelled oligonucleotide and protein extract, the protein-DNA complexes will form retarded bands on electrophoresis as already explained. If, when the reaction is set up, as well as adding radiolabelled binding site oligonucleotide, a 100-fold molar excess of unlabelled oligonucleotide is added, then the DNA binding protein will be able to bind the unlabelled oligonucleotide and the radioactive probe. This will result in a decrease in the amount of factor available for binding to the probe by competition from the unlabelled site, and subsequent reduction in the intensity of the retarded band. If, rather than introducing an unlabelled oligonucleotide identical to the probe, an oligonucleotide corresponding to a different site to which the factor cannot bind is added, there should be no difference in the intensity of the retarded band because the competing oligonucleotide is not able to bind the factor. This is the standard means by which a transcription factor is demonstrated to bind in a sequence-specific manner (Figure 3). This kind of experiment may then be taken a stage further by introducing, as competitor, variations on the same binding site. The new sequences chosen may be related binding sites found in other genes in order to study the comparative affinities of natural binding sites, or they may be mutations of the consensus binding site designed to study the effect of base substitutions on the DNA-protein interaction. If a related site that binds the same factor is introduced at an excess into a binding reaction, the extent of competition should vary from that obtained using the site identical to the probe. A higher affinity binding site would compete more protein away from the labelled site, resulting in a stronger reduction in the intensity of the retarded band. A lower affinity site would be less efficient in competition, leaving more protein for complex formation with labelled oligonucleotide, resulting in less reduction in the retarded band. Some related sequences may be totally unable to bind, resulting in the competed binding reaction giving a retarded band identical to that obtained when no competitor DNA is added (or indeed, competitor DNA for a totally unrelated binding site). The comparative affinity of the protein for different sites can be determined by adding different amounts of competitor for the sites to a uniform binding reaction, and measuring the amount of competitor required to reduce the level of protein bound to the radioactive probe by a particular amount. For example, if one sequence competed all of the protein off the probe at a 10-fold molar excess of competitor, but another related sequence required a 100-fold excess, it could be concluded that the first sequence comprised the higher affinity binding site for the factor in question. Experiments of this type have been used in our laboratory to provide 18

1: The DNA mobility shift assay

Figure 5. DNA mobility shift assay using four different labelled octarner oligonucleotides (a-d) and extracts prepared from ND7 neuronal cells (A) and Daudi B cells (B). Arrowheads indicate the positions of Oct-1 and Oct-2, Note the different affinities of B cell and neuronal Oct-2 for the different oligonucleotides. For further details see ref. 19.

evidence that the neuronal and B-cell forms of Oct-2 are not identical (19). The affinities of the proteins for a panel of four oligonucleotidc variants of an overlapping octamer/TAATGARAT consensus sequence (Figure 5) were measured. A mobility shift assay was carried oul on oligonucleotide C and competed with varying amounts (1-, 10-, and 100-fold molar excess) of the four oligonucleotides and an unrelated Spl binding site as a negative control (Figure 6), Oct-2 from B cells binds oligonucleotide A with a higher affinity than any other sequence, as demonstrated by the fact that A competes for binding more efficiently. However, Oct-2 from neuronal cells behaves differently, binding to and competing most efficiently with C, thus demonstrating a higher affinity for this sequence.

5. Characterization of DNA binding proteins 5.1 Addition of antibodies Antisera raised against a transcription factor may affect its binding by one of two methods. First, the antibody may bind to a site on the transcription factor that is essential for DNA binding, thus totally blocking the ability of the factor to bind DNA and resulting in the complete absence of a DNA-protein complex from the gel. Second, if the antibody binds to a non-essential site on the factor, DNA binding may not be impaired, but the mobility of the complex 19

C. L. Dent et al.

Figure 6. Competition analysis of the B cell (B) and ND cells
will be altered by the binding of the antibody. The resulting complex will possess a lower mobility than the DNA-protein complex alone, due to the involvement of an additional factor. This will result in a super-shift of the DNA-protein complex to a position of lower mobility, due to binding of the antibody, Antisera may be used to determine whether a mobility shift is caused by a previously identified transcription factor. Evidence such as the size of the shifted band and binding site specificity of a protein could suggest that it may be a previously characterized factor. The binding of an antibody may help towards confirming that the protein is indeed the factor that it is suspected of being (or may demonstrate that it is unique). In one series of experiments, antibodies raised against the transcription factor Oct-1 were found to bind to NFIII, an adenovirus DNA replication factor (20). This provided strong evidence that NFIII and Oct-1 were indeed identical, and was supported by additional experiments demonstrating that the anti-Oct-I antibodies could block NFIII stimulation of adenovirus DNA replication, and that Oct-1 could functionally substitute for NFIII in this system. This work also serves to demonstrate how the results obtained from the DNA mobility shift assay are particularly valuable when coupled with evidence obtained by other means. If an antiserum does bind to a factor, it does not necessarily prove that it is identical to the protein to which the antiserum was raised; binding may result from the proteins sharing a common domain that includes the epitope for the 20

1: The DNA mobility shift assay antibody. However, antisera may demonstrate that two proteins that appear to be different in a mobility shift assay may be related if they share a domain containing the epitope. Antibody binding experiments are carried out as already described in Protocol 6, with the only modification being that the antisera and protein extract should be mixed together and pre-incubated for about 30 min on ice before the addition of the extract to the binding reaction. When setting up such an experiment, it is also important to include a control binding reaction containing the antisera in the absence of protein extract. This is because serum may contain proteins capable of binding to the DNA probe, resulting in extra shifted bands that may affect the interpretation of results. Similarly, a control should be included containing the extract and an unrelated antibody or pre-immune serum.

5.2 Addition of potential ligands This approach involves attempting to modulate the binding characteristics of a protein by adding to or removing from the binding reaction certain cofactors that may be required for binding. Such ligands would be added to the standard binding reaction mix described in Protocol 6. Factors that may influence the binding reaction include the presence of ions such as calcium, magnesium, and zinc. These may be removed by the addition of chelating agents such as EDTA and EGTA to the binding reaction mix. The activity of a transcription factor in a depleted extract can yield important information about its structure. Spl was demonstrated to be inactive in HeLa extracts that had been depleted of zinc by treatment with EDTA. This suggests that Spl probably binds to DNA by a zinc-dependent structure known as the zinc ringer (21). In the same series of experiments, Oct-1 was shown to be fully active in zinc-free extracts, demonstrating that its activity is not dependent on zinc. Some DNA binding proteins also act as receptors for endogenous substances, examples being the steroid and retinoic acid receptors. It would thus be obvious in these cases to study the effect of adding the natural ligand to the binding reaction, along with studies involving the addition of ligand analogues which may have either a stimulatory or inhibitory effect compared with the natural ligand (see Chapter 8).

5.3 Proteolytic clipping band-shift assay The proteolytic clipping band-shift assay can be used alongside antibody binding to examine the relatedness of two proteins. This type of assay involves the addition of various dilutions of a specific protease, e.g. trypsin or chymotrypsin, to a DNA-protein complex, resulting in the production of partial cleavage products that become more completely cleaved with the addition of higher concentrations of protease. The pattern of partial and complete 21

C.L. Dent el al.

Figure 7. Protease clipping assay in which DMA mobility shifts were carried out with an oligonucleotide which binds the octamer-binding protein Oct-1 (arrowed) using an untreated extract (track 1) and the same extract pre-treated with various amounts of chymotrvpsin (tracks 2-5).

cleavage products gives a typical 'fingerprint' to a protein. If two proteins are identical, they will give identical cleavage products. If two proteins are related, sharing some homologous domains, then some similar cleavage products may be present within a somewhat different fingerprint. Figure 7 shows a typical result from a proteolytic clipping band-shift assay, with the first track showing the binding reaction in the absence of protease, and the following tracks showing the cleavage products on addition of increasing amounts of protease. A method for proteolytic clipping using trypsin and chymotrypsin is given in Protocol 9. Protocol 9.

Proteolytic clipping band-shift assay

Equipment and reagents As in Protocol 6 Stock solutions of chymotrypsin: 0.004, 0.002, 0.001, and 0.0005 units/ul

Stock solutions of trypsin; 30, 10, 3, and 1 ng/ul

22

1: The DNA mobility shift assay Method 1. Set up binding reaction exactly as described in Protocol 6. Incubate at room temperature for 10 min (not on ice as in Protocol 6). Five tubes should be prepared for each type of protease to be included, for four different concentrations of protease and a control without protease. 2. During the binding incubation, prepare serial dilutions of the proteases in water. Keep a stock solution of protease at 4000 units/ml for chymotrypsin and 30 ug/ml for trypsin at 4°C. Working stock solutions of chymotrypsin and trypsin should be prepared by dilution before use. 3. Add 1 ul of each dilution of protease to the relevant binding reaction tube, and 1 ul of water to the no-protease control tube. Incubate at room temperature for a further 10 min. 4. Load samples on to a 4% polyacrylamide gel in 0.25 x TBE and electrophorese as described in Protocol 6.

6. Study of protein—protein interactions Many DNA binding proteins interact with other proteins during their activity. Factors such as Apl are active as a heterodimer of two proteins, Fos and Jun, that will then together recognize and bind to a site in the DNA. A different type of protein cofactor is Vmw65, a HSV-1 protein that binds to Oct-1, forming a complex which can then bind to the HSV-1 immediate-early gene TAATGARAT consensus and enable transactivation of viral RNA synthesis. The Vmw65 protein itself is not capable of binding DNA, although it does contact the DNA sequence when complexed with Vmw65 (2,3). The inclusion of additional proteins into DNA binding complexes may be studied using the gel retardation assay. The binding of an extra protein into a complex will further increase its apparent molecular weight on electrophoresis. Vmw65 is by far the most extensively studied cofactor involved in DNA binding complexes (2,3). Nuclear extracts made from cells that have been infected with HSV-1 are compared with extracts from mock-infected cells. The mockinfected extract will bind to a TAATGARAT probe, giving a shift typical of the Oct-1 protein, but the infected extract will give a further shift of lower mobility due to the binding of an additional factor, Vmw65, into the complex. Figure 8 shows the type of mobility shift given by Oct-1 when complexed with Vmw65 to form a larger complex. The shift typical of Oct-1 disappears and a new, lower mobility shift appears, providing further evidence that Oct-1 is the protein involved in the interaction with Vmw65 (3). 23

C.L.Dent et al.

Figure 8. DNA mobility shift assay using a radioactively labelled oligonucleotide containing the octamer-related TAATGARAT sequence. Note that incubation with HeLa cell extract alone results in Oct-1 binding to the probe [track 1). However, addition of increasing amounts of HSV virion extract containing Vmw65 results in the formation of the larger infected extract complex (IEC) containing both Oct-1 and Vmw65 (tracks 2-5K This complex is not formed upon incubation with HSV virion extract alone, indicating that it requires both Oct-1 and Vmw65, and that Vmw65 alone does not bind to DNA (tracks 6-10). Data kindly provided by Dr C. M. Preston, For further details see ref. 3.

24

1: The DNA mobility shift assay

7. Concluding comments The methods and examples given in this chapter demonstrate that the DNA mobility shift assay is a simple yet powerful technique for the identification and study of DNA binding, although it must be remembered that the technique does have its limitations and that the demonstration of DNA binding in vitro does not necessarily mean that a particular gene is under the control of a particular transcription factor in vivo. The following chapters give a detailed description of the types of experiment that can, and should, be carried out to support evidence gained using the DNA mobility shift assay.

References 1. Wheatley, S. C, Dent, C. L.( Wood, J. N., and Latchman, D. S. (1991). Exp. Cell. Res., 194,78. 2. O'Hare, P. and Coding, C. R. (1988). Cell, 52,435. 3. Preston, C. M., Frame, M. C., and Campbell, M. E. M. (1988). Cell, 52,425. 4. Gerster, T. and Roeder, R. G. (1988). Proc. NatlAcad. Sci. USA, 85,6347. 5. Lillycrop, K. A., Dent, C. L., Wheatley, S. C., Beech, M. N., Ninkina, N. N., Wood, J. N., and Latchman, D. S. (1991). Neuron, 7,381. 6. Fried, M. and Crothers, D. M. (1981). Nucleic Acids Res., 9,6505. 7. Garner, M. M. and Revzin, A. (1981). Nucleic Acids Res., 9, 3047. 8. Dent, C. L., Mclndoe, G. A., and Latchman, D. S. (1991). Nucleic Acids Res., 19, 4531. 9. Janson, L. and Pettersson, U. (1990). Proc. NatlAcad. Sci. USA, 87, 4732. 10. Sambrook, J., Fritsch, E. T., and Maniatis, T. (ed.) (1989). Molecular cloning, a laboratory manual (2nd edn). Cold Spring Harbor Press, NY. 11. Dignam, J. D., Lebovitz, R. M., and Roeder, R. G. (1983). Nucleic Acids Res., 11, 1575. 12. Manley, J. L., Fine, A., Cano, A., Sharp, P. A., and Gefter, M. L. (1980). Proc. NatlAcad. Sci. USA, 77, 3855. 13. Bradford, M. (1976). Anal. Biochem., 72, 248. 14. Schreiber, E., Matthias, P., Muller, M., and Schaffner, W. (1989). Nucleic Acids Res., 17, 6419. 15. Roberts, S. B., Segil, N., and Heintz, N. (1991). Science, 253, 1022. 16. Neuberg, M., Adamkiewicz, J., Hunter, J. B., and Muller, R. (1989). Nature, 341, 243. 17. Green, M. R., Maniatis, T., and Melton, D. A. (1983). Cell, 32, 681. 18. Kristie, T. M., Le Bowitz, J. H., and Sharp, P. A. (1989). EMBO J., 8, 4229. 19. Dent, C. L., Lillycrop, K. A., Estridge, J. K., Thomas, N. S. B., and Latchman, D. S. (1991). Mol. Cell. Biol., 11, 3925. 20. Pruijn, G. J. M., van der Vliet, P. C., Dathan, N. A., and Mattaj, I. W. (1989). Nucleic Acids Res., 17, 1845. 21. Kadonaga, J. T., Garner, K. R., Musiarz, F. R., and Tjian, R. (1987). Cell, 51,1079.

25

This page intentionally left blank

2 Footprint analysis of DNA-protein complexes in vitro and in vivo CRAIG SPIRO and CYNTHIA T. McMURRAY

1. Overview Protein-DNA complexes mediate essential processes such as transcription and DNA replication. Footprint experiments use chemicals and enzymes that modify DNA to obtain detailed information on the individual nucleotides in protein-DNA complexes, even inside living mammalian cells. Chemical modification of a DNA molecule can inhibit binding of sequencespecific proteins because of changes at critical nucleotides (Figure 1A, step 1, 'Interference'). On the other hand, when sequence-specific proteins are bound to DNA they can protect nucleotides in the binding site from chemical modification and from nucleases (Figure IB, step 1, 'Protection'). When visualized by denaturing gel electrophoresis, interference or protection will bring about an empty spot that looks like a footprint in the pattern of bands (Figure 1A, steps 4 and 5; Figure IB, steps 3 and 4)). The footprint thereby identifies specific nucleotides in the protein-DNA complex and can indicate changes in the conformation of the DNA. Footprinting is a direct approach for analysing the interactions of individual proteins with DNA, the structure of DNA in control regions, and the interactions of binding sites (and their bound proteins). In 1978, Galas and Schmitz called footprinting a 'simple conjoining of the Maxam-Gilbert DNA-sequencing method and the technique of DNAaseprotected fragment isolation' for detecting protein-DNA binding specificity (1). Mirzabekov, Gilbert, and colleagues had previously studied proteinDNA contacts and chromatin structure using the alkylating agent dimethyl sulfate (DMS) (2). DMS and DNAase I remain the most commonly used agents, but other chemicals and nucleases are also valuable in the study of protein-DNA interaction and of DNA structure. The availability of thermostable DNA polymerases and thermal cyclers has simplified footprint experiments. Protein complexes with large molecules (such

Craig Spiro and Cynthia T. McMurray

as closed, supercoiled plasmids) can be analysed directly and, using ligationmediated polymerase chain reaction (PCR) (3), modification of individual nucleotides in living mammalian cells can be detected. While in vivo analysis requires many steps (described below), the individual steps are not technically difficult, so that footprinting can be used to study protein-DNA complexes and the interactions of DNA sites (and their bound proteins) during transcription. 28

2: Footprint analysis of DNA-protein complexes Figure 1. Schematic view of interference and protection footprinting methods. (A) Interference. 1. Double-stranded DNA molecule is 32P-end-labelled on one strand only and then modified with a chemical agent to alter at most one residue per molecule. Dots indicate positions of modification. The radioactively labelled, chemically modified DNA is mixed with sequence-specific DNA binding protein (oval). 2. The binding reaction is loaded on to a 5% native polyacrylamide gel to resolve protein-DNA complex from free DNA. The wet gel is exposed to film for several hours, and the film is used as template to cut the protein-bound DNA and free DNA from the gel. 3. The DNA is eluted from the gel, purified, and cleaved. Certain fragments will be under-represented in the plus protein sample (arrow), since modification of certain bases will have inhibited protein binding. 4. DNA fragments are resolved on sequencing gels. Number of counts loaded in each lane should be the same to allow comparison of bands. Bands represented in the free DNA lane but absent in the bound DNA lane indicate residues important for binding as their modification made protein binding less likely. 5. DMS is useful in both protection and interference reactions. It does not cleave the DNA backbone, but it modifies bases. Subsequent treatment with hot piperidine causes cleavage of the DNA specifically at modified Gs. The gel image shows that each G in the DNA is not equally susceptible to DMS modification. Comparison of naked and plus protein DNA indicates bases critical in DNA-protein contact (arrows). (B) Protection. 1. Double-stranded DNA molecule is 32Pend-labelled on one strand and is used in two separate binding reactions—one lacking and one containing a sequence-specific DNA binding protein (oval). Arrows indicate potential cleavage or modification sites. 2. The binding reactions are modified chemically (e.g. DMS) or treated with a nuclease (e.g. DNAase I), and the DNA is then purified. Chemically modified DNA is then cleaved. Certain fragments will be under-represented in the plus protein sample (arrow) because certain sites will have been protected from modification or cleavage. 3. The fragments from chemical modification/cleavage or nuclease cleavage are resolved on sequencing gels. Absence of certain fragments in plus protein lane (seen as a footprint) indicates base(s) protected from modification or cleavage by the bound protein. 4. DNAase I has been the most commonly used footprinting reagent. While not every nucleotide is equally susceptible to DNAase I cleavage, cuts are fairly evenly distributed along the length of DNA. Gel image shows DNA cleavage in the absence or presence of protein and the footprint created by bound protein.

2. In vitro footprinting 2.1 Analysis at base resolution of binding sites on 32P-endlabelled fragments Identification of the nucleotides modified in a footprint experiment (see Figure 1A, steps 4 and 5; Figure IB, steps 3 and 4) requires an unambiguous reference point. Footprints are visualized from the sequencing gels on which radioactive DNA fragments have been resolved. Orientation can be achieved by 32P-end-labelling a double-stranded molecule on one strand only (Figure 1, top; Figure 2). To analyse both strands, complementary sets of reactions are set up, each set with only one of the strands labelled (Figure 2B, bottom). Extension of a single 32P-end-labelled primer along the DNA is an alternative means of detection and can be used to analyse both in vivo and in vitro modified DNA in protection experiments (as described in subsequent sections; Figure 3, Figure 4, Figure 5). 29

Craig Spiro and Cynthia T. McMurray

Figure 2. Preparation of DNA molecule labelled on one strand only. (A) Region to be analysed ('promoter'), which is on a restriction fragment with unique ends, is inserted into multiple 'cloning site of a vector with additional unique sites (e.g. Xhol and fcoRI, indicated by arrows) outside of the sites used to insert promoter fragment. (B) New plasmid CpPromoter') is digested with enzymes that recognize outside sites (Xhol and EcoRI). The resulting fragment is 32P-end-labelled at both ends, and one of the inside sites (Hindlll or Xba\) is used to remove the label from one end only.

2.1.1 Linear DNA molecule for footprint Linear DNA molecules from 75 to 600 base pairs are suitable. The binding site can be between 20 and 400 bases from the end of the molecule, with a distance of 50 to 200 best for analysis on a 40 cm long, 0.4 mm thick, 6% poly30

2: Footprint analysis of DNA-protein complexes

Figure 3. DNAase I protection on closed circular plasmid (Protocol 7.) (A) For example, the DNA molecule is a closed, circular plasmid containing a mammalian cAMP-responsive enhancer (CRE; shaded rectangle in detail) and the coding region of the chloramphenicol acetyl transferase gene (filled region; wavy arrow shows direction of transcription). Analysis requires a primer for sequencing through the enhancer region on each strand (CREfp5' and CREfp3' primers). (B) Sequence of a section of the rat prodynorphin enhancer region which contains three CREs (enclosed in rectangles) (20). The primers are indicated by horizontal bars [arrow indicates direction of extension). (C) Protection of the top strand by increasing amounts of purified CREB (CRE-binding) protein (9) is visualized by extension of dynfp3' primer, for which fragments of the top strand are templates. Dideoxy G sequencing reaction (Protocol 5) of the bottom strand (i.e. using the dynfp3' primer) is used to identify specific bases. Note the size of the area protected from DNAase I, the contrast between protected and unprotected bands, and the greater resolution of smaller fragments. (C) adapted from ref. 9, with permission. 31

Craig Spiro and Cynthia T. McMurray

acrylamide, denaturing gel. Longer fragments are less well resolved than shorter fragments (see, for example, Figure 3C). The double-stranded DNA molecule 32P-labelled on just one end can be prepared from a restriction fragment subcloned into the multiple cloning site of a plasmid vector (e.g. Figure 2). The restriction fragment with your region of interest ('promoter' in Figure 2A) must have unique sites on the ends. If convenient sites do not flank your region of interest, create them by amplification of the region with primers designed to introduce the sites (e.g. HindIII and XbaI in Figure 2A). Using the unique sites (HindIII and XbaI), subclone the fragment into the multiple cloning site of the vector between two other unique restriction sites (Figure 2A, arrows) to create the plasmid with your region of interest ('pPromoter' in Figure 25). Cut the fragment from the plasmid using the outer restriction sites (XhoI and EcoKL in Figure 2B), dephosphorylate with calf intestinal phosphatase (CIP), heat inactivate at 15°C, and gel purify the fragment containing the region of interest (Xhol-EcoRl 'promoter' fragment in Figure 25). Store this fragment at -20°C until needed for footprinting. Then, end label the fragment using [32P]ATP and T4 polynucleotide kinase. (Protocol 1). Cut with one of the inside enzymes (i.e. 32

2: Footprint analysis of DNA-protein complexes Figure 4. LMPCR experimental design (re Protocols 8-14). (A) Flow chart for preparation of DNA for LMPCR. (1) Prepare duplicate sets of cultured cells (and treat with inducers of transcription if necessary). (2) Treat one set of cells with PBS only and the other set with DMS. (3) Harvest cells and purify DNA. (4) Modify some of the DNA from PBS-treated control cells in vitro with DMS. (Control DNA can be used for genomic sequencing or methyl cytosine analysis as well.) (5) Cleave DMS-modified DNA with piperidine and proceed to LMPCR. (B) Design of primers for LMPCR. LMPCR requires three nested primers, with successively higher Tm. Primers are shown along one strand only, but a set of primers should be prepared for each strand to flank the region being studied. Primer 3 (the final one used) should be about 30-100 bases (not less than 10) from the 5' end of the region being examined (shaded rectangle). Primer 1, used for primer extension to create bluntended molecules, should be 20 bases long with 12 G + C (60% GC). Primer 2, the amplification primer, must be 3' to primer 1. The distance between primer 2 and primer 1 is not critical. They can be separated by as much as 50 to 100 bases, but they can also overlap a little. Primer 2 should never overlap primer 1 by more than 5 bases. Primer 2 should be 25 bases long with 15 G + C (60% GC) (as is LMPCR.1). Primer 3 should overlap primer 2 by at least 15 bases (it is fine if they completely overlap) and must extend a few bases 3' to primer 2. Primer 3 should be 28 bases long with 17 or 18 G + C (60-65% GC). If the sequence of the gene does not allow these precise lengths and G + C content, then adjust the primer length and GC content so that the calculated Tm is the same. (C) Amplification and detection of one gene-specific fragment by LMPCR. 1. Genomic DNA is modified and cleaved. 2. Gene-specific primer 1 anneals to fragment from gene, and primer extension creates blunt-ended molecule. 3. Unidirectional linker is ligated to the blunt-ended fragments. 4. Gene-specific primer 2 anneals to fragment from gene and primer extension creates double-stranded molecule with sequence of linker primer (LMPCR.1) on the other strand. 5. During PCR amplification, gene-specific primer 2 and LMPCR.1 anneal to the complementary strands. 6. Fragment is amplified. 7. Extension of 32P-end-labelled genespecific primer 3 allows detection of fragment by gel electrophoresis and autoradiography.

HindllL or XbaI; Figure 2B) to create a double-stranded molecule labelled on just one end (one strand) (Figure 2B, bottom). DNA molecules labelled on just one end can be prepared by alternative methods, including amplification by PCR with one 32P-labelled primer (Protocol 1} and one unlabelled primer. Protocol 1. End labelling DNA molecules using T4 polynucleotide kinase and [32P]ATP Equipment and reagents End-labelling grade [32P]ATP: 259 TBq mmol, 22.7 uM, 5.92 GBq/ml (ICN) Spin desalting column (e.g. Bio-Spin 6; BioRad)

T4 polynucleotide kinase (New England Biolabs, 10U/ul 10 x T4 polynucleotide kinase buffer. 100 mM MgCI2, 50 mM DTT, 1 mM spermidine, 1 mM EDTA, 500 mM Tris-HCI pH 7.6 (store in aliquots at -20°C)

Method 1. Synthetic oligonucleotides with 5'-OH groups are ready for end labelling. Restriction fragments with 5'-PO4 must be dephosphorylated.8 33

Craig Spiro and Cynthia T. McMurray Protocol 1.

Continued

2. To label 40 pmol oligonucleotide primer, combine 19 ul H2O 3 ul 10 x T4 polynucleotide kinase buffer 3 (J end-labelling grade [32P]ATP 4 ul 10 (uM oligonucleotide (or DNA fragment) 1 ul T4 polynucleotide kinase (10 U/ul, New England Biolabs) 3. Incubate at least 15 min at 37 °C. 4. Stop with 20 ul 25 mM EDTA. 5. Remove unincorporated [32P]ATP with Bio-Spin 6 desalting column.6 6. Count 1 ul in scintillation counter (Cerenkov). Expect about 2 x 106 to 4 x 106 c.p.m./pmol for labelled oligonucleotides of 20-40 bases. Longer molecules will have lower specific activity. Store at 4°C. (If counts are significantly lower, e.g. less than 5 x 105 c.p.m./pmol, check reaction components.) "Dephosphorylate, if necessary, with calf intestinal alkaline phosphatase (which is active in restriction enzyme buffers) and then heat inactivate the phosphatase for 10 min at 75°C. (To completely remove phosphatase, gel purify the DNA or extract once with phenol: trichloromethane:iso-amyl alcohol (25:24:1) and ethanol precipitate the DNA.) ''Some unincorporated [32P]ATP will not be retained by column, but it will not interfere with subsequent reactions. Other methods, such as ethanol precipitation, can more effectively remove [32P]ATP, but they are more likely to remove labelled DNA as well, so the spin column is a good compromise.

2.1.2 Source of protein for binding reaction The protein-DNA complex being studied will affect the clarity of the footprint. In a protection experiment, a footprint will be difficult to detect if there is low occupancy of the binding site. Furthermore, the hypothesis being tested will dictate what source of protein is suitable. For example, crude protein may be used to locate the binding site or to study relative binding affinity, but purified protein is required for determination of absolute equilibrium binding constants (4). Nuclear extracts can be used for studies of interactions among sites. Dignam and colleagues described a method for extracting transcriptionally competent nuclear extracts (5), but more rapid preparation of nuclear extract from a relatively small number of mammalian cells can be satisfactory for evaluating the effect of a series of treatments (e.g. 6-8) (Protocol 2). From 10 to 150 ug nuclear extract may used for binding in a protection experiment. Figure 5. LMPCR. (Section 3.4, Protocol 14) (A) Portion of the rat proenkephalin enhancer region analysed by LMPCR (9). cAMP-response elements are enclosed in rectangles. Nested primers (Figure 4B) are numbered in order of use. (B) In vivo footprint of proenkephalin enhancer region in C6 glioma and R2C Leydig tumour cell lines. Proenkephalin gene is expressed in C6 but not in R2C cells. Naked DNA was purified and 34

2: Footprint analysis of DNA-protein complexes

then modified by DMS in vitro; '+ forsk' cells were treated with forskolin before DMS modification in vivo. Residues that are less reactive (open triangles) or more reactive (filled triangles) in cells expressing proenkephatin are indicated along the images of the gels and in the representation of the sequence. Top strand gel image was produced from reactions with bottom strand primers (Figure 5A). Reproduced from ref, 9, with permission. 35

Craig Spiro and Cynthia T. McMurray Protocol 2. Nuclear extract from small number of cultured mammalian cells Equipment and reagents PBS Low-salt buffer: 1.5 mM MgCI2, 10 mM KCI 0.5 mM DTT, 10 mM Hepes, pH 7.9 10% NP40 High-salt extraction buffer: 25% (v/v) glycerol 1.5 mM MgCI2, 1 mM DTT, 420 mM NaCI 20 mM Hepes pH 7.9

Dilution buffer: 5 mM Hepes pH 7.9, 20% glycerol, 1 mM DTT Protease inhibitors: e.g. PMSF (to 0.5 mM) may be used to supplement buffers Refrigerated microcentrifuge or refrigerated superspeed centrifuge with adapters to hold microcentrifuge tubes

Method 1. Wash cells (5 x 105 to 1 x 107 cells) with PBS and collect in microcentrifuge tube. (For cells in suspension resuspend in 1 ml PBS and transfer to microcentrifuge tube. For monolayers, scrape into 1 ml PBS and transfer to microcentrifuge tube.) Keep on ice until all cells have been harvested. 2. Pellet cells by centrifugation for 10 sec in microcentrifuge. 3. Resuspend cells in 400 ul low-salt buffer by pipetting up and down a few times. 4. Let swell 15 min on ice. 5. Add 25 (ul 10% NP40.

6. Vortex 10 sec. Return tubes to ice. 7. Spin 10 sec to collect nuclei. (Discard supernatant.) 8. Resuspend pellet in 50 ul high-salt buffer by gently stirring with pipette tip. Pellet of nuclei may be clumpy. The suspension should not become viscous. Keep tubes on ice. 9. Rock tubes at 4°C for 20 min for extraction of proteins. 10. Spin tubes at 16000 x g for 45 min at 4°C in refrigerated microcentrifuge or in refrigerated superspeed centrifuge. 11. Supernatant is the nuclear extract. Dilute with equal volume dilution buffer. 12. Aliquot; freeze in dry ice/ethanol bath, then store at -70°C until ready for use.

In a protection experiment, bases are protected from modification by the bound protein (Figure 1B, steps 3 and 4; Figure 3C). Sensitivity of DNA to modifying agents depends on local structure/environment and on the points of contact with the bound protein. Therefore, in a 'protection' experiment, some of the bases in or near the binding site may not be protected at all but instead 36

2: Footprint analysis of DNA-protein complexes may become more reactive (indicated by darker bands) (e.g. Figure 3C; Figure 5B). Using more than one probe can provide supplementary information on the site of binding and points of contact (Section 2.1.4). 2.1.3 Protection from DNAase I digestion Protection from DNAase I is the most commonly reported in vitro footprinting assay (Figure 1B; Protocol 3). DNAase I has relatively little sequence preference, so it cleaves along the entire length of the DNA backbone. Another advantage of DNAase I is that digestions can be carried out over a range of temperature, pH, and salt conditions. DNAase I footprints tend to be large because of the exclusion of the enzyme by bound protein, so the footprints are, therefore, often easily detected (e.g. Figure 1B, step 4; Figure 3C). Most footprint analysis is qualitative, but protection experiments can also yield thermodynamic profiles of protein-DNA interactions. Because DNAase I can probe complexes without perturbing the equilibrium, quantitative analysis of carefully controlled footprint experiments can provide data on equilibria in solution (4). Protocol 3. Protection from DNAase I of linear DNA molecule Equipment and reagents • 0.5 ml and 1.5 ml slliconized microcentrifuge tubes (Fisher Scientific) . DNAase I (RQ1; Promega, 1 U/ul) « Mg2+/Ca2+ supplement: 50 mM MgCI2, 10 mM CaCI2, 10 mM Tris-HCI pH 7.5 • Phenol:trichloromethane:iso-amyl alcohol (25:24:1) • DNAase I stop solution: 400 mM NaCI, 30 mM EDTA, 1% SDS, 200 ng/ml yeast tRNA • DNAase I dilution buffer: 1 mM MgCI2, 1 mM CaCI2,10 mM Tris-HCI pH 7.5

• DNAase I dilution (prepare just before use and keep on ice): 2 ul DNAase I (1 U/M.I, RQ1, Promega) + 98 ul DNAase I dilution buffer

(to 0.02 U/ul working

cyanol, 0.3% bromophenol blue, 0.37% EDTA made up in deionized formamide (store aliquots at -20°C) • Non-specific DNA (e.g. sonicated salmon sperm DNA)" . BSAb

Method 1. Prepare double-stranded DNA molecule strand only (see text; Figure 2; Protocol 1).

32

P-end-labelled on one

2. To generate G sequence ladder, methylate 32P-end-labelled DNA molecule with DMS and cleave with hot piperidine (Protocol 4). (If necessary. A, C, T sequence ladders can also be prepared by basespecific modification of 32P-end-labelled DNA molecule [13].) 3. Set up binding reactions in 0.5 ml siliconized tubes with 20000 c.p.m. 32 P-end-labelled DNA and protein (include a 'no protein' control) in total volume of 40 ul. These samples are radioactive and must be handled appropriately. Use binding conditions suitable for proteinDNA complex being studied.b 37

s

o

l

u

t

i

Craig Spiro and Cynthia T. McMurray Protocol 3.

Continued

4. Bind at temperature (4°C to 37°C) and for time (few min to overnight) suited to the protein-DNA complex being studied. Prepare reagents for DNAase I digestion and tubes for enzyme dilutions.c 5. Digest for 2 min at room temperature with DNAase I as follows (with appropriate pacing, 10 to 12 reactions can be conveniently handled as a group by treating samples in order): (a) (b) (c) (d) (e)

Put samples at room temperature. Begin 2 min timer countdown. Add 5 (ul Mg2+/Ca2+ supplement to each 40 ul binding reaction. When timer goes off, begin 2 min timer countdown. Add 5 ul appropriate DNAase I dilution to each tube in order and stir with pipette tip. (f) When timer goes off, add 50 ul DNAase I stop solution to each tube in order.

6. Extract with 100 ul phenol:trichloromethane:iso-amyl alcohol. Vortex. Separate phases by centrifugation for several min at full speed in microcentrifuge. 7. Transfer aqueous (upper) phase to 1.5 ml siliconized tube. 8. Precipitate by addition of 250 ul ethanol. Mix. Keep at -20°C for at least 30 min. 9. Collect pellet by full speed centrifugation in microcentrifuge for at least 10 min. 10. Dry pellet in SpeedVac concentrator. 11. Resuspend pellet in 6 ul sequencing stop/load dye. Vortex and spin briefly to collect sample. Repeat vortex and spin. Check that DNA is resuspended by holding 1 ul in pipette tip in front of Geiger monitor. Sample may be stored at -20°C. 12. Before analysis on sequencing gel (Protocol 75) denature sample for 5 min at 90°C and then put on ice. "In the presence of certain protein extracts, the correct amount of DNAase I could be 20 x as much and must be empirically determined. Use gel electrophoresis and autoradiography to determine the dilution of DNAase I that leaves half the molecules at full length. b Excess non-specific DNA (0.4-2.0 pg sonicated salmon sperm DNA) keeps constant the substrate for DNAase I and may reduce excess nicking outside the binding site, but it is not essential and at very high levels can compete for protein binding. BSA can also be included (2-20 ug). c Dilute enzyme just before beginning digestions. Keep diluted enzyme on ice.

2.1.4 Other reagents useful for protection analysis Other enzymes, such as nuclease P1 (9), and chemical nucleases, such as phenanthroline-copper, methidium-propyl EDTA, and ferrous EDTA (re38

2: Footprint analysis of DNA-protein complexes viewed in 10) are also useful reagents in protection experiments. Phenanthroline-copper has been used in a combination gel shift-protection experiment that can overcome the difficulty posed by low occupancy of sites in a binding reaction (11). After allowing complexes to form, protein-DNA complexes are resolved from free DNA on a native polyacrylamide gel (as in Figure 1A, step 2). The protein-DNA complex is digested with the chemical nuclease while still embedded in the gel matrix (11). This approach can also be used to analyse the individual retarded complexes observed in a gel shift (10). Ferrous EDTA produces the smallest chemical nuclease for footprinting— the hydroxyl radical, which is like a reactive water molecule (12). Hydroxyl radical has access which is denied the larger molecules such as DNAase I and it can, therefore, provide valuable complementary information by identifying very small patches of protection. In experimental design, hydroxyl radical footprinting is the same as DNAase I protection (Protocol 3) except for the cutting reagent used. Binding reactions are set up and then exposed to the nuclease, which is prepared from ferrous ammonium sulfate and EDTA stock solutions, mixed at a molar ratio of 1:2. The reaction is typically initiated by addition of 50 mM Fe(II)/100 mM EDTA solution, 0.84% H2O2, and 20 mM sodium ascorbate to final concentrations of 50 uM Fe(II):100 uM EDTA, 0.03% hydrogen peroxide, and 1 mM ascorbate (12). (To add the reagents, place a drop of each of the stock solutions on the side of the tube, and then tap the tubes to mix them with the DNA simultaneously.) Allow the reaction to proceed for 1-2 min at room temperature, then terminate by the addition of an equal volume of stop solution containing 135 mM thiourea (12). (Appropriate cutting conditions must be confirmed so that there is not more than one cut per molecule.) DNA can be purified by phenol:trichloromethane extraction, as in the DNAase I experiment (Protocol 3), then analysed by gel electrophoresis (Protocol 75). Chemical probes which modify specific bases provide information complementary to that obtained with nucleases. Protein-DNA complexes are treated with base-modifying reagents, such as those described for Maxam-Gilbert chemical sequencing (13,14). Dimethyl sulfate (DMS), which modifies purine bases, has been most widely used in protection and interference assays (e.g. 14-15) (see Protocols 4 and 5 and in vivo footprinting section, below). Depending on the method of cleavage after base modification, DMS can detect changes in susceptibility of G (Protocols 4 and 5) or of both G and A residues (13). Reagents that modify other bases have been used, as well, to probe DNA-protein interaction and DNA structure (reviewed in 16, 17). 2.1.5 Interference assays The interference assay is closely related to the protection assay, but the DNA is modified before it is incubated with binding protein so that bases critical for binding can be identified (Figure 1A). (The interference assay cannot be carried out on the genome of a mammalian cell.) Interference assays do have 39

Craig Spiro and Cynthia T. McMurray some advantages over protection assays. Inefficient binding does not hamper an interference experiment, since all of the probe in the slow-migrating complex is protein bound (Figure 1A, step 2). In the protection experiment, on the other hand, low occupancy of a site might result in only small changes in the intensity of bands on the gel, so that changes can be difficult to detect. One problem in the interference experiment is that recovery of the protein-bound and free DNA bands from polyacrylamide may be inefficient. For interference assays, the modification reagent must leave the backbone intact. Protein-DNA complexes are resolved on native gels (Figure 1A, step 2). The protein-bound DNA and free DNA bands are purified, cleaved, and then analysed by gel electrophoresis. Commonly used reagents for interference are DMS, which alkylates purines (Protocol 4; see Section 3.1), and ethylnitrosourea (ENU), which alkylates the backbone phosphates (14). DMS modification followed by cleavage in hot piperidine is G-specific (Protocol 4; Figure 1A, step 5; Figure 5B; ref. 13). A variation on the interference assay is the 'missing nucleoside' experiment (18). The DNA molecule is first treated with the hydroxyl radical to randomly remove nucleosides, then incubated with binding protein. Free DNA and protein-bound DNA are separated by gel retardation and analysed by denaturing gel electrophoresis to identify nucleosides important for binding (18). Protocol 4. Alkylation by DMS (and G sequence reaction)" Equipment and reagents • 200 ml of 5 M NaOH in 600 ml (or larger) beaker for DMS waste . DMS (Fluka) . TE: 10 mM Tris-HCI pH 8, 1 mM EDTA . Piperidine (Fluka)

• DMS stop buffer: 1 M B-mercaptoethanol, 1.5 M sodium acetate, 200 ug/ml yeast tRNA • Lid locks for microcentrifuge tubes (Fisher Scientific)

Method 1. Prepare double-stranded DNA molecule 32P-end-labelled on one strand only (see text; Figure 2; Protocol 1). (NB these samples are radioactive and must be handled appropriately.) 2. Bring 2 x 106to 1 x 107 c.p.m. of 32P-end-labelled DNA molecule to 175 ul in TE. 3. Immediately before using, prepare 1.5% DMS in TE (1.5 ul DMS + 98.5 ul TE) and vortex. Use fume hood for all manipulations of DMS. Decontaminate pipette tip by pipetting 5 M NaOH and then discard to solid DMS waste beaker in hood.a 4. Add 25 ul freshly prepared 1.5% DMS to DNA sample (step 2). Vortex. (Decontaminate pipette tips.) 5. Incubate 3 min at room temperature.b 40

2: Footprint analysis of DNA-protein complexes 6. 7. 8. 9.

Stop the reaction by adding 50 ul DMS stop buffer. Add 750 ul ethanol and vortex. Freeze 10 min in dry ice/ethanol bath. Pellet DNA by spinning 10 min at full speed in microcentrifuge. Discard supernatant to radioactive waste. 10. Wash with 1 ml 70% ethanol. Vortex. Spin at full speed in microcentrifuge and discard supernatant to radioactive waste. 11. Resuspend the sample in 50 ul water. (Sample should be at least 15000c.p.m. per ul.) 12. Molecules (step 11) are ready for use in interference assays (Protocol 5). Cleave a portion of the sample with piperidine for use as G sequencing standard and to check extent of modification. (Continue with step 13 for cleavage and analysis.) 13. Mix 9 ul methylated DNA (step 11) with 91 ul 11% piperidine (prepared fresh by dilution with water). (Manipulations with piperidine must be in properly functioning fume hood.) Put lid locks on tubes. Vortex samples. 14. Heat 30 min at 90°C. (If lid locks were not put on tubes, cover tubes with weights so volatile piperidine does not force lids to pop open.) 15. Spin tubes briefly to collect any condensation and then freeze in dry ice. 16. Evaporate samples to dryness in SpeedVac concentrator (at least 1 h). 17. Dissolve precipitate in 100 ul H2O and evaporate to dryness in SpeedVac. 18. Dissolve precipitate in 50 ul H20 and evaporate to dryness in SpeedVac. 19. Resuspend pellet in 12 ul TE (vortex, spin, vortex, spin) and use Cerenkov scintillation counting to determine c.p.m./ul. Add 8 ul sequencing stop/load dye (Protocol 3) and mix. Samples can be stored at -20°C. 20. For analysis on sequencing gel (Protocol 75), denature samples for 5 min at 90°C and then put on ice. Estimate extent of modification by electrophoresis of unmodified DNA, modified, uncleaved DNA, and DMS-modified, piperidine-cleaved DNA. To be sure of no more than one modified base per molecule, half the molecules should not be cleaved. a Dimethyl sulfate (DMS) is extremely hazardous, but it can be used safely and effectively if handled properly. Always wear protective clothing and use a fume hood with satisfactory air flow when handling DMS. DMS is very unstable in water, ethanol, p-mercaptoethanol, and sodium hydroxide. Therefore, DMS is effectively inactivated during the procedures described. Use a 600 ml (or larger) beaker containing 200 ml 5 M NaOH for disposal of DMS waste. Decontaminate pipettes and pipette tips by pipetting NaOH up and down a couple of times. 6 These conditions have worked well for molecules of different sizes and for different amounts of DNA, but check cleaved molecules by gel electrophoresis to confirm that molecules are modified no more than once each (step 20).

41

Craig Spiro and Cynthia T. McMurray Protocol 5.

DMS Interference

Equipment and reagents Elution buffer: 0.20 M NaCI, 1 mM EDTA, 20 (ig/ml yeast tRNA, 10 mM Tris-HCI pH 8 Piperidine (Fluka)

Spin filtration device (e.g. 0.22 (um Spin-X; Corning/Costar)

Method 1. Prepare 5% polyacrylamide native gel for resolving naked DNA and protein-bound DNA (Figure 1A, step 2).a 2. Incubate 1 X 105 to 2 X 105 c.p.m. methylated probe (Protocol 4; step 11) with sequence-specific DNA-binding protein in 50-100 ul (or in the appropriate volume as determined).a Samples are radioactive, so handle accordingly. 3. After appropriate binding period for the complex being studied, load the binding reaction on to the native gel (step 1), using 1-3 lanes, as necessary. 4. Separate by electrophoresis. 5. Identify and recover the free DNA (fast band) and protein-bound DNA (retarded band). (a) Remove one of the glass plates from the gel. (b) Wrap the gel with cling film and mark with fluorescent paint to allow unambiguous orientation. (c) Expose X-ray film to the wet gel in a light-proof container for about 4 h. (d) Develop X-ray film. (e) Cut spots corresponding to DNA-protein complex and naked DNA from the film. (f) Align film (using spots from fluorescent paint) with gel, and use film as template to cut bands from gel. 6. DNA molecules up to 300 bp can be recovered by crushing and soaking. (Longer molecules should be recovered by electroelution.) (a) Place the band into a microcentrifuge tube and crush by using a pipette tip to roll and press the gel slice against the sides of the tube. (b) Incubate in 0.5 ml elution buffer at 37°C with shaking for 12-16 h. (c) Spin the tube for at least 10 min at full speed in a microcentrifuge to pellet gel fragments as much as possible. (d) Transfer the supernatant to a spin filtration device (e.g. 0.22 um Spin-X; Corning/Costar), and centrifuge 2 min to collect solution containing DNA and to remove gel pieces. 42

2: Footprint analysis of DNA-protein complexes 7. Precipitate DNA by addition of 1 ml ethanol. Keep at -20°C at least 1 h. 8. Spin at least 15 min at full speed in microcentrifuge and discard supernatant. 9. Wash pellet with 1 ml 70% ethanol. Spin at full speed in microcentrifuge. Remove as much of the supernatant as possible. 10. Resuspend pellet in 100 ul 10% piperidine (prepared fresh by dilution with water) (NB all manipulations with piperidine should be in hood). Put lid locks on tubes. Vortex and leave at room temperature 15 min to allow DNA to dissolve. Vortex. Spin briefly to collect. 11. Heat 30 min at 90°C. (If lid locks not put on tubes, cover tubes with weights so volatile piperidine does not force lids to pop open.) 12. Spin tubes briefly to collect condensation and then freeze in dry ice. 13. Evaporate to dryness in SpeedVac concentrator (at least 1 h). 14. Dissolve in 100 ul H20 and evaporate to dryness in SpeedVac. 15. Dissolve in 50 ul H2O and evaporate to dryness in SpeedVac. 16. Resuspend pellet in 6 ul TE (vortex, spin, vortex, spin to resuspend) and use Cerenkov scintillation counting to determine c.p.m./ul. Add 4 ul sequencing stop/load dye (Protocol 3). 17. Analyse DNA from protein-bound and naked DNA bands by denaturing gel electrophoresis (Protocol 75). (Denature samples for 5 min at 90°C and then put on ice prior to loading on gel.) It is important to load the same number of c.p.m. per lane for comparison of proteinbound and naked DNA (try to load at least 5000 c.p.m. per lane.) aThe first several steps are essentially a gel shift experiment scaled up 5-10 times. Appropriate gel and binding conditions (including salt concentration and amount of non-specific DNA) can first be determined using smaller amounts of DNA and protein.

2.2 Analysis of binding on closed, circular plasmid Thermal cyclers and thermostable DNA polymerases allow a simple, direct approach to footprinting (Figure 3). Complexes of protein with closed, circular plasmids or with linear DNA fragments can be analysed by primer extension. Analysis of larger DNA molecules can provide information different from that obtained on shorter fragments, including information on interaction of sites (9, 19). The only special reagents required are sequencing primers that flank the region to be studied (Figure 3A). Test the primers in thermal cycle sequencing reactions (see Protocol 6). Then the region can be analysed to determine the site of and the effects of bound protein (Protocol 7; Figure 3). Reporter plasmids containing fragments of transcriptional control regions are ideal for this analysis. Typically, control regions in such plasmids have 43

Craig Spiro and Cynthia T. McMurray been modified by point mutation and deletion to study the effect of mutations on transcription. These mutant plasmids can be used to address questions about protein-DNA interaction during transcription, including the effect of mutation on the structure and interaction of DNA binding sites (9,19). After incubating protein with DNA, treat the reaction with DNAase I. Then purify the DNA and use the DNA as template for extension of 32P-end-labelled primers (Protocol 7; Figure 3). A single primer was used to analyse a region containing three CREs (cAMP responsive enhancers) in the rat prodynorphin gene (Figure 3B, C). Primers flanking other regions (such as the cap site) can be used to analyse the same DNA samples. While the requirements for the sequencing primers are not rigorous, both primers should have the same or similar Tm. [Tm = 81.5°C + 16.6 log M + 0.41 (%GC) - (500/n) - 0.61(%formamide), where M is the molar salt concentration, %GC is the percentage of Gs + Cs in the primer, and n is the number of bases in the primer (ref. 21).] Primers used for ligation-mediated PCR (described below) are suitable for sequencing and primer extension in these protocols as well. Primers with Tm > 50 consistently work well. Protocol 6. Thermal cycle sequencing of region of interest Equipment and reagents • 10 x Circum vent buffer: 100 mM KCI, 100 mM (NH4)2S04, 200 mM Tris-HCI pH 8.8 at 25°C, 50 mM MgS04 (ref. 22); store in aliquots at -20°C . 3% Triton X-100 • Deep Vent (exo-l DNA polymerase (New England Biolabs, 2 U/ul) • dNTP/ddNTP mixes (prepared in 1 x circumvent buffer):a A: 900 uM dideoxy ATP, 30 uM dATP, 100 uM each dCTP, dGTP, TTP

C: 250 (uM dideoxy CTP, 30 uM dATP, 37 (uM dCTP, 100 uM each dGTP, TTP G: 250 uM dideoxy GTP, 30 uM dATP, 37 (uM dGTP, 100 uM each dCTP, TTP T: 600 uM dideoxy TTP, 30 uM dATP, 37 uM TTP, 100 (uM each dCTP, dGTP . Template-primer mix (for each primer to be used): 1 ul plasmid template (10-100 ng/ul), 2 ul 32P-end-labelled primer (Protocol 7), 1.5 ul 10 x circum vent buffer, 1 ul 3% Triton X-100, 8.5 ul H2O"

Method 1. End label primers (Protocol 1). 2. Prepare thermal cycler by heating block to 95°C. Sequencing reactions will require 2 min denaturation at 95°C, followed by 20 cycles of 30 sec at 95°C, 15 sec at calculated Tm (see text) for the primer, and 40 sec at 72°C; cool to 4°C.C 3. For each primer, label a set of 0.5 ml tubes A, C, G, T, and to each tube add 3 ul of the appropriate dNTP/ddNTP mix. Keep tubes on ice. 4. Prepare template-primer mix for each primer to be used.b Stir with pipette tip and put on ice. 5. To template-primer mix, add 1 (ul Deep Vent (exo-) DNA polymerase (2 U/ul, New England Biolabs) and mix by gentle stirring or pipetting. 44

2: Footprint analysis of DNA-protein complexes 6. Put 3.2 (ul of the template-primer-polymerase mix into each of the four dNTP/ddNTP-containing tubes (on ice) and cover with 1 drop of mineral oil. 7. Put reactions into hot block of thermal cycler and start programme (see step 2).c 8. To completed reactions add 4 pl sequencing stop/load dye (Protocol 3) under the oil. 9. Spin tubes briefly. Samples can be stored at -20°C. 10. Prior to analysis on sequencing gel (Protocol 75), heat samples for 5 min at 90°C, then put on ice. a

These nucleotide concentrations are suitable for Deep Vent (exo-) DNA polymerase (New England Biolabs). Adjust as appropriate if another DNA polymerase is used. 6 For templates with very high GC content, use 1-2 p.1 formamide and reduce the amount of water. "Shorter extension and annealing steps or a two-step (denaturation, extension) programme may also work well (even for primers with calculated Tm as low as 50 °C).

Protocol 7.

DNAase I footprint on plasmid

Equipment and reagents Siliconized microcentrifuge tubes (0.5 ml, 1.5 ml) (Fisher Scientific) DNAase I reagents (Mg2+/Ca2+ supplement; DNAase I stop solution; DNAase I dilution buffer; see Protocol 3) DNAase I dilution to 0.013 U/ul: 2 (ul DNAase I (RQ1, Promega; 1 U/ul) + 148 ul DNAase I dilution buffer (Protocol 31" 2.5 mM dNTP mix: 2.5 mM each dATP, dCTP, dGTP, TTP

Non-specific DNA (e.g. sonicated salmon sperm DNA)b BSA (purified, not acetylated; New England Biolabs)* Phenol:trichloromethane: iso-arnyl alcohol (25:24:1) Deep Vent (exo-) DNA polymerase (New England Biolabs)

Method 1. Label primers (Protocol 7). 2. Check primer extension reactions and prepare sequence standard (Protocol 6).c 3. Set up binding reactions in total of 40 PLul(in0.5 ml siliconized tube) under conditions suited to protein-DNA interaction being studiedb [e.g. prodynorphin enhancer-CREB (cAMP responsive enhancer binding protein) binding (Figure 36-C): 100 ng (30 fmol) CAT reporter plasmid, sequence-specific DNA binding protein (0, 0.05, 0.2, 0.6, 1.2 uM), 1 ug sonicated salmon sperm DNA, 20 (ug BSA (not acetylated), 30 mM KCI, 4% glycerol, 0.5 mM DTT, 10 mM Tris-HCI pH 7.5). Bind 10 min to overnight at 4°C to 37°C as appropriate for protein-DNA complex being studied. 45

Craig Spiro and Cynthia T. McMurray Protocol 7. Continued 4. Prepare reagents for DNAase I analysis (Protocol 3) and tube for enzyme dilution. Dilute DNAase I immediately before use and keep on ice. 5. Digest with DNAase I (see Protocol 3, step 5, for suggestion on processing samples). (a) Equilibrate 2 min at room temperature. (b) Add 5 ul Mg2+/Ca2+ supplement. ( c ) A d d 5 ( u l diluted DNAase I a n d stir with pipette (e) Stop with 50 ul DNAase I stop solution. 6. Extract with 100 ul phenol:trichloromethane:iso-amyl alcohol (25:24:1). Vortex well and spin in microcentrifugeto separate phases. 7. Transfer aqueous to 1.5 ml siliconized microcentrifuge tube and precipitate the DNA with 260 ul ethanol. Keep at -20°C at least 1 h. 8. Spin sample at least 20 min at full speed in microcentrifuge to pellet the DNA. 9. Aspirate all the supernatant and wash with 185 ul 70% ethanol. Vortex and spin at full speed in microcentrifuge. Then aspirate the supernatant. 10. Dry the pellet in SpeedVac concentrator. 11. Resuspend the pellet in 12 ul 1 x circum vent buffer (Protocol 6). Let the pellet dissolve on ice (at least 1 h). 12. Prepare thermal cycler for primer extension by heating block to 95°C. Primer extension will require 2 min at 95°C, followed by 7 cycles of 30 sec at 95°C, 15 sec at calculated Tm for the primer, and 40 sec at 72°C.C 13. Aliquot 3 ul DNA (step 11) to duplicate sets of tubes (on ice) for analysis by primers flanking the region of interest (e.g. Figure 3B, C, dynfp5' and dynfp3'). Store remaining DNA at 4°C. The remaining DNA can be used to analyse other areas of the plasmid with different sets of primers. 14. Prepare master mixes (on ice) for each primer to be used in the analysis (e.g. Figure 36, one mix for dynfp5' primer and another for dynfp3' primer). In the primer extension master mix include, for each DNA sample, 0.5 (ul32P-end-labelled primer, 0.32 ul 10 x circum vent buffer (Protocol 6), 0.2 .ul 3% Triton X-100, 0.3 ul 2.5 mM dNTP mix, 1.58 ul H2O. Mix gently and chill on ice. For regions of very high GC content include 0.25-0.5 JJL! formamide and reduce amount of water. 46

2: Footprint analysis of DNA-protein complexes 15. Add to the master mix (for each sample) 0.3 ul Deep Vent (exo-) DMA polymerase and stir. Keep mix on ice. 16. Add 3.2 pj appropriate master mix (step 15) to each DNA sample on ice and cover samples with one drop of mineral oil. 17. Put samples in 95°C block of thermal cycler and run programme (step 12).c 18. Add 4 ul sequencing stop/load dye (Protocol 3) under the oil. 19. Spin samples briefly. Samples can be stored at -20°C. 20. Before analysis on sequencing gel (Protocol 15), heat samples to 90°C for 5 min then cool on ice. Use sequence reactions (Protocol 6) to identify bases on the gel (e.g. 'G' lane in Figure 3d. a

ln the presence of certain protein extracts, the correct amount of DNAase I could be 20 x as much and must be empirically determined. 6 Excess non-specific DNA (0.4-2.0 ug sonicated salmon sperm DNA) keeps constant the substrate for DNAase I and may reduce excess nicking outside the binding site, but it is not essential and at very high levels can compete for protein binding. BSA can also be included (2-20 M-g)c Primer extension conditions are the same as the conditions that yield good sequence ladders (Protocol SI but with fewer cycles.

3. In vivo footprinting Ligation-mediated polymerase chain reaction (LMPCR) uses (i) nested genespecific primers for specificity and (ii) ligation of a linker to allow amplification of fragments of genomic DNA (3, 23) (Figure 4C). The procedure is specific and sensitive enough for nucleotide-resolution analysis of DNA structure and protein binding within single-copy genes of mammalian cells. Whole genome analysis and genomic footprinting were previously hindered by the complexity and methodological difficulty of the steps (15, 24; reviewed in 25). In contrast, while LMPCR requires many steps, the individual steps are not technically difficult. This section describes the steps necessary for in vivo footprinting: modification of the DNA in vivo, purification of the DNA, and LMPCR (including design of the gene-specific primers, preparation of the unidirectional linker, and amplification of gene-specific fragments). LMPCR allows visualization of each nucleotide (genomic sequence), making possible the direct comparison of naked DNA and DNA in vivo (Figure 5B). In vivo footprinting is a variation of the protection experiment (Figure 1B; Figure 4), in which DNA is modified in vivo, most often with DMS (Protocols 8 and 9}. (Alternatives may also be used, as discussed below.) Genomic DNA is purified and cleaved to generate a collection of fragments (Figure 4A) (Protocols 10 and 12). Gene-specific primers (Figure 4B; Figure 5A) are used to amplify fragments and to label the fragments for analysis on sequencing gels (Protocol 15) (Figure 4C; Figure 5B). 47

Craig Spiro and Cynthia T. McMurmy • Primer 1 is extended to create blunt-ended molecules (Figure 4C, step 2). • A universal linker is ligated to the blunt-ended fragments (Figure 4C, step 3). • PCR with the linker primer and primer 2 amplifies the fragments (Figure 4C, steps 4-6). • The fragments are 32P-end-labelled (primer 3) for visualization by denaturing gel electrophoresis (Figure 4C, step 7). [Footprints may also be visualized by electrophoresis of the unlabelled amplified DNA, electrotransfer to nylon membrane, and hybridization with radiolabelled probe (23, 26, 27)]. Modification of DNA in cultured mammalian cells by DMS is a simple procedure (Protocols 8 and 9), and careful application of standard protocols will provide high-quality DNA (Protocol 10; other procedures, including kits such as G Nome, from Bio 101, can also produce satisfactory DNA). Thus, the source of greatest variability and uncertainty for generating genomic sequence is the design of gene-specific primers (Figure 4B; Figure 5A, and see text, below). Once the DNA is prepared, LMPCR comprises reactions of gene-specific primers with the purified DNA (Figure 4B, 4C; Figure 5). Use of high quality reagents and proper laboratory technique in applying the protocols will generate the genomic sequence (Figure 4; Figure 5). The procedure described for LMPCR (Protocol 14) is adapted from modifications of the original procedure (3) that use Vent DNA polymerase exclusively (28, 29). In addition to simplifying the procedure, LMPCR using Vent DNA polymerase allows direct analysis of DNA with 3'-hydroxyl ends (e.g. DNAase I-cut DNA) (28).

3.1 Modification of DNA in vivo DMS is a small molecule that can freely enter living cells. It does not disrupt protein-DNA complexes and its pattern of modification of bases is well understood: DMS is a very reactive alkylating agent that methylates the N7 of guanines (which is in the major groove of DNA) and to a lesser extent the N3 of adenine (which is in the minor groove) (13, 25, 30). N3 of cytosine and Nl of adenine, both potential sites for reaction, are protected from modification by hydrogen bonds of double-stranded DNA. Cleavage of DMS-modified DNA with hot piperidine (Protocols 8, 9, 11, 12} is G-specific (13, 31). While DMS is an effective probe for in vivo analysis, it is not perfect. DNA interaction with proteins that bind weakly, that bind within the minor groove, that interact with the backbone of DNA, or that bind to sites lacking Gs will be poorly detected by DMS modification. Modified cleavage chemistry allows detection of both Gs and As after DMS modification (32). Other reagents can provide supplementary information (reviewed in refs 25, 30). Reagents used for in vivo modification of DNA must have access to the nucleus and they must be reactive enough to modify the DNA efficiently. The reagent-modified DNA must be compatible with LMPCR. Both potassium 48

2: Footprint analysis of DNA-protein complexes permanganate (33) and UV irradiation (34) have been used to modify the DNA of intact cells for subsequent analysis by LMPCR. Enzymes such as DNAase I and micrococcal nuclease can be used with LMPCR, but they cannot freely enter cells. The enzymes can be introduced into cells after treatment with lysolecithin (33, 35), or the enzymes can be used to treat isolated nuclei (36, 37). Enzymes may also be introduced into intact nuclei attached to tissue culture dishes. Treatment of monolayers of cultured cells with NP40 will remove the plasma membrane, leaving intact nuclei on the dish and accessible to enzymes (38). • Treat monolayers for 10 sec with 1% NP40, 200 mM sucrose, 40 mM NaCl, 5 mM MgCl2, 5 mM EGTA, 30 mM Hepes pH 7.9. • Aspirate NP40-containing buffer. • Wash dish with 20 ml Hepes-buffered saline (150 mM NaCl) supplemented with 2.5 mM CaCl2,1 mM MgCl2. (Inspect the nuclei by light microscopy.) • Digest nuclei with enzyme (e.g. DNAase I), stop reaction with buffer containing SDS and EDTA, and proceed with DNA purification (Protocol JO). DNAase I-cleaved DNA can be used directly for LMPCR with Vent DNA polymerase (28). Micrococcal nuclease leaves 5'-OH groups, so the DNA must first be 5'-phosphorylated (by kinase reaction with T4 polynucleotide kinase and ATP) before proceeding with LMPCR (Protocol 14). Characteristics of reagents for in vivo footprinting have been reviewed (25, 30). Protocol 8. DMS treatment of monolayers of mammalian cells Equipment and reagents Cold PBS Lysis solution: 300 mM NaCI, 10 mM EDTA, 0.2% SDS, 50 mM Tris-HCI pH 8

> DMS(Fluka) . 5 M NaOH for DMS waste (200 ml in 600 ml or larger beaker).

Method 1. Grow cells from 1/2 to nearly confluent (or as suitable for maintaining health of cells) on one or two 15 cm dishes per treatment. Include 1-2 dishes for in vitro DMS treatment ('naked DNA' control). A minimum of 2 x 107 mammalian cells should yield an adequate amount of DNA. 2. Treat with inducers of transcription, as appropriate. (NB Carry out all manipulations with DMS in fume hood.) 3. Bring dishes of cells to fume hood and wash with PBS. 4. Prepare 0.1% DMS immediately before use by dilution into PBS, followed by vortex mixing in fume hood (e.g. add 25 (ul DMS to 25 ml PBS and vortex). Decontaminate pipette tip by taking up and expelling NaOH. 49

Craig Spiro and Cynthia T. McMurray Protocol 8.

Continued

5. Pipette 5 ml 0.1% DMS on to each dish for in vivo modification (decontaminate pipettes by taking up and expelling NaOH). Remember to treat some dishes with PBS alone for naked DMA control. 6. Leave DMS on dish for 10 min at room tempera ture.a 7. Pour DMS from dishes into DMS waste beaker. 8. Wash twice with 8 ml cold PBS (for each wash, allow PBS to cover dish and then pour into DMS waste beaker). 9. Use pipette to aspirate last of second PBS wash and dispose into NaOH/DMS waste beaker. 10. To each dish add 1.5 ml lysis solution and rock to let lysis solution cover dish. 11. Collect viscous lysate with help of rubber policeman and transfer to 15 ml polypropylene conical centrifuge tube (viscous lysate should drop into tube). Purify DNA (Protocol 10). "While these conditions have worked well for several cell types and genes, they may not be appropriate for all. Check size of DNA fragments after cleavage (Protocol 12). Modify if necessary. Incubation time beyond 10 min will not increase modification because of hydrolysis of DMS.

Protocol 9.

DMS treatment of suspension cells

Equipment and reagents Cold PBS Lysis solution: 300 mM NaCI, 10 mM EDTA, 0.2% SDS, 50 mM Tris-HCI pH 8

DMS (Fluka) 5 M NaOH for DMS waste (200 ml in 600 m or larger beaker)

Method 1. Treat cells with inducers of transcription as appropriate. Harvest and wash (3 X 107 to 5 X 107) cells with PBS and then resuspend in 0.5 ml PBS in 15 ml conical polypropylene centrifuge tube (include PBS-only control, which will not be modified with DMS in vivo). 2. Bring cells to fume hood and prepare 0.2% DMS immediately before use (e.g. add 20 p,l DMS to 10 ml PBS and vortex). Decontaminate pipette tip by pipetting NaOH. 3. Add 0.5 ml 0.2% DMS to 0.5 ml cells in PBS (step 1) (include PBS-only control). Tap and swirl tube to mix. 4.Incubate5minatroomtemperature.a 5. Add 9 ml ice-cold PBS. Spin 5 min at 500 x gto collect cells. Discard supernatant to DMS waste. 50

2: Footprint analysis of DNA-protein complexes 6. Wash pellet with 10 ml ice-cold PBS. Spin 5 min at 500 x gto collect cells. Discard supernatant to DMS waste. 7. Resuspend pellet in 1.5 ml lysis solution and purify DNA (Protocol 10). ' Check size of DNA fragments after cleavage (Protocol 72). Modify if necessary.

Protocol 10. Purification of genomic DNA Equipment and reagents • • . •

10 mg/ml RNAase (DNAase-free) 20 mg/ml proteinase K Trichloromethane:iso-amyl alcohol (24:1) Phenol equilibrated to pH > 7.6

• Phenol:trichloromethane:iso-amyl (25:24:1) * TE (Protocol 4}

alcohol

Method 1. To cells in 1.5 ml lysis solution (in 15 ml tubes) (Protocols 8 and 9) add 10 ul 10 mg/ml RNAase (DNAase-free). 2. Rock 20 min at room temperature. 3. Add 45 ul 0.5 M EDTA and 15 ul 20 mg/ml proteinase K. Incubate at 30°C for at least 2 h. (Extensive incubation at 37°C or higher may cause depurination of alkylated DNA.) 4. Lysate (step 3) should be clear. If it has not cleared by 2 h, supplement with 15 ul 20 mg/ml proteinase K, swirl tubes, and continue incubation for another 2 h. Continue incubation until lysate clears. The DNA is not in danger during this incubation, since EDTA, SDS, and proteinase K all inhibit DNAase activity. 5. When lysate has cleared, begin extractions (for solvents, see steps 6-10). Perform extractions (at room temperature) as follows: (a) Add indicated organic solvent, and mix by gentle inversion (30 X) of tubes (do not vortex). (b) Spin 1500 X g for 10 min to resolve phases. (c) Remove the lower (organic) layer using a glass Pasteur pipette. To remove the lower layer insert the pipette to the interface; if aqueous (DNA-containing) phase starts to enter the organic layer, force out a few bubbles to help get through to the organic layer. Leave the interface behind (err on the side of leaving behind some of the organic layer, since the viscous DNA can be at the interface and subsequent extractions will remove traces of organic solvent). 6. Extract with 2.0 ml phenol as described in step 5 (do not vortex samples). After separating the phases from first phenol extraction,

51

Craig Spiro and Cynthia T. McMurray Protocol 10.

Continued

the interface is typically fairly large and the aqueous layer is cloudy. The aqueous will clear and the interface sharpen with subsequent extractions. 7. Extract a second time with 2.0 ml phenol (as described in step 5). Aqueous phase should be clearer. 8. Extract twice with 2.0 ml phenol:trichloromethane:iso-amyl alcohol (25:24:1) (as described in step 5). If the aqueous phase is not clear extract a third time with phenol:trichloromethane:iso-amyl alcohol, otherwise extract with trichloromethane (step 9). 9. Extract with 2.0 ml trichloromethane:iso-amyl alcohol (as described in step 5). Remove all of lower (organic) phase (interface between trichloromethane and aqueous layer will be sharp). 10. For each sample, prepare a 15 ml tube containing 3 ml TE. 11. To aqueous phase (step 9), add 3 volumes cold ethanol. Invert the tube (at least 20 times); stringy, white precipitate will appear (if the precipitated DNA is not visible, collect by centrifugation). Do not let the DNA dry during the following purification procedure. Large genomic DNA will be very difficult or impossible to resuspend if it dries. 12. Remove stringy DNA by spooling around a plastic pipette tip. Carefully lift and transfer to tube containing TE (step 10). The DNA will be difficult to see on the pipette tip when it is lifted out of the liquid. It should come off the tip easily in TE. If it does not come right off, just turn the tip to let the DNA unwind. 13. The DNA (step 12) may dissolve right away, but it may be slow to dissolve. Rock slowly at room temperature for 2-4 h. 14. Prepare tube containing 1.5 ml TE. 15. To samples (step 13), add 330 ul 3 M sodium acetate, and rock to mix. Then add 6.7 ml ethanol and invert the tube (as in step 11) to precipitate the DNA. Spool the DNA (as in step 12) and transfer to tube containing 1.5 ml TE (step 14). Rock at room temperature several hours to dissolve. 16. The samples (step 15) will generally dissolve. If a sample does not dissolve, cut off the end of a 1 ml pipette tip and use that tip to pipette the DNA up and down several times. Rock at room temperature. 17. In vivo DMS-treated DNA must be cleaved by piperidine (Protocol 12). To prepare a portion for cleavage (scale up as needed): (a) Transfer 360 ul DNA (step 16) to 1.5 ml siliconized microcentrifuge tube. (b) Add 40 ul 3 M sodium acetate. 52

2: Footprint analysis of DNA-protein complexes (c) Precipitate with 1 ml ethanol. Collect by centrifugation. (d) Wash pellet with 70% ethanol. (For piperidine cleavage see Protocol 72). 18. DNA from control dishes must be modified in vitro for naked control DNA (see Protocol 11).

Protocol 11. In vitro modification with DMS for naked genomic DNA control (G sequence reaction) Equipment and reagents • DMS stop buffer: 1 M p-mercaptoethanol, 100 ug/ml yeast tRNA, 1.5 M sodium acetate pH 7.

• 5 M NaOH for DMS waste

Method 1. Resuspend 75-175 ug purified genomic DNA from control, PBStreated cells, (Protocol 10) in 175 ul TE in a 1.5 ml tube. (This DNA will be viscous. Cut off end of disposable pipette tip and pipette up and down to help resuspend.) 2. Prepare fresh 1.5% DMS immediately before use (add 3 ul DMS to 197 uJ TE and vortex). Decontaminate pipette tips with NaOH. 3. Add 25 ul 1.5% DMS to 175 ul DNA and vortex. 4. Incubate 2 min at roomtemperature.a 5. Stop with 50 ul DMS stop buffer. 6. Add 750 ul ethanol to precipitate. 7. Put tube in dry ice/ethanol bath for 10 min. 8. Spin tube 10 min at top speed in microcentrifuge. 10. Aspirate supernatant to NaOH/DMS waste. (Remove as much as possible.) 11. Wash pellet with 0.9 ml 70% ethanol (dislodge pellet). 12. For piperidine cleavage of modified DNA, see Protocol 12. "These conditions have worked well for a variety of samples, but they may have to be adjusted to allow direct comparison between in vivo and in vitro alkylated DNA.

Protocol 12. Cleavage of modified DNA with piperidine and preparation for LMPCR Equipment and reagents • Piperidine (Fluka) « TE (Protocol 4)

• 3 M sodium acetate pH 7 • 8 M ammonium acetate

53

Craig Spiro and Cynthia T. McMurray Protocol 12.

Continued

Method NB Transfer piperidine in fume hood. Piperidine is volatile, toxic, and may cause burns. 1. Spin nucleic acid (in 70% alcohol wash; see Protocols 10 and 11) at least 5 min in microcentrifuge to pellet. 2. Aspirate all of the supernatant, but do not allow pellet to dry. 3. Resuspend pellet in 200 ul 10% piperidine (diluted with water immediately before use). Transfer piperidine in fume hood. 4. Put lid locks on tubes. 5. Let pellet dissolve about 15 min at room temperature with intermittent brief vortexing. Make sure pellet is dissolved (it will be clear, so check carefully for floating pellet after vortexing). 6. Incubate at 90°C for 30 min (weights may be put on tubes if lid locks are not available to keep volatile piperidine from forcing lids open). Note: DNA will be cleaved at modified residues by this treatment; RNA will be hydrolysed. 7. Spin tubes briefly to collect precipitate. 8. Place tubes into powdered dry ice or dry ice/ethanol bath to freeze. 9. Evaporate to dryness in SpeedVac concentrator (at least 2 h). 10. Resuspend pellet in 360 ul TE plus 40 ul 3 M sodium acetate. This pellet will generally dissolve easily. If it does not, leave at room temperature for around 30 min and then proceed. 11. Add 1 ml ethanol, mix, and put at -20°C for about 1 h. 12. Spin at least 15 min at full speed in microcentrifuge to pellet DNA. 13. Aspirate all of the supernatant. 14. Resuspend pellet in 500 (ul TE plus 170 ul 8 M ammonium acetate. This pellet should dissolve rapidly. 15. 16. 17. 18.

Add 670 ul iso-propanol to precipitate. Put at -20°C for about 1 h. Spin at least 15 min at full speed in microcentrifuge to pellet. Aspirate supernatant with care (pellets may be loose). Wash once with 180 p.1 of 70% ethanol. Vortex. Spin at least 5 min at full speed in microcentrifuge. Aspirate all of the supernatant. 19. Resuspend pellet in 50 ul H2O. Evaporate in SpeedVac concentrator. This is the final step for removal of piperidine. 20. Resuspend pellet in 30 ul TE. 21. When pellet has dissolved, spin 10 min at full speed in microcentrifuge. Transfer the supernatant, which contains the DNA, to a new 1.5 ml microcentrifuge tube. (Discard any pellet.) 54

2: Footprint analysis of DNA-protein complexes 22. Quantitate DMA by measuring A260. Following the piperidine treatment, assume that the DNA is predominantly single stranded and therefore 40 ug/ml for A260 = 123. Dilute the DNA with TE to A260 = 10 (0.4 (ug/ml) for ligation-mediated PCR. 24. Distribution of DNA fragment sizes can be estimated by electrophoresis on denaturing agarose gels (50 mM NaOH). Average size of 600 bases is good.

3.2 Nested gene-specific primers Amplification of specific, rare fragments of DNA requires nested genespecific primers with successively increasing Tm (Figure 4B; Figure 5A). A set of top strand and bottom strand primers should be synthesized, as follows: (a) Primer 1 should be 20 bases long with 12 G + C. (b) Primer 2 should be 25 bases with 15 G + C. (c) Primer 3 should be 28 bases with 17 or 18 G + C (see Figure 4B and figure legend; Figure 5A). Primer 3 (the labelling primer; Figure 4C, step 7), should be at least 30 bases from the 5' end of the region of interest (see Figure 4B\ legend to 4B; Figure 5A). Avoid primers likely to form structures due to complementary ends (e.g. GGG at 5' end and CCC at 3' end). Because some genes will not accommodate the specific prescription for primers, the length and composition for primers 1-3 will need to be adjusted. Maintain the prescribed Tm (see Figure 4B legend) by compensating for changes in GC composition by changes in length.

3.3 Linker for LMPCR A staggered linker is ligated to the blunt-ended molecules created by extension of gene-specific primer 1 (Figure 4C, step 3) (3, 29). The staggered linker should be prepared in advance and can be stored in aliquots at -20°C. Protocol 13. Preparation of staggered, unidirectional linker for ligation-mediated PCR Method 1. Synthesize oligonucleotides LMPCR.1 and LMPCR.2" (ref. 29), using the 0.2 (jimol synthesis (Applied Biosystems). 2. Desalt by gel filtration (e.g. Econo Pac 10DG column, Bio-Rad) and elute with water. 3. Dry in SpeedVac concentrator. Resuspend in 150 p.I TE and determine molar concentration by measuring absorbance at 260 nm (extinction 55

Craig Spiro and Cynthia T. McMurray Protocol 13.

4.

5. 6.

7.

Continued

coefficients for the primers are 240, 500 for LMPCR.1 and 112, 200 for LMPCR.2). Dilute appropriate volumes of LMPCR.1, LMPCR.2, and 1 M Tris-HCI pH 7.7 to achieve these final concentrations: 20 uM LMPCR.1, 20 uM LMPCR.2, and 250 mM Tris-HCI pH 7.7. For example, to prepare 750 ul of unidirectional linker combine 30 ul 500 uM LMPCR.1, 30 ul 500 (uM LMPCR.2,187.5 ul 1 M Tris-HCI pH 7.7, and 502.5 ul H20. Vortex. After mixing well, aliquot 50 ul into 0.5 ml microcentrifuge tubes. Denature for 5 min at 95°C, cool to 70°C over 2 min, and hold at 70°C for 2 min. Then cool to 22°C over 1 h and hold at 22°C for 1 h. Cool to 4°C over 1 h and keep at 4°C for at least 1 h. Spin tubes briefly. Then store at -20°C.

• LMPCR.1: 5'-GCGGTGACCCGGGAGATCTGAATTC-3';LMPCR.2: 5'-GAATTCAGATC-3'.

3.4 Visualization to nucleotide resolution by LMFCR LMPCR allows visualization of rare changes within single-copy genes for high-resolution genomic sequencing. Changes between naked DNA (not modified in vivo) and in v/vo-modified DNA indicate alterations in DNA sensitivity due to protein binding or structural alteration (Figure 5B). DMSpiperidine treatment produces G ladders, so that protein binding and alterations in DNA structure are detected by changes in the pattern of modification of Gs (Figure 5B). Changes in charge distribution may cause appearance of certain As in this protocol due to alkylation of N7 of adenine (Figure 5B; ref. 9). Amplification with top strand primers yields the sequence of the bottom strand (Figure 5). Amplification through GC-rich regions may be hampered by secondary structure formation. If the pattern on the sequencing gel shows bias toward shorter fragments, include deaza-dGTP in nucleotide mix in place of some of the dGTP (see Protocol 14; recipe for 2.5 mM dNTP mix with deaza-dGTP). If the signal appears weak, reduce the amount of DNA from 2 n_g to 1 ug. If that does not solve the problem, modify the number of cycles of amplification. The labelling step (two cycles) should not be increased. Protocol 14. LMPCR Equipment and reagents 2.5 mM dNTP mix: 2.5 mM each dATP, dCTP, dGTP, TTP; store at -20°C Vent DNA polymerase (New England Biolabs, 2 U/M.I)

10 X First strand buffer: 400 mM NaCI, 50 mM MgSO4, 0.1% gelatin, 100 mM Tris-HCI pH 8.9 (at room temperature); store in aliquots at -20°C

56

2: Footprint analysis of DNA-protein complexes . 2.5 mM dNTP with deaza-dGTP mix: 0.625 mM dGTP; 1.875 mM 7-deaza-dGTP (Boehringer-Mannheim), and 2.5 mM each dATP, dCTP, TTP;' store at -20°C • 10 x Amplification buffer: 400 mM NaCI, 50 mM MgS04, 0.1% gelatin, 1% Triton X-100, 200 mM Tris-HCI pH 8.9 (at room temperature); store in aliquots at -20°C . T4 DNA ligase at SU/^I . 100 mM ATP

10 mg/ml BSA (not acetylated; use, e.g. purified DNA, New England Biolabs) 10 mg/ml yeast tRNA 3 M sodium acetate Siliconized 0.5 ml and 1.5 ml microcentrifuge tubes (Fisher Scientific) unidirectional linker (Protocol 13} primer sets (Figure 4B, Section 3.2)

Method Note: Use master mixes in setting up reactions. Put all common components together and then aliquot to sample tubes. Chill master reaction mixes and add enzymes last, as described below. Volumes given in the protocol are for each sample to be included in the master mix. 1. Label two sets of 0.5 ml siliconized tubes. One set is for top primers and the other for bottom primers (Figure 5A). Each set should contain naked DNA control and the in vivo-modified DNA samples. 2. Put 5 ul DNA (0.4 (ug/ul) into a 0.5 ml siliconized tube.b 3. Equilibrate heating block or thermal cycler to 95°C. (First strand synthesis will require denaturation for 5 min, followed by 30 min at 60°C and 10 min extension at 76°C; cool to 4°C.) 4. Prepare two first strand master mixes—one for top primer set and one for bottom primer set. Include in the first strand mix the following for each reaction: 3 uJ 10 X first strand buffer, 0.3 ul 1 uM primer 1 (top or bottom, as appropriate), 2.4 ul 2.5 mM dNTP mix (or 2.5 mM dNTP with deaza-dGTP mix), 19.05 ul H20. Chill. 5. Add to the chilled master mix, for each reaction, 0.25 ul 2U/ul Vent DNA polymerase. Stir the mix with the pipette tip and keep on ice. 6. Add 25 ul master mix to each tube with 5 ul DNA in 0.5 ml tube on ice. 7. Transfer samples from ice to 95°C heat block. Carry out first strand synthesis (see step 3). 8. During first strand synthesis prepare ligation dilution and ligation reaction master mixes. 9. Include in ligation dilution mix, for each sample: 2.2 ul 1 M Tris-HCI pH 7.5, 0.35 ul 1 M MgCI2,1 ul 1 M DTT, 0.25 ul 10 mg/ml BSA (do not use acetylated BSA), 16.2 ul H2O. Chill and keep on ice. 10. Prepare ligation reaction mix containing, for each sample: 0.25 ul 1 M MgCI2, 0.5 ul 1 M DTT, 0.75 ul 100 mM ATP, 0.125 ul 10 mg/ml BSA, 17.775 ul H2O. Mix and chill, then add in order to master mix, for each reaction: 5 ul unidirectional linker (see Protocol 73) and 0.6 ul 5U/(ul T4 DNA ligase. Stir with pipette tip and keep on ice. 57

Craig Spiro and Cynthia T. McMurray Protocol 14.

Continued

11. To each sample tube (step 7} add in order 20 ul ligase dilution mix (step 9), stir with pipette tip, and 25 ul ligation reaction mix (step 10). Then mix by stirring. 12. Incubate at least 6 h at 16°C. 13. Prepare precipitation salt mix containing, for each reaction: 8.4 ul 3 M sodium acetate pH 7 and 1 ul 10 mg/ml yeast tRNA. Keep on ice. 14. Put ligation reactions on ice (after step 12) and add 9.4 ul precipitation salt mix (step 13); stir. Add 220 ul ethanol and vortex. Put samples at -20°C for at least 1 h. 15. Spin samples for 15 min at full speed in microcentrifuge and aspirate all of the supernatant. 16. Wash pellet once with 180 ul of 70% ethanol. Spin at least 5 min at full speed in microcentrifuge. Aspirate all of supernatant. Dry in SpeedVac concentrator. 17. Resuspend pellet in 70 ul H20. Vortex. Let dissolve 30 min at room temperature with intermittent vortexing. Then place samples on ice. Prepare amplification and Vent DNA polymerase mixes during this time. 18. Equilibrate thermal cycler to 95°C. (Amplification programme: 4 min at 95°C, then 20 cycles of 1 min at 95°C, 2 min at 66°C, 3 min at 76°C with an increase of 5 sec per cycle of the extension time at 76°C; then a final additional extension of 5 min 30 sec at 76°C; followed by cooling to 4°C.) 19. Prepare two amplification mixes (one for top and the other for bottom primer set). Include in the amplification mix, for each sample: 10 ul 10 x amplification buffer, 1(1ul10 pM LMPCR.1 (Protocol 73), 1 ul 10 uM primer 2 (top or bottom, as appropriate), 8 ul 2.5 mM dNTP mix (or dNTP with deaza-dGTP mix), 10 ul H20. Stir and chill on ice. 20. Prepare Vent DNA polymerase mix containing, for each sample: 0.3 (xl 10 x amplification buffer, 2.2 ul H2O, and 0.5 ul 2U/ul Vent DNA polymerase. Stir. Keep on ice. 21. To DNA samples on ice (step 17) add 30 ul of appropriate (top or bottom) amplification mix (step 19). 22. Add 3 ul Vent DNA polymerase mix (step 20) to each sample and cover reaction mix with two drops mineral oil. 23. Place samples in thermal cycler (block at 95°C). Run programme (step 18). 24. During amplification, prepare bottom) (Protocol 1).

32

P-end-labelled primer 3 (top and

58

2: Footprint analysis of DNA-protein complexes 25. At conclusion of amplification (step 23), put samples on ice and heat thermal cycler to 95°C. (End labelling will require 3 min at 95°C, followed by 2 cycles of 1 min at 95°C, 2 min at 69°C, 10 min at 76°C, followed by cooling to 4°C). 26. Prepare two end-label master mixes (top and bottom) containing, for each sample: 0.8 (xl 10 x amplification buffer, 2.3 ul end-labelled primer (step 24), 4 ul 2.5 mM dNTP mix (or dNTP with deaza-dGTP mix), 0.4 ul H2O. Stir and chill, then add, for each sample, 0.5 ul 2 U/ul Vent DNA polymerase. 27. Pipette 8 ul appropriate (top or bottom) end-labelling mix under the oil to each sample on ice. 28. Place samples in thermal cycler (block at 95°C). Denature 3 min at 95°C, then extend 32P-labelled primer (see step 25). (One cycle is probably sufficient, and no more than two are needed to label the amplified DNA.) 29. Prepare Vent reaction stop solution. Include for each reaction: 25 ul 3 M sodium acetate pH 7, 2.5 ul 1 M Tris-HCI pH 7.5, 2 ul 0.5 M EDTA, 2 ul 10 mg/ml yeast tRNA, 20.5 ul H2O. 30. To end-labelled samples (step 28) introduce 52 ul Vent reaction stop solution under the oil. 31. Extract with 250 ul phenol:trichloromethane:iso-amyl alcohol (25:24:1). Vortex. Spin at full speed in microcentrifuge. 32. Transfer aqueous to 1.5 ml siliconized tube. Add 2.5 volumes (about 425 ul) ethanol to precipitate. Place samples at -20°C for at least 1 h. 33. Spin samples 15 min at full speed in microcentrifuge. Aspirate the supernatant. 34. Resuspend the sample in 6.2 ul TE, vortex, spin briefly. (You can check that sample is resuspended by taking up 1 ul in pipette and holding in front of Geiger monitor.) 35. Add 4 ul sequence stop/load dye (Protocol 3), vortex, spin briefly. Sample is ready for analysis on sequencing gel (Protocol 15) or may be stored at -20°C. * May be necessary for analysis of GC-rich regions; see text. 'As little as 1 ug DNA can suffice and lowering the amount of DNA may increase the signal.

Protocol 15. Sequencing gel Equipment and reagents Rain-X car windshield treatment Gel lift paper (Schleicher and SchuelD*

For in vitro footprints: 0.4 mm thick, 40 cm long gelc Well-forming combs (recommended)d

59

Craig Spiro and Cynthia T. McMurray Protocol 1.5.

Continued

Notes 1. Estimate proper length of run according to migration of dyes. Bromophenol blue migrates with 26 base fragment in 6% gel and with 19 base fragment in 8% gel; xylene cyanol migrates with 106 base fragment in 6% gel and with 76 base fragment in 8% gel. 2. In the initial run, use only a portion (perhaps one-quarter) of sample in order to allow later adjustment of sample loading for equal signal intensity. Method 1. Treat the inside (gel side) of inner plate with Rain-X or other siloxanecontaining windshield treatment to prevent gel from sticking to glass. Assemble sandwich of plates with spacers. 2. Prepare 7 M urea, 6% (or 8%) acrylamide sequencing gel. 3. Add 10% ammonium persulfate. Pour gel using a syringe. Hold sandwich at about 45° angle at start and pour gel along spacer to avoid trapping bubbles. Lower plates as you pour the gel. 4. Lay gel flat on bench and insert comb. Let gel polymerize at least 45 min. 5. Prepare 1 x Tris-borate-EDTA running buffer. Wet area around comb and carefully remove comb from gel. 6. Set up gel in apparatus. Rinse wells using pipette and upper reservoir buffer. 7. Pre-run gel at 50 W (40 cm gel) or 75 W (62 cm gel) for at least 30 min to allow gel to warm. (Constant power will maintain temperature during the run.) 8. Denature samples (in sequencing stop/load dye (Protocol 3); 90°C for 5 min) then place on ice before loading. 9. Very important: after pre-running gel, rinse wells (using pipette and buffer from upper reservoir) to clear out urea etc. before loading samples. 10. Run denatured samples for appropriate time (at least until bromophenol blue runs off the gel). 11. After electrophoresis, remove gel (remember that lower buffer will be radioactive) and cool under running water. 12. Carefully remove inner (siliconized) plate. Gel should stick to other plate. 13. For gels longer than 40 cm (i.e. too long for gel dryer), put spacer across gel at 40 cm from the bottom of the gel and cut gel by sliding spatula along the spacer. Sections should be peeled separately. Bottom (40 cm) section is the most important. 60

2: Footprint analysis of DNA-protein complexes 14. Press gel lift paper against 40 cm bottom section of gel and apply pressure along entire surface to ensure good contact. For 6% gel, peel the paper with attached gel right from the glass. For 8% gels, turn the paper/gel/glass sandwich over and bring out over edge of bench. Then peel paper with attached gel off the glass while gradually sliding glass with attached gel over edge of bench. 15. Dry gel.

16. Expose film or phosphor storage screen to dried gel.e • Better than 3MM, especially with 8% gels. b This mixture costs little more than the ingredients, reduces handling of acrylamide, and eliminates the need to dissolve urea. C 6% gels are easier to handle than 8% gels, they stick better to paper for drying; but 7% or 8% gels can help resolution of smaller fragments. For in vivo footprints, longer gels are usually required (62 cm should suffice). The amount of information in a sequencing gel can be increased by running electrolyte gradient gels, as described by Sheen and Seed (39). The bottom 40 cm will be most informative. ^Shark's tooth combs will work as long as there are blank wells between samples. (In contrast to sequencing reactions in which neighbouring lanes would have different bands, in a footprint reaction the bands in neighbouring lanes are the same; therefore, space between lanes can be helpful.) " Kodak Biomax MR single emulsion film is recommended as it gives good resolution with low background, but the sensitivity is lower than with double emulsion film, so that exposures will be longer.

References 1. Galas, D. J. and Schmitz, A. (1978). Nucleic Acids Res. 5,3157. 2. Kolchinsky, A. M., Mirzabekov, A. D., Gilbert, W., and Li, L. (1976). Nucleic Acids Res. 3,11-18. 3. Mueller, P. R. and Wold, B. (1989). Science 246,780. 4. Brenowitz, M., Senear, D. F., Shea, M. A., and Ackers, G. K. (1986). In Methods in enzymology (ed. C. H. W. Hirs and S. N. Timasheff), Vol. 130, p. 132. Academic Press, London. 5. Dignam, J. D., Lebovitz, R. M., and Roeder, R. G. (1983). Nucleic Acids Res. 11, 1475. 6. Dyer, R. B. and Herzog, N. K. (1995). Biotechniques 19, 192. 7. Schreiber, E., Matthias, P., Muller, M. M., and Schaffner W. (1989). Nucleic Acids Res. 17,6419. 8. Andrews, N. C. and Faller, D. V. (1991). Nucleic Acids Res. 19, 2499. 9. Spiro, C., Bazett-Jones, D. P., Wu, X., and McMurray, C. T. (1995). J. Biol. Chem. 270, 27702. 10. Sigman, D. S. and Chen, C. B. (1990). Ann. Rev. Biochem. 59, 207. 11. Kuwabara, M. D. and Sigman, D. S. (1987). Biochemistry 26, 7234. 12. Price, M. A. and Tullius, T. D. (1992). Methods in enzymology (ed. D. M. J. Lilley and J. E. Dahlberg), Vol. 212, p. 194. Academic Press, London. 13. Maxam, A. M. and Gilbert, W. (1980). Methods in enzymology (ed. L. Grossman and K. Moldave), Vol. 65, p. 499. Academic Press, London.

61

Craig Spiro and Cynthia T. McMurray 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29.

30. 31. 32. 33. 34. 35. 36. 37. 38. 39.

Siebenlist, U. and Gilbert, W. (1980). Proc. NatlAcad. Sci. USA 77, 122. Church, G. M. and Gilbert, W. (1984). Proc. NatlAcad. Sci. USA 81,1991. Palacek, E. (1991). Crit. Rev. Biochem. Mol. Biol. 26, 151. Hayes, J. J. (1995). Chemistry and Biology 2,127. Hayes, J. J. and Tullius, T. D. (1989). Biochemistry 28, 9521. Spiro, C. and McMurray, C. T. (1997). /. Biol. Chem. 272, 33145. Collins-Hicok, J., Lin, L., Spiro, C., Laybourn, P. J., Tschumper, R., Rapacz, B., and McMurray, C. T. (1994). Mol. Cell. Biol. 14, 2837. Meinkoth, J. and Wahl, G. (1984). Anal. Biochem. 138,267. Sears, L. E., Moran, L. S., Kissinger, C., Creasy, T., Perry-O'Keefe, H., Roskey, M., et al. (1992). Biotechniques 13, 626. Pfeifer, G. P., Steigerwald, S. D., Mueller, P. R., Wold, B., and Riggs, A. D. (1989). Science 246, 810. Church, G. M., Ephrussi, A., Gilbert, W., and Tonegawa, S. (1985). Nature 313, 798. Saluz, H. P. and Jost, J. (1993). Crit Rev. Eukaryotic Gene Exp. 3, 1. Pfeifer, G. P. and Riggs, A. D. (1993). In Methods in molecular biology, Vol. 23. DNA sequencing protocols (ed. H. and A. Griffin), p. 169. Humana Press, Totowa, NJ. Hornstra, I. K. and Yang, T. P. (1993). Anal. Biochem. 213,179. Garrity, P. A. and Wold, B. J. (1992). Proc. NatlAcad. Sci. USA 89, 1021. Mueller, P. R., Garrity, P. A., and Wold, B. (1992). In Current protocols in molecular biology (ed. F. A. Ausubel, R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith, and K. Struhl), p. 15.5.1. John Wiley & Sons, New York. Cartwright, I. L. and Kelly, S. E. (1991). Biotechniques 11, 188. Mattes, W. B., Hartley, J. A., and Kohn, K. W. (1986). Biochim. Biophys. Acta 868, 71. Strauss, E. C., Andrews, N. C., Higgs, D. R., and Orkin, S. H. (1992). Mol. Cell. Biol. 12, 2135. Pfeifer, G. P. and Riggs, A. D. (1991). Genes Dev. S, 1102. Pfeifer, G. P., Drouin, R., Riggs, A. D., and Hohnquist, G. P. (1992). Mol. Cell. Biol. 12, 1798. Zhang, L. and Gralla, J. D. (1989). Genes Dev. 3, 1814. Rigaud, G., Roux, J., Pictet, R., and Grange, T. (1991). Cell 67, 977. McPherson, C. E., Shim, E., Friedman, D. S., and Zaret, K. S. (1993). Cell 75, 387. Leeds, J. M., Slabaugh, M. B., and Mathews, C. K. (1985). Mol. Cell. Biol. 5, 3443. Sheen, J. Y. and Seed. B. (1988). Biotechniques 6, 942.

62

3

In vitro transcription and characterization of transcription factors AUSTIN J. COONEY, SOPHIA Y. TSAI, and MING-JER TSAI

1. Introduction Once a cw-element and transacting factor have been determined to be important for regulation of transcription of a gene, gel retardation, DNAase I footprinting, methylation interference analyses, and in vitro transcription assays are generally employed for characterization of a factor during purification. The majority of transcription factors are present in very low quantities within a cell, making purification long and arduous, and requiring large quantities of starting material. Thus, the initial characterization of the factor is undertaken with partially purified material. This permits elucidation of vital information about the factor, such as the molecular weight, subunit composition, and interactions, if any, with non-DNA binding factors, before complete purification of the factor (1). This information is mostly derived from analysis of data from a combination of in vitro transcription assays and chromatographic and electrophoretic techniques, which will be covered in this chapter.

2. In vitro transcription assays In vitro transcription assays involve the reconstitution of gene expression in a test tube. They can be used for many different functional purposes, including mechanistic and kinetic studies of transcription, analysis of promoter and response elements, and identification and purification of transcription factors. In vitro transcription assays can also be used as functional assays to monitor the activity of a transcription factor during purification. In vitro transcription assays can also be used to monitor and study the activities of both DNA binding and non-DNA binding transcription factors (1). Using a chicken ovalbumin upstream promoter-transcription factor (COUP-TF)-dependent in vitro transcription assay, it was shown that the ovalbumin promoter activity was depen-

Austin J. Cooney et al. dent on an additional factor S300-II, a non-DNA binding factor. An S300-IIdependent in vitro transcription assay was used to purify and clone S300-II, which turned out to be identical to TFIIB (1,2). The basic components of a reconstituted in vitro transcription assay are a source of general transcription factors, specific transcription factors, and reporters. Reporters can be either driven by endogenous promoters such as the oestrogen responsive vitellogenin or mouse mammary tumour virus (MMTV) promoters or driven by minimal promoters containing specific response elements (3, 4). Minimal promoters containing only TATA boxes from the ovalbumin promoter have been successfully used with steroid response elements in in vitro transcription assays (5-8). Generally, multiple copies of a single response element are required to yield a robust response because endogenous promoters are complex and contain multiple response elements for various types of transcription factors, which generally interact synergistically to regulate gene expression (3, 4). For example, using a single progesterone response element led to 5-fold stimulation above the basal promoter activity while two copies led to a 30-fold induction (5). The reporter or template can be either linear or circular and similar results are obtained with either template (9). Several different types of in vitro transcription assays have been developed that are either direct, i.e. radiolabel is incorporated into the transcript, or indirect where the transcript is detected using radiolabelled probes. Direct assays include run-off assays or the G-free cassette assay. Indirect assays include nuclease SI analysis, primer extension, or RNAase protection. The G-free in vitro transcription assay is a direct approach to monitor gene expression and is based on the properties of the Tl RNAase which preferentially digests RNA containing G residues. Thus, Sawadogo et al. developed a G-free cassette, which when expressed would be resistant to Tl RNAase activity (10, 11). The transcription initiation site for the template lies 17 nucleotides within the 5' boundary of the G-free cassette; this permits distinction between specific and non-specific (read-through) transcriptional events after Tl RNAase digestion (Figure 1). Specific transcripts correctly initiated will be 360 nt long, while non-specific events, initiated anywhere within the plasmid, will yield a product of 377 nt, distinguishable on a denaturing urea polyacrylamide gel (Figures 1 and 2). A second G-free template is generally incorporated into the assay as an internal control. It contains a shorter G-free cassette (i.e. 190 nt specific and 200 nt non-specific transcripts) to permit easy distinction from the test plasmid (360 and 377 nt) (Figure 1). The second Gfree cassette is put under the regulation of a second general promoter, such as the adenovirus major late promoter. Such an internal control will allow the researcher to control for effects on basal transcription, which would have similar affects on both the control and test promoters. The G-free assay depicted in Figure 1 is given in detail in Protocol 2. Figure 2 is an illustration of the use to which the G-free in vitro transcription assay can be put. In this experiment progesterone receptor (PR) activity 64

3: In vitro transcription

Figure 1. Schematic diagram of the G-free in vitro transcription assay. (A) The test template (PRE TATA) which, after RNAase T1 processed transcription, yields correctly (360 nt) and incorrectly initiated (read-through) transcripts (377 nt). (B) The internal control, adeno major late promoter driven template, which yields correctly (190 nt) and incorrectly (200 nt) initiated transcripts.

is being assayed. The templates are PRE2 pLovTATA (progesterone response element two times upstream of the ovalbumin TATA box, lanes 1-6) and pLovTATA (lanes 7-9). When increasing amounts of progesterone receptor, in the presence of progesterone, are added, there is a dose-dependent increase in the specific transcript (arrow 3, lanes 2-6) relative to basal activity in the absence of receptor (lane 1). This increase is specific for the PRE elements as no increased expression is observed with the parent vector, pLovTATA which lacks the PREs (lanes 7-9) or indeed with the internal control (arrows 1 and 2). Note that the progesterone receptor only affects the level of correctly initiated transcripts (arrow 3, lanes 1-6) and not the non-specific transcription (arrow 4, lanes 1-6). Thus, in one test tube one can control for non-specific and specific effects with two templates that each yield two processed transcripts, which can be distinguished by the correctly and incorrectly initiated transcriptional events. This assay can be applied to many problems and has been used for functional and kinetic studies. A typical in vitro transcription assay involves the incubation of purified templates with RNA polymerase and ribonucleotides, the basal transcription machinery, and purified transcription factors under the appropriate temperature, salt, and buffer conditions. Transcription proceeds for 45-60 min, subsequently the reaction is terminated and the transcripts purified. We provide protocols for the preparation of transcriptionally active nuclear extracts and for the G-free assay itself which will be covered in this chapter. The con65

Austin J. Cooney et al.

Figure 2. Progesterone receptor dependent in vitro transcription. Progesterone receptor and progesterone (10 7 M) were added in increasing amounts to either PRE 2 pLovTATA (lanes 1-6) or the control pLovTATA (-PREI (lanes 7-9). Arrow 1, correctly initiated transcript from the adeno major late internal control; arrow 2, incorrectly initiated transcript from the adeno major late promoter; arrow 3, correctly initiated transcript from either PRE; pLovTATA or pLovTATA; and arrow 4, the incorrectly initiated transcript from either of the test templates PRE2 pLovTATA or pLovTATA.

ditions for optimum in vitro transcription will vary from factor to factor analysed and templates, assay, and extracts utilized. However, the optimum conditions can be determined empirically by systematically varying each of the components and analysing the results in in vitro transcription. Some of the variables that can be altered include the amount of transcription factor added, the concentrations of salt and Mg 2+ > as well as the amount of nuclear extracts. An excellent reference for such systematic optimization of in vitro transcription is the paper by Bagchi etal., who optimized PR-dependent in vitro transcription (9). The general transcriptional machinery, which includes RNA polymerases, basal transcription factors TF1IB-H, and associated transcription activation factors (TAFs) are generally contributed from nuclear extracts. Nuclear extracts that support high levels of in vitro transcription have been successfully generated form HeLa cells and Drosophila embryos (4, 12). Below we describe the generation of nuclear extracts from HeLa cells (Protocol I). See Chapter 9 for the methods used for preparation of such extracts from rat brain and their use in neuronal promoter analysis. 66

3: In vitro transcription Protocol 1. Preparation of HeLa cell nuclear extract Eauioment and reagents PBS: 150 mM NaCI, 16 mM Na2HP04,16nM NaH2P04 pH 7.3 Buffer A: 10 mM Hepes pH 7.9, 1.5 mM MgC12, 10 mM KCI, 2 mM DTT8 Buffer C: 20 mM Hepes pH 7.9, 20% glycerol, 1.5 mM MgC12, 0.2 mM EDTA, 2 mM DTT3

Buffer D: 20 mM Hepes pH 7.9,20% glycerol, 0.1 M KCI, 0.2 mM EDTA, 5 mM MgC12, 2 mM DTP Dialysis bag, M, cut-off 12000-14000 Da Beckman centrifuge, JA20 or JA14 rotor Dounce homogenizer (glass, B-type pestle)

Method 1. Harvest HeLa cells from culture by centrifuging at 1000 x g at 4°C for 10 min in a Beckman JA20 or JA14 rotor, depending on culture volume. 2. Resuspend the cell pellet in 5-10 packed cell volumes (PCV) of icecold PBS. 3. Centrifuge at 500 x g at 4°C for 10 min. 4. Resuspend the cell pellet in 5 x PCV of buffer A. 5. Centrifuge at 500 x g at 4°C for 10 min. 6. Aspirate the supernatant and resuspend the cells in 2 x PCV of buffer A. 7. Transfer the cells to a Dounce homogenizer. Lyse the cells with 10 strokes. 8. Examine the homogenate microscopically for maximum cell lysis (90% of cells should be lysed). 9. Centrifuge at 500 x g at 4°C for 10 min. Carefully aspirate the supernatant and recentrifuge the pellet at 27000 x g in a Beckman JA20 rotor at 4°C for 20 min to remove any residual supernatant. 10. Resuspend the nuclei pellet in 0.5 X PCV of buffer C containing 0.6 M NaCI or 0.66 x PCV of buffer C containing 0.55 M NaCI. 11. Transfer to an appropriately sized Dounce homogenizer and lyse the nuclei with 10 strokes. 12. Transfer the lysate to 50 ml polypropylene centrifuge tubes and cover with parafilm. Incubate on ice for 30 min, vortexing gently at 5 min intervals. 13. Centrifuge at 27 000 x g for 30 min at 4°C in a JA20 rotor. 14. Recover the supernatant (nuclear extract) and dialyse against 100 volumes of buffer D for 4 h, with one change of buffer. 15. Centrifuge the nuclear extract at 27000 x g for 20 min at 4°C in a JA20 rotor. Recover the supernatant, aliquot, and freeze at -80°C. " DTT is added fresh.

67

Austin J. Cooney et al. Protocol 2.

In vitro transcription using the G-free cassette

Equipment and reagents Transcription reaction mixture: 12.5-200 ng of control and test template DNAs;3 60 mM KCI, 12 mM Hepes pH 7.9, 5 mM MgC12, 12% glycerol, 0.12 mM EDTA, 2 mM DTT; ribonucleotide mix:500 nM ATP, CTP; 20 uM UTP; 2 (uCi/reaction [32P]UTP (800 Ci/mM); 10 lU/reaction nuclease T1; 1 fimol/reaction 3'-O-methyl GTP; herring sperm DNA (0.5-1 ug/reaction); 3-5 ul nuclear extract Stop solution: 10 mM Tris-HCI pH 8.0, 1 mM EDTA, 8 M urea

Transcription termination mixture: 30 mM Tris-HCI pH 8.0,10 mM EDTA, 0.5% SDS; 120 ug/H.1 yeast tRNA; 400 ug/ml proteinase K Phenol:trichloromethane:/so-pentanol (25:24:1 v/v; phenol buffered to pH 5.0 with 10 mM sodium acetate) 3 M sodium acetate pH 5.2 80% deionized formamide, 1 mM EDTA, 1 ng/ml xylene cyanol, 1 ng/ml bromophenol blue Dry ice/ethanol bath

Method 1. Make up the transcription reaction mixture in an Eppendorf tube to a final reaction volume of 25-30 uJ. Mix by gentle vortexing and spin down briefly. Start the reaction by adding the ribonucleotide mix and incubate at 30°C for 45 min. 2. Terminate transcription by adding 85 uJ transcription termination mixture. 3. Incubate at 37°C for 30 min. 4. Add 200 ul urea stop solution per reaction. 5. Extract twice with an equal volume of phenol:trichloromethane:/sopentanol. 6. Aspirate the upper aqueous phase and add one-tenth volume 3 M sodium acetate (pH 5.2),10 pig yeast tRNA, and 2.5 volumes of absolute ethanol. Incubate in a dry ice/ethanol bath for 30-40 min. 7. Centrifuge at 13000 x gfor 60 min at 4°C. Wash the pellet once with 80% ethanol (-20°C). 8. Centrifuge for 5 min, aspirate the supernatant, and dry the pellet. 9. Resuspend the pellet in 4 ul of 80% deionized formamide, 1 mM EDTA, 1 ng/ml xylene cyanol, 1 ng/ml bromophenol blue. Heat at 95°C for 5 min and cool quickly on ice. 10. Run the samples on a 7% acrylamide urea sequencing gel. Electrophorese at 1800 V. 11. Fix the gel in 10% methanol and 10% acetic acid for 5 min and dry the gel. 12. Autoradiograph overnight at —70°C with two intensifying screens. " The amount of DNA added depends on the strength of the promoter.

68

3: In vitro transcription

3. Determination of the molecular weight of native transcription factors Molecular weight determination under native conditions allows the researcher to estimate the total size of the factor/complex binding to an element and may aid in identifying whether the binding species is a monomer, dimer, or higher complex. Separation of factors on the basis of size, under native conditions, has the advantage of allowing direct assay of DNA binding and transcriptional activities, to assign the factor's molecular weight relatively easily with samples where purity is not very high. In order to determine accurately the molecular weight of the factor under native conditions, the factor must be fairly globular in conformation. Transcription factors exist as an arrangement of globular domains and thus lend themselves to this type of analysis.

3.1 Molecular weight determination using gel filtration chromatography Gel nitration chromatography separates proteins on the basis of their size, where size refers to the physical dimensions or Stokes' radius, rather than the molecular weight, which is related to mass. However, in the case of globular proteins that have a roughly spherical tertiary structure, the Stokes' radius is proportional to the molecular weight. Thus a semi-logarithmic standard curve can be produced plotting Km versus log molecular weight, where Kav is the partition coefficient (see Figure 3). The partition coefficient is defined by the equation

where Ve = elution volume of the protein, V0 = void volume, and Vt = total volume.

Figure 3. The sigmoidal dependence of Kav, the partition coefficient, on the logarithm of the molecular weight (Mol. Wt.).

69

Austin J. Cooneyet al. Thus the molecular weight of an unknown factor can be determined by linear regression from such a standard curve constructed from the chromatography of standard proteins of known molecular weight. The gel nitration matrix consists of porous beads that sieve proteins molecularly from the largest to the smallest. Gel filtration achieves two purposes: it permits determination of the molecular weight of a native factor binding to a response element (13,14) and it also achieves an additional step in the purification of the transcription factor (15, 16). Gel filtration columns can be calibrated with gel filtration protein standards of known molecular size (Bio-Rad Laboratories, Pharmacia LKB Biotechnology, or Sigma Chemical Co.). Identification of the elution volume of the DNA binding or transcriptional activities allows determination of the molecular weight of the factor by reference to the standard curve. Table 1. Fractionation ranges of some commercially available gel filtration matrices Matrix type9

Fractionation rangeb

SephadexG-10 SephadexG-15 Bio-Gel P-2 Bio-Gel P-4 Sephadex G-25 Bio-Gel P-6 Bio-Gel P-10 Sephadex G-50 Bio-Gel P-30 Bio-Gel P-60 Sephadex G-75 SephacrylS-100 Bio-Gel P-100 Sephadex G-100 Sephacryl S-200 Sephadex G-150 Sephadex G-200 Bio-Gel A-0.5M Sephacryl S-300 Bio-Gel A-1.5M Sepharose CL-6B Bio-Gel A-5M Bio-Gel A-15M Sepharose CL-4B Sepharose CL-2B Bio-Gel A-50M Bio-Gel A-150M

<7 x 102 <1.5X 103 1 X10 2 -1.8X 103 8 X 1 02-4 X 103 1 X 103-5X 103 1 X10 3 -6X10 3 1.5 X 103-2X 104 1.5 X 103-3X 104 2.5 X 103-4X 10" 3X 103-6x 104 3x 103-8X 104 1 X 1 03-1 X 105 5X 103-1 X 106 4X 103-1.5X 105 5X 103-2.5X 10 5 5x 103-3x 105 5 x l03-6 x 105 1 X 104-5X 105 1 X10 4 -1.5X10 6 1 x 104-1.5x 106 1 x 104-4x 106 1 X 104-5X 106 4x 104-1.5X 107 6x 104-2X 107 7 x 104-4X 107 1 X10 5 -5X10 7 1 X10 6 -1.5X10 8

" Sephadex, Sephacryl, and Sepharose are supplied by Pharmacia LKB Biotechnology and Bio-Gel is supplied by Bio-Rad. 6 The fractionation range denotes the molecular sizes of proteins that elute in a volume equal to the bed volume to proteins that elute in the void volume. 70

3: In vitro transcription Selection of the appropriate gel filtration matrix is important. Initially, for an uncharacterized transcription factor whose molecular weight is unknown, a matrix with a broad fractionation range is appropriate. However, once the molecular weight of the factor has been estimated, the most accurate determination can be achieved with a matrix from which the factor will elute midway in the fractionation range, where the plot of Km vs. log molecular weight is linear (see Figure 3). Sephacryl S-300 (Pharmacia) or Bio-Gel A-1.5M (Bio-Rad) have excellent broad fractionation ranges from 104 to 1.5 X 106 Da relative molecular mass (Mr), and are useful in the initial determinations of molecular weights. A number of different matrices, with different fractionation ranges and characteristics, are produced by Pharmacia LKB and BioRad (see Table 1). A potential, but infrequent, problem with gel filtration matrices is nonspecific interaction between a transcription factor and the gel matrix, leading to retardation of its elution or the opposite effect. This has been observed for some steroid receptors when agarose A15M or A-1.5M (Bio-Rad) is used; where this phenomenon is suspected, i.e. Km > 1, change the matrix from Sepharose to acrylamide or vice versa. However, the molecular weights of many native factors have been determined by gel filtration (1). Pharmacia also produces pre-poured gel filtration columns (Superose) for use with their FPLC system, which can be used for all the purposes described here. Protocol 3. Molecular weight determination by gel filtration chromatography Equipment and reagents Gel matrix Gel filtration column Automated fraction collector with on-line UV detector

Electrophoresis apparatus Dextran blue, potassium bichromate Protein standards

Method 1. Swell the dry gel matrix in the appropriate buffer or according to the manufacturer's directions. Allow the gel matrix to settle and aspirate the fines. Alternatively, if the gel matrix comes pre-swollen, wash the matrix in a sintered-glass funnel to remove the preservatives and to equilibrate it in the gel filtration buffer. De-gas the slurry. 2. Erect the column vertically in a 4°C cold room. Optimum resolution of proteins is best achieved with columns greater than 50 cm in length with a matrix diameter of 1 cm. 3. Fill the outlet tube with buffer and clamp it. Then add some buffer to the column gel matrix suspension and pour slowly to the desired bed height. Great care should be taken in pouring the column; it should be 71

Austin ]. Cooney et al. Protocol 3.

Continued

poured in a single motion to produce an homogeneous column matrix with no air bubbles. 4. Layer 1 cm of buffer carefully on top of the column bed, after the gel matrix has settled, without disturbing the matrix. Then connect the gel filtration buffer reservoir to the column. Open the column outlet and wash the gel matrix with 2-3 column volumes of the gel filtration buffer. 5. Determine the void volume (V0) and the total volume (Vt) of the column by filtration of dextran blue and potassium bichromate, respectively. 6. Carefully allow the buffer in the column to drop to the level of the bed and close the outlet. Disconnect the reservoir and gently load the protein standards (10-20 u,g each in a volume equivalent to 1-2% of the total column volume.) Open the outlet and let the protein standards enter the column, then close the outlet. Carefully pipette a 1 cm layer of gel filtration buffer on to the column bed and reconnect the buffer reservoir. 7. Open the outlet and commence elution of the protein standards at a flow rate of 8 ml/h.a Collect fractions equal to 1-2% of the total column volume using an automated fraction collector with an on-line UV detector. Alternatively, aliquots of each fraction can be taken to determine the protein peaks of the standards by measuring the absorbance at 280 nm (A2go). 8. Electrophorese aliquots of each fraction containing a protein peak on a denaturing 10% polyacrylamide-SDS gel, with a stacking gel using the discontinuous buffer system of Laemmli (17), to determine their molecular weight. This calibrates the column and allows the preparation of a standard curve to predict the size of a polypeptide which elutes in a particular fraction from the column. 9. Load the protein sample at a concentration of 0.2-0.5 mg/ml in a volume equal to 1-2% of the column volume. The loading and elution of the transcription factor preparations are the same as for the protein standards (see steps 5 and 6). 10. Measure the A2eo of aliquots from each fraction. Take 5-10 uJ aliquots from alternate fractions to assay for specific DNA binding, using gel retardation, DNAase I footprinting assays (see Chapters 1 and 2 respectively), and/or transcriptional activity. "The flow rate varies from column matrix to matrix and is determined by multiplying the linear flow rate (supplied by the manufacturer) by the cross-sectional area of the column in cm2 to yield the flow rate in ml/h.

72

3: In vitro transcription

3.2 Molecular weight determination using glycerol gradient centrifugation An alternative method can be used to determine and to confirm the native size determined by gel filtration and to eliminate possibilities of non-specific interactions between the gel matrix and the factor. This method uses density gradient sedimentation, developed by Martin and Ames (18). The formation of a glycerol gradient and sedimentation of the applied sample by ultracentrifugation permits the fractionation of partially purified factors on the basis of their sedimentation coefficients, which approximates to their molecular weight (14). Determination of molecular weight by sedimentation velocity has the advantage over gel chromatography of being a primary technique which permits the direct determination of hydrodynamic parameters, such as the sedimentation coefficient (s), from first principles. See Siegel and Monty (19) for a detailed discussion of the derivation and use of the equations. Protein standards of known s can be used to determine molecular weights rapidly and easily by glycerol gradient centrifugation. The use of glycerol or sucrose to form the gradient has the advantage, because of their viscosity and density, that essentially linear migration of the majority of biological macromolecules results. Thus, the ratio (R) of the distances sedimented by any two substances will always be constant. Given a standard protein of known s, the s of an unknown factor can be calculated from the equation

As the rate of movement of any macromolecule is nearly constant and assuming they have the same partial specific volumes, then

where S20,w is the sedimentation coefficient at 20°C in water. An approximate value for the molecular weight can be determined from

where MW = molecular weight. The glycerol gradient can be calibrated by sedimenting protein standards of known molecular weight (Bio-Rad) whose position in the gradient can be determined by electrophoresing aliquots from alternate fractions on an SDSpolyacrylamide gel (17). Identification of the position in the glycerol gradient to which the transcription factor sediments is determined by assaying fractions of the gradient for specific DNA binding and/or transcriptional activity. Measurement of the distance travelled by the unknown factor and standards 73

Austin J. Cooney et al. permits calculation of the factor's molecular weight. This technique has been commonly used to determine the native molecular weight of transcription factors (1,14, 20). Protocol 4. Determination of molecular weights by glycerol gradient centrifugation Equipment and reagents Buffer A: 20 mM Hepes pH 7.9, 100 mM KCI, 4 mM DTT, 0.2 mM EDTA Density gradient former (Beckman Instruments, Inc.) and mixing chamber SW50.1 tubes and rotor (Beckman) SP6 polymerase (Promega)

Auto Densi-Flow (Labconco Corp.) and fraction collector Molecular weight standards (Bio-Had) Electrophoresis apparatus

Method 1. Pour 4.4 ml linear 7-23% (w/v) glycerol gradients in buffer A, in SW50.1 tubes using a density gradient former and a mixing chamber. 2. Layer a 200 uJ aliquot of partially purified or affinity purified transcription factor carefully on to the pre-formed glycerol gradients. This sample should contain 15 ID of SP6 polymerase as an internal control. 3. Layer 10 to 20 ug of molecular weight standards carefully on to a balance gradient to construct a standard curve to calibrate the gradient. 4. Centrifuge the samples at 25000 x g (45000 r.p.m.) in an SW50.1 rotor (Beckman) for 16 h at 4°C. 5. Collect fractions of 150-200 ul from each gradient from the top to the bottom with an Auto Densi-Flow (Buchler) connected to a fraction collector, or puncture the bottom of the tube and collect fractions from the bottom up. 6. Assay each fraction for SP6 polymerase activity as per the manufacturer's instructions. It has a molecular weight of 98 kDa. 7. Electrophorese an aliquot of each fraction from the standard curve glycerol gradient under denaturing conditions in a 10% SDS-polyacrylamide gel (17). The positions of the standards are detected by silver or coomassie blue staining (21) to determine in which fractions they are located, to calibrate the glycerol gradients. 8. Assay 5-10 ul aliquots of each fraction for specific DNA binding, by gel retardation or DNAase I footprinting, and/or transcriptional activity to identify the fraction in which the transcription factor of interest sediments. This allows determination of the molecular weight by comparison with the standard curve. 74

3: In vitro transcription

3.3 Determination of molecular weight of native factors using non-denaturing gradient gel electrophoresis Non-denaturing gradient gel electrophoresis can be used to determine the native molecular weight of a transcription factor (22, 23). Gradients of polyacrylamide yield greater resolution of a protein mixture compared with a single polyacrylamide concentration and are thus more accurate for molecular weight determinations of native proteins. If the factor has been purified to homogeneity, it can be identified by silver staining (21), and if it has only been partially purified and an antibody is available, it can be identified by Western analysis (22,24). Here we will describe non-denaturing gradient electrophoresis and Western analysis. Although we describe the traditional use of radioiodinated protein A for the detection of proteins in Western analysis, many excellent non-radioactive kits are available from several manufacturers, with accompanying protocols, which allow rapid sensitive, safe, and reproducible analysis of Western blots. Many transcription factors in their native form are associated with other factors in relatively large complexes or as dimers, thus to facilitate the electrophoretic transfer of the factor, 20% of the N,N'-methylenebisacrylamide cross-linker can be replaced with the reversible cross-linker diallyltartardiamide (DATD): incorporation of DATD allows the polyacrylamide gel to be permeabilized by brief incubation with 5 mM periodic acid. A drawback of non-denaturing gradient electrophoresis is that it yields the best results with highly purified transcription factor preparations. Crude nuclear or cytosolic extracts tend to form insoluble aggregates which do not enter the gel and remain in the well, and also tend to produce ill-defined smeared bands on the Western immunoblot. Protocol 5. Molecular weight determination by non-denaturing gradient electrophoresis Equipment and reagents 0.5 x TBE: 50 mM Tris, 50 mM boric acid, 1 mM EDTA Transfer buffer: 49 mM Tris, 39 mM glycine, 0.04% SDS pH 9.2 5 mM periodic acid Pre-stained molecular weight markers (BioRad) Gradient gel casting chamber (Bio-Rad) Peristaltic pump (Gilson, Oberlin OH) Electrophoresis apparatus

Nitrocellulose or nylon membrane (e.g. Zeta-Probe or Immobilon P) Whatman No 3 paper Trans-blot apparatus Blotto: 2-3% non-fat dry milk (w/v), 10 mM Tris-HCI pH 8, 150 mM NaCI Western buffer: 10 mM Tris-HCI pH 8.0, 150 mM NaCI 125 l-labelled protein A (1.11 MBq/ng ICN)

Method 1. Cast a 3-25% gradient polyacrylamide gel (acrylamide:bisacrylamide: DATD = 79:0.8:0.2 w/w) in 0.5 x TBE with a gradient gel casting chamber. Maintain a 12 ml/min flow rate using a peristaltic pump.

75

Austin J. Cooney et al. Protocol 5.

Continued

2. Pre-run the gel at 300 V for 3 h to overnight in 0.5 x TBE at 4°C. 3. Incubate the samples at room temperature for 10-15 min. At this stage, DNA or ligands for particular experiments can be added. 4. Load 200 ng to 1 ug of transcription factor preparation and pre-stained molecular weight markers (Bio-Rad). Run at 300 V for 5 h at 4°C. 5. Rinse the gel in distilled H2O three times. 6. Incubate the gel in 5 mM periodic acid for 15 min at 30°C. 7. Repeat step 5. 8. Incubate in transfer buffer for 10 min. 9. Place the gel on a pre-cut, pre-wetted nitrocellulose or nylon membrane and sandwich between pre-wetted Whatman No 3 paper, three sheets on each side. 10. Place the sandwich in trans-blot apparatus and run at 20 V for 1 h in transfer buffer. 11. Disassemble the blot and block the membrane for 1 h at room temperature with blotto. 12. Decant the blotto and add the appropriate amount of antibody in 8-10 ml blotto. Incubate for 3-4 h at room temperature with gentle shaking. 13. Decant the antibody solution and wash the membrane with 40 ml Western buffer for 5 min with shaking. Decant and repeat twice. 14. If a monoclonal antibody was used, add the secondary antibody (rabbit anti-mouse IgG, commercially available) in 10 ml blotto and incubate at room temperature for 90 min with gentle shaking. If a polyclonal antibody was used, go straight to step 16. 15. Repeat step 13. 16. Add 2 ul of 125l-labelled protein A (1.11 MBq/ug) to 10 ml blotto. Incubate at room temperature with gentle shaking for 90 min. 17. Decant and dispose of the radioactive blotto carefully. Wash the membrane with 40 ml Western buffer for 10 min at room temperature. Decant and repeat twice. 18. Dry the membrane and develop the signal by autoradiography.

4. Determination of the molecular weight of denatured transcription factors Analysis of a factor under native conditions only permits estimation of the molecular weight of the total complex binding to an element. Electrophoresis 76

3: In vitro transcription under denaturing conditions (SDS-polyacrylamide gel electrophoresis (SDSPAGE)) allows accurate determination of the molecular weight of the individual transcription factors involved in that complex (14, 17, 25, 26). Comparison of the results will also yield invaluable insights into the factors binding to an element. If the factor binds as a dimer, then the native molecular weight will be approximately double that of the denatured factor. Also, if the complex is composed of a heterodimer, it may be possible to identify two species by SDS-PAGE if they have different molecular weights. Determination of the molecular weight of a factor under denaturing conditions requires the identification of a specific band on an SDS-polyacrylamide gel which corresponds to the binding or transcriptional activity being studied. The two methods outlined here differ conceptually as to whether the factor is identified before or after running the SDS-polyacrylamide gel. Ultraviolet (UV)-cross linking to DNA identifies the binding activity prior to running the SDS gel, while renaturation and assay for binding or transcriptional activity identifies the factor after running the gel. Both of these procedures have been employed successfully to identify the molecular size of many transcription factors early in the purification procedure.

4.1 UV cross-linking of transcription factors to DNA Irradiation of DNA with UV light produces purine and pyrimidine free radicals that are chemically reactive and can form covalent bonds such as thymidine dimers. This reactive property of UV-irradiated DNA can be used to link transcription factors covalently to their response elements. When a proteinnucleic acid complex is irradiated with UV light, it causes the formation of covalent bonds between pyrimidines and certain amino acid residues in the DNA binding domain in close proximity to the DNA. Thus a transcription factor can be selectively labelled indirectly by UV cross-linking as a consequence of its specific binding to a DNA sequence (27). Labelling the transcription factor in this fashion allows the easy and rapid determination of its molecular weight, under denaturing conditions in an SDS-polyacrylamide gel, even in crude extracts. We describe two techniques here, one involves incorporation of deoxybromouridine into the DNA prior to cross-linking, the other utilizes UV cross-linking to an unsubstituted oligonucleotide and gel retardation assay.

4.2 UV cross-linking with a bromodeoxyuridine-substituted DNA Halogenated analogues of thymidine (e.g. bromodeoxyuridine) are significantly more sensitive to UV-induced cross-linking because replacement of the thymidine methyl group with a bromine atom creates a molecule more susceptible to UV-light-induced free-radical formation (28). Bromodeoxyuridine can be functionally utilized in place of thymidine by many enzymes including 77

Austin J. Cooney et al. the Klenow fragment, which is used to prepare the substituted probe. An additional advantage of using bromodeoxyuridine is that a UV light of longer wavelength can be used for cross-linking than can be used with unsubstituted DNA. Longer wavelength UV light is less damaging to DNA because it causes fewer nicks in the phosphate backbone. After cross-linking of the factor and DNA, the probe is digested to remove the majority of the probe not protected by and covalently linked to the factor. This is a precautionary measure since UV light may also induce the formation of thymidine dimers in the DNA, leading to electrophoretic anomalies. Thus, the removal of excess DNA will minimize this problem. Protocol 6. Cross-linking with a bromodeoxyuridine-substituted DNA Equipment and reagents • 10 x Klenow buffer: 50 mM Tris-HCI pH7.5, . 10 mM MgCI2, 1 mM DTT, 50 mg/ml bovine serum albumin (BSA) • . [a32P]dCTP(110TBq/mmol) • . dNTP mix: 2.5 mM each of dATP, dGTP, 5bromo 2'-deoxyuridine trisphosphate (Pharmacia), and 0.25 mM dCTP • . 0.1 M DTT • Klenow fragment enzyme (25 IU, Promega) . • Phenol:trichloromethane:/so-pentanol (25:24:1 v/v) • Microcentrifuge • • 0.3 M ammonium acetate • Dry ice/ethanol bath • « 10 x loading buffer: 0.25% bromophenol blue, 0.25% xylene cyanol, 25% ficoll (type • 400) in H20 • Transcription buffer: 20 mM Hepes pH 7.9, . • 100 mM KCI, 20% glycerol, 4 mM DTT, 0.2 mM EDTA

TE buffer (pH 8) 10mM Tris-HCI ph 8.0, ImMEDTA Agarose gel electrophoresis set-up UV transilluminator emitting at 305-310 nm, with a maximum intensity of -7000 W/cm2 (e.g. Fotodyne UV lamp) 0.5 M CaCI2, DNAase I (Worthington Diagnostics), micrococcal nuclease (Worthington Diagnostics) ~~ 2 x SDS sample buffer: 0.125 M Tris-HCI pH 6.8, 6% SDS, 10% p-mercaptoethanol, 20% glycerol, 0.025% bromophenol blue SDS-PAGE set-up with 10% polyacrylamide gel and stacking gel "C-iabelled protein standards (BRL, Life Technologies, Inc.) Fluor (e.g. Enhance, MEN) Kodak XAR-5 X-ray film Dupont Cronex Lightening Plus intensifying screen

Method 1. Subclone a DNA fragment or oligonucleotide containing the specific response element for the transcription factor into M13. 2. Mix 5 ug of the recombinant single-stranded M13 clone containing the binding site with an equimolar amount of the 17 nucleotide M13 universal primer. Add 10 (JL! of 10 x Klenow buffer. 3. Heat at 95°C for 5 min and cool very slowly at room temperature, this should take several hours. 4. Make the following additions to the annealed mixture: 50 uI [a32P]dCTP (110 TBq/mmol) 3.5 (ul of dNTP mix 1.75(11 0.1 MDTT 78

3: In vitro transcription 7.5 ul 10 x Klenow buffer 7 ul H20 5 ul Klenow fragment enzyme (25 1U, Promega) 5. Incubate at 16°C for 90 min. 6. Heat the reaction to 68°C for 10 min to inactivate the Klenow enzyme. 7. Digest the double-stranded M13 recombinant vector with an excess of enzymes, under appropriate reaction conditions, which will liberate the inserted response element. 8. Extract the reaction once with phenol:trichloromethane:iso-pentanol. Microcentrifuge for 5 min at 11 000 X g and aspirate the aqueous phase. 9. Precipitate the DNA with 0.3 M ammonium acetate and 2 volumes of 100% ethanol, incubate in a dry ice/ethanol bath for 10 min and microcentrifuge for 15 min at 11000 x g. Wash the pellet with 0.5 ml 70% ethanol. Dry the pellet and resuspend in 20 ul TE buffer (pH 8) and add 2 ul 10 x loading buffer. 10. Load the DNA on to an agarose gel and electrophorese. Isolate the required fragment by your favourite technique (or use Geneclean (Bio 101) or NA45 DEAE membrane (Schleicher and Schuell), according to the manufacturers' protocols). 11. Dilute and count an aliquot of the probe using a scintillation counter and determine the DNA concentration by measuring the A260. Then determine the specific activity of the probe. 12. Set up a number of binding reactions. Incubate at room temperature under conditions determined previously (generally those used for gel retardation assays), for 15-30 min. The reactions should contain 105 c.p.m. of labelled probe, the extract or fraction containing the transcription factor, transcription buffer and 10-20 ug of non-specific competitor, e.g. poly(dl-dC)-poly(dl-dC); bring the volume up to 50 ul with H2O. Check for successful binding by running a gel retardation assay. 13. Cover the top of the tube with UV-transparent cling film. 14. Irradiate the binding reaction from a distance of 5 cm (see Figure 4) with a UV transilluminator for 5-60 min. The effect of UV irradiation can also be checked by running a gel retardation assay. 15. Add 1 ul 0.5 M CaCI 2 ,4 mg DNAase I, and 1 IU micrococcal nuclease to each reaction and incubate at 37°C for 30 min. 16. Add an equal volume of 2 x SDS sample buffer to each binding reaction and boil for 5 min. 17. Load the samples on a 10% polyacrylamide-SDS gel with a stacking gel, include a marker lane containing 14C-labelled protein standards. 79

Austin J. Cooney et al. Protocol 6. Continued 18. Electrophorese for 2-3.5 h at 35 mA (6). Remove the portion of the gel at the dye front, which contains the majority of the radioactivity in the digested nucleic acids. 19. Fix the gel with 7.5% acetic acid and 50% methanol for 60 min. Then impregnate the gel with a fluor, to enhance detection, according to the manufacturer's instructions. 20. Dry the gel for 1 h at 60°C and then 1 h at 80°C. Then autoradiograph with Kodak XAR-5 X-ray film and a Dupont Cronex Lightening Plus intensifying screen. Expose the film at —70°C for 1-4 days to visualize the cross-linked protein.

Figure 4. Experimental set-up for UV irradiation of binding reactions.

4.3 UV cross-linking of transcription factors to nonsubstituted probes DNA can be cross-linked to a bound transcription factor in the absence of substitution with bromodeoxyuridine, however, at lower efficiency (1). In this case cross-linking occurs directly between the DNA bases and the polypeptide. In the protocol we describe here, an end-labelled oligonucleotide is UV cross-linked in situ as the final step in a gel retardation assay. A short oligonucleotide with overhangs should be chosen, to minimize the effects of UV light on the electrophoretic mobility of the DNA, because the probe will not be digested with nucleases, and to allow easy radiolabelling by the Klenow frag80

3: In vitro transcription ment of DNA polymerase 1 by a fill-in reaction. The advantage of this procedure is that labelling an oligonucleotide obviates the time-consuming subcloning and probe preparation in M13. The gel retardation assay is a very useful analytical tool for identifying and studying the binding of specific factors. It can also be used as a preparative procedure. Gel retardation both identifies and separates a factor from the milieu of crude extracts or partially purified fractions. Use is made of these dual properties to determine the molecular weight of a factor which binds to a specific element. In order to increase the success of cross-linking a transcription factor to DNA, the binding sequence must be A:T rich, as thymidine is the most reactive of the bases. Also certain response elements bind multiple proteins and the gel retardation assay allows individual factor DNA complexes to be visualized and isolated, and then analysed by SDS-PAGE. Protocol 7. UV cross-linking of non-substituted DNA Equipment and reagents [a32P]dCTP Klenow fragment 10 x Klenow buffer (Protocol 6) TE buffer pH 7.5 10mM Tris-HCI pH 7.5, 1mM EDTA 0.5 M EDTA Phenol:trichloromethane

Sodium acetate pH 5.2 Dry ice/ethanol bath Microcentrifuge 1 M ammonium acetate Electrophoresis set-ups as in Protocol 6 dNTP mix (Protocol 6)

Method 1. To end-label the oligonucleotide, make the following additions: 100 ng of double-stranded oligonucleotide 100 uCi of [a32P]dCTP 2.5 ul dNTP mix 5 ul 10 x Klenow buffer H2O to a final volume of 50 ul Incubate at room temperature for 60 min. 2. Stop the reaction by adding 4 ul 0.5 M EDTA, bring the volume up to 100 ul with TE buffer (pH7.5) and extract with an equal volume of phenol:trichloromethane:iso-pentanol (25:24:1, v:v:v). Aspirate the aqueous phase. 3. Add one-tenth of a volume of sodium acetate (pH 5.2) and 3 volumes of ethanol. Incubate in a dry ice/ethanol bath for 30 min. 4. Microcentrifuge at 11000 x g at 4°C for 40 min and discard the radioactive supernatant carefully. 5. Resuspend the pellet in 100 ul 1 M ammonium acetate, vortex vigorously, add 250 ul ethanol and incubate in a dry ice/ethanol bath for 30 min. 6. Microcentrifuge for 40 min at 11000 X g at 4°C and aspirate the supernatant. 81

Austin J. Cooney et al. Protocol 7.

Continued

7. Repeat steps 5 and 6. 8. Wash the pellet with 0.5 ml 80% ethanol, dry the pellet and resuspend it in 50 ul TE buffer (pH 7.5). 9. Determine the specific activity of the probe (see Protocol 6, step 11). 10. Incubate the partially purified transcription factor with the probe under binding conditions determined previously (see Protocol 6, step 12). 11. Cross-link the probe and transcription factor. The UV light wavelength used should be 254 nm. The duration of irradiation can be varied from 5 min to over 3 h, as this reaction is less efficient. 12. Electrophorese each reaction in a native 5% polyacrylamide gel to separate the protein-DNA complexes from the unbound probe. 13. Autoradiograph the gel for 1-3 h at 4°C with Kodak XAR-5 X-ray film. 14. Excise the retarded protein-DNA complexes. 15. Layer the gel slices on to a 10% polyacrylamide-SDS protein gel with a stacking gel. Load a lane of l4C-labelled molecular weight standards. 16. Electrophorese and autoradiograph the gel (as in Protocol 6, steps 17-20).

4.4 Renaturation of transcription factors from SDSpolyacrylamide gels Fractions containing partially purified transcription factors can be separated on the basis of size, with high resolution, by SDS-PAGE. Proteins eluted and renatured from the gel can be assayed for specific DNA binding and/or transcriptional activity. This technique allows accurate determination of the molecular weight of individual factors that bind to an element. However, this is dependent on the factor being able to bind to DNA as a monomer or a homodimer. If the active form is a heterodimer, specific DNA binding or transcriptional activity may not be detected by this procedure. The most critical parameter of this procedure is the ability of a transcription factor to renature into its active form. Many transcription factors are composed of domains with distinct structural conformations, which aids the process of renaturation. However, the likelihood of successful renaturation increases with decreased size. The recovery of active transcription factor after this procedure is very low. On average, a 5% recovery is assumed (this will vary from factor to factor, and also with size) to calculate the amount of the factor that has to be loaded on a gel to yield enough active factor after renaturation to be detectable by in vitro transcription or gel mobility shift assays. The method of Hager and Burgess (29), with a few modifications, is recommended and has been successfully employed to determine the molecular size of transcription factors (30). 82

3: In vitro transcription Protocol 8. Renaturation from SDS-polyacrylamide gels Equipment and reagents • 5 x SDS loading buffer: 0.25 M Tris-HCI pH 6.8,15% SDS, 50% glycerol, 0.025% bromophenol blue; before boiling add to each sample B-mercaptoethanol to 5% • SDS-PAGE apparatus and silver staining reagents

• • • •

Small dialysis bag (Mr cut-off 10000 Da) BSA, nuclease free (Boehringer-Mannheim) Transcription buffer (Protocol 6) 6 M guanidine-hydrochloride in transcription buffer

Method 1. Add one-fifth volume of 5 x SDS loading buffer to an aliquot of up to 100 ul of partially purified or affinity-purified transcription factor and boil for 2 min. 2. Load the supernatant on to a denaturing SDS-polyacrylamide gel (7.5-10%, depending on protein size) with a stacking gel (17). 3. Electrophorese for 3-5 h at 30 mA (17). 4. Remove the molecular weight markers and a portion of the sample lane and silver stain (21). 5. Cut the unstained sample lane into 5-10 mm horizontal strips from the bottom to the top of the gel. 6. Place each gel strip in an individual small dialysis bag (Mr cut-off 10000 Da) containing 0.5 ml of SDS-gel running buffer and 100 ug/ml BSA. Place the dialysis bags perpendicular to the electric current in a large horizontal gel trough, filled with SDS-gel running buffer. Apply the current at 50 mA for 3 h to elute the proteins from the gel. 7. Recover the buffer containing the electroeluted proteins from the dialysis bags and precipitate by the addition of acetone (4:1 v/v) in 15 ml Corex tubes and incubate at -70°C for 30 min or -20°C overnight. Centrifuge the precipitate at 10000 x g for 30 min at 4°C. This step concentrates the protein and removes the SDS. 8. Rinse the pellet once with 80% acetone and 20% transcription buffer and dry. 9. Dissolve the pellets in 100 ul of 6 M guanidine-hydrochloride in transcription buffer and incubate at room temperature for 20 min. This treatment denatures the polypeptides. 10. Place the resulting solutions in individual small dialysis bags and dialyse against transcription buffer for 16 h at 4°C, with one change of buffer during that time. 11. Assay 5-10 ul of the dialysed samples for specific DNA binding by gel mobility shift assay or DNAase I footprinting and/or transcriptional activity. 83

Austin J. Cooney et al.

5. Analysis of the monomer/dimer structure of the DNA binding forms of a transcription factor Another characteristic of transcription factors requiring examination is the subunit composition of the active DNA binding form. Many transcription factors exist as monomers, dimers, or even trimers as their active DNA binding forms. At this stage of characterization of a transcription factor there should be substantial evidence to indicate its subunit composition. One indicator is the sequence of the DNA response element itself. Symmetry within the sequence of the response element, whether it is palindromic or directly repeated, suggests that the DNA binding form of the transcription factor may be a dimer or a higher order structure. Symmetric response elements are the binding elements for many groups of transcription factors, including helixloop-helix factors and the steroid/thyroid hormone receptor superfamily. Both families bind to their cognate response elements as dimers. DNAase I protection and methylation interference analysis will provide evidence as to whether a factor protects and contacts the repeated sequence motif. The response element will then indicate the subunit structure of the transcription factor. Comparison of the molecular weights of a transcription factor determined under both native and denaturing conditions will provide more direct evidence of the subunit structure of the DNA binding form. If a factor binds as a monomer, its molecular weight determined under native and denaturing conditions will be the same. If the factor binds as a dimer, its native molecular weight will be approximately double the molecular weight determined under denaturing conditions, and so on for higher order structures. If the native molecular weight is not an exact whole-number multiple of the denatured molecular weight, this may indicate that the DNA binding form is a heterodimer of two subunits with different molecular sizes. Heterodimers have been observed for many transcription factors. With this preliminary information, more direct experiments are required to determine the subunit structure of the transcription factors. These experiments involve analysis by gel retardation and chemical cross-linking. In addition, cooperative DNA binding can be observed between dimers bound to adjacent response elements. This cooperativity can be analysed by gel mobility shift assays.

5.1 Analysis of transcription factors by DNA mobility shift assay The tertiary structure of transcription factors can be analysed indirectly by gel retardation. The use of double-stranded oligonucleotides as probes allows easy and rapid analysis of factor-binding requirements by incorporation of base mutations into the sequence during synthesis. Methylation interference will often highlight the symmetric nature of the purine contact points. These nucleotides are good points to start mutational analysis, by changing them to 84

3: In vitro transcription pyrimidines and analysing their effect on factor binding by gel retardation. In this fashion, mutation of certain residues of a half site of the symmetric response element may lead to loss of binding activity. This suggests that the factor binds as a dimer, requiring binding of both subunits. To provide absolute proof that the active DNA binding form is a dimer, the factor needs to be cloned. The cloned gene permits genetic manipulations, including deletion analysis and formation of fusion proteins, which allows analysis of structure-function relationships, such as dimer formation. The subunit structure of a transcription factor can be analysed by attempting to form heterodimers between the full-length factor and either truncated (31) or extended (32) forms of the factor that retain the capability of binding DNA. These experiments can be performed using an in vitro transcription/translation system (Promega) to synthesize the factor and these mutants. By mixing the translation products or cotranslating their mRNA, the subunit composition can be assayed by gel retardation (33). If a factor binds as a monomer, two retarded complexes will be formed, corresponding to the full-length factor and either a truncated or an extended form. However, if the factor forms a dimer, then three retarded complexes will be observed. There will be two homodimers and a heterodimer of intermediate mobility, which contains one molecule of full-length factor and a molecule either of the truncated or extended factor.

5.2 Analysis of transcription factor subunits by chemical cross-linking The monomer/dimer structure of a factor and interactions between transcription factors can be analysed directly by chemical cross-linking. If a factor exists as a dimer, cross-linking of the two subunits will double its molecular weight as observed under denaturing conditions such as SDS-PAGE. Cross linking compounds react with chemical groups in proteins, including those found in the amino acid backbone and those found in the side chains. They form a covalent bond between two reactive groups and thus stably link them. Owing to the small size of these compounds, the cross-linked chemical groups must be in close proximity, inferring a direct interaction between their cognate polypeptides. Thus cross-linking is a sensitive method for directly assaying protein-protein interactions, such as dimer formation, and other interactions. There are many different chemical cross-linkers that are characterized and grouped according to their cross-linking properties (34). Each compound of this extensive array of cross-linkers has distinctive chemical characteristics and is used to cross-link different chemical groups separated by varying distances. They include homobifunctional and heterobifunctional compounds, some of which are cleavable or reversible. Extensive review of these cross-linking reagents is beyond the scope of this chapter. The Pierce catalogue contains an extensive list of cross-linking 85

Austin J. Cooney et al. reagents and protocols for their use. However, we will give an example of this powerful technique. The choice of a particular cross-linker depends on the groups that are to be cross-linked. However, with uncharacterized transcription factors, the information to make the best choice is not available, thus more general cross-linkers are used. The best results are achieved by trial and error. For example, Aranyi et al. (35) used several related compounds, bisimidates (e.g. dimethyl succinimidate, dimethyl adipimidate, and dimethylpimelidate) to cross-link the progesterone receptor (PR) to show that it exists as a dimer in solution. However, only dimethylpimelidate (DMP) successfully cross-linked the PR. DMP (Sigma) is a homobifunctional imidoester which forms a covalent bond between opposing primary amines. Protocol 9. Cross-linking transcription factors with DMP Equipment and reagents • 2.2 M triethanolamine pH 8 . 200 mM ethanolamine * 100 mM DMP/HCI: 100 mM DMP/HCI in 200 mM ethanolamine pH 8, adjusted with NaOH . 2 x SDS loading buffer

• TESH buffer: 10 mM Tris, 1 mM Tris, 1 mM EDTA, 12 mM 1-thioglycerol pH 7.6 • 7.5% polyacrylamide-SDS gel with stacking gel • Nitrocellulose or Zeta-Probe membrane

Method 1. Make the following additions:

2. 3. 4. 5. 6. 7.

0.1 volumes of 2.2 M triethanolamine pH 8 0.2 volumes of 100 mM DMP/HCI 100-200 ng of transcription factor Make up the volume to 30 ul with TESH buffer. Incubate the reaction for 30 min at room temperature. Stop the reaction with 0.1 volumes 200 mM ethanolamine. Add an equal volume of 2 x SDS loading buffer and boil the sample for 2 min. Run the sample on a 7.5% polyacrylamide-SDS gel with a stacking gel. Transfer the proteins to nitrocellulose or Zeta-Probe membrane. Detect the cross-linked factors by Western immunoblot analysis (see Protocol 5).

5.3 Analysis of cooperative binding between dimers bound to adjacent response elements Often enhancer elements act synergistically to enhance transcription. This phenomenon has been studied extensively using transient transfections and in 86

3: In vitro transcription vitro transcription. As an example, tandemly linked progesterone-response elements (PREs) upstream of a basal heterologous promoter fusion gene (TK-CAT) confer synergistic progesterone inducibility when compared with induction from single PREs (36). This observed synergism could be due to cooperative binding of two PR dimers to adjacent elements. To determine if PR dimers bind cooperatively to tandem PREs, varying concentrations of PR DNA binding domains were assayed by gel retardation and two retarded complexes were observed, II and IV (see Figure 5). Methylation interference showed they were formed by binding of a single dimer and two dimers to the probe, respectively. At low concentrations of receptor, only complex II was detected; however, at increasing concentrations complex IV was detected. With increasing receptor concentrations, the formation of complex IV was progressively enhanced to become the predominant protein-DNA complex. The levels of complexes II and IV and free DNA can be quantitated using a betascope 603 blot analyser (Betagen, Waltham, MA), phosphor imaging, or excision of the respective bands and scintillation counting. Quantitation showed that complex II reached approximately 10% of bound DNA, but no higher. In contrast, complex IV is present only at higher concentrations of receptor, where it rapidly becomes the predominant form. Analysis of the binding indicates that cooperative DNA binding displays sigmoidal kinetics (see Figure 5) (36). If no cooperativity is observed between the binding of the dimers, then complex II will reach 50% before the appearance of complex IV.

Figure 5. Quantitation of the two mobility shift complexes II and IV at different receptor concentrations and plotted as the percentage of DNA bound in complexes II and IV.

87

Austin J. Cooney et al.

6. Identification and initial characterization of non-DNA binding transcription factors Interaction of a transcription factor with itself or other transcription factors to form DNA binding dimers is relatively easy to detect and analyse. In contrast, the interaction of a transcription factor with a non-DNA binding transcription factor is not as easy to detect. However, techniques have been developed to study these interactions. The requirement for a non-DNA binding transcription factor is generally discovered during the purification of the DNA binding transcription factor by a loss of transcriptional activity as the two factors are separated. Transcriptional activity can be restored by mixing two different fractions. In contrast to a requirement for an additional DNA binding factor, the other fraction required for transcription activity will not bind DNA specifically, as determined by gel mobility shift assays and DNAase I footprinting. The interactions between DNA binding and non-DNA binding transcription factors can be either strong or weak and necessitate different methods of analysis depending on their strength. Strong interactions are stable under the conditions of native electrophoresis and can be studied by gel mobility shift analysis and also by DNAase I footprinting. Weak interactions are unstable and fall apart under the conditions employed for native gel electrophoresis and thus necessitate less stringent assays. These methods include immunoprecipitation, chemical cross-linking, or kinetic analysis. The requirement for a non-DNA binding factor is generally discovered fortuitously during the purification of a DNA binding factor. There may be a significant or total loss of transcriptional activity as the DNA binding transcription factor is progressively purified and assayed. An authentic loss of activity indicates the requirement for an additional transcription factor in addition to the DNA binding factor. DNAase I footprinting and gel retardation can be employed to determine if it is capable of binding specifically to DNA. The requirement of a DNA binding transcription factor for a non-DNA binding partner for transcriptional activity is not predictable. These requirements are peculiar and individual to each pair studied, thus it is difficult to describe protocols for the initial identification and analysis of non-DNA binding transcription factors. The first indications of a requirement for a non-DNA binding transcription factor may not be as dramatic as a total loss of activity, it may first manifest itself as a non-alignment of DNA binding and transcriptional activities in chromatographic fractions, as was observed for COUP-TF (a specific DNA binding factor) and S300-II (a non-DNA binding transcription factor) when chromatographed over a Sepharose S300 column (1-38). The two activities (DNA binding and stimulation of transcription) overlapped, but the peaks of each activity were not coincident. Gel mobility shift assays and DNAase I 88

3: In vitro transcription footprinting can be used to determine whether one or both peaks contain a specific DNA binding activity. Where a requirement for a non-DNA binding factor can be established, additional chromatographic steps can be used to further separate the two activities, leading to a total loss of transcriptional activity from the specific DNA binding activity. COUP-TF and S300-II fractions containing the DNA binding and non-DNA binding activities were re-chromatographed on a smaller 33 ml Sephacryl S-300 column to obtain a better resolution and separation of the two activities (see Figure 6). The fractions collected from this column, containing the COUP-specific DNA binding activity, were incapable of supporting transcriptional activity in the reconstituted system. However, when a fraction containing the non-DNA binding factor, i.e. No 57, was added to these assays, a peak of transcriptional activity was detected around fractions 41 and 42 (see Figure 6), which corresponded to the peak of DNA binding activity. Fraction 57 was chosen to avoid COUPTF contamination, as it was located on the side of the peak distal to the DNA binding activity. Similarly, the S300-II fractions supported only a low level of transcriptional activity; addition of fraction 41, containing COUP-specific DNA binding activity most distal from the non-DNA binding activity required for transcription, elicited a dramatic increase in transcriptional activity. Thus, the second Sephacryl S-300 chromatographic step achieved separation of the specific binding and non-DNA binding transcription factors and clearly demonstrated that transcriptional activity was dependent on both factors. Complementation of chromatographic fractions to the DNA binding activity will lead to the restoration and identification of the fractions containing the non-DNA binding activity. This outline highlights the approaches used to identify and study non-DNA binding transcription factors. Such an initial characterization of the requirement for a non-DNA binding factor establishes

Figure 6. Dependence of the ovalbumin promoter expression on two factors, COUP-TF I (peak I) and S300-II (peak II). COUP-TF is a DNA binding transcription factor, S300-II is TFIIB, which does not bind directly to the ovalbumin promoter.

89

Austin J. Cooney et al. a reconstituted assay system for the non-DNA binding factor, which will be required for subsequent purification and analysis of its function.

6.1 Assay of direct interactions between DNA binding and non-DNA binding transcription factors Having established a functional requirement of a non-DNA binding factor for the transcriptional activity of a DNA binding factor, the next step is to analyse this functional interaction. As it does not bind the response element, the nonDNA binding transcription factor must interact directly, in some fashion, with the factor bound at this site. It may also interact with the general transcriptional machinery, by bridging or adapting the binding of the DNA binding transcription factor with a functional response from the RNA polymerase II transcriptional machinery. Thus, the interactions of the DNA binding and nonDNA binding transcription factors need to be studied. Interactions between factors can be analysed by such methods as co-immunoprecipitation, chemical cross-linking, or using 'supershift' gel mobility shift assays, or kinetic analysis of transcription factor and DNA binding stability. Generally, more than one technique is employed in the analysis of direct interactions, and the particular choices are guided by the strength of the observed interaction.

6.2 'Supershift' gel mobility shift analysis If the interaction between the DNA binding transcription factor and the nonDNA binding transcription factor is sufficiently strong, it can be detected indirectly in a conventional gel mobility shift assay. The non-DNA binding transcription factor, by its nature, will be incapable of binding to the probe, but it will bind to the DNA binding transcription factor bound to the probe, further retarding the migration of the complex during electrophoresis. This 'supershift' is similar in nature to that observed upon addition of specific antibodies. The conditions for the 'supershift' have to be determined empirically for each pair of DNA and non-DNA binding transcription factors; however, the interaction generally occurs under the transcription conditions used to assay their functional interaction (37). Protocol 10. 'Supershift' gel mobility shift assay Equipment and reagents • Hinfl-digested pBR322 or poly(dl-dC)poly(dl-dC) . Native 4 or 5% polyacrylamide gel • 0.5 X TBE

• Whatman 3MM paper . Kodak XAR-5 X-ray film . Dupont Cronex Lightening Plus intensifying screens

Method 1. Mix the partially purified DNA binding and non-DNA binding transcription factors with the labelled DNA fragment or oligonucleotide with 90

3: In vitro transcription non-specific competitor (0.5 ug Hinfl-digested pBR322 or 1 ug poly(dldC)-poly(dl-dC) under transcription conditions determined previously. 2. Incubate at room temperature for 15-30 min. 3. Load the samples on to a native 4 or 5% polyacrylamide gel (the acrylamide percentage depends on the probe size). 4. Electrophorese for 1.5-3 h at 160-180 V in 0.5 x TBE. 5. Transfer the gel to Whatman 3MM paper and dry the gel at 80°C for 30 min. 6. Autoradiograph the dried gel with Kodak XAR-5 X-ray film with two Dupont Cronex Lightening Plus intensifying screens at-70°C.

6.3 Analysis of the dissociation of transcription factor-DNA complexes in the presence and absence of the non-DNA binding transcription factor If the interaction between the specific DNA binding and the non-DNA binding transcription factors is insufficiently strong, they will not be stable enough to detect as a 'supershift' in a gel retardation assay. However, non-DNA binding factors affecting the DNA binding stability can be assayed in a gel retardation assay by analysing the stabilization of the binding of the DNA binding transcription factor to its response element. Stabilization of binding by the non-DNA binding transcription factor can be determined and quantified by analysing the 'off rate' of the DNA binding transcription factor after it has bound to the response element. The 'off rate' is determined by challenging the transcription factor-labelled DNA complex with unlabelled response element in a kinetic experiment (38). The amount of unlabelled DNA added must be sufficient to eliminate completely the gel mobility shift if co-incubated with the probe and factor. Thus, the amount used will vary from factor to factor and will depend on its DNA binding affinity. The amounts of transcription factor and labelled response element used in each reaction in this experiment should be similar to those used to detect the transcription factor by gel mobility shift assay and should be incubated under similar conditions. Protocol 11. Analysis of the dissociation of transcription factor-DNA complexes in the presence and absence of non-DNA binding transcription factor Equipment and reagents • Transcription factor, partially or affinity purified • 32P-labelled DNA fragment or oligonucleotide • Specific competitor DNA fragment or oligonucleotide, unlabelled

91

• BSA • Native electrophoresis gel, 4 or 5% polyacrylamide • Densitometer (or phosphor imaging or beta scanner)

Austin J. Cooney et al. Protocol 11.

Continued

Method 1. Mix the partially or affinity-purified transcription factor with 32P-labelled DNA fragment or oligonucleotide under conditions determined previously for a gel mobility shift assay (Chapter 1). Carry out one reaction for each time point. The reaction at each time point should be duplicated using an equivalent amount of BSA instead of non-DNA binding factor as a control for non-specific stabilization. (Non-specific stabilization is sometimes observed when a highly purified factor is mixed with other fractions or proteins.) 2. Incubate at room temperature for 10 min to allow the transcription factor to bind to the response element. 3. Add a 50- to 200-fold molar excess (depending on the DNA binding affinity of the individual factor) of unlabelled specific competitor DNA fragment or oligonucleotide to each reaction after 10 min. 4. Terminate each reaction at specified time points by loading on to the gel and applying a very low voltage (this separates the bound and unbound probe immediately and stops the binding reaction). The addition of unlabelled competitor and the termination of the reaction should be timed accurately and spaced between reactions to produce an accurate and reproducible time curve. Typical kinetic time points may include 0, 2, 4, 8, 12, 16, 24, 32, and 60 min. 5. Analyse the reactions by native electrophoresis in a 4% or 5% polyacrylamide gel (depending on the probe size) and autoradiography as in a typical gel mobility shift assay. 6. Scan the autoradiograph with a densitometer, or scan the gel by phosphor imaging or with a beta scanner to quantitate the radioactivity in the retarded factor-DNA complexes. 7. Plot the amount of undissociated factor-DNA complex versus time to determine the tl/2 of the dissociation of the complexes in the presence and absence of the non-DNA binding transcription factor (see Figure 5 for example).

6.4 Analysis of direct interactions by coimmunoprecipitation Co-immunoprecipitation is a powerful technique that can detect direct interactions between two factors (39). It relies on the specificity of antibodies to recognize one of the components and protein A-Sepharose to precipitate the cross-linked immunocomplexes. Protein A-Sepharose is commercially available (Pharmacia). If the interaction between two factors is sufficiently strong, immunoprecipitation of one will bring down the other. However, if this inter92

3: In vitro transcription action is not sufficiently strong, it can be stabilized by using a reversible chemical cross-linker such as dithiobis(succinimylpropionate) (DSP) (35, 40). It is a homobifunctional cross-linking reagent which reacts with primary amine groups. It can be cleaved with excess thiol and is thus reversible under denaturing SDS-PAGE conditions in the presence of B-mercaptoethanol. Thus, SDSPAGE will separate the cross-linked species into their individual components. This procedure is especially useful if the factors have been cloned and can be labelled with [35S]methionine by in vitro transcription/translation, or if antibodies are available. A drawback of co-immunoprecipitation is that non-specific entrapment of proteins can occur, especially when using polyclonal antibodies, which recognize many antigens, leading to highly cross-linked aggregates. Controls are required to distinguish between specific co-immunoprecipitation and nonspecific entrapment, such as use of another factor which is known not to react with the immunoprecipitation target. The immunoprecipitated pellets are washed several times with buffer to remove non-specific proteins, but the stringency of these washes can be increased by adding small amounts of non-ionic detergents (e.g. 0.5% NP40 or 0.1% Triton X-100). In order to identify the co-precipitating factor, an antibody raised against it is required for Western immunoblot analysis, unless a clear identification can be made on the basis of molecular size, or the cloned protein can be labelled with [35S]methionine during in vitro translation of its mRNA. Protocol 12. Detection of direct interactions between factors by co-immunoprecipitation Equipment and reagents Partially purified transcription factors Primary antibody: mouse monoclonal or rabbit polyclonal Secondary antibody: rabbit anti-mouse IgG Protein A-Sepharose TE buffer pH 7.5

• Microcentrifuge • Wash buffer: 20 mM Tris-HCI pH 7.4, 100 mM NaCI • 1 x SDS sample buffer • 7.5-10% polyacrylamide-SDS gel with stacking gel

Method 1. Pre-incubate the partially purified DNA binding and non-DNA binding transcription factors in a 100 ul reaction volume under transcription conditions at a concentration of approximately 5 ug/ml for 1 h at 4°C. 2. Add 5 ug primary antibody (mouse monoclonal) or rabbit polyclonal antibody to the reaction. The final antibody concentration is approximately 50 ug/ml, but will depend on the individual affinity and titre of antibodies used. 3. Incubate for 2 h at 4°C with gentle agitation. 4. If a mouse monoclonal antibody has been used, add the secondary

93

Austin J. Cooney et al. Protocol 12.

Continued

antibody to a final concentration of 100-200 ug/ml. Incubate for 2 h at 4°C with gentle agitation. If using a rabbit polyclonal primary antibody, go directly to step 5. 5. Add 100 uJ protein A-Sepharose 1:1 suspension in TE buffer pH 7.5. Incubate for 1 h at 4°C with gentle agitation. 6. Microcentrifuge at 11 000 X g for 5 min at 4°C. 7. Carefully aspirate the supernatant and wash the pellet three times with 1 ml wash buffer, to remove non-specifically bound proteins. 8. Resuspend the pellet in 50 ul 1 X SDS sample buffer, boil for 2 min, load on to a polyacrylamide-SDS gel (7.5-10%, depending on protein size) with a stacking gel and electrophorese (17). 9. The co-immunoprecipitation of the non-DNA binding factor can be detected by Western transfer and immunoblot analysis using an antibody raised to the non-DNA binding factor or by autoradiography if it has been labelled with [35S]methionine.

Protocol 13. Stabilization of weak interactions with DSP Equipment and reagents • Transcription buffer • DSPa

• Ethanolamine • 3 x SDS-PAGE buffer, B-mercaptoethanol

Method 1. Incubate the transcription factors in transcription buffer at 37°C for 30 min in a total volume of 50 ul, to allow association. 2. Add DSP to a final concentration of 2 mM. 3. Incubate at room temperature for 1 h. 4. Add ethanolamine to a final concentration of 0.1 M, to stop the crosslinking reaction. 5. Immunoprecipitate the cross-linked species (see Protocol 11). 6. Reduce and reverse the cross-linked immune complexes by boiling in 3 x SDS-PAGE buffer containing 15% B-mercaptoethanol. 7. Analyse by SDS-PAGE (6) and fluorography, if a [35S]methionine labelled factor has been used, or by Western analysis. •Buffers used with DSP should be free of Tris as it contains primary amino groups which will quench the DSP reaction.

94

3: In vitro transcription

References 1. Sagami, I., Tsai, S. Y., Wang, H., Tsai, M.-J., and O'Malley, B. W. (1986). Mol. Cell. Biol. 6, 4259. 2. Ing, N. H., Beekman, J. M., Tsai, S. Y., Tsai, M.-J., and O'Malley, B. W. (1992). J. Biol. Chem. 267, 17617. 3. Corthesy, B., Hipsking, R., Theulaz, I, and Wahli, W. (1988). Science 239, 1137. 4. Freedman, L. P., Yoshinaga, S. K., Vanderbilt, J. N., and Yamamoto, K. R. (1989). Science 245, 298. 5. Klein-Hitpass, L., Tsai, S. Y., Weigel, N. L., Allan, G. F., Riley, D., Rodriguez, R., et al. (1990). Cell 60, 247. 6. Bagchi, M. K., Tsai, S. Y., Tsai, M.-J., and O'Malley, B. W. (1990). Nature 345, 547. 7. Elliston, J. F., Fawell, S. E., Klein-Hitpass, L., Tsai, S. Y., Tsai, M.-J., Parker, M. G., and O'Malley, B. W. (1990). Mol. Cell. Biol. 10, 6607. 8. Klaff, M., Gross, B., and Beato, M. (1990). Nature 344, 360. 9. Bagchi, M. K., Tsai, S. Y., Weigel, N. L., Tsai, M.-J., and O'Malley, B. W. (1990). J. Biol. Chem. 265, 5129. 10. Sawadogo, M. and Roeder, R. G. (1985). Cell 43, 165. 11. Sawadogo, M. and Roeder, R. G. (1985). Proc. Natl Acad. Sci. USA 82, 4394. 12. Dignam, J. D., Lebovitz, R. M., and Roeder, R. G. (1983). Nucl. Acids Res. 11, 1475. 13. Andrews, P. (1970). In Methods of biochemical analysis, Vol. 18 (ed. D. Glick), pp. 1-53. Interscience, New York. 14. Laue, T. M. and Rhodes, D. G. (1990). In Methods in enzymology, Vol. 182 (ed. M. P. Deutscher), pp. 566-87. Academic Press, London. 15. Fischer, L. (1980). Gel-filtration chromatography. Elsevier, Amsterdam. 16. Stellwagen, E. (1990). In Methods in enzymology, Vol. 182 (ed. M. P. Deutscher), pp. 317-28. Academic Press, London. 17. Laemmli, U. K. and Favre, M. (1973). J. Mol. Biol. 80, 575. 18. Martin, R. and Ames, B. (1961). J. Biol. Chem. 236, 1372. 19. Siegel, L. M. and Monty, K. J. (1966). Biochim. Biophys. Acta 112, 346. 20. Chodosh, L. A., Baldwin, A. S., Carthew, R. W., and Sharp, P. A. (1988). Cell 53, 11. 21. Merril, C. R. (1990). In Methods in enzymology, Vol. 182 (ed. M. P. Deutscher), pp. 477-88. Academic Press, London. 22. Rodriguez, R., Weigel, N. L., O'Malley, B. W., and Schrader, W. T. (1990). Mol. Endocrinol. 4, 1782. 23. Slater, G. G. (1968). Anal. Biochem. 24, 215. 24. Timmons, T. M. and Dunbar, B. S. (1990). In Methods in enzymology, Vol. 182 (ed. M. P. Deutscher), pp. 679-88. Academic Press, London. 25. Garfin, D. E. (1990). In Methods in enzymology, Vol. 182 (ed. M. P. Deutscher), pp. 425-41. Academic Press, London. 26. Weber, K. and Osborn, M. (1969). J. Biol. Chem. 244, 4406. 27. Chodosh, L. A., Carthew, R. W., and Sharp, P. A. (1986). Mol. Cell Biol. 6, 4723. 28. Lin, S.-Y. and Riggs, A. D. (1974). Proc. Natl Acad. Sci. USA 71, 947. 29. Hager, D. A. and Burgess, R. R. (1980). Anal. Biochem. 109, 76. 30. Bagchi, M. K., Tsai, S. Y., Tsai, M.-J., and O'Malley, B. W. (1987). Mol. Cell. Biol. 7, 4151. 95

Austin J. Cooney et al. 31. Kumar, V. and Chambon, P. (1988). Cell 55, 145. 32. Tsai, S. Y., Carlstedt-Duke, J., Weigel, N. L., Dahlman, K., Gustafsson, J.-A., Tsai, M. J., and O'Malley, B. W. (1988). Cell 55, 361. 33. Ladias, J. A. A. and Karathanasis, S. K. (1991). Science 251, 561. 34. Pierce catalogue and references therein. 35. Aranyi, P., Radanyi, C., Renoir, M., Devin, J., and Baulieu, E.-E. (1988). Biochemistry 27, 1330. 36. Tsai, S. Y., Tsai, M.-J., and O'Malley, B. W. (1989). Cell 57, 443. 37. Buratowski, S., Hahn, S., Guarente, L., and Sharp, P. A. (1989). Cell 56, 549. 38. Tsai, S. Y., Sagami, I., Wang, H., Tsai, M.-J., and O'Malley, B. W. (1987). Cell 50, 701. 39. Lee, W. S., Kao, C. C., Bryant, G. O., Liu, X., and Berk, A. J. (1991). Cell 67, 365. 40. Yang-Yen, I. I.-F., Chambard, J.-C., Sun, Y.-L., Smeal, T., Schmidt, T. J., Drouin, J., and Karin, M. (1990). Cell 62, 1205.

96

4

Purification and cloning of DNA binding transcription factors R. H. NICOLAS, G. HYNES, and G. H. GOODWIN

1. Introduction If you have just identified a protein binding to a novel regulatory sequence, the next step is to clone the cDNA encoding the protein. Most practitioners in this field would first attempt direct screening of expression libraries with the DNA binding site (see Chapter 5) before embarking on the protein purification/peptide sequence/oligonucleotide screening approach. However, in many cases, direct screening of libraries has not proved successful or has not been possible (for example when the protein is part of a multiprotein complex and more than one protein is required for DNA binding) and protein purification may then have been the only method available. The purification of DNA binding proteins has been remarkably successful with the development of sequence-specific DNA affinity chromatography and the band-shift technique for rapidly analysing chromatography fractions. With the recent advances in protein-sequencing techniques (e.g. sequencing peptides immobilized on blots following electrophoretic separation), the protein purification approach has become less arduous and one can expect to get good sequence results from 50 pmol of protein. Also, mass spectroscopy techniques are being used to mass fingerprint proteins for the identification of protein sequences in databases, and tandem mass spectroscopy is now able to obtain peptide sequences from femtomole amounts of protein. One of the major difficulties faced by the protein purifier is the abundance of the transcription factor in the cell. Eukaryotic transcription factors are present in the cell nucleus in widely varying quantities, ranging from 103 molecules per cell for some retinoic acid receptors to well over 105 per cell for the GATA1 factor in erythroid cells. If at least 100 pmol of pure protein are required for production of peptides for sequencing, it is apparent that for a protein of average abundance at 104 molecules per cell, more than 1011 cells of starting material are required, assuming 5% recovery of protein. The successful isolation of protein also requires that the molecular weight of the protein has been determined (for monitoring the final stages of purification

R. H. Nicolas et al. of the protein) and that a DNA sequence with high affinity and specificity for the protein has been identified. As a rough guide, the ratio of the strengths of binding to specific versus non-specific DNA sequences should be at least 100. This can be determined using competitive DNA binding analyses in the bandshift assay. The molecular weight of the protein can be determined by twodimensional electrophoretic techniques (1), by photoaffinity labelling (2), or by Southwestern blotting techniques (3). In this chapter, we describe the techniques used to purify moderately abundant factors, as illustrated by the purification of a zinc-finger factor termed CTCF, a ubiquitous protein which binds to a GC-rich element in the promoter region of the chicken c-myc gene (4), and chicken GATA1 (formerly termed EF1, NF, El, EryFl), a protein expressed in erythroid cells, megakaryocytes, and platelets (1, 5). This is outlined schematically in Figure 1. In these purifications, DNA-cellulose chromatography is carried out to isolate DNA binding proteins prior to DNA affinity chromatography. Other chromatographies commonly used to achieve partial purification before affinity chromatography include heparin-sulfate, phosphocellulose, or ion-exchange.

Figure 1. The steps required to clone a transcription factor via protein isolation.

98

4: Purification and cloning In our view, gel filtration rarely has sufficient resolution or capacity to make it worthwhile. Following an affinity chromatography step, final purification is achieved by further cycles of affinity chromatography or by reverse-phase chromatography.

2. Buffers and solutions When the aim of purifying a transcription factor is to obtain some primary sequence of the protein, it is important that all solutions should be scrupulously clean. Any glassware used should be acid washed by soaking overnight in nitric acid (one part nitric acid, two parts HPLC-grade water), then rinsed four times in HPLC-grade water and dried in an oven at 150°C. Sterile disposable plasticware is a useful alternative, and tissue-culture flasks make good reagent bottles. Only high-quality chemical reagents should be used, and stock solutions should be filtered through 0.22 um filters. Many of the buffers contain glycerol for protein stability, but it is very difficult to measure viscous reagents by volume, so glycerol is added by weight. It is also a good idea to make a glycerol-containing stock solution such as Ns (Table 1) which can be used to make all the buffers for DNA chromatography.

3. Preparation of nuclear extract Consideration should be given first to which tissue or cell line will give the greatest yield. In general, tissue culture is expensive, but relatively pure nuclei can be prepared in high yield in a couple of steps using just low-speed centrifugations. To prepare nuclei and obtain undegraded nuclear proteins from tissues, the starting material must be fresh. This may present difficulties when working with human tissue. Animal tissue may be more readily available, but it should be remembered that tissues do not contain a homogeneous cell population and only a few cell types may be expressing the protein of interest. Some tissues, such as thymus and spleen, may be gently disrupted in phosphatebuffered saline (PBS) and filtered through gauze to yield a crude cell suspension and then Protocol 1 may be followed to prepare a nuclear extract. With tougher tissues, it is better not to prepare cells but prepare pure nuclei directly by centrifugation through a high-sucrose gradient as described in Protocol 2. The preparation of nuclei gives between 10- and 100-fold enrichment of a nuclear-localized protein, depending on the size of the cells. The cytoplasmic supernatant fractions should be assayed for activity because in some cases weakly bound factors may be lost from the nuclei. Some transcription factors may be susceptible to denaturation by the detergent treatment used to lyse the cell membrane, in which case cells may alternatively be lysed by hypotonic shock (6). In many cases, lower concentrations of Triton are sufficient to lyse cells and this can be investigated for each cell type, but for optimum yields it should be ensured by microscopic examination that all the cells are lysed. 99

R. H. Nicolas et al. Table 1. Solutions (a) Stock solutions 200 mM Hepes-NaOH pH 7.9 200 mM EDTA pH 8.0 1 M Tris-HCI pH 7.6 1 M MgCI2 4MNaCI 2 M (NH4)2SO4 10% Triton X-100 (Pierce) 10% Brij-35(Pierce) Glycerol 1 M DTT (store in aliquots at -20oC and add to solutions on day of use) (b) Stock of protease and phosphatase inhibitors 50 mM PMSF (dissolve in propan-2-ol and store in a dark bottle at 4°C for up to 3 months, as long as it is kept free of water) 50 mm benzamidine in water 100 mg/ml leupeptin in water 100 mg/ml aprotinin in water 100 mg/ml pepstatin in methanol 200 mM levamisole in water 1 M B-glycerophosphate in water All these solutions are 100 x concentration and should be added to the pre-cooled solutions on the day of use. All except PMSF can be stored as aliquots at -20°C for up to 3 months or at 4°C for 1 week. (c) Solutions for preparation of nuclear extracts TM TTM PBS

N100 NDil

10 mM Tris-HCI pH 7.6, 5 mM MgCI2 10 mM Tris-HCI pH 7.6, 5 mM MgCI2, 0.5% Triton X-100 Dulbecco's phosphate-buffered saline. Dissolve 1 tablet (OXOID) in 100 ml or dissolve 0.2 g KCI, 0.2 g KH2P04, 8.0 g NaCI and 2.16 g Na2PO4.7H2O in 1 litre. The pH should be 7.3-7.4 See Table 1(d) See Table 1(d)

All the above solutions should have the inhibitors added just before use. (d) Buffers required for DNA chromatography Buffer

N100 N200 N1000 NDil E250 WOO/50b

Reagent required (ml) to make 100 ml of buffer Nsa Water 4 M NaCI 2 M (NH4)2SO4 1 M DTT

50 50 50 75 50 50

40.4 37.9 17.9 17.9 30.4 10.4

2.5 5.0 25 2.5

12.5 -

a

0.1 0.1 0.1 0.1 0.1 0.1

Inhibitors

7x 1 7 X1 7X1 7 X1 7 X1 7x 1

Ns = 40% v/v glycerol (= 50% w/v glycerol), 10 mM MgCI2, 0.2 mM EDTA, 40 mM Hepes (pH 7.9), 0.1% Brij-35. b Make up to volume with glycerol (37.5 g).

100

4: Purification and cloning Transcription factors are then extracted from the nuclei with a buffer containing 0.3 M NaCl. The highest concentration of salt that can be used without extracting histone H1 is 0.35 M NaCl. Extraction of H1 at higher salt concentrations results in chromatin decondensation, nuclear lysis, and formation of a gel that can only be effectively removed by ultracentrifugation. H1 also causes problems during the subsequent chromatography steps since it is very abundant and will bind to the DNA-cellulose and bind non-specifically to the DNA affinity columns. Although H1 can be removed by ion-exchange chromatography, it is better to avoid extracting the H1 in the initial nuclear extraction, provided the factor of interest is extracted in high yield by 0.30.35 M NaCl. Since the nuclear extraction buffer contains other components which contribute to the ionic strength, the NaCl should be added to only 0.3 M final concentration to minimize H1 solubilization. Following extraction of the nuclei, the NaCl concentration of the extract is reduced either by dilution or dialysis before loading on to the DNA-cellulose column. This causes precipitation of proteins, and some transcription factors may come out of solution at this point. Since the solubility is usually inversely proportional to NaCl concentration, the highest NaCl concentration that will still allow binding to the DNA-cellulose column should be determined. Conversely, if the factor remains soluble at low salt concentration, then by optimizing this precipitation a two- to three-fold purification can be achieved. The use of a non-ionic detergent such as Brij-35 does keep some factors soluble, but it may also be detrimental to other binding activities. Ammonium sulfate fractionation can be included before the DNA-cellulose chromatography. This has the advantage of concentrating the protein and removing some nucleic acid, but since the precipitate has to be redissolved in buffer and then dialysed against N100 buffer before loading on to the column, there is little saving in the time taken for the preparation. Ammonium sulfate precipitation can be detrimental; e.g. in the case of GATA1, some activity was lost due to denaturation or incomplete dissolution of the protein precipitate. Interestingly, when 3.5 g (NH4)2SO4 is added per 7.5 ml of 0.3 M nuclear extract, the majority of the proteins precipitate, leaving only the high mobility group (HMG) proteins in solution. Thus, after pelleting the ammonium sulfate precipitate by centrifugation (40000 X g, 30 min), the HMG proteins can simply be prepared in an undenatured form by extensively dialysing the supernatant. Protocol 1.

Preparation of nuclear extract

Equipment and reagents * Centrifuge • PBS * TTM * TM buffera

*N100 buffera

buffera . NDil buffera • Dounce glass homogenizer

101

R. H. Nicolas et al. Protocol 1.

Continued

Method 1. Collect cells by centrifugation at 2000 x g. 2. Resuspend the cells in PBS, centrifuge 2000 x g for 10 min. 3. Resuspend in 10 volumes of TTM buffer, homogenizing with a loosefitting Dounce glass homogenizer, five strokes is usually enough. 4. Check that cells have lysed by examining a methylene blue stained sample under the microscope. 5. Centrifuge at 5000 x g for 20 min. 6. Resuspend the crude nuclei in 10 volumes TM buffer and centrifuge at 5000 x g for 10 min. 7. Resuspend in TM buffer. Use 1 ml for every 5 X 108 starting cells. Make the suspension 0.3 M NaCI by adding 4 M NaCI. Ensure that the extract is mixed thoroughly while adding the 4 M NaCI. Leave at 4°C for 30 min and then centrifuge at 40000 x g for 30 min. 8. (a) Dialyse the supernatant versus N100 buffer overnight, or (b) Add 2 volumes of zero salt dilution buffer (NDil). Mix by gentle stirring. Leave on ice for 1 h. 9. Clarify the nuclear extract by centrifugation for 1 h at 40000 x g and collect the supernatant. Freeze rapidly on solid C02 in aliquots or load directly on to the DNA-cellulose column. " See Table 1.

Protocol 2.

Preparation of nuclei from whole tissue, e.g. liver

Equipment and reagents • 2.4 M sucrose in TM buffera • Surgical gauze • Centrifuge

• Domestic blender • TM buffera with 0.25 M sucrose • TTM buffera with 0.25 M sucrose

Method 1. Dice 100 g liver and homogenize for 2 min at low speed in a domestic blender or equivalent with 1 litre of TM buffer plus 0.25 M sucrose. 2. Filter through four layers of surgical gauze. 3. Centrifuge at 2500 x g for 15 min. 4. Discard the supernatant, add 1 litre of TTM buffer containing 0.25 M sucrose and stir or very gently homogenize. Stir for 15 min and centrifuge as above. 102

4: Purification and cloning 5. Discard the supernatant. Gently homogenize the pellet with 5 volumes of 2.4 M sucrose in TM buffer and then centrifuge at 35000 x g for 1 h. 6. Remove supernatant and resuspend the nuclear pellet in 70 ml of TM buffer. ' See Table 7.

4. DNA-cellulose chromatography Calf thymus DNA-cellulose chromatography is a soft-gel technique and does not require any special apparatus apart from a peristaltic pump, column, fraction collector, and UV monitor. The sample should be loaded in a reasonably low ionic strength buffer; most DNA binding proteins bind at 100 mM NaCl. In some cases (e.g. for the GATA1 factor), yields from the DNA-cellulose column can be low unless the detergent Brij-35 is included in the buffer. After loading, the eluate from the column should be monitored by measuring the absorbance at 280 nm (A280) and washed until the absorbance returns to baseline. It is advisable not to stop the buffer flow for long during the chromatography as this results in an increase in absorbance, presumably due to some DNA eluting off the column. The DNA binding proteins are eluted from the column at high ionic strength (E250 buffer, 250 mM (NH4)2SO4), and usually all the protein is eluted in less than a column volume. There is usually little advantage in using more complex chromatographic elutions (i.e. using gradient or multiple salt steps). The ratio of bound to non-bound protein is usually 1:10 and, in the case of GATA1, an approximate yield of 70% is usually obtained. Protocol 3. DNA-cellulose chromatography Equipment and reagents • DNA-cellulose (Sigma) • Column: 1.6 mm diameter, 20 ml volume * 10 mM Tris-HCI pH 7.6, 1 mM EDTA *

. N100 buffera • E250 buffera N200 buffera

Method 1. Weigh out 10 g of DNA-cellulose (Sigma) and allow it to swell in 300 mM NaCl, 10 mM Tris-HCI pH 7.6, 1 mM EDTA. Suspend in 100 ml of buffer and allow to stand for a few minutes. Remove the fines by aspiration. Pack the cellulose into a 20 ml column (1.6 mm diameter) and place in a 4°C cold room. 2. Wash column with 1 volume of 2 M NaCl, 10 mM Tris-HCI pH 7.6, 1 mM EDTA at a flow rate of 0.2-0.5 ml/min. 103

R. H. Nicolas et al. Protocol 3.

Continued

3. Wash column with 1 volume 100 mM NaCI, 10 mM Tris-HCI pH 7.6, 1 mM EDTA. 4. Wash column with at least 3 volumes of N100 buffer. 5. Pump clarified nuclear extract on to the column. Large volumes can be loaded at a slow flow rate overnight. Approximately 200-400 mg of protein is obtained from 1011 cells. 6. Once loaded, wash the column with at least 2 column volumes N100 buffer until the A280 returns to background level. 7. Elute the DNA binding proteins with E250 buffer.b Most of the protein is eluted in 1 column volume as judged by A280. 8. Pool the fractions and dialyse against at least 20 volumes N200 buffer for a minimum of 5 h. 9. Flash freeze in aliquots at -70°C to store or continue with affinity chromatography. aSee Table 1. b Aliquots of fractions should be kept for band-shift analysis. Also, an estimate of the yield of protein is useful at this stage. This can be determined using the Bio-Rad protein assay with minimal interference by non-ionic detergents. Check that there is no activity remaining on the column following elution with E250 by eluting with a higher (NH4)2SO4 concentration (e.g. 500 mM).

5. DNA affinity chromatography In this method, originally developed by Kadonaga and Tjian (7), a large number of binding sites are linked to the column by first polymerizing an oligonucleotide containing a high affinity protein binding sequence and this is then covalently attached to the column matrix (Sepharose). One should aim for more than 2 nmol per millilitre of matrix. This maximizes the concentration of protein binding sites in the column so that there are a large number of specific sites to compete with the competitor DNA that is required for selectivity. Commercially available cyanogen bromide (CNBr)-activated Sepharose 4B (Pharmacia) works well, provided it is not close to its expiry date (Protocol 4). Double-stranded DNA is very unreactive towards CNBr Sepharose 4B, so the oligonucleotide has to have an overhang of at least four base pairs at the end of the DNA and should contain a G base. ACGT is commonly the overhang used. In order to anneal the complementary strands of the oligonucleotide, the extinction coefficients of both strands must be calculated and their concentrations determined accurately. Great care should be taken that the two strands are annealed in equimolar amounts so that no single-stranded DNA remains to be attached to the Sepharose. 104

4: Purification and cloning The binding site for the transcription factor should have been thoroughly characterized by band-shift analysis (see Chapter 1), footprinting, and related techniques (see Chapter 2). The sequence chosen for affinity chromatography should, if possible, bind only one factor specifically and should be the strongest site that has been found. It may be necessary to screen a number of potential sites and also generate mutants to find a binding site with high specificity and affinity. Ideally, the affinity column should bind the protein strongly enough so that it does not elute off the column below 0.5 M NaCl; if it elutes below 0.3 M NaCl, the protein will be very difficult to purify without a number of other steps in the purification procedure. The selectivity of the method is determined by the immobilized polynucleotide and by the concentration and character of the competitor DNA. All the proteins loaded on to the column have the potential to bind to the column. The trick is to add as much competitor DNA as possible so that most non-specific DNA binding proteins partition on to the mobile competitor DNA. In practice, one is limited in how much non-specific DNA can be added without precipitating some of the factor of interest. As a rule of thumb, one can add up to an equal weight of poly(dI-dC) as total weight of protein in the sample to be loaded on to the column. More complex DNA such as E. coli DNA is a more effective competitor but usually starts to compete for the specific protein at lower concentrations. Other synthetic DNA polymers that may help the factor bind more specifically can also be included (8). The sequence used to purify the CTCF factor also binds a poly(G)-binding protein and SP1 (4). Since band-shift analysis showed that poly(dG)-poly(dC) would compete for the poly(G)-binding factor and not CTCF, this was included In the final mix and this prevented binding to the column. Typically, in the purification of CTCF, 16 ml of the E250-eluted fraction from the DNA-cellulose column (1 mg/ml protein) was mixed with 8 ml of poly(dI-dC) and 160 ml of poly(dG)-poly(dC), both at 1 mg/ml in 100 mM NaCl. Non-specific binding of proteins to the affinity column can have several causes. (a) Another protein(s) may bind to a site adjacent to or overlapping the sequence of interest. This problem can be corrected by band-shift analysis and redesigning the oligonucleotide bound to the column; remember that on ligation a new site may be formed. (b) Some proteins may have an affinity for Sepharose 4B. For example, the GATA1 factor binds to Sepharose and elutes gradually at salt concentrations between 100 and 300 mM NaCl. This may be solved by including a pre-column of Sepharose 4B before the affinity column. (c) There are proteins that bind to the DNA of the affinity column because they do not have a high affinity for the competitor DNA. The nature of these contaminants will depend on the nuclear extract, the DNA of the affinity column, and the competitor DNA. One such example are the 105

R. H. Nicolas et al. single-strand nucleic acid binding proteins associated with nuclear hnRNA. During the purification of GATA1, four or five of these proteins consistently co-purified with GATA1. GATA1 could be separated from their contaminants by reverse-phase chromatography (see below), or by recycling through the affinity column, although this resulted in lower yields since contaminated side fractions had to be discarded. If there is evidence that these 40 kDa proteins are binding to the column, then they can be removed by passing the extract through a single-strand DNAcellulose column after the double-stranded DNA-cellulose column (5). (d) It is possible that a DNA binding protein has one or more non-DNA binding proteins associated with it. Such large molecular complexes may be manifested by slowly migrating bands in band-shift analysis. Indeed, a slow broad band might hide a number of complexes. Cross-linking experiments with deoxybromouridine (2) can be used to determine the molecular weight and complexity of the DNA binding components but may give no indication of the size and complexity of the non-DNA binding components (see Chapter 3). It has been shown that a modification of a method relating mobility to molecular weight described by Ferguson (9) for proteins can be applied to the DNA-protein complexes (10). This type of approach may help in determining the overall composition of the complex. If one is sure that all the proteins bound to the column are components of the complex, then in principle they could all be isolated, sequenced, and hence cloned. In practice, one should consider raising antibodies to the complex to generate reagents that will aid in the purification and cloning, and help elucidate the structure of the complex. The pre-column step is included in Protocol 5 but is not mandatory; it may help to remove some of the contaminants from the final products. The precolumn matrix is usually Sepharose 4B with an immobilized polymerized oligonucleotide having a sequence unrelated to that bound by the factor being purified. In practice, we have used the run-through peak from one DNAaffinity-chromatography column to load on to a second different DNA affinity column to purify a second transcription factor. Thus a number of factors were prepared from the same nuclear extracts. The run-through peaks are especially valuable as they have competitor DNA added and should be stored by fast-freezing and kept at —70°C until required. Figure 2 shows an SDS-polyacrylamide gel and the band-shift analysis of the fractions from DNA affinity chromatography during the purification of CTCF from chicken tissue-culture cells. The contaminating hnRNA binding proteins can be seen in the early fractions and elute from the column at approximately 300 mM NaCl. The sample loaded on the column had already been passed through a GATA1 binding column which effectively removed most of these contaminating proteins. CTCF eluted at 500 mM in a total volume of 4 ml. It can be seen from the analysis of Figure 2 that a major protein 106

4: Purification and cloning

Figure 2. (A) SDS-PAGE analysis of fractions obtained following affinity chromatography of CTCF. 10 ul of each sample plus an equal volume of SDS loading buffer was electrophoresed on a Pharmacia Excel gel and stained with silver. Numbering of fractions starts with the start of the gradient. Most of the CTCF eluted at 500 mM NaCI in fractions 17 and 18. M, protein markers (BCL) loaded at 1 ug and 100ng per band; B, the pooled protein from the previous DNA-cellulose chromatography; L, the protein loaded on to the affinity column after adding competitor DNA and centrifuging; W, the wash fractions eluted with N200 buffer. (B) Band-shift analysis of CTCF DNA binding activity in chromatography fractions. The 32 P-labelled DNA probe FPV (2 fmol) was incubated with various competitors and protein fractions and electrophoresed on a 4% polyacrylamide gel (1). From left to right: probe alone; Load, protein loaded on to affinity column; RT, protein not bound to the column; Wash, protein eluted with N200 buffer; PPT, precipitate formed on addition of competitor DNA prior to chromatography (-5% of total CTCF); 15-20, fractions of the affinity chromatograph. Fractions 15, 17, and 18 were also incubated with a 100-fold molar excess of unlabelled FPV probe or with the same amount of an oligonucleotide containing an SP1 binding site, as indicated.

107

R. H. Nicolas et al. of 130 kDa co-elutes with the DNA binding activity. Since photoaffinity labelling of CTCF had characterized CTCF as a 130 kDa protein, it was assumed that the protein band seen on the SDS gel was the CTCF protein. Band-shift and footprinting analyses demonstrated that the purified protein had the same characteristics as the original CTCF in crude nuclear extracts. The protein was finally purified by reverse-phase chromatography as described in Section 6. An alternative affinity purification technique first described by Gabrielsen et al. (11) utilizes biotinylated oligonucleotides which can be immobilized on streptavidin beads. In this method one of the oligonucleotide strands containing the binding site for the protein is chemically synthesized with a biotin moiety at the 5' end. After annealing the complementary strand, the doublestranded oligonucleotide is incubated with the nuclear extract in the presence of poly(dI-dC) using the same conditions used for the band-shift analysis. Magnetic streptavidin beads (Dynabeads M-280 streptavidin, Dynal) are then added and these can be washed rapidly several times using a magnet to draw the beads to the side of the incubation tube. Protein bound to the beads can then be eluted with 0.3-1.0 M NaCl steps. The advantage of the method is the simplicity of the oligonucleotide immobilization and the speed with which the binding, washing, and elution can be performed. Furthermore, since the initial binding of the protein to the oligonucleotide is carried out with the oligonucleotide in solution, the conditions for binding can readily be extrapolated from the band-shift conditions. Protocol 4. Preparation of DNA affinity columns Equipment and reagents • 10 x K buffer: 400 mM Tris-HCI pH 8.0, 100 mM MgCI2, 0.1 mM EDTA . 10 X ligation buffer: 500 mM Tris-HCI pH 7.8, 100 mM MgCI2 • 5 M ammonium acetate . 10 mM ATP pH 7.5 . 370 GBq/ml [32P]ATP (Amersham) • 100 mM potassium phosphate pH 8.0

1 mM HCI 1 M ethanolamine-HCI pH 8.0 1 M KCI, 1 mM EDTA pH 7.6 300 STE: 300 mM NaCl, 10 mM Tris-HCI pH 7.6, 1 mM EDTA NAP-5 column (Pharmacia) CNBr-activated Sepharose 4B (Pharmacia) 0.22 um 115 ml filter (Nalgene)

Method 1. Phosphorylate the two DNA oligonucleotides separately with polynucleotide kinase and ATP. For each strand mix the following: 10 ul 10 x K buffer 20 nmol DNA oligonucleotide 10 ul 10 mM ATP 0.1 ul 370 GBq/ml [y32P]ATP H2O to make up to 100 ul 2. Incubate for 2 h at 37°C with 100 IU of T4 polynucleotide kinase.

108

4: Purification and cloning 3. Add 50 ul of 5 M ammonium acetate, followed by 300 ul ethanol. Leave on dry ice for 5 min and collect the precipitate by centrifugation. Wash the precipitate with 80% ethanol and dry under vacuum. 4. Dissolve the DNA oligonucleotides in 40 ul of water and mix them together. Add 10 ul of 10 X ligase buffer. 5. Heat to 90°C for 5 min and allow to cool to room temperature. Anneal at room temperature for at least 30 min. 6. Add and mix the following: 1 M dithiothreitol (DTT) T4 DNA ligase

1 ul 8 IU

10 mM ATP

10 ul

and incubate for 16 h at 16°C in a water bath in the cold room. 17. Extract the DNA with phenol/trichloromethane followed by trichloromethane. 18. Purify the DNA from any buffer components containing amino groups and ATP by gel filtration chromatography by either of these methods: (a) Load the aqueous phase on to a Pharmacia NAP-5 gel filtration column that has been equilibrated with 100 mM potassium phosphate buffer pH 8.0. Collect 200 ul fractions and monitor the absorbance at 260 nm or use a beta counter to monitor the 32Plabelled DNA. (b) Alternatively, use a spinning column (12). Pack the matrix from a NAP-5 column into a 1 ml syringe and equilibrate with 100 mM potassium phosphate pH 8.0. 9. Estimate the yield of DNA from its A260 and retain an aliquot for scintillation counting the 32P. A sample should also be analysed by agarose gel electrophoresis to measure the extent of ligation, which is usually between 10 and 30 oligonucleotides per chain. 10. Swell 1 g CNBr-activated Sepharose in 100 ml 1 mM HCI for 1 h. 11. Filter off the supernatant under a slight vacuum but do not allow the Sepharose to become dry. 12. Wash the Sepharose with a further 200 ml of 1 mM HCI. 13. Wash the Sepharose with 100 ml of water and transfer to a 15 ml screw-cap plastic tube. Allow the 3.5 ml of Sepharose to settle out and remove all but 1.0 ml of the supernatant. 14. Add 0.5 ml 100 mM potassium phosphate buffer pH 8.0, immediately followed by the oligonucleotide solution from step 8. Rotate the tube on a roller for 20 h at room temperature. 15. Remove supernatant and add 100 ml of 1 M ethanolamine-HCI pH 8.0 and rotate for a further 3 h.

109

R. H. Nicolas et al. Protocol 4. Continued 16. Transfer to a filter and wash slowly (approximately 10 ml/min flow) with: (a) 100 ml 1 M ethanolamine-HCI pH 8.0 (b) 100 ml 10 mM potassium phosphate buffer pH 8.0 (c)

100 ml 1 M KCI, 1 mM EDTA pH 7.6

(d) 100 ml 300 STE plus 0.04% w/v NaN3 17. The material is now ready to be packed into a column. 18. Estimate the yield of oligonucleotide bound to the Sepharose by counting the radioactivity and compare with the results obtained at step 8. Usually between 20 and 40% of the DNA is bound to the column.

Protocol 5. Affinity chromatography Equipment and reagents • N200 and N1000 . 300 STE

buffersa

. N100/50 buffer: N100 buffer containing 50% (v/v) glycerolb

Method 1. Pack a 3-5 ml column with the specific binding oligonucleotide of interest linked to Sepharose 4B. Pack an equivalent sized pre-column with a non-binding oligonucleotide linked to Sepharose 4B. Both columns should be 1 cm diameter and washed with 300 STE. 2. Connect the columns in tandem and wash with N200 buffer at a flow rate of 0.2 ml/min at 4°C. 3. Thaw the DNA binding protein sample and add the solution of the non-specific DNA competitor slowly with careful mixing. Leave on ice for 1 h to allow any precipitate to form. 4. Centrifuge at 40 000 x g for 6 min. 5. Pump the supernatant on to the affinity columns. 6. Wash the columns with N200 buffer until the A280 returns to baseline. 7. Disconnect the pre-column and keep for later elution. 8. Wash the specific column with N200 buffer and then elute with a 60 ml linear salt gradient between 0.2 and 1.0 M NaCI (buffers N200 and N1000) and collect 2 ml fractions. 9. Take aliquots of each fraction for band-shift analysis and SDS-PAGE. Store fractions on ice until these analyses are complete.

110

4: Purification and cloning 10. If required, elute pre-column with a similar gradient of N200 vs. N1000. 11. Wash both columns with 2 M NaCI TE followed by 300 mM NaCI TE, 0.04% NaN3, and store at 4°C. 12. Protein fractions containing the DNA binding activity may be (a) stored (step 13) (b) recycled (step 14) (c) finally purified by reverse-phase chromatography (see below) 13. Dialyse active samples against 10 volumes of N100/50 buffer. After 4 h the sample can be centrifuged out of the dialysis bag and then stored at -20°C. Do not store at -70°C because this results in a large loss of activity, especially with these dilute protein samples. 14. Active samples from step 13 may be reapplied to the affinity column. Normally one would only use the specific column with a small amount of competitor, i.e. 1:1 ratio of competitor DNA to protein. So repeat steps 2-12 omitting steps 7 and 10. a b

See Table 1. 50% v/v glycerol = 62.5% w/v glycerol.

6. Reverse-phase chromatography Ideally, after the affinity chromatography there should be one band on an SDS gel in the fractions that contain all the DNA binding activity. This is not usually achieved in one cycle of affinity chromatography. One can recycle the protein through the affinity column, in which case the amount of competitor DNA should be reduced. The yields of protein are usually quite high; in principle, therefore, this procedure can be continued until the protein is pure. At any stage when there is a simple mixture of proteins (say five bands) several methods (see Chapter 3) can be used to identify which protein is the one of interest: (a) renaturation of protein eluted from SDS gel slices (13); (b) photoaffinity cross-linking to labelled DNA (2); or (c) by two-dimensional electrophoresis of the band-shift gel (1). Once the protein has been unambiguously identified on the SDS-polyacrylamide gel, one is no longer reliant on nondenaturing forms of chromatography, and the high resolution of preparative electrophoresis and reverse-phase chromatography can now be used. Electrophoresis requires that the protein eluted from the affinity columns should be concentrated into a small volume of low-salt loading buffer. Unacceptable losses usually occur on concentration and dialysis of dilute solutions. Some of the buffer components such as non-ionic detergents may also concentrate with the protein and interfere with the electrophoresis. Biotinylated oligonucleotides can be used to concentrate the protein as outlined in the previous section. 111

R. H. Nicolas et al. Reverse-phase chromatography will remove the detergents and other buffer components, remove other protein contaminants, and concentrate the protein into a volatile buffer, Figure 3 shows an elution profile of protein from the previous CTCF affinity chromatograph loaded on to a Ci/Cx Pharmacia Pro-RPC column and eluted with an acctonitrile gradient in 0.08% trifluoroacetic acid (TFA). The sample was diluted with TFA and then loaded by three injections using a 2 ml loop. As can be seen, the vast majority of the optical density does not bind to the column. Most of the other major peaks do

Figure 3. (A) Elution profile of the Pro-RPC reverse-phase chromatography of CTCF. Flow rate, 0,7 ml/min; h 214 nm. Buffer A 0.08% TFA; buffer B 80% acetonitrile, 0.08% TFA. Three 2 ml loadings were made at 20, 14, and 7 min and the gradient started at time zero; 1 min (0.8 ml) fractions were collected. (B) SDS-PAGE analysis of reversephase chromatography fractions. Lanes 1-20 correspond to fractions 21-40; 20 ml of each fraction was loaded. Lane C is 20 ml of affinity-purified CTCF and the marker lanes M were loaded with 100 ng and 25 ng per band. The Excel gel was stained with silver by the Pharmacia standard protocol.

112

4: Purification and cloning not contain protein and are due to detergent and other buffer components used in the previous steps. The 130 kDa protein elutes at 35% acetonitrile in a volume of 0.7-1.4 ml. This protein is then available for direct sequence analysis and chemical or enzymatic cleavage. Protocol 6.

Reverse-phase chromatography

Equipment and reagents • Apparatus: HPLC with two pumps, large (2 ml) loop, and gradient programmer, and a detector set at 214 nm (the Beckman System Gold was used in Figure 3) • Column: Pharmacia Pro-RPC . Solvent A: 0.08% (v/v) TFA (Pierce)a • Solvent B: -0.08% (v/v) TFA in 80% acetonitrile, 20% H2Oa

• Gradient programme: Time (min) % Solvent B 5 Start fraction collector 0 3 5 3-5

5-15

5-45 45 45-50 55-60

15-70 70 Stop fraction collector 70-100 100-5 5 5 Stop flow

60 70

Method 1. Connect the Pro-RPC column to the HPLC. The 5/2 (5 mm diameter x 2 cm) column is used at a flow rate of 0.7 ml/min. Up to 4 ml of detergent-containing buffers have been loaded on to this column with good results. The 10/10 (10 mm x 10 cm) column is used for 10-20 ml of sample and can be run at 2 ml/min. 2. Pre-wash the column alternately with 100% solvent B and 100% solvent A a few times until the baseline stabilizes, then return to the loading condition (5% solvent B). 3. Return the HPLC to manual control mode so that multiple loading can be carried out. 4. Remove one 2 ml protein sample from ice and hand-warm it to room temperature. Add 20 ul 10% (v/v) TFA to the sample, load it into the loop, and inject the column. 5. After 3 min, reload the loop with protein as in step 3 and inject the column once the absorbance has returned to zero. 6. Set up the gradient programme. Start the fraction collector, recorder, and gradient at the same time and collect 1 min fractions. 7. Dry down 40 ul of each sample between fractions 15 and 40 using vacuum centrifugation (Savant SpeedVac concentrator). Redissolve in SDS sample solvent and analyse by SDS-PAGE, silver staining the gel.b 8. The remainder of the sample should be rapidly frozen in solid CO2 and stored at -70°C.

113

R. H. Nicolas et al. Protocol 6.

Continued

9. At the end of the run, wash out the loop with solvent A then wash the column and HPLC with methanol. aThe solvents should be made up and purged with helium for 10 min and then the absorbance at 214 nm measured. Small aliquots of 10% TFA can be added to the A or B solvent with the lower absorbance until they balance. b Note some proteins do not stain well with silver. In these cases, gels may be electroblotted on to PVDF (polyvinylidene flouride) membranes such as Immobilon-P (Millipore) and stained with the gold stain Aurodye (Janssen) (1).

7. Production and isolation of peptides In order to clone the cDNA, one should aim to get several peptide sequences since not all oligonucleotide probes (based on peptide sequence) are suitable for screening libraries. Some reasons why an oligonucleotide may not be suitable are: (a) if they contain a high GC content (b) if they are very redundant (c) if they contain cross-hybridizing sequences (d) if they code for the N-terminal end of the protein and the 5' end of the gene is not represented in the library Also, with two or more sequences one has the option of using a polymerase chain reaction (PCR) approach to obtain a larger probe to screen the library (see Chapter 6). The rather sparing quantities of the pure protein that are usually produced mean that there is very little opportunity to optimize the production and purification of the peptides. A test protein (e.g. bovine serum albumin) can be used but the peptide cleavage pattern of each protein is unique so one usually has to just 'go for it'. We have used two methods for separating peptides after cleavage: microbore reverse-phase chromatography and polyacrylamide gel electrophoresis followed by protein blotting (14). Both these methods will give sufficient peptides for sequencing using a minimum of 100 pmol of protein. There will always be losses associated with fractionating the peptides following cleavage, and not all the peptide immobilized on the support in the sequencer is chemically available. With 100 pmol of protein we obtained 5-10 pmol initial yield of amino acids at the start of a sequencing run, and a sequence of up to 20 amino acids can be obtained, depending on the nature and purity of the peptides. To obtain reliable sequences it is essential that sufficient pure protein is accumulated before committing it to cleavage and a sequence run. The two choices that have to be made are the type of cleavage and the type of purification. Cleavage with CNBr is commonly used, because it is very specific for methionine which is a rare amino acid (-2%) in most proteins. 114

4: Purification and cloning This generates few but large peptides, which are best separated by SDSpolyacrylamide gel electrophoresis (SDS-PAGE). This method is especially useful for large proteins. Enzymatic digestions normally cut more frequently and this can make it more difficult to purify peptides in one reverse-phase chromatography step. However, we have used the enzyme Glu-C, which cleaves the carboxyl terminus of glutamic and aspartic acid, successfully to generate peptides from GATA1 and have purified the peptides by reversephase chromatography. Microsequencing is a field of its own and beyond the scope of this chapter. There are two good guides to microsequencing and related techniques (14, 15) and the second describes the use of microbore columns for chromatographies other than reverse-phase chromatography. The sequence analysis of CTCF and GATA1 was carried out on an Applied Biosystems gas-phase sequentor 120A with a 900A data-handling system. Applied Biosystems also provides useful application notes. A complementary approach is to use mass spectroscopy to obtain the masses of tryptic peptides of the SDS-PAGE purified protein (16, 17). The data can then be used to search theoretically derived databases derived from established sequence databases (18). This peptide mass fingerprinting approach has been refined to include short partial peptide sequences that can be obtained from tandem mass spectroscopy techniques. Thus after the reversephase purification of the protein described in Protocol 6 the protein fraction can be evaporated to dryness, redissolved, and digested with trypsin then analysed by mass spectrometry. If the protein is still not pure after the reverse-phase chromatography the protein can be run out on an SDS electrophoretic gel; Protocol 9 describes the procedure that can be used to obtain tryptic peptides from gel-purified protein for matrix-assisted laser desorptionmass spectrometry (MALD-MS) (19 ). This protocol could be used after the affinity chromatography step instead of the reverse-phase chromatography for the final purification but the protein would have to be concentrated prior to loading on to the electrophoretic gel. This is carried out using an Amicon Centricon concentrator or using a biotinylated oligonucleotide and streptavidin-coated magnetic beads (see above). Other modifications to this protocol include using silver-stained gels and in-gel trypsin digestions (20). Protocol 7. Preparation of peptides by CNBr cleavage and purification by reverse-phase chromatography Equipment and reagents • SpeedVac (Savant) • CNBr: 25 mg/ml in 70% v/v formic acid * • 8 M guanidine-HCI, 20 mM Hepes pH 7.9, 100 mM DTT • Microbore reverse-phase column: Brownlee C18 RP-300, 2.1 x 100 mm

• Solvent A: 0.08% TFA Solvent B: -0.08% TFA, 80% acetonitrile • HPLC syringe pump designed for microbore work, programmable • UV detector: h = 214 nm, or variable wavelength

115

R. H. Nicolas et al. Protocol 7.

Continued

Method 1. Concentrate purified protein from the preceding reverse-phase chromatography to 200 ul by evaporation using a SpeedVac (Savant).a Do not evaporate to dryness. 2. Add 0.4 ml formic acid (analytical grade), followed by 0.2 ml of 25 mg/ ml CNBr in 70% v/v formic acid. 3. Flush with nitrogen and incubate in the dark at 25°C for 6 h. 4. Add a further 0.2 ml CNBr solution and incubate overnight. 5. Dry down sample in a SpeedVac with NaOH in the trap. 6. Add 100 ul of HPLC grade H2O and dry down again. Continue to step 7 for reverse-phase chromatography or to Protocol 8 for SDS-PAGE separation of peptides. 7. Redissolve sample in 100 ul 8 M guanidine-HCI, 20 mM Hepes (pH 7.9), 100 mM DTT. 8. Centrifuge sample and load on to microbore reverse-phase column with a flow rate of 100 ul/min. Collect 1 min fractions or collect peaks manually. (Remember to determine the time delay from the detector to the fraction collector.) Use the following gradient programme: Time (min)

% Solvent B

0 0-5 5-55 55-60 70-75 90

5 Start 5 70 100 5 5 Stop

9. Freeze all peak fractions rapidly and store at -70°C. Losses tend to increase with storage, so if possible analyse the best peaks immediately. a

For enzyme cleavage, neutralize the protein sample from the preceding reverse-phase chromatography by adding Hepes to 20 mM. Evaporate off 50% of the solvent; this removes most of the acetonitrile. Check the pH and add enzyme. For these dilute solutions, use a ratio of 1 ug sequence-grade enzyme (Boehringer) to 10 ug protein. Both Glu-C and trypsin will work well in 10% acetonitrile. After digestion, reduce with 100 mM DTT and continue to step 8.

Protocol 8. Purification of peptides by SDS-PAGE and blotting on to PVDF membrane Equipment and reagents • SDS-polyacrylamide (Laemmli) gel: 0.75 mm thick, 15-20% polyacrylamide • 100 uM thioglycollic acid

• SDS gel loading buffer: 2% (w/v) SDS, 62.5 mM Tris-HCI pH 6.9, 10% (v/v) glycerol, 100 mM DTT

116

4: Purification and cloning • Transfer buffer: 10 mM CAPS (3-[cyclo• Transfer apparatus hexylamino]-1-propanesulphonic acid) pH . 0.1% coomassie blue R250 (PAGE blue 83, 11.0, 10% methanol BDH) * PVDF (polyvinylidene flouride) membranes * Destain solution: 50% (v/v) methanol, 10% (Applied Biosystems) (v/v) acetic acid • Whatman 3MM paper

Method 1. Make up a 0.75 mm thick SDS-polyacrylamide (Laemmli) gel with 15-20% polyacrylamide, depending on the sizes of the peptides in the mixture. Use highly purified acrylamide, ultra-clean apparatus, and HPLC-grade water for the buffers. Store the gel overnight. 2. Redissolve peptide sample in 20-40 ul SDS gel loading buffer and heat to 90°C for 5 min. 3. Load gel and run as normal with markers but include 100 uM thioglycollic acid in the top reservoir buffer. 4. At the end of the electrophoresis, place the gel in transfer buffer for 5 min. 5. At the same time wet three PVDF membranes in methanol then soak in transfer buffer. 6. Make up the transfer sandwich with Whatman 3MM paper (soaked in transfer buffer) at the anode side, followed by two membranes, the gel, a protecting back-membrane of PVDF, and more sheets of Whatman 3MM paper. 7. Place the sandwich in the apparatus and start the transfer. The time and current required depend on the size of the protein and design of the apparatus. The conditions should first be optimized for some test peptides in the same molecular weight range before committing the transcription factor peptide sample. 8. The backing membrane will have picked up some of the protein by passive blotting: stain with Aurodye for a permanent record. 9. Wash transfer membranes in HPLC water, then stain with 0.1% coomassie blue for 5 min and destain with 50% (v/v) methanol, 10% (v/v) acetic acid and, finally, with methanol. Allow to dry. The detection limit is about 200 ng of peptide. Since this is the lower limit for sequencing, there is no need for more sensitive staining. 10. Cut bands out and place in gas-phase sequencer for sequencing. If protein bands are detected on the second transfer membrane, load them also into the sequencer. Bands can be stored at -20°C, if necessary.

117

R. H. Nicolas et al.

8. Design of oligonucleotides for cDNA isolation Typically, the sequencing of peptides will give you several sequences, 10-30 residues in length, which can be used to design oligonucleotides for PCR or library screening. At some positions, there may be doubt as to the identity of the amino acid; e.g. histidine and serine are sometimes not detected and tryptophan may be missed unless suitable precautions are taken. Two approaches should be used to isolate a cDNA. The first is to use highly redundant oligonucleotides in the PCR reaction to amplify a short fragment of DNA which (after cloning) can be used to screen a cDNA library for the full-length cDNA. This approach was first used to isolate a cDNA encoding the CREB transcription factor (21). The oligonucleotide primers can be 20-30 base pairs in length and can be up to 106-fold degenerate to accommodate all the codon possibilities, although we would advise restricting degeneracy to less than 105. The template is cDNA, preferably random primed. The PCR methodology is described in more detail in Chapter 6. The second strategy is the use of labelled oligonucleotides to screen the libraries. The problem here is that, unlike the PCR technique, it is difficult to use fully degenerate probes. If the complexity of the probe is very high and only a small percentage of the oligonucleotides in the mixture is sufficiently homologous to the target sequence to hybridize, then the concentration of hybridizing probe is not sufficiently high to drive the annealing kinetics at a reasonable rate. Also, a large percentage of the radioactivity will be on incorrect oligonucleotides and this will contribute to high backgrounds on the filters. In practice, researchers have used either short, mixed oligonucleotides to reduce degeneracy or single longer probes, guessing the third base at each codon or using inosine. Lathe (22) has advocated the use of a longer single probe, the sequence of which is optimized using codon usage considerations. The reader is well advised to consult this paper, not only for probe design but also for hybridization and washing conditions. Briefly, in order to screen a cDNA library of average complexity, the probe should be designed such that it has sufficient specificity to detect its target, and there should only be a 0.1 frequency of chance binding to spurious sequences in the library. This can be achieved with a 36-mer oligonucleotide, choosing the third base from a codon usage table and avoiding sequences encoding leucine, arginine, and serine. The average probability of correct choice of the third position is then 0.55 and gives an average predicted probe-target homology of 85%. Such a 36-mer is the minimum probe length required. In practice, one is likely to design a larger probe since there may be amino acids within the peptide chosen that were not identified or were not unambiguously identified, and there is a high chance that the chosen peptide has a serine, leucine, or arginine. Improved homology can be achieved by introducing a moderate level of complexity in the probe (giving a mixture of less than 128 oligonucleotides) by using two preferred bases at a few positions. Also, since deoxyinosine can base-pair with A and C (23), where the 118

4: Purification and cloning third base is likely to be T or G, this base can be incorporated. Inosine can also be used for the three bases of an unknown amino acid in the peptide sequence. With such considerations in mind, the CTCF cDNA was isolated by designing several oligonucleotides which were end-labelled and first checked by probing a Northern blot. One of the probes detected a band of sufficiently long mRNA to encode the protein. The sequence of this oligonucleotide was: ATG M GAG/A E

GAG/A E GCC/T A

GGC/A G ATT I

GAG/A E GTG V

GCT A GAG/A E

GTG V GA E

This oligonucleotide was end-labelled with polynucleotide kinase to a specific activity of more than 108 c.p.m./mg (Protocol 10) and used to probe a cDNA library. The most convenient procedure for screening high-complexity libraries (~106 recombinant) is to plate out the phage at high density in large rectangular dishes and grow the phage until the plaques are less than 1 mm diameter. Duplicate filters are lifted and hybridized (24-48 h) in pairs, back to back, in sealed bags (Protocol 11) at a temperature 10-25°C below the expected Tm of the hybrid. This can be calculated from the formula: where Tm is the melting temperature in 6 X SSC, l is the length of the probe, and h is the percentage homology between probe and target. The filters are washed with SSC buffers at increasing temperatures, autoradiographing after each wash. Phage are extracted from positive plaques and rescreened through two more rounds. The cloned phages are then amplified and the DNA extracted. The inserts can be amplified and isolated by PCR and checked by Northern blot hybridization for correct size or tissue distribution, before subcloning and sequencing Protocol 9. Trypsin digestion of electrophoretically purified proteins for mass spectrometry Equipment and reagents • SpeedVac (Savant) • SDS-PAGE and electroblotting apparatus as in Protocol 8 . 0.1% Coomassie blue R-250, 50% methanol • 70% Acetonitrile

• Trypsin (Promega, modified sequence grade) • Digest buffer: 25 mM ammonium bicarbonate pH 7.8, 1% octyl B-glucoside, 10% methanol • Formic acid:ethanol (1:1)

Method 1. After the reverse-phase chromatography dry down the protein fractions using a SpeedVac. 2. Redissolve the protein in SDS sample buffer, run on SDS-PAGE (up to

119

R. H. Nicolas et al. Protocol 9.

Continued

six wells of an analytical gel can be loaded), then electroblot on to PVDF membrane as described in Protocol 8. 3. Stain the membrane with 0.1% Coomassie blue in 50% methanol, then destain with 50% methanol. 4. Rinse the membrane with HPLC-grade water and air dry between sheets of Whatman filter paper. Excise the protein band and place in an Eppendorf tube. 5. Destain the band with 70% acetonitrile, dry, and add 10 ul of trypsin, 20 ug/ml (Promega, modified sequence grade) in digest buffer. Incubate the digest at room temperature for 5 h then at 27°C overnight. 6. Extract peptides from the membrane with 20 ul formic acid:ethanol (1:1) for 2 h at 27°C and dry down the extract. 7. Resuspend the peptides in 10 ul methanol for MALD-MS analysis.

Protocol 10. End labelling of oligonucleotide Equipment and reagents • Buffer K (Protocol 4) * [y32P]ATP, 370 MBq/ml (185 GBq/umol) • T4 polynucleotide kinase (New England Biolabs)

• Spinning-column apparatus • Sepharose G-25 (DNA grade) or NAP-5 column . 10 mM Tris pH 7.6, 1 mM EDTA

Method 1. Mix the following on ice: 2 ul buffer K (Protocol 4) 2 ul oligonucleotide (360 pmol, -0.7 mg) 15 ul [y32P]ATP, 370 MBq/ml (185 GBq/umol) 1 ul (10 IU) T4 polynucleotide kinase 2. Incubate at 37°C for 45 min. 3. Remove unincorporated nucleotides by the spinning-column method using DNA-grade G-25 Sepharose or a NAP-5 column equilibrated in 10 mM Tris (pH 7.6), 1 mM EDTA. The specific activity of the probe should be > 108 c.p.m./ug.

Protocol 11. Hybridization of oligonucleotide to library filters Equipment and reagents • 24.5 x 24.5 X 2 cm dishes (Nunc) * • 20 x 20 cm nitrocellulose filters (Amersham)

UV Stratalinker 1800 (Stratagene) • 20 x SSC: 3 M NaCI, 30 mM sodium citrate

120

4: Purification and cloning *• Pre-hybridization buffer: 6 x SSC, 0.5% SDS, 100 ug/ml sonicated denatured salmon sperm DNA, 5 5x x Denhardt's Denhardt's solution . Hybridization buffer: 6 x SSC, 0.1% SDS

• 0,1% w/v Ficoll Type 400, Pharmacia .• 0.1% w/v polyvinyl-pyrrolidone .•0.1% bovine serum albumen 0.1% bovine serum albumen

Method 1. Plate out the cDNA library at a density of approximately 5 X 105 plaques per 400 cm2 in 24.5 x 24.5 x 2 cm dishes. Lift the plaques in duplicate on to 20 x 20 cm nitrocellulose filters, alkali denature, and neutralize as described by Maniatis et al. (12). 2. Cross-link the filters with UV light using the UV Stratalinker. Use the auto cross-linking conditions (1.2 x 105 mJ). 3. Pre-hybridize the filters overnight in sealed bags; two duplicate filters back to back per bag, in pre-hybridization buffer. Use the same temperature for pre-hybridization as that chosen for hybridization (40-55°C). 4. Replace the solutions with 20-30 ml per bag hybridization buffer containing the labelled probe (106-107 c.p.m./ml). The temperature of hybridization should be 10-25°C below the Tm. Hybridize for 24-48 h, depending on the complexity of the probe. 5. Wash the filters four times for 5 min each in 6 X SSC at room temperature, then at steps of increasing temperatures until the Tm is reached, autoradiographing the filters between washes and exposing to the Xray film for approximately 4 h at -70°C with an intensifying screen. Each high-stringent wash should be for 1-5 min.

Acknowledgements The work described in this chapter was supported by the Cancer Research Campaign. The authors thank A. Carne for peptide sequencing and N. Perkins, C. Heath, and E. Klenova for their contributions in developing the techniques described in this chapter.

References 1. Perkins, N. D., Nicolas, R. H., Plumb, M. A., and Goodwin, G. H. (1989). Nucleic Adds Res., 17, 1299. 2. Wu, C., Wilson, S., Walker, B., Dawid, I., Paisley, T., Zimarino, V., and Ueda, H. (1987). Science, 238, 1247. 3. Wang, J., Nishigama, K., Araki, K., Kitamura, D., and Watanabe, T. (1987). Nucleic Acids Res., 15, 10105. 4. Lobanenkov, V. V., Nicolas, R. H., Adler, V. V., Paterson, H., Klenova, E. M., Polotskaja, A. V., and Goodwin, G. H. (1990). Oncogene, 5, 1743. 121

R. H. Nicolas et al. 5. Evans, T. and Felsenfeld, G. (1989). Cell, 58, 877. 6. Dignam, J. D., Lebovitz, R. M., and Roeder, R. G. (1983). Nucleic Acids Res., 11, 1475. 7. Kadonaga, S. T. and Tjian, R. (1986). Proc. Natl Acad. Sci. USA, 83, 5889. 8. Kadonaga, J. T. (1991). In Methods in enzymology, Vol. 208 (ed. R. T. Sauer), pp.10-23. Academic Press, London. 9. Ferguson, K. A. (1964). Metabolism, 266, 3052. 10. May, G. E., Sutton, C., and Gould, H. (1991). J. Biol. Chem., 266, 3052. 11. Gabrielsen, O. S., Homes, E., Korsnes, L., Ruet, A., and Oyen, T. B. (1989). Nucleic Acids Res., 17, 6253. 12. Maniatis, T., Fritsch, E. F., and Sambrook, J. (eds) (1982). Molecular cloning, a laboratory manual. Cold Spring Harbor Press, Cold Spring Harbor, NY. 13. Hager, D. A. and Burgess, R. R. (1980). Anal. Biochem., 109, 76. 14. Matsudaira, P. T. (ed.) (1989). A practical guide to protein and peptide purification for microsequencing. Academic Press, London. 15. Nice, E. C. (1990). Nature, 348, 462. 16. Patterson, S. D., Aebersold, R. (1995). Electrophoresis, 16, 1791. 17. Wilm, M., Shevchenko, A., Houthaeve, T., Breit, S., Schweigerer, L., Fotsis, T., and Mann, M. (1996). Nature, 379, 466. 18. Mann, M. and Wilm, M. (1994). Anal Chem., 66, 4390. 19. Hynes, G., Sutton, C. W., U, S., and Willison, K. R. (1996). FASEB J., 10, 137. 20. Varga-Welsz, P. D., Wilm, M., Bonte, E., Dumas, K., Mann, M., and Becker, P. B. (1997). Nature, 388, 598. 21. Gonzalez, G., Yamamoto, K. K., Fischer, W. H., Karr, D., Menzel, P., Biggs, W., et al. (1989). Nature, 337, 749. 22. Lathe, R. (1985). J. Mol Biol., 183, 1. 23. Kawase, Y., Iwai, S., Inoue, H., Miura, K., and Ohtsuka, E. (1986). Nucleic Acids Res., 14, 7727.

122

5 Cloning transcription factors from a cDNA expression library IAN G. COWELL and HELEN C. HURST

1. Introduction The initial chapters in this book describe methods of investigating the DNA binding activity of a transcription factor and how to use these properties to purify the protein. However, further studies concerning the functioning of the factor in transcription can be extremely limited unless a cDNA clone encoding the factor is isolated. There are a number of ways of cloning DNA binding factors and the most suitable method will vary from case to case. If, for example, the factor of interest has been isolated in sufficient quantity and purity for a reliable peptide sequence to be obtained, then the peptide sequence data may be used to design oligonucleotide probes to screen cDNA libraries by hybridization (see Chapter 4). However, a number of alternative approaches utilize cDNA expression libraries, where specific clones are detected through some property of the encoded polypeptide such as its DNA binding activity or immunological properties. Commonly, cDNA expression libraries are constructed using bacteriophage vectors such as hgt11 where the cloned cDNA is expressed as a fusion protein with B-galactosidase upon addition of an inducer (1). The advantages of such a vector include rapid screening of large numbers of recombinants and large library size. Pertinent points concerning the handling of bacteriophage libraries are included in Section 2. As alluded to above, one method of screening bacteriophage expression libraries is the use of DNA probes comprising the binding site of the factor in question. This approach has been particularly successful in the cloning of members of the leucine zipper family of factors (2-5), and has also been used for other classes of DNA binding protein (6-11). The relative merits and shortcomings of this approach and of immunological screening are discussed in Section 3.1 and protocols for each method appear later in the section. A recent addition to the list of strategies available for cloning DNA binding proteins is the yeast one-hybrid system of Inouye et al. (12). This also utilizes the DNA binding properties of the factor of interest. However, in this

Ian G. Cowell and Helen C. Hurst method, as described in more detail in Section 3.4, detection occurs in vivo through activation by 'positive' cDNA clones of a marker gene, allowing cells to form colonies on selective minimal media.

2. Handling bacteriophage expression libraries This topic is covered exhaustively in a previous edition of this series (1). However, the main points are presented here for completeness and particular issues relevant to this type of screening are discussed.

2.1 Library selection An obvious prerequisite is that the cDNA expression library should be generated from a cell line or tissue known to express the factor of interest in reasonable abundance. It is then the investigator's choice whether to make their own library or purchase one. In either event, the library should have a complexity of at least 106: in theory only one in six inserts will be in the correct orientation and frame to generate a B-galactosidase fusion protein, although increasing use of directionally cloned cDNA libraries may increase the odds to one in three. In practice we have found, however, that proteins expressed from such libraries are often not expressed as fusion proteins, as translation starts from an internal AUG codon. This can mean that the genuine initiator AUG may be used in cDNA clones complete at their 5' end. The reason for this is presumably because 5'-untranslated regions often contain in-frame stop codons, thus preventing read-through from B-galactosidase. If screening with DNA binding-site probes (see Section 3.2), a full length cDNA is not usually required, but sufficient coding information to generate a complete DNA binding/dimerization domain is essential. As this domain may lie at the very N terminus or C terminus of the protein, the best way to try to ensure that these sequences are represented in the library is to screen a mixed library generated using both random priming and oligo-dT priming. This is particularly relevant if previous characterization of the factor (by Southwestern blotting for example, see Section 3.2.2) has indicated that it is quite a large protein (say > 60 kDa). However, smaller proteins may also be difficult to clone if they have exceptionally long mRNAs. For example, the 46 kDa CREB protein is encoded in 1023 bases within a 7 kb mRNA. The following protocols assume the use of a hgt11 library, but libraries constructed in hZAP (Stratagene) can also be used by modifying the phage growth protocol to comply with the supplier's instructions. We have used a Clontech human placenta library in hgt11 which we have found to have a high complexity, containing cDNA clones for a wide range of transcription factors. However, few of these clones are full length, apparently due to incomplete EcoRI methylation of the nascent cDNA prior to cloning. 124

5: Cloning from a cDNA expression library

2.2 Library plating for screening The home made (preferably unamplified) or commercial library should first be titred on the bacterial expression strain Y1090 (Clontech). Make fresh plating cells by inoculating one colony into 20 ml of L-broth containing 0.2% maltose. For all h work use L-broth containing 10 mM MgSO4. Shake the culture vigorously until stationary phase is reached, then pellet the cells and resuspend in half their original volume using SM buffer (see Protocol 7). Plating cells may be stored on ice until required (up to 24 h). Once the titre has been ascertained the library can be plated on 15 cm plates and protein expression induced for transfer to nitrocellulose membranes as detailed in Protocol 1. Protocol 1. Library plating and replica lifts Equipment and reagents • Well-dried L-agar plates, 15 cm and 9 cm • Fresh Y1090 plating cells diameter . Nitrocellulose filters (Schleicher and • Autoclaved top agarose: 0.7 g agarose in Schuell): 132 and 82 mm circles 100 ml L-broth. Melt in microwave and • Isopropyl B-D-thiogalactopyranoside (IPTG): maintain at 45°C for 20 min prior to use 1 M stock in water; store at -20°C . SM buffer to dilute phage: 100 mM NaCI, 10 • Wash buffer: 50 mM NaCI, 10 mM Tris-HCI mM MgSO4, 50 mM Tris-HCI pH 7.5, 0.01% pH 7.5, 1 mM EDTA, 1 mM DTT, 0.05% lauryl gelatin; autoclaved dimethylamine oxide (LDAO; Calbiochem)

Method 1. Plate out at a density of 5 x 104 per 15 cm plate. For each plate dilute phage in 200 ul of SM buffer and add 600 ul of fresh Y1090 plating cells. Leave to infect for 15 min at 37°C then add 7.5 ml top agarose and pour on to pre-warmed plates. Aim to analyse 10-20 15 cm plates in the primary screen. 2. When all plates are set, transfer to a 42°C incubator until plaques are just visible as pinpricks on the bacterial lawn. This takes 3-4 h. Meanwhile, soak one nitrocellulose filter per plate in 10 mM IPTG and pat dry. 3. Working quickly so that the agar temperature does not drop below 37°C, take one plate at a time and place a still-damp numbered filter on the surface, then transfer the plate to a 37°C incubator. Incubate for 1-2 h. 4. Again taking one plate at a time, pierce the filter and agar with three to four orientation holes using a syringe needle dipped in India ink. Carefully remove the filter with forceps and transfer to a lunch box containing 500 ml of wash buffer. Place a second IPTG-soaked filter on the agar and incubate at 37°C for a further 2-3 h.

125

Ian G. Cowell and Helen C. Hurst Protocol 1. Continued 5. With a permanent marker pen, mark the second set of filters with dots over the syringe holes in the agar. A light box makes this easier. Remove these filters to wash buffer. All filters should be washed for 5-10 min to remove adherent bits of agarose and so reduce background. The filters are now ready for blocking (see Section 3.2 or 3.3 depending on screening method to be employed).

2.3 Plaque purification The screening process is described in the next section. Once positives are detected on duplicate filters, corresponding plugs of agar should be picked from the plate for further rounds of screening and for phage purification. Using the orientation holes in the agar, align the plate with the developed autoradiograph and use the wide end of a Pasteur pipette to pierce the agar. Transfer the resulting agar plug into 1 ml of SM buffer containing a drop of trichloromethane in a bijou bottle. The bottles can be shaken gently to speed phage elution (2 h to overnight). Ideally the eluted phage should be titred before plating for second round screens. Plate out the titred, eluted phage as described in Protocol 1, but this time at a density of 500-1000 plaques on a 9 cm plate. It is possible to use less dense plates, but it is more reassuring if second round screens contain several positive signals per plate! Duplicate lifts are not always necessary for second and further screens. Screen these second round filters and pick positive areas of agar using the narrow end of a Pasteur pipette, again into 1 ml of SM plus a drop of trichloromethane. The phage eluted from these plugs should be titred and replated at low density on 9 cm plates for third round screens. After this stage it should be possible to pick a pure, clean phage plaque; however, it is wise to perform a fourth round screen to check for purity, as phage particles do diffuse readily on agar plates.

3. Screening methods 3.1 Introduction Once the filters carrying the expression library have been generated they may be probed using one of the methods described in this section. Two approaches will be considered in detail: the use of DNA binding-site probes and immunological screening. Each of these methods has its own merits and drawbacks; the use of DNA binding-site probes has the advantage that little prior characterization of the factor or factors of interest is required before embarking on library screening and it is not necessary to obtain highly purified fractions of the protein. This feature can be useful if it is simply desired to clone factors that bind a particular sequence, for example a defined regulatory element in a 126

5: Cloning from a cDNA expression library promoter, when little detail is known about these factors. However, this lack of specificity can be a problem if one particular factor is to be cloned, since it is frequently the case that a family of factors rather than a single species bind any given regulatory element (see Section 3.2.2). However, as discussed in Section 3.2.1 this problem can often be minimized by careful selection of probe sequence. Some other limits to the effectiveness of binding site screening are outlined below. (a) If post-translational modification such as phosphorylation or specific proteolytic cleavage is required for efficient DNA binding activity, then the factor expressed in E. coli may not bind DNA with high affinity. Careful selection of binding and washing conditions may circumvent this problem (see Section 3.2). (b) Problems might arise from poor solubility or protein folding of factors expressed in E. coli. However, this need not be intractable as described in Section 3.2.5. (c) A third problem with binding site screening arises if the factor of interest binds DNA efficiently only as a heterodimeric or heteromeric complex, since only a single species can be expressed from any one clone. Similarly this method is of no value in the cloning of components of a multisubunit transcription factor complex if they do not stably bind DNA themselves in the absence of the other components. Immunological screening has the advantage that DNA binding activity is not required from the expressed product of the cDNA clone and so the potential problems discussed above do not arise. However, there are potential drawbacks associated with this method as well. First, prior purification of the protein is required in order to raise an antibody for screening. Second, as described in Section 3.3.3, unexpected results can arise due to lack of specificity or cross-reactions with the antibody. A third drawback of immunological screening compared to screening with DNA binding-site probes concerns the ease of analysing the clones that are obtained. Since a universal feature of transcription factors is DNA binding, the first feature to test when checking the identity of a clone is its DNA binding activity and specificity. This is straightforward if DNA binding activity was the feature used to select the clone (see Section 4), but this is not necessarily the case for 0immunological screening, since the clones that are obtained may be partial and will not always contain the sequences required for DNA binding. This may be overcome easily by rescreening the library, based on hybridization to the sequence isolated, to obtain longer cDNA clones.

3.2 Screening with DNA binding-site probes Before embarking on library screening it is advisable to determine the optimum conditions for binding in terms of specific probe sequence, non-specific 127

Ian G. Cowell and Helen C. Hurst competitor DNA, and binding buffer composition. These points are considered in Section 3.2.1. A procedure for screening with binding-site probes is given in Section 3.2.4 and a variation of the protocol appears in the following section. Further notes and fault-finding hints are given in Section 3.2.6. 3.2.1 Selection of probe sequence and binding conditions The use of appropriate binding-site probes is obviously essential for successful cDNA library screening. In general, the binding site that is used will be one that is known from gel retardation or other experiments to be bound by the factor of interest. Probes containing multiple binding sites give good signals and may take the form of a single, long oligonucleotide containing several sites for the factor of interest, a restriction fragment containing multiple cloned binding-site oligonucleotides, or alternatively, and in our hands preferably, a concatenated binding-site-containing oligonucleotide. The rest of this section assumes the use of this third type of probe; however, the use of such probes can be problematic, as discussed below, and successful screens can be performed using single binding-site probes. A recent example is the cloning of the factor E4F (13). For the construction of concatenated probes the specific binding sequence for the factor of interest should be contained in a 20-25-mer double-stranded oligonucleotide with free 3'-hydroxyl groups to facilitate labelling with [y32P]ATP and compatible 'sticky ends' to allow ligation. The sequence of the oligonucleotide should be chosen, if possible, such that it does not contain binding sites for DNA binding factors other than the ones of interest. The specificity of a new oligonucleotide sequence can easily be tested by electromobility shift assay (EMSA). However, another consideration when designing a probe is the nature of the new sequences that are generated at the ligated junctions of the concatenated probe. The generation of binding sites for other factors at these positions should also be avoided as much as possible in order to simplify the screening procedure. EMSA analysis can also be used to determine the best non-specific competitor DNA to use. We have found poly(dA)-poly(dT) to be suitable, but poly(dI)-poly(dC) or sheared salmon sperm DNA may be more suitable for some factors. We found TNE-50 (see Protocol 2) to be a suitable binding and washing buffer. Its low salt concentration encourages the binding of bacterially synthesized proteins which may bind DNA relatively poorly due to lack of posttranslational modifications, etc. Binding activity in TNE-50 should preferably be tested in advance for the factor of interest by either EMSA or Southwestern blot analysis. 3.2.2 Southwestern blotting As the name implies, Southwestern blotting is a variation of the traditional Western blotting technique. Following transfer on to a nitrocellulose membrane, electrophoresed DNA binding proteins are detected with a labelled 128

5: Cloning from a cDNA expression library DNA probe (rather than by immunological means). The method therefore embodies the technique used to screen the library. A protocol for Southwestern blotting is given below (Protocol 2). The sample material could be whole cell extract, nuclear extract, or partially purified DNA binding protein. Protocol 2. Southwestern blotting Equipment and reagents • SW block: 2.5% (w/v) dried milk powder, 25 mM Hepes pH 8.0, 1mM DTT, 10% (v/v) glycerol, 50 mM NaCI, 0.05% (v/v) LDAO (Calbiochem), 1 mM EDTA

• TNE-50: 10 mM Tris pH 7.5, 50 mM NaCI, 1 mM EDTA, 1 mM DTT

Method 1. Subject sample(s) to SDS-polyacrylamide gel electrophoresis and transfer on to a nitrocellulose membrane using standard Western blotting techniques. It is best to load quite a lot of protein (50-100 ug) per lane, especially for crude extracts. 2. After transfer, gently wash the membrane in TNE-50 and place in a tray or sandwich box containing enough SW block to immerse the membrane completely. Block overnight at 4°C. 3. Remove the filter from the SW block solution and wash it briefly in TNE-50. 4. Immerse the membrane in a probe mixture consisting of (a) 10 ug/ml non-specific competitor DNA (b) 2 x 106c.p.m./ml labelled DNA probe (see Section 3.2.3) (c) TNE-50 5. Use a minimum volume, just sufficient to immerse the blot and incubate for 1-2 h at room temperature in a small tray or lunch box (or, alternatively, in a sealed plastic bag). 6. Remove the membrane from the probe solution and wash in 100-200 ml of TNE-50 for 5-10 min at room temperature. (Hybridization and washing at 4°C can be tried if initial attempts are unsuccessful.) 7. Repeat this washing two to three times or until the radioactive level of the membrane no longer falls appreciably between washes. 8. Blot the filter dry, wrap in cling film, and expose to X-ray film.

A successful Southwestern blot would reveal a single hybridizing band with a mobility consistent with the molecular weight of the factor of interest as determined from other studies such as UV cross-linking (see Chapter 3). Two 129

Ian G. Cowell and Helen C. Hurst important pieces of information that are revealed by Southwestern blotting are: (a) If the factor of interest is able to bind the probe DNA after electrophoretic separation, it must be able to bind as a single species, that is, a heterodimeric or heteromeric complex is not required for binding. This is clearly important when screening a library with binding site probes. (b) If more than one band is present on the autoradiograph after Southwestern blotting, the probable explanation is that the probe sequence is bound by a family of factors. In this event it may be advisable to try and determine a more selective probe or be prepared to clone and characterize several members of a family of factors. Multiple bands may also be due to proteins binding the DNA non-specifically. This can be tested by repeating the Southwestern but including a 100-fold excess of cold probe as competitor; remember that if your probe is concatenated then the cold competitor should be too. After autoradiography, any remaining bands are likely to be due to non-specific DNA binding proteins. 3.2.3 Preparation of oligonucleotide probe Described below is the preparation of concatenated binding-site probes. i. Annealing Having decided on the sequence of the oligonucleotide probe to use, it is important to obtain efficient annealing of the two synthesized strands. Take 5 ug each of the upper and lower strand oligonucleotides and place in a 1.5 ml microcentrifuge tube. Add 10 ul of 10 X T4 kinase buffer (500 mM Tris pH 8.0, 100 mM MgCl2, 5 mM DTT, 1 mM spermidine). Make the volume up to 100 ul and place in a water bath at 90°C. Turn the water bath off and allow to cool to room temperature overnight. Store the double-stranded oligonucleotide at -20°C until required. ii. Probe labelling and ligation Double-stranded binding-site oligonucleotides are labelled with T4 polynucleotide kinase and [y32P]ATP and then ligated with T4 DNA ligase. The concatenated probe should then be used within 24 h. Protocol 3. Probe preparation Equipment and reagents • 10 x polynucleotide kinase buffer: 500 mM Tris pH 8.0, 100 mM MgCI2, 5 mM DTT, 1 mM spermidine (make up this solution freshly for maximum labelling efficiency)

. TE: 10 mM Tris pH 8.0, 0.1 mM EDTA

130

5: Cloning from a cDNA expression library Method 1. Mix in a microcentrifuge tube: Double stranded oligonucleotide (0.1

ug/ul)

10 x polynucleotide kinase buffer 32

4 ul 2 ul

[y P]ATP (10 uCi/ul; 5000 Ci/mmole)

5 ul

T4 polynucleotide kinase (10 U/uI)

2 ul

H2O

7 ul

Incubate at 37°C for 60 min. 2. Add 80 ul of TE and separate the labelled oligonucleotide from the excess ATP by spun column chromatography (we use a 1 ml syringe plugged with polyallomer wool and filled with Sephadex G-50M). 3. Check the incorporation of radioisotope by counting 1 ul of the spun column eluate by liquid scintillation counting. (The labelling reaction should yield 1-2 x 108 c.p.m. in total.) 4. To the labelled oligonucleotide (approximately 100 ul) add: 10 x T4 DNA ligase buffer (Biolabs)

11 ul

10 mM ATP (if not present in ligation buffer)

2 ul

T4 DNA ligase (Biolabs, 10 U/ul)

2 ul

5. Ligate overnight at 15-18°C.

3.2.4 Library screening Filter lifts (from Section 2) must be blocked prior to screening with labelled probe. For 10-20 filters place 500 ml of SW block (see Section 3.2.2) in a tray or large sandwich box. Immerse the filters in the blocking solution one at a time ensuring that both surfaces of each filter come into contact with the solution. Block overnight at 4°C. Filters may then be probed as described in Protocol 4. Protocol 4. Filter screening Method 1. Make up the probe mixture by adding the concatenated probe (Protocol 3) to 100 ml of TNE-50 (Protocol 2) containing non-specific competitor DNA at 10 ug/ml. The final radioactive concentration should be 5 x 105-1 X 106c.p.m./ml. 2. Place 50 ml of the probe solution into each of two 150 mm Petri dishes. 3. Lay filters one at a time in the probe mixture, protein side up, placing

131

Ian G. Cowell and Helen C. Hurst Protocol 4.

Continued

only five to six filters in each dish.a It is best to do this as the buffer swirls on an orbital shaker; it is important to have a layer of buffer between each filter to allow efficient hybridization. 4. Incubate for 1 h at room temperature, on a slowly moving orbital platform shaker to keep the filters floating freely. 5. Remove the filters to a large tray or lunch box and wash three to four times for 5-10 min each time in 200-300 ml of TNE-50 at room temperature, making sure that the filters remain free in the solution and do not stick together. Washing can be performed on an orbital platform shaker with fairly vigorous movement.b 6. Blot the filters dry, cover with cling film, and expose to X-ray film including fluorescent or radioactive markers to facilitate orientation of the filters to the developed film. a Process the filters in batches if necessary, re-using the probe. bThe probe mixture may be retained and kept at -20°C for up to 1 week and may be used for second round screens after adding fresh DTT immediately prior to use.

Using Kodak X-omat AR film and an intensifying screen, an overnight exposure at -70°C should be sufficient to detect clones giving strong signals. However, if the background allows, a second, longer exposure (1 week) is also recommended in order to detect clones giving weaker signals. Figure 1 shows the result of typical first and second round screens after a 24 h exposure. In this particular case a concatenated ATF site probe was used to screen for members of the CREB/ATF family. The hybridization conditions described above are given as a guide and represent conditions that we have found to work well. Longer incubation times at a lower temperature may result in a better signal from a poorly binding clone. Similarly, the washing conditions described above are the ones we found to be the most effective. Longer washing times at room temperature are detrimental due to disassociation of the bound probe, but longer washing times with cold buffer can reduce background without significant signal loss. 3.2.5 Guanidinium chloride denaturation/renaturation Some DNA binding proteins may be inactive as bacterially produced fusion proteins due to insolubility or incorrect protein folding. DNA binding activity can sometimes be recovered by denaturation in guanidinium chloride (GnHCl) followed by slow renaturation. This procedure may be applied immediately after making the filter lifts or, if initial screening by the standard method has been unfruitful, the filterbound proteins may be denatured as described below and then re-probed. 132

5; Cloning from a cDNA expression library

Figure 1. An example of a typical first and second round screen of a hgt11 cDNA expression library using a DNA binding site probe, in this case a concatenated oligonucleotide containing an ATF site from the adenovirus E4 promoter. Duplicate positive signals on the primary screen are ringed for clarity. A representative secondary screen is shown beneath.

Protocol 5. Guanidinium chloride denaturation Method 1. Immerse the filters one at a time in 6 M GnHCI using a small tray or sandwich box containing 40 ml of solution per 150 mm filter, 2. Agitate on a slowly moving orbital shaker for 10 min at room temperature. 3. Remove half of the denaturing solution and replace with an equal volume of TNE-50. Shake for a further 10 min, making sure all the filters are moving freely in the buffer and that they are not stuck together. 133

Ian G. Cowell and Helen C. Hurst Protocol 5. Continued 4. Repeat this two-fold dilution of the GnHCI every 10 min for a further seven times and finally discard the solution and replace with TNE-50. 5. Block the filters overnight in SW block. 6. Probe as described in Section 3.2.4.

3.2.6 Faultfinding A list of troublesome symptoms and their possible explanations are given in Table 1. Table 1. Fault finding Symptom

Possible explanation

Spotty background on all filters

Too much labelled DNA in probe mixture Labelled DNA not separated fully from unincorporated radioisotope Insufficient blocking or washing Filters stuck together during washing Filters stuck together during probing Too vigorous washing or too high a temperature—try 4°C Filters require denaturation/renaturation to recover DNA binding activity Factor of interest not represented in the library

High and even or blotchy background on all or some of the filters Light patches on all or some of the filters Blank filters

3.3 Immunological screening If it is possible to purify reasonable amounts of the factor of interest, then two routes to cloning the cognate gene are open. These are either to derive peptide sequence information and thence screen a library with redundant oligos, as described in Chapter 4, or to use the purified protein to raise antibodies for use in screening expression libraries. In some cases, it may be possible to adopt both routes, as peptide sequence information may also be used to make synthetic peptides that, once coupled to a suitable carrier, can be used to raise antibodies in rabbits. 3.3.1 Antibody considerations The use of synthetic peptide to raise polyclonal antisera has the advantage that a large amount of immunogen and hence antiserum can be generated, but in practice peptide antisera often do not recognize the native protein. Consequently, antisera raised against the purified protein are usually preferable. If a small amount of protein has been purified, but the quantity is insufficient to ensure reliable peptide sequencing, then a good compromise may be to use 134

5: Cloning from a cDNA expression library the material to raise polyclonal antisera in mice or rats. The use of mice and rats has the advantage that only small amounts of antigen are required to activate the immune system and additionally, if useful antisera are obtained, they can be used to generate monoclonal antibodies. The obvious disadvantage is that the yield of polyclonal serum is very small and, while it is possible to use monoclonal antibodies for screening expression libraries, generally speaking polyclonal antisera are considered to give better results. One way round this may be to use a mixture of monoclonal antibodies in the screening process. The other problem is whether the purified factor will be sufficiently antigenic to produce an immune response in the first place, especially as nuclear factors are generally highly conserved across species. In practice, though, many factors have been used to raise good antisera. In Protocol 6 we describe a method adapted for using small amounts of purified material which we have used successfully to raise mouse antisera against ATF-43 (14) and AP-2 (15). Protocol 6. Generating nuclear factor antibodies in mice Equipment and reagents • Two or three F1 hybrid mice, e.g. Black 10 x Balbc. Pre-bleed to obtain pre-immune serum

• Affinity purified factor. Each mouse must be injected at least twice and possibly three times with 50-200 ng of material, as estimated by EMSA assay (see Chapter 1)

Method 1. Precipitate 50-200 ng of purified protein by adding cold (-80°C) acetone to 80%. Resuspend in 300 ul of 0.9% saline. If the protein is thought not to precipitate well, then try to concentrate to approximately 100 ng in 300 ul of dialysis buffer (up to 10% glycerol, but no salt) using a Centricon spin concentrator (Amicon). 2. To each 300 ul sample add 90 ul of a sterile aqueous solution of 10% potassium alum (aluminium potassium sulfate) and mix. 3. Carefully add 12 ul of 4 M NaOH and mix well. A thick white precipitate forms consisting of a complex of AI(OH)3 and protein. This adjuvant/ protein suspension can be injected subcutaneously into the mouse at multiple sites. 4. Repeat the sample preparation and injections 3 weeks after the primary injections. Test bleeds can be taken 2 weeks after the secondary injections and if further boosts are required these can be given at 3-4 weekly intervals.

The easiest way to test antibody efficacy is to include 1-2 ul of serum in an EMSA sample. Incubate for 2 h to overnight at 4°C alongside control 135

Ian G. Cowell and Helen C. Hurst incubations containing pre-immune serum, before adding the probe and then proceed as usual (see Chapter 1). If the antibody cross-reacts with the purified factor then the normal protein-DNA complex will either be abolished or further retarded in its migration, depending on which protein domain(s) are recognized (see refs 14 and 15). Promising antisera can then be further tested by Western blotting. This is a good guide as to how well the antibody will recognize protein expressed from a bacterial expression library, as in both cases the antigen is presented on nitrocellulose filters in a partially denatured state. However, if the factor is normally post-translationally modified (e.g. heavily phosphorylated), then the antiserum may not recognize the same protein devoid of these modifications made in bacteria. One possibility is to treat the purified factor before using it as an immunogen—for example, phosphate may be removed using calf intestinal phosphatase immobilized on a column support (Scotlab). Before using polyclonal antisera to screen an expression library it is best to remove any contaminating antibodies to bacterial and phage proteins that these preparations often contain. This will reduce the background in the primary screens. Concentrated bacterial/phage lysate can be obtained commercially and is often provided as part of a screening kit (e.g. Stratagene's picoblue immunoscreening kit). Incubate three to four strips of nitrocellulose in the lysate at room temperature then rinse in PBS and air dry. The filters can then be used sequentially to pull out cross-reacting antibodies from polyclonal serum previously diluted 1:5 or 1:10 in PBST (PBS plus 0.05% Tween-20). This serum may represent the bleed out of one mouse, so care must be taken not to waste the sample. The cleaned up serum can be stored at -20°C until needed. 3.3.2 Library screening Wash the filter lifts from Section 2 (Protocol 1) in PBST and block overnight at 4°C in 500 ml of PBST supplemented with 1% BSA. Place 50 ml of a suitable dilution (1:10 to 1:25) of the cleaned up, diluted serum in blocking solution in a 15 cm Petri dish. Add batches of five to six filters one at a time so that they do not stick together and incubate for 1 h at room temperature with orbital shaking. Wash the filters three times in PBST. The primary antibody can be stored at —20°C and re-used for subsequent screens. This re-use will also reduce the background. The filters are developed either using an enzyme-conjugated second antibody followed by a colour reaction, or 125I-labelled protein A followed by autoradiography. Commercial kits are available to facilitate these steps. Either way, plaques giving positive signals should be picked and rescreened as described in Section 2.3. Once a pure preparation of a positive phage has been achieved then steps should be taken to establish whether this truly encodes the factor of interest. This is most easily done by preparing protein extract from a lysogen and test136

5: Cloning from a cDNA expression library ing this in an EMSA assay (see Section 4). However, if the cDNA clone is incomplete, the DNA binding domain may not be contained within the purified phage and further screening steps using the initial clone as a hybridization probe may be required to obtain a full length clone. 3.3.3 Potential problems with immunological screening One particular problem with immunological screening is that it is possible to clone a protein completely distinct from the intended one. This arises when the material used to inject the animals to raise the antiserum is not totally free of contamination. A co-purifying protein may be a very minor impurity and may not be detected on silver-stained gels, but it could be very immunogenic. Consequently, the polyclonal serum will contain antibodies not only to the factor of interest, but also to this contaminant. We have found that the 68 kDa subunit of human Ku-antigen, which binds to the ends of dsDNA molecules, produces a very strong immunogenic response even when the protein is present in undetectable amounts in highly purified factor preparations (A. Skinner and H. C. H., unpublished data). Unfortunately, this can mean that the investigator ends up cloning a rogue contaminant rather than the factor of interest. Consequently, when picking plaques from primary screens for further purification, it is important to persevere with weak as well as strong signals. Of course the accidental cloning of a co-purifying protein that influences the function of the factor of interest could be highly advantageous!

3.4 Other approaches to library screening and their relative merits We described in Sections 3.3 and 3.2 methods for screening cDNA expression libraries immunologically and with DNA binding-site probes. A third method which we have not discussed so far involves screening of an expression library with a probe consisting of a labelled protein, possibly a subunit of a multicomponent transcription factor. The factor or subunit of interest is then detected by virtue of its protein-protein interaction with the labelled probe. This method is exemplified by MacGregor et al. (16) who screened a cDNA expression library with a biotinylated Jun probe to detect other members of the leucine zipper family with which Jun could dimerize. This screening strategy has value in the cloning of components of a transcription factor complex which alone bind DNA poorly or not at all and in the cloning of ancillary proteins with which transcription factors interact. For example, this so-called farWestern approach has also been adapted to isolate clones for cofactors of the oestrogen receptor which interact only with the ligand-bound, transcriptionally active receptor (17). Expression of cDNA clones in eukaryotic cells provides an alternative to screening bacteriophage libraries. In this approach cDNA libraries are constructed using a high level transient expression vector plasmid. COS cells or 137

Ian G. Cowell and Helen C. Hurst other cells allowing high level expression are transfected with pools of cDNAcontaining plasmids. Nuclear extracts are then prepared from transfected cells and analysed by EMSA for the presence of a binding activity corresponding to the factor of interest. Plasmid DNA is recovered from transfected cells expressing the expected DNA binding activity and recovered DNA is used for successive rounds of selection, until a pure clone is obtained. The use of this approach is exemplified by Tsai et al. (18) who cloned the erythroid specific factor GF-1 (GATA) by transient expression in COS cells. This approach is most appropriate where the DNA binding activity of interest is highly tissue specific or inducible, and it has the advantage that cloned factors expressed in cultured cells are far more likely to receive appropriate post-translational modification than bacterially synthesized factors. In addition this procedure allows rapid comparison between the cloned factor and the endogenous protein in terms of DNA binding specificity and mobility in EMSA. However, as with screening bacteriophage expression libraries with binding-site probes, this strategy is only useful if the binding activity of the factor is contained in a single protein species. The yeast one-hybrid method of Inouye et al. (12) is a relatively new screening method that is gaining popularity for cloning transcription factors. This method employs a yeast (Saccharomyces cerevisiae) strain harbouring a defective HIS3 gene which, as a result, requires histidine-containing medium for growth. First, yeast are transformed with a reporter plasmid containing an intact HIS3 gene under the control of a promoter containing synthetic cis-sites for the factor of interest. In the absence of a transcriptional activator bound at these sites the HIS3 gene is silent. The next step is to transform the resulting yeast strain with a plasmid cDNA expression library constructed such that cDNA inserts are expressed constitutively from a yeast promoter (pADH1) as fusion proteins with a strong activation domain such as the activation domain from the yeast GAM transcription factor. Upon plating on media lacking histidine, only transformants expressing a fusion protein that can bind to the promoter of the plasmid-born HIS3 gene multiply to form colonies. Hopefully these will include factors that bind specifically to the cis-elements. False positives that, for example, bind to other parts of the promoter or flanking DNA are identified by isolating expression plasmid DNA from the resulting 'positive' clones and transforming them back into yeast containing reporter genes lacking the cis-elements. Some workers have designed complementation approaches to obtain cDNA clones for particular transcription factors. For example, Becker et al. (19) cloned the CCAAT binding factor CP1 by complementation of a yeast HAP2 mutant. HAP2 is the yeast homologue of CP1; HAP2 mutants are viable but suffer from a respiratory deficiency. Transformants expressing a clone capable of replacing HAP2 were selected for their ability to grow on minimal agar. Another example of cloning by complementation is provided by Zhou and Thiele (20) who cloned the Candida glabrata metal-activated transcription 138

5: Cloning from a cDNA expression library factor in S. cerevisiae and by Steimle and Mach (21) who describe cloning by complementation in mutant mammalian cell lines. Cloning by complementation is clearly not a universally applicable means of cloning transcription factors and the strategy and detailed methodology will be different in each case. However, it has the advantage that the clone is selected on the relatively stringent basis of expressing a particular function.

4. Proving the identity of the factor When cloning transcription factors from h phage cDNA expression libraries or by other means it is usually desirable to prove as rapidly as possible the identity of the cloned factors. The criteria that are readily tested are listed below: • • • •

DNA binding specificity immunological properties amino acid sequence transcriptional activating/repressing potential

Using the hgt11 system it is possible to generate bacterial colonies carrying an integrated copy of the purified factor-encoding phage (a lysogen). The advantage of this is that a liquid culture can then be grown up and induced to synthesize the cloned factor (usually, but not necessarily, as a fusion protein with B-galactosidase) within the cells. Crude protein extracts prepared from these cells will be enriched in the cloned recombinant protein and can be assayed in much the same way as crude nuclear extracts prepared from eukaryotic cells. A detailed method for generating a lysogen using the bacterial strain Y1089 (Clontech) is given in Protocol 7. To provide a concentrated phage stock for this procedure, make a plate lysate from the purified phage by inoculating at a high density such that total lysis of the Y1090 lawn occurs overnight. Cool the plate and add 4-5 ml of cold SM buffer. Gently shake or tip at 4°C for several hours. Recover the supernatant from the plate and transfer to a sterile tube with a drop of trichloromethane. This can be stored short term at 4°C, but for longer term storage add glycerol to 50% and maintain at —20°C. The plate lysate will contain 5 X 1010 to 1011 p.f.u./ml, but should be titred ready for lysogen generation. Protocol 7.

Generating a lysogen from purified phage stock

Equipment and reagents • L-broth, maltose • L-broth/ampicillin plates

. Shaker at 32°C • Incubators at 32°C and 42°C

139

Ian G. Cowell and Helen C. Hurst Protocol 7.

Continued

Method 1. Inoculate one colony of Y1089a from a freshly streaked L-broth/ ampicillin plate into 5 ml of L-broth plus maltose. Shake to an A600 of 0.4 at 32°C. 2. Take 1 ml of culture and dilute 10-fold with L-broth. 200 ul of this dilution will contain roughly 6.4 x 106 bacterial cells. 3. Infect 200 ul of cells at a multiplicity of 100-200 (i.e. 10 ul of a plate lysate with a titre of 1011 p.f.u./ml) for 20 min at room temperature. 4. Dilute each sample to 10 ml with L-broth and take 15 ul of this dilution and dilute again with 10 ml of L-broth. This final dilution will have approximately 103 bacterial cells/ml. 200-300 ul can be spread on an L-broth/ampicillin plate and grown overnight at 32°C. 5. Plates should contain 200-300 colonies. Replica plate 20-30 of these colonies on to two L-broth/ampicillin plates in an asymmetric grid pattern. Grow one plate at 32°C (the master) and the other at 42°C overnight. 6. Colonies which grow at 32°C, but not at 42°C, can be assumed to contain an integrated copy of the purified phage. a We have found that it was preferable always to maintain cultures of Y1089 at 32°C except when inducing lysogenesis as in step 5.

Using this method we have found that 10-50% of the replica plated colonies in step 5 will be lysogens. The next step is to grow up a colony, induce protein synthesis, and prepare an extract: this is described in Protocol 8. Protocol 8.

Using a lysogen to make a crude protein extract

Equipment and reagents Master plate from Protocol 7 *L-brothwith ampicillin Falcon tubes, 50 ml Shakers at 32°C, 38°C, and 43°C Water baths at 32°C and 43°C *IPTG Lysozyme

* Microcentrifuge, 1.5 ml microcentrifuge tubes . Buffer A: 50 mM Tris pH 7.5, 1 mM EDTA, 1 mM PMSF, 5 mM DTT (ref. 6) • 5 M NaCI * Rotary shaker

Method 1. Inoculate an appropriate colony from the master plate into 5 ml of Lbroth plus ampicillin in a 50 ml Falcon tube. Shake vigorously at 32°C to an A600 of approximately 0.5 (about 2 h). 140

5: Cloning from a cDNA expression library 2. Immerse the tubes in a 43°C water bath to raise the temperature rapidly, then transfer to a 43°C shaker for 20 min. 3. Add IPTG to 10 mM and shake for 1 h at 38°C. 4. Transfer 1.25 ml of culture into each of two 1.5 ml microcentrifuge tubes. Pellet the cells and remove the supernatant. Resuspend each pellet in 100 ul of buffer A. 5. Recombine the two aliquots and lyse the cells with three rounds of freeze-thaw lysis. Add lysozyme to 0.5 mg/ml; incubate for 15 min on ice. 6. Add 55 ul of 5 M NaCI (to 1 M) and place on a rotary shaker for 15 min at 4°C. Spin for 30 min in microcentrifuge at 4°C. The supernatant is a crude bacterial protein extract and will contain the cloned recombinant protein. Process wild type Y1089 alongside to provide a control extract.

The extract can be either diluted or dialysed against buffer A supplemented with 10% glycerol to reduce the salt concentration for binding assays. The protein concentration of a dialysed sample is usually in the range of 3-5 ug/ul, but the exact value is readily determined using a commercial assay (Bio-Rad) and 1-5 ul should be sufficient for a binding assay. Lambda lysogen extracts may be used to test the DNA binding properties of a cloned factor by gel retardation or Southwestern analysis. It may be of value, for example, to compare the DNA binding-site preferences of the cloned factor with the known binding specificity of the factor in question. If antibodies are available, the identity of the cloned factor may also be checked immunologically using the A. lysogen extract in an EMSA assay, since an antibody that recognizes the native protein will usually abolish or alter the mobility of the complex formed between the factor and its binding-site oligonucleotide. Alternatively or in addition to the use of A lysogen extracts, cloned factors may be translated in vitro following in vitro transcription of the cDNA and the translated protein may be used for DNA binding studies as detailed in Chapter 8. Obviously the cloned factor will also need to be sequenced to check that there is an open reading frame of a suitable length to encode the factor of interest. Depending on the result, more cloning of 5' and 3' ends may then be necessary. However, initial clones obtained from bindingsite screens should contain the DNA binding domain, so the derived amino acid sequence can be checked for known DNA binding structures. This is particularly relevant if initial characterization of the factor and its binding site has indicated that this protein should be a member of one of the known families of DNA binding proteins. If a transcription factor is known to activate (or repress) transcription of a given gene, then the ability of the cloned factor to duplicate this effect may be tested by transient expression in eukaryotic cells in culture. In such 141

Ian G. Cowell and Helen C. Hurst experiments a suitable reporter plasmid would contain a marker gene such as chloramphenicol acetyl transferase (CAT) or luciferase, driven by a promoter containing binding sites for the factor of interest. The reporter is cotransfected into cells with a second plasmid which expresses the cloned factor. The transcriptional activating/repressing activity of the cloned factor is then extrapolated from the activity of the reporter gene. Methods of analysing cloned transcription factors are described in more detail in Chapter 8. A more time consuming, but ultimately definitive approach to decide whether the cloned factor equates to the protein initially identified in nuclear extracts is to raise antibodies against the newly cloned factor. This can be done by synthesizing suitable peptides (C- or N-terminal ones work best) determined from the derived amino acid sequence. Alternatively, all or part of the cloned factor may be expressed as a bacterial fusion protein (to facilitate purification, see Chapter 8) which may be used as an immunogen. Clearly if these antibodies interact with both the cloned factor and the purified or partially purified preparations of the factor of interest, then one is virtually home and dry!

References 1. Huynh, T. V., Young, R. A., and Davis, R. W. (1988). In DNA cloning: A practical approach (ed. D. M. Glover), Vol. 1, pp. 49-78. IRL Press, Oxford. 2. Maekawa, T., Sakura, H., Kanei-Ishii, C., Sudo, T., Yoshimura, T., Fujisawa, J.-L, et al. (1989). EMBO J., 8, 2023. 3. Katagiri, F., Lam, E., and Chua, N.-H. (1989). Nature, 340, 727. 4. Hai, T. W., Liu, F., Coukos, W. J., and Green, M. R. (1989). Genes Dev., 3, 2083. 5. Poli, V., Mancini, F. P., and Cortese, R. (1990). Cell, 63, 643. 6. Singh, H., LeBowitz, J. H., Baldwin, A. S., and Sharp, P. A. (1988). Cell, 52, 415. 7. Kageyama, R. and Pastan, I. (1989). Cell, 59, 815. 8. Klemsz, M. J., McKercher, S. R., Celada, A., Van Beveren, C., and Maki, R. A. (1990). Cell, 61, 113. 9. Xiao, J. H., Davidson, I., Matthes, H., Gamier, J.-M., and Chambon, P. (1991). Cell, 65, 551. 10. Williams, T. M., Moolten, D., Burlein, J., Romano, J., Bhaerman, R., Godillot, A., et al. (1991). Science, 254, 1791. 11. Lum, L. S. Y., Sultman, L. A., Kaufman, R. J., Linzer, D. I. H., and Wu, B. J. (1990). Mol. Cell. Biol., 10, 6709. 12. Inouye, C., Remondelli, P., Karin, M., and Elledge, S. (1994). DNA Cell Biol., 13, 731. 13. Fernandes, E. R. and Rooney, R. J. (1997). Mol. Cell. Biol., 17, 1890. 14. Hurst, H. C., Masson, N., Jones, N. C., and Lee, K. A. W. (1990). Mol. Cell. Biol., 10, 6192. 15. Bosher, J. M., Williams, T., and Hurst, H. C. (1995). Proc. Natl Acad. Sci. USA, 92, 744. 16. MacGregor, P. F., Abate, C., and Curran, T. (1990). Oncogene, 5, 451. 17. Cavailles, V., Dauvois, S., L'Horset, F., Lopez, G., Hoare, S., Kushner, P. J., and Parker, M. G. (1995). EMBO J., 14, 3741. 142

5; Cloning from a cDNA expression library 18. Tsai, S. F., Martin, D. I. K., Zon, L. I., D'Andrea, A. D., Wong, G. G., and Orkin, S. H. (1989). Nature, 339, 446. 19. Becker, D. M., Fikes, J. D., and Guarente, L. (1991). Proc, Natl Acad. Sci. USA, 88, 1968. 20. Zhou, P. B. and Thiele, D. J. (1991). Proc. Natl Acad. Sci. USA, 88, 6112. 21. Steimle, V. and Mach, B. (1995). Curr. Opin. Genet. Dev., 5, 646.

143

This page intentionally left blank

6

Cloning transcription factors by sequence similarity ALAN ASHWORTH

1. Introduction Knowledge of the primary structure of different classes of transcription factors has revealed the modular nature of these proteins. Functional domains, such as DNA binding motifs, are often highly conserved between different transcription factors. The sequence conservation of these domains can be taken advantage of to isolate clones related to a particular transcription factor. In some cases, these gene families may have more than 500 members (1).

2. Methods for cloning related transcription factors Two methods are commonly used to isolate clones for related transcription factors. The traditional approach has been to use low-stringency hybridization with a probe for a member of the gene family. More recently, the polymerase chain reaction (PCR) using degenerate oligonucleotides has become popular. Both of these methods have advantages and disadvantages, and the choice of method would depend on the particular gene family to be analysed. Some examples of the different approaches are listed in Table 2. Increasingly, however, with the advent of the various genome projects, computer methods are becoming increasingly important for the isolation of novel members of gene families. These are discussed in Section 2.3.

2.1 Low-stringency hybridization Low-stringency hybridization has been used to isolate clones for members of several transcription factor gene families (Figure 1). It is also particularly suitable for the isolation of the equivalent (orthologous) gene from phyloenetically diverse organisms. An obvious requirement for the use of this method is the availability of a cDNA or genomic clone for the factor in question. This can be generated by the PCR methods described in Protocols 2 and 8. A typical screening procedure is shown in Protocol 1. Note that the most important

Alan Ashworth

Figure 1. Low-stringency hybridization screening of a cDNA library. (A) Approximately 50000 p.f.u, of an 8.5 day mouse embryo cDNA library in Agt10 were plated on to a 140 mm agar plate. Duplicate lifts were hybridized to a Drosophila Kruppel probe at low stringency (see Protocol 1). Filters were washed at low stringency and autoradiographed. Sharp dots are orientation marks. The single 'positive' is indicated by arrows. (B) The area corresponding to the positive plaque in (A) was excised and the phage eluted. Approximately 100 plaques were plated on a 90mm dish and duplicate lifts hybridized as in (A).

parameter is the signal-to-noise ratio and particular attention, therefore, should be paid to reducing background hybridization. Both genomic and cDNA libraries can be screened at low stringency, although the lower complexity of the latter leads to the isolation of fewer hybridization artefacts. Furthermore, eukaryotic genomes contain large numbers of pseudogenes and these can complicate the analysis of the products of a low-stringency screen. 2.1.1 Determining optimal conditions of hybridization It is often useful to determine the optimal conditions for hybridization, before library screening is initiated. This is achieved by probing Southern and/or Northern blots under various stringency conditions. It is convenient to start 146

6: Cloning by sequence similarity with the lowest stringency conditions of hybridization and post-hybridization washing, expose the blot to film and then wash at higher stringency until the background is acceptable. Suggested conditions are given in Protocol 1. In addition, the screening of Northern blots of RNA isolated from different tissues can give an indication of which cDNA libraries could profitably be screened. 2.1.2 Choice of probe Both cloned probes and labelled oligonucleotides have been used successfully. Cloned probes are generally used when the region of conservation is extensive. Probes from lower organisms such as Drosophila have frequently been used for the isolation of related mammalian genes containing motifs such as homeoboxes and zinc fingers (2, 3). Oligonucleotides are used when the region of homology is more restricted. As with PCR, degeneracy of the oligonucleotides can be reduced by considerations of codon preference (see Section 2.2.1 (c)) (4). Oligonucleotides are particularly useful for the isolation of clones containing multiple copies of a motif, as with the zinc finger (5) (Table 2). Protocol 1. Low-stringency hybridization screening of cDNA and genomic libraries Equipment and reagents • 5 X SSC,a 0.5% SDS, 1 mM EDTA

> 6 x SSC, 0.1% SDS, 5 x Denhardt's reagent,b 100 ug/ml salmon sperm DNA,C (50% formamide, optional)

Method 1. Plate out a bacteriophage, or a plasmid or cosmid library, on to appropriate agar plates and take lifts with nylon membranes using standard protocols (6). Duplicate lifts are strongly recommended to distinguish genuine signals (see Figure 1). 2. Fix DNA to the filters according to the manufacturer's instructions. 3. Pre-wash the filters for 2 h in 1 litre of a solution containing 5 x SSC, 0.5% SDS, and 1mM EDTA at 65°C. This removes much of the bacterial debris and considerably reduces the non-specific background. Briefly rinse the filters in the same solution. 4. Pre-hybridize the filters in either 6 x SSC, 0.1% SDS, 5 x Denhardt's reagent, and 100ug/ml salmon sperm DNA at 50-65°C or the same solution containing 50% formamide at 30-42oC. At least 4 h of prehybridization are required for minimal background hybridization. 5. Add the 32P-labelled probe and hybridize overnight at the above temperatures.

147

Alan Ashworth Protocol 1.

Continued

6. The filters should be washed extensively with large volumes and several changes of buffer. A reasonable starting point for washing is 2 x SSCa/ 0.1% SDS at room temperature, although higher salt concentrations (and hence lower stringency) can be used. Care should be taken not to allow the filters to dry out before autoradiography as this makes subsequent higher stringency washing ineffective. a1 x SSC is 0.15 M NaCI and 0.015 M sodium citrate (pH 7.0). b 50 x Denhardt's reagent is 5% ficoll, 5% polyvinylpyrrolidone, and 5% bovine serum albumin (fraction V). cSalmon sperm DNA should be sonicated and denatured before use (6).

2.2 Polymerase chain reaction approaches The advent of PCR technology (7, 8) has facilitated the isolation of clones of many members of gene families which could not have been isolated by lowstringency hybridization. This is due, in part, to the minimal overall sequence homology required to construct PCR primers (Figure 2). These approaches also have the advantage that many samples can be screened for homology and that the procedure is relatively rapid. The disadvantage is the considerable number of artefactual sequences that can be generated.

Figure 2. Degenerate PCR primers for POU-domain genes. The figure shows a schematic diagram of the structure of the bipartite POU domain. Degenerate primers are shown which are complementary to a consensus of previously isolated POU-domain proteins (19, 24).

2.2.1 Selection of PCR primers The selection of PCR primers is of critical importance. Some general principles are: (a) Select highly conserved sequences as primers. If possible, select them on the basis of conservation across species (the more distantly related the better) as well as conservation within the gene family. 148

6: Cloning by sequence similarity (b) Choose sequences of minimal degeneracy (see Table 1): (i) peptides containing tryptophan and methionine are preferred as they are specific by unique codons (ii) avoid peptides containing leucine, arginine, and serine, if possible, as they result in a large increase in the number of degeneracies required. (c) If the primer sequences are extremely degenerate, the degeneracy can reasonably safely be reduced by taking account of the under-representation of CpG dinucleotides in vertebrate genomes. For exmple, in a primer coding for a peptide containing histidine followed by glutamic acid, CA(T,C)GA(A,G) would normally be used. However, it is statistically unlikely that the CAC codon is used as it is followed by a G. Degeneracy at other positions could be reduced by utilizing codon preference tables (4) or by including inosine at positions of degeneracy. During the course of the PCR, many mismatched primers are, in fact, incorporated into the product (see Figure 3). (d) Residues at the 3' end of the primers are most important in conferring specificity on the PCR. It is best, therefore, to have as many unamiguous residues at the 3' end as possible. Alternatively, and perhaps surprisingly, it appears that T as the 3' residue is relatively neutral when mismatched to G, C, or T (9). (e) The length of the PCR product should optimally be 200-l000bp, depending on the degeneracy of the primers. Fewer artefactual bands appear to be generated when amplifying small products with highly degenerate primers. (f) If more than two regions in a protein are highly conserved, several sets of PCR primers can be constructed. This allows the use of 'nested' PCR which increases the specificity considerably. Multiple PCR primers have been used to clone a human TFIID gene, with the yeast sequence being used before it was known which parts of the protein are highly conserved (10). Examples of the use of highly degenerate primers to isolate clones related to various transcription factors are discussed in Section 4. 2.2.2 PCR conditions In general, it is worth trying the standard conditions given in Protocol 2 first. These have frequently been successful for cross-species PCR (see reference 11). However, it is often necessary to optimize the efficiency of the PCR when highly degenerate primers are used. Some of the parameters that can be varied are given below. i. Annealing temperature In general, the lower the annealing temperature the more the chance of nonspecific amplification. This will, of course, depend on the degeneracy of the 149

Table 1. Degeneracy of amino acid codons for selection of PCR primers Number of codons

One

Two

Three

Four

Sixa

Methionine (ATG) Tryptophan (TGG)

Asparagine (AAQ) Aspartic acid (GAQ) Cysteine (TGQ) Glutamic acid (GAP) Glutamine (CAP) Histidine (CAQ) Lysine (AAP) Phenylalanine (TTQ) Tyrosine (TAQ)

Isoleucine (ATR)

Alanine(GCX) Glycine(GGX) Proline (CCX) Threonine(ACX) Valine (GTX)

Arginine (SGX) Leucine (QTX) Serine (VWX)

Q = A or T; P = C or T; R = T, C, or A; S = C or A; V = A or T; W = G or C; X = T, C, G, or A. aThese amino acids are encoded by six codons but introduce eight (arginine and leucine) or sixteen (serine) degeneracies.

6: Cloning by sequence similarity primers. Annealing temperatures as low as 37°C can be used with some highly degenerate primers with little artefactual amplification in a standard PCR (P. Denny and A. Ashworth, unpublished observations). A method, called 'Touchdown PCR' (12) has been developed to attempt to reduce spurious amplification. In this procedure, amplification starts off above rather than at or below the expected annealing temperature. The annealing temperature is then decreased (by 1-2°C) every 1-2 cycles to a 'Touchdown' temperature at which 10-20 cycles are performed. This gives a selective advantage early in the PCR to highly matched sequences. This method was utilized in the cloning of the small subunit of the general trancription factor TFIIE (13). Highly degenerate PCR primers were designed based on the sequences of purified TFIIE peptides. A PCR starting at an annealing temperature of 54°C and dropping by 2°C each cycle to a touchdown of 34°C was used. This annealing temperature was used for a further 40 cycles. The PCR product was sequenced and shown to be specific for TFIIE. ii. Number of cycles of amplification The use of an excessive number of cycles can lead to artefactual amplification. The minimum number necessary (usually 30-50) should be determined by removing an aliquot and testing the specificity of amplification every 5 cycles or so. iii. Ramp time The rate of cooling from the denaturing to the annealing stage can be varied on some PCR machines. In general, a slower ramp time should favour mismatch hybridization. iv. PCR buffer composition The major component of the PCR buffer that can usefully be varied is the concentration of Mg2+ ions. Multiple PCRs should be set up with final Mg2+ concentrations varying in 0.25 mM increments between 1.5 mM and 3.0 mM. This can have a considerable effect on the final yield of PCR product (see Figure 4) as well as the specificity, depending on the primers used. 2.2.3 Choice of template Genomic DNA, cDNA reverse-transcribed from RNA, and cDNA libraries are all suitable for PCR with degenerate primers. i. Genomic DNA About 0.25-1.0 ug genomic DNA can be included in a standard PCR reaction (Protocol 2). Clearly, large introns should not be present within the region to be amplified if this approach is to be successful. In addition, there is the danger that pseudogenes, of which there are many in vertebrate genomes, will also be amplified when genomic DNA is used as template. The advantage 151

Alan Ashworth

Figure 3. Incorporation of mismatched primers into PCR products. Primer sequences incorporated into PCR products are compared with the relevant template sequences for the genes Oct-1, Oct-2, and Oct-11. DP1 and DP2 correspond to the forward and back degenerate primers in Figure 2 (19, 24). Clearly, many of the PCR products have incorporated mismatched primers (mismatches are indicated by underlining). Some of the products have incorporated bases not present in the degenerate oligonucleotides (indicated by lower-case letters). Dots indicate bases not determined (A.S. Goldsborough, personal communication).

152

6: Cloning by sequence similarity

Figure 4. Determination of optimal Mg2+ concentration for PCR. PCRs were set up as described in Protocol 2 using buffers differing only in their Mg2+ concentration. The Mg2+ concentrations were: 1, 1.5 mM; 2, 1.75 mM; 3, 2.0 mM; 4, 2.25 mM; 5, 2.5 mM; 6, 2.75 mM; 7, 3.0 mM. The optimal concentration for this set of primers was 2.25 mM,

IS that one docs not need prior knowledge of where the related gene is expressed.

ii. cDNA This can be synthesized using the method described in Protocol 8. The amount of cDN A included in the PCR should be varied as the yield of reverse transcription reactions is variable. Priming of the cDNA can be with oligo-dT, random hexamers. or a specific oligonucleotide. In addition, the use of the RACE (rapid amplification of cDNA ends) protocol should allow PCR with only a single degenerate oligonucleotide, although we have not tried this.

iii. cDNA libraries High-titre bacteriophage and plasmid cDNA libraries can be used as temlates for PCR. This has several advantages. The PCR product generated can be used directly to screen the Library for longer clones by filter hybridization. The fact that the insert in the cDNA library is flanked by known vector sequences allows alternative PCR approaches to be used. PCR can be performed with one degenerate primer and a vector primer ('in-and-out' PCR). If a specific product is generated, it can be sequenced directly using the vector primer (see below). An extension of this technique is to fractionate a cDNA library and analyse the fractions by PCR, cither with two degenerate primers or by the in-and-out procedure (Protocol 3). The PCR product itself can then be used as a hybridization probe to screen the appropriate fraction. Protocol 2.

Polymerase chain reaction—standard conditions

1. Assemble the components of the reaction at room temperature as follows: • 10 X PCR buffer: 150 mM Tris-HCI pH 8.8, 600 mM KCI, 22.5 mM MgCI2 • 10 mM dNTPs (Pharmacia or Boehringer-Mannheim) • of each primer at 10 pmol/ul • Taq Polymerase (Perkin-Elmer Cetus, many other sources are also suitable)

•

H2O 153

5 ul 1.25 ul 1 ul 2.5 units

to 49 ul

Alan Ashworth Protocol 2.

Continued

It is often convenient to prepare a master mix of these components which can then be divided into aliquot parts before addition of template. 2. Add the template last to reduce chances of cross-contamination. Thorough mixing appears to be unnecessary. The template may be genomic DNA (50-500 ng), cDNA (see Protocol 8), or cDNA libraries (1-5 ul). Care should be taken over the buffer in which the template is added as the addition of extraneous Mg2+ ions and EDTA could alter the efficiency of the PCR. 3. Overlay the reactions with two drops (50-100 ul) of light mineral oil (Sigma No M-3516) and close the tubes. 4. Subject the samples to cycles of denaturation, annealing, and extension on a thermal cycler. Exact conditions will depend on the cycler used, but the following conditions have proved suitable for a range of primers and templates. Thirty cycles of: (a) denaturation at 94°C for 30 sec (b) annealing at 50-55°C for 30 sec (c) extension for 60 sec at 72°C followed by 5min at 72°C. The extension time has been found to be sufficient for the synthesis of greater than 1.5 kb. 5. The PCR products can be loaded, with care, directly on to agarose (11.6% regular agarose or 3% NuSieve GTG agarose (FMC) + 1% regular agarose) gels or polyacrylamide (10-15%) gels. If the mineral oil is a problem, it can be extracted by the addition of 100 ul trichloromethane. Alternatively, place the PCR tubes at —20°C until the aqueous phase freezes. The mineral oil remains liquid and can be removed with a pipette. This is particularly useful if the PCRs are performed in microtitre well dishes (14).

Protocol 3. Screening libraries by PCR 1. Plate out cDNA or genomic library in a A vector at a density of 5000-40000 plaques per 140 mm bacterial dish. 2. Grow for 12-16 h to near confluence. Lifts can be taken at this stage with nylon filters, or one can proceed directly to the next step. 3. Pipette approximately 8 ml of 100 mM NaCI, 10 mM MgCI2, 10 m TrisHCI (pH 8.0) on to the surface of each of the plates and incubate at room temperature with shaking for 1-2 h. 154

6: Cloning by sequence similarity 4. Recover the buffer containing the amplified fractions of the library and analyse by PCR as described in Protocol 2. 5. PCR products are subcloned and sequenced. In addition, they can be labelled with 32P and used as probes to screen the filters generated in step 2. Alternatively, the positive fractions of the library can be replated and fresh filters generated for hybridization with the 32Plabelled probe.

2.2.4 Sequencing PCR products Most PCR products derived by the methods described above will be mixtures of products of more than one gene. As such they cannot be sequenced directly and must be subcloned. i. Subcloning PCR products Various methods are available for subcloning PCR products. The addition of restriction enzyme recognition sites to the end of the oligonucleotide primers used for amplification should allow the cloning of the digested PCR products into similarly digested vectors. However, many restriction enzymes fail to cleave efficiently if their recognition sequences are located too close to the ends of the oligonucleotide. We find, however, that direct cloning of PCR products either by blunt-end ligation (Protocol 4) or by ligation into T-tailed vectors (15) (Protocol 5) is very efficient. Both these approaches can be used for the subcloning of essentially any PCR product. T vectors take advantage of the A residue frequently added to the 3 end of the PCR product during the PCR itself. Protocol 4. Cloning PCR products by blunt-end ligation 1. PCR products electrophoresed in agarose or polyacrylamide gels are purified using the Geneclean or Mermaid (BIO101) procedures. Mermaid is used for fragments of less than 400 bp. The DNA is eluted into 13 ul H2O. 2. Add 2 ul 10 x kinase buffer (700mM Tris-HCI pH 7.5/100mM MgCI2), 2 ul 100mM dithiothreitol (DTT), 2 ul 10mM ATP, and 1 ul ((5 units) T4 polynucleotide kinase. Mix and incubate at 37°C for 30min. 3. Add 3 ul 0.5 mM dNTPs and 1 ul Klenow polymerase (5 units). Mix and incubate at room temperature for 15 min. 4. Purify DNA by the Geneclean or Mermaid procedures. Elute into 16 ul H2O. 5. Mix with 2 ul 10 x ligase buffer (Boehringer), 1 ul (5-10 units) T4 DNA ligase, and 1 ul (20 ng) of a suitable blunt-ended vector that has been treated with alkaline phosphatase. A control ligation containing H2O

155

Alan Ashworth Protocol 4. Continued instead of the PCR product should also be set up to assess the level of vector background. A suitable vector is pBluescript (Stratagene) cut with EcoRV and treated with calf intestinal alkaline phosphatase. 6. After a minimum of 2 h, transform 10 ul of the ligations into E. coli competent cells (JM83 or XL-1Blue, or others for pBluescript) and plate on to L-agar plates containing appropriate supplements such as ampicillin and X-gal.

Protocol 5. Cloning PCR products using T-tailed vectors 1. Digest 2 ug pBluescript or other suitable plasmid with EcoRV and purify the linearized plasmid by agarose gel electrophoresis and the Geneclean procedure. 2. Incubate the digested plasmid with Taq polymerase (2 units) in 1 x PCR buffer (see Protocol 2) in the presence of 2 mM TTP for 2 h at 72°C. Purify by the Geneclean procedure and elute into 50 ul H2O. 3. Set up ligations of 1 ul of this vector with aliquots of the gel-purified PCR products and transform into suitable bacteria as in steps 5 and 6 of Protocol 4.

H. Sequencing subcloned PCR products It is most convenient to sequence directly from bacterial colonies harbouring plasmids containing the subcloned PCR product. This is done by PCR on the bacterial colonies with primers flanking the cloning site in the plasmid used (Protocol 6). The PCR product may then be sequenced directly using one of the methods (such as Protocol 7) available for this. These procedures allow the rapid sequence analysis of the products of the PCR (Figure 5). Protocol 6. Amplification of DNA from bacterial colonies-'wiggles' 1. Make up PCR mixtures as in Protocol 2. Primers should be complementary to regions flanking the restriction site used for subcloning. They should be further than 20 bp away from the cloning site to enable the full sequence of the insert to be determined (Protocol 7). 2. Using a disposable microbiological loop, pick a bacterial colony, twirl it in the PCR mix for 2-3 sec and then place the loop into a culture tube containing 5 ml L-broth + 50 ug/ml ampicillin. 3. Add mineral oil to the PCR tubes and cycle as in Protocol 2. Shake the cultures at 37°C for 5 h to overnight.a

156

R: Cloning by sequence similarity 4. Analyse the PCRs by agarose gel electrophoresis. Excise bands, purify the DNA by the Genecfean or Mermaid (BIO101) procedures, and sequence directly by the method described in Protocol 7. aAlternatively, PCR can be performed on the bacterial culture after a few hours of growth as follows: centrifuge 100 ul bacterial culture for 1min in a microfuge, discard the supernatant and resuspend the pellet in 50 ul H2O, Boil the tube for 5min, centrifuge for 1min, and use VI of the supernatant in the PCR.

Figure 5. Direct sequencing of PCR products. DNA was amplified as described in Protocol 2 and sequenced directly as described in Protocol 7. The sequencing reactions were electrophoresed on a thin polyacrylamide/urea wedge gel and the gel was dried and autoradiographed.

Protocol 7.

Direct sequencing of PCR products

1. Mix 7 ul (0.25-1.0 ug) PCR product (Protocol & with 1 ul (10 pmol) of the appropriate sequencing primer and heat at 95-100oC for 5 min. This is most conveniently done in a heating block. One of the primers from the PCR (Protocol 6) or an internal primer may be used for sequencing. 157

Alan Ashworth Protocol 7. Continued 2. Place the tube on dry ice for 1min to freeze, then centrifuge briefly (2-3 sec) in a microcentrifuge. 3. Add 7.5 ul of a previously prepared reaction mix composed of:a • 5 X sequencing buffer: 200 mM Tris-HCI pH 7.5, 100 mM MgCI2, 250 mM NaCI

2 ul

• 100 mM DTT

1 ul

• diluted labelling mix (7.5 uM dCTP, dGTP, and TTP; this is diluted 1:4 in water) • [a35S]dATP (Amersham, 37 x 1012 Bq/mmol)

2 ul 0.5 ul

• T7 DNA polymerase (Pharmacia; diluted in 10 mM Tris-HCI pH 7.5, 5 mM DTT, 0.5 mg/ml bovine serum albumin) 2 ul (2 units) 4. Centrifuge briefly and incubate at room temperature for 2 min. During this time add 3.5 ul of the mixture to the lip of each of four microcentrifuge tubes containing T, C, G, or A termination mixes.b The termination reaction is initiated by brief centrifugation. The tubes are placed in a 37°C water bath and incubated for 5min. 5. Following addition of 4 ul of formamide/dye mix (95% formamide, 20 mM EDTA, 0.05% bromophenol blue, 0.05% xylene cyanol FF) the reactions are heated at 95°C for 2-3 min and loaded on to a 6% polyacrylamide/7 urea sequencing gel. a0.5 ul MnCI2/sodium isocitrate buffer (16) can be included if sequences close to the primer site are required. bTermination mixes have the following composition: • T termination mix: 80 uM dNTPs, 50 mM NaCI, and 8 uM ddTTP • C termination mix: 80 uM dNTPs, 50 mM NaCI, and 8 uM ddCTP . G termination mix: 80uMdNTPs, 50 mM NaCI, and 8uMddGTP • A termination mix: 80 uM dNTPs, 50 mM NaCI, and 8 uM ddATP.

2.2.5 Artefacts Great care should be taken with PCR approaches to avoid and recognize artefacts. (a) The PCR is very prone to artefacts due to its extreme sensitivity. Hence the usual precautions (see PCR, A Practical Approach in this series) should be taken to avoid contamination from plasmid stocks containing amplifiable sequences. If necessary, pipettes, tubes and buffers can be treated with UV light to eliminate contamination from nucleic acids (17). (b) Depending on the conditions used, the use of degenerate primers can produce large numbers of artefactual sequences due to mispriming. These can be reduced by careful titration of the PCR conditions (see Section 2.2.2 above). 158

6: Cloning by sequence similarity (c) The inherent error rate of Taq polymerase is significant and all sequences should be assembled from more than one PCR product. The use of thermostable DNA polymerases with much lower error rates (such as Pyrococcus DNA polymerase) should circumvent this problem. (d) Often a specific or artefactual product will amplify preferentially, making the isolation of other related sequences difficult. This can be partially alleviated by restriction digestion of the PCR product with an enzyme specific for the preferentially amplified product. (e) Occasionally, sequence analysis of a single subcloned PCR product can yield a mixed sequence where bands are observed in two tracks of a sequencing gel. This is almost certainly due to cloning of a heteroduplex. This would be replicated after transformation, resulting in the bacterium in question containing two distinct plasmids. (f) Chimeric or recombinant sequences can be formed in the PCR if closely related sequences are present. This will be a particular problem at high product concentration towards the end of the PCR. Increasing the elongation time has been reported to reduce the frequency of this phenomenon. (g) Yeast sequences are represented in some commercially available cDNA libraries, probably due to the use of yeast RNA as a carrier in cDNA preparation (18). As a consequence, care should be taken to confirm that the cDNA is derived from the appropriate organism. (h) The isolation of a PCR product from a particular cDNA sample does not necessarily indicate that this gene is significantly expressed in that tissue. We isolated a novel octamer binding protein gene, Oct-11, by PCR from testis cDNA (19). Subsequently, however, we were unable to detect expression of Oct-11 in the testis by Northern blotting. Furthermore, we were unable to isolate Oct-11 cDNA clones after extensive screening of testis cDNA libraries (A. S. Goldsborough and A. Ashworth, unpublished observations). We presume that the extreme sensitivity of the PCR technique led to the isolation of the Oct-11 gene which is not significantly expressed in the testis. Thus care should be taken to confirm that the isolated PCR product is meaningfully expressed. Specific PCR primers can be designed and the expression of the gene studied by RT-PCR (Protocol 8) and the results can be confirmed by Northern blotting. Protocol 8. Reverse transcription of RNA for PCR analysis 1. Prepare RNA (total or poly-A enriched) from tissues or cells by any standard method. 2. Set up reverse transcription by mixing the following components on ice: • 5 X reverse transcription buffer (Gibco-BRL)

159

4 ul

Alan Ashworth Protocol 8. Continued • 10 mM dNTPs

1 ul

. 100 mM DTT

2 ul

• Primer-oligo dT12-18, random hexamers, or a specific primer 1 ul (10 pmol • Human placental ribonuclease inhibitor

• H2O

0.5 ul (5 units)

10 ul

• Moloney murine leukemia virus reverse transcriptase (Gibco-BRL) 0.5 ul (100 units) 3. Heat ul RNA at 70°C for 2 min, chill on ice and add to the reaction mix. Incubate at 37°C for 30 min. 4. Dilute to 50-100 ul with H2O. Include various amounts (usually 1 ul) of the cDNA in PCR reactions with suitable primers. Analyse the products by gel electrophoresis.

2.3 In silico approaches to identifying novel transcription factors The various genome projects that have been initiated over the past few years have led to an explosion in the amount of DNA sequence data available. The complete sequences of the genomes of many microorganisms, including the eukaryote Saccharomyces cerevisiae, are now available. The sequence of the nematode C. elegans is likely to be completed soon, with the majority of the human genome sequence being available by the first few years of the twenty-first century. In the meantime, other sequence databases generated by the random sequencing of cDNA libraries, the EST (expressed sequence tag) databases, are proving to be rich sources for the identification of novel members of gene families. These databases (available at the National Center for Biotechnology Information (NCBI); web site http://www.ncbi. nlm.nih.gov/) can be interrogated via text searching for genes that have been pre-assigned to families or by sequence similarity. Software is available at the NCBI to search for sequence similarity to either DNA or peptide sequences. More complex searches, such as searching the translation of a DNA sequence in all reading frames against the translation of a database in all reading frames, are also available. This type of search can, in some cases, prove more efficient in detecting weaker similarities. An example of a database search performed at NCBI is shown in Figure 6. The amino acid sequence of mouse Sox6, a member of the Sry-related family of transcription factors (20), was searched against a translation (in all possible reading frames) of the dbEST database at NCBI, using the programme tBLASTn (21). Multiple significant 'hits' were obtained (Figure 6a) including some previously undescribed genes. Of particular interest was an EST clone

160

Table 2. Examples of novel transcription factors isolated by PCR or low-stringency hybridization Class of transcription factor Homeobox proteins

Zinc-finger proteins

POU-domain proteins

Helix-loop-helix proteins

Conserved domain

Notes

References

60 amino acid DNA binding domain (homeobox)

Conserved domain originally described in Drosophila homeotic genes. Present in a wide range of eukaryotic organisms. Low-stringency hybridization with Drosophila probes, hybridization with degenerate oligonucleotides, and PCR have been used to isolate

2, 22, 23

The zinc-finger motif was originally described in transcription factor IIIA from Xenopus. The Drosophila Kruppel gene and other probes have been used to isolate colonies for zinc-finger proteins. In addition, hybridization with degenerate oligonucleotides and PCR have also been used. Mammalian genomes may contain more than 500 genes of the zinc-finger class Conserved region originally described in mammalian proteins Pit-1, Oct-1, and Oct-2 and in the product of the nematode Unc-86 gene, hence referred to as POU family of genes. PCR has been utilized to isolate large numbers of POU genes. Low-stringency hybridization has also been used Homology first identified among c-myc, the muscle determination gene MyoD, and genes of the Drosophila achaete-scute complex. Many other members now identified.

1,3

members of this gene family 28-30 amino acid motif frequently tandemly repeated and based on the motif CX2CX3FX5LX2HX3HTGEKP(F,Y)

Tripartite structure; 75-80 amino acid POU-specific domain, a non-conserved linker, and a 60 amino acid carboxy-terminal divergent homeodomain Two segments of 12-15 amino acids capable of forming amphipathic a helices connected by a non-conserved loop region of varying length. Associated with a 10-20 amino acid basic region

19, 24-26

27, 28

Alan Ashworth

Figure 6. Identification of novel members of gene families by database searching. The complete 827 amino acid sequence of mouse Sox6 (20) was searched against an allframe translation of the dbEST database at NCBI using the tBLASTn programme (21). (a) List of some of the significant matches generated. Some correspond to previously described genes, such as the mouse Sox6 gene itself (AA646726), whereas others are novel. (b) Alignment of mouse Sox6 with an EST (AA428135) identified in the above search. The aligned sequence corresponds to the carboxyl terminus of mouse Sox6 and probably encodes the previously uncharacterized human SOX6.

162

6: Cloning by sequence similarity AA428135, derived from a cDNA library made from a human fetus, which showed high sequence similarity to the carboxyl terminus of mouse Sox6 (Figure 6b). Given the very high sequence similarity, this clone seems likely to encode human SOX6. This clone could be a starting point for the isolation of the complete human SOX6 cDNA, either via further searching of the dbEST database or via more conventional library screening approaches. Similarly other clones identified via this search could be characterized to identify novel Sox genes. It seems likely, as more sequence becomes available for various organisms, that database searching will be the most important route for the identification of novel members of gene families.

3. Some examples of the cloning of transcription factors by homology Some examples of the use of low-stringency hybridization and PCR to isolate members of transcription factor families are given in Table 2. This list is not intended to be exhaustive, but represents the best-documented examples of the use of these techniques.

Acknowledgements Research in my laboratory is supported by the Cancer Research Campaign, the Medical Research Council, and the Wellcome Trust. Thanks to Andy Goldsborough and Paul Denny for unpublished information and evolving many of the protocols described here.

References 1. Ashworth, A. and Denny, P. (1991). Mammalian Genome, 1, 196. 2. McGinnis, W., Garber, R. L., Wirz, J., Kuroiwa, A., and Gehring, W. J. (1984). Cell, 37, 403. 3. Chowdhury, K., Deutsch, U., and Grass, P. (1987). Cell, 48, 771. 4. Lathe, R. (1985). J. Mol. Biol., 1830, 1. 5. Bellefroid, E. J., Lecocq, P. J., Benhida, A., Poncelet, D. A., Belayew, A., and Martial, J. A. (1989). DNA, 8, 377. 6. Sambrook, J., Fritsch, E. F., and Marcatis, T. (ed.) (1989). Molecular cloning, a laboratory manual (2nd edn). Cold Spring Harbor Press, Cold Spring Harbor, NY. 7. Saiki, R., Scharf, S., Faloona, F., Mullis, K. B., Horn, G. T., Erlich, H. A., and Arnheim, N. (1985). Science, 230, 1350. 8. Saiki, R. K., Gelfand, D. H., Stofel, S., Scarf, S. J., Higuchi, R., Horn, G. T., Mullis, K. B., and Erlich, H. A. (1988). Science, 239, 487. 9. Kwok, S., Kellogg, D. E., McKinney, N., Spasic, D., Goda, L., Levenson, C., and Sninsky, J. J. (1990). Nucleic Acids Res., 18, 999. 163

Alan Ashworth 10. Hoffmann, A., Sinn, E., Yamamoto, T., Wang, J., Roy, A., Horikoshi, M., and Roeder, R. G. (1990). Nature, 346, 387. 11. Ashworth, A., Rastan, S., Lovell-Badge, R., and Kay, G. (1991). Nature, 351, 406. 12. Don, R. H., Cox, P. T., Wainwright, B. J., Baker, K., and Mattick, J. S. (1991). Nucleic Acids Res., 19, 4008. 13. Peterson, M. G., Inostroza, J., Maxon, M. E., Flores, O., Admon, A., Reinberg, D., and Tjian, R. (1991). Nature, 354, 369. 14. Lennon, G. G., Drmanac, R., and Lehrach, H. (1991). Biotechniques, 11, 185. 15. Marchuk, D., Drumm, M., Saulino, A., and Collins, F. S. (1991). Nucleic Acids Res., 19, 1154. 16. Tabor, S. and Richardson, C. C. (1989). Proc. Natl Acad. Sci. USA, 86, 4076. 17. Sarker, G. and Sommer, S. S. (1990). Nature, 343, 27. 18. Lovett, M., Kere, J., and Hinton, L. M. (1991). Proc. Natl Acad. Sci. USA, 88, 9628. 19. Goldsborough, A. S., Ashworth, A., and Willison, K. R. (1990). Nucleic Acids Res., 18, 1634. 20. Connor, F., Wright, E., Denny, P., Koopman, P., and Ashworth, A. (1995). Nucleic Acids Res., 23, 3365. 21. Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. J. (1990). J. Mol. Biol. 215, 403. 22. Crompton, M. R., MacGregor, A. D., and Goodwin, G. H. (1991). Leukemia, 5, 357. 23. Singh, G., Kaur, S., Stock, J. L., Jenkins, N. A., Gilbert, D. J., Copeland, N. G., and Potter, S. S. (1991). Proc. Natl Acad. Sci. USA, 88, 10706. 24. He, X., Treacy, M. N., Simmons, D., Ingraham, H. A., Swanson, L. W., and Rosenfeld, M. G. (1989). Nature, 340, 35. 25. Herr, W., Sturm, R. A., Clerc, R. G., Corcoran, L. M., Baltimore, D., Sharpe, P. A. et al. (1988). Genes Dev., 2, 1513. 26. Scholer, H. R., Ruppert, S., Suzuki, N., Chowdhury, K., and Grass, P. (1990). Nature, 344, 435. 27. Benezra, R., Davis, R. L., Lockshon, D., Turner, D. L., and Weintraub, H. (1990). Cell, 61, 49. 28. Davis, R. L., Weintraub, H., and Lassar, A. B. (1987). Cell, 51, 987.

164

7

Identification of target genes for a transcription factor by genomic binding-site cloning SATOSHI INOUE, SHIGERU KONDO, and MASAMI MURAMATSU

1. Introduction Whereas a great number of genes or cDNAs are isolated by molecular cloning, relatively few genes are known with respect to their function. In order to gain more insight into the life processes of higher eukaryotes, including humans, a greater understanding of gene function is of prime importance. Transcription factors are known to be involved in a variety of cell functions including growth, development, and differentiation. Little is known, however, of the molecular mechanisms by which they regulate these phenomena. In this context, identification of the target genes for a cloned transcription factor is at present one of the crucial issues in molecular biology. We have established a method to isolate the binding sites for a transcription factor in genomic DNA. Using genomic DNA fragments containing the transcription factor binding sites, target genes can be screened and isolated. Here we describe, as an example, our attempt to isolate oestrogen responsive genes from human DNA. Oestrogen receptor (ER), a transcription factor, mediates oestrogen action by binding to the oestrogen responsive element (ERE) that exists in the enhancer region of target genes, regulating their ligand-dependent transcription (1, 2). Despite the wide variety of oestrogen actions, only a few genes are identified that are directly responsive to this hormone. In order to isolate more oestrogen responsive genes, we isolated ER-binding fragments from human genomic DNA with recombinant ER protein and identified novel oestrogen-responsive genes present in the adjacent regions (3-5).

2. Method for genomic binding-site (GBS) cloning Among several methods for isolation of the target gene for a transcription factor, the GBS cloning method (3-5) is applicable if the recombinant protein

Satoshi Inoue et al. of the transcription factor is available in sufficient amount. In this method, target elements can be isolated depending on their binding activities, not on the expression levels of the associated genes. This indeed constitutes a remarkable feature of this method since most other procedures require a certain abundance of mRNA, A schematic diagram for GBS cloning is shown in Figure 1. First, the

Figure 1. The strategy used for the screening of transcription factor (TF) binding fragments in genomic DNA, A schematic diagram of the selection cycle is shown. First, the genomic DNA fragments digested with restriction enzymes are bound with recombinant transcription factor, selected by nitrocellulose filter, and cloned into a plasmid vector. Next, from the library of cloned plasmids, the TF-binding fragments are again trapped by nitrocellulose filter, and concentrated by repeating the selection cycle.

166

7: Identification of target genes by GBS cloning genomic DNA fragments digested with restriction enzymes are mixed with recombinant transcription factor, selected by nitrocellulose filter, and cloned into a plasmid vector. Next, from the library of cloned plasmids, the ER-binding fragments are again mixed with the protein trapped by nitrocellulose filter and concentrated by repeating the selection cycle.

2.1 Preparation of a transcription factor protein Purified transcription factor protein expressed from a cloned gene is utilized for GBS cloning. If the DNA binding specificity of the transcription factor is in a domain, expression of the DNA binding domain alone may be sufficient. Expression of the protein in Escherichia coli is recommended when the DNA binding activities are confirmed. Epitope tags such as histidine are useful not only for the purification but also for the selection of the DNA-protein complex. Expression of proteins in baculovirus, yeast, or mammalian cells may be utilized when the protein expressed in E. coli has insufficient DNA binding activity. If the transcription factor needs other factors for DNA binding, proteins of such factors may also be prepared. An example for expression of the DNA binding domain (DBD) of oestrogen receptor (ER-DBD) is described below. The oestrogen receptor is a member of a superfamily of nuclear receptors for small hydrophobic ligands including the steroid hormones, thyroid hormone, and retinoic acids (1, 2). As a class, these receptors are transcription factors whose activities are regulated by ligand binding. Nuclear receptors bind to their responsive elements in the enhancer regions of target genes and regulate directly the transcription of these genes. Utilizing homology comparison among ERs of different species and transient co-transfection assays of the mutated ERs and reporter plasmids, several functional domains of the oestrogen receptor have been defined. These include a DNA binding domain, which is essential for it to bind to its specific responsive elements, a hormone binding domain, and transactivation domains (6). This domain model of the oestrogen receptor is applicable to other nuclear receptors. The structure and the position in the protein of the DNA binding domain and hormone binding domain are rather well conserved among the nuclear receptors (1, 2). Functional chimeric receptors have been constructed which show the steroid binding specificity of one receptor and the DNA binding specificity of the other receptor. These domain-swap experiments have shown that the DNA binding activity and specificity exist exclusively in this domain (1, 2). The DNA binding domain contains two 'zinc finger' motifs (finger 1 and finger 2) that are critical for the interaction of these proteins with their cognate sequences. Site-directed mutagenesis experiments have shown that relatively few amino acids in the domain can determine the specificity of its binding (7, 8). The P (proximal) box (5 amino acids) in finger 1 and the D (distal) box (5 amino acids) in finger 2 have been shown to be essential to the DNA binding specificity of these domains (8) (Figure 2). The DNA binding domains of the glucocorticoid (9), progesterone (10), and oestrogen (3, 11) 167

Satoshi Inoue et al.

Figure 2. DNA binding domain of the oestrogen receptor. Single-letter amino-acid codes are used. This domain contains two 'zinc finger' motifs (finger 1 and finger 2) and is fully conserved among rat, chicken, and human receptors. The residues thought to be involved in discriminating specific responsive elements (P box and D box) are shown in bold type. The corresponding residues of the P box and D box in the glucocorticoid and thyroid hormone receptors are shown below.

receptors expressed in E. coli are sufficient to recognize and bind specifically to their responsive elements. Thus, ER-DBD was expressed in E. coli producing inclusion bodies (3). The protein was purified as given in Protocol 1 and the DNA binding activity was confirmed by filter binding assay and footprinting (Figure 3). Protocol 1. Expression and purification of a transcription factor protein when inclusion bodies are produced Equipment and reagents • Sonicator for bacterial cell disruption • Transcription factor expression vector • Dialysis buffer 1: 4 M urea, 30 mM Tris-HCI pH 7.5, 30 mM NaCI, 0.05 mM ZnCI2, 1 mM DTT

Dialysis buffer 2: 30 mM Tris-HCI pH 7.5, 30 mM NaCI, 0.05 mM ZnCI2, 1 mM DTT . Lysis buffer: 30 mM Tris-HCI pH 7.5, 30 mM NaCI, 0.05 mM 2nCI2

Method 1. Transform E. coli JM109 cells with the transcription factor expression vector.

168

7: Identification of target genes by GBS cloning 2. Grow cells to OD (wavelength) = 600 nm in 200 ml of LB-ampicillin medium. 3. Induce the culture by 1 mM IPTG for 10 h. 4. Harvest the induced cells in a 50 ml tube, rinse with lysis buffer, suspend in 20 ml of the same buffer, and disrupt by sonication for 5 min at approximately 75% of maximum setting in a sonicator (time and setting may vary depending on what sonicator one uses). 5. Centrifuge the sonicated lysate at 10000 g for 30 min, suspend the precipitate in 20 ml of 1 M sucrose with gentle stirring, and centrifuge at 10000 g for 30 min. 6. Suspend the precipitate in 40 ml of 2% Triton X-100, 10 mM EDTA for 12 h and centrifuge again at 10000 g for 30 min. 7. Dissolve the precipitate in 5 ml of 8 M urea and dialyse against 2.5 litres of dialysis buffer 1 for 12 h, then against 2.5 litres of dialysis buffer 2 for 12 h. 8. Store at -80°C

2.2 Source of DNA High molecular weight DNA is required as the source DNA for GBS cloning. The efficiency of obtaining binding sites depends on genome size. If an animal with relatively small genome size is used as the source of genomic DNA, one can isolate specific binding sites at a higher frequency. These include Drosophila, C. elegans, Arabidopsis thaliana, and some fish such as Fugu (puffer fish). For example, we calculate the estimated number of binding sites in a whole genome, when a cloned transcription factor recognizes a specific 10 bp sequence, as follows. Assuming a random appearance of the 10 bp sequence, there may be one perfect binding sequence in about every 1000 kb (410) bp. Thus the human genome (3 X 109 bp) could potentially have a few thousand binding sequences. The use of CpG islands as source DNA for isolation of the target elements has a significant advantage over the use of whole genomic fragments, because CpG islands are enriched in active genes (12). The human CpG island library (12) contains 45 000 copies of the gene and the average insert size is 1 kb. In the above case, the CpG island library would contain only about 50 binding sequences, but a cloned transcription factor recognizes a 10 bp sequence. Furthermore, most CpG islands exist as single copies in the genome. Thus, the isolated CpG islands could be used efficiently as probes in Northern and Southern blots for detecting their associated genes. Of course, it has to be noted that CpG islands include a maximum of 60% of human genes, so that target elements not associated with CpG islands can not be isolated by this strategy. 169

Satoshi Inoue et al.

Figure 3. The purity and binding specificity of the ER-DBD protein. (a) SDS-polyacrylamide gel electrophoresis of the purified ER-DBD protein. The ER-DBD protein is in lane 1 and molecular weight marker proteins (Bio-Rad) in lane M. The arrow shows the ER-DBD protein and the numbers refer to the molecular masses of reference proteins (kDal). (b) Filter binding assay with the ER-DBD protein. The viteltogenin ERE (vitERE) sequence was synthesized and cloned into pUC18 to construct a pUC/vitERE plasmid. The pUC/vitERE was digested with Hpall and EcoRI and labelled with [a32P]dCTP and Klenow fragment. An aliquot of the labelled fragments was loaded in lane M (5000 c.p.m.). The remainder [100000 c.p.m.) was incubated without the ER-DBD protein (lane 1) or with 0.2 pmol (lane 2) or 5 pmol (lane 3) of the ER-DBD protein. The fragments trapped by the nitrocellulose filter were analysed on a 5% polyacrylamide gel. The arrow shows the fragment containing vitERE. (c) DNAase I footprinting with the ER-DBD protein. The single-end labelled DNA fragment containing vitERE was incubated without the protein (lane 1) or with 1 pmol (lane 2) or 2 pmol (lane 3) of the ER-DBD protein. After DNAase I digestion, the resulting fragments were analysed on a 5% denaturing polyacrylamide gel. The sequence ladder A (> C) (lane M) and the sequence protected by the ER-DBD protein are shown.

Some reports have utilized genomic DNA in native chromatin as a source for isolating DNA binding sites of transcription factors using immunoprecipitation methods (13-15).

2.3 Selection procedure Several methods can be used to concentrate the DNA-protein complex. Filter binding is a simple method used to trap the complex on a nitrocellulose filter. 170

7: Identification of target genes by GBS cloning The gel mobility shift assay can be used to isolate the complex from the shifted fraction. If antibodies to the transcription factor are prepared, they can detect the complex by immunoprecipitation. Expression of a tagged, cloned transcription factor protein may be useful, then the DNA-protein complexes can be detected and concentrated by immunoprecipitation using antibodies to the tag or by utilizing the binding properties of the tag. For amplification of selected DNA, amplification by plasmid DNA is recommended. Amplification by polymerase chain reaction (PCR) can also be used. A whole genome PCR method has been used for the screening of Xenopus transcription factor TFIIIA binding sites (16). This method is also applicable in combination with the filter binding method (Figure 4). However, only short DNA fragments of less than a few hundred base pairs may be obtained by this method, with the shorter fragments being selected more efficiently. In contrast, when amplification of plasmid DNA is utilized, relatively long fragments of up to 2-3 kb may be obtained. These sizes of fragments are suitable for sequencing, walking in the genome, analysing for enhancer activity, or Northern blotting.

Figure 4. Whole genome PCR, First, the genomic DNA fragments are digested to completion with Sau3AI. The DNA is then ligated to 'catch linkers' consisting of phosphorylated catch A oligomer (5'-GATCGTCGACGAATTCGACGAT-3 r ) and phosphorylated catch B oligomer (5'-CAGCTGCTTAAGCTGCTA-3r). The DNA fragments bound to the transcription factor (TF) protein were selected by the filter binding procedure then amplified by PCR. This cycle can be repeated as often as is necessary.

171

Satoshi Inoue et al.

2.4 GBS cloning procedure For example, the screening of the ER-binding sites in human genomic DNA was performed in the following manner. The fragments bound to the ERDBD protein were selected by passing them through a nitrocellulose filter. The selection process shown in Figure 1 was repeated six times and six independent clones were isolated. These plasmids contained inserts ranging from 200 bp to 2 kb and were named E0-E5. Sequence analysis has shown that all of these fragments contain perfect palindromic ERE sequences, termed here ERE0-ERE5, respectively (Figure 5).

Figure 5. The ER binding fragments and their ERE sequences. The six ER binding fragments (E0-E5) were sequenced. The position of the perfect palindromic ERE is shown by the box. The palindromic sequences (ERE0-ERE5) are shown with underlines and compared with the vitellogenin ERE sequence (vitERE).

Protocol 2. Isolation of transcription factor binding sites in genomic DNA Equipment and reagents • Vacuum apparatus for filter binding assay . • Nitrocellulose filter (Schleicher and Schuell, BA85, 0.45 um, 25 mm) • * Genomic DNA . Binding buffer: 40 mM Tris-HCI pH 7.5, 100• • mM KCI, 1 mM EDTA, 1 mM DTT, 10% glycerol, 0.1 mg/ml BSA

Washing buffer: 40 mM Tris-HCI pH 7.5, 1 mM EDTA Elution buffer: 20 mM Tris-HCI pH 7.5, 1 mM EDTA, 20 mM NaCI 0.1% SDS Transcription factor protein High efficiency, competent E. coli cells

Method 1. Digest genomic DNA (10 ug) with restriction enzymes to obtain genomic fragments at a size of 0.5-3 kb, for example.

172

7: Identification of target genes by GBS cloning 2. Incubate the genomic DNA fragments with the transcription factor protein (10 pmol) for 30 min on ice in a volume of 400 ul of binding buffer. 3. Pass this solution slowly through a pre-soaked nitrocellulose filter under vacuum. 4. Wash the filter with 500 ul of washing buffer five times. 5. Elute the trapped DNA from the filter in a volume of 400 ul of elution buffer. 6. Ligate the trapped DNA (10 ng) with a plasmid vector (e.g. pUC18). 7. Transform the plasmid in high efficiency, competent E. coli cells (e.g. JM109; Stratagene). 8. Culture the cells on an LB-ampicillin plate for 24 h and prepare the plasmid DNA by alkaline treatment followed by centrifugation in a caesium chloride-ethidium bromide gradient. 9. Incubate the plasmid DNA (10 ug) with the transcription factor protein (10 pmol) as described in step 2 and perform the selection procedure as described in steps 3-5, 7-8. 10. Repeat selection cycle (step 9) several times.

2.5 GBS cloning from CpG islands CpG islands are short stretches of DNA containing a high density of nonmethylated CpG dinucleotides and distributed throughout the genome as 45 000 short regions of about 1 kb (17, 18). About 60% of human genes are found to be associated with CpG islands, including most of the housekeeping genes and 40% of the tissue-specific genes. These regions often include the promoter region and one or more exons of associated genes (19-21). Human CpG islands selected by the methyl-CpG binding domain (MBD) of MeCP2, which binds DNA methylated at CpG (22), were inserted into a plasmid vector for the construction of a human CpG island library (12). We can utilize the human CpG island library as a source of genomic DNA for GBS cloning. It has been shown that binding fragments isolated from CpG islands contain exons of associated genes at high frequency (5). Protocol 3. Isolation of transcription factor binding sites in genomic DNA from a CpG island library Equipment and regents • Human CpG island library (12) (kindly provided by the HGMP Resource Centre, Clinical Research Centre, Watford Road, Harrow HA1 3UJ, UK)

• Vacuum apparatus for filter binding assay • Nitrocellulose filter (Schleicher and Schuell, BA85, 0.45 um, 25 mm) « Transcription factor protein

173

Satoshi Inoue et al. Protocol 3. Continued • Washing buffer: 40 mM Tris-HCI pH 7.5, 1 mM EDTA . Elution buffer: 20 mM Tris-HCI pH 7.5, 1 mM EDTA, 20 mM NaCI, 0.1% SDS

• Binding buffer: 40 mM Tris-HCI pH 7.5, 100 mM KCI, 1 mM EDTA, 1 mM DTT, 10% glycerol, 0.1 mg/ml BSA

Method 1. Incubate the plasmid DNA (10 ug) prepared from the CpG island library with the transcription factor protein (10 pmol) for 30 min on ice in 400 ul of binding buffer. 2. Pass this solution slowly through a pre-soaked nitrocellulose filter under vacuum. 3. Wash the filter with 500 ul of washing buffer five times. 4. Elute the trapped DNA from the filter in 400 ul of elution buffer. 5. Transform the plasmid in high efficiency, competent E. coli cells (e.g. JM109; Stratagene). 6. Culture the cells on an LB-ampicillin plate for 24 h and prepare the plasmid DNA by alkaline treatment followed by centrifugation in a caesium chloride-ethidium bromide gradient. 7. Repeat the selection cycle as described in steps 1-6, using the plasmid DNA (10 ug) obtained.

Shago and Giguere (23) described the use of CpG-island-enriched DNA for the isolation of target genes. They enriched the CpG island fraction using restriction enzymes and selected binding sites by utilizing a combination of PCR amplification and gel mobility shift assay. Although their procedure is different from that given in Protocol 3, they showed the importance of using CpG islands for isolating the target elements of transcription factors.

3. Molecular cloning of target genes using the DNA fragments obtained by GBS cloning 3.1 Zoo blotting If a transcription factor binding site is functional in vivo and associated with a target gene, the region near the binding site, especially the coding region of the target gene, should be conserved among different species. Thus one can select the fragment for which the sequence is conserved among different species by zoo blot analysis. If the sequence is conserved, the region is more likely to be functional in vivo. In addition, by zoo blotting one can determine whether the fragments are suitable for further screening or not. When the DNA fragment contains repetitive sequences, the signal may be detected as a smear. Even in this case, it may be possible to find specific signals in another species that does not have the repetitive sequences. 174

7: Identification of target genes by GBS cloning Protocol 4. Zoo blotting Reagents • •

Genomic DNA from several species Hybridization buffer 750 mM NaCI, 20 mM Tris (pH 8.0), 2.5 mM EDTA, 1 x Denhardt solution, 1% SDS, 50 ug/ml salmon sperm DNA 1 x SCC (saline-sodium citrate)

150 mM NaCI, 15 mM sodium citrate 1 x Denhardt solution 0.2% BSA fraction V (Sigma), 0.2% polyvinylpyrrolidone, 0.2% Ficoll 400 (Pharmacia)

Method 1. Prepare high molecular weight DNA from different species such as human, mouse, Xenopus, Drosophila, and fish. 2. Digest DNA (10 ug) with one of the six-base recognizing restriction enzymes; e.g. BamHI or EcoRI. Resolve digested DNA on 0.7% agarose gels. 3. Transfer DNA to a nitrocellulose filter. 4. Prehybridize the filter in hybridization buffer for 2 h at 65°C. 5. Hybridize the filter to the probe of the genomic DNA fragment in hybridization buffer for 18 h at 65°C or lower. 6. Wash the filter twice in 2 x SSC for 15 min at room temperature. 7. Expose to X-ray film for one day and develop film. If non-specific signals are too strong, wash the filter at a higher temperature and repeat step 7.

3.2 Northern blot analysis When transcription factor binding DNA fragments are obtained, target genes may be identified in the adjacent region. To find the target genes associated with the genomic DNA fragments containing binding sites for the transcription factor, Northern blotting is useful. Organs, tissues, and cells in which the particular transcription factor is expressed should be used as the sources of mRNA. Genomic DNA fragments are used as the probes hybridizing to each mRNA on Northern analysis. If the signals are detected, the genomic DNA fragments may contain coding as well as non-coding regions of the target gene adjacent to the binding sites. Transcriptional enhancers often exist in introns and in 3'-trailer sequences. To avoid false positives due to cross-hybridization by repetitive sequences, zoo blotting is useful as mentioned above. When the DNA fragment is likely to be a part of a gene, so-called 'gene walking' may be carried out in either 5' or 3' direction.

3.3 Sequencing and DNA database searches On the other hand, the sequences of the DNA fragments should be determined. This is also required for the identification of the binding sequences in the genomic DNA fragment. The binding sites are often multiple. 175

Satoshi Inoue et al. The DNA fragments can contain CpG islands. Computer-assisted DNA analysis is applied to determine the CG contents. The DNA fragment may contain enhancer regions and exons of target genes. The homologous sequences can be looked for in DNA databases by BLAST and FASTA analysis using a Macintosh or PC computer connected to the Internet with Netscape (Netscape Communication Corporation) or Internet Explorer (Microsoft) software to access the web site http://www.ncbi.nlm.nih.gov/BLAST/.

3.4 Binding activity of the genomic fragment The DNA fragments should be reconfirmed for their binding to the transcription factor. Fragments can be end labelled with [y32P]ATP and the binding capacities analysed by filter binding assay, comparing with the negative control. Gel mobility shift assay and footprinting are also utilized. Protocol 5. Nitrocellulose filter binding assay Equipment and reagents • Vacuum apparatus for filter binding assay • « Nitrocellulose filter (Schleicher and Schuell, BA85, 0.45 um, 25 mm) • • Labelled DNA fragments 1 mM * Binding buffer: 40 mM Tris-HCI pH 7.5, • 100 mM KCI, 1 mM EDTA, 1 mM DTT, 10% . glycerol, 0.1 mg/ml BSA

Washing buffer: 40 mM Tris-HCI pH 7.5, 1mM EDTA Elution buffer: 20 mM Tris-HCI pH 7.5, EDTA, 20 mM NaCI, 0.1% SDS Transcription factor protein poly(dl-dC)

Method 1. Using [y32P]ATP, end label the DNA fragments, as well as positive and negative control DNA fragments with or without the binding site. 2. Incubate the labelled DNA fragments (10000-100000 c.p.m.) with the transcription factor protein (0-5 pmol) for 30 min on ice. Carry out the binding reaction in 200 ul of binding buffer in the presence of 1 ug of poly(dl-dC) as carrier DNA; add other competitor DNA when required. 3. Pass this solution slowly through a pre-soaked nitrocellulose filter under vacuum. 4. Wash the filter with 500 ul of washing buffer five times. 5. Elute the trapped DNA from the filter in a volume of 400 ul of elution buffer. 6. Resolve the eluted DNA on a 5% polyacrylamide gel and count the radioactivity by liquid scintillation counter to determine the recovery.

3.5 Enhancer activity of the genomic fragment To determine whether the genomic DNA fragments have any enhancer activity that is dependent on the transcription factor, the fragments should be inserted 176

7: Identification of target genes by GBS cloning into a reporter vector such as thymidine kinase (tk)-CAT or tk-luciferase having a Herpes simplex virus tk promoter. The reporter plasmids will be transfected with or without the transcription factor expression vector into mammalian cells. As controls, the reporter plasmids without insert and with the control enhancer should be assayed simultaneously. After transfection, the cells are cultured and harvested, then the CAT or luciferase activities assayed. Thus one can check the enhancer activity of the genomic DNA fragments.

3.6 Molecular cloning of target genes using the DNA fragment obtained by GBS cloning According to the methods mentioned above, candidates for the probes to screen the target genes are selected. In some cases, sequence analysis may be sufficient to identify the adjacent gene. Usually, one can screen cDNA libraries of some organs, tissues, and cells in which the transcripts are detected by the probe. When the cDNA associated with the binding sites is isolated, it should be further analysed to determine whether it is a target gene in vivo or not. With the aid of GBS cloning we isolated an oestrogen-responsive finger protein (EFP) (Figure 6A) as a candidate target gene of oestrogen receptor from a placenta cDNA library using an ER-binding fragment (shown as E0 in Figure 4) as a probe. We showed (i) upregulation of EFP mRNA after oestrogen treatment in cultured cells (4) and in animals (24) and (ii) colocalization of EFP mRNA with ER mRNA by in situ hybridization histochemistry (24). In the case of two other genes, COX7RP (Figure 6B) and EBAG9 (Figure 6C), isolated by GBS cloning with a CpG island library (5), we showed upregulation of both mRNAs by Northern blotting. The increase was inhibited by actinomycin D but not by cycloheximide. Their binding sites indeed bound to ER as shown by filter binding assay and gel mobility shift assay (5). They had significant oestrogen-dependent enhancer activities in CAT assay, when they were inserted into the 5'-upstream of the chicken B-globin promoter. So, these genes must also be among the important genes involved in the variety of oestrogen action. Actually, recent findings have suggested that EFP and EBAG9 are involved in the growth and development of uterus and/or uterine cancer (unpublished data).

4. Concluding remarks Here we have described a method designated 'GBS cloning' and some of the results obtained by this procedure. Several genes in addition to those described in this chapter have already been isolated and are under detailed study. The human genome project is supposed to finish the sequencing of 3 X 109 177

Satoshi Inoue et al.

Figure 6. Schematic diagram of isolated ER binding fragments and their associated cDNA. ER binding fragments and their associated cDNA are indicated. A: EFP (oestrogenresponsive finger protein); B: COX7RP (cytochrome oxidase c subunit Vll-related protein); C: EBAG9 (oestrogen receptor-binding fragment-associated gene). The open reading frame (ORF) of the cDNA is indicated by an open box and the region containing each ERE is indicated by a filled box.

nucleotides by AD 2003-5, which will reveal 50-100 X 103 genes. The elucidation of the functions of these genes, however, will not be easy if only the amino acid sequences of the proteins are known. A significant proportion of these proteins will be various transcription factors, since approximately 7% of the 6000 yeast Saccharomyces cerevisiae genes are likely to be transcription factors. Even if only 5% of the human genes represent transcription factors, scientists will be faced with 5000 genes, each of which may regulate at least several downstream genes that in turn express various phenotypes. The complexity of the networks formed by these genes would become enormous. Although a number of different approaches must be taken to resolve gene networks, the GBS cloning method presented here could play a crucial role as a fundamental technique along this line.

References 1. Evans, R. M. (1988). Science, 240, 889. 2. Green, S. and Chambon, P. (1988). Trends Genet., 4, 309. 3. Inoue, S., Kondo, S., Hashimoto, M., Kondo, T., and Muramatsu, M. (1991). Nucleic Acids Res., 19, 4091. 178

7: Identification of target genes by GBS cloning 4. Inoue, S., Orimo, A., Hosoi, T., Kondo, S., Toyoshima, H., Kondo, T., et al. (1993). Proc. Natl Acad. Sci. USA, 90, 11117. 5. Watanabe, T., Inoue, S., Hiroi, H., Orimo, A., Kawashima, H., and Muramatsu, M. (1998). Mol. Cell. Biol., 18, 442. 6. Kumar, V., Green, S., Stack, G., Berry, M., Jin, J.-R., and Chambon, P. (1987). Cell, 51, 941. 7. Mader, S., Kumar, V., Verneuil, H., and Chambon, P. (1989). Nature, 338, 271. 8. Umesono, K. and Evans, R. M. (1989). Cell, 57, 1139. 9. Freedman, L. P., Luisi, B. F., Korszun, Z. R., Basavappa, R., Sigler, P. B., and Yamamoto, K. R. (1988). Nature, 334, 543. 10. Eul, J., Meyer, M. E., Tora, L., Bocquel, M. T., Quirin-Stricker, C., Chambon, P., and Gronemeyer, H. (1989). EMBO J., 8, 83. 11. Schwabe, J. R., Neuhaus, D., and Rhodes, D. (1990). Nature, 348, 458. 12. Cross, S., Charlton, J., Nan, X., and Bird, A. (1994). Nature Genet., 6, 236. 13. Gould, A. P., Brookman, J. J., Strutt, D. L, and White, R. A. (1990). Nature, 348, 308. 14. Phelps, D. E. and Dressier, G. R. (1996). J. Biol. Chem., 271, 7978. 15. Tomotsune, D., Shoji, H., Wakamatsu, Y., Kondoh, H., and Takahashi, N. (1993). Nature, 365, 69. 16. Kinzler, K. W. and Vogelstein, B. (1989). Nucleic Acids Res., 17, 3645. 17. Bird, A., Taggart, P. M., Frommer, M., Miller, O. J., and Macleod, D. (1985). Cell 40, 91. 18. Cooper, D., Taggart, M., and Bird, A. (1983). Nucleic Acids Res., 11, 647. 19. Antequera, P. and Bird, A. (1993). Proc. Natl Acad. Sci. USA, 90, 11995. 20. Gardiner-Garden, M. and Frommer, M. (1987). J. Mol. Biol., 196, 261. 21. Larsen, F., Gundersen, G., Lopez, R., and Prydz, H. (1992). Genomics, 13, 1095. 22. Lewis, J. D., Meehan, R. R., Henzel, W. J., Maurer-Fogy, I., Jeppesen, P., Klein, F., and Bird, A. (1992). Cell, 69, 905. 23. Shago, M. and Giguere, V. (1996). Mol. Cell. Biol., 16, 4337. 24. Orimo, A., Inoue, S., Ikeda, K., Noji, S., and Muramatsu, M. (1995). /. Biol. Chem., 270, 24406.

179

This page intentionally left blank

8 Analysis of cloned factors ROGER WHITE and MALCOLM PARKER

1. Introduction The analysis of a cloned transcription factor and the characterization of its functional properties is carried out primarily by the identification and mapping of domains within the protein using deletion and point mutagenesis. This is accomplished by expression of either the factor itself, or altered forms of the protein, in vitro and in vivo in different cell types. The aim of this chapter is to describe some of the approaches and methods used in this analysis. Methods will also be described by which the cloned factor itself, or fragments of the factor, may be overexpressed with the aim of producing large amounts of the protein for biochemical and structural studies. Finally a number of assay systems will be described which may be used to measure the transcriptional activity of the cloned factor and its ability to interact with different components of the cell, in particular other proteins and DNA. The chapter will be illustrated by reference to studies using two different proteins; the mouse oestrogen receptor (MOR), a member of the nuclear receptor family of transcription factors (1), and a coactivator protein, the steroid receptor coactivator protein SRC1a (2), which interacts with nuclear receptors in vitro and in cells where it enhances receptor mediated transcription. The nuclear hormone receptors are multifunctional proteins which have the ability to respond to signals from the extracellular environment and to act as transcription factors by directly regulating gene expression (3). These proteins are characterized by a number of different functional properties including ligand binding, dimerization, nuclear localization, DNA binding, and transcriptional activation. In the identification of specific residues involved in individual functions it is important to ensure that mutational analysis does not disrupt the overall protein structure. The colocalization of ligand binding, dimerization, and transactivation functions in the hormone binding domain of the oestrogen receptor facilitates a detailed analysis of this protein by mutagenesis since changes may be introduced to alter only one function while the others are unaffected. The recent determination of models for the structure of the ligand binding domain by X-ray crystallography has emphasized the importance of the combination of mutagenesis and structural studies in developing a

Roger White and Malcolm Parker detailed understanding of the function of a cloned factor (4). In addition to the properties described above, the ligand binding domain of the receptor also forms a surface by which the receptor interacts with other proteins, including the coactivator SRC1a. These cloned factors therefore provide an ideal model to illustrate a number of different methods of analysis. A flow chart describing some of the techniques included in this chapter and the order in which they may be applied is shown in Figure 1.

1.1 Identification of conserved domains When the protein sequence of the factor has been determined from analysis of cDNA clones, the first step is to analyse this sequence for homology with other known factors. Sequence comparison with clones encoding the same factor from other species allows the identification of the most highly conserved regions. Sequence alignments with other proteins will also indicate the relative positions of conserved amino acids, motifs, and functional domains. The relative importance of these may then be considered when selecting targets, either particular regions or specific residues, for mutagenesis studies. Additional information may also be derived from the study of models of the structure of all or parts of the protein, as and when these are determined. Many sequence and structure analysis programs and databases are accessible on the Internet (5).

2. Mapping and analysis of domains by deletion and point mutagenesis The organization of functional domains within a transcription factor is determined by deletion mutagenesis. This involves either progressive deletion of sequences from the amino terminus or carboxyl terminus of the protein or the construction of a series of internal deletions in which small regions of the sequence are removed. Deletion analysis determines the positions of the boundaries of individual domains which then allows the analysis of a particular functional property to be investigated, either in isolation or as part of a chimeric protein. This section will describe methods used in the construction of a series of deletion mutants of a cloned transcription factor, in this case the coactivator protein SRC1a.

2.1 Preparation of deletion mutants using PCR The simplest and most reliable method of generating a series of deletion fragments of a cloned DNA is by using the polymerase chain reaction (PCR) and oligonucleotide primers. In using this approach, two points should be considered: (a) Primer design—where possible, the pair of primer sequences should be: (i) unique and anneal to only one site on the template DNA 182

8: Analysis of cloned factors

Figure 1. Methods and procedure for the analysis of a cloned factor.

(ii) approximately 50% GC and 50% AT (iii) lacking in direct or inverse repeat sequences (iv) similar in Tm to ensure both primers anneal at the same temperature (b) Availability of restriction sites. Suitable restriction sites must be available in the original sequence to reinsert the altered sequence generated by PCR, if this is required, and these sites may be designed into the sequence of the PCR primers. In general the size of the fragments produced in the PCR reaction should be as short as possible since all the newly synthesized sequence must be verified by DNA sequence analysis. 183

Roger White and Malcolm Parker A number of high-fidelity DNA polymerases with a proof-reading capability to limit the production of errors are currently available. The following protocol describes the use of PCR to generate a series of deletion fragments of SRC1a. Protocol 1.

PCR deletion analysis

Equipment and reagents • Oligonucleotides to amplify the region(s) required (10 uM) • GeneAmp thin-walled amplification tubes (Perkin-Elmer) . ELONGase enzyme mix (Gibco-BRL Life Technologies Ltd) • Supercoiled DNA template (0.1 ug/ml)

Buffer A: 300 mM Tris-SO4 pH 9.1, 90 mM (NH4)2SO4, 5 mM MgSO4 Buffer B: 300 mM Tris-SO4 pH 9.1, 90 mM (NH4)2SO4, 10 mM MgSO4 dNTP mix: dATP 10 mM, dCTP 10 mM, dGTP 10 mM, TTP 10 mM; all Amersham Life Science Ltd.

Method 1. Add 1 ul template DNA to 32 ul distilled water and heat at 94°C for 3 min to denature the DNA. 2. Add 1 ul dNTP mix, 2 ul of each primer and 5 ul each of buffer A and buffer Ba and 1 ul ELONGase enzyme mix to the denatured DNA. 3. Amplify by PCR by denaturing at 94°C for 30 sec, annealing at 5°C below the T m b f or 30 sec, and extending at 68°C for 60 seec per kb of DNA to be synthesized. 4. Load an aliquot of the sample on to an agarose gel to check the synthesis. 5. Add an equal volume of a 1:1 mix of phenoktrichloromethane to the remaining sample, vortex, and spin in a microcentrifuge for 1 min. Remove the aqueous phase and recover the DNA by ethanol precipitation. 6. Digest with the appropriate restriction enzymes to generate compatible ends for cloning the fragments into expression vectors into the full length cDNA clone. • Buffers A and B may be mixed in different ratios to optimize the magnesium concentration to improve synthesis of DNA. b The Tm is calculated from the nucleotide composition of the primers, allowing 4oC for every C or G and 2°C for every A or T. c Increasing the extension time can improve DNA synthesis in some cases.

The relative positions and sizes of a series of deletion fragments of SRC1 generated by PCR and designed to analyse and map some of the functional domains within this protein are shown in Figure 2. 2.1.1 Selection of methods to construct deletion mutations The advantages of PCR compared to other methods of creating deletion fragments, such as using specific restriction enzyme sites or exonucleases, are that 184

8: Analysis of cloned factors (a) precise deletions may be constructed with the boundaries defined by the user (b) the 5'-primers may be designed to contain a Kozak consensus sequence (6) to optimize the initiation of translation (c) appropriate restriction sites may be inserted into the primer sequence to simplify the cloning of the fragments into a number of different vectors The fragments obtained may be inserted into vectors for in vitro expression and translation, expression in bacteria or yeast, and expression in insect cells or mammalian cells. Vectors can be chosen in which the fragments are expressed as fusions with other proteins or short peptide sequences to simplify purification and detection. These sequences are termed 'tags' and include glutathione 5-transferase (GST), homopolymers of the amino acid histidine (His tags), and epitope tags such as the myc peptide or the FLAG peptide. Many vector systems and specific antibodies for purification are available from commercial suppliers. In those cases where short peptide tags are required, the DNA sequence encoding the tag may even be incorporated into the PCR primer sequences so that every fragment generated may be analysed for its expression using a single detection system. A number of different types of expression system are described in Section 3.

2.2 Point mutagenesis PCR also provides a relatively rapid and convenient method for introducing SRC1a

Figure 2. Deletion analysis of SRC1a. The SRC1a protein is 1441 amino acids in length. A to L are the positions of PCR primers designed to create a series of non-overlapping cDNA clones. Each primer is designed to contain unique restriction enzyme sites to facilitate cloning into expression vectors. The relative positions and the amino acid numbers of each fragment are shown.

185

Roger White and Malcolm Parker

Figure 3. Construction of point mutations using PCR. P1 and P4 are primers homologous to the cloned sequence and contain unique restriction enzyme sites RE 1 and RE 2. P2 and P3 are mutagenic primers with the position of the altered sequence shown in black. In this region P2 and P3 are complementary in sequence allowing these sequences to anneal and produce the full length mutated fragment in step D.

186

5: Analysis of cloned factors specific point mutations into a cloned DNA sequence. Similar principles may be used to create mutations which (a) produce novel restriction enzyme sites without altering the coding sequence (b) alter the coding sequence of a cDNA clone to change one or more amino acids (c) generate specific internal deletions The basic method and relative positions of primers are shown in Figure 3. PCR reactions are carried out as described in Protocol 1. 2.2.1 Oligonucleotide cassettes If it is necessary to create a large number of different mutations in a short region of a sequence of 20-80 nucleotides, the simplest method is to use Oligonucleotide cassettes which may be cloned directly into unique restriction enzyme sites flanking the region. Suitable sites may be created using the strategy described in Figure 3 if these are unavailable in the original DNA sequence. Different point mutations may also be introduced at the position of a single amino acid in a cDNA clone by using degenerate oligonucleotides. A detailed description of these mutagenesis techniques may be found in Gene probes 2 (7).

3. Expression systems 3.1 The use of in vitro translation systems and overexpression systems to analyse transcription factor function 3.1.1 Synthesis of proteins using coupled in vitro transcription and translation A number of the functional properties of a cloned transcription factor may be analysed using in vitro synthesized protein. In the case of the mouse oestrogen receptor, a coupled transcription and translation system has been used to synthesize small amounts of protein for investigating DNA binding, ligand binding, and determining the ability of the protein to interact with other cloned factors in vitro. In vitro translation systems are also particularly suitable for screening a large number of different proteins generated in the mutational analysis of a transcription factor. A number of plasmid vectors are available for the synthesis of RNA for the cloned factor using in vitro transcription. The vectors consist of: (a) a cloning site or multiple cloning sites for the insertion of the cloned cDNA (b) promoter element(s) for RNA polymerase(s) 187

Roger White and Malcolm Parker (c) an origin of replication and an antibiotic resistance gene for the growth and maintenance of the vector in bacterial cells RNA polymerases which have been purified and are commercially available for in vitro transcription systems include RNA polymerase SP6, RNA polymerase T3, and RNA polymerase T7. A suitable vector will therefore contain a promoter region for one of these enzymes flanking a multiple cloning site into which the coding region of the cDNA is inserted. The sequence between the start site of transcription in the promoter and the beginning of the coding region should not contain other translation initiation sites or open reading frames which may interfere with subsequent translation of the RNA. It is therefore advisable that the distance between the start site and the first codon is small, approximately 20-50 nucleotides. The efficiency of translation is also improved by inclusion of a Kozak consensus sequence. In addition a termination codon, which is in frame with the coding sequence of the protein, must be present at the 3' end of the sequence to prevent translation of sequences transcribed from the plasmid DNA. Protocol 2. Coupled in vitro transcription and translation Equipment and reagents • TNT® coupled transcription/translation system (Promega) • RNasin ribonuclease inhibitor (Promega)

. L-[35S]methionine at TNT® 37 Bq/mmol (Amersham) • 1 mM methionine

Method 1. Thaw the reticulocyte lysate and keep on ice with the other reaction components. 2. Pipette into a microcentrifuge tube 25 ul of the lysate, 2 ul TNT® reaction buffer, 1 ul TNT® RNA polymerase (SP6, T7, or T3 depending on the promoter of the vector), 1 ul amino acid mixture minus methionine, 1 ul RNasin ribonuclease inhibitor, 1 ug DNA template, and nuclease-free water to a final volume of 47 ul. 3. Transfer 23 ul to a clean tube and add 2 ul L-[35S]methionine. Add 1 ul unlabelled methionine to the remaining mixture. Incubate both samples at 30°C for 60 min. 4. Add glycerol to a final concentration of 15% and store the translated protein at -70°C.

Translated proteins labelled with [35S]methionine are used to assess the size, purity, and yield of the protein by SDS-polyacrylamide gel electrophoresis. Radiolabelled proteins may also be used in protein-protein interaction assays as are described later (Section 6). Unlabelled protein can be used for 188

8: Analysis of cloned factors DNA binding assays (Section 5) and, in the case of the oestrogen receptor, for hormone binding studies. Although suitable for the rapid production and analysis of a large number of different samples, in vitro transcription and translation systems have a number of disadvantages. In particular the quantity of protein produced is small and a large proportion of the synthesized protein may be incorrectly folded. Estimates indicate that less than 5% of the receptor protein produced in a reticulocyte lysate is folded correctly. The protein may also lack normal post-translational modifications such as phosphorylation and glycosylation. The suitability of coupled transcription and translation systems for general analysis of transcription factor function must therefore be determined empirically.

3.2 Expression in bacteria Expression in bacteria provides a rapid and inexpensive way of isolating large amounts of protein. However, some problems may arise which result in the failure to express and isolate particular polypeptides. This may be a consequence of one or more of the following: (a) the size of the protein (b) the stability of the protein (c) the solubility of the protein Large proteins in particular may be difficult to express, subject to proteolysis and cleavage by bacterial proteases, and form insoluble complexes in the bacterial cells, making purification of the protein difficult. These problems may be overcome to some extent by using bacterial strains as host cells which are deficient in certain proteases. In general, therefore, bacterial expression systems are ideal for overexpressing and analysing small proteins or fragments of large proteins. It must be remembered, however, that bacterially expressed proteins will lack the modifications which may occur in mammalian cells such as phosphorylation and glycosylation. A variety of different vector systems are currently available to express polypeptides as fusion proteins with different types of tag sequences; described in Section 2.1.1. These vectors contain: (a) an inducible promoter for high levels of expression (b) a coding sequence for specific tag(s) at the 5' and/or 3' end (c) a polylinker sequence to insert the cloned fragment in frame with the sequence of the tag(s) (d) protease cleavage sites to remove the tag sequences if required An example of one type of expression vector, pGEX-2TK (Pharmacia) containing a cDNA insert encoding the MOR ligand binding domain (LBD) is shown in Figure 4. The following protocol has been used to express and purify fusion proteins of GST and the LBD of the oestrogen receptor and fusion 189

Roger White and Malcolm Parker

Figure 4. pGEX2TK bacterial expression vector containing the MOR ligand binding domain (LBD). Ptac is the inducible promoter. The internal kinase site allows the fusion protein to be labelled in vitro with 32P to generate probes for far-Western blotting. The fusion protein may be cleaved with thrombin to purify the LBD from GST if required.

proteins between GST and the fragments of SRC1a (described earlier in Section 2.1 and Protocol 1). Protocol 3.

Expression of GST-fusion proteins.

Equipment and reagents • Isopropyl-b-D-thiogalactoside (IPTG) (1 M) • . NETN: 20 mM Tris-HCI pH 8.0, 100 mM NaCI, 1 mM EDTA, 0.5% NP40, 1 mM DTT • L-broth (LB) medium

Protease inhibitors: bacitracin 100 mg/ml, PMSF 10 mg/ml in iso-propanol, leupeptin 2.5 mg/ml, aprotinin 1 mg/ml in 10 mM Hepes pH 8.0 (Sigma)

Method 1. Grow a 5 ml culture of bacteria (transformed with the pGEX recombinant plasmid to express the fusion protein) in LB medium containing ampicillin at 50 ug/ml by shaking at 37°C for 12-15 h. 2. Subculture the 5 ml culture into 45 ml of pre-warmed LB containing ampicillin at 50 ug/ml and grow at 37°C with shaking for 60 min. 3. Add 25 ul IPTG to a final concentration of 0.5 mM and continue shaking at 37°C for 3 h.

190

8: Analysis of cloned factors 4. Spin at 4000 g at 4°C in a swing-out rotor to pellet the bacteria. 5. Pour off and aspirate the medium. The pellet may be stored at -70°C at this stage. 6. Resuspend the pellet in 5 ml NETN containing protease inhibitors using 5 ul 100 mg/ml bacitracin, 4 ul 10 mg/ml PMSF, 2 (ul 2.5 mg/ml leupeptin, and 2 ul 1 mg/ml aprotinin per 1 ml of buffer. Sonicate on ice for two 10 sec bursts at power/amplitude 22. 7. Spin at 10000 gat 4°C for 5 min. 8. Remove the supernatant from the pelleted debris to a clean tube and store on ice. If long-term storage is desirable, add glycerol to a final concentration of 10% and store the lysate at -70°C.

Protocol 4. Purification of GST fusion proteins Equipment and reagents • NETN (Protocol 3) with protease inhibitors: 5 ul of 100 mg/ml bacitracin,4 ul of 10 mg/ml PMSF, 2 ul of 2.5 mg/ml leupeptin, and 2 ul of 1 mg/ml aprotinin per 1 ml of buffer (Sigma) . NETN milk: NETN with protease inhibitors containing 0.5% milk powder

• Glutathione-Sepharose 4B beads (Pharmacia) mixed gently to resuspend immediately prior to use • Glutathione elution buffer: 20 mM glutathione, 100 mM Tris-HCI pH 8.0, 120 mM NaCI » Bacterial lysate (Protocol 3, step 9)

Method 1. Add 100 ul NETN milk to 100 ul glutathione-Sepharose beads and mix gently. 2. Spin down the beads for 5 sec at 10000 x g in a microcentrifuge, aspirate the supernatant, and add 125 ul NETN milk solution. 3. Repeat step 2 three times. 4. Add 125 ul NETN milk solution to the beads and add this to 5 ml of the bacterial lysate from step 9 of Protocol 3. Mix gently on a rocker or rotating platform at 4°C for 60 min. 5. Spin down the beads with the bound protein at 1200 g for 10 min at 4°C, aspirate the supernatant and add 5 ml of NETN with protease inhibitors. 6. Repeat step 5 three times. The protein remaining on the beads may now be eluted or the protein-bead complex may be used directly in a GST pull-down assay (Protocol 11). 7. Elute the protein by adding 1 ml of the glutathione solution to the beads and rocking gently for 60 min at 4°C. 8. Spin at 10000 x g in a microcentrifuge to pellet the beads and transfer the supernatant containing the purified protein to a clean tube.

191

Roger White and Malcolm Parker

3.3 Expression in mammalian cells 3.3.1 Cell lines There are two main criteria for selection of cell lines in which to study transcription factor activity: (a) The cell line should either lack or have very low endogenous levels of the factor. (b) The cells should transfect easily and efficiently. A number of different cell lines have been used to study the mouse oestrogen receptor including L-cells, NIH-3T3 cells, COS-1, and HeLa cells. All of these cell lines lack endogenous receptor and suitable methods for the introduction of DNA by transfection have been established. These will be described later in Section 4.3. COS-1 may be used for overexpression of protein and this is described in detail in Section 3.4 and Protocol 5. 3.3.2 Organization of expression vectors The most important feature of an expression vector is a strong promoter which is active in a wide variety of eukaryotic cells. The vector should also contain 5'-untranslated leader, splicing, and termination sequences to generate stable and functional mRNA as well as a polylinker sequence to allow insertion of the cDNA clone of the transcription factor and an origin of replication and antibiotic resistance gene for growth and amplification in bacterial cells. A number of different expression vectors have been developed and are suitable for the expression of transcription factors with the majority of vectors containing promoters and enhancers of viral origin (8). Promoter elements contain DNA sequences which determine both the site and the rate at which transcription is initiated. Enhancer elements consist of binding sites for proteins involved in the stimulation of transcription. Many viral promoters and enhancers have a wide host range and therefore stimulate transcription efficiently in a variety of cell types. These include the SV40 early promoter and the human cytomegalovirus (CMV) major immediate-early promoter. Other enhancers may be cell type specific or, as in the case of the mouse mammary tumour virus long terminal repeat (MMTV LTR), stimulate transcription efficiently only in the presence of an inducer. This type of expression system may be advantageous if expression of the protein is toxic to the transfected cells. Two examples of mammalian expression vectors are pMT2 (8) and pSG5 (9). pSG5, shown in Figure 5, has the additional advantage that it also contains a T7 promoter and may therefore also be used for in vitro translation as described in Protocol 2. Neither of these plasmid vectors, however, has an in-frame tag sequence for detection and/or purification of the translated protein. 192

8: Analysis of cloned factors

Figure 5. pSG5 mammalian and in vitro expression vector. The vector contains the SV40 early promoter and a b-globin intron, the presence of which stabilizes the mRNA transcripts. The T7 promoter which may be used in coupled in vitro transcription and translation (Protocol 2) is also shown together with the position of cloned inserts.

3.4 Overexpression in mammalian cells Cloned factors may be overexpressed to high levels in mammalian cells using COS-1 cells. Recombinant vectors containing an SV40 early promoter and origin of replication are amplified in COS-1 cells, resulting in high levels of expression of mRNA and protein (10). This method is simple to perform and is suitable for the analysis of a large number of mutant forms of a protein; however, it is not suitable for sustained high level production of the protein since replication of the vector DNA makes the cells unviable. If it is necessary to express the cloned factor in specialized cell types then a different vector system must be used. Other eukaryotic viral expression systems to consider are those based on vaccinia, bovine papillomavirus (BPV), and retroviruses. The following protocol describes a method for the high efficiency transfection of COS-1 cells by electroporation. This technique may be applied to many different types of cells but the precise voltage and capacitance settings used must be determined empirically. 193

Roger White and Malcolm Parker Protocol 5.

Electroporation of COS-1 cells

Equipment and reagents • Gene Pulsar electroporation cuvettes (Bio-Rad)

Method 1. Harvest cells at 70% confluence by scraping and resuspend in phosphate-buffered saline (PBS) at 8 x 106 cells/ml. 2. Pipette 0.8 ml of the cell suspension containing 6.4 x 106 cells into a Gene Pulsar cuvette. Add 20 ul of a 1 mg/ml solution of DNA (20 ug) and place on ice for 5 min. 3. Electroporate at 450 V and 250 uF using a Bio-Rad Gene Pulsar apparatus or equivalent. 4. Place the cuvette on ice for 5 min. 5. Plate the cells from the cuvette in 4 x 9cm dishes and leave at 37°C for 48-60 h until ready for harvesting.a * Preparation of cell lysates is described in Protocol 6.

3.5 Overexpression in yeast and insect cells Alternative sources of protein for a cloned factor include overexpression of the cDNA in either yeast or insect cells. Yeast systems may be used to give inducible expression which may be advantageous if the protein in question is toxic. Expression in insect cells with baculovirus vectors is used to produce large quantities of protein suitable for purification and structural analyses, and for assessing protein-DNA or protein-protein interactions in vitro. Detailed methods for the maintenance of insect cells and the manipulation of baculovirus vectors have been described (11,12) and vector systems which allow high level expression of tagged proteins have been developed and are commercially available. In studying the functional properties of a transcription factor, one major advantage of synthesizing large amounts of material is that sufficient quantities are made available to add to in vitro transcription reactions allowing a direct analysis of transcriptional activation (see Chapter 9). The large amounts of protein produced may also be used for biochemical studies and for structural studies using biophysical techniques. Very high levels of expression, however, may also result in the protein becoming insoluble, difficult to purify, and possibly unsuitable for further analysis. A method for preparing a whole cell extract from either mammalian or insect cells is described in Protocol 6. 194

8: Analysis of cloned factors Protocol 6.

Preparation of a whole cell extract

Equipment and reagents 1 x High-salt buffer: 0.4 M KCI, 20 mM Hepes pH 7.4, 1 mM DTT, 20% glycerol, 0.5 mg/ml bacitracin, 40 fig/ml PMSF, 5 ug/ml pepstatin, 5 ug/ml leupeptin (Sigma)

Method 1. Remove the medium from the cell monolayer and wash the cells three times with ice-cold PBS. Scrape the cells from the dish and transfer the cell suspension in PBS into a sterile plastic universal. Recover the cells by centrifugation at 2000 x g for 5 min. Cell pellets may be stored at -70°C.a 2. Resuspend the cell pellet in 10 volumes of 1 x high-salt buffer. 3. Pass the lysed cell suspension through a 25 gauge needle five times. 4. Centrifuge at 50000 x g for 15 min to pellet insoluble debris. 5. Transfer the supernatant to a clean tube, determine the protein concentration, and store in aliquots at -70C. a Infected Sf9 cells are loosely attached to culture flasks so the cells can be harvested by shaking the flask then recovering the cells from the medium by centrifugation at 2000 g for 10 min. Wash the cell pellet three times with ice-cold PBS. The pellet may be stored at -70°C until required.

3.6 Comparison of expression systems In summary, no single method of expression is ideally suitable for the detailed analysis of a cloned factor. In vitro transcription and translation systems allow the rapid analysis of either a large number of proteins or an extensive series of mutant forms of a single protein. In vitro synthesized protein, however, is produced in small amounts and may lack normal processing and modification. Expression in bacteria may be used to synthesize large quantities of protein which may then be analysed in functional assays or used for the determination of protein structure. The protein produced, however, may lack authentic processing and other post-translational modifications such as glycosylation and phosphorylation. This problem may be overcome by expression in eukaryotic cells. The baculovirus overexpression system allows the production of large amounts of material but, compared to other expression systems, is not suitable for the analysis of a large number of different proteins. Finally, expression in mammalian cells using vectors based on viral systems is most likely to produce correctly processed or modified protein, although high levels of expression may be restricted to specialized cell types or require the isolation of stable cell lines. It is clear therefore that a detailed analysis of a cloned 195

Roger White and Malcolm Parker factor requires a number of different expression systems and vectors to be used.

4. Analysis of the properties of cloned factors This section will describe some of the assay systems used to analyse the functional properties of a cloned transcription factor and in particular will concentrate on methods for assessing transcriptional activation and the interactions between the factor and other proteins and DNA.

4.1 Transient transfection in mammalian cells A detailed functional analysis of a proposed transcription factor requires confirmation of both the identity of the factor and its ability to regulate gene expression. Identification may be accomplished directly by using a cotransfection assay, where the factor together with an appropriate target gene are introduced into heterologous cells in which the factor is normally absent or expressed at very low levels. This type of analysis depends on the specific DNA sequence of the binding site of the transcription factor being known. The methods described in the following section have been used in the analysis of the transcriptional activity of the MOR following transient transfection into a number of receptor-negative cell lines. The major considerations to be taken into account in the design of this type of procedure are: (a) (b) (c) (d)

the selection of a suitable cell line the choice of expression vector the type of target or reporter gene the method of transfection

The selection of cell lines and expression vectors has been described in Sections 3.3.1 and 3.3.2. The following section describes different types of reporter genes, methods of transfection, and reporter gene assays designed to study the activity of a cloned transcription factor.

4.2 Reporter genes and control plasmids A reporter contains three major regions: a binding site or sites forming a response element for the transcription factor being analysed, a basal promoter to provide binding sites for general transcription factors and to determine correct initiation of transcription, and a suitable marker gene for which a simple assay system is available. A number of different reporter genes have been used in co-transfection assays to study transcription factor activity. These include chloramphenicol acetyl transferase (CAT) (13), the lac-Z gene encoding p-galactosidase (14), and luciferase (15). Transcriptional activation by the MOR has been analysed using a reporter 196

8: Analysis of cloned factors consisting of a single oestrogen response element (ERE) contained within a 32 bp oligonucleotide isolated from the vitellogenin A2 gene from Xenopus laevls. The ERE has been characterized in detail and the binding site consists of the 13 nucleotide sequence GGTCANNNTGACC which forms a palindromic response element (16). In the reporter plasmid pEREBLCAT the 32 bp oligonucleotide has been synthesized and cloned as an Xbal fragment into the polylinker of the vector pBLCAT2 (17). This vector contains the basal promoter of the herpes simplex virus (HSV) thymidine kinase (tk) gene from -105 to +51 linked to the coding region of the bacterial chloramphenicol acetyl transferase (CAT) gene and is therefore suitable for use with any response element. Chimeric CAT fusion genes are particularly useful because there is no endogenous CAT activity in eukaryotic cells and the CAT enzyme can be monitored by a rapid and sensitive assay. Similar reporter vectors have been constructed containing identical promoter elements linked to the luciferase gene from the firefly Photinus pyralis. Different types of luciferase reporters are now available which allow the measurement of more than one form of the enzyme in the same cell extract and modifications have been made to the luciferase gene, for example in the pGL3 series of reporters (Promega), which ensure that the enzyme is expressed and processed efficiently in mammalian cells. To reduce background activity most vectors have also been modified to minimize background transcription derived from cryptic promoters within vector sequences. A diagram of a typical reporter vector is shown in Figure 6. One potential criticism of assays which depend on the measurement of an enzyme activity such as CAT is that this is not a direct measurement of transcriptional activation. A number of studies have shown, however, that the level of CAT activity correlates with the steady state level of mRNA produced. Levels of mRNA may be analysed directly using the techniques of primer extension or RNAase mapping and these methods also confirm correct initiation of transcription. This is more easily accomplished using p-globin as a reporter gene which produces relatively stable RNA transcripts compared to CAT. These assay systems, however, are relatively time consuming and contrast with enzyme assays which are both sensitive and simple to perform and may be used to analyse a large number of samples. Nevertheless it is important to establish directly that any alteration in activity results from a change in the rate of initiation of transcription. 4.2.1 Controls for transfection efficiency Analysis of transcription factor function in a co-transfection experiment also requires the introduction of an internal control, for example a different reporter gene expressed from a promoter unaffected by the factor being analysed. This is particularly important where comparison is being made either between mutated forms of the transcription factor and the normal protein, or with a factor where transactivation is measured in the presence and absence of an inducer. In these experiments the data obtained from the reporter must be 197

Roger White and Malcolm Parker

normalized for both transfection efficiency and for general effects on transcription. A suitable control plasmid expressing B-galactosidase, which has been used to normalize data in the analysis of oestrogen receptor activity, consists of the lac-Z gene controlled by the CMV promoter. Both luciferase and B-galactosidase activity may be measured using luminescence assays with the reactions carried out in luminometer cuvettes or microtitre plates and measured in a luminometer where the peak or continuous light emission is recorded. The reporter gene activity is then normalized using the control gene activity to obtain a comparison of different samples. A simple procedure is used to obtain cell extracts suitable for the measurement of CAT, luciferase, and B-galactosidase activity and a protocol for these assays is described in Protocols 8 and 9.

4.3 Methods of transfection A number of different methods have been developed for the introduction of DNA into cells, all of which are suitable for the analysis of transcription factor activity. The method chosen depends to a large extent on the particular cell line to be transfected. The most commonly used methods are (a) calcium phosphate mediated transfection (b) electroporation (c) lipofection 198

8: Analysis of cloned factors Calcium phosphate mediated transfection may be used on a wide variety of cell lines, but the efficiency of uptake of DNA may vary between cell types. Electroporation, which involves applying a brief high-voltage electric pulse to the cells to generate pores in the plasma membrane, results in a large percentage of cells taking up DNA, although the precise conditions for each cell type must be determined empirically. Lipofection has been found to be the method of choice for a number of cell lines which transfect inefficiently by other techniques and a number of reagents are commercially available, but most of these are relatively expensive compared to calcium-phosphate-based techniques. This section will describe a calcium phosphate co-precipitation method of transfection (18) which has been used to analyse oestrogen receptor function in HeLa cells. Similar results have been obtained using lipofection. A method for introducing DNA into COS-1 cells using electroporation is described in Protocol 5. In all cases, the DNA used should be of high quality, purified from contaminating RNA and nicked or open circular DNA, and all traces of ethanol and buffers remaining from column purification methods must be eliminated. The best method for preparing large amounts of DNA is by ultracentrifugation in caesium chloride gradients. Protocol 7. Transient transfection using calcium phosphate coprecipitation Equipment and reagents • 2.5 M CaCI2 (filter sterilize through 0.45 um filter and store at -20°C)

. 2 x BBS: 50 mM BES (N,N-bis[2-Hydroxyethyl]-2-amino-ethanesulfonic acid) (Sigma B6137), 280 mM NaCI, 1.5 mM Na2HPO4; adjust to pH 6.95 with 1 M NaOH, filter sterilize, and store at -20°C

Method 1. Plate cells at approximately 30-40% confluence in 24-well microtitre plates. Handle the plate carefully and do not swirl the plate to ensure even plating. Leave for 24 h to allow the cells to attach. 2. Remove the medium from the wells by aspiration and replace with fresh medium. 3. Prepare DNA solutions for transfection by diluting 1 ug reporter plasmid, 0.1-0.3 ug internal control plasmid, and 0.5 ug expression plasmida with sterile distilled water to a final volume of 45 ul. Add 5 ul of 2.5 M CaCI2 and mix gently, then add 50 ul of 2 x BBS, mix again and leave at room temperature for 15-20 min. At this point the solution appears slightly cloudy as the DNA precipitate begins to form. 4. Add the 100 ul suspension dropwise to the cells and leave at 37°C for 18-24 h for the precipitate to continue to form and for DNA uptake to occur.

199

Roger White and Malcolm Parker Protocol 7.

Continued

5. Remove the medium by aspiration and wash the cells with fresh medium carefully 2-4 times to remove the precipitate.b 6. Re-feed the cells with fresh medium and leave at 37°C for 24-48 h. a Amounts of expression vector must be varied to optimize expression of the cloned factor. The amount of expression vector DNA must be constant and made up to 0.5 ug using the vector lacking the cloned insert. b The wash medium can be left on the cells for 10-20 min at 37°C to allow the precipitate to dissolve.

4.4 Assays for reporter genes in transfected cells The following protocols describe methods for analysing luciferase activity, b-galactosidase activity, and CAT activity in transfected cells. Protocol 8.

Preparation of cell lysates to assay luciferase and (3-galactosidase activity

Equipment and reagents • Lysis buffer: 0.65% NP40, 10 mM Tris-HCI pH 8.0, 1 mM EDTA pH 8.0, 150 mM NaCI • Luminometer cuvettes ILife Sciences International (UK) Ltd] • Luciferase reaction buffer: 25 mM glycyl glycine pH 7.8, 5 mM ATP pH 8.0, 15 mM

MgSO, Luciferin (Sigma) at 3 mM Galacto-light™ B-galactosidase assay system (Applied Biosystems Ltd) containing Galacton" substrate, reaction buffer, and Accelerator"

Method 1. Wash the transfected cells twice with 1 ml PBS at room temperature, carefully removing all the buffer from the final wash by tilting the plate. 2. Pipette 60 ul of lysis buffer on to the cell monolayer of each well of the microtitre plate and leave the plate horizontal for 2 min until the cells have lysed and only intact nuclei are visible under the microscope. Tilt the plate for 1-2 min to allow the lysate to drain to one side of the wells. 3. Transfer the lysate to a pre-cooled microcentrifuge tube with an automatic pipette and store on ice. Spin for 1 min in a microcentrifuge to pellet cell debris and transfer the supernatant to a clean tube. Store at 4°C on ice.3 4. Measure luciferase activity by adding 20 uJ of extract to 350 (ul luciferase reaction buffer in a polystyrene luminometer cuvette.6 5. Load samples into the luminometer (LKB 1251) and inject 33 uJ of 3 mM luciferin using an LKB 1291 dispenser. Record the peak light emission. 200

8: Analysis of cloned factors 6. Measure b-galactosidase activity by adding 2-10 ul of lysate from step 3 to 100 ul of reaction buffer containing a 1:100 dilution of Galacton" in a luminometer cuvette. Leave at room temperature for 60 min. 7. Add 150 ul Accelerator" and leave at room temperature for 10 min. 8. Place the samples into the luminometer and read the level of light emission. "Luciferase is unstable and the activity in the lysate decreases significantly if the samples are allowed to warm to 25-30°C. b Reagents are commercially available with extended light emission kinetics.

The lysate produced by steps 1 to 3 of Protocol 8 may be used directly to determine CAT activity. This is measured using either chromatography or an assay which is non-chromatographic and directly quantitative (19). Protocol 9. Measurement of CAT activity Equipment and reagents • Chloramphenicol: 8 mM stock in water diluted from a 0.1 M stock in 50% ethanol • [14C]acetyl CoA (Amersham International, CFA 729), 1.96 GBq/mmol

. Lysis buffer: 0.65% NP40, 10 mM Tris-HCI pH 8.0, 1 mM EDTA pH 8.0, 150 mM NaCI

Method 1. Heat 20 JJL! of extract to 68°C for 5 min in safe-seal or multi-click microcentrifuge tubes. 2. Add 20 fil of 8 mM chloramphenicol. 3. Add 20 ul of 0.5 mM acetyl CoA containing 0.1 HCi of [14C]acetyl CoA. Add 10 ul lysis buffer and 30 ul 0.25 M Tris-HCI pH 7.8. 4. Heat samples at 37°C for 1 h. 5. Add 100 nl of ice-cold ethyl acetate (ethyl ethanoate) to each sample. If a large number of samples is being processed, do no more than 8-12 samples at one time as ethyl acetate is extremely volatile. Vortex, and spin samples in a microcentrifuge for 2 min. Chloramphenicol and its acetylated derivatives are soluble in ethyl acetate while the acetyl CoA remains in the aqueous layer. 6. Remove 80 (ul of the upper organic layer and pipette directly into a scintillation vial containing 5 ml of liquid scintillant. Ensure that no material from the aqueous layer is transferred to the scintillation vial. 7. Add another 100 ul of ice-cold ethyl acetate to the aqueous layer and repeat the extraction process, this time transferring 100 ul of the organic layer to the scintillation vial.

201

Roger White and Malcolm Parker Protocol 9. Continued 8. Count the samples in a scintillation counter.a a Mock transfected cells give an assay background of approximately 250 c.p.m. Complete conversion gives a value of about 150000 c.p.m. but the assay is only linear up to 60000 c.p.m. Extracts which give higher values than this must be diluted in lysis buffer and re-assayed within the linear range. The data obtained from an analysis of the CAT activity in transfected cells are normalized using the results measured from an internal control derived from the same extracts. The final results may be presented either graphically or in a tabular form.

4.5 Identification of transactivation domains using chimeric proteins The construction of cDNA clones encoding fragments of a cloned factor has been described in Section 2.1 and the fusion of these fragments to create chimeric polypeptides for expression in bacterial cells has been described in Section 3.2. In a similar way the transactivation properties of the factor may be analysed by fusing either fragments or the entire factor itself to a heterologous DNA binding domain and expressing the chimeric protein in mammalian cells or yeast. In this way the boundaries of individual transactivation domains may be determined. An example of this approach is shown below in which either the ligand binding domain of the oestrogen receptor and fragments of SRC1 (constructed in Section 2) are linked to a 147 amino acid sequence identified as the DNA binding domain of the yeast protein GAL4 in the expression vector pSG424 (20). The chimeric polypeptides formed are expressed in HeLa cells (Protocol 7) and assayed for their ability to activate transcription of a reporter gene controlled by a promoter containing a series of GAL4 binding sites (21). Diagrams of these vectors and the results of these assays are shown in Figure 7. This demonstrates that SRC1a contains a single transactivation domain which is encoded by sequences between amino acids 781 and 988 (2).

5. Analysis of protein—DNA interactions 5.1 Analysis of DNA binding activity using a gel retardation assay Methods have been described in Section 3.1.1 for the preparation of in vitro translated proteins (Protocol 2) and of whole cell extracts (Protocol 6). The DNA binding activity of a factor is determined using a gel retardation assay in which a 32P-labelled oligonucleotide containing a response element or binding site for the transcription factor is first incubated with the whole cell extract. The protein-DNA complexes which form are separated from the unbound or free DNA on vertical acrylamide gels and visualized by autoradiography. The single-stranded oligonucleotides forming the oestrogen receptor binding site 202

8: Analysis of cloned factors

Figure 7. (A) Diagrams of the reporter vector p5xGAL4CAT and the expression vector pSG424. p5xGAL4CAT contains five GAL4 binding sites linked to a simple promoter consisting of a TATA box. pSG424 contains the GAL4 DNA binding domain (DBD) (1-147) which may be fused to different fragments of SRC1a. (B) Transactivation assay of SRC1a deletion mutants.

203

Roger White and Malcolm Parker (ERE) have the sequence 5'-CTAGAAAGTCAGGTCACAGTGACCTGATCAAT-3' and 5'-CTAGATTGATCAGGTCACTGTGACCTGACTTT-3'. These are annealed and labelled with [32P]dCTP using the Klenow fragment of DNA polymerase to form a double-stranded probe containing a consensus oestrogen receptor binding site. Protocol 10. Preparation of 32P-labelled oligonucleotide for DNA binding assays Equipment and reagents • 1 x Annealing buffer: 0.1 M NaCI, 10 mM Tris-HCI pH 8.0, 1 mM EDTA « 10 X Repair buffer: 0.5 M Tris-HCI pH 7.6, 0.1 M MgCI2

• [32P]dCTP in aqueous solution at 110 TBq/ mmol and 10 mCi/ml (Amersham)

Method 1. Anneal the DNA to form the probe by mixing in a microcentrifuge tube 0.5 ug of each oligonucleotide in a total volume of 100 ul of annealing buffer. 2. Heat the sample to 85°C for 3 min in a small water bath, switch off the bath and allow the sample to cool for 30 min to below 45°C. 3. Add 200 ul of ethanol, vortex, and place the sample on dry ice for 30 min to precipitate the DNA. 4. Spin in a microcentrifuge for 15 min at 4°C, carefully remove the supernatant and wash the pellet with cold 70% ethanol. Spin for 15 min at 4°C, remove the supernatant, and allow the pellet to dry. 5. Dissolve the annealed oligonucleotides in 10 ul distilled water to give a concentration of 100 ng/ul. 6. Label the DNA using the following reaction: 2 ul annealed DNA, 2 ul 10 X repair buffer, 10 ul distilled water, 5 ul [32P]dCTP, 1 ul Klenow DNA polymerase. Incubate at 37°C for 30 min. 7. Add an equal volume of 50:50 phenohtrichloromethane, vortex, spin for 2 min in a microcentrifuge, and remove the upper aqueous layer to a clean tube. Repeat the phenol:trichloromethane extraction and remove the aqueous layer to a clean tube. 8. Purify the labelled DNA from the unincorporated nucleotides by precipitation (steps 9-11) or by using a spin column. 9. Add 20 ul 5 M ammonium acetate and 80 ul ethanol. Place on dry ice for 30 min to precipitate and recover the DNA by centrifugation in a microcentrifuge for 15 min. 10. Resuspend the pellet in 20 ul of distilled water, add 20 ul 5 M ammo-

204

8: Analysis of cloned factors nium acetate and 80 ul ethanol. Place on dry ice for 30 min to precipitate and recover the DNA by centrifugation as before. 11. Wash the labelled DNA probe with 70% ethanol, spin for 15 min, and carefully remove the supernatant and leave the pellet to dry. Resuspend the DNA in 20 ul to give a final concentration of 10 ng/ul.

Protocol 11. Analysis of DNA binding activity using a gel retardation assay Equipment and reagents • 2 x Binding buffer: 40 mM Hepes pH 7.4, . Poly[d(l-C)-d(l-C)](1 mg/ml) 100 mM KCI, 2 mM 2-mercaptoethanol, . 5 x TBE: 450 nM Tris-borate (pH 8.3), 10 nM 20% glycerol EDTA . 32P-labelled oligonucleotide (Protocol 70)

Method 1. Pipette 10 ul of 2 x binding buffer into a microcentrifuge tube. 2. Add 1 »J of 1 mg/ml poly[d(l-C)-d(l-C)] and 1 ul of 10% BSA. 3. Add lysate or whole cell extract. The amount of protein added is determined empirically, and in the example of the oestrogen receptor suitable values are 2 ul of in vitro translated receptor, 2 ul of COS-1 cell extract or the equivalent of 0.1 ul of extract from recombinant baculovirus-infected insect cells. To measure non-specific binding, prepare a parallel set of reactions using equivalent amounts of either control lysate or whole cell extract lacking receptor. 4. Make the total volume up to 19 ul with distilled water and incubate the samples at room temperature for 15 min. 5. Add 1 ul of [32P]oligonucleotide at 1 ng/ul to each sample and incubate for 30 min at room temperature and then 30 min at 4°C. 6. Prepare a 6% polyacrylamide gel (30% acrylamide, 0.8% bisacrylamide) in 0.5 x TBE buffer. 7. Before running the gel, rinse the slots with 0.5 X TBE buffer to remove unpolymerized acrylamide. Pre-run the gel in 0.5 x TBE at 100 V for 30 min. 8. Load the samples directly on to the gel and run at 250 V. Load marker dye in the outside track(s) to monitor the electrophoresis. 9. Fix the gel in 10% acetic acid, 30% methanol for 15 min, dry, and autoradiograph.

5.1.1 Application of the gel retardation assay to transcription factor analysis The gel retardation assay may be applied to determine the affinity of the protein-DNA interaction by using increasing concentrations of input DNA 205

Roger White and Malcolm Parker probe and measuring the amount of the protein-DNA complex formed with a phosphor imager. To directly compare either different samples or different reaction conditions it is important that the DNA probe is in excess. Specific binding of the protein to DNA is confirmed by the inclusion of specific antisera. This may either generate a decrease in mobility (or 'supershift') of the protein-DNA complex in the gel as a result of the binding of the antibody or may result in the loss of the complex due to interference with DNA binding. In the example of the oestrogen receptor the gel retardation assay may be used to observe the effect of the addition of ligand on protein-DNA complexes. Hormone agonists, for example oestradiol, or antagonists, such as tamoxifen, may be added to the binding reaction either before or after the addition of the DNA probe and alterations in the mobility of the complex may reflect conformational changes in the protein.

6. Analysis of protein-protein interactions in vitro Interactions between the factor and other proteins may be analysed in vitro using a number of different techniques. The following protocol describes methods of measuring protein-protein interactions in the absence or presence of DNA. The GST pull-down assay uses the potential to immobilize GST fusion proteins on glutathione-Sepharose beads. The immobilized protein is then incubated with 35S-labelled in vitro translated proteins and, after washing to remove unbound material, the bound proteins are eluted and analysed by SDS-PAGE and autoradiography.

6.1 Analysis of interactions between proteins using a GST pull-down assay Protocol 12.

GST pull-down assay

Equipment and reagents .

35

S-labelled in vitro translated protein (Protocol 2) • NETN with protease inhibitors (Protocol 4) at 40°C

• GST bead suspension (Protocol 4, step 6) . 2 x SDS-PAGE buffer: 4% SDS, 20% glycerol, 125 mM Tris-HCI pH 6.8, 2% bmercaptoethanol, 0.1% bromophenol blue

Method 1. Add 1 ml of cold NETN with protease inhibitors to the beads and divide solution over five multi-click or safe-seal microcentrifuge tubes. 2. Dilute the 35S-labelled protein in NETN containing 10% glycerol and protease inhibitors and add to the GST bead suspension. 3. Incubate for 3 h at 4°C on a rotating wheel. 206

8: Analysis of cloned factors 4. Spin beads down for 5 sec at 10000 x g in a microcentrifuge, aspirate supernatant, and add 500 ul of cold NETN. Mix gently by inverting and spin again. 5. Repeat step 4 twice. 6. Dry in a SpeedVac for approximately 30 min and resuspend in 20 ul 2 X SDS-PAGE buffer. 7. Boil samples for 3 min and spin beads down for 15 sec to pellet the beads. 8. Load 10 ul of each sample and run on an SDS-PAGE gel to resolve the bound proteins. 9. Fix gel in 10% acetic acid/10% methanol, amplify the signal with Amplify (Amersham), and expose to film at -70°C.

An example of this type of analysis is shown in Figure 8 where fragments of SRC1a (Protocol 1) have been cloned into the pGEX vector and expressed as fusion proteins with GST (Protocol 3). These have then been tested in the pull-down assay for interaction with in vitro translated MOR or a transcriptionally inactive MOR mutant in which two leucine residues in the hormone binding domain have been mutated to alanine. Using this technique, liganddependent interactions are observed between the receptor and the SRC fragments 570-780 and 1240-1441 (2).

Figure 8. GST pull-down analysis of SRC1a deletion fragments with the wild type and a mutant MOR, Fragments 570-780 and 1241-1441 bind in a hormone-dependent manner.

207

Roger White and Malcolm Parker

6.2 Analysis of protein-protein interactions using immunoprecipitation The following protocol describes a method for determining protein-protein interaction which in this case uses immunoprecipitation to measure dimerization between two different forms of the same protein in solution. This is a form of a co-immunoprecipitation assay suitable for analysing interactions between any two factors where a specific antiserum is available for one of the proteins. Protocol 13. Analysis of dimerization using immunoprecipitation Equipment and reagents • 1 x Immunoprecipitation buffer (IB): 100 mM KCI, 20 mM Hepes pH 7.4, 1 mg/ml BSA, 0.1% NP40, 10% glycerol

• Protein A-Sepharose beads

Method 1. Select two different forms of the protein, only one of which contains the epitope for the antibody to be used. 2. Co-translate the proteins in vitro in the presence of [35S]methionine (Protocol 2). 3. Prepare Protein A-Sepharose beads (PAS) by equilibrating the beads in IB overnight at 4°C. Spin the beads for 10 sec in a microcentrifuge and wash 3 times with IB. 4. Dilute aliquots of the lysate containing 105 c.p.m. in IB to a final volume of 50 ul. 5. Microcentrifuge the sample for 5 min to pellet aggregates. 6. Add antiserum or preimmune serum. The amount used depends on the quality of the serum and should be determined empirically. Typically 1-5 ul of antiserum is used. 7. Incubate at 4°C overnight, or for 30 min at room temperature. 8. Add 50 ul of a 50% slurry of PAS and incubate at 4°C for 30 min. 9. Pellet the beads by spinning for 10 sec in a microcentrifuge. 10. Wash the beads three times with IB. 11. Add 2 x SDS-PAGE sample buffer and analyse using SDS-PAGE.

This type of experiment is illustrated in Figure 9. A 35S-methionine labelled N-terminal deletion mutant of the mouse oestrogen receptor 121-599 is cotranslated with either the labelled full length protein 1-599 or full length protein containing specific point mutations. A specific antipeptide antiserum MP15, which has been raised to amino acids 17-28, is then used to immuno208

8: Analysis of cloned factors

Figure 9, Analysis of protein-protein interaction by immunoprecipitation.

precipitate the full length protein, together with the truncated form of the protein where this deletion mutant is associated with the full length form in a heterodimer (22). This approach can be used to map the boundaries of dimerization domains and assess the effect of specific point mutations on dimer formation.

6.3 Analysis of protein-protein interactions using the gel retardation assay Protein-protein interactions and dimerization may also be analysed in the presence of DNA using the gel retardation assay (Protocol II) by the formation of complexes between proteins of different sizes. Co-translation (Protocol 2) or co-expression (Protocols 5 and 7) of a full length protein and a deletion mutant results in the formation of heterodimcrs which migrate in an intermediate position in the gel relative to the two different homodimers of the protein. This type of analysis is illustrated in Figure 10 by an assay in which the full length oestrogen receptor MOR 1-599 is co-translated with an N-terminal deletion mutant MOR 121-599, both of which retain residues required for high affinity DNA binding (22). Although this assay is measuring a dimerization function, detection of the protein-DNA complexes depends on the ability of the proteins to bind to DNA.

6.4 Analysis of protein-protein interactions on DNA using an ABCD assay The interaction of proteins on DNA can be assayed using an avidin-biotin complex on DNA (ABCD) assay in which the DNA binding site for one protein is synthesized as a biotinylated oligonuclcotide and immobilized on streptavidin paramagnetic beads (23, 24). The DNA binding protein is allowed to bind to the oligonucleotide and this protein-DNA complex incubated w i t h either 35S-labelied proteins or cell lysates. After washing to remove unbound proteins the bound protein(s) are eluted and detected by autoradiography or Western blotting. 209

Roger White and Malcolm Parker

Figure 10. Analysis of protein-protein interaction by gel retardation.

Protocol 14. Analysis of protein-protein interactions on DNA using an ABCD assay Equipment and reagents • Streptavidin-coated magnetic beads Istreptavidin MagneSphere paramagnetic particles, Promega Z5481, containing 15 X 0.6 ml aliquotsl a • Buffer A: 8 mM trisphosphate pH 7.4, 120 mM KCI, 8% glycerol, add DTT to 4 mM and the following protease inhibitors immediately prior to use: 5 ul of 100 mg/ml bacitracin, 4 ul of 10 mg/ml PMSF, 2 ul of 2.5 mg/ml leupeptin and 2 ul of 1 mg/ml aprotinin per 1 ml of buffer. Store at 4"C

. Buffer H: 20 mM Hepes pH 7.7, 50 mM KCI, 20% giycerol, 0.1% NP40 containing DTT to 4 mM and protease inhibitors added immediately prior to user 5 ul of 100 mg/ml bacitracin, 4 ul of 10 mg/ml PMSF, 2 ul of 2.E mg/ml leupeptin and 2 ul of 1 mg/ml gprotinin per 1 ml of buffer. Store at 4°C • MagneSphere magnetic separation stand (Promega, 25332)

Method 1. Wash the streptavidin-coated magnetic beads from one 0.6 ml aliquot a three times with PBS and once with buffer A. 2. Resuspend the beads in 800 ul buffer A and divide into 4 x 200 ul aliquots.

210

8: Analysis of cloned factors 3. In a separate tube add 1 ug of annealed double-stranded oligonucleotide (Protocol 10, steps 1 and 2), of which one strand is synthesized biotinylated at the 5' end, with 50 ul of a lysate containing overexpressed protein (Protocol & in a total volume of 200 ul buffer A. 4. Incubate at room temperature for 10 min. 5. Add 200 ul of the bead suspension to the binding site/protein mix and incubate at room temperature for 10 min. 6. Wash four times in buffer H, once in buffer A, and resuspend in 500 ul of buffer A.c 7. Add in vitro translated protein or 1.0-1.5 mg of a whole cell lysate (Protocol & and incubate for 60 min at 4°C. 8. Wash three times with buffer H and then add 20 ul 2 x SDS-PAGE sample buffer to the beads and boil for 3 min. 9. Place the tube(s) on the magnetic stand to attract the beads and analyse 10 ul of each sample by SDS-PAGE. Detect bound proteins by autoradiography/fluorography or Western blotting. a One tube is sufficient for four samples. b All washing is done by adding buffer, vortexing slowly, and allowing beads to be attracted to the side of the tube on the magnetic separation stand. c To investigate hormone-dependent interaction in the case of the oestrogen receptor, add 500 ul buffer A with or without hormone (final concentration 1 uM) and incubate at room temperature for 30 min.

6.5 Analysis of protein-protein interactions in cells using two-hybrid analysis The basic principles of the two-hybrid assay are shown in Figure 11. This technique requires the co-expression of two different proteins and allows measurement of the ability of the proteins to interact with each other (25-27). One, termed the bait, is a chimeric protein which contains a DNA binding domain but lacks transcriptional activation domains. The second protein, the target, must lack DNA binding activity and either contain its own activation function or be fused to a strong transactivation domain, for example the acidic activation domain of the herpes virus transcriptional activator VP16. The coexpression assay may be carried in yeast or mammalian cells, with yeast having the advantage that a large number of different proteins may be screened for their ability to interact with the DNA bound protein. Co-expression of two interacting proteins results in the recruitment of the activation sequence to the promoter of the reporter gene by the DNA-bound bait and an increase in reporter activity which can be measured using the appropriate assay system. The two-hybrid assay may be carried out in mammalian cells by co-expression of two proteins, for example the GAL4 LBD of the oestrogen receptor and full length SRC1 using transient transfection as described in Protocol 7, 211

Roger White and Malcolm Parker

Figure 11. Model of a two-hybrid assay.

Reporter gene activity and the ability of the two proteins to interact in cells may be measured using the methods described in Protocols 8 and 9.

7. Summary In this chapter we have described the approaches and some of the methods used to analyse the properties of a cloned factor. These have been illustrated using two different proteins; the oestrogen receptor which binds directly to DNA, and SRC1a, a coactivator which is recruited to nuclear receptors and may act as an intermediary factor between the receptors and the basal

212

8: Analysis of cloned factors transcriptional machinery. The basic techniques described may, however, be applied to the study of any novel factor.

References 1. White, R., Lees, J. A., Needham, M., Ham, J, and Parker, M. (1987). Mol. Endocrinol. 1, 816. 2. Kalkhoven, E., Valentine, J. E., Heery, D. M., and Parker, M. G. (1998). EMBO J., 17, 232. 3. Parker, M. G. (1993). Curr. Opin. Cell Biol., 5, 499. 4. Brzozowski, A. M., Pike, A. C. W., Dauter, Z., Hubbard, R. E., Bonn, T., Engstrom, O., et al. (1997). Nature, 389, 753. 5. Lewitter, F. (1997). Trends Genet. 13, 286. 6. Kozak, M. (1987). Nucleic Acids Res., 15, 8125. 7. White, R., Butler, A., and Parker, M. G. (1995). In Gene probes 2: a practical approach (ed. D. Rickwood and B. D. Hames), p. 329. IRL Press, Oxford. 8. Maniatis, T., Fritsch, E. F., and Sambrook, J. (eds) (1989). Molecular cloning, a laboratory manual. Cold Spring Harbor Press, Cold Spring Harbor, NY. 9. Green, S., Issemann, I., and Sheer, E. (1988). Nucleic Acids Res., 16, 369. 10. Mellon, P., Parker, V., Gluzman, Y., and Maniatis, T. (1981). Cell, 27, 279. 11. Summers, M. D. and Smith, G. E. (1987). A manual of methods for baculovirus vectors and insect cell procedures. Texas Agrigulture Experiment Station Bulletin, No. 1555. Texas Agricultural Experiment Station, Texas A&M University, College Station, Texas. 12. Kitts, P. A., Ayres, M. D., and Possee, R. D. (1990). Nucleic Acids Res., 18, 5667. 13. Gorman, C. M., Moffet, L. F., and Howard, B. H. (1982). Mol. Cell. Biol., 2, 1044. 14. Hall, C. V., Jacob, P. E, Ringold, G. M., and Lee, F. (1983). J. Mol. Appl. Genet., 2, 101. 15. De Wet, J. R., Wood, K. V., Pehuea, M., Helinski, P. R., and Subromani, S. (1987). Mol. Cell. Biol., 7, 725. 16. Klein-Hitpass, L, Schorpp, M., Wagner, U., and Ryffel, G. (1986). Cell, 46, 1053. 17. Luckow, B. and Schutz, G. (1987). Nucleic Acids Res., 15, 5490. 18. Chen, C. and Okayama, H. (1987). Mol. Cell. Biol., 7, 2745. 19. Sleigh, M. J. (1986). Anal. Biochem., 156, 251. 20. Sadowski, I. and Ptashne, M. (1989). Nucleic Acids Res., 17, 7539. 21. Lillie, J. W. and Green, M. P. (1989). Nature., 338, 39. 22. Fawell, S. E., Lees, J. A., White, R, and Parker, M. G. (1990). Cell, 60, 953. 23. Kurokawa, R., Soderstrom, M., Horlein, A., Halachmi, S., Brown, M., Rosenfeld, M. G., and Glass, C. K. (1995). Nature, 377, 451. 24. L'Horset, F., Dauvois, S., Heery, D. M., Cavailles, V. and Parker, M. G. (1996). Mol. Cell. Biol., 16, 6029. 25. Fields, S. and Song, O. (1989). Nature, 340, 245. 26. Finkel, T., Duc, J., Fearon, E. R., Dang, C. V., and Tomaselli, G. F. (1993). J. Biol. Chem, 268, 5. 27. Nagpal, S., Saunders, M., Kastner, P., Durand, B., Nakshatri, H., and Chambon, P. (1993). Cell, 70, 1007. 213

This page intentionally left blank

9 Neuronal promoter analysis by in vitro transcription using nuclear extracts from rat brain M. L. SCHWARTZ and W. W. SCHLAEPFER

1. Introduction We present a reliable method for making nuclear extracts from rat brain and for using these extracts in an in vitro transcription assay. Our nuclear extract technique allows for the isolation of nuclear proteins from post-mitotic neurones that presumably contain all of the factors required for the transcriptional regulation of neurone-specific genes. The use of brain as a source of nuclear extract is in contrast to the more commonly used cultured cell lines where expression of neurone-specific genes may be poor or lacking in tissue specificity. In vitro transcription using nuclear extracts from expressing (brain) and non-expressing (liver) tissues provide a direct assessment of promoter strength and level of tissue-specificity apart from the effects of higher order DNA structure (i.e. chromatin). In vitro transcription is also amenable to DNA competitor studies and to the addition of purified or partially purified protein factors and antibodies. In this way, true neurone-specific transcriptional activator or repressor sequences or factors can be characterized. The approach is rapid, reliable, and easily quantifiable. Finally, advancements in the field of nucleosome reconstitution have allowed the coupling of in vitro transcription to nucleosome assembly systems and provide a novel and promising approach to the understanding of neuronal-specific gene expression. In the first half of this chapter we describe a procedure for making nuclear extracts from tissues (specifically brain) that produce efficient and reliable extracts for use in in vitro transcription studies. The second half of this chapter describes our methodology for performing in vitro transcription using these nuclear extracts.

M. L. Schwartz and W. W. Schlaepfer

2. Background The first cell-free transcription systems using tissue culture cells were described over 10 years ago (1-3). Unfortunately, the efficiency of these systems and those derived from tissues (4) is relatively low (< 3% template usage). Since these initial procedures, modifications, such as rapid isolation and subsequent stabilization of nuclei in sucrose to prevent loss of transcription factors due to leakage, have been introduced. These modifications have dramatically improved the efficiency and reliability of extracts for in vitro transcription (5, 6). However, most systems still utilize extracts derived from cultured cells (see Chapter 3 for methods for preparing such extracts from HeLa cells). This may be a limitation in the study of neuronal genes that are poorly expressed in vitro or that lack tissue specificity when transfected into proliferating cells (7). Our laboratory has adapted the preparation of nuclear extracts from brain and used in vitro transcription for the study of the brain-specific neurofilament (NF) genes. The use of brain extract as a source of tissue extract provides a model system that potentially contains all of the components necessary for high-level NF transcription in post-mitotic neurones. Parallel experiments with extracts from non-expressing tissues (liver) or low-expressing cell lines provide a measure of tissue or cell specificity. Our method has been successful in assessing the strength and tissue-specific aspects of the NF promoters (8, 9, 12).

3. Applications 3.1 Comparison to transfection and transgenic mouse studies In vitro transcription allows the process of transcription initiation to be studied apart from the effects of chromatin. This is significant since nucleosomes or higher order structures can block the access of transcription factors or the initiation complex to promoter regulatory elements, thus making it difficult to study basic promoter elements. We have assayed neurofilament (NF) promoter constructs that are not able to function in transgenic mice (10) or that showed aberrant start sites in transient transfection studies (7). These studies allowed us to define clearly positive and negative acting regions or sites within these promoters (8, 9, 12). From a practical aspect, in vitro transcription offers many advantages over transfection or transgenic mouse studies since it is faster and less labour intensive to perform. We have, in one day, studied 15 different promoter constructs using two different tissue extracts and obtained well controlled and easily quantifiable results. The reaction is also amenable to the addition of competitor DNAs, purified or partially purified proteins, or antibodies to transcription 216

9: Neuronal promoter analysis factors. In addition, it is possible to obtain a more direct comparison between functional data and physical data, such as gel shift or DNAase I footprints, since all of these procedures can be conducted using the same nuclear extract preparations.

3.2 Studies of the effect of chromatin structure on transcription In vitro transcription can also be used to study the effects of chromatin structure. Recently, it has been shown that nucleosome reconstitution systems coupled with in vitro transcription can be used to study the relationship of transcription to a variety of important nuclear processes such as DNA replication, nuclear scaffolding, higher order chromatin organization, and cell cycle events (11). This methodology has successfully reproduced the appropriate erythroid-specific pattern of regulation of the adult bA-globin gene (11). A combined system may also allow the study of chromatin remodelling. Alterations in chromatin structure are likely to account for the activation of most, if not all, neurone-specific genes that arise in post-mitotic neurones. For further details of the methods used to analyse the effect of chromatin on transcription, see Chapter 10.

4. Preparation of nuclear extracts 4.1 Overview Briefly, nuclear protein extracts that are competent for in vitro transcription and gel shift assays are obtained by homogenizing tissue, pelleting nuclei through sucrose, and gently extracting proteins out of the nuclei using salt. The proteins are then precipitated with ammonium sulfate and dialysed against KCl/glycerol and stored at -80°C. It generally requires 8 h from removal of brains to the addition of ammonium sulfate. Sedimentation of the protein and dialysis is usually carried out on day two. We generally obtain 1.5 ml of 12 mg/ ml extract, enough to run approximately 375 in vitro transcription reactions at 4 u1 per reaction. This procedure represents a combination of two other procedures (4, 5) as well as some of our own modifications.

4.2 Purification of nuclei from rat brain All solutions should be made a day in advance using sterile water and kept at 4°C. Phenylmethylsulfonyl fluoride (PMSF) and dithiothreitol (DTT) are added just prior to use. Final volumes are per 100 g of tissue (50 brains). Manipulations can be done on the bench top but all tubes and solutions must be kept on ice. 217

M. L. Schwartz and W. W. Schlaepfer Protocol 1. Purification of nuclei Equipment and reagents • 50 rats at 4-5 weeks of age • 100 ml NIB-B: 9:1 NIB-A:100% glycerol, v/v • Motor driven 30 ml serrated Teflon-glass . 20 ml 2 X NRB: 20 mM Hepes pH 7.9, 6 mM homogenizer MgCI2, 200 uM EDTA, 2 mM DTT, 200 uM PMSF (Sigma), 20 ug/ml bestatin/leu• 7 ml Dounce homogenizer with loosepeptin/aprotinin/antipain (all Sigma), 30% fitting pestle . 500 ml NIB-A: 10 mM Hepes pH 7.9, 25 mM glycerol KCI, 150 uM spermine, 500 uM spermidine, 1 mM EDTA, 10 % glycerol, 2 M sucrose

Method 1. Process 50 rat brains (95-100 g) in two batches of 25 brains each. 2. Expose brain tissues in C02-euthanized animals by mid-line skull separation. Excise tissues rostral to the cervical spinal cord, section serially, and immerse in ice-cold Dulbecco's modified Eagle's medium (DMEM). 3. Wash brain tissues further in ice-cold DMEM and weigh in a large plastic weigh dish. 4. Pour 10 ml of NIB-A into the weigh dish and mince the brain tissue. Take care not to mince the tissue too finely. 5. Split minced tissue into six parts and place each part into a 30 ml homogenizer. Add NIB-A to about 20 ml and homogenize the tissue by two strokes of a slowly turning Teflon pestle (serrated edge). 6. Combine the six homogenates into a 250 ml beaker and dilute to 150 ml with NIB-A. 7. Layer 25 ml of the homogenate above 10 ml of fresh NIB-A in a SW27 tube (six tubes total) and centrifuge at 100 000 x g for 30 min at 4°C in a SW27 rotor. 8. While the nuclei from the first batch of 25 brains is being collected by centrifugation, repeat the above procedure (steps 2-7) for the second batch of 25 brains. 9. After centrifugation, remove the supernatant (myelin, red blood cells, etc.) by aspiration, leaving a white nuclear pellet. 10. Resuspend each pellet with 0.5 ml of NIB-B using a plastic bacti-loop and gentle vortexing and transfer to a 30 ml homogenizer using a glass Pasteur pipette with the tip mostly removed. Nuclei stick to the insides of plastic pipettes and are virtually impossible to recover. 11. Rinse tubes with an additional 0.5 ml of NIB-B. 12. Combine resuspended nuclei from both batches, homogenize with one stroke of a Teflon pestle, and dilute to 25 ml with NIB-B.

218

9: Neuronal promoter analysis 13. Layer this homogenate above 10 ml of fresh NIB-B and centrifuge at 100 000 g for 30 min at 4°C in an SW27 rotor. 14. Remove the supernatant by aspiration, leaving a white pellet of brain nuclei at the bottom of the tube. 15. Resuspend the pellet of nuclei with 1 ml of 1 x NRB using a bacti-loop and transfer into a 7 ml Dounce homogenizer. 16. Rinse the tube three times with 1 ml of 1 x NRB. 17. Disperse the nuclei well with 5-7 strokes of a loose-fitting pestle and transfer to a 50 ml plastic centrifuge tube. 18. Rinse the homogenizer twice with 0.5 ml 1 x NRB and note the volume. 19. Dilute a 5 ul sample of the nuclei to 500 ul with 1 x NRB and count the number of nuclei/ml directly using a haemocytometer.

4.2.1 Yield and quality of the nuclei The nuclei should appear round with very defined nucleoli and be free of any contaminants or debris. Generally, concentrations range from 2-4 X 108 nuclei/ml in 6-7 ml of 1 X NRB, giving a total of about 2-3 X 109 nuclei. Large neuronal nuclei represent approximately 15% of the total nuclei. We have not rigorously tested concentrations of nuclei well above or below these numbers; however, our best extracts have been obtained when KC1 extraction of protein (see Protocol 2) was performed at these concentrations. Too few total nuclei makes for smaller volumes and more difficult handling as well as limiting the final amount of extract. Numbers of nuclei in excess of 4 X 109 should be fine as long as the concentration (nuclei/ml) is kept within the guidelines stated above. We should also note here that, for reasons we have not pursued, we have been unable to obtain sufficient numbers of nuclei from 60 adult mice. 4.2.2 The effect of myelin on yield Because brain tissue contains a considerable amount of myelin, it is easy to overwhelm the homogenization process, reducing the yield of nuclei. This probably occurs either because nuclei are not freed from the cells or nuclei become entangled in the myelin and do not pellet. To this end we have chosen young rats (4-5 weeks) because, although their brains have less mass than adult brains, they have similar numbers of neuronal cells and significantly less myelin. We also perform the homogenization step in two batches of 25 brains and further divide each batch among six sucrose/glycerol gradients. In this way, we attempt to keep the homogenate dilute, thereby increasing the yield of nuclei by minimizing their entanglement with myelin. Since the nuclei are isolated in a sucrose/glycerol solution (NIB) we assume that the benefits of 219

M. L. Schwartz and W. W. Schlaepfer dilution significantly outweigh any potential loss of soluble transcription factors due to leakage (5, 6). The yield of nuclei can also be enhanced by as much as 25% by collecting the material at the boundary between the overlaid homogenate and the fresh NIB-A (Protocol 1, step 7). The material from this layer, which includes trapped nuclei and red blood cells, is collected from both sets of initial purifications (keeping the volume to a minimum), re-homogenized in NIB-B, and processed as described in Protocol 1, step 12. These nuclei are added to the main batch of nuclei before counting (Protocol 1, step 15).

4.3 Extraction of nuclear proteins Protocol 2.

Extraction of nuclear proteins by KCI

Reagents • Dialysis buffer: 25 mM Hepes pH 7.9, 100 mM KCI, 15% glycerol, 250 um EDTA, 1 mM DTT, 500 uM PMSF; sterile filtered

Method 1. Dilute nuclei to 2.0-4.0 X 108 nuclei/ml with 1 x NRB. 2. Obtain nuclear proteins by slowly adding (over about 20 min) an equal volume of 1 X NRB + 0.7 M KCI to the suspension of nuclei to a final KCI concentration of 0.35 M. Add the 1 x NRB + 0.7 M KCI by peristaltic pump while swirling the resuspended nuclei in a 50 ml plastic centrifuge tube in an ice bucket on a rotating platform shaker. 3. After the KCI is added, place the tube on its side and very slowly shake for 15 additional minutes. At the end of this time the solution will be non-viscous except for a 'glob' of nuclei. 4. Pellet nuclei by centrifugation at 125000 x g for 20 min at 4°C in a SW50.1 rotor. Nuclei form a tight, greyish opaque pellet. 5. Combine the supernatants from each tube, determine the volume, and transfer into a 25 ml flask containing a spin bar. 6. Add ammonium sulfate slowly to 0.33 g/ml and allow the proteins to precipitate at 4°C overnight. 7. Transfer the precipitated proteins to 13 ml quick-seal tubes and sediment by centrifugation at 85000 g for 30 min at 4°C in a 75Ti rotor. 8. Resuspend protein pellets with 1 ml of dialysis buffer per2.5 x 109nuclei. 9. Dialyse the extract twice for 45 min against 250 ml dialysis buffer. 10. After dialysis, transfer the extract to a 2 ml microcentrifuge tube and clarify by centrifugation at 14000 g for 10 min. A small dark precipitate is visible after this step. 11. Determine protein concentration (10-14 mg/ml) by Bradford assay. 220

9: Neuronal promoter analysis 12. Freeze extracts with dry ice in small aliquots (100 ul) and store at —80°C. We have found little loss of activity over 1 year and with 2-3 freeze/thaw cycles. 13. Determine the optimal amount of extract (generally 3-4 ul) needed for in vitro transcription (see Protocol 3) for each extract.

4.3.1 The effect of the rate and amount of KC1 addition Extraction of nuclear proteins with 0.35 M KC1 is the most critical step of the procedure. If too little salt is added, insufficient amounts of protein are extracted. Too much salt and the nuclei break open, spilling chromatin and histones into the nuclear protein preparation, causing the extracts to be of poor quality and efficiency for in vitro transcription and gel shift assays. We have arrived at 0.35 M KC1 through much trial and error and are confident that this concentration of salt yields the best quality brain and liver extracts. Salt must also be added slowly to prevent local pockets of high KC1 concentration that will cause nuclei to break open. Interestingly, the amount of salt needed to extract nuclear proteins seems to vary widely between cultured cells and tissues (unpublished observations). For example, Kamakaka et al. (6) use only 0.1 M KC1 to extract proteins from Drosophila embryos. In addition, the sensitivity of different nuclei to breakage by salt also varies widely, with nuclei from tissues being generally more sensitive than those from cultured cells (unpublished observations). 4.3.2 Post-extraction centrifugation notes Breakage of the nuclei can also occur during pelleting. At the suggestion of Dr Jim Kandonaga (personal communication) we pellet the nuclei in a swinging bucket rotor rather than a fixed-angle rotor. This minimizes shearing of the nuclei against the walls of fixed angle tubes. Nuclear breakage is characterized by a viscous solution (chromatin) that will not pellet under the conditions described above. The quality of the post-salt-extracted nuclear pellet can further be checked by visual inspection. This pellet should be opaque grey and contain 'ghost' nuclei that have lost defined internal structures (i.e. nucleoli) but maintain a round shape. If there is significant breakage of nuclei, the extract can still be salvaged by pelleting the chromatin with a high speed spin at 4°C in a 75Ti Beckman fixed angle rotor spun at 150000 g for 90 min The ammonium sulfate precipitation can then be carried out as described. This extract should still be useful for transcription as well as physical studies (i.e. gel shift assays) or factor purification but may be contaminated with small amounts of histones and/or DNA. 4.3.3 Resuspension of nuclear proteins This part of the procedure is fairly straightforward protein work. We attempt to keep the extract concentration high (approx. 1 mg/ml) in order to keep the 221

M. L. Schwartz and W. W. Schloepfer proteins and protein complexes from falling apart. Yet, consideration is given to keeping working volumes manageable. The final dilution in dialysis buffer (1 ml per 2.5 X 109 nuclei) was difficult to determine. Generally, 3-5 ul of extract are needed per 25 u1 reaction for optimal in vitro transcription, thus leaving room for pipetting error. From our experience, protein concentrations that are too high (> 18 mg/ml) tend to have a very narrow range of amount needed (in ul) for full activity of the extract, thus making pipetting error very significant. Concentrations that are too low (< 5 mg/ml) need large ul amounts of extract and will raise the amount of glycerol and KCl per in vitro transcription reaction. In addition, these extracts may not work at all, perhaps due to the loss of proteins or protein complexes at low protein concentrations, 4,3.4 Extract quality We have not rigorously tested the efficiency of our extracts by determining the number of transcripts per template. We do know that using the in vitro transcription procedure described below, we get a very strong signal from 0.5 nM adenovirus major-late promoter on an overnight exposure at room temperature without an enhancing screen (Figures 1 and 2). In addition, we have tested our extracts against a commercially available HeLa cell extract (Promega) that is obtained by the method of Dignam et al. (3). In direct comparisons, less of our extract is required (3-4 ul vs. 8-9 ul) for a significantly stronger maximal signal.

5. In vitro transcription 5.1 Overview of the reaction The in vitro transcription reaction is a several step procedure that requires approximately one full day to complete. Nuclear extract and promoter DNA

Figure 1. In vitro transcription was performed as described in the text on the adenovirus major-late promoter (pML) and the NF-H promoter (pH) using increasing amounts of a brain nuclear extract. Shown is an autoradiograph of cDNA products from the pML (+406 nt) and pH (+234 nt) promoters. The amount of nuclear extract added is shown along the top of the panel.

222

9: Neuronal promoter analysis are initially incubated for 30 min to form initiation complexes and then rNTPs added for a 30 min elongation reaction. The RNA products are precipitated, hybridized to 32P-labelled antisense oligonucleotides, and reverse transcription performed. cDNA products are then separated by electrophoresis on acrylamide/urea gels and the gels exposed to film. Protocol 3. In vitro transcription Reagents • 10 x TB: 400 mM Hepes pH 7.9, 10 mM . 10 x dNTPs: 5 mM each dATP, dCTP, dGTP, MgCI2, 10 mM DTT dTTP (Boehringer-Mannheim) . 5 x rNTPs: 2.5 mM each ATP, CTP, GTP, • 40 units/ul RNAsin (Prornega) UTP (Boehringer-Mannheim), 50 mM • 1 mg/ml actinomycin D in H20 (Sigma 1410) MgCI2, 2 units/ul RNAsin • 9.5 units/ul AMV reverse transcriptase • Stop buffer: 40 mM Hepes pH 7.9, 20 mM (Promega M510A) EDTA, 25 ug/ml tRNA, 1% SDS, 0.2 M • 5 nM adenovirus major-late promoter as NaOAc control (pML) . TE: 10 mM Tris pH 8.0, 0.1 mM EDTA • 50 nM test promoter (pTest) . 10 x HB: 100 mM Tris pH 8.0, 10 mM EDTA, • 2 x 105 c.p.m./ul 32P-labelled antisense 1.5 M KCI oligos (300-600 nM) . 10 x RT: 200 mM Tris pH 8.0, 100 mM MgCI2, 100 mM DTT

Method 1. Set up the transcription reaction by mixing 2.5 ul 10 x TB, 2.5 ul 5 nM pML, 2.5 ul 50 nM pTest, and H20 to 20 ul minus the volume of extract to be used. 2. Initiate the reaction by the addition of nuclear extract and incubate for 30 min at 30°C. 3. Add 5 x rNTPs and incubate for an additional 30 min at 30°C. 4. Quench the reaction by the addition of 175 ul of stop buffer (to 200 ul), extract with phenol:trichloromethane, and precipitate the RNA products with 500 ul 100% ethanol. 5. Wash the RNA pellet with 80% ethanol, dry briefly (1 min) at 95°C, resuspend in 12.5 ul TE, vortex, and place at 95°C for 1 min to dissolve. 6. Add 1.5 ul 10 x HB and 2 x 105 c.p.m./ul of oligos to 15 ul.

32

P-labelled antisense

7. Hybridize the RNA product and antisense oligos by heating at 95°C for 5 min, transferring to 64°C for 15 min, then cooling the reaction to 42°C. 8. Carry out reverse transcription by adding 4.5 ul 10 x RT, 4.5 ul 10 X dNTPs, 1.5 ul 1 mg/ml actinomycin D, 0.25 units of AMV reverse transcriptase, and H2O to 45 (xl. Incubate at 42°C for 20 min. 223

M. L. Schwartz and W. W. Schlaepfer Protocol 3. Continued 9. Add 5 ul of 2 M sodium acetate pH 7.0 to quench the reaction and precipitate the cDNA products with 125 ul 100% ethanol. 10. Wash the cDNA pellet in 70% ethanol, dry briefly (1 min) at 95°C, resuspend in 25 ul of formamide gel dye at 95°C, and separate the cDNA products on a 7.5% acrylamide/8 M urea gel. 11. Dry gels and expose to X-ray film.

5.2 Typical results of in vitro transcription Figure 1 shows the results of a titration for a brain nuclear extract using from 1 to 6 ul in an in vitro transcription assay. Maximal activity of both the adenovirus major-late promoter (pML) and the NF-H (pH) promoters occurs between 3 and 5 ul. Below 3 ul there is no activity. Above 5 ul activity begins to drop. These data show the narrow range of optimal activity for the extract. This gel was exposed to film overnight without an intensifying screen. Figure 2 shows how in vitro transcription using brain extract can be used to define areas of a neuronal promoter (NF-H) that are important in basal (pH 28-65) and enhanced (pH 85-1314) promoter function. The cDNA products from each promoter are clear and easily quantifiable apart from any potential aberrant bands. Again, this signal was obtained by exposing the gel to film overnight without an intensifying screen. Similar constructs transfected into neuronal cells yielded aberrant transcription initiation sites and showed no enhanced promoter activity between 65 and 115 (7). Even the TATA-less construct (pH 20) produced significant product in CAT assays (7). Any attempts to define positive or negative areas using transgenic mice would be far too time consuming and prohibitively expensive. In addition, the short constructs are not likely to be functional due to the effects of chromatin structure overriding any effects of basic transcriptional factors (10).

5.3 Notes on in vitro transcription The in vitro transcription is a fairly straightforward assay to perform once usable extract has been obtained. We have empirically determined optimum times, temperatures, and amounts of promoter DNA using previously published protocols as a starting point (3-5). As a rule, we mix the buffer and the Figure 2. In vitro transcription using either brain or liver nuclear extract was performed on plasmids containing 5' deletions of the NF-H proximal promoter and the adenovirus major-late promoter and assayed by a reverse transcriptase reaction using32P-labelled antisense oligos to NF-H (to +234) and ML (to +406) as described in Protocol 3. (A) An autoradiograph of cDNA products from a reverse transcriptase assay. The products corresponding to adenovirus major-late (ML) (406 nt) and NF-H (234 nt) cDNAs are marked

224

on the right. Deletion end-points are marked along the top of the panel. (B) The amount of radioactivity in each product band was quantified by a Molecular Dynamics phosphoimaging system. The ratio of NF-H product (pH) 1o adenovirus major-late product (pML) in either brain or liver extracts was calculated using the average of at least two experiments and displayed graphically. (C) The transcription data and standard deviations. Figure reprinted with permission from Schwartz et al. (8).

225

M. L. Schwartz and W. W. Schlaepfer DNAs together first and then initiate the first incubation by adding extract, thus, both the test and control promoters are exposed to the extract for the same amount of time. For experiments where two extracts are mixed, the two extracts are mixed separately in a tube and then added to the buffer and promoter DNAs. We describe the amount of plasmid in the transcription reaction in terms of nanomolar rather than nanogram quantities. In this way, each reaction has the same number of moles of promoter, thus allowing for more direct analysis of autoradiograms and quantification of radipactivity. The addition of promoterless vector to maintain constant total microgram amounts of DNA between reactions containing large and small constructs has little effect on the results and is therefore not added. In all tested cases, the overall level of transcription may have been affected by the addition of promoter-less vector but the ratio of control to test promoter products was unaffected. The use of 5 nM test promoter is not fixed and can range from 2.5 to 10 nM. It should be noted, however, that in vitro transcription requires a certain minimum amount of DNA (in ug) to function. This may be related to the presence of DNAases in the extract (see below). In contrast, too much DNA can lower the overall level of transcription. This is probably due to the titration of transcription factors by non-specific DNA sequences. Somewhat surprisingly, we have not observed any RNAase contamination problems with our extracts. Transcription assays performed with and without RNAsin (Promega) are indistinguishable. Generally, we add RNAase inhibitors as a precaution. Interestingly, the extracts are contaminated with low levels of DNAases. However, their effects seem to be minimal over the time course in which the incubation between extract and DNA is performed as long as the concentration of MgCl2 in the incubation buffer is not well above 1 mM.

5.4 Promoter templates It is extremely important to use supercoiled or at least circular plasmid DNA in the in vitro transcription reaction. In fact, we purify DNA originally isolated from a Wizard Maxi-prep (Promega) through a caesium chloride gradient. Our reasons are three-fold. First, the Wizard kit leaves residual chemicals that interfere with the reaction (ethanol precipitation can remove these impurities). Second, in some cases a significant amount of chromosomal DNA and RNA can contaminate the plasmid preparation making concentration determination difficult and inaccurate. Third, we have found that linear DNAs make poor templates since it appears that proteins can bind to free DNA ends and affect the transcription complex positively or negatively (unpublished observations). Our initial studies used DNA fragments that had increasing amounts of upstream region 'deleted' from the proximal promoter regions (TATA box) by digestion with a unique restriction enzyme. The results suggested that the 226

9: Neuronal promoter analysis promoter increased in strength as the upstream regions were deleted (i.e. as the restriction cut came closer to the TATA box). In contrast, supercoiled templates containing the same deletions showed similar activity profiles to those observed in transient transfections (although certain neurone-specific enhancements (8) were observed). We also noticed that the addition of linearized promoter-less vector DNA significantly diminished transcription from a supercoiled template while the addition of an equal amount of the supercoiled promoter-less vector had no effect.

6. Concluding remarks We have developed a method for making nuclear extracts from rat brain which allows us to conduct in vitro transcription experiments on the neuronespecific NF gene promoters. Using this methodology we have been able to define positive and negative cis-acting regions of the promoters and to characterize sequences required for brain-specific high level transcription (8) that were not observed by transient transfection into neuronal cell lines (7). We have also been able to determine the strength of the minimal NF promoters that either produced aberrant start sites in transient transfection assays (7) or did not function in transgenic mice (10). In addition, these same extracts were used to define brain-specific bands on gel shift assays using wild type and mutant templates that had been functionally characterized (8). These data can now be used with assays that evaluate the effects of higher order DNA structures to gain further understanding into the complexities underlying neuronal-specific gene expression. Hence, improvements in nuclear 'extractology' to obtain good quality extracts from brain provide a critical advance towards characterizing the strength and function of neurone-specific promoters.

References 1. Weil, P. A., Luse, D. S., Segall, J., and Roeder, R. G. (1979). Cell 18, 469. 2. Manley, J. L., Fire, A., Cano, A., Sharp, P. A., and Gefter, M. L. (1980). Proc. Natl Acad. Sci. USA 77, 3855. 3. Dignam, J. D., Lebovitz, R. M., and Roeder, R. G. (1983). Nucleic Acids Res. 11, 1475. 4. Gorski, K., Carneiro, M., and Schibler, U. (1986). Cell 47, 767. 5. Shapiro, D. J., Sharp, P. A., Whali, W. W., and Keller, M. J. (1988). DNA 7, 47. 6. Kamakaka, R. T., Tyree, C. M., and Kandonaga, J. T. (1991). Proc. Natl Acad. Sci. USA 88, 1024. 7. Shneidman, P. S., Bruce, J., Schwartz, M. L., and Schlaepfer, W. W. (1992). Mol. Brain Res. 13, 127. 227

M. L. Schwartz and W. W. Schlaepfer 8. Schwartz, M. L., Katagi, C., Bruce, J., and Schlaepfer, W. W. (1994). J. Biol. Chem. 269, 13444. 9. Schwartz, M. L., Hua, Y., and Schlaepfer, W. W. (1997). Mol Brain Res. 48, 305. 10. Beaudet, L., Charron, G., Houle, D., Tretjakoff, L, Peterson, A., and Julien, J. P. (1992). Gene 116, 205. 11. Barton, M. C., Madani, N., and Emerson, B. M. (1993). Genes Dev. 7, 1796. 12. Schwartz, M. L., Hua, Y., Canete-Soler, R., and Schlaepfer, W. W. (1998). Mol. Brain Res. 57, 21.

228

10 Preparation of chromatin templates for transcription studies JOAN BOYES

1. Introduction Transcription in the eukaryotic nucleus takes place within a highly packaged chromatin environment. Both genetic and biochemical approaches have been used to investigate how this process differs from transcription of protein-free DNA in vitro. This chapter describes various methods by which chromatin templates are generated for biochemical studies and discusses the merits of these different templates for particular studies. The preparation of histones and mononucleosome templates is described in full experimental detail. Moreover, the use of mononucleosome templates as a model to investigate nucleosome disruption over promoters is discussed.

2. Experimental approaches using chromatin templates 2.1 Introduction to chromatin structure DNA in the nucleus of a eukaryotic cell exists in a highly packaged chromatin structure (reviewed in 1). At the first level of this packaging, 146 bp of DNA are wrapped 1.75 times around the histone octamer which consists of two molecules each of the four core histones H2A, H2B, H3, and H4. This complex is defined as the nucleosome core particle. Depending on the cell type, nucleosome core particles are spaced 160-260 bp apart to give the 'beads-ona-string' structure or 10 nm chromatin fibre. This structure is condensed further by binding the linker histone, H1, to DNA at the entry and exit points from the nucleosome core particle. Histone H1 shields the charges along the DNA backbone to allow compaction to the 30 nm fibre. The exact structure of the 30 nm fibre is unknown but the most favoured model is a solenoid structure with six nucleosome core particles per helical turn. The 30 nm fibre is then compacted over 250-fold more to generate the final condensed chromatin structure of metaphase chromosomes. The mechanism of these higher

Joan Boyes levels of chromatin compaction is unclear, although some evidence exists that the 30 m fibre is folded into looped domains.

2.2 Initiation and elongation of transcription on chromatin templates Although our knowledge of chromatin structure remains incomplete, experiments to address chromatin function can be designed. Since all nuclear processes such as transcription, replication, recombination, and repair take place in a chromatin environment, a fundamental question is to understand how the proteins involved in these processes contend with the chromatin packaging. Numerous studies have shown that a static, packaged chromatin structure is incompatible with gene transcription. Genetic studies in yeast revealed that a reduction in cellular histone levels caused increased transcription of certain genes (reviewed in 2). Moreover, the promoters and enhancers of transcriptionally active genes were shown to be associated with nuclease hypersensitive sites. These regions have a disputed nucleosome structure and are bound by a variety of transcription factors (reviewed in 3). Therefore, a key question in understanding the initiation of transcription is to determine how nucleosomes are disrupted to allow access to the transcription machinery. Recently, there has been significant progress in this area. The ability of various transcription factors to bind to their cognate site on a nucleosome has been determined (reviewed in 4) and the roles of different nucleosome disrupters and histone acetyltransferases in assisting this process are being investigated (reviewed in 5). Chromatin disruption is also required during transcriptional elongation. Electron micrographs have indicated that the highly packaged chromatin structure is incompatible with the passage of the transcribing polymerase (6). A further key question therefore is to determine the mechanism of transcription elongation on chromatin templates. Current technology permits quite detailed biochemical analysis of transcription initiation and elongation at the nucleosome level of chromatin packaging. Both mononucleosome templates and arrays of nucleosomes can be prepared. The advantages of these two different types of chromatin templates are outlined below, followed by examples of how they have been used to investigate transcriptional initiation and elongation.

2.3 Mononucleosome templates versus nucleosome arrays Mononucleosome templates have the advantage that a highly purified, distinct population can be prepared. This means that the exact position of the histone octamer with respect to the DNA sequence of interest can be defined and the mechanism of nucleosome disruption during the different transcription processes can be determined to quite high resolution. Mononucleosome templates can be used, in addition, to begin to define the influence of additional 230

10: Preparation of chromatin templates levels of complexity on the transcription process. For example, proteins which bind to nucleosomes such as histone H1 and HMG 14/17 can be added to mononucleosomes and their ability to promote or impede transcription of these templates can be determined. However, in vivo, transcription does not take place on mononucleosomes and a further level of understanding of the transcription process requires the use of nucleosome arrays. Although these templates do not provide the same level of detail of the process, they do allow different questions to be addressed. Most significantly, transcription of the chromatinized templated can be analysed by linking a reporter gene to the promoter of interest. Changes in chromatin structure which allow productive initiation of transcription as well as full elongation of transcription can be thus characterized. 2.3.1 Analysis of transcription initiation To investigate transcription factor access to promoters, mononucleosomes have proved to be a particularly valuable model. These studies can determine not only whether a particular transcription factor can bind to its site on a nucleosome, but also determine the influence of rotational and translational positioning of the factor binding site on the ability of the factor to bind to the nucleosome (7, 8). Moreover, various mechanisms to assist transcription factor binding to nucleosome templates can be investigated. These include modification of the histones, e.g. by histone acetylation (9), and the use of auxiliary factors such as the nucleosome assembly factor, nucleoplasmin, and nucleosome disruption complexes such as SWI/SNF, NURF (nucleosome remodelling factor), and RSC (remodel the structure of chromatin) (reviewed in 5, 10). Studies with mononucleosome templates can be complemented by experiments with nucleosome arrays. In this case, chromatin disruption is analysed by nuclease-hypersensitive site formation or loss of a regular nucleosome ladder. Generally, similar results as with mononucleosomes were obtained (but see 10). Nucleosome arrays permitted additional control experiments to investigate whether nucleosome disruption was specific to regulatory regions. Moreover, the use of arrays uncovered factors which promote chromatin accessibility via nucleosome mobilization (reviewed in 5); this would not have been possible using only mononucleosomes. 2.3.2 Analysis of transcription elongation Mononucleosome templates have also proved to be an invaluable tool to analyse transcription elongation through a nucleosome. In this case, careful analysis of the mononucleosome positions before, during, and after transcription has elucidated a detailed mechanism of the process. Initially, using phage polymerase, the histone octamer was shown to transfer from in front to behind the transcribing polymerase (11). Further detailed analysis with uniquely positioned nucleosomes mapped the point of transfer of the histone octamer 231

Joan Boyes to within 10 bp (12, 13). This mechanism has recently been confirmed using yeast RNA polymerase III (14). Studies using nucleosome arrays have uncovered additional factors involved in transcriptional elongation of chromatin templates: A heterodimeric complex, FACT (facilitates chromatin transcription) has been purified as a specific complex which facilitates elongation of RNA polymerase II through nucleosome templates (15). In addition, nucleosome arrays were found to induce polymerase pausing in vivo similar to that observed in vivo on the human hsp 70 promoter. The transactivation domain of human heat shock factor, in combination with the nucleosome disruption complex SWI/SNF, could overcome this nucleosome-mediated pausing (16). The precise mechanism by which these components assist polymerase passage through nucleosomes remains to be determined.

3. Preparation of chromatin templates 3.1 Technical considerations Both mononucleosome templates and nucleosome arrays can be prepared with relative ease in vitro. However, an important consideration is that if histones are simply mixed with DNA at physiological salt concentrations, aggregation occurs. To prevent this aggregation, and to form a proper nucleosome structure, three broad methods can be used: (a) association of histones with an assembly factor (b) mixing histones with DNA at high salt concentrations (c) exchange of histones from chromatin at high salt concentrations An advantage of chromatin assembly using assembly factors at physiological salt concentrations is that 'order of addition' experiments could be performed whereby components of the transcription machinery are added either prior to, during, or following chromatin assembly. A disadvantage is that the assembly factors will still be present during the transcription analysis. This was particularly important previously since the most efficient chromatin assembly reactions at physiological salt conditions relied on crude nuclear extracts. Although this problem could be overcome, to some extent, by sucrose gradient purification of the chromatin templates, the presence of contaminating proteins could not be completely eliminated. As described below, a major recent advance is the purification of assembly factors from Drosophila embryo extracts. The purified assembly system will also overcome a second major problem for the assembly of nucleosome arrays in vitro: the ability to achieve correct spacing of the nucleosomes. This had relied on crude nuclear extracts and suffered from the same problems as described above. It would be anticipated that, in the future, the cloned Drosophila assembly system will provide the 232

10: Preparation of chromatin templates ideal system for the preparation of purified, regularly spaced nucleosome arrays. Therefore methods to prepare and assemble chromatin with Drosophlla extracts will not be described in detail here. Currently, the best method for the preparation of purified chomatin templates relies on mixing histones with DNA at high salt. This method is also likely to continue to be the method of choice for the assembly of mononucleosome templates. Although this method does not allow 'order of addition' experiments to be performed and assembles close-packed nucleosomes, it has the advantage that the reagents are highly purified. Moreover, no specialized reagents are required other than purified histones. This technique is described in full detail below. For comparison, the basis and merits of other chromatin reconstitution methods are described.

3.2 Xenopus nuclear extracts The earliest methods developed to assemble chromatin under physiological conditions relied on Xenopus oocyte or egg extracts (17-19). These extracts are rich in the acidic proteins, nucleoplasmin and N1/N2, which bind histones and permit ordered nucleosome assembly on to added DNA. The extracts are prepared by high speed centrifugation of lysed Xenopus oocytes or eggs. The crude extracts contain both histones and assembly factors. In the presence of ATP and magnesium, a regularly spaced nucleosome array is assembled. Components of chromatin assembly could be partially separated from the crude extract by preparation of a heat-treated supernatant. This supernatant is highly enriched in the thermostable histone binding protein, nucleoplasmin, and is capable of nucleosome assembly. In this case, however, exogenous histones must be added and the nucleosome arrays are no longer regularly spaced. This method can be used for the assembly of both mononucleosome templates and nucleosome arrays. A drawback is that the templates produced are not necessarily homogeneously assembled. This was a particular problem in some early transcription studies where the efficiency of in vitro transcription extracts was very low (approximately 1% of templates). Thus, these studies were subject to the criticism that the low level of transcription of the chromatinized templates was due to incompletely assembled arrays and/or initiation of transcription in linker regions between nucleosomes.

3.3 Drosophila embryo nuclear extracts Significant advances in achieving more purified, correctly spaced chromatin have come from the use of Drosophila embryo extracts. Originally, a high speed extract (S-150) was developed from pre-blastoderm embryos (20). This was shown to be capable of assembly of regularly spaced nucleosome arrays when supplemented only with ATP, magnesium, and an ATP-regenerating system. A slight modification of this extract was developed in which a high 233

Joan Boyes speed supernatant was prepared from 0-6 h Drosophila embryo extracts (S-190; 21). These extracts reconstituted a regularly spaced nucleosome array (190-220 bp) very efficiently (10 min compared to 1-2 h with Xenopus extracts) upon addition of ATP and exogenous histones (22). This latter S-190 extract has been extensively fractionated. The components required for assembly of a correctly spaced nucleosome array have recently been narrowed down to an ATP-utilizing chromatin assembly activity (ACF), a histone chaperone (NAP-1 or a nucleoplasmin-like protein), purified histones, and ATP (23, 24). ACF has been purified and found to consist of four subunits (23). When cloned, these reagents will undoubtedly be the system of choice for correct assembly of nucleosome arrays for transcriptional analysis. A further advance relating to the transcription of chromatin templates using the Drosophila embryo system was the development of a very efficient in vitro transcription extract. Using this extract, up to 25% of the templates could be transcribed (25). This overcame previous criticisms that transcription was not representative of most of the molecules in the population.

3.4 Minichromosome preparation Nucleosome arrays with properties very close to those within the cell can be prepared by isolating minichromosomes from transfected cells. Minichromosomes can be prepared either from yeast or from mammalian cell lines by transfection with a vector carrying an origin of DNA replication. Replication within eukaryotic cells results in chromatinization of the vector. Recovery of the chromatinized vectors from cell lines (26) is via gentle lysis of the cell membrane and low salt (120 mM) extraction of the nuclei to maintain the minichromosomes as intact as possible. Cellular debris is removed by sucrose gradient purification to yield a relatively purified population of minichromosomes. Isolation of minichromosomes from yeast is slightly more complex since spheroplasts must first be formed, followed by isolation of nuclei. Minichromosomes are then gently extracted from the nuclei and are subject to a number of centrifugation steps to remove nucleases and proteases (27). This method has the advantage that the minichromosomes have correct nucleosome spacing and, in theory, should retain many of the properties of nucleosome arrays in vivo. The main disadvantage is the low yield (from cell lines) and the presence of nucleases and proteases (in yeast). In addition, the minichromosome preparation is not necessarily a completely homogeneous population and the minor contaminants present in the chromatin preparation are unknown.

3.5 Mononucleosome preparation 3.5.1 Salt-urea dialysis This procedure is the method of choice for preparation of medium to large scale amounts of chromatin templates (microgram to milligram quantities). It is described in full detail later in this chapter. In this technique, DNA is mixed 234

10: Preparation of chromatin templates with histone under conditions of high salt and urea where the histones will not precipitate the added DNA. The salt and urea concentrations are gradually lowered by stepwise dialysis to generate a nucleosome reconstituted on to the added DNA (28). The advantage is that one is working with purified reagents. In addition, the correctly reconstituted nucleosome can be purified from inappropriately reconstituted material. The main disadvantages are that the high salt conditions preclude 'order of addition' experiments and physiological spacing of nucleosome arrays is not achieved. 3.5.2 Histone exchange In this procedure, the DNA fragment for reconstitution is mixed with a large molar excess of nucleosome core particles, or of Hl-depleted longer chromatin in the presence of 1 M salt. The DNA fragment is radiolabelled so it can be distinguished. Under the high salt conditions, the histone octamer becomes detached from the DNA of the donor nucleosomes and is free to exchange to the radiolabelled fragment. The salt concentration is then lowered gradually to 0.1 M salt, by slowly diluting the salt solution (29). This method is efficient and can yield almost 100% reconstitution of small (microgram to nanogram) quantities of DNA. A minor drawback is that the donor nucleosomes or chromatin are present during subsequent analyses of the radiolabelled reconstitute. Moreover, the use of non-physiological salt conditions precludes 'order of addition' experiments, and the generation of correctly spaced nucleosome arrays as discussed above. 3.5.3 Polyglutamic acid To circumvent the problems associated with chromatin assembly at nonphysiological salt, various histone chaperones can be used. Polyglutamic acid is such an acidic chaperone. It is pre-bound to histones under high salt conditions. The salt concentration is gradually lowered to 0.1 M by dialysis and the histone/polyglutamic acid complex is then added to DNA under physiological salt conditions (0.1 M). Relatively efficient chromatin assembly has been reported by this technique (30). However, correct nucleosome spacing has not been achieved. Moreover, since polyglutamic acid can act as a histone sink, its presence could interfere with subsequent transcription analyses of the chromatin template. 3.5.4 Nucleoplasmin A second chaperone-based assembly method utilizes nucleoplasmin, the physiological histone chaperone from Xenopus eggs. Nucleoplasmin is purified from unfertilized Xenopus eggs on the basis of its thermostable properties (31). Reconstitution is achieved at physiological salt concentration by mixing histones with the nucleoplasmin and then subsequently adding the DNA. This method is, unfortunately, relatively inefficient (29) and, moreover, does not result in correctly spaced nucleosome arrays (an average of 155 bp). It has now been largely superseded by the chromatin assembly system from Droso235

Joan Boyes

phila embryo extracts where histone chaperones like nucleoplasmin work in conjunction with energy dependent nucleosome assembly factors (above). The salt-urea reconstitution method, described in detail later, uses purified histones and DNA. The purification of these reagents is described. In addition, the use of these templates to analyse the binding of transcription factors is discussed.

4. Preparation of histones 4.1 Principle of hydroxyapatite purification In theory, histones can be prepared from any nucleated cell. Chicken erythrocytes have the advantages that they have low levels of endogenous nucleases and are readily available. Various methods to purify histones have been described including acid extraction, ultracentrifugation, and salt elution from DNA bound to a hydroxyapatite column (32). The latter is preferred since highly purified histone subunits (free of linker histones, H1 and H5) can be obtained. This procedure is based on that initially described by Simon and Felsenfeld (33). Nuclei are initially prepared and then the chromatin is cleaved into fragments of roughly 2-3 kb. Cleavage of the chromatin can be achieved by either mechanical shearing, sonication, or by limited digestion with micrococcal nuclease. Following release of the chromatin from nuclei, the DNA is bound to a hydroxyapatite column. Histone H1 and other proteins which are bound relatively weakly to the chromatin are released by salt washes. A higher salt concentration is then used to release the lysine-rich histones H2A/H2B, which are bound less strongly within the nucleosome than the arginine-rich histones H3/H4, which are released by very high salt. Between the elution of histones H2A/H2B and histones H3/4, a salt gradient is used to prevent cross-contamination of histone peaks. Some protocols use a single salt elution step to release all four core histones. However, histones H2A/ H2B are slightly depleted in this procedure. Since proper stoichiometry of histones is essential for efficient reconstitution, the two-step elution protocol described here is preferred. Protocol 1. Preparation of chromatin for histone isolation Equipment and reagents • Chicken blood (can normally be obtained fresh from a slaughter house) mixed with an equal volume of 20 mM EDTA or 1000 U/ml heparin • Buffer A: 0.34 M sucrose, 15 mM Tris pH 7.5, 15 mM NaCI, 60 mM KCI, 15 mM b-mercaptoethanol, 0.5 mM spermidine, 0.15 mM spermine • Buffer A/NP40: as above but containing 0.5% NP40

• Buffer A/Ca: buffer A with 1 mM CaCI2 • Micrococcal nuclease (Worthington) dissolved at 10000 U/ml in 10 mM Hepes pH 7.5, 50 mM KCI, 50% glycerol, 1 mM CaCI2 • Dialysis buffer: 0.63 M NaCI, 0.1 M potassium phosphate pH 6.7, 1 mM EGTA • Dialysis tubing: Spectra/Por 7, Spectrum; molecular weight cut off 8000 Da, 24 mm flat width

236

10: Preparation of chromatin templates Method 1. Place 100 ml of chicken blood into four 50 ml plastic tubes and centrifuge at 600 x g for 5 min at 4°C. 2. Pipette off the supernatant and remove as much of the buffy coat (white cell layer above the erythrocyte pellet) as possible. 3. Resuspend the cells in PBS (without calcium and magnesium), centrifuge as in step 1, and remove the supernatant as in step 2. 4. Repeat step 3 two more times. 5. Resuspend the cell pellets each in 50 ml of buffer A/NP40 and invert the tubes 2-3 times. It is important to use a large volume of this buffer to get complete lysis of the erythrocyte membrane. 6. Transfer the suspension to suitable plastic tubes and centrifuge at 4100 X g for 5 min at 4°C. 7. Remove the red supernatant, resuspend the nuclear pellet in a total volume of 150 ml buffer A, and centrifuge as in step 6. 8. Repeat step 7. 9. Remove the supernatant, resuspend the nuclear pellet in 150 ml buffer A/Ca and centrifuge as in step 6. 10. Remove the supernatant and resuspend the nuclei in 25 ml buffer A/Ca. 11. Measure the concentration of the nuclei: dilute 2 ul of the resuspended nuclei into 1 ml 1 M NaOH; measure the A260 against a blank of 1 M NaOH. 12. Put the nuclei into a small conical flask. Adjust the concentration of nuclei to A260 =100 with buffer A/Ca. Store on ice. 13. Empirically determine the amount of micrococcal nuclease to digest the chromatin by taking five 1 ml aliquots of the nuclei. Add 0.5, 1, 1.5, 2, 4 ul of micrococcal nuclease. Incubate at 37°C for 20 min. 14. Stop the reaction by adding EDTA to a final concentration of 10 mM. 15. Centrifuge the nuclei at 4100 X g for 5 min at 4°C. Remove as much of the supernatant as possible. 16. Lyse the nuclei in 300 ul 0.2 mM EDTA/0.25 mM PMSF. Centrifuge at 10550 x g for 10 min 4°C. 17. Remove the supernatant and measure the A260 of the released chromatin by diluting 2 ul of chromatin in 1 ml H2O and use H2O as a blank. Calculate the concentration of released chromatin. The optimal concentration of micrococcal nuclease is that where 30-60% of the chromatin is released (compared to the amount in step 11). Calculate the amount of micrococcal nuclease needed for digestion of the bulk nuclei from the volume in step 12.

237

Joan Boyes Protocol 1. Continued 18. Check the size of the released chromatin on a 0.8% agarose gel. The optimal average size range is 2-3 kb. 19. Place the conical flask containing the nuclei (step 12) at 37°C for 5 minutes. Add the amount of micrococcal nuclease calculated in step 17 and incubate at 37°C for 20 min. 20. Stop the reaction by adding 0.5 M EDTA to a final concentration of 10 mM and repeat steps 15-17. 21. Place the released chromatin into washed dialysis tubing. Dialyse against a 50-fold greater volume of dialysis buffer. The chromatin will initially come out of solution, but then is resolubilized. Dialyse for one hour after the chromatin is resolubilized. Change the dialysis buffer; dialyse one more hour.

The hydroxyapatite column chromatography is performed using a short, broad column. This prevents any aggregated chromatin from blocking the column and permits a reasonable flow rate. In addition, since the histones are eluted from the hydroxyapatite column in distinct steps, a longer column is not required for careful separation of closely related species. Hydroxyapatite (BioGel HTP, Bio-Rad) has an extremely high capacity (> 500 mg DNA per dry gram) and it is very important not to underload the column; otherwise considerable losses of material are incurred. The optimal chromatin to hydroxyapatite ratio can be determined empirically by mixing aliquots of chromatin (0.5-1 mg) with increasing amounts of hydroxyapatite. By measuring the A260 of the supernatant, it can be determined when most of the chromatin is bound by the hydroxyapatite. The dried hydroxyapatite is resuspended in 0.63 M column buffer (Protocol 2). Following thorough mixing, the hydroxyapatite is allowed to settle and the fines are removed according to the manufacturer's instructions. The hydroxyapatite is then resuspended in approximately twice its packed volume of column buffer. The column is packed, equilibrated, and transferred to 4°C. Protocol 2. Hydroxyapatite chromatography Equipment and reagents • Hydroxyapatite (BioGel HTP, Bio-Rad) . 1.2 M column buffer: 1.2 M NaCI, 0.1 M resuspended in 0.63 M column buffer and potassium phosphate pH 6.7, 0.5 mM PMSF packed into a short, broad column • 2 M column buffer: 2 M NaCI, 0.1 M potassium phosphate pH 6.7, 0.5 mM PMSF • 0.63 M column buffer: 0.63 M NaCI, 0.1 M potassium phosphate pH 6.7, 0.5 mM PMSF . 0.93 M column buffer: 0.93 M NaCI, 0.1 M potassium phosphate pH 6.7, 0.5 mM PMSF

238

10: Preparation of chromatin templates Method 1. Centrifuge the dialysed chromatin (Protocol 1, step 21) at 16 500 x g for 5 minutes at 4°C to remove any aggregated chromatin. 2. Load the supernatant on to the hydroxyapatite column. A low flow rate is used to ensure efficient binding and to minimize clogging of the column. 3. Wash the column with three column volumes of 0.63 M column buffer. Collect fractions which are 1/10 of the column volume and measure the A230. 4. Change the column buffer to the 0.93 M buffer. Collect fractions and measure the A230 as above. The peak of absorbance from this column should be histones H2A/H2B. Continue to wash the column with this buffer until the A230 has returned to the baseline. Take aliquots of 4 ug from each of the peak fractions and TCA precipitate (Protocol 3). Freeze the fractions with peak A230 at -80°C. 5. Place 2.5 column volumes of 0.93 M column buffer in the near side of a gradient maker and 2.5 column volumes of 1.2 M column buffer in the other side. Prepare a linear gradient by gradually mixing the 1.2 M buffer into the 0.93 M buffer and apply buffer from the 0.93 M chamber to the column as the buffers are mixing. 6. Collect fractions and read the A230 as before. 7. Change the buffer to 2 M column buffer and pass a total of 2-3 column volumes over the column. Continue to collect fractions and measure the A230. The peak of the absorbance should correspond to histones H3/4. Take aliquots of 4 ug from each of the peak fractions and TCA precipitate (Protocol 3). Freeze the peak fractions at -80°C. 8. Change the column buffer to 0.5 M potassium phosphate pH 6.7 and elute DNA from the column using 2-3 column volumes of this buffer.

Integrity of the prepared histones is checked using polyacrylamide gel electrophoresis of the TCA-precipitated samples. Although chicken erythrocytes have relatively low levels of proteases, some degradation of histones can occasionally occur. The best measure to prevent this is to increase the speed of all steps prior to loading the column. The most susceptible to degradation is histone H1 (which elutes during the 0.63 M wash). Histone H3 is the next most susceptible whereas the relative rate of degradation of the remaining histones depends on the source of the chromatin (34). All four histones are of similar low molecular weights (e.g. calf thymus histones, H3: 15.27 kDa, H2A: 13.96 kDa, H2B: 13.77 kDa, and H4: 11.24 kDa). Efficient separation of these proteins is achieved using an 18% polyacrylamide gel with a very high acrylamide to bis ratio (360:1). 239

Joan Boyes Protocol 3. TCA precipitation of histones Equipment and reagents • 50% Trichloracetic acid (w/v, TCA)

• Protein loading buffer: 10% glycerol, 5% water-saturated bromophenol blue solution, 50 mM Tris pH 6.8, 142 mM b-mercaptoethanol, 1% SDS

Method 1. Four micrograms of each pair of histones are TCA precipitated for electrophoresis. The extinction coefficient for histones at 230 nm is 4.2. Calculate the volume of the histone fractions for TCA precipitation from the equation: volume/ul = 2. In a 1.5 ml plastic tube, add an equal volume of 50% TCA. Place on ice for at least 30 min. 3. Centrifuge at 12600 X gat 4°C for 5 min. 4. Remove the supernatant carefully and wash the pellet thoroughly with 500 ul acetone/HCI with vortexing. 5. Centrifuge as in step 3. 6. Remove the supernatant carefully and wash the pellet with 750 ul acetone with vortexing. 7. Centrifuge as in step 3. 8. Remove the supernatant carefully and air dry the pellet. 9. Resuspend the pellet in 20 ul protein loading buffer.

Protocol 4.

Electrophoresis of histones

Equipment and reagents • Tris/glycine running buffer: 0.05 M Tris, 0.373 M glycine, 0.1% SDS • Coomassie staining solution: 50% methanol, 10% acetic acid, 0.05% w/v coomassie blue

• Coomassie destaining solution: 50% methanol, 10% acetic acid

Method 1. Sandwich two 0.5 mm spacers along the long edges of two 20 cm x 20 cm glass plates and tape the sides and bottom for polyacrylamide electrophoresis. 2. Prepare separating gel mixture (25 ml for a 20 x 20 cm gel) with final concentrations (18% arcylamide, 0.05% bisacrylamide, 0.75 M Tris pH 8.8, 0.1% SDS, 0.1% ammonium persulfate). 240

10: Preparation of chromatin templates 3. Add TEMED to a final concentration of 0.02% and pour the gel, leaving a space of at least 2.5 cm from the top of the smaller glass plate. Stand the gel upright. Allow to polymerize for 20 min. 4. Prepare the stacking gel mixture (5 ml for a 20 x 20 cm gel) with final concentrations (3% acrylamide, 0.15% bisacrylamide, 0.12 M Tris pH 6.8, 0.1% SDS, 0.1% ammonium persulfate). 5. Add TEMED to a final concentration of 0.05%, pour the stacking gel on to the polymerized separating gel and insert the comb. Allow to polymerize for 20 min. 6. Assemble the gel in the electrophoresis apparatus and add Tris/ glycine running buffer to the chambers. 7. Boil the histone samples (from Protocol 3, step 9) for at least 1 min. 8. Load the gel using a duck-billed pipette tip and electrophorese at 1.25 mA/cm until the bromophenol blue dye has run off the bottom of the gel. 9. Disassemble the gel apparatus and incubate with coomassie stain for 20 min at room temperature with shaking. 10. Destain the gel by shaking in destaining solution for 20 min to visualize the histone bands. Change the buffer and continue to destain until the gel background is no longer blue.

The histones recovered from the hydroxyapatite column are considerably more dilute than is useful for nucleosome reconstitution. Therefore, once the integrity of the histones has been established, they are concentrated. The highly charged histones are bound to a cation exchange matrix and are eluted using high salt. By using a small column with a high binding capacity, significant concentration of the histones can be achieved. Protocol 5. Concentration of purified histones Equipment and reagents • Column, of approximate dimensions 1 cm . 0.2 M NaCI/HE: 0.2 M NaCI, 10 mM Hepes in diameter and 5 cm high pH 8.0, 1 mM EDTA pH 8.0 • CM-52 cation exchange matrix (Whatman) . HE: 10 mM Hepes pH 8.0, 1 mM EDTA . 2 M NaCI/HE: 2 M NaCI, 10 mM Hepes pH pH 8.0 8.0, 1 mM EDTA pH 8.0

Method 1. Weigh 3 g of pre-swollen CM-52 and resuspend in 30 ml 2 M NaCI/HE. Mix well. Allow to settle and decant the fines. 2. Add a further 20 ml 2 M NaCI/HE. Transfer to a 25 ml measuring cylinder. Allow to settle and decant the fines.

241

Joan Boyes Protocol 5.

Continued

3. Add 2 ml of 2 M NaCI/HE and use the slurry to pack a small column. 4. Wash the column thoroughly with 0.2 M NaCI/HE. 5. Thaw the histone H2A/B-containing fractions which are in 0.93 M NaCI/0.1 M potassium phosphate (Protocol 2, step 4). Dilute with HE so that the final concentration of salt is 0.2 M. Mix well and pump on to CM-52 column. 6. Wash the column with 5 column volumes 0.2 M NaCI/HE. 7. Elute histones H2A/B from the column with 3 column volumes 2 M NaCI/HE. Collect fractions which are 1/5 column volume. 8. Measure the A230 of these fractions. 9. Pool the peak fractions and measure the concentration of the histones in triplicate: dilute three aliquots 10-fold in HE and measure the A230 against a blank of 0.2 M NaCI/HE. 10. Calculate the concentration of histones using the extinction coefficient at A230 of 4.2. (NB for successful reconstitution of nucleosomes, it is important that the concentration of histones is calculated very carefully and that the triplicate A230 readings agree closely. If they do not, mix the pooled histones again, take care to wipe excess histones from the pipette tip on to the side of the Eppendorf tube, and repeat the A230 readings.) 11. Re-equilibrate the column with 0.2 M NaCI/HE by washing with at least 5 column volumes of this buffer. 12. Repeat steps 5-9 with the H3/4 fractions, remembering that the H3/4 fractions are in 2M NaCI and must be diluted to 0.2 M NaCI/HE prior to loading on to the CM-52 column.

Successful nucleosome reconstitution relies on the histones being free from contaminating nucleases and proteases. The presence of these contaminants is assayed very stringently. If these contaminants are present, the preparation is discarded. (a) The presence of proteases is tested by diluting 4 ug of the histones into 400 ul sterile, deionized water and incubating at 37°C overnight. The histones are then TCA precipitated and electrophoresed as described in Protocols 3 and 4. (b) To assay for the presence of nucleases, 0.5 ug of histones are incubated with 5 ug of supercoiled plasmid DNA in the presence of 1 mM CaCl2 and 3 mM MgCl2 overnight at 37°C. One tenth of this mix is then loaded on a 1% agarose gel and assayed for the presence of nicks and degradation. 242

10: Preparation of chromatin templates

5. Reconstitution of mononucleosomes by salt—urea dialysis (35) 5.1 Considerations for the DNA fragment for reconstitution The advantage of reconstitution of mononucleosomes in vitro is that virtually any fragment of interest can be packaged into a nucleosome. The exact fragment chosen will obviously depend on the experimental question being asked. However, several points are worth bearing in mind. 5.1.1 Length of DNA fragment reconstituted As discussed above, the technique of salt-urea dialysis does not deposit nucleosomes with physiological nucleosome spacing. Therefore, to promote reconstitution of only one nucleosome on a fragment and not close-packed dinucleosomes, the fragment should not be longer than 250 bp and ideally should be much shorter, in the range 160-200 bp. The fragment of interest should be cloned between convenient restriction sites in a plasmid from which it can be easily exercised and labelled. 5.1.2 Position of the transcription factor binding site The structure of nucleosomal DNA is not the same at all positions on the nucleosome surface and nucleosome structure can profoundly affect the binding of transcription factors. Both the rotational positioning (whether the transcription factor binding site faces towards or away from the histone octamer; see ref. 8) and translational positioning (the position of the binding site along the nucleosomal DNA; see ref 7) can alter protein binding to the nucleosome. During reconstitution, the histone octamer can be positioned differently relative to the sequence of interest in various ways: (a) Some sequences naturally position the histone octamer, e.g. the Xenopus 5S gene (36). (b) The salt-urea dialysis reconstitution procedure frequently deposits nucleosomes on to the ends of the DNA fragments (37). (c) The nucleosome position can be manipulated using the positioning sequences developed by Crothers (38). These sequences bend the DNA and make nucleosome deposition more favoured at certain positions. The fact that nucleosome positions are not always easily predicted means that the nucleosome position should always be checked carefully following reconstitution (Protocol 10). 5.1.3 Relation of transcription factor binding site(s) DNA sequences carrying single and multiple transcription factor binding sites can be reconstituted into a nucleosome. These can be either from natural promoter regions or from artificial sequences. Since nucleosome deposition on 243

Joan Boyes most DNA fragments is fairly random (apart from the end preference), initial investigations of the ability of an individual transcription factor to bind to a nucleosome can be made by cloning the factor binding site of interest approximately 80-120 bp from the end of a 160-200 bp DNA fragment. This should ensure that the factor binding site is within the nucleosome-bound DNA in all or most of the nucleosomes in the population. 5.1.4 Technical considerations for preparation of DNA reconstitution The most important technical considerations concerning the DNA for salturea reconstitution are: (a) to use purified DNA which is free from contaminants, such as ethidium bromide, which intercalate into the DNA and disrupt histone/DNA contacts (b) to accurately determine the concentration of both histones and DNA: mixing these components at the wrong ratio can result in either no nucleosome reconstitution or in the formation of aggregates. Protocol 6. Preparation of the DNA fragment for nucleosome reconstitution Equipment and reagents • Qiagen Por . HE: 10mMTris pH 8.0, 1 mM EDTA « Loading buffer: 15% ficoll (type 400, Pharmacia), 10 mM Tris pH 8.0, 1 mM EDTA pH 8.0, 0.25% bromophenol blue, 0.25% xylene cyanol

• Spectra/Plasmid Mega Kit 7 dialysis tubing (Spectrum); molecular weight cut off 8000 Da, 24 mm flat width . TE: 10 mM Tris pH 8.0, 1 mM EDTA • TAE: 40 mM Tris-acetates, 1 mM EDTA

Method 1. Prepare a plasmid carrying the fragment of interest from a 1-2 litre bacterial culture according to the instructions of the Qiagen Plasmid Mega Kit.a 2. Digest the plasmid with restriction enzymes which flank the fragment of interest. Digest sufficient plasmid to yield 30-40 ug of final purified fragment.b 3. Continue the digestion for at least 3 h to overnight to ensure that the digestion is complete. 4. Stop the reaction by addition of 0.5 M EDTA to 10 mM final concentration. 5. Extract the DNA twice with an equal volume of a 1:1 mixture of phenol/chloroform. 6. Ethanol precipitate the supernatant from step 5, wash the pellet with 70% ethanol, and resuspend in 150 ul TE. 244

10: Preparation of chromatin templates 7. Add 1/10 volume loading buffer, load into a 3 cm well in a 1% agarose gel, and electrophorese using TAE as the running buffer. To achieve good separation of the fragment, the gel should contain 1 ug/ml ethidium bromide. Cover the gel box with aluminium foil to prevent nicking of the DNA. 8. Briefly expose the gel to long-wave UV light and excise the DNA fragment. 9. Wash a piece of dialysis tubing (Spectra/Por 7) thoroughly with deionized water. Close one end with a dialysis clip. Place the fragment into the bag and add 1-2 ml TAE. Close the other end of the dialysis tubing with a dialysis clip and place into a gel tank with fresh TAE. Place the gel to one side of the electrophoresis bag and place this side of the bag closest to the negative electrode. Electroelute at 100 V for 30 min. Cover the gel box with foil to prevent nicking of the DNA. 10. Check the fragment has eluted from the gel slice by briefly exposing the dialysis bag to long-wave UV light. 11. Put the liquid from the dialysis bag into a 15 ml polypropylene tube. Wash remaining traces of DNA from the dialysis bag by adding 3 ml TAE, rinse the inside of the dialysis bag, and add the liquid to the 15 ml tube. 12. Remove traces of agarose by filtering the contents of the 15 ml tube through a 0.22 um Millipore filter into a fresh 15 ml polypropylene tube. 13. Add 300 ul 5 M NaCI to the sample. 14. Add 10 ml butanol, mix well, and centrifuge at 670 x g for 2 min at room temperature. 15. Remove the top butanol layer and discard. 16. Repeat steps 14 and 15 three times, taking care to ensure that the volume of the DNA sample is not reduced below 0.5 ml. If the volume is less than 1 ml prior to a butanol extraction, add 0.5-1 ml H2O. Continue the butanol extractions until the volume of the DNA sample is 0.5 ml. 17. Put the sample into a 1.5 ml Eppendorf tube. Extract once with chloroform. Put the upper phase into a fresh tube. 18. Transfer to dialysis tubing (washed well with deionized water). 19. Dialyse against 1500 ml 3 M NaCI for at least 2 h. 20. Dialyse against fresh 3 M NaCI for at least a further 4 h. 21. Dialyse against 1500 ml HE for at least 2 h. 22. Remove contents from dialysis bag to 1.5 ml Eppendorf tube. (NB It is essential to wash all powder from globes prior to removing contents of bag as this can interfere with subsequent measurements.)

245

Joan Boyes Protocol 6.

Continued

23. Centrifuge the Eppendorf tube 12600 x g for 30 sec at room temperature and remove the supernatant to a fresh Eppendorf tube. 24. Measure the A260 of the DNA sample in triplicate: take three aliquots, dilute 1 in 10, and measure the A260 against a blank of HE. 25. Calculate the concentration of the DNA. (NB It is crucial that the A260 measurement and calculation are very accurate: precise histone/DNA ratios are essential for efficient reconstitution.) a This plasmid preparation procedure is preferred since the DNA obtained is reproducibly of a high quality. Moreover, the DNA is not exposed to large quantities of ethidium bromide during plasmid isolation, which can cause nicking of the DNA. b Typically, for a 200 bp fragment cloned into the polylinker of pUC 19 plasmid, 1 mg of plasmid DNA is digested. Restriction enzymes which generate 5' protruding ends are chosen since these ends can easily be radioactively labelled with DNA polymerase (Protocol 7).

5.2 Amount of DNA needed in the reconstitution reaction The self-urea reconstitution reaction requires relatively high amounts of DNA and a typical reaction uses 10 ug of purified fragment. This can be reduced to 8 ug. However, at lower weights of DNA, losses to the surfaces of the dialysis tubing occur and effective reconstitution is not achieved. The most convenient way to visualize nucleosomes is by radioactively labelling the nucleosomal DNA. This can be achieved by any standard endlabelling technique. The highest specific activity is achieved, however, using the large fragment of E. coli DNA polymerase to 'fill-in' protruding 5' ends with a number of radioactively labelled nucleotides. It is also preferable to follow the radioactive labelling with a 'chase' of non-radioactive nucleotides. In this way, completely blunt-ended DNA fragments can be generated. Protocol 7. End labelling of DNA fragments for reconstitution Equipment and reagents • 10 X Klenow buffer: 100 mM Tris pH 7.5, 50 mM MgCI2, 75 mM DTT • HE (see Protocol 6)

. [32P]dNTPs, 110 or 220 TBq/mmol (Amersham; Dupont)

Method 1. Mix, on ice, 10 ug of purified fragment DNA with one-tenth final volume of 10 x Klenow buffer, 7.5 ul of each of the required radioactive dNTPs, and 1 ul of DNA polymerase large fragment in a final volume of 150-250 ul (depending on the concentration of the purified fragment). 2. Incubate at 10°C for 10 min.

246

10: Preparation of chromatin templates 3. Prepare a mixture of non-radioactive dNTPs, each at a concentration of 2 mM. Add 2 ul of this to the reaction and incubate for 10 min at 25°C. 4. Stop the reaction by adding 4.5 ul 5 M NaCI and 3 ul 0.5 M EDTA. 5. Inactivate the DNA polymerase by incubating at 70°C for 5 min. 6. Cool the mixture on ice for 10 min. 7. Extract twice with an equal volume of a 1:1 mixture of phenol/ chloroform. Precipitate with ethanol. (NB Do not add carrier (tRNA or glycogen.) 8. Wash the pellet with 70% ethanol. 9. Resuspend in 100 ul HE. 10. Vortex well and allow to resuspend for at least 10 min.

In the salt-urea reconstitution protocol, histones and DNA are mixed under conditions of high salt and urea, where histones will not aggregate the DNA. The salt and urea concentrations are gradually decreased by sequential dialysis steps. Initially, the arginine-rich histones, H3 and H4, associate with the DNA as a tetramer. The positioning of this tetramer plays a major role in determining the final nucleosome position along the DNA fragment (39). As the salt and urea concentrations are lowered further, H2A/B dimers associate with the H3/4 tetramer on the DNA. This generates the fully reconstituted histone octamer. It too little of histones H2A/B are used, hexamer complexes result in which only one H2A/B dimer is associated with the H3/4 tetramer. Hexamer complexes can also be produced if the H3/4 tetramer initially positions itself close to one end of the DNA fragment such that there is only sufficient DNA remaining to bind one H2A/B dimer. The presence of hexamer complexes is detected as described in Section 5.5. Technical considerations in setting up the reconstitution reaction: (a) to keep the sulfydryl group in histone H3 reduced until the nucleosome is reconstituted, b-mercaptoethanol is added to the reconstitution buffers (b) to prevent histone loss to surfaces, siliconized Eppendorf tubes are used throughout the procedure The compositions of the buffers used for nucleosome reconstitution are given in Table 1. The urea for these buffers is prepared as follows: (a) Weigh 800 g urea in a 2000 ml beaker, add water; warm to dissolve, make up the volume to 1667 ml. This gives an 8 M stock solution. (b) Deionize for 1 h by stirring with AG 501-X8 resin (Bio-Rad). (c) Remove the resin by filtering through a 0.45 um filter unit (Nalgene). 247

Joan Boyes Table 1 Buffers for nucleosome reconstitution Buffer 8 M Urea (ml)

1M Hepes 0.5 M EDTA (ml) (ml)

1 2 3 4 5 6 7

5 5 5 5 5 10 10

312.5 312.5 312.5 312.5 312.5 -

1 1 1 1 1 2 2

NaCI (g) b-Mercaptoethanol Total volume (ml) (ml) 58.44 35.06 29.22 23.35 17.53 35.06 -

0.35 0.35 0.35 0.35 0.35 0.7 -

500 500 500 500 500 1000 1000

These buffers are prepared in conical flasks in the absence of b-mercaptoethanol and chilled at 4°C.

5.3 Calculation of the amount of histones in the reconstitution reaction The histones used in the reconstitution reaction are calculated as a molar ratio of the DNA fragment. The optimum is to reconstitute all the free DNA into nucleosomes but without causing aggregation of DNA fragments due to excess histones. A ratio of 0.8 of histones to DNA has been found to best achieve these criteria. Since the histones are calculated as a molar ratio, the amount used varies with the length of the fragment and the longer the fragment the less the weight of histones added. The amount is calculated from the equation:

where L = length of DNA fragment, 667 = M r /bp DNA, 109 790 = Mr histone octamer, and 0.8 = percentage occupancy (above). For example, for a 227 bp fragment, this equation indicates that the amount of histones needed is 65% the weight of DNA. Thus, when reconstituting 10 ug DNA, 6.5 ug of histones are needed in total, i.e. 3.25 ug H2A,B and 3.25 ug H3,4. The importance of accurately adding the correct amounts of histones in this procedure cannot be overemphasized. Protocol 8. Mononucleosome reconstitution reaction Equipment and reagents • Buffers 1 7 (Table 7) • HE (see Protocol 6)

• Dialysis tubing: Spectra Por 7 (Spectrum), molecular weight cut off 8000 Da, 12 mm flat width

Method 1. Mix, on ice, in a siliconized Eppendorftube, 120 ul 5 M NaCI, 100 ul end-labelled DNA (Protocol 7, step 10), 2.1 ul p-mercaptoethanol

248

10: Preparation of chromatin templates

2.

3.

4. 5. 6.

7.

8. 9. 10. 11. 12.

13.

(diluted 1 in 10), HE to a total volume 300 ul (taking into account the volume of histones which will be added in steps 3 and 4). Mix very well by vortexing and centrifuge the tube 12600 X g for 30 sec at room temperature to collect the residue from the sides of the tube. Add the calculated amount of histones H2A/B (above). Take care to remove excess histones from the outside of the pipette tip. Mix very well with the DNA by flicking the tube. Centrifuge as in step 2. Add the calculated amount of histones H3/4 (above), mix, and centrifuge as in step 3. Add 350 ul b-mercaptoethanol to reconstitution buffer 1. Wear gloves and wash powder from the outside of the gloves. Wash approximately 9 cm of dialysis tubing thoroughly with deionized water. Tie a double knot at one end of the tubing. Pipette the reconstitution mix into the dialysis tubing using an automatic 1 ml pipette. A 100-200 ul pipette tip can be used to hold open the top of the dialysis tubing. Remove as much air as possible from the dialysis mixture and close the top of the dialysis tubing using two dialysis clips. Dialyse against reconstitution buffer 1 at 4°C overnight. Cover the top of the flask with a Perspex shield. Add 350 ul b-mercaptoethanol to reconstitution buffer 2. Transfer the dialysis to this buffer and dialyse for 1 h at 4°C. Add 350 ul b-mercaptoethanol to reconstitution buffer 3 and dialyse against this buffer for 1 h at 4°C. Continue changing the dialysis buffer each hour until the dialysis bag is in buffer 6. Dialyse against this buffer for at least 5 h at 4°C. Don't forget to add the b-mercaptoethanol at each step. Transfer the dialysis to HE and dialyse at 4°C overnight.

5.4 Purification of mononucleosome reconstitutes The reconstitution procedure does not generate a homogeneous population of mononucleosomes. Typical other products include non-reconstituted free DNA and aggregated material. In addition, hexamer complexes (Section 5.2) are occasionally produced. Accurate analysis of the ability of the transcription machinery to interact with nucleosomal templates is best performed on purified templates. Purification of mononucleosomes from free DNA and aggregated material can most simply be achieved using 10-40% sucrose gradients. Typically, 4 ml gradients, buffered by HE, are run in an SW60 rotor at 485000 X g for 17 h at 4°C. Fractions from the sucrose gradient can be used directly or dialysed to remove the sucrose. 249

Joan Boyes

Sucrose gradients have the advantage that very few additional contaminants are introduced. However, better separation of nucleosomes at different translational positions along the DNA template and of hexamer complexes is achieved by gel purification of the reconstitutes. In this procedure, it is particularly important to remove contaminants from the acrylamide gel which might destabilize the nucleosome. This is achieved by extensive pre-electrophoresis. In addition, precautions should be taken to keep the nucleosome intact both during the preparative gel electrophoresis and during elution of the nucleosome from the gel. During electrophoresis: (a) the preparative gel is electrophoresed using a low current (b) the nucleosome is loaded at high concentration at which it is more stable During elution of the nucleosome: (a) BSA is added to the elution buffer (b) Low elution volumes are used to keep the nucleosome as concentrated as possible (c) Siliconized Eppendorf tubes are used to prevent histone loss to surfaces Protocol 9. Preparative electrophoresis of nucleosome core particles (13) Equipment and reagents • Running buffer: 20 mM Hepes pH 7.5, 0.2 mM EDTA • 50% sucrose/HE: 50% sucrose w/v, 10 mM Hepes pH 7.5, 1 mM EDTA pH 8.0

. HE/BSA: 10 mM Hepes pH 7.5, 1 mM EDTA pH 8.0, 200 ug/ml BSA • Loading buffer (see Protocol &

Method 1. Sandwich two 1.5 mm spacers between the side edges of glass plates (in the size range 16 cm x 16 cm) and tape the sides and bottom edges for preparation of a polyacrylamide gel. 2. Prepare 50 ml of non-denaturing polyacrylamide gel mixture to give a final gel concentration of 4.5% acrylamide with acrylamide to bisacrylamide ratio of 40:1 (4.5% acrylamide, 0.056% bisacrylamide, 2.5% glycerol, 20 mM Hepes pH 7.5, 0.1 mM EDTA pH 8.0, 0.1% ammonium persulfate). 3. Mix well, add TEMED to a final concentration of 0.2%, and pour into the prepared glass plates. Insert a preparative gel comb (with well widths of 1.5 cm) and allow to polymerize for at least 30 min. 4. Pre-electrophorese for 6 h at 7.5 V/cm. 5. Replace the running buffer with fresh buffer and electrophorese for a further 30 min.

250

10: Preparation of chromatin templates 6. Replace running buffer again and decrease the voltage to 5 V/cm. 7. Wear gloves and wash excess powder from the outside of the gloves. Transfer the reconstitute (protocol 8 step 13) from the dialysis bag to a siliconized Eppendorf tube on ice as follows: (a) Wipe excess liquid from the outside of the dialysis tubing. (b) Cut the dialysis tubing just below the dialysis clips. (c) Invert the Eppendorf tube over the cut end of the dialysis tubing. (d) Invert the tubing and Eppendorf tube and squeeze the liquid into the Eppendorf tube. 8. Measure volume of reconstitute with a 1 ml automatic pipette. Mix gently and take a 1 ul aliquot for scintillation counting. This is important to determine the concentration of the reconstitutes at later stages. 9. Add 1/5 volume 50% sucrose/HE to the measured volume in step 8 to give a final concentration of 10% sucrose. 10. Load into two wells of the pre-electrophoresed preparative gel. Do not dilute the dialysis mixture by loading more lanes than this or the reconstitute will fall apart. Load an empty lane on the gel with loading buffer as a marker.a 11. Electrophorese at 5 V/cm until the bromophenol blue is close to the bottom of the gel (approximately 7 h for a 16 cm x 16 cm gel). 12. Dismantle the gel so that the polyacrylamide gel remains attached to one glass plate only. Cover the gel with cling film and expose the gel to X-ray film. A typical exposure time is 15 sec. 13. Cut out the minimum amount of acrylamide which contains the reconstituted nucleosome and place in a siliconized Eppendorf tube on ice. 14. Excise the free DNA band and place in a siliconized Eppendorf tube on ice. If a free DNA band is not visible, cut out the aggregated material which migrates above the reconstituted nucleosome band. This can be extracted later to yield DNA which is labelled to the same specific activity as the nucleosome. 15. Add a minimum volume of HE/BSA (usually 200 ul per cut out band). 16. Crush the acrylamide in the Eppendorf tube and rotate at 4°C overnight. 17. Centrifuge the Eppendorf tubes at 12 600 x g for 2 min at 4°C. 18. Remove the liquid to a fresh siliconized Eppendorf tube on ice. 19. Add 75 ul HE/BSA to each of the tubes containing the crushed acrylamide. Mix, centrifuge as above, and add the liquid to the siliconized

251

Joan Boyes Protocol 9. Continued Eppendorf tubes in step 18. Monitor the residual crushed acrylamide. If more than 50% of the total counts remain, repeat this step. 20. Mix the recovered liquids. Take a 1 ul aliquot. Count using a scintillation counter and calculate the concentration of recovered nucleosomes. 21. Store the reconstitutes undiluted at 4°C. a Radiolabelled, low molecular weight size markers can also be loaded in this lane to assist identification of the nucleosome and free DNA bands.

5.5 Characterization of the reconstitutes Following gel purification, the quality of the reconstitutes should be assessed. Typical contaminants are: (a) Free DNA: gel purification can remove a significant amount of contaminating free DNA but a low level of nucleosome disintegration during purification generally results in approximately 5% contamination. (b) hexamer complexes: poor excision of the nucleosome band above can result in residual contamination by these complexes. 5.5.1 Characterization by native gel electrophoresis The presence of these contaminants can most simply be analysed by electrophoresis of the reconstitute on a 4.5% non-denaturing gel. This is prepared exactly as described in Protocol 9 except that a 15-20 tooth comb is used. The samples are prepared and electrophoresed as follows: (a) Take aliquots of 10 000 counts of the samples from Protocol 9, step 20. (b) Add 50% sucrose in HE to final concentration of 10% sucrose. (NB Do not use dyes such as bromophenol blue in the loading buffer since this can disrupt histone/DNA interactions.) (c) Load the gel; load a separate lane with radioactively labelled low molecular weight markers (size range 50-500 bp) in loading buffer (Protocol 6). (d) Electrophorese in running buffer (Protocol 9) at 5 V/cm until the bromophenol blue is 85% of the way down the gel. (e) Dismantle the gel and dry on to one sheet of DE-81 paper and one sheet of Whatman 3 MM paper. (f) Detect the radioactive bands with X-ray film or a phosphor imager. A typical nucleosome reconstituted on to a 200 bp fragment has a mobility similar to 500 bp of unreconstituted DNA. Hexamer complexes electrophorese faster. Nucleosomes often occupy different positions along the length 252

10: Preparation of chromatin templates of the DNA fragment. In some cases, differences in nucleosome position can also be detected by the mobility on native gels: reconstitutes where the histone octamer is in the centre generally migrate slightly slower than those where the octamer is at the ends of the DNA fragment. 5.5.2 Characterization by micrococcal nuclease digestion (11) The best method to characterize the nucleosome positions and to determine whether hexamer complexes are present, is via micrococcal nuclease digestion. Micrococcal nuclease generates a typical 146 bp stop when it encounters the histone octamer. Hexamer complexes generate a 105-120 bp stop. These different species are detected as described in Protocol 10. To map the position of the histone octamer, the DNA which is protected from micrococcal nuclease digestion is isolated. Restriction enzymes which cut, only once, the original reconstituted fragment are then used to characterize the micrococcal nuclease-protected fragment. Three pieces of information can be gained from such restriction mapping: (a) If a restriction enzyme cuts the nucleosome-protected DNA fragment, then this indicates that the particular site was within the nucleosome. (b) The percentage of 146 bp fragments that are cut by a given restriction enzyme indicates the percentage of nucleosomes in that position. (c) The sizes of the restriction fragments generated indicate the distance of the particular restriction enzyme site from the end of the nucleosome. By combining these pieces of information using 3-6 restriction enzymes across a DNA fragment, a detailed map of the nucleosome positions can be generated. Protocol 10. Micrococcal nuclease analysis of reconstitutes Equipment and reagents • 10 x Micrococcal nuclease digestion buffer: 350 mM NaCI, 100 mM Hepes pH 7.5, 10 mM DTT, 20 mM CaCI2 • Micrococcal nuclease (Worthington; Protocol 1) • Low molecular weight DNA markers (e.g. pUC19 digested with Hpall) • HE (see Protocol 6)

• Loading buffer (see protocol 6) • 10 x Polynucleotide kinase buffer: 700 mM Tris pH 7.6, 100 mM MgCI2, 50 mM DTT • T4 Polynucleotide kinase (New England Biolabs) « [y32P]ATP, 220 TBq/mmol (Amersham; DuPont)

Method 1. Remove an aliquot of the purified reconstitute containing 0.45 ug of DNA (usually 50-100 ul) to a fresh siliconized Eppendorf tube. 2. Prepare a mixture of 15 ul 10 x micrococcal nuclease digestion buffer and HE (to give a final volume of 150 ul when added to the aliquot in

253

Joan Boyes

Protocol 10. Continued step 1). It is important to dilute the 10 X buffer in HE prior to addition to the reconstitute to avoid adding a high concentration of salt directly to the reconstitute. 3. Add the mixture to the reconstitute and place the Eppendorf in a water bath at 37°C. 4. Dilute the stock micrococcal nuclease 1 ul in 1 ml of 10 mM Hepes pH 7.5. 5. Add 2 ul of micrococcal nuclease to the pre-warmed Eppendorf. 6. Remove 50 ul aliquots from the Eppendorf at 1, 2, and 5 min after addition of the micrococcal nuclease to tubes containing 5 ul of 50 mM EDTA. Mix well. 7. Add an equal volume of a 1:1 mixture of phenol/chloroform. Extract and remove the supernatant to a fresh tube. 8. Ethanol precipitate using 20 ug glycogen as a carrier. 9. Resuspend the pellet in 20 ul TE. Add loading buffer. 10. Prepare an acrylamide gel with a final concentration of 5% acrylamide and an acrylamide to bisacrylamide ratio of 40:1 in 1 x TAE. 11. Load the gel; load a lane with low molecular weight DNA markers. 12. Electrophorese at 7.5 V/cm until the bromophenol blue is 80% of the way down the gel. 13. Dismantle the gel and stain in 2 ug/ml ethidium bromide in 1 x TAE for 10-20 min. 14. Destain the gel in 1 x TAE for 10 min. 15. Expose the gel to a UV light transilluminator and photograph. 16. To map the positions of the nucleosome along the DNA fragment, excise the 146 bp and/or 105-120 bp bands. Add 400 ul HE and 4 ul 10% SDS. Crush the acrylamide in the tube and elute the DNA for 2 h. 17. Centrifuge the tube at 12 600 x g at room temperature for 1 min. Remove the supernatant to a fresh tube. 18. Add an equal volume of a 1:1 mixture of phenol/chloroform. Extract. Remove the supernatant to a fresh tube and ethanol precipitate. 19. Resuspend the pellet in 15 ul TE. Add 2 ul 10 x kinase buffer, 2 ul [y32P]ATP, and 1 ul polynucleotide kinase. 20. Incubate at 37°C for 20 min. Heat inactivate the polynucleotide kinase at 70°C for 10 min. Increase the volume to 100 ul, phenol/chloroform extract, and ethanol precipitate as in step 18. 21. Resuspend the pellet in 400 ul TE.

254

10: Preparation of chromatin templates 22. Take 44 ul aliquots, add 5 ul of the appropriate 10 X restriction enzyme buffer and 5-10 U of the chosen restriction enzyme. Incubate at the appropriate restriction enzyme temperature for 2-4 h. 23. Prepare an acrylamide gel with a final concentration of 8% acrylamide (40:1 acrylamide to bis ratio) in 1 x TAE. 24. Load 20 ul of each restriction digestion in 1 x loading buffer. Load also a marker lane using radioactively labelled low molecular weight markers. 25. Electrophorese at 10 V/cm until the bromophenol blue is 70% of the way down the gel. 26. Dismantle the gel. Place on to one piece of DE-81 paper and one piece of Whatman 3 MM paper. Cover the gel with cling film, dry, and expose to X-ray film. 27. From the restriction enzyme digestion pattern, determine the relative nucleosome positions.

6. Transcription factor binding to nucleosome templates Initiation of transcription in vivo requires that the transcription machinery gains access to promoters in chromatin. Mononucleosome templates provide a good model to test the ability of transcription factors to interact with chromatin templates. Moreover, as described above, various factors which modulate nucleosome structure can be analysed in a defined system. In principle, the reaction of binding a transcription factor to a nucleosome is very straightforward. The transcription factor is mixed with the nucleosome under salt conditions typical for transcription factor/DNA interactions and the binding is analysed by either gel mobility shift assay or by DNAase I footprint analysis. Care should be taken, however, to ensure that the nucleosome is stable under the reaction conditions used. It has been shown that there is considerable dissociation of nucleosome cores into histones and DNA at low nucleosome concentrations. This problem is increased with increasing ionic strength and temperature. It has been suggested that at the nucleosome concentrations used in some studies (of 0.1 ug/ml), considerable nucleosome instability would occur (40). This, in turn, could lead to erroneous interpretations of the ability of transcription factors to bind to nucleosomes. We found that the problem of nucleosome instability could be considerably reduced by the use of siliconized Eppendorf tubes and the presence of BSA (400 ug/ml) in the reactions. In addition, the presence of the tetravalent cation, spermidine, at a final concentration of 2 mM significantly stabilized nucleosomes when binding reactions were performed at relatively low nucleo255

Joan Boyes some concentrations (0.5 ug/ml). Control reactions showed that equivalent results were obtained at high nucleosome concentrations in the absence of spermidine (41). Highly purified transcription factors are generally used in nucleosome binding studies. Nuclear extracts can be used, but care should be taken to avoid artefacts from the binding of HMG14/17. These are relatively abundant proteins which have a higher affinity for nucleosomes than DNA. Unlike HMG14/17, the binding of a transcription factor to a nucleosome is generally less efficient than to free DNA. Obviously, this varies with the transcription factor. A typical transcription factor which can bind to nucleosome templates does so with approximately 10-fold lower efficiency than to uncomplexed DNA. In primary experiments to investigate the ability of a particular transcription factor to bind to nucleosome templates, the concentration of factor is typically titrated over a 100- to 1000-fold range.

Protocol 11. Transcription Factor Binding to a Nucleosome Equipment and reagents • 10 x Binding buffer: 450 mM Hepes pH 7.5, 300 mM NaCI, 30 mM MgCl2, 20 mM spermidine, 10 mM b-mercaptoethanol, 200 ug/ml BSA • Running buffer: 20 mM Hepes pH 7.5, 0.2 mM EDTA

. Dilution buffer: 20 mM Hepes pH 7.5, 20% glycerol, 20 mM KCI, 2 mM MgCI2, 0.2 mM EGTA, 0.5 mM DTT • Loading buffer (see Protocol 6)

Method 1. Prepare a binding mixture on ice by adding in the following order: H20 to a final volume of 120 ul, 20 ul 10 X binding buffer, 20 ul 40% ficoll, 8 ug non-specific competitor DNA, 60 ug BSA, and 80 fmoles reconstitute. Prepare an equivalent, control, binding mixture containing 80 fmoles of free DNA instead of reconstitute. 2. Prepare two series of tubes, on ice, containing 10, 9, 7, 9.7, 9, 7, and 0 ul dilution buffer. 3. Aliquot 15 ul of the binding mixture into each tube. 4. Dilute 1 ul of the transcription factor into 100 ul dilution buffer. Add 1 and 3 ul, respectively, to the second and third tubes of the series. 5. Add 0.3, 1, 3, and 10 ul of undiluted transcription factor to the fourth, fifth, sixth, and last tubes of the series. Bind the transcription factor to the nucleosome or DNA for 10 min on ice. 6. Prepare an acrylamide gel with final concentration 4.5% as in Protocol 9, except use a normal 15-20 tooth comb instead of a preparative gel comb.

256

10: Preparation of chromatin templates 7. Load the samples into the bottom of the wells using a tapered loading tip. Include a marker lane where only loading buffer (Protocol 6) diluted in water is loaded. (NB Do not add dyes to the binding mixture as these can destroy protein/DNA interactions.) 8. Electrophorese at 5 V/cm until the bromophenol blue marker has migrated 75% of the length of the gel. 9. Dismantle the gel; place on to DE-81 paper and Whatman 3 MM paper and dry as in Protocol 10, step 26. 10. Expose to X-ray film or a phosphor-imager screen.

A supershift of the nucleosome following addition of the transcription factor is indicative that the transcription factor can bind to its site on a nucleosome. Once this has been established, further questions concerning the mechanism of nucleosome disruption can be addressed. For example, to determine whether the histones remain attached to the nucleosomal DNA following factor binding, a competitor oligonucleotide (carrying the transcription factor binding site) can be added to the factor/nucleosome binding mix. If the transcription factor causes histone removal, then protein-free DNA will result following electrophoresis. If the histones remain attached, then the nucleosome will be regenerated. If the transcription factor in question is incapable of independently binding to a nucleosome template, then various mechanisms to assist factor binding to a nucleosome can be investigated. For example, if a natural promoter is used with a number of different factor binding sites, then it can be investigated whether pre-incubation with the other factors assists binding of the transcription factor in question (42, 43, 44). In addition, use of the nucleosome positioning signals can reconstitute a nucleosome with the factor binding site in a slightly different orientation with respect to the underlying nucleosome structure; this can facilitate access of some factors to their binding site. Moreover, by reconstituting nucleosomes with acetylated histones (prepared as described in ref. 45), the role of histone acetylation in aiding the particular transcription factor to bind to the nucleosome can be investigated. By using a combination of such approaches, mononucleosomes can be used to determine, in considerable detail, the potential mechanisms of nucleosome disruption during gene transcription.

Acknowledgements I am very grateful to David Clark and Vasily Studitsky who taught me the methods described here and to Gary Felsenfeld, in whose laboratory many of these techniques were developed. I also gratefully acknowledge support from the Kay Kendall Leukaemia Fund. 257

Joan Boyes

References 1. Wolffe, A. P. (1992). In Chromatin structure and function. Academic Press, London. 2. Grunstein, M, Durrin, L. K., Mann, R. K., Fisher-Adams, G., and Johnson, L. M. (1992). In Transcriptional regulation, p. 1295. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 3. Gross, D. S. and Gerrard, W. T. (1988). Ann. Rev. Biochem. 57, 159. 4. Owen-Hughes, T. and Workman, J. L. (1994). Crit. Rev. Eukaryotic Gene Exp. 4, 403. 5. Kadonaga, J. T. (1998). Cell 92, 307. 6. McKnight, S. L. and Miller, O. L. (1979). Cell 7, 551. 7. Li, Q. and Wrange, O. (1993). Genes Dev. 7, 2471. 8. Li, Q. and Wrange, O. (1995). Mol. Cell. Biol 15, 4375. 9. Grunstein, M. (1997). Nature 389, 349. 10. Pazin, M. J. and Kadonaga, J. T. (1997). Cell 88, 737. 11. Clark, D. J. and Felsenfeld, G. (1992). Cell 71, 11. 12. Studitsky, V. M., Clark, D. J., and Felsenfeld, G. (1994). Cell 76, 371. 13. Studitsky, V.M., Clark, D. J, and Felsenfeld, G. (1995). Cell 83, 19. 14. Studitsky, V. M., Kassavetis, G. A., Geiduschek, E. P., and Felsenfeld, G. (1997). Science 278, 1960. 15. Orphaides, G., LeRoy, G., Chang, C.-H., Luse, D. S., and Reinberg, D. (1998). Cell 92, 105. 16. Brown, S. A., Imbalzano, A. N., and Kingston, R. E. (1996). Genes Dev. 10, 1479. 17. Shimamura, A., Jessee, B., and Worcel, A. (1989). In Methods in enzymology (ed. P. M. Wassarman and R. D. Kornberg), Vol. 170, p. 603. Academic Press, London. 18. Laskey, R. A., Honda, B. M., Mills, A. D., and Finch, J. T. (1978). Nature 275, 416. 19. Workman, J. L., Taylor, I. C. A., Kingston, R. E., and Roeder, R. G. (1991). In Functional organisation of the nucleus (ed. B. A. Hamkalo and S. C. R. Elgin). Academic Press, San Diego. 20. Becker, P. B. and Wu, C. (1992). Mol. Cell. Biol. 12, 2241. 21. Kamakaka, R. T., Bulger, M., and Kadonaga, J. T. (1993). Genes Dev. 7, 1779. 22. Bulger, M., Ito, T., Kamakaka, R. T., and Kadonaga, J. T. (1995). Proc. Natl Acad. Sci. USA 92, 11726. 23. Ito, T., Bulger, M., Pazin, M. J., Kobayashi, R., and Kadonaga, J. T. (1997). Cell 90, 1479. 24. Ito, T., Bulger, M., Kobayashi, R., and Kadonaga, J. T. (1996). Mol. Cell. Biol. 16, 3112. 25. Kamakaka, R. T., Tyree, C. M., and Kadonaga, J. T.(1991). Proc. Natl Acad. Sci. USA 88, 1024. 26. Batson, S. C., Sundseth, R., Heath, C. V., Samuels, M., and Hansen, U. (1992). Mol. Cell. Biol. 12, 1639. 27. Dean, A., Pederson, D. S., and Simpson, R. T. (1989). In Methods in enzymology (ed. P. M. Wassarman and R. D. Kornberg), Vol. 170, p. 26. Academic Press, London. 28. Camerini-Otero, D. and Felsenfeld, G. (1976). Cell 8, 333. 29. Rhodes, D. and Lskey, R. A. (1989). In Methods in enzymology (ed. P. M. Wassarman and R. D. Kornberg), Vol. 170, p. 575. Academic Press, London. 30. Stein, A. (1989). In Methods in enzymology (ed. P. M. Wassarman and R. D. Kornberg), Vol. 170, p. 585. Academic Press, London.

258

10: Preparation of chromatin templates 31. Sealy, L., Burgess, R. R., Cotten, M., and Chalkley, R. (1989). In Methods in enzymology (ed. P. M. Wassarman and R. D. Kornberg), Vol. 170, p. 26. Academic Press, London. 32. von Holt, C., Brandt, W. F., Greyling, H. J., Lindsey, G. G., Retief, J. D., Rodrigues, J. et al. (1989). In Methods in enzymology (ed. P. M. Wassarman and R. D. Kornberg), Vol. 170, p. 431. Academic Press, London. 33. Simon, R. H. and Felsenfeld, G. (1979). Nucl. Acids Res. 6, 689. 34. Boehm, L. and Crane-Robinson, C. (1984). Bioscience Reports 4, 365. 35. Studitsky, V. M., Clark, D. J., and Felsenfeld, G. (1996). In Methods in enzymology (ed. S. Adhya), Vol. 274, p. 246. Academic Press, London. 36. Rhodes, D. (1985). EMBO J, 4, 3473. 37. Neubauer, B. and Horz, W. (1989). In Methods in enzymology (ed. P. M. Wassarman and R. D. Kornberg), Vol. 170, p. 630. Academic Press, London. 38. Shrader, T. E. and Crothers, D. M. (1989). Proc. Natl Acad. Sci. USA 86, 7418. 39. Dong, F. and van Holde, K. E. (1991). Proc. natl Acad. Sci USA 88, 10596. 40. Godde, J. S. and Wolffe, A. P. (1995). J. Biol. Chem. 270, 27399. 41. Boyes, J., Omichinski, J. G., Clark, D. J., Pikaart, M. J., and Felsenfeld, G. (1998). J. Mol. Biol., 379, 529. 42. Steger, D. J. and Workman, J. L. (1997). EMBO J 16, 2463. 43. Pina, B., Brueggemeier, U., and Beato, M. (1990). Cell 60, 719. 44. Adams, C. C. and Workman, J. L.(1993). Cell 72, 305. 45. Krajewski, W. A. and Becker, P. B. (1998). In Chromatin protocols (ed. P. B. Becker). Humana Press, Totowa, USA. In press.

259

This page intentionally left blank

11 Analysis of transcription factor modifications N. SHAUN B. THOMAS

1. Introduction Transcription factors are regulated by being modified by phosphorylation and glycosylation. Each can cause changes in protein charge, shape, hydrophilicity, and mass, all of which can be exploited in differentiating the modified from the unmodified protein. This chapter will cover how these post-transcriptional modifications can be analysed, the methods most frequently used, and the problems and limitations associated with each. I will assume that the transcription factor has been cloned and sequenced and that you have basic reagents to detect your factor—in particular, antibodies that can be used for Western blotting and immunoprecipitation. Thereafter, the questions to be answered will necessarily dictate which methods are most appropriate. In this chapter I shall start by introducing simple techniques to determine whether a given protein is phosphorylated or glycosylated before covering more complex and costly procedures to determine which sites are modified and the physiological relevance of such modifications.

2. Phosphorylation The examples in this chapter are based on work from my laboratory which has been directed towards an understanding of the molecular mechanisms involved in controlling entry into the cell cycle and I will illustrate various methods by using members of the E2F and octamer transcription factor families. These are important proteins which are required in order for cells to respond to a mitogenic stimulus and protein phosphorylation is crucial in regulating their activities (reviewed in ref. 1). Activation of mitogenic signalling pathways in a quiescent cell via ras, Raf, and MAPK leads to the synthesis of D-type cyclins. These in turn activate cyclin-dependent kinases (cdks) which then phosphorylate proteins such as members of the retinoblastoma family. The retinoblastoma protein (pRb) is a nuclear phosphoprotein which is phosphorylated

N. Shaun B. Thomas progressively on more than 10 sites as quiescent cells progress through Gl, reaching a hyperphosphorylated form at the G1/S border. pRb is phosphorylated on different sites (2; and see Section 2.5) by a succession of cdks which are activated at different times during G1: for example, cdk6 and cdk4 in early/mid G1; followed by cdk2 in late G1. The pRb protein shares regions of homology with two other proteins, p107 and p130, and like pRb they are phosphoproteins which have consensus sites for phosphorylation by cdks. Hypophosphorylated pRb binds members of the E2F transcription factor family (reviewed in ref. 3) and it recruits a histone deacetylase (HDAC1) to the E2F DNA binding site (reviewed in ref. 4). This causes histone deacetylation which in turn represses transcription. Each member of the E2F family is composed of a heterodimer of an E2F protein (E2F-1 to E2F-5) together with a member of the DP family (DP-1 or DP-2) (reviewed in refs 5-7). p130 and p107 also bind and repress E2F and different members of the pRb family have some specificity for binding different E2Fs: E2F-1, -2, and -3 bind pRb rather than p107 whereas E2F-4 or -5 bind p130 and p107 in preference to pRb. The functional consequences are that E2F-1 overexpression preferentially overcomes cell-cycle arrest by pRb, and E2F4/DP-l overcomes arrest caused by p130. Thus the balance between particular E2Fs and pRb or p130 is clearly important in determining whether a cell proliferates or remains quiescent. Members of the E2F family of transcription factors are also regulated by phosphorylation. For example, E2F-1 is a phosphoprotein which can be phosphorylated in vitro only by certain cdks and it is likely that E2F-1 phosphorylation during S phase switches off its transcriptional activation, thereby downregulating certain E2F-dependent genes which were activated during late G1 (G1B) when the unphosphorylated E2F was released from repression by pRb. Indeed, there is a check point during S phase which requires E2F-1 phosphorylation by cyclinA-cdk (8). In contrast to E2F-1, E2F-4 is multiply phosphorylated and the hyperphosphorylated form is predominant in G0 (9). This form is replaced by multiple hypophosphorylated forms as the cells progress through G1; which may be required for its activation. The DP-1 partner bound by these E2Fs is also a phosphoprotein which is dephosphorylated as cells enter the cell cycle and this correlates with an increase in E2F activation. This introduction to the pRb and E2F families illustrates the importance of the phosphorylation of a number of different proteins in regulating transcription from a promoter containing an E2F site. Further, the methods used by a number of groups to investigate a system of this complexity serve to illustrate how to study transcription factor phosphorylation and many of the protocols which have been used will be described later. The other transcription factor which I use here as an example is Oct-2. This factor is also required for cell proliferation and B lymphocytes from Oct-2 -1-'~ mice have a defect in cell proliferation early in the G1 phase of the cell cycle 262

11: Analysis of modifications (G1A) such that they will not progress through to G1B in response to stimulation either by lipopolysaccharide or anti- u-chain antibody (10). We have shown also that Oct-2 (and the related factor Oct-1) are downregulated during cell-cycle arrest of a B-cell line by a-interferon (11). Oct-2 is a transcription factor (reviewed in ref. 12) belonging to the POU (Pit, Oct, Unc) family (reviewed in ref. 13), which is expressed primarily in B lymphocytes and neuronal cells (14, 15). It is encoded by a single gene, but a family of isoforms is produced by alternative splicing (16). The complement of proteins produced by alternative splicing differ in different cell types: Oct-2.1 is the predominant form in B lymphocytes and Oct-2.4 and Oct-2.5 in neuronal cells (17). Members of the Oct-2 family bind to the consensus octamer DNA motif 5'-ATGCAAAT-3', which is present in all immunoglobulin gene promoters and in the immunoglobulin heavy chain enhancer (18, 19). However, the B lymphocyte and neuronal forms of Oct-2 have different effects on transcription from octamer motifs (20). Oct-2.1 has a C-terminal activation domain which is dominant over the inhibitory domain at the N terminus, and so Oct2.1 stimulates transcription. In contrast, Oct-2.4 and Oct-2.5 do not have the C-terminal activation domain and so they repress transcription (17, 21). Indeed, when expressed together Oct-2.4 and Oct-2.5 interact with Oct-2.1 and convert it from an activator into a repressor (22). The activities of the Oct-1 and Oct-2 transcription factors are also regulated by phosphorylation. Oct-1 is phosphorylated by protein kinase A, which inhibits its DNA binding activity (23), while a Ca2+/calmodulin-activated phosphatase, calcineurin, augments Oct-1-dependent transcription (24). Oct-1 is also phosphorylated by the cdc2 kinase, which terminates Oct-1-dependent transcription during mitosis (25). Members of the Oct-2 family are also phosphorylated (unpublished data; see also Sections 2.2 and 2.3) although the precise function of site-specific phosphorylations has yet to be determined. In addition to regulating activity by affecting DNA binding or transcriptional activation, changes in phosphorylation can also modulate the specificity of binding to different DNA sites. For example, we have shown that several different kinases can affect binding of Oct-1 and Oct-2 to a variety of different octamer DNA binding sites (26). This phenomenon is also true for p53 (27). Thus, phosphorylation of Oct-1 and the Oct-2 family is important in regulating their activity and specificity. Furthermore, the methods used to analyse the phosphorylation of Oct-2 serve to illustrate the analysis of a family of homologous proteins which are produced by alternative splicing. In the sections below I will give detailed protocols for a variety of different methods for analysing protein phosphorylation. I shall use as examples two transcription factors, namely E2F-4 and Oct-2, as well as pRb.

2.1 Gel electrophoresis There are many forms of gel electrophoresis which have been used successfully to resolve phosphorylated from unphosphorylated proteins, or indeed 263

N. Shaun B. Thomas proteins phosphorylated on different sites. These include two-dimensional gels and one-dimensional polyacrylamide gel electrophoresis of proteins denatured by boiling in the presence of SDS (SDS-PAGE). The simplest method for detecting potential phosphorylated forms of the protein is to look for a change in its mobility by SDS-PAGE. 2.1.1 Electrophoresis in one dimension Western blotting is a standard procedure in most cell and molecular biology laboratories, but there are potential pitfalls in sample preparation and gel conditions which can markedly affect the results obtained. For example, selective protein degradation and dephosphorylation of the transcription factor of interest can occur even during lysis in SDS sample buffer. Measures to try and circumvent these problems are detailed in Protocol 1. From the data shown in Figure 1 it can be seen that changing the acrylamide: bisacrylamide ratio from 38:1 to 200:1 alters the migration of E2F-4 which is present in quiescent primary T cells such that at 38:1 this form (lane 1) migrates slower than those forms present in Daudi cells arrested in G1 with a-interferon (IFN) (lane 2). However, at 200:1, E2F-4 from T cells (lane 3) migrates with that in Daudi cells (lane 4). The forms of E2F-4 are due to differences in phosphorylation (see 9, 32), therefore we have determined simple conditions under which the different phosphorylated forms of the E2F-4 component of the transcription factor can be separated. This assay has been used, for example, to determine when the phosphorylation state of E2F-4 changes in CD34+ haemopoietic progenitor cells as they enter the cell cycle (9). Protocol 1. Analysis by Western blotting In this protocol it is assumed that the transcription factor of interest is present in sufficient abundance to be detected in an extract of up to 1 X 106 cells. A. Sample preparation 1. Grow cells in culture or isolate primary cells as appropriate. 2. Pellet 1 x 106 cells at 800 g for 7 min in a 1.5 ml screw-cap microcentrifuge tube (e.g. Sarstedt 72.692.005). Remove all of the supernatant without disturbing the pellet.a 3. Flick the tube several times to dislodge the cell pellet and then add 25-50 ul 2 x SDS sample buffer (containing protease and phosphatase inhibitorsb), flick hard or vortex briefly, and heat immediately at 100°C for 10 min. 4. Centrifuge in a microcentrifuge at 10000 x g for 5 sec, flick to mix the solution, and repeat the centrifugation. The sample should not be viscous and is now ready to load.

264

11: Analysis of modifications B. SDS-PAGE There are many variations on the original method described in refs 30, 31. These are standard in most laboratories so I will not describe the method in detail. However, try using different acrylamide:bisacrylamide ratios from 20:1 to 200:1, keeping the percentage of acrylamide the same.0 These conditions can alter the relative migration of different forms of many proteins, thereby resolving phosphorylated from

unphosphoryl-

ated species. Examples of the effects of altering the acrylamide:bisacrylamide ratio are shown in Figure 7 for E2F-4, * Depending on the type of cells being used, it may be necessary to put each sample into icecold PBS containing diisopropyl fluorophosphate (DIFP) to prevent degradation, followed by lysis in boiling 2 X SDS sample buffer (containing N-ethyl maleimide (NEM), protease and phosphatase inhibitors: see footnote b). Note that we have been able to detect and analyse E2F complexes in human primary mongcytes but only after pre-incubating cells with DIFP before lysis (28). (I wish to thank P. J. Roberts, Department of Haematology, UCL Medical School for advice on the use of NEM and DIFP.) b ln general we add 50 mM NaF, 2 mM phenyl methyl sulfonyl fluoride (PMSF), 1 mM Na3VO4.UH2O, 5 U/ml aprotinin, 5 ug/ml pepstatin A, 5 ug/ml leupeptin, 1 mM DIFP, although this list is by no means exhaustive. Cells such as neutraphils contain significant protease activity which can degrade many cellular proteins during cell lysis. In this case, pre-incubate the cells on ice in PBS containing DIFP before the addition of boiling SDS sample buffer containing DIFP and NEM. Note that protein degradation during lysis may not be uniform and that the protein of interest may be completely degraded even when others are not. Addition of 2 mM NEM, which covalently modifies SH groups in enzymes including proteases, may prevent such degradation (29). Great care should be taken when using DIFP, NEM, and the other inhibitors shown above: always read the data sheet and use appropriate safety measures to avoid personal exposure. c Note that it may be necessary to alter the final percentage of acrylamide to keep the protein of interest on the gel.

There are many other examples where phosphorylation changes alter the migration of a transcription factor, such as E2F-1 (33) and DP-1 (34). It should be noted that phosphorylation can either cause proteins to be retarded

Figure 1. Electrophoresis in one dimension. Proteins in total cell lysates of human primary T cells (tanes 1 and 3) and Daudi B cells cultured in a-IFN for 3 days (lanes 2 and 4) were separated by SDS-PAGE as described in Protocol 1 with an acrylamide:bisacrylamide ratio of either 38:1 (lanes 1 and 2) or 200:1 (lanes 3 and 4). E2F-4 in each sample was then detected by Western blotting.

265

N. Shaun B. Thomas relative to their un- or less- phosphorylated counterparts, as shown above, or to migrate faster (see 35). However, if phosphorylation of a given protein does not change its migration by one-dimensional SDS-PAGE, there are other electrophoretic methods which are more powerful in resolving phosphorylated forms of a protein, such as two-dimensional gels, examples of which are in the next section. 2.1.2 Electrophoresis in two dimensions The most common technique which works well for a large number of proteins employs isoelectric focusing in the first dimension followed by SDS-PAGE in the second (36). Isoelectric focusing separates proteins on the basis of their isoelectric points, either using carrier ampholines to establish a pH gradient or by using an immobilized pH gradient (for example, see 37). Therefore two forms of a protein which differ because they contain different numbers of phosphates can be separated. If isoelectric focusing is sufficient to separate the different forms of the protein of interest then a second dimension may not be required. A good example of the use of two-dimensional gel electrophoresis to analyse the phosphorylation state of the Oct-1 transcription factor during the cell cycle is given by Roberts et al. (38). Two methods of detection were used: the gels were either subjected to Western blotting with an anti-Oct-1 antibody, or Oct-1 immunoprecipitated from 32P-labelled cells with an antibody covalently coupled to protein A-Sepharose beads (to avoid overloading the gels with immunoglobulin light and heavy chains) was detected by autoradiography. These data illustrate the power of two-dimensional gel electrophoresis to resolve different phosphorylated forms of a protein which are not separated well by SDS-PAGE. However, it should be noted that some proteins do not focus well and some do not denature completely in the presence of SDS or even remain as dimeric or multimeric complexes. Therefore alternative gel systems are important in such circumstances and a couple which could be tried are acid/urea gels (39) or gels which have been used to separate ribosomal proteins (40).

2.2 Dephosphorylation in vitro In the examples given above I have shown how different forms of the same protein can be separated by electrophoresis and the implication has been that these are forms of the same protein carrying different post-translational modifications. However, differences in migration could be due to a change in the size of the protein caused by specific cleavage (e.g. clipping of pRb when cells are triggered to undergo apoptosis), or to the production of different forms of the protein by alternative splicing or RNA editing. In order to determine that differences in migration in SDS-PAGE can be accounted for by changes in phosphorylation, dephosphorylation of some proteins can be carried out in vitro. 266

11: Analysis of modifications 2.2.1 Dephosphorylation in vitro with endogenous cellular phosphatases Phosphatases which are active in a cell lysate can be used to dephosphorylate the protein of interest. The procedures for cell lysis are described in Protocol 2 and an example of the type of experiment which can be done is shown in Figure 2A and in ref. 41. Incubating a lysate of proliferating Daudi B cells at 30°C for up to 10 min in the presence of protease inhibitors but in the absence of phosphatase inhibitors leads to a gradual change in the migration of the retinoblastoma protein from a hyperphosphorylated form (pRbp; lane 3) to a dephosphorylated form (pRb; lane 6). Incubation for 10 min in the presence of a cocktail of phosphatase inhibitors prevents this change in migration (lane 2), showing that the shift is due to dephosphorylation and not degradation. The shift in the migration of pRb shown in Figure 2A can be inhibited by the addition of EGTA or EDTA to the reaction (Figure 2B, top panel, lanes 8-10, 11-13), indicating that there is a cellular type 2B/C phosphatase requiring a divalent cation(s) which can dephosphorylate pRb (42). Note that this experiment only shows that a shift in pRb migration is caused by a change in its phosphorylation state and does not prove that a particular type of phosphatase is responsible for dephosphorylating pRb in viable cells. The appropriate experiments to prove which phosphatase(s) dephosphorylate a particular protein of interest are beyond the scope of this chapter. This type of experiment shows that the protein is phosphorylated and the protocol used does not require the protein to be purified or the purchase of expensive enzymes. However, not all proteins are dephosphorylated: under these conditions the phosphorylation state and migration of p130 does not change (Figure 2B, bottom panel, lanes 1-3) whereas pRb is dephosphorylated in the same lysates (Figure 2B, top panel, lanes 1-3). In such circumstances it is necessary to carry out dephosphorylation experiments with specific phosphatases. 2.2.2 Dephosphorylation in vitro with specific phosphatases A variety of phosphatases can be used to dephosphorylate phosphoproteins in vitro, including acid and alkaline phosphatases, \-protein phosphatase, which is a dual-specificity tyrosine and serine/threonine phosphatase, and other more specific phosphatases. All can be purified from the appropriate source and many are available commercially. We have used y-protein phosphatase to dephosphorylate E2F-4 (9) but here I will give an example of its use to show that p130 and some of the forms of the alternatively spliced Oct-2 family have different mobilities in SDS-PAGE due to phosphorylation. An example is shown in Figure 2C, D in which the effects of dephosphorylating p130 and Oct-2 and subsequent analysis by one-dimensional SDSPAGE and Western blotting are illustrated. The migration of p130 was altered by incubation with \-phosphatase (Figure 2C, lane 2) but not when 267

N. Shaun B. Thomas

268

11: Analysis of modifications Figure 2. Dephosphorylation in vitro. Cell lysates of proliferating Daudi cells were prepared as described in Protocol 2 and incubated at 30°C with or without phosphatase inhibitors, as detailed below, but in the presence of protease inhibitors. pRb, p130, or Oct1 and Oct-2 in each sample was detected by Western blotting. The hyperphosphorylated form of pRb is labelled pRbp and the predominant phosphorylated forms of p130 as 1, 2, and 3. (A) Dephosphorylation of pRb. Lane 2: lysate incubated for 10 min with phosphatase inhibitor cocktail (5 mM EDTA, 1 mM EGTA, 50 mM NaF, 1 mM b-glycerophosphate, 1 mM Na3VO4, 5 mM Na4P2O7.10H2O); lanes 3-6: lysates incubated for 0, 0.5, 3, and 10 min in the absence of phosphatase inhibitors. Lanes 1 and 7 contain lysates of Daudi cells cultured for 24 h with a-IFN, which contain hyper- and hypophosphorylated forms of pRb (41). (B) pRb and p130: effects of specific phosphatase inhibitors. Lysates were incubated for 0 (lanes 1, 4, 8, and 11), 5 (lanes 2, 5, 9, and 12), and 10 (lanes 3, 6, 10, and 13) min in the absence of phosphatase inhibitors (lanes 1-3), with the inhibitor cocktail (lanes 4-6), or with EGTA (lanes 8-10) or EDTA (lanes 11-13) alone. Lane 7: total cell lysate of proliferating Daudi cells; lane 14: a-IFN treated Daudi cells. pRb and p130 in each sample were detected by Western blotting. (C) Dephosphorylation of p130 with \-phosphatase. Lysates were incubated for 60 min with y-phosphatase (lanes 2, 3) or without (lanes 1, 4) in the absence (lanes 1 and 2) or presence (lanes 3 and 4) of 50 mM NaF and 2 mM Na3VO4. p130 was detected by Western blotting and 0 denotes the dephosphorylated form. (D) Dephosphorylation of Oct-2 with y-phosphatase. Lysates of Daudi cells stably transfected with constructs expressing Oct-2.1 (lanes 3, 4) or Oct-2.5 (lanes 7, 8, 11, 12), or not expressing exogenous Oct-2 (lanes 1, 2, 5, 6, 9 and 10) were incubated for 60 min with (lanes 1-8, 10, and 12) or without (lanes 9, 11) y-phosphatase in the presence (lanes 1, 3, 5, and 7) or absence (lanes 2, 4, 6, 8, and 9-12) of 50 mM NaF and 2 mM Na3V04. Oct-2 in each sample was detected by Western blotting and the Oct-2.1 and Oct-2.5 isoforms are indicated. The antibody also detects cellular Oct-1, the position of which is shown.

sodium fluoride and orthovanadate were added (lane 3). These data suggest that p130 is phosphorylated and that different phosphorylated forms can be separated by SDS-PAGE. To analyse the phosphorylation state of different isoforms of Oct-2, Daudi cells were transfected with constructs expressing Oct-2.1 or Oct-2.5 and lysates were then treated with y-phosphatase (Figure 2D). This reduced the three predominant phosphorylated forms of Oct-2.1 (lane 3) to two faster-migrating forms (lane 4). This was also true of the endogenous cellular Oct-2 (lanes 5, 6 and 9, 10). Similarly, Oct-2.5 was resolved into two forms (lanes 7, 11), which were reduced to one, fastermigrating form by y-phosphatase (lanes 8, 12). Since these changes in migration of Oct-2.1 and Oct-2.5 caused by y-phosphatase were prevented by adding NaF and Na3VO4, the data suggest that these forms are produced by differences in their phosphorylation state. Protocol 2. y-protein phosphatase Reagents . Buffer A: 20 mM Hepes pH 7.8, 450 mM 1 ug/ml trypsin inhibitor, 0.5 ug/ml aprotinin, 40 ug/ml bestatin, 2 mM DIFP NaCI, 25% (v/v) glycerol, 0.2 mM EDTA, 0.5 mM DTT, 0.5 mM PMSF, 0.5 ug/ml leu. 50 mM Tris-HCI pH 7.8 at 25°C, 2 mM MnCI2, 100 ug/ml BSA peptin, 0.5 ug/ml Sigma protease inhibitor,

269

N. Shaun B. Thomas Protocol 2.

Continued

A. Sample preparation 1. To a pellet of 1 x 107 cells add 100 ul lysis buffer A.a Note that this buffer contains a variety of protease inhibitors but no phosphatase inhibitors. Freeze/thaw rapidly three times in dry ice/ethanol and at 37°C. Centrifuge at 13000 x g for 2 min and remove the supernatant to a fresh tube. Freeze in aliquots on dry ice and keep at -80°C or maintain on ice if the extract is to be used immediately. 2. Measure the protein content of the extract by a suitable method, such as that of Lowry or Bradford. B. Phosphatase reaction 1. Aliquot a suitable amount (4 ug protein in this case) into each of eight microcentrifuge tubes on ice and adjust each to 50 mM Tris-HCI pH 7.8 at 25°C, 2 mM MnCI2, 100 ug/ml BSA. The following should be added in a final volume of 25 ul: Sample

50 mM NaF

1

2

--

2 mM Na3V04

---

100 U \-phosphataseb

-

+

3

4

5

+

-

+

6

7

8

+

+

-

+

+

-

+

+

+

+ + ---

Incubate at 30°C for up to 60 min. a

It may be necessary to pre-incubate some cells in PBS containing 2 mM DIFP to prevent partial or complete proteolysis. b y-phosphatase is a dual-specificity phosphatase (43) which will dephosphorylate both phosphoserine/-threonine as well as phosphotyrosine. Dephosphorylation of the former is inhibited by NaF and the latter by Na3V04. It should be noted, however, that \-phosphatase will also dephosphorylate phosphohistidine (44), which may be important if phosphoamino acid analysis (Protocol 8) shows that the protein of interest is phosphorylated on amino acids other than serine, threonine, or tyrosine.

2.3 Radioactive labelling The protocols described above are important as they provide simple and robust methods for analysing putative phosphorylation changes. Such methods have been used in my laboratory, for example, for characterizing changes in the phosphorylation state of E2F-4 and p130 as primary haemopoietic CD34+ progenitor cells enter the cell cycle from G0 (9, 44a). However, the definitive test for phosphorylation is by labelling the protein with [32P]orthophosphate, and this is also crucial if one is to map phosphorylation sites (see Section 2.4). Relatively large amounts of radioactivity are usually required for each experiment, typically 5-20 mCi/185-740 mBq, and this has profound effects on the 270

11: Analysis of modifications procedures which have to be employed. Before starting you should discuss the health and safety issues with the appropriate Safety Officer. In order to label the transcription factor of interest, cells should be cultured for a few hours in phosphate-free medium in the presence of [32P]orthophosphate. The duration of labelling will be dictated by the type of experiment you need to perform—as long as possible for mapping, or brief incubations for characterizing phosphorylation changes in response to agonists. However, cells require phosphate and so the labelling time has to be a balance between the minimum possible so as not to unduly stress or kill the cells and a prolonged period to ensure maximum labelling, which has to be determined experimentally. In order to analyse the 32P-labelled protein of interest, this has to be isolated, usually by immunoprecipitation, although one of a number of column chromatography or other physical methods could be used if appropriate. Methods for labelling and isolating [32P]Oct-2 are described in Protocol 3. Protocol 3. Labelling with [32P]orthophosphate This method is for labelling Oct-1, Oct-2, or pRb in the Daudi B-cell line. However, we have used very similar conditions to label a number of other proteins in a variety of cell types. Equipment and reagents • Centrifuge, 50 ml conical screw-capped tubes • Microcentrifuge, 1.5 ml screw-capped tubes with rubber O-ring • Phosphate-free DMEM/10% (v/v) dialysed fetal calf serum (FCS)a (Sigma) • 37°C Incubator (humidified, 5% C02) • High-salt extraction buffer: as low-salt buffer, but containing 450 mM NaCI . [32P]orthophosphateb (1-10 mCi/37-370 mBq)

• Low-salt extraction buffer: 50 mM Tris pH 8.0 at 4°C, 150 mM NaCI, 0.5% (v/v) NP40, 5 mM EDTA, 0.1 mM Na3V04, 50 mM NaF, 1 mM EGTA, 5 mM NaP3, 1 mM b-glycerophosphate, 10 mM K2HP04, 0.5 mM PMSF, 0.5 ug/ml leupeptin, 0.5 ug/ml Sigma protease inhibitor, 1 ug/ml trypsin inhibitor, 0.5 ug/ml aprotinin, 40 ug/ml bestatin, 2 mM DIFP

Method 1. Pellet 1 X 107 proliferating Daudi cells (20 ml at 5 X 105 cells/ml) at 800 g for 7 min in a disposable 50 ml conical, screw-capped centrifuge tube. 2. Wash three timesc with 25 ml each phosphate-free DMEM/10% (v/v) dialysed FCS. 3. To remove as much unlabelled phosphate from the cells as possible, culture in 20 ml phosphate-free DMEM/10% (v/v) dialysed FCS for 30 min at 37°C in a humidified atmosphere of 5% C02. 4. Pellet cells as in step 1 and resuspend in 10 ml of the same medium containing 1-10 mCi/37-370 mBq [32P]orthophosphate. 5. Culture for 4 hd at 37°C in a humidified atmosphere of 5% CO2.

271

N. Shaun B. Thomas Protocol 3. Continued 6. Pellet as in step 1e and dispose of the supernatant. 7. Transfer the tube to ice and lyse the cells for immunoprecipitation.f For the analysis of Oct-2 or pRb, lyse the cells in 800 ul low-salt buffer, transfer to a screw-capped 1.5 ml microcentrifuge tube with rubber Oring and centrifuge at 13000 X g for 1 min to pellet nuclei and insoluble matter. Extract proteins in the nuclear pellet with 80 ul high-salt buffer for 20 min with end-over-end mixing, centrifuge at 13 000 g for 3 min, and combine the low- and high-salt extracts. 8. Pre-clear the lysates of proteins which stick non-specifically to IgGg for 1 h with end-over-end mixing, centrifuge at 13000 g for 5 sec, then immunoprecipitate Oct-2 or pRb with 1-10 ug of the appropriate antibodyh and 20 ul protein A-agarose for 1 h (or overnight if necessary). Centrifuge at 13000 x g for 5 sec and retain the supernatant i then wash the pellets four times with 800 ul each of the low-salt buffer before boiling for 10 min in 30 ul 2 X SDS sample buffer. a Use phosphate-free medium other than DMEM as appropriate to the cells being used. The phosphate-free FCS is produced by dialysing 50 ml heat-inactivated FCS in a 10000 Da molecular-weight cut-off dialysis membrane against four changes of 5 litres 10 mM Hepes pH 7.0, 137 mM NaCI, 2.7 mM KG to remove small molecules, including orthophosphate. Alternatively, use Tyrodes' medium and incubate the cells in a water bath (45, 46). b We have generally used cell-labelling grade [32P]orthophosphate from Amersham (PBS-13), but an equivalent product from any other manufacturer would be fine. The amount of [32P]orthophosphate required for labelling the protein of interest should be determined experimentally and kept to a minimum. c After removing the medium flick the tube several times to disaggregate the cell pellet into a single-cell suspension before adding the next wash. d Since Daudi cells grow in suspension they settle out during the culture period. Therefore shake to resuspend them in the medium every 30 min. An adherent cell line can be left undisturbed during the labelling period. e We use a 50 ml screw-capped tube even for 10 ml. Volumes > 30 ml can leak during centrifugation. f All procedures should be carried out in a designated radioactive area, preferably in a 4°C cold-room. Otherwise, manipulate all samples on ice and use a refrigerated microcentrifuge. g Use 10 ug rabbit IgG as a control for a polyclonal antibody, or the antipeptide antibody preincubated with a 100-fold excess of the cognate peptide, or a matched amount of an isotypematched immunoglobulin when using a mouse monoclonal antibody. h The amount should be determined by titration. i The maximum amount of a particular "P-labelled protein may be obtained by repeated immunoprecipitation from the retained supernatant. Alternatively, additional 32P-labelled proteins can be immunoprecipitated sequentially. If the extent of labelling of a particular protein in cell extracts cultured under different conditions is to be compared, then the incorporation of 32 P into total cell protein has to be determined. This is done by precipitating a small volume (5-10 ul in triplicate) in 50% (w/v) TCA (with 0.01% (w/v) BSA as carrier), collecting the precipitate on GF/C filters (Whatman) which are washed in 5% (w/v) TCA then ethanol before quantification by scintillation counting.

An example of the results from such a preparation are shown in Figure 3. Note that Oct-2 (see Figure 2D) runs as a 32P-labelled smear (Figure 3A, lane 3),

272

Figure 3. 32P-Labelling of Oct-1, Oct-2, and pRb in whole cells. (A) [32P)Oct-1 and Oct-2. Anti-Oct-1 (lanes 2 and 5} and anti-Oct-2 (lanes 3 and 6) antibodies were used to immunoprecipitate each from a 32P-labelled lysate of proliferating Daudi cells according to Protocol 3. Control immunoprecipitates with rabbit IgG are in lanes 4 and 7. Lane 1: 14 C-labelled molecular weight markers. Proteins were separated by SDS-PAGE and blotted on to a membrane (Immobilon-P). Labelled proteins were detected by autoradiography (left panel) and Oct-1 and Oct-2 specifically by Western blotting with an antibody which recognizes both proteins (right panel). (B) [32P]pRb. pRb from 32P-labelled proliferating Daudi cells (lanes 1 and 4) or cells cultured with a-IFN (lanes 2, 5) or aphidicolin (lanes 3, 6) for 24 h was visualized by autoradiography (lanes 1-3) and pRb protein on the same membrane was detected by Western blotting. pRb in total lysates of the cells prior to labelling is shown in lanes 7-9.

273

N. Shaun B. Thomas consistent with there being several phosphorylated forms (see Figure 2V). In contrast, the principal form of pRb which is labelled corresponds to the hyperphosphorylated form (pRbp), which predominates in cells arrested at the G1/S border with aphidicolin (Figure 3B, lanes 3, 6, and 9).

2.4 Mapping phosphorylation sites Having 32P-labelled the protein of interest you are now in a position to be able to determine the site(s) phosphorylated. There are several excellent examples which illustrate the use of the methodology for mapping the sites on E2F-1 (33), DP-1 (34), C-JUN (47), and pRb (48). I will illustrate one method by showing how phosphopeptide maps of pRb can be generated. These protocols are adapted from methods provided by J. R. Woodgett, The Ontario Cancer Institute, and J. Hsuan, The Ludwig Institute for Cancer Research, London. First, purify the 32P-labelled protein. This can be done by chromatographic methods such as FPLC or by elution from a polyacrylamide gel. We have used the latter and the method is described briefly in Protocol 4. Protocol 4.

Preparation of [32P]pRb

Method 1. Separate [32P]pRb (labelled in Daudi cells and immunoprecipitated as described in Protocol 3) by SDS-PAGE (6%).a 2. Transfer on to PVDF membrane (Immobilon-P, Millipore) by semi-dry electroblotting for 45 min at 500 mA. Allow the membrane to dry. 3. Wrap the membrane in clear film (cling film/Saran wrap), stick on to a card sported with orientation marks,b and visualize [32P]pRb on the membrane by autoradiography or using a phosphor imager.c 4. Cut out a piece of membrane containing the [32P]pRb,d dice into small pieces with a clean scalpel blade, and put into a screw-topped 1.5 ml microcentrifuge tube. 5. Proceed as in Protocol 5. * Reduce samples with DTT (not b-mercaptoethanol) and boil 3 min in SDS sample buffer. Alkylation is unnecessary as this occurs naturally in the gel by free acrylamide. Load directly on to a single track of an SDS-PAGE gel with 0.1 mM sodium thioglycolate in the upper reservoir. Leave at least one track empty either side to avoid contamination. b This can be prepared by mixing a little [32P)orthophosphate with black India ink. A spot should give a signal on X-ray film in 3-4 h. c Note that the phosphor imager must be capable of printing the image accurately at full size. d The piece of membrane should be as small as possible to contain all the [32P]pRb: re-expose the remaining membrane to check.

Next, the 32P-labelled protein has to be cleaved to generate a number of phosphopeptides. The protease or chemical cleavage method used will depend 274

11: Analysis of modifications on the sequence of the protein of interest. However, many maps are produced after cleavage with trypsin, which usually generates a number of small peptides that are ideal for mapping. However, in order to analyse specific phosphorylation changes, independent maps should be generated after cleavage with different proteases as peptides derived by cleavage with one protease may contain several phosphorylation sites. A list of proteases and their cleavage sites is given in ref. 49. Protocol 5.

Protease digestion

Reagents • PVP buffer: 0.5% polyvinylpyrrolidone, 100 mM acetic acid • 50 mM ammonium carbonate pH 8, stored in aliquots at -80°C

• 1 mg/ml TPCK trypsin (Calbiochem), stored in aliquots at -80°C

Method 1. Re-wet the pieces of membrane in 200 ul methanol, wash twice with 500 ul water,3 and add 300 ul PVP buffer. Vortex and incubate for 30 min at 37°C to block non-specific binding sites on the membrane. 2. Discard the PVP solution and wash the membrane pieces twice with 500 uJ water. Discard. Add to the membrane 50 ul 50 mM ammonium carbonate pH 8 and 10 ul 1 mg/ml TPCK trypsin. Incubate overnight at 37°C, then add 10 ul TPCK trypsinb and incubate for a further 3 h. 3. Centrifuge in a microcentrifuge for 5 sec and remove the supernatant, which contains [32P]peptides digested by the trypsin, to a fresh tube. Add 100 ul ammonium carbonate solution to the membrane, vortex, centrifuge, and combine the supernatants. Add 500 ul water and lyophilize.c 4. When dry, place the tube on ice and prepare performic acid (PA).d Add 100 ul PA9 to the [32P]peptide pellet, vortex, and place on ice for 1 h. Chill 1 ml water on ice at the same time. 5. Add 400 ul ice-cold water, freeze on dry ice, lyophilize.c 6. Add 300 (J water, vortex for 1 min, lyophilize.c At this point there should not be obvious salt crystals at the bottom of the tube. If there are, add 300 ul water and lyophilize again. "Unless stated otherwise, all water used for any of the methods in this chapter is assumed to be double-distilled and deionized or HPLC-grade. "Other proteases can be used such as thermolysin (see 33). Alternatively, proteins can be cleaved chemically with cyanogen bromide (50). c Preferably, this should be done while the tube is being centrifuged (using an apparatus such as a SpinVac with the heat on) to prevent losses due to bubble formation. d Add 100 nl 30% H2O2 to 900 (il formic acid and incubate for 45 min at 20°C, then chill on ice. "PA is used to oxidize Met and Cys. Once the PA has been added you must continue to the end of step 5.

275

N. Shaun B. Thomas There are several other methods for recovering proteins which have been separated by electrophoresis. An efficient method for recovering and digesting a protein in a gel slice is given below in Protocol 6. Protocol 6. Recovery and digestion of a protein in a gel slice Reagents • Stain: 2.5 g coomassie blue G250, a 400 ml HPLC-grade methanol, 80 ml acetic acid, 320 ml water . Destain: 10% (v/v) acetic acid, 5% (v/v) methanol • Wash buffer: 50% (v/v) acetonitrile, 10 mM Tris-HCI pH 8.8, 0.1% Thesit b (BoehringerMannheim)

• UltrafreeMC insert (Millipore) • Alkylated trypsin (Promega) or achromobacter lys-C (Wako or Promega) . 10 mM Tris-HCI pH 8.0 or pH 8.8 . 10 mM Tris-HCI pH 8.8, 0.1% Thesit . 0.08% (v/v) TFA, 50% (v/v) acetonitrile

Method 1. Separate proteins by SDS-PAGE as described for Protocol 4. 2. Transfer the gel directly to fresh staining solution and stain 5-10 min, or more only if necessary. 3. Immediately destain with rapid changes until bands are just apparent on light-box illumination. Excise the protein using a clean scalpel in a minimum gel volume and transfer to a 1.5 ml microcentrifuge tube. Also excise a comparable volume of protein-free gel from a blank lane as a control. Store the excised pieces without free liquid at -20°C. 4. Wash twice in 1 ml water for 20 min each at 20°C, then destain in 1 ml wash buffer for 10 min at 37°C, mixing every 2 min. 5. Remove the supernatant and all remaining liquid, then fragment the gel into small pieces, preferably by spinning through an UltrafreeMC (Millipore) insert with the filter removed.c 6. Add alkylated trypsin to ~5% (w/w) in 10 ul 10 mM Tris-HCI pH 8.0 or achromobacter lys-C (Wako or Promega) to ~2% (w/w) in 10 ul 10 mM Tris-HCI pH8.8 and allow to absorb completely into gel pieces at 20°C. Add sufficient 10 mM Tris-HCI buffer containing 0.1% Thesit to cover the gel pieces and incubate overnight at 37°C. 7. Centrifuge for 5 sec at 13000 g in a microcentrifuge and remove all free supernatant to a fresh tube. Extract the gel pieces with 200 ul 10 mM Tris-HCI pH 8.8, 0.1% Thesit for at least 2 h at 37°C with mixing. Centrifuge for 5 sec at 13000 g and add the supernatant to the first extract. Repeat the extraction of gel pieces with 200 ul 0.08% (v/v) TFA, 50% (v/v) acetonitrile in a sonicating water bath at 20°C for 2 hd. 8. Pool the extracts, centrifuge for 1 min at 13000 g and transfer the supernatant to fresh tubes (leave the last 20 ul which contains micro-

276

11: Analysis of modifications scopic gel pieces). Dry by lyophilization (SpinVac) to ~50 ul without heating. The peptides are then ready to be separated by HPLCd or another suitable method. a Do not use coomassie blue R250 as this stains too heavily. b Thesit should be omitted when using thin layer chromatography (Protocol 7) or mass spectrometry (see Section 2.8). c Keratins tend to contaminate the preparation when direct mashing/cutting procedures are used. This is important when the phosphorylation site is to be determined by direct sequencing (see Section 2.8). d Check the recovery of [32P]peptides at each stage by Cerenkov counting. If the [32P]peptides remain in the gel slice, sonification in a sonicating water bath can aid their recovery. a For HPLC, add an equal volume of 0.08% (v/v) TFA, 1% (v/v) acetonitrile and load directly on to an octadecyl reverse phase column.

The next step is to 'map' the phosphorylation sites. This involves separating the phosphorylated peptides from each other by some suitable method and is usually the first step in identifying the specific amino acids that are phosphorylated. There are several ways of resolving peptides to produce a peptide 'map', such as by HPLC or electrophoresis followed by chromatography. In this section I will describe the method for producing a two-dimensional 'map' on a crystalline cellulose thin-layer plate by electrophoresis followed by ascending chromatography (51; also see 52). Protocol 7. Peptide maps Equipment and reagents • Buffer A: 50 ml 88% (v/v) formic acid, 156 ml acetic acid, 1794 ml water; pH 1.9 • Microcentrifuge • Cellulose TLC plate (20 cm x 20 cm, Merck or Kodak)

• Whatman 3MM paper • Electrophoresis apparatus a • Buffer B: 187.5 ml n-butanol, 125 ml pyridine, 37.5 ml acetic acid, 150 ml water

Method 1. Resuspend the pellet from Protocol 6 in 500 ul buffer A.b Vortex and centrifuge for 3 min at 13 000 g in a microcentrifuge to pellet any insoluble material.c Remove 400 ul to a new microcentrifuge tube. Add 500 ul buffer A to the previous pellet. Vortex, centrifuge for 3 min, remove 400 ul, and combine with the previous supernatant. 2. Lyophilize the samples in both tubes (soluble and insoluble material) in a SpinVac. Count each dry tube in a b-counter by Cerenkov counting. A count of > 500 c.p.m. for multiply phosphorylated proteins will give a good mapd (> 50 c.p.m. for phosphoamino acid analysis, Protocol 8). 3. Resuspend the soluble fraction in 10 ul buffer A, vortex, centrifuge for 3 min, remove 8 ul, and load 1-2 ul a timee on to a cellulose TLC plate,f/

277

N. Shaun B. Thomas Protocol 7.

Continued

drying in between with a stream of air (pressurized air line or a hair dryer). The loading position should be ~3 cm in from each edge at the bottom left-hand corner of the plate. 4. Wet the plate in buffer A with two sheets of 3MM paper with a 1 cm diameter hole cut to encircle the loaded sample. The sample is concentrated into a tight spot as the buffer wets the plate. 5. Load the plate into the electrophoresis apparatus according to the manufacturers' instructions and electrophorese at 1000 V, 17 mA for 40 min (or as appropriate to resolve your peptides). Air dry the plate. 6. Separate the peptides in the second dimension by ascending chromatography for 16-20 h in buffer B. All operations should be carried out in a fume hood. Air dry the plate, cover with plastic film (cling film/ Saran wrap), attach to a card marked with [32P]ink and expose to preflashed X-ray film or a phosphor-imager plate. 7. If you want to compare 32P-labelled peptides derived originally from your protein of interest in cells treated under different conditions, you should run half of each sample on plates and mix the remainder and run the mixed sample on another plate (Figure 4). Similarly, spots which you know or suspect correspond to particular phosphopeptides should be verified by mixing and then running the [32P]peptide preparation with a pure synthetic phosphopeptide (see below), which can be visualized by staining the plate with ninhydrin. a The electrophoresis apparatus is made for the job (HTLE-7000, C.B.S. Scientific Co., Del Mar, Ca), has a cooled stage with electrode buffer tanks on either side, and the plate is held in place during the run by a pressurized air bag. b This is the buffer which will be used for the first dimension (electrophoresis) of the tryptic map. Buffers with a different pH, such as pH 3.5, pH 4.72, or pH 8.9, may be more suitable for your particular protein. This must be determined by trial and error. pH 3.5 buffer: 100 ml acetic acid, 10 ml pyridine, 1890 ml water; pH 4.72 buffer: 100 ml n-butanol, 50 ml pyridine, 50 ml acetic acid, 1800 ml water; pH 8.9 buffer: 2 g ammonium carbonate, water to 2000 ml. c Insoluble material which is left at this stage may cause smearing in the two-dimensional (2D) map. d Care should be taken that static on the plastic does not produce spuriously high counts. Wiping each tube with a tissue wetted with ethanol before counting reduces this problem. I have produced 2D maps of pRb, which is multiply phosphorylated, with ~50 c.p.m., so don't throw the sample away if the counts are < 500 c.p.m. a The surface of the plate is fragile and it is best to load the sample with a long, flexible gel-loading tip on a 20 ul automatic pipette. f Check the plate against the light for imperfections in manufacture before use. The loading position can be marked lightly with a soft pencil.

An example of tryptic phosphopeptide mapping of [32P]pRb is shown in Figure 4. This figure illustrates the reproducibility of the method: proliferating cells (panel A) and cells arrested at the G1/S border (panel B) contain pRb 278

11: Analysis of modifications

Figure 4. Tryptic peptide map of [32P]pRb. Tryptic peptide maps of 32P-labelled pRb from proliferating Daudi cells (A) or cells cultured for 24 h with aphidicolin (B) were generated according to Protocols 3, 4, 5, and 7. Panel C is of a 1:1 mix of samples in A and B. Panel D is a [32P]phosphoamino acid analysis (see Protocol 8) of a protein phosphorylated predominantly on serine. The positions of phosphotyrosine (Yp), phosphothreonone (Tp), and free phosphate (P1) are indicated. Spots marked * are undetermined. In each case the sample was applied at the origin and the direction of electrophoresis and chromatography are indicated.

which is phosphorylated on the same sites, as shown in panel C which is a map of a 1:1 mix of each sample. It is important to know whether your protein is phosphorylated on serine, threonine, or tyrosine, particularly if any of the 32P spots in the phosphopeptide map change specifically in samples of cells treated, say with a particular cytokine. Determine the location of any spots of interest and scrape off the cellulose from the plate into a screw-capped microcentrifuge tube (re-expose the plate to determine whether you have scraped the correct place). The type of amino acid which is phosphorylated can then be determined according to Protocol 8. An example of phosphoamino acid analysis of a protein phosphorylated predominantly on serine is shown in Figure 4D. It might be possible to use the phosphoamino acid data to analyse phosphorylatton changes in the protein of interest using antibodies to phosphotyrosine, phosphoserine, or phosphothre279

N. Shaun B. Thomas onine. Good antibodies to phosphotyrosine have been available for several years and an example of their use in analysing the phosphorylation of Stat91 is described by Shuai et al. (53). Antibodies to phosphoserine and phosphothreonine have also been generated, but to date such antibodies are not widely used because a single antibody is not available which universally recognizes either phosphoserine or phosphothreonine in a variety of proteins. Protocol 8. Phosphoamino acid analysis Equipment and reagents 5.7 M 'constant-boiling' HCI (Sigma) Heating block, 110°C SpinVac Buffer A (Protocol 7) Phosphoamino acids: Serp, Thrp, Tyrp

• Cellulose TLC plate • pH 3.5 buffer: 100 ml acetic acid, 10 ml pyridine, 1890 ml water • Ninhydrin

Method 1. Resuspend the sample in 100 ul 5.7 M 'constant-boiling' HCI, vortex, and pulse in a microcentrifuge. 2. Incubate at 110°C in a heating block for 60 min. Dry the sample immediately in a SpinVac.a 3. Resuspend in 10 ul buffer A containing 0.7 ug each Serp, Thrp, Tyrp. Centrifuge in a microcentrifuge (13000 x g, 1 min) and remove 8 ul of the supernatant. Count to determine that the procedure has liberated the phosphoamino acids. 4. Spot on to a cellulose TLC plate and wet (as in Protocol 7). Four samples can be run on each plate. Electrophorese in buffer A for 20 min at 1.5 kV. Air dry the plate. 5. Turn the plate 90° anticlockwise, wet the plateb in pH 3.5 buffer containing 0.5 mM EDTA and carry out electrophoresis for 16 min at 1.3 kV. 6. Dry the plate in a fume hood, spray with ninhydrin and bake for 15 min at 65°C to develop the stain. Wrap the plate in cling film, tape to a card spotted with [32P]ink and expose to X-ray film or phosphor imager. Localization of the 32P spots with respect to the standards will show whether serine, threonine, or tyrosine is phosphorylated.c a Use an NaOH trap to protect the pump. b Wetting the plate is tricky: use Whatman 3MM strips immersed in buffer A (Protocol 7) according to the instrument manufacturers' protocol. c Other amino acids can be phosphorylated, such as histidine. If the spots obtained do not colocalize with the phospho-Ser, -Thr or -Tyr standards, then seek advice.

Having obtained a phosphopeptide map in which one of the spots is of interest because, for example, its phosphorylation changes during the cell 280

11: Analysis of modifications cycle, you may wish to map the precise site that is phosphorylated. This will allow you to determine the possible function of phosphorylation at this site (see Section 2.7) or to raise an antibody which recognizes the protein only when it is phosphorylated at this site (Section 2.6). The phosphorylation site(s) can be determined by cleaving the protein with different proteases (see ref. 49) until the site can be mapped unequivocally to a small fragment. If there are one or two potential sites in this fragment then one of two approaches may be adopted. Phosphopeptides can be synthesized either by an in-house facility or purchased from one of a number of companies and their migration in 2D peptide mapping and/or HPLC can then be compared with the [32P]peptide of interest. Alternatively the phosphorylation site(s) can be mutated (see Section 2.7) and the modified protein then analysed either after transfection into whole cells or in suitable assays in vitro. Of course, it is possible to narrow the possible phosphorylation sites if, for example, you suspect or know that a particular kinase is involved in phosphorylating your protein (see ref. 33).

2.5 Phosphorylation in vitro Labelling a protein in whole cells is important if you wish to determine which of perhaps many sites become phosphorylated in response to a physiologically relevant stimulus, but it may be more appropriate for studies on kinase activation to phosphorylate the protein of interest in vitro. For example, this has been done for the phosphorylation of Oct-2 by cdc2/cyclin B (26) and C-JUN by SAPK/JNK (53). If the substrate is to be the whole protein then it must be cloned, expressed, and purified. Expression is usually in bacteria or in insect cells and the expressed protein is frequently engineered with an additional cleavable peptide 'tag' at the amino or carboxyl terminus to aid in purification. There are drawbacks to each system: proteins produced in bacteria may be insoluble except under denaturing conditions and even if they are soluble their conformation may not be the same as in eukaryotic cells, whereas proteins produced in insect cells may become partially phosphorylated by endogenous kinases. Alternatively, synthetic peptides corresponding to predicted phosphorylation sites in the protein of interest can be used. This is covered in Protocol 9 and examples of the use of different cyclin-dependent kinases (cdk6 and cdk2) to phosphorylate synthetic peptides corresponding to sites in pRb (2) are shown in Figure 5. Active kinases can be isolated from eukaryotic cells by a number of biochemical techniques including immunoprecipitation. However, caution must be exercised as a kinase preparation may be contaminated with other kinases. It may be possible to overcome this by preparing the active kinase by different methods or by using antibodies to different sites in the protein. Alternatively, a range of inhibitors which preferentially inhibit specific kinases can be included in any kinase reaction to try and inhibit contaminating kinases but not the kinase of choice. 281

N. Shaun B. Thomas

Figure 5. Labelling peptides. Cdk6 (A) and cdk2 (B) from cells in G0 (lane 2) and in late G1/early S phase (lane 3) were used in vitro to phosphorylate two peptides corresponding to specific phosphorylation sites in pRb. In each case, lane 1 is of the rabbit IgG control. Kinases and 32P-labelled peptides in each reaction were separated by electrophoresis in a tricine gel and detected by Western blotting (lower panel) and by autoradiography (upper panel) as described in Protocol 9.

Protocol 9. Phosphorylation of synthetic peptides Equipment and reagents • 2 x kinase buffer: 50 mM Tris pH 7.4, 10 mM DTT, 2 mM EGTA, 20 mM MgCI2 . ATP and [y 3 2 P]ATP • p81 paper (Whatman)

* 75 mM H3PO4 • SDS sample buffer . 10% (v/v) propan-2-ol, 10% |v/v| acetic acid

Method 1. Immunoprecipitate the kinase of interest, using the appropriate conditions to maintain kinase activity. 2. Wash the protein A-agarose-immune complexes twice with 400 ul of 2 x kinase buffer, 3. Incubate the immunopurified cdks at 30°C for 30 min a in 1 x kinase buffer containing 2.5 ug of peptide substrate, 50 uM ATP, and 5 uCi (185 kBq) [y32P]ATP, 10 mM NaF, and 1 mM sodium orthovanadate, in a final volume of 20 ul. (The substrate used for cdk6 kinase assay was RAAPLSPIPHIPR and that used for the cdk2 reaction was AKAKKTPKKAKK; see ref. 2). 282

11: Analysis of modifications 4. The incorporation of 32P on to each peptide can be determined by quantifying the signal obtained from the gel (see step 5) or as follows: remove three 2 ul aliquots of each reaction, spot on to p81 paper, wash six times in 75 mM H3PO4, and quantify by liquid scintillation counting. 5. To the remainder of the reaction, add 10 ul of 2 x SDS sample buffer and heat at 100°C for 5 min. Load 15 ul on to each lane of a three-stage tricine 16.5% polyacrylamide gel system, made as described in (54). Carry out electrophoresis at 100 V, until the bromophenol blue dyefront reaches the bottom edge of the gel. This gel system allows separation of the [32P]peptide in the resolving gel and the cdk in the spacer gel. The separating gel containing the [32P]peptide should be cut away, fixed in four changes of 10% (v/v) propan-2-ol, 10% (v/v) acetic acid,b dried in a gel dryer for 45 min at 80°C and then exposed to pre-flashed X-ray film or to a phosphor-imager plate. The relative amount of 32P incorporated into the peptide in each sample should be quantified by scanning the film or using image quantification software. In order to ensure that the same amount of kinase was used in each assay, protein in the spacer gel should be transferred to a nitrocellulose membrane for Western blotting with the appropriate antibody (to cdk2 or cdk6). a The time of incubation should be adjusted to ensure linearity with the peptide and kinase being used. b Some peptides may elute from the gel under these conditions. If this occurs, alternative fixation conditions should be tried, such as 50% (w/v) TCA.

2.6 Generation and use of phosphorylation-site-specific antibodies Once the sites of phosphorylation have been determined then synthetic phosphopeptides corresponding to these sites can be used to generate antibodies. Specific antibodies which only recognize the protein when it is phosphorylated at that site can then be identified. A suitable method is described by Alberta and Stiles in (55). The antibodies generated are very powerful as they can then be used in Western blotting, ELISA, or immunocytochemical assays to investigate the presence or absence of a particular phosphorylated form of a protein in cell lysates. For example, such a strategy has been used for CREB (56, 57), pRb (2), and for p53 (58).

2.7 Functional analysis To test the functional importance of phosphorylation at a specific site, that amino acid has to be mutated in the cDNA of interest. This can be carried out using one of a variety of site-directed mutagenesis strategies (see Chapter 8) and it is usual to mutate the codon(s) encoding serine or threonine to alanine 283

N. Shaun B. Thomas and tyrosine to phenylalanine. The mutated amino acids have similar structures of the same polarity and it is assumed therefore that the mutated protein will have the same overall conformation as the wild-type protein. However, neither alanine nor phenylalanine can be phosphorylated. These constructs can be used to confirm that phosphorylation does indeed occur at the specific site(s) identified in Section 2.4 in whole cells and to determine whether this phosphorylation has a functional consequence. If a cell line is available which lacks the protein of interest then cDNAs encoding the mutant or wild-type protein can be stably transfected, otherwise a tagged construct may have to be used. After treating the transfectants in the appropriate manner, 32P-labelling and mapping can then be carried out as described above. The map generated from the mutated protein should lack the specific spot which it is believed corresponds to the phosphopeptide of interest. This strategy assumes that mutation of the protein of interest does not cause apoptosis or inhibit cell proliferation, in which case transient transfection (if transfection efficiency is high) or inducible vectors should be used. The transfected cells can also be used to test the effect of the mutated transcription factor on other aspects of cell physiology which will depend on the factor being investigated. Also, the effect on various molecular mechanisms can be tested (for example, see refs 33, 51). There are a number of assays which can be used, such as electrophoretic mobility shift assays (Chapter 1) and interaction with other proteins and reporter assays (Chapter 8). It should be noted that particular amino acids in certain transcription factors can either be phosphorylated or glycosylated (see Sections 3.1 and 3.1.4), so caution should be exercised in claiming that an effect on cell physiology is due solely to phosphorylation.

2.8 Other methods for analysing phosphorylated proteins This section briefly covers more powerful methods for the analysis of phosphorylated proteins. However, the techniques require a number of expensive pieces of equipment and specialist knowledge in their use. I shall describe below how such analyses can be done, but without detailed protocols. In brief, peptides are purified by HPLC, analysed by multidimensional electrospray mass spectrometry (ESMS), and fractions containing phosphorylated peptides are then sequenced (see, for example, ref. 59). Alternatively, phosphopeptides may be analysed by matrix-assisted laser desorption time-of-flight mass spectrometry (MALDI-MS) (60). The phosphorylated protein has first to be isolated either from whole cells or phosphorylated in vitro with candidate kinases, as described above. The methods for determining the precise sites that are phosphorylated are given in Protocol 10. 284

11: Analysis of modifications Protocol 10. Alternative methods for determining phosphorylation sites Method 1. Starting with 500 pmol of a suitable protein, digest overnight at 37°C with trypsin (such as TPCK trypsin, Calbiochem; or modified trypsin, Promega) at a ratio of 7:1 (w/w) protein:trypsin. Other suitable enzymes can also be used as appropriate. Methods for analysing the peptides produced are described in detail in ref. 60a. 2. Separate the peptides produced by reverse-phase HPLC and detect using an in-line UV detector set at 214 nm. 3. To identify which fractions contain phosphopeptides, analyse a small amount (10%) of each fraction by mass spectroscopy (negative-ion ESMS) optimized to detect PO3~ ions (61, 62). 4. Determine the molecular weights of the phosphopeptides in positive fractions by nanoelectrospray mass spectroscopy. Such accurate molecular weight measurement may yield sufficient data to determine which peptide is in each fraction. 5. Determine which sites are phosphorylated by sequencing by mass spectroscopy. These methods are very powerful and precise and provide direct methods for determining which sites in a protein, are phosphorylated.

3. Other post-translational modifications In the sections above I have described a number of methods which can be used to analyse the phosphorylation of transcription factors. Certain transcription factors have also been shown to be modified by O-linked glycosylation on serines and threonines. Like phosphorylation, this is a dynamic process which may regulate transcriptional activity and below I will deal briefly with the main analytical methods that have been used.

3.1 O-Linked glycosylation Over 50 nuclear proteins have been shown to be modified by O-linked glycosylation. This involves the attachment of the O-linked monosaccharide Nacerylglucosamine (O-GlcNAc) to the hydroxyl groups of either serine or threonine. Many of the proteins that are modified in this way are also phosphorylated and it has been suggested that the sugar may block the sites of phosphorylation and so may have a reciprocal effect to that caused by phosphorylation (63-65). Attachment of O-GlcNAc is regulated and can turn over 285

N. Shaun B. Thomas rapidly and enzymes capable of adding and removing the sugar from proteins have been purified. Such dynamic changes in glycosylation have been shown to occur in a signal-dependent manner, suggesting that they may be crucial for regulating cellular functions (66). Transcription factors are amongst the proteins that are modified with O-GlcNAc (67) and examples of these are the serum response factor (SRF) (68), c-Myc (65), Spl (67), and p53 (69). For SRF and c-Myc, the modified amino acids are in the transcriptional activation domains and for c-Myc the site modified (Thr-58) is also known to be phosphorylated in vivo (reviewed in ref. 70). Phosphorylation at this site is important for cell proliferation and Myc harbouring a mutation of Thr-58 has enhanced transformation potential in cell lines and increased tumour-inducing potential in animals. This site is mutated in Burkitt's and other lymphomas as well as in v-myc, which provides evidence for the potential importance of reciprocal O-GlcNAcylation (i.e. attachment of O-GlcNAc)/ phosphorylation in neoplasia. However, as the same site can be modified by two different mechanisms it is not clear whether phosphorylation and/or OGlcNAcylation is a critical determinant in cancer. The p53 tumour suppressor, which is mutated or deleted in about 50% of all human cancers, is also O-GlcNAcylated. The DNA affinity of p53 as well as its activity are dependent on unmasking a basic region in the carboxyl terminus which normally acts as a repressor domain. Unmasking this region is dependent on O-GlcNAc-modification and so dynamic O-GlcNAcylation may well be important in regulating the affinity of p53 for DNA (69). O-GlcNAc-modification of Spl is regulated by levels of glucose and adenylate cyclase: reduced glucose and elevated cAMP cause hypoglycosylation of Spl and proteasome-mediated degradation (71). Glycosylation of Spl not only controls its stability in the cell but is also important in regulating protein-protein interaction (72). Also, Spl can be isolated or its function as a transcription factor can be blocked in vitro by binding to the lectin wheatgerm agglutinin (73). Indeed, lectin binding is a common method for isolating proteins modified with specific sugars (reviewed in 74 and see below). Glycosylation probably occurs as proteins are translated in the cytoplasm and glycosylation of Spl at a specific site may prevent it binding to other proteins until it is on the DNA. The O-GlcNAc-modification of serine and threonine was discovered almost 15 years ago (75), however, the literature on the role of O-GlcNAcylation of transcription factors is scant compared with that for phosphorylation. Thus development and application of suitable methods for the analysis of O-GlcNAc-modification has been pursued by comparatively few groups. Below I will give an overview of some of the methods that have been used. 3.1.1 Use of lectins Glycosylated proteins will bind to lectins and their use in associating with various different sugar moieties has been reviewed (74). Thus different lectins 286

11: Analysis of modifications are valuable as probes for different substituents; however, they have overlapping ligand specificity and so cannot be used to determine unequivocally which modification exists on the protein of interest. For example, wheatgerm agglutinin (WGA) binds both sialic acid and 0-GlcNAc residues. In spite of this, lectin binding has proved very useful. For example, WGA-agarose has been used to purify O-GlcNAc-modified Spl (73) and biotinylated WGA was used as a probe to identify 0-GlcNAc-Spl on Western blots (67), as described in Protocol 11. Protocol 11. Use of lectins Equipment and reagents • WGA-agarose column (Vector Labs) . BSA/TBST: 3% (w/v) BSA in TBST . Buffer 1: 50 mM Tris-HCI pH 7.5, 420 mM• 50 mM sodium citrate pH 5.0 KCI, 20% (v/v) glycerol, 10% (w/v) sucrose, • b-/V-acetylglucosaminidase 5 mM MgCI2, 0.1 mM EDTA, 1 mM PMSF, 1 • b-N-acetylglucosamine or a-methyl-D-manmM sodium metabisulfite, 2 mM DTT noside or . Buffer 2: 25 mM Hepes pH 7.6, 100 mM KCI, • Succinylated WGA or biotinylated con12.5 mM MgCI2, 20% (v/v) glycerol, 0.1% canavalin A (con A) (v/v) NP40, 10 nM ZnS04, 1 mM DTT • 10 mM sodium phosphate pH 7.4, 0.1% • GlcNAc (Sigma) (w/v) SDS, 0.5% (v/v) Triton X-100, 140 mM . TBST: 10 mM Tris pH 8.0, 150 mM NaCI, NaCI 0.05% (v/v) Tween-20 • Avidin-alkaline phosphatase or peroxidase

A. Isolation using a lectin column This method is as described in (73) for Sp1 purification, giving ~1% pure Sp1 after WGA-agarose chromatography, which can then be purified by binding to a DNA affinity column (yield ~4.5 ug/g HeLa cells). 1. Prepare a crude nuclear extract (e.g. according to ref. 76). 2. Apply to a WGA-agarose column (~135 mg protein extract/ml column) in buffer 1. 3. Wash twice with 5 ml buffer 1 and four times with 5 ml buffer 2. Elute with free 300 mM GlcNAc in the same buffer. B. Blotting using biotinylated lectin (67) 1. Separate proteins by SDS-PAGE and blot on to nitrocellulose. 2. Block in BSA/TBST. 3. As a control for non-specific binding, remove the O-GlcNAc from proteins on one strip of the membrane as follows: wash in 50 mM sodium citrate pH 5.0, incubate with 2.5 U/ml b-N-acetylglucosaminidase at 37°C for 24 h; wash in TBST, then BSA/TBST. 4. Incubate the membrane (with or without b-/V-acetylglucosaminidase treatment) with 10 ug/ml biotinylated WGA in TBST. Controls should include blocking biotin-WGA with 200 mM b-N-acetylglucosamine or

287

N. Shaun B. Thomas Protocol 11.

Continued

a-methyl-D-mannoside or probing with succinylated WGA or biotinylated concanavalin A (add 1 mM CaCI2 to all buffers when using con A). 5. Wash in 10 mM sodium phosphate pH 7.4, 0.1% (w/v) SDS, 0.5% (v/v) Triton X-100, 140 mM NaCI; incubate in BSA/TBST containing avidin-alkaline phosphatase or peroxidase; wash in TBST, then develop. C. Anti-O-GlcNAc Monoclonal antibodies which recognize O-GlcNAc have been raised (77) and the RL-2 antibody has been used to identify O-GlcNAc-Spl by Western blotting (see 71).

Lectins can also be used to test the potential function of O-GlcNAcylation in a number of in vitro assays, such as DNA footprinting (Chapter 2) or transcription (see ref. 67 and Chapter 9). 3.1.2 Radioactive labelling The methods for in vitro glycosylation of proteins are described in refs 78, 79 and are covered briefly in Protocol 12. Protocol 12. Radioactive galactosylation in vitro Glycosylation site(s) can be labelled in vitro by treating the glycosylated protein or protein lysates (see ref. 80) with [3H]galactose and galactosyltransferase. Equipment and reagents • Galactosyltransferase (bovine; Sigma) • Galactose buffer: 100 mM galactose, 50 mM Hepes pH 7.3, 150 mM NaCI, 50 mM . 25 mM Hepes pH 7.3, 5 mM MnCI2, 50% MnCI2 (v/v) glycerol . UDP-[6-3H]galactose (1-5 uCi, (37-185 kBq) 40Ci/mmol in 25 mM 5'-AMP)

Method 1. Autoglycosylate galactosyltransferase (75), concentrate by precipitation in 85% ammonium sulfate, and store at -20°C in 25 mM Hepes pH 7.3, 5 mM MnCI2, 50% (v/v) glycerol. 2. Prepare the protein of interesta or a cell extract. 3. Label the purified transcription factor or protein extract (100 ug) in 50 mU autoglycosylated galactosyltransferase, galactose buffer, and UDP-[6-3H]galactose. Incubate at 37°C for 30-60 min. The GlcNAc residues are modified with [3H]galactose.b

288

11: Analysis of modifications 4. Separate the [3H]galactose-labelled protein from unincorporated UDP[6-3H]galactose by chromatography through Sephadex G-50 or immunoprecipitation. The labelled peptide may then be determined as described briefly below. a Note that proteins produced by infecting insect cells with recombinant baculovirus or by translation in reticulocyte lysates are glycosylated. "Galactosyltransferase is active in 0.1% (w/v) SDS and 2% (v/v) Triton X-100 (79). Therefore high concentrations of detergents can be used for protein extraction which should then be diluted prior to labelling.

3.1.3 Determination of the modification and the site glycosylated The glycosylated transcription factors identified to date are O-GlcNAcylated and the methods described above will indicate whether this is also likely to be the case for the protein of interest. However, many different protein glycosylations are possible (see ref. 74) and below I will describe the methods which can be used to confirm O-GlcNAcylation or to determine which of the other modifications has occurred. Protocol 13. Determination of the type and site of glycosylation The type of glycosylated linkage (N- or O-linked) can be determined by enzymatic deglycosylation (see methods in 81, 82). 1. Perform carbohydrate analysis after digesting galactosyltransferaselabelled protein with PNGase F (material resistant to digestion is Olinked), separating the released glycans from the protein on a Sephadex G-50 column followed by alkali-induced b-elimination (81). 2. After desalting and size fractionation by FPLC, identify the saccharide(s) present by anion-exchange HPLC in comparison with known standards (see 80, 83). 3. Determine the site(s) of glycosylation as follows (for example see 68, 72):a separate peptides/glycopeptides derived by cyanogen bromide and enzyme proteolysis by reverse-phase HPLC and analyse 10% of each fraction by mass spectrometry. This allows peptides to be identified by their molecular masses and fractions containing peptides differing by 203 mass units potentially contain a GlcNAc residue. 4. To verify the presence of a GlcNAc-modified residue in such fractions, label by galactosyltransferase, as described in Protocol 12. Further verify the modified peptides by automated sequencing of [3H]Gal-labelled peptides. ° As for similar analysis of phosphorylated peptides the type of analysis described in brief here is specialized and requires expensive equipment and a detailed knowledge of the procedures being used.

289

N. Shaun B. Thomas 3.1.4 Functional assays Having determined the site(s) of glycosylation, the individual amino acid(s) can be changed (e.g. Ser —» Ala) by site-directed mutagenesis, as described for protein phosphorylation. Thereafter, the analysis of such modification(s) is the same as described previously in Section 2.7. However, individual amino acids have been shown to be phosphorylated or O-GlcNAcylated under different conditions and so caution should be exercised in claiming that any functional effects of mutating a particular residue are due solely to preventing glycosylation.

3.2 Other modifications There are other modifications that are important for controlling transcription: some factors require coordination with metal ions such as Zn2+ to become active (reviewed in ref. 84), and dynamic hyper-acetylation of histones and non-histone proteins occurs in active chromatin (reviewed in refs 85, 86). Since none of these are covalent modifications of transcription factors per se I will not deal with their analyses in this chapter. However, if neither phosphorylation nor O-GlcNAcylation can account for the modification of the particular transcription factor under study then it is worth considering whether the factor is modified with some other sugar or indeed whether it could be acetylated.

4. Conclusions The role of phosphorylation in regulating transcription factor activity has been studied for many years by a large number of groups and it is clear that this particular modification is crucial for controlling a wide range of cellular functions. The methods I have covered in brief in this chapter are well established and in general there are many simple methods for assaying phosphorylation changes which can be carried out by most laboratories equipped for cell and molecular biology techniques. In contrast, glycosylation of transcription factors has largely been ignored, even though the function of O-GlcNAcylation is important in regulating transcriptional activity, protein-protein interaction, as well as protein half-life. This may reflect the fact that there is a perception that glycosylation is restricted to receptors at the cell surface or to proteins in intracellular organelles and that the methods for analysing glycosylation are complex. I have described here that this need not be the case and I hope that by the time the next edition of this book appears the functions of transcription factor modification(s) other than phosphorylation will have been more widely investigated and that simple techniques will be in general use in non-specialist laboratories. 290

11: Analysis of modifications

Acknowledgements I should like to thank Sanjay Tiwari, Tony Best, and Stephanie Grenfell for providing reagents or data and Tony Best for critical comments on the manuscript. I should also like to thank Pam Roberts, Justin Hsuan, and Jim Woodgett for providing methods and for helpful suggestions. I am funded by the Kay Kendall Leukaemia Trust.

References 1. Tiwari, S., Jamal, R., and Thomas, N. S. B. (1996). In Protein phosphorylation in cell growth regulation (ed. M. J. Clemens), pp. 255-82. Academic Press, London. 2. Kitagawa, M., Higashi, H., Jung, H.-K., Suzuki-Takahashi, I., Ikeda, M., Tamai, K., et al. (1997). EMBO J. 15, 7060. 3. Lam, E. W.-F. and LaThangue, N. B. (1994). Curr. Opin. Cell Biol. 6, 859. 4. DePinho, R. A. (1998). Nature 391, 533. 5. LaThangue, N. B. (1994). Curr. Biol. 6, 443. 6. Peeper, D. S. and Bernards, R. (1997). FEBS Lett. 410, 11. 7. Nevins, J., Jakoi, L., and Leone, G. (1997). In Methods in enzymology (ed. W. G. Dunphy), Vol. 283, pp. 205-19. Academic Press, London. 8. Krek, W, Xu, G., and Livingston, D. M. (1995). Cell 83, 1149. 9. Williams, C. D., Watts, M., Linch, D. C., and Thomas, N. S. B. (1997). Blood 90, 194. 10. Corcoran, L. M. and Karvelas, M. (1994). Immunity 1, 635. 11. Dent, C. L., Lillycrop, K. A., Bybee, A., Latchman, D. S., and Thomas, N. S. B. (1991). J. Biol. Chem. 266, 20888. 12. Latchman, D. S. (1996). Int. J. Biochem. 28, 1081. 13. Robertson, M. (1988). Nature 336, 522. 14. Singh, H., Sen, R., Baltimore, D., and Sharp, P. A. (1986). Nature 319, 154. 15. He, X., Treacy, M. N., Simmons, D. M., Ingraham, H. A., Swanson, L. S., and Rosenfeld, M. G. (1989). Nature 340, 35. 16. Wirth, T., Staut, L., and Baltimore, D. (1997). Nature 329, 174. 17. Lillycrop, K. A. and Latchman, D. S. (1992). J. Biol. Chem. 267, 24960. 18. Falkner, F. G. and Zachau, H. G. (1984). Nature 310, 71. 19. Parslow, T. G., Blair, D. L., Murphy, W. J., and Granner, D. K. (1984). Proc. Natl Acad. Sci. USA 81, 2650. 20. Dent, C. L., Lillycrop, K. A., Estridge, J. K., Thomas, N. S. B., and Latchman, D. S. (1991). Mol. Cell Biol. 11, 3925. 21. Lillycrop, K. A., Dawson, S. J., Estridge, J. K., Gerstner, T., Matthias, P., and Latchman, D. S. (1994). Mol. Cell Biol. 14, 7633. 22. Lillycrop, K. A., Estridge, J. K., and Latchman, D. S. (1994). Biochem. J. 298, 245. 23. Segil, N., Roberts, S. B., and Heintz, N. (1991). Science 254, 1814. 24. Clipstone, N. A. and Crabtree, G. R. (1992). Nature 357, 695. 25. Roberts, S. B., Segil, N., and Heintz, N. (1991). Science 253, 1022. 26. Grenfell, S. J., Latchman, D. S., and Thomas, N. S. B. (1996). Biochem. J. 315, 889. 291

N. Shaun B. Thomas 27. Wang, J. and Prives, C. (1995). Nature 376, 88. 28. Williams, C. D., Linch, D. C., Sorensen, T. S., LaThangue, N. B., and Thomas, N. S. B. (1997). Brit. J. Haem. 96, 688. 29. Roberts, R. J., Khwaja, A., Lie, A. K. W., Bybee, A., Yong, K., Thomas, N. S. B., and Linch, D. C. (1994). Blood 84, 1064. 30. Laemmli, U. K. (1970). Nature 227, 681. 31. Harlow, E. and Lane, D. (1988). In Antibodies: a laboratory manual. CSH Press, NY. 32. Vario, G., Livingston, D. M., and Ginsberg, D. (1995). Genes Dev. 9, 869. 33. Fagan, R., Flint, K. J., and Jones, N. (1994). Cell 78, 799. 34. Bandara, R. L., Lam, E. W., Sorensen, T. S., Zamanian, M., Girling, R., and LaThangue, N. B. (1994). EMBO J. 13, 3104. 35. Gu, Y., Rosenblatt, J., and Morgan, D. O. (1992). EMBO J. 11, 3995. 36. O'Farrell, P. H. (1975). J. Biol. Chem. 250, 4007. 37. Gorg, A., Postel, W., and Gunther, S. (1988). Electrophoresis 9, 531. 38. Roberts, S. B., Segil, N., and Heintz, N. (1991). Science 253, 1022. 39. Rovera, G., Magarian, C., and Borun, T. W. (1978). Anal. Biochem. 85, 506. 40. Cannon, M., Schindler, D., and Davies, J. (1977). FEBS Lett. 75, 187. 41. Burke, L., Bybee, A., and Thomas, N. S. B. (1992). Oncogene 7, 783. 42. Cohen, P. (1989). Ann. Rev. Biochem. 58, 453. 43. Cohen, P. T. W. and Cohen, P. (1989). Biochem J. 260, 931. 44. Zhuo, S., Clemens, J. C., Hakes, D. J., Barford, D., and Dixon, J. E. (1993). /. Biol. Chem. 268, 17754. 44a. Thomas, N. S. B., Pizzey, A. R., Tiwari, S,, Williams, C. D., and Yang, J. (1998). J. Biol. Chem. 273, 23659. 45. Cockroft, S. (1984). Biochim. Biophys. Acta 795, 37. 46. Thomas, N. S. B. (1989). J. Biol. Chem. 264, 13697. 47. Kyriakis, J. M., Banerjee, P., Nikolakaki, E., Dai, T., Rubie, E. A., Ahmad, M. F., et al. (1994). Nature, 369, 156. 48. Lees, J. A., Buchkovich, K. J., Marshak, D. R., Anderson, C. W., and Harlow, E. (1991). EMBO J. 10, 4279. 49. Carrey, E. A. (1990). In Protein structure: a practical approach (ed. T. E. Creighton), pp. 117-44. IRL Press, Oxford. 50. Luo, K., Hurley, T. R., and Sefton, B. (1991). In Methods in enzymology (ed. T. Hunter and B. M. Sefton). Vol. 201, pp. 149-52. Academic Press, London. 51. Boyle, W. J., van der Geer, P., and Hunter, T. (1991). In Methods in enzymology (ed. T. Hunter and B. M. Sefton). Vol. 201, pp. 110-49. Academic Press, London. 52. Woodgett, J. R. (1992). In Cell lines in neurobiology: a practical approach (ed. J. Wood), pp. 133-59. IRL Press, Oxford. 53. Shuai, K., Stark, G. R., Kerr, I. M., and Darnell, J. E. (1993). Science 261, 1694. 54. Schagger, H. and vonJagow, G. (1987). Anal. Biochem. 166, 368. 55. Alberta, J. A. and Stiles, C. D. (1997). BioTechniques 23, 490. 56. Ginty, D. D., Kornhauser, J. M., Thompson, M. A., Bading, H., Mayo, K. E., Takahashi, J. S., and Greenberg, M. E. (1993). Science 260, 238. 57. Hagiwara, M., Brindle, P., Harootunian, A., Armstrong, R., Rivier, J., Vale, W., et al. (1993). Mol. Cell. Biol. 13, 4852. 58. Shieh, S. Y., Ikeda, M., Taya, Y., and Prives, C. (1997). Cell 91, 325. 292

11: Analysis of modifications 59. Verma, R., Annan, R. S., Huddleston, M. J., Carr, S. A., Reynard, G., and Deshaies, R. J. (1997). Science 278, 455. 60. Annan, R. S. and Carr, S. A. (1996). Anal. Chem. 68, 3413. 60a. Annan, R. S. and Carr, S. A. (1997). J. Protein Chem. 16, 391 61. Carr, S. A., Huddleston, M. J., and Annan, R. S. (1996). Anal. Biochem. 239, 180. 62. Wilm, M. and Mann, M. (1996). Anal. Chem. 68, 1. 63. Holt, G. D., Haltiwanger, R. S., Torres, C.-R., and Hart, G. W. (1987). J. Biol. Chem. 262, 14847. 64. Hart, G. W. (1997). Ann. Rev. Biochem. 66, 315. 65. Chou, T. Y., Hart, G. W., and Dang, C. V. (1995). J. Biol. Chem. 270, 18961. 66. Kearse, K. P. and Hart, G. W. (1991). Proc. Natl Acad. Sci. USA 88, 1701. 67. Jackson, S. P. and Tjian, R. (1988). Cell 55, 125. 68. Reason, A. J., Morris, H. R., Pacino, M., Marais, R., Treisman, R. H., Haltiwanger, et al. (1992). J. Biol. Chem. 267, 16911. 69. Shaw, P., Freeman, J., Bovey, R., and Iggo, R. (1996). Oncogene 12, 921. 70. Amati, B. and Land, H. (1994). Curr. Opin. Genet. Dev. 4, 102. 71. Han, I. and Kudlow, J. E. (1997). Mol. Cell Biol. 17, 2550. 72. Roos, M. D., Su, K, Baker, J. R., and Kudlow, J. E. (1997). Mol. Cell Biol. 17, 6472. 73. Jackson, S. P. and Tjian, R. (1989). Proc. Natl Acad. Sci. USA 86, 1781. 74. Hart, G. W., Haltiwanger, R. S., Holt, G. D., and Kelly, W. G. (1989). Ann. Rev. Biochem. 58, 841. 75. Torres, C.-R. and Hart, G. W. (1984). J. Biol. Chem. 259, 3308. 76. Dignam, J. D., Lebovitz, R. M., and Roeder, R. G. (1983). Nucleic Acids Res. 11, 1475. 77. Snow, C. M., Senior, A., and Gerace, L. (1987). J. Cell Biol. 104, 1143. 78. Holt, G. D. and Hart, G. W. (1986). J. Biol. Chem. 261, 8049. 79. Roquemore, E. P., Chou, T-Y., and Hart, G. W. (1994). In Methods in enzymology (ed. W. J. Lennarz and G. W. Hart). Vol. 230, pp. 443-60. Academic Press, London. 80. Haltiwanger, R. S. and Philipsberg, G. A. (1997). /. Biol. Chem. 272, 8752. 81. Lin, A. I., Philipsberg, G. A., and Haltiwanger, R. S. (1994). Glycobiology 4, 895. 82. Gribben, J. G., Devereux, S., Thomas, N. S. B., Keim, M., Jones, H. M., Goldstone, A. H., and Linch, D. C. (1990). Lancet 335, 434. 83. Roquemore, E. P., Chevrier, M. R., Cotter, R. J., and Hart, G. W. (1996). Biochemistry 35, 3578. 84. O'Halloran, T. V. (1993). Science 261, 715. 85. Grunstein, M. (1997). Nature 389, 349. 86. Wolffe, A. P. (1997). Nature 387, 16.

293

This page intentionally left blank

Al List of suppliers Amersham International plc, White Lion Road, Amersham, Bucks HP7 9LL, UK. Applied Biosystems, Birchwood Science Park North, Warrington, Cheshire WA3 7PB, UK. Beckman Instruments Inc., 2500 Harbor Blvd., Fullerton, CA 92634, USA. Betagen, Waltham, MO, USA. BIO101, Inc., 1070 Joshua Way, Vista, CA92083, USA. Bio-Rad Laboratories, 3300 Regatta Blvd, Richmond, CA 94804, USA. Boehringer-Mannheim, Sandhoferstrasse 116, Postfach 310120, D6800, Mannheim, Germany. British Drug House (BDH) Ltd., PO Box 8, Dagenham, Essex RMS 1RY, UK. Calbiochem, PO Boc 12087, San Diego, CA 92112, USA. Clontech laboratories, 4030 Fabian Way, Palo Alto, CA 94303, USA. Corning Costar Corp., One Alewife Center, Cambridge, MA 02140, USA. DuPont U.K. NEN Life Science Products, BRU/BRU/40349, PO Box 66, Hounslow, TW5 9RT. USA NEN Life Science Products, 549-3 Albany Street, Boston, MA 02118. Eastman-Kodak Co., 343 State Street, Bldg 701, Rochester, NY 14652, USA. Fisons, Bishop Meadow Road, Loughborough, Leics LE11 ORG, UK. Gibco/BRL Life Technologies Inc., 3175 Staley Road, Grand Island, NY 14072, USA. ICN Biomedicals Ltd, Unit 18, Thame Park Business Centre, Wenman Road, Thame, Oxfordshire OX9 3XA, UK. ICN Flow, 3300 Hyland Avenue, Costa Mesa, CA 92626, USA, ICN Pharmaceuticals, Inc., 3300 Hyland Ave, Costa Mesa, CA 92626, USA. Labconco, 8811 Prospect Avenue, Kansas City, MO 64132, USA. Nalgene U.K. Marathon Laboratory Supplies Ltd, Units 6-8, 55-57 Park Royal Road, London NW10 7LP. USA Nalge Nunc International, 75, Panorama Creek Drive, P.O. Box 20365, Rochester, NY 14602-0365. National Diagnostics, 305 Patton Drive, Atlanta, GA 30336, USA; Unit 4, Fleet Business Park, Mings Lane, Hessle, Hull HU13 9LX, UK.

List of suppliers New England Biolabs U.K. New England Biolabs (UK) Ltd, Knowl Pierce, Wilbury Way, Hitchin, Herts, SG4 0TY. USA New England Biolabs Inc, 32 Tozer Road, Beverly, MA 01915-5599. Nunc, 2000 North Aurora Road, Naperville, IL 60653, USA. Perkin Elmer Cetus, 761 Main Avenue, Norwalk, CT 08659, USA. Pharmacia/LKB, PO Box 175, Bjorkgarten 30, 75182 Uppsala, Sweden; Piscataway, NJ, USA. Pierce, 3747 North Meridian Road, PO Box 117, Rockford, IL 61105, USA. Promega Biotec, 2800 Woods Hollow Road, Madison, WI53711, USA. Schleicher and Schuell, 10 Optical Avenue, Keene, NH 03431, USA. Sigma Chemical Co., PO Box 14509, St. Louis, MI 63178, USA. Spectrum, Medicell International Ltd, 239 Liverpool Road, London Nl 1LX, UK; 23022 La Cadena Drive, Suite 100, Laguna Hills, CA 92653, USA. Stratagene, 11099 North Torrey Pines Road, La Jolla, CA 92037, USA. Worthington Biochemical Corp., Halls Mill Road, Freehold, NJ 07728, USA. Worthington, Lome Laboratories, 7 Tavistock Estate, Ruscombe Business Park, Ruscombe Lane, Twyford, Reading, Berkshire RG10 9NJ, UK; Worthington Biochemical Corporation, Halls Mill Road, Freehold, NJ 07728, USA.

296

Index ABCD (avidin-biotin complex on DNA) assay, protein-protein interactions 209-11 adenovirus major-late promoter, and in vitro transcription 222, 224-5 affinity chromatography, method 110-11 alkylation of DNA with dimethyl sulphate (DMS), method 40-1 amino acids, degeneracy of codons 149-50 amplification of DNA from bacterial colonies, method 156-7 antibodies characterization of DNA-binding proteins 19-21 detection of DNA-transcription factor complexes 171 method for binding 21 method for production 135 phosphorylation-site-specific 283 for screening bacteriophage expression libraries 134–6 specificity 20–1 bacteriophage expression libraries, see cDNA expression libraries baculovirus, expression of cloned transcription factors 17, 194, 195 biotinylated oligonucleotides, concentration of DNA-binding proteins 108, 111 blotting peptides to PVDF membranes, method 116-17 blunt-end ligation, cloning of PCR products 155–6 Bradford assay for protein concentration, method 13 brain in vitro transcription assay 215-17, 222-6 nuclear extracts from 215-6, 217-22 bromodeoxyuridine-substituted DNA, crosslinking with ultraviolet light 77-80 calcium phosphate-mediated transfection 198-200 cations, requirement for binding of transcription factors 21 cDNA expression libraries cloning by complementation 138–9 cloning transcription factors 123–6 expression in eukaryotic cells 137–8 library selection 124 plaque purification 126

plating 125–6 screening methods 126–39 use in polymerase chain reaction (PCR) 153 cDNA and genomic libraries, method for lowstringency hybridization 147-8 cDNA isolation, design of oligonucleotides 118–21 cell cycle control 261-3 cell membranes, permeabilization 49 chloramphenicol acetyl transferase (CAT), assay 201-2 chromatin structure 229-30 transcription from 217, 229–59 cloning cDNA expression libraries 123–6, 138–9 methods for PCR products 155–6 target genes for transcription factors 174–7 transcription factors 145–64 c-Myc, protein modification 286 CNBr, cleavage of proteins to peptides 114–15 CNBr-activated Sepharose 4B, for DNA affinity chromatography 104–5,108–9 codons, degeneracy 149–50 cofactors, DNA-binding proteins 21 co-immunoprecipitation, analysis of interactions between transcription factors 92–4 cooperative binding, analysis 86–7 COUP-TF, interaction with non-DNA-binding transcription factor S300-II 88-9 CpG islands enrichment in genes 169 genomic binding-site cloning 173–4 cross-linking of transcription factors 85–6, 94 transcription factors to DNA 78–80, 81–2 CTCF factor, purification 98,105–8,112 dephosphorylation of proteins, method 269-70 dimerization, transcription factors 84–6 dimethylpimelidate (DMP), cross-linking agent 86 dimethyl sulphate (DMS) interference assay 42-3 modification of DNA 39, 40–1, 48–51, 53 dithiobis(succinimylpropionate)(DSP), protein cross-linker 93–4 DNA cross-linking to transcription factors 77–82 dephosphorylation methods 8

Index DNA (cent.) end labelling methods 7–8, 33–4, 120, 246–7 in vitro modification of DN A with DMS 53 packing into chromatin 229–30 piperidine cleavage 53-5 purification 51–3 supercoiling and promoter activity 226–7 DNA affinity chromatography, purification of DNA-binding proteins 104–11 DNAase I footprinting 37–8, 84 binding of transcription factors to nucleosomes 255 non-DNA-binding transcription factors 88 see also footprint analysis DNA-binding activity, mobility shift assay 1–25, 202, 204–6 DNA-binding assays, preparation of 32Plabelled oligonucleotides 204–5 DNA-binding proteins characterization 19-23 cloning 123–43 cofactors 21 concentration with biotinylated oligonucleotides 108, 111 detection 2-14 DNA mobility shift assay 1-25 DNA sequence specificity 3-5 guanidinium chloride denaturation and renaturation 132–4 identification 4,19-23, 111 interactions with other proteins 23–4 molecular weight determination 2–3,17, 98 production of peptides from 114–17 proteolytic clipping band-shift assay 21-3 purification 17, 97—114 requirement for cations 21 see also transcription factors DNA-binding specificity transcription factors 17-19 use of competitor DNA 3–5,18–20 DNA-cellulose chromatography, purification of DNA-binding proteins 103–4 DNA fragments, nitrocellulose binding assay 176 DNA labelling with DNA polymerase 8–10 end labelling 28–9, 32–4 single strand 29-30 DNA mobility shift assay analysis of tertiary structure 84–5 applications 2–6 binding reaction 12-13 binding of transcription factors to nucleosomes 255 for DNA-binding proteins 1–25, 202, 204–6 identification of transcription factors 4–6

limitations 6 measurement of transcription factor levels 4–5 preparation of labelled oligonucleotide probes 7–8 principle 2–4 selection of DN A probe 6–7 sequence specificity 3–5,18–20,128,135–7, 138, 141 DNA modification dimethyl sulphate (DMS) 39, 40–1, 48–51 ethylnitrosourea (ENU) 40 in vivo 48–51 DNA probes, selection for DNA mobility shift assay 6–7 DNA-protein complexes footprint analysis 27–62 see also DNA-binding proteins DNA-protein interactions, determination of affinity 205–6 Drosophila embryo, nuclear extracts 233 E2F family, transcription factors 261-2, 264–5 electro-mobility shift assay (EMSA) see DNA mobility shift assay electroporation, for transfection of transcription factors 198-9 electroporation of COS-1 cells, method 194 enhancer activity, genomic DNA fragments 176–7 enzymes introduction into cells 49 production of peptides from DNA-binding proteins 115 ethylnitrosourea (ENU), DNA modification 40 expression vectors, organization 192-3 far-Western approach, screening cDNA expression libraries 137 ferrous EDTA, use in footprint analysis 38–9 footprint analysis of binding sites on closed circular plasmids 31, 43–7 of binding sites on end-labelled fragments 29–13 choice of linear DNA molecules 30–4 DNA-protein complexes 27-62 interference assays 27–9, 39–43, 84 in vitro 29–47 in vivo 47–61 ligation-mediated PCR (LMPCR) 32-3, 34–5, 47–9 method 45–7, 59–61 protection analysis 27–9, 37–9

298

Index protection from DNase I digestion 37–8,84 protein sources for binding 34–7 fragment probes, labelling of 8–10

homeobox proteins, properties 161 hydroxyapatite, purification of histones 236, 2389,241

fj-galactosidase, assay 2001 GATA1 factor purification 98, 101, 103, 105–6 quantity in cells 97 gel filtration chromatography molecular weight determination 69-72 properties of matrices 70–1 gel mobility shift assay analysis of protein-protein interactions 209, 210 non-DNA-binding transcription factors 88, 90–1 gel retardation assay, see DNA mobility shift assay; gel mobility shift assay gene promoters, TAATGARAT sequence 1–2 genes, enrichment in CpG islands 169 genome sequences, databases 160,162 genomic binding-site cloning cloning procedure 172-3 concentration of DNA-protein complex 170–1 identification of target genes 165-79 preparation of transcription factor protein 167-9 principle 165-7 source of DNA 169–70 use of CpG islands 169,173–4 G-free cassette assay, for transcription 64–6, 68 glycerol gradient centrifugation, molecular weight determination 73–4 glycosylation, transcription factors 285–90 GST-fusion proteins method for expression 190–1 purification 191 GST pull-down assay, analysis of protein-protein interactions 206–7 guanidinium chloride, denaturation and renaturation of DNA-binding proteins 132–4

immunoprecipitation analysis of protein-protein interactions 208–9 detection of DNA-transcription factor complexes 171 insect cells, overexpression of cloned transcription factors 194 in vitro transcription and adenovirus major-late promoter 222, 224–5 method 68, 223–4 and neurofilament (NF) promoter 222, 224–5, 227 and promoter templates 226–7 in vitro transcription and translation, method 188

HeLa cells, preparation of nuclear extracts 66–7 helix-loop-helix proteins, properties 161 herpes simplex virus 1 (HSV-1) 1–2,23–4 histone acetylation, effects on transcription 257,262 histones packing of DNA into chromatin 229–30 preparation 236–42 and transcription 230, 257, 262

lectins, analysis of protein glycosylation 286–8 library plating and replica lifts, method 125–6 ligation-mediated polymerase chain reaction (LMPCR) method 56–9 preparation of DNA for 53-5 preparation of linkers 55–6 resolution 56–61 for in vivo footprinting 32–3,34–5,47–9 linear DNA, method for protection from DNase I 37–8 lipofection, for transfection of transcription factors 198–9 luciferase, assay 200–1 lysogen, generation from purified phage stock 139–40 mammalian cells, expression of transcription factors 15–16, 192–4, 195 mass spectrometry analysis of protein phosphorylation 284-5 identification of protein sequences 97 measurement of mass of peptides 115 method for preparing proteins 119-20 methidium-propyl EOT A, use in footprint analysis 38 microbore reverse-phase chromatography, separation of peptides 114 micrococcal nuclease, analysis of reconstituted nucleosomes 253-5 minichromosomes, preparation of nucleosome arrays 234

299

Index molecular weight determination calibration 69-70,73-4 denatured transcription factors 76-83 DNA-binding proteins 2-3, 98 methods 3,17,69-77 transcription factors 69-76,98 mononucleosome templates analysis of transcription elongation 231 analysis of transcription initiation 231 binding of transcription factors 255-7 position of transcription factor binding site 243^1 preparation 232-3,234-6,243-55 transcription 230-2 myelin, effect on isolation of nuclei 219-20 neurofilament (NF) promoter, and in vitro transcription 222,224-5,227 non-denaturing gradient electrophoresis, molecular weight determination 75-6 non-DNA-binding transcription factors DNAase footprinting 88 gel mobility shift analysis 88, 90—1

restoration of activity 89 S300-II 88-9 'supershift' gel mobility shift analysis 90-1 Northern blot analysis, transcription factor binding sites 175 nuclear extracts Drosophila embryo 233 preparation 66-7,99-103,215-6,217-22 Xenopus 233 nuclear proteins extraction from rat brain 220-2 methods of preparation 11,36,67,101—3 nuclease PI, use in footprint analysis 38 nuclei effect of myelin on isolation from brain 219-20 isolation from rat brain 217—20 preparation from whole tissue 102-3 nucleoplasmin, reconstitution of nucleosomes 235-6 nucleosome arrays, see nucleosome templates nucleosome core particles, preparation method 250-2 nucleosomes characterization of reconstituted 252-5 packing of DNA into chromatin 229-30 reconstitution 247-9 and transcription 230 nucleosome templates analysis of transcription elongation 232 analysis of transcription initiation 231 binding of transcription factors 255-7

essential components 234 preparation 232-4 octamer family, transcription factors 1—2,20, 23–1, 261-3, 266, 268-9, 273 oestrogen receptor 165, 177-8, 181-2, 208-9 DNA-binding domain 167-8, 170, 172 interaction with steroid receptor coactivator protein SRC1 a 207 in vitro transcription and translation 187-9 ligand-binding domain 189-90 transcriptional activation by 196-7 zinc fingers 167-8 oestrogen receptor binding site, nucleotide sequence 202,204 oestrogen response element, binding site 197 oestrogen-responsive genes, isolation 165, 172,177-8 oligonucleotides design for cDNA isolation 118-21 for DNA mobility shift assay 7-8 method for hybridization to library filters 120-1 for screening libraries 114,130-1 orthologous genes, isolation from phylogenetically diverse organisms 145 p53 tumour suppressor, glycosylation 286 PCR, see polymerase chain reaction (PCR) PCR deletion analysis, method 184 PCR products cloning 155-6 method for direct sequencing 157-8 peptide mapping for phosphorylation sites, method 277-8 peptides measurement of mass by mass spectroscopy 115 preparation methods 114-17 peptide tags, for purification of proteins 185 phenanthroline-copper, use in footprint analysis 38-9 phosphatases, use in analysis of protein phosphorylation 267-70 phosphoamino acid analysis, method 280 phosphorylation retinoblastoma (pRb) protein 261-2 synthetic peptides 282-3 transcription factors 261-85 piperidine, cleavage of modified DNA 48, 53-5 plasmids, footprint analysis 43-7 polymerase chain reaction (PCR) annealing temperature 149,151 artefacts 158-60 choice of template 151,153-5 300

Index cloning of products 155-6 and cloning transcription factors 148-60 'in-and-out'PCR 153 methods for screening libraries 154-5 preparation of deletion mutants of transcription factors 182-5 preparation of point mutations of transcription factors 185-7 reaction conditions 151 selection of primers 148-9,150,152,182-3, 185 sequencing of products 155-8 standard method 153-4 'touchdown PCR' 151 whole genome PCR method 171 POU-domain proteins, properties 161 progesterone receptors assay 64-6 cooperative binding to progesteroneresponse elements (PREs) 87 progesterone-response elements (PREs) cooperative binding of progesterone receptors 87 synergistic progesterone inducibility 87 promoters TAATGARAT sequence 1-2 for transcription assays 64 promoter templates and in vitro transcription 226-7 and supercoiling of DNA 226-7 protein binding to DNA, method 12 protein concentration, measurement 13 protein extraction, method using lysogen 140-1 protein glycosylation determination of type and site of glycosylation 289 methods of analysis 286-90 protein phosphorylation in vitro phosphorylation 281-3 labelling with radioactive phosphorus 270-4 mapping phosphorylation sites 274-81 methods of analysis 263-70,283-5 protein preparation homogenization of tissue 11-12 mini-extracts 13-14 nuclear extracts 10-11 whole-cell extracts 10-12, 13-14 proteins bacterial synthesis 15 measurement of concentration 13 method for labelling with [32P]orthophosphate 271-2 recovery and digestion from a gel slice 276-7 protein sequences, identification using mass spectrometry 97

proteolytic clipping band-shift assay, DNAbinding proteins 21-3 reporter genes assays 200-2 structure 196-8 for study of activity of transcription factors 196-8 retinoblastoma (pRb) protein phosphorylation 261-2, 278-9 preparation of radioactive 274 retinoic acid receptors, DNA-binding proteins 21 reverse-phase chromatography purification of DNA-binding proteins 111-14 separation of peptides 114 RNA, method for reverse transcription for PCR analysis 159-60 RNA polymerases, for in vitro transcription 188 S300-II interaction with DNA-binding factor COUP-TF 88-9 non-DNA-binding transcription factors 88-9 Saccharomyces cerevisiae, percentage of genes that are transcription factors 178 SDS-polyacrylamide gel electrophoresis analysis of protein phosphorylation 263-6 identification of DNA-binding proteins 111 molecular weight determination 76-7, 78-80,82-3 separation of peptides 114, 115 sequence databases, identification of transcription factors 160, 162-3 serum response factor (SRF), glycosylation 286 Southwestern blotting, for screening cDNA expression libraries 128-30 Sox genes, identification from database 160, 162-3 steroid receptor coactivator protein SRCla 181-2 deletion analysis 185 interaction with oestrogen receptor 207 transactivation domain 202, 203 steroid receptors, DNA-binding proteins 21 supercoiling, and promoter activity of DNA 226-7 'supershift' gel mobility shift analysis, nonDNA-binding transcription factors 90-1

301

Index T4 DNA polymerase, DNA labelling 9-10 T4 polynucleotide kinase, end labelling of DNA 7-8, 32-4 TAATGARAT sequence gene promoters 1-2 and immediate-early gene transcription 1-2 protein binding 1-2, 23-4 target genes binding activity 176-7 molecular cloning 174-7 Northern blot analysis 175 sequencing 175-6 zoo blotting 174-5 thermal cycle sequencing, method 44-5 tissues, homogenization for protein extraction 11-12 transactivation domains identification using chimeric proteins 202, 203 steroid receptor coactivator protein (SRCla) 202, 203 transcription effect of chromatin structure 217, 229-32 effect of histone acetylation 257, 262 from chromatin 229-59 andhistones 230 in vitro assays 63-8, 215-17, 222-6 and nucleosomes 230-2 transcriptional machinery, components of 66 transcription assays choice of promoters 64 G-free cassette assay 64-6 in vitro 63-8, 215-17, 222-6 transcription elongation analysis using mononucleosome templates 231 analysis using nucleosome arrays 232 transcription factor binding sites binding activity 176 methods of isolation 172-4 Northern blot analysis 175 position on mononucleosome template 243-4 sequencing 175-6 zoo blotting 174-5 transcription factor-DNA complexes, analysis of dissociation 91-2 transcription factors analysis of 181-213 analysis of tertiary structure 84-7 analysis using gel retardation assay 205-6 Apl 23 binding to nucleosome templates 255-7 cloning 15-16,123–13, 145-64 cross-linking of dimers 85-6 cross-linking to DNA 77-82 dephosphorylation in vitro 266-70 dimerization 84-6

DNA-binding specificity 17-19 E2F family 261-2,264-5 expression in bacteria 14-15,189-91,195 expression in baculovirus 17,194,195 expression by in vitro transcription and translation 16-17,187-9,195 expression in mammalian cells 192-5 extraction from nuclei 100-3 glycosylation 285-90 see also protein glycosylation identification 4-6,75-6,160,162-3 identification of conserved domains 182 identification of target genes 165-79 interactions with other proteins 23-4,92-4, 206-12 ABCD (avidin-biotin complex on DNA) assay 209-11 analysis using immunoprecipitation 92-4, 208-9 DNA mobility shift assay 209 GST pull-down assay 206-7 two-hybrid analysis 211-12 in vitro transcription and translation 187-9 mapping of functional domains 182-7 method for expression and purification 168-9 molecular weight determination 69-83, 98 non-DNA-binding 88-94 octamer (Oct) family 1-2, 20, 23-4, 261-3, 266, 268-9, 273 overexpression 193-5 percentage of genes in yeast 178 phosphorylation 261-85,290 see also protein phosphorylation preparation of deletion mutants 182-5 preparation of point mutations 185-7 proving identity 139-42 purification 17,97-114 quantity in cells 97 renaturation from SDS-polyacrylamide gels 82-3 sources 10-17 subunit composition 84-7 use of cDNA clones 15 use of transfection for study of function 196, 197-200 see also DNA-binding proteins; non-DNAbinding transcription factors transcription factor Spl glycosylation 286 requirement for zinc 21 transcription initiation, analysis using nucleosome arrays 231 transcription in vitro, RNA polymerases for 188 transfection controls for efficiency 197-8 of mammalian cells 196, 197-200

302

Index methods 198-200 for study of function of transcription factors 196, 197-200 transient transfection, calcium phosphate coprecipitation 199-200 T-tailed vectors, cloning of PCR products 155-6 two-dimensional electrophoresis, analysis of protein phosphorylation 266 two-hybrid analysis, protein-protein interactions 211-12

Western blotting analysis of protein phosphorylation 264-6 identification of transcription factors 75-6 whole cell extract, method of preparation 195 whole genome PCR 171 Xenopus, nuclear extracts 233 yeast, overexpression of cloned transcription factors 194

U2 small nuclear RNA enhancer, cooperative binding to 7 ultraviolet light, binding transcription factors toDNA 77-82 viral component Vmw65 1-2,23

zinc-finger proteins oestrogen receptor 167-8 properties 161 zoo blotting, transcription factor binding sites 174-5

303