Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Structure-function analysis of the MAT[Alpha]2 protein of Saccharomyces cerevisiae Ho, Chi-Yip 1993

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1994-89427x.pdf [ 4.48MB ]
JSON: 831-1.0088291.json
JSON-LD: 831-1.0088291-ld.json
RDF/XML (Pretty): 831-1.0088291-rdf.xml
RDF/JSON: 831-1.0088291-rdf.json
Turtle: 831-1.0088291-turtle.txt
N-Triples: 831-1.0088291-rdf-ntriples.txt
Original Record: 831-1.0088291-source.json
Full Text

Full Text

STRUCTURE-FUNCTION ANALYSIS OF THE MATh2 PROTEIN OF SACCHAROMYCES CEREVISIAE  /  by CR1-YIP HO B.Sc., Concordia University, 1986 A THESIS SUBMITTED iN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR TUE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF BIOCHEMISTRY AND MOLECULAR BIOLOGY  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA January 1994 © Chi-Yip Ho, 1993  ____  in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my In presenting this thesis  department  or  by  his  or  her  representatives.  It  is  understood  that  copying  or  publication of this thesis for financial gain shall not be allowed without my written permission.  (Signature)  Department of  Clr&  The University of British Columbia Vancouver, Canada  Date  DE.6 (2/88)  MO1e(ciir  (QfOjy  11  ABSTRACT  In Saccharomyces cerevisiae, the M4 Tc2 gene encodes a negative regulatory protein responsible for determining cell-specific mating-type. The !vL4 To2 protein, x2, represses two sets of mating-type-specific genes by interacting with two other factors. c2 forms a heterodimer with the repressor protein al, and binds a DNA-target site located upstream of all haploid-specific genes, thereby repressing these genes.  In  combination with a second factor, MCM1, cc2 recognizes another DNA-target site found in all the a-specific genes. The binding of cL2-MCM1 to these sites leads to the repression of a-specific genes. Therefore, x2 carries out a dual regulatory function in mating-type control through protein-protein interactions with its co repressors and through the changes in DNA-binding specificities upon interactions with these partners. The objective of this thesis is to gain further insight (i) into the heterodimerization of c2 with al and (ii) into DNA-binding specificity of c2 when acting with MCMI. In studies to localize the important structures within o.2 for functional al-cc2 heterodimerization, two sequences in the N-terminal half of cL2, containing the 3,4-hydrophobic heptad repeat pattern characteristic of coiled-coils, were identified and examined. Amino acid replacements showed that particular positions within the heptads are important in a1-c2 repression and in stabilizing the a1-c2 heterodimer.  These  mutational analyses are consistent with the proposal that the two heptad sequences of x2 associate with al in a coiled-coil-like manner. To determine the DNA-binding specificity of cc2 in the presence of MCM1, a variant of the wild-type x2 operator was used as the DNA-target in an in vitro electrophoretic band-shift assay. This variant cv2 operator was also used in a heterologous reporter gene system to test the in vivo repression mediated by i2-MCM1. The results together indicate that the nucleotide differences in the variant operator caused a decrease in both x2-MCM1 binding and repression.  This suggests that nucleotides outside the core  consensus sequence of the cL2 operator are important for its in vivo recognition by c2-MCM1.  111  TABLE OF CONTENTS  Abstract  ii  List of Tables  x  List of Figures  xi  Acknowledgements  xiii  List of Abbreviations  XiV  Introduction  1  1.1.  1  Cell-Type-Specific Control Regulated by M4 T Products 1.1.1.  Control of Mating-Type by MAT Loci  1  1.1.2.  Combinatorial Regulations by MAT Products  1  TheRoleofM4Tc1  3  The Dual Function of the MA Tcj.2 Product  3  Regulatory Mechanism and DNA-Binding Specificity of the MAT Products  4  Requirement of the MCM1 Product for al and o2 Fuction  4  Binding of cd-MCM1 to the cL-specific Genes and Transcriptional Activation  5  Binding of cL2-MCM1 to the a-specific Genes in c2 Repression  7  Binding of al-cL2 to the Haploid-Specific Genes  10  1.1.4.  MCM1 and STE12 in the Pheromone Response Pathway  12  1.1.5.  Regulation of the Expression ofMAT Locus  13  1.1.6.  Yeast Cell-type Specific Transcription as a Model System  14  1.1.3.  1,2,  Transcriptional Regulation in Eukaryotes  14  1.2.1.  A Common Theme in Eukaryotic Transcription  14  1.2.2.  Cis-Elements on Promoters  15  1.2.3.  Trans-Acting Factors  17  1.2.4.  Positive and Negative Regulations  18  iv  1.2.5.  Structure-Function of Transcriptional Factors The Modular Structure Concept  DNA-BindingMotifs  21  Activation Motifs  24  Oligomerization Motif  25  Modular Structures  25  Structure-Function Aspects of the x2 Protein  26  1.3.1.  Overall Strucure of the x2 Protein  26  1.3.2.  Structure-Functional Analysis of x2  28  Nuclear Targeting  28  Rapid Selective Protein Degradation  28  Separate Domains for Dimerization and DNA Binding  29  1.3.3.  1.4.  1.5.  Interaction with MCM1 by the Linker Regions of c2  30  Random Mutagenesis on the Homeodomain  30  DNA-Binding Homeodomain of the cL2 Repressor  32  1.4.1.  c2 Homeodomain  32  1.4.2.  DNA-Binding Specificity of x2 Homeodomain  33  Putative Coiled-Coil Dimerization Motifs of x2 1.5.1.  1.5.2.  Requirement of the N-terminal Domain for Protein-Protein Interactions  1.5.3.  40 40  cQHomodimers  40  a1-c2Heterodimers  41  Putative Coiled-Coil Motifs  1.6.  20 1.3.  20  3,4-Hydrophobic Heptad Repeats in cL2 and al  Coiled-Coil (Leucine Zipper-Like) Motifs  Objectives of Thesis Project  42 42 45 52  1.6.1.  Characterization of the Dimerization Domain of cL2  52  1.6.2.  Study the DNA-Binding Specificity of cL2  53  V  Materials and Methods  54  2.1.  54  Reagents 2.1.1. Enzymes  54  2.1.2. Nucleotides  54  2.1.3. Oligonucleotides  54  2.1.4. Autogradiography Materials  54  2.1.5. Purified cL2 Protein from E. coil  58  2.1.6. Antibodies  58  2.1.7. In vitro Translation Materials  58  2.1.8. Media Components  58  2.2. Bacterial and Yeast Strains  58  2.3. Media and growth conditions  59  2.4. Transformation of E. coil  61  2.4.1. Calcium Chloride Method  61  2.4,2. Electroporation Method  62  2.5. Transformation of S. cerevisiae  62  2,6. Techniques for DNA Manipulation and Analysis  63  2.6.1. Restriction Digests and DNA Modification  63  2.6.2. Gel Electrophoresis  63  2.6.3. Isolation of DNA Fragments  63  2.6.4. Ligations  63  2.6.5. DNA Sequence Determination  64  2.6.6. Isolation of Plasmids and Bacterial Phage DNA  64 Isolation of Plasmid DNA from E.coli  64  2.6,6.2. Isolation of Single-Stranded DNA  64  2,6.7. In vitro Oligodeoxyribonucleotide-Directed Site-Specific Mutagenesis  65  2.6.8. Polymerase Chain Reaction  65  vi  2.7. In vitro Transcription and Translation  68  2.7.1. In vitro Transcription  68  2.7.2. In vitro Translation  68  2.8. Isolation and Analysis of Yeast RNA  69  2.8.1. Isolation of Yeast RNA  69  2.8.2. Primer Extension Analysis of Yeast RNA  69  2.9. Preparation of Crude Yeast Extracts  70  2.10. SDS Polyacrylamide Gel Electrophoresis (SDS-PAGE) of Proteins  70  2.10.1. Fluorography 2.11. Yeast Functional Assays  71 71  2.11.1. Mating Assay  71  2.11.2. Sporulation  71  2.11.3. -Galactosidase Assays  72  2.11.4. Growth Rate Determination  72  2.12. In vitro DNA-Binding Assays  72  2.12.1. DNA-Binding Gel Electrophoresis Mobility-Shift Assay  72  2.12.2. DNA-Binding Band-Shift Assay  73  2.12.3. DNase I-Protection Footprinting Assay  73  2.13. In vitro Immunological Assays  74  2.13.1. Affinity Purification of Rabbit Anti-MATx2 Serum  74  2.13.2. Co-Immunoprecipitation of in vitro Translated Proteins  74  2.13.3. Electrophoresis and Western Blot Analysis  75  2.14. Plasmids  75  2.14.1.pE162  75  2.14.2. Plasmid pLG669-Z and the YpSN Series  76  2.14.3. CYC]: URA3 and BAR]: URA3 Reporter Plasmids  76 Plasmid YHE36  76  vii PlasmidsYHE38toYHE4l  76 Plasmids YHE43 to YHE46  80  2. 14.3.4. Plasmid YHE62  80  2. 14.3.5. Plasmids of YFIE64 and YHE65  80  2.14.4. Construction of YRP-MATx2 Plasmid and its Derivatives  81  2.14.5. E. coil Expression Plasmid pLacMATcL2  81  2.14.6. Construction of pEMBLMATc-(BamHI-BciJ)  81  2.14.7. Construction of pEMBL-Yr-MATc  82  2.14.8. Construction of pGEM3Zf(+)-MATcL2  82  2.15. Construction ofM4Tx2 Mutants 2.15.1. Mutations Generated by in vitro Oligonucleotide-Directed Site-Specific Mutagenesis  82 82 Generation of Single or Multiple Replacements  84 Screening of C33F and C33Y matcL2 Mutants  84 Screening of C34F1L291 and C34F matcx2 Mutants  84  2.15.2 Mutations Generated by PCR Amplification  85  Results and Discussion (Chapter 1)  86  3.1. Mutational Analysis of the MATa1-oe.2 Dimerization Domain  86  3.1.1. Identification of Two Hydrophobic Heptad Repeat Motifs: x2A and cc2B  86  3.1.2. Functional Assays for a1-c2 Dimerization  86  3.1.3. Construction and Analysis of the Internal Deletion Mutants of IvL4 Tc2  88  (i) Construction of the u2A and x2B Deletion Mutants of MA Tci2  88  (ii) Analysis of the Effect of the Internal Deletions  89  3.1.4. Production and Functional Analyses of Multiple Replacements in cc2A and c2B  89  (i) Construction of Mutiple Argimne or Alamne Replacements  89  (ii) The Effect of Multiple Replacements on a1-o2 Repression  92  3.1.5. Studies on the Function of x2A and cL2B in a 1-x2 Heterodimerization Construction of c2A and cL2B Mutant cL2 Genes  92 92  viii Determining the Effect of Heptad Mutations on the Dominant Negative matcv2t4-7 Allele  94 Mutational Analysis on the dPositions of x2B  94 Mutational Analysis of cQA  100  (i) Investigation of Funtional Significance of Cys33 and Cys34  100  (ii) The Role of Leucine Residues in a Positions of x2A  100  3.2. Discussion on the Significance of cL2A and x2B in Dimerization  103  3.3. Implication of the Mutational Studies on the Dimerization Domain  106  3.3.1. Models on the a1-c2 Heterodimers  106  3.3.2. Mechanism to Define the al-cL2 Heterodimenzation Specificity  107  3.3.3. Functional Relationship of Putative Coiled-Coil Motifs and Homeodomain Proteins  110 Evidence for Functional Coiled-Coil Motifs in Homeodomain Proteins  110 Position of Leucine-Zipper-like Motifs with Respect to Homeodomains  110  3.4. Future Prospectives  111  3.4.1. Evaluation of the al-x2 Model  111  3.4.2. Further Study on c42A and x2B of cL2  112  3.4.3. Competition Study on al-ct2 Heterodimer  114  Results and Discussion (Chapter 2)  116  4.1. Characterization of x2-MCM1 DNA-Binding Specificity  116  4.1.1. In vitro a2-MCM1 DNA-Binding  116 E. coil-Expressed c2 Proteins  117 DNA-Binding Band-Shift Assay of the cL2 Proteins  117 DNase I-Protection Footprinting with MCM1 and cL2  120  4.1.3. c2-MCM1 DNA Binding to the pseudo BARd Operator  121  4.1.4. In vivo Function of the pseudo BAR 1 Operator  121  4.2. Discussion on x2-MCMI Cooperative DNA Binding  133  4.3. Conclusions and Future Prospectives  137  ix  Appendix A  139  5.1. In vitro a1-c2 DNA-Binding of Missense Substituted cL2 Mutants  139  5.1.1. Missense Mutions in the c2B Coiled-Coil Dimerization Motif  139  5.1.2. Serine Replacement of Leu196 of c2  141  Appendix B  143  5.2. BAR]: URA3 Reporter System for Selecting x2 Function  143  References  151  x  LIST OF TABLES  Table I.  List of Oligonucleotides  55  Table II.  £ cerevisiae Strains  60  Table III.  The Effect of Acc2A and iS.cx2B Mutations on al-cL2 Repressor Function  Table IV.  90  The Effect of Multiple Heptad Mutations on al-0L2 Repressor Function  93  Table V.  Summary on Functional Analysis of c2B Mutations  95  Table VI.  Summary on Functional Analysis of x2A Mutations  102  Table VII.  Physiological Functions ofADH]:MA TcL2  126  Table VIII.  In vivo Repressor Functions of the ADH]:MA Tc’.2 Fusion  127  Table IX.  The Mating-Type Control of the CYC]: URA3 Reporters  130  xi  LIST OF FIGURES  Figure 1.  Regulatory Functions of the Mating-Type Loci.  Figure 2.  DNA-Binding Sites of the Regulatory Proteins in Cell-Type-Specific  2  Transcription.  6  Figure 3.  The DNA and Amino Acid Sequence ofM4 Tc,2.  27  Figure 4.  The ct2 Homeodomain-DNA Co-Crystal Model.  35  Figure 5.  DNA-Binding Specificity of the ci2 Homeodomain.  38  Figure 6.  The Putative Dimerization Motifs of c2.  43  Figure 7.  A Scheme of Two-Stranded Coiled-Coil Structure.  44  Figure 8.  Putative Overlapping Coiled-Coil Motifs of al.  46  Figure 9.  The Two-Stranded Parallel Coiled-Coil Structure of GCN4.  47  Figure 10.  In vitro Oligonucleotide-Directed Site-Specific Mutagenesis.  66  Figure 11.  CYCJ:LacZ Reporter Plasmids.  77  Figure 12.  Construction of Plasmid YHE36.  78  Figure 13.  Construction of CYC]: URA3 and BAR]: URA3 Reporter Plasmids.  79  Figure 14.  Plasmids pEMBL-MATh.-(BamHI-BclI) and pEMBL-Yr-MATx.  83  Figure 15.  The Identities of Various cL2A and cc2B Mutants.  91  Figure 16.  Effect of cL2B-l Mutation on a1-c2 DNA-Binding.  96  Figure 17.  Co-Immunoprecipitation of al and c2.  97  Figure 18.  The al-x2 DNA-Binding Assays of the cx2B Mutants.  99  Figure 19.  The al-cL2 DNA-Binding Assays of the x2A Mutants.  101  Figure 20.  A Schematic Model of the al-x2 Heterodimer.  104  Figure 21.  The Possible Arrangments of the Putative Coiled-Coil Helices in  Figure 22.  the al-x2 Dimer.  108  Hypothetical Coiled-Coil Interactions of al Homodimers.  109  xii  Figure 23.  The Expression of the x2 protein in E. coil and Dimer Analysis of the Expressed Protein.  118  Figure 24.  The Competition Analysis of the cL2-MCM1 DNA Binding.  119  Figure 25.  The DNase I Footprinting Assay on the BAR] x2-Operator.  122  Figure 26.  Analysis on a2-MCM1 Binding to the pseudo BAR] Operator.  123  Figure 27.  Primer Extension Analysis of the ADH]:IvIA Tcs2 Expression.  128  Figure 28.  The c42 and a1-c2 Repressions on the CYC]: URA3 Reporter Genes.  131  Figure 29.  Effect of the cs2 Amino Acid Replacements on al-x2 DNA Binding.  140  Figure 30.  5FOA Selection of x2 Repression.  144  Figure 31.  The Growth Curves of the BAR]: UR.43 Reporter Strains on  Figure 32.  5FOA Medium.  146  5FOA Selection of x2 Repression.  149  xlii  ACKNOWLEDGEMENTS  I would like to thank my supervisor, Michael Smith, for his invaluable supervision and for the freedom and generous support as well as the stimulating research enviromnent he has provided. I also thank members of my advisory committee, Dr. Patrick Dennis, and Dr. Ivan Sadowski, for their constructive criticism and timely advice. Particularly, I am indebted to Dr. Patrick Dennis for his friendliness and help proofreading this dissertation; I am also grateful to Ivan Sadowski for his assistance on experimental approaches, as well as for sharing his enthusiasm in science. My gratitude also goes to the helpful faculty staff and students in the Department of Biochemistry & Molecular Biology and in the Biotechnology Lab. Special thanks go to Dr. Jim Kronstad for critically reading my thesis. I would like to express my appreciation to the past and present Smith lab members who made my life as a graduate student memorable. Especially, I thank Rob Cutler, Sarbjit Ner, Ian Lorimer, Evan McIntosh, Peter Davies, David Goodin, Chris Overall and Mark Ring for their helpful discussions and materials needed for this thesis. To my friend, Marianne Huyer, thanks for the liffle garden she set up by the window in our lab and for proofreading the early drafts of my thesis. To Louis Lefebvre, “Bart Simpson” (in real life), in addition to his wit, insight, and simulating discussion as well as his company on so many late nights over the years, I wish to thank him for extending his philosophical search of “identity” to me. I thank my parents who brought me to this land and gave me the courage to keep on exploring new ways of living. Thank you to all my family members whose love and care has never failed. To my brother, Jimmy, thanks for supporting me in so many ways. To my special friends, Jeffrey Leung, Robert Boissy and Raymond Tong, thanks a lot for showing me how one can bridge the gap between east and west. Last but not least, thank-you, Mei!  xiv  LIST OF ABBREVIATIONS  al  the 1v14 Tal gene product the M4 Tcd gene product  c2  the M4 TcL2 gene product  ATP  adenosine triphosphate  bp  base pair  BSA  bovine semm albumin  CTP  cytosine triphosphate  d.s.  double-stranded  dATP  deoxyadenosine tnphosphate  ddNTP  2’,3’-dideoxynucleoside triphosphate  dNTP  2’-deoxynucleoside triphosphate  Dfl’  dithiothreitol  dUMP  deoxyuridine monophosphate  EDTA  ethylenediaminetetraacetic acid  5F-dUMP  5-fluoro-deoxyurindine monophosphate  5F-UMP  5-fluoro-urindine monophosphate  5FOA  5-fluoroorotic acid  ECL  enhanced chemiluminescence (Amersham)  GTP  guanosine triphosphate  GRM  the General Regulator of Mating type  HEPES  N-2-hydroxyethylpiperazine-N-2-ethanesulphonic acid  HO  the homothallic gene  IgG  immunoglobulin G  IPTG  isopropyl f3-D-thiogalacto-pyranoside  xv  kb  kilobase  kD  kiloDalton  KLH  keyhole limpet hemacyanin  MCM1  the MCM] gene product  OMP  orotidine 5’-monophosphate  ONPG  orthomtrophenylgalactoside  PAGE  polyacrylamide gel electrophoresis  PMSF  phenyl methyl sulfonyl fluoride  PRTF  the Pheromone Receptor Transcription Factor  s.s.  single-stranded  SSN6  the Suppressor of Sucrose Nonfermentation 6 gene  SDS  sodium dodecylsulfate  TUPI  the Thymidine Uptake 1 gene  UMP  uridine momophosphate  UTP  uridine triphosphate  X-gal  5-bromo-4-chloro-3-indoly1-3-D-galactoside  1  INTRODUCTION  1.1.  1.1.1.  CELL-TYPE-SPECIFIC CONTROL REGULATED BY MAT PRODUCTS  Control of Mating-Type by MAT Loci The yeast Saccharomyces cerevisiae has three distinct cell types. The two haploid cell types,  designated a and x, are capable of mating with each other to form the third cell type, the a/tx diploid which is non-mating but is capable of meiosis and sporulation. Two partially nonhomologous alleles, a and ct, are responsible for determining these cell types (Mortimer and Hawthorne, 1969). These alleles are on a single regulatory locus tenned M4 T, which is located on the right arm of chromosome III. Three complementation groups were identified genetically by analysing mutations that affect mating or sporulation (MacKay and Manney, 1974a,b; Kassir and Simchen, 1976; Kiar et at., 1979; Strathern et a!., 1981). Two groups were mapped within the M4 Ta. locus. The first of these, MA Ta.], positively regulates a series of unlinked genes responsible for physiological aspects specific to a. cells (a.-specific genes). The second group, MATc2 (Strathern et at., 1981), negatively regulates a constitutively expressed set of genes responsible for the a cell type (a-specific genes). MA Ta], the third complementation group, was mapped to the M4Ta locus (Kassir and Simchen, 1976), and is essential for determining the diploid cell type (Strathern et a!., 1981). MA Ta] in conjuntion with IvL4Ta.2 indirectly derepresses a set of diploid-specific genes. The M4 T loci have been cloned and sequenced and the location and direction of transcripts specific to each loci has been determined (Hicks et at., 1979; Nasmyth and Tatchell, 1980; Astell et at., 1981; Kiar eta!,, 1981; Nasmyth et at., 1981; Tatchell eta!., 1981). 1.1.2.  Combinatorial Regulations by MATProducts Although the functions of mating-type genes, schematically illustrated in Figure 1, were proposed  based on genetic analyses, this functional scheme was later confinned by molecular studies. The roles of the three MAT products are described below.  2  a Cell  genes (STE3, MFaL..)  genes (BARI, STE6..) :__haploid- specific genes (HO, RME1 ..) MATa  specific genes (spot1ation)  a Cell  o -specific genes :  al  a-specific genes  = _haploid-specific genes M4Ta  diploid- specific genes  a Ia Cell  o -specific genes  OD  haploid-specific genes _diploid-specific genes  Figure 1. Regulatory Functions of the Mating-Type Loci. The three panels show how the products of the MAT loci, namely cd, cL2 and al, regulate the transcription of relevant sets of genes in different cell-types. In the c cell type (top panel), (xl is a positive regulator which activates the cc-specific genes, and cc2 is a negative regulator which represses the a-specific genes. In the a cell type (middle panel), the a-specific genes are constitutively expressed. The haploid-specific genes are expressed in both a and cc mating-types. A subset of the haploid-specific genes (e.g. RAtE]) is necessary for the inhibition of certain diploid-specific genes. In the a/cL diploid cell type (bottom panel), the cc2 protein still represses the a-specific genes; in addition, c.c2 acts with al (al-cc2) to repress the haploid specific genes. Diploid-specffic genes can then be expressed.  3  The Role of MATtJ  The role of the MATcc1 product was initially proposed by Strathern et at. (1981) to be a positive regulator of cL-specific genes because matci.] mutants failed to produce the cc-factor (a pheromone secreted by cc cells for mating with a cells). Molecular cloning of two cc-specific genes, namely the STE3 gene encoding the a-factor receptor (Hagen et al., 1986) and the MFcc] gene [an cc-factor structural gene (Kurjan and Herskowitz, 1982)], has made possible the confirmation of this proposal.  Sprague et at. (1983) detected  STE3-specific RNA in M4 Tcc cells and matcc2 mutants, but not in MA Ta or matcc] mutants. Their results demonstrated that expression of cc-specific genes requires functional ccl proteins, but not cc2 proteins.  The Dual Function of the MATa2 Product  The MA Tcc2 product plays two roles in the regulation of cell-type-specific genes. In cc cells, cc2 represses the transcription of a-specific-genes (Strathern et a!., 1981; Wilson and Herskowitz, 1984). Mutations in MA Tcc2 cause expression of the a-specific phenotypes (MacKay and Manney, 1974a,b) in cc cells. These mutant phenotypes include mating as a cells, the expression of a-specific functions such as afactor production and the barrier function (the ability to degrade cc-factor externally) (Hicks and Herskowitz, 1976; Sprague and Herskowitz, 1981; Manney, 1983).  Northern blot analysis for the expression of a-  specific STE6 RNA (from a gene necessary for a-factor secretion) directly confirmed that transcription of STE6 is negatively regulated by cc2 (Wilson and Herskowitz, 1984). This transcriptional regulation of the aspecific genes is generally called cc2-repression. A second function of cc2 is observed in a/cc diploid cells. It works with the product of the MA Ta] gene to repress transcription of M4 Tccl and of haploid-specific genes (Klar et a!., 1981; Strathern et at., 1981; Jensen et a!., 1983; Miller et a!,, 1985a). M?4Tcc2 is required with M4Ta] in the regulation of haploid-specific genes that is needed for sporulation in diploid cells. This is shown by the inability of either MATcc2 (Strathern et at., 1981) or MATa] (the mata* mutant; Kassir and Simchen, 1976) mutants to mediate sporulation. On the other hand, sporulation is not affected by mutations in MA Tccl. However, recessive mutations including rme (Kassir and Simchen, 1976), sca (Gerlach, 1974), and csp] (Hopper and Hall, 1975) allowed sporulation in the absence of functional MA Tcc2 and MA Ta]. This observation led to  4  the hypothesis that c2 and al permit sporulation by repressing RIvIE and the other negative regulators which normally inhibit sporulation. This role of al and ft2 in transcriptional repression in diploids was verified by demonstration of the requirement for functional MATcL2 and M4 Ta) alleles for negative regulation of the RJv1E gene (Mitchell and Herskowitz, 1986), the HO gene (Jensen et at., 1983) and the A’L4 Tad gene (K.lar et a!., 1981; Nasmyth et at., 1981) at the transcriptionaL level. These genes are referred to as haploid-specific genes because they are expressed only in haploid cell types. The negative regulation of these genes in diploid cells is called al-x2 repression. Regulatory Mechanism and DNA-Binding Specificity of the MAT Products  1.1.3.  Requirement of the MCMJ Product for ctl and c2 Function  Although the mating-type genes first were identified as transcriptional regulators through genetic analyses, later manipulation using molecular biology and biochemical studies of the DNA-binding activities of the MAT products revealed the requirement for a non-cell-type specific factor, MCM1 (also called PRTF or GRM). In order to carry out the transcriptional regulation of cell-type specific genes, the MATcL gene products, namely cd and a2, must cooperate with MCM1. MCMJ (Minichromosome maintenance 1) was originally identified by Maine et at. (1984) as a gene required for proper fiinctiomng of autonomously replicating sequences (ARSs), which are the yeast DNA replication origins (Brewer and Fangman, 1987; Huberman et at., 1987). A role for MCM1 in the expression of cell-type-specific genes was initially suspected because of the two-fold reduction in the mating efficiency of c cells caused by the mcml-] mutation (Passmore et a!., 1988). MCM1 was subsequently shown to be the same protein as the two known factors, PRTF and ORM, which are involved in cd- and x2-dependent DNA-binding, respectively (Bender and Sprague, 1987; Keleher et a!., 1988; Jarvis eta!., 1989; Passmore et at., 1989; Ammerer, 1990). By using gel-shift assays, Jarvis et a!. (1989) demonstrated that MCM1 is able to bind to the PRTF DNA-binding site. Moreover, antibodies raised against a MCM1 polypeptide retarded the migration of PRTF-DNA complexes, indicating that PRTF shares at least an epitope with MCM1.  Also, Keleher et at. (1989) showed that the GRM  dependent DNA-binding activity contains a protein recognized by anti-MCM1 antiserum. Furthermore,  5  Ammerer (1990) purified a single protein which exhibited GRM and PRTF functions, as defined by its ability to bind the a2 and a 1 DNA-binding targets, respectively. By using antibodies against this purified protein to screen a yeast gt1 1 expression library, the gene encoding this protein was cloned.  DNA  sequence analysis showed an identical sequence to MCM1.  Binding of al-MCM1 to the a-Specific Genes and Transcriptional Activation Identification of a role for MCM1 (PRTF) in mating-type control was first demonstrated by  Bender and Sprague (1987) through the use of DNA-binding electrophoretic gel-shift and DNase I footprinting assays. Initially, to define the site of action for al, Jarvis et a!. (1988) constructed deletion mutations in the control region of the a-specific STE3 gene which encodes the a-factor receptor (Nakayama et al., 1985; Hagen et a!., 1986). This analysis identified a 43 base-pair (bp) long sequence from the control region of STE3 that confers both al-dependent and a-factor-induced activation to a heterologous test gene, a CYCI:iacZ fusion (Jarvis et a!., 1988). This short sequence from STE3 contains a 26 bp sequence that is highly conserved among all a-specific genes (Jarvis et a!., 1988). Within this conserved sequence, two regulatory elements were noted: an imperfect palindrome of 16 bp, designated the P box, and a 10 bp element, designated the  Q box (Figure 2).  Bender and Sprague (1987) demonstrated that al binding to this  26 bp sequence needs an additional factor that binds at the P box and is present in extracts derived from a, a, and a/a cells. The strict requirement for this non-cell-type specific factor, called Pheromone Receptor Transcription Factor (PRTF), was shown by the fact that al produced in Escherichia coil is unable to bind to DNA unless supplemented with PRTF-contaimng yeast extracts. MCM1 (PRTF) was shown to bind to its dyad symmetric DNA-target site [a synthetic, perfectly palindromic version of the P-box, termed P(PAL)] by gel-shift assays (Jarvis et at., 1989). Passmore et a!. (1989) determined the DNA-binding specificity of MCM1 by gel-shift assay in which a polymer of the 10 bp dyad symmetry element (CCTAATTAGG) was used as a DNA-binding probe. The ability of variants of this sequence to compete with the DNA probe was tested. The result indicated that the six residues (CCTAAT) is most important for MCM1 recognition. Deviations from the original 10 bp dyad symmetry element occuring at any position other than the CCTAAT are tolerated.  Chemical methylation and DNase I  6  Consensus sequence  Representative site  A) x1site  I  (MCM1 MCM1)  oJ  CTGtCATTGAtGACTAATTAGGAAA  STE3  CrGTCATTGTGACACTAATTAGGAAA  Q B)  MCM 1 site  P  STE6  CMcc11iD P(PAL)  TTACC[AATAGGGAAA  C) cx2 operator site cAATfACCNAATAAGGAAATfIjN  CMcMcMiD CATGTAATTACCfAATAGGGAAATTTACACGC a2  D)  a I -cL2 operator site  MATftL  TC Air TG1INNN Aff NANN1jCA  a2  a1  GCTTCCCAATGTAGAAAAGTACATCATAG  Figure 2. DNA-Binding Sites of the Regulatory Proteins in Cell-Type Specific Transcription. The cell-type specific regulators of the MA Tci locus, cd and c2, require a non-cell-type specific transcription factor, MCM1 (also called PRTF and GRM) to activate the c-specffic genes and repress the a-specific genes, respectively. c2 also interacts with the MATa product, al, to repress the haploid-specific genes. (A) The cd-MCM1 binding site upstream of the cc-specific STE3 gene is shown in this figure. The MCM1 binding site is designated as the “P-element” and the ccl binding site as the “Q-element’. The PQ elements are located upstream of all cc-specific genes and are required for transcriptional activation. ccl recognizes the Q-site and helps MCM1 dimers bind efficiently to the imperfect palindromic P-element, thus leading to transcriptional activation. (B) In a-specific genes (e.g. the STE6 gene), MCM1 dimers alone can bind efficiently to the almost perfect palindromic P sequences located upstream of STE6, thus activating transcription. (C) In addition, the cc2 operator consists of two cc2-binding sites flanking a central MCM1 binding site. cL2 homodimers bind cooperatively with the MCM1 dimers to the cc2 operator and repress expression of this a-specific gene. (B) In haploid-specific genes (for instance, the MA Tccl gene shown in the figure), al and cc2 form heterodimers and bind to the al-cc2 operators to repress the transcription of these haploicl-specific genes. The dyad symmetry shown in the al-cc2 operator (ATGT ACAT) is also present in the cc2 operator. However, the spacing is longer in the a-specific genes because of the presence of the MCM1 recognition sites. -  7  interference experiments subsequently showed that MCM1 interacts with each ATTAGG half site on both strands of the DNA target. The two G residues seem particularly important since methylation of either one of them abolishes MCM1 binding (Passmore eta!., 1989). Although MCM1 (PRTF) can bind to the palindromic P(PAL) alone, it requires cooperative interactions with oci in order to bind the imperfect palindrome of the P-box found in STE3, termed P(STE3). Similarly, these additional interactions with MCM1 allow xl to bind the  Q box.  The relationships between  binding at these sites and transcriptional activation were demonstrated by the fact that P(STE3) functions to activate transcription only in  X  cells when adjacent to a  Q  box, whereas the P(PAL) sequence mediates  transcriptional activation in a non-cell-type specific manner (Jarvis eta!., 1988). However, increased binding affinity to the PQ-element was thought to be only one of the consequences of MCM1-cd interaction.  Tan and Richmond (1990) suggested that cd also alters the  conformation of MCM1. Experiments showed that the DNA footprinting of the ct-specific gene target for the MCM1-cd. complex was broader than for MCM1 alone (Tan et a!., 1988). Also, proteolytic experiments suggested that MCM1 has a different conformation when bound to an a-specific gene than when bound to an x-specilic gene (Tan and Richmond, 1990). These results led to the speculation that cd binding may change an inactive MCM1-DNA complex into an active conformation. However, by using LexA-fusion proteins and an artificial reporter gene containing a LexA DNA-binding site, Sengupta and Cochran (1991) showed that LexA-MCM1 or LexA-cd alone is sufficient for trans-activation. As well, Grayhack (1992) discovered that MCM1 is capable of binding to one of two strands of the x-specific (PQ) or the a-specific (P) binding site in a sequence-specific manner,  Interestingly, in  parallel to the effect of x1 on the duplex-DNA binding activity of MCM1, cd also enhances the singlestranded DNA binding by MCM1, but c1 alone does not bind to the single-stranded target DNA.  Binding of a2-MCM1 to the a-Specific Genes in 2 Repression  The fact that x2 regulates transcription through specific DNA-binding to the target gene was first demonstrated by Johnson and Herskowitz (1985). A 32 bp DNA sequence was found to be protected from DNase I digestion by the cL2 protein (Figure 2, panel C, CATGTAATTACCTAATAGGGAAATT  8  TACACGC).  This cL2 operator directs negative regulation in a x2-dependent manner.  Sequences  homologous to this x2 binding site are present in the upstream regulatory regions of all a-specific genes (Johnson and Herskowitz, 1985; Miller et al., 1985a). Although a hybrid x2-J3-galactosidase synthesized in E. coil bound to the DNA (Johnson and Herskowitz, 1985) in the absence of any other yeast proteins, Keleher et a!. (1988) have shown the involvement of a yeast protein (designated as General Regulator of Mating type, or GRM) in high affinity cooperative binding to the DNA target site. The x2 protein present in crude yeast extracts protects a large region of DNA (32 bp) from cleavage by DNase I, whereas x2 purified from E. coil protects only the ends of the 32 bp site (binding to two 13 bp sites) leaving the middle of the site exposed. This observation led to a search for a second yeast protein capable of protecting the middle of the c2 binding site, and resulted in the identification of GRM.  These two proteins show overlapping regions of protection but can bind  independently (Keleher et al., 1988; Saner et a!., 1988). However, on a wild-type operator, GRM and c2 bind cooperatively and form a ternary complex, with GRM estimated to raise the affinity of cL2 for the operator by approximately 50-fold. In vivo evidence for the absolute requirement of GRM in repression by x2 comes from the fact that mutant operators that abolish GRM binding without affecting x2 binding are unable to repress transcription (Keleher et a!., 1988). On the other hand, a mutant operator with terminal deletions, which is incapable of cL2 binding but capable of GRM binding, exhibits weak but significant activation of transcription. GRM was postulated to be the co-repressor with c2 and a transcriptional activator in the absence of x2, GRM was subsequently found to be identical to PRTF, and is encoded by MCMI (Passmore et a!., 1988) (discussed above). Two models were proposed by Johnson and colleagues to explain the repression by x2-MCM1. In the first model, a.2 was proposed to interact with MCM1 and mask the domain of MCM1 which interacts with the general transcriptional machinery, thus blocking activation. In the second model, x2-MCM1 binds with the transcriptional machinery and forms a very stable complex. The interaction of x2-MCM1 with the transcriptional machinery was also proposed to be much stronger than that of a normal transcriptional  9  activator so that c2-MCM1 would compete off the normal activator and lock the transcriptional machinery in place, thus preventing a subsequent step in initiation from proceeding. Simpson and colleagues (Roth et a!., 1990; Shimizu et at., 1991; Roth et a!., 1992) suggested that the (x2-MCM1 complex may cause repression by a precise and stable positioning of nucleosomes adjacent to the cL2 operator.  Positioning a nucleosome in promoter regions might inhibit transcription through  occlusion of cis-acting elements of the target a-specific genes (Simpson, 1990). The best support for such a view caine from an experiment using amino-terminal deletion and point mutants of the histone H4 gene (which encodes a protein component of the core particle in the nucleosome) (Roth et at., 1992). These mutations altered the location of nucleosomes adjacent to the cL2 operator and partially derepressed the expression of a 3-galactosidase reporter construct that is usually under c2 control. Roth et a!. proposed that the mutation in the H4 histone might specifically prevent normal interaction between the nucleosome core particle and the x2-MCM1 complex, even though x2-MCM1 still binds to the operator site. The mechanism of oL2 repression of a-specific genes appears to be a complicated process that requires multiple factors. At least two other genes, SSN6 (Carlson et a!., 1984; Schultz and Carlson 1987; Schultz eta!., 1990) and TUPI (Lemontt eta!., 1980; Trumbly, 1986; Williams and Trumbly, 1990; Mukai et at., 1991) are also known to be involved in c2 repression and in a1-c2 repression. SSN6 was first identified as a negative regulator of the expression of the SUC2 gene which codes for yeast invertase.  The requirement of SSN6 for cL2 repression was first noticed because ssn6-mutant  ot-cells mate at low efficiency compared to wild-type x cells (Carlson et a!., 1984) and express a-specific genes (Schultz and Carlson, 1987). Likewise, the low mating efficiency of the tupi pleiotropic mutant phenotype led to the conclusion that TUPI is also involved in mating-type control (Lemontt et a!., 1980). A direct role for SSN6 in x2 repression was suggested by the recent work of Keleher et a!. (1992). The cL2 operator was demonstrated to confer mating-type control only in c SSN6 or a SSN6/ci. SSN6 yeast cells but not in isogenic strains carrying the non-functional ssn6 mutation. Keleher et a!. also showed that SSN6 is sufficient to carry out TUP1-dependent repression, and suggested that the x2-MCM1 complex recruits SSN6-TUP1 to execute repression, based on the fact that SSN6 and TUP1 form a complex in vitro (Williams et a!., 1991).  In a similar experiment using the a1-c2 operator sequence which is located  10  upstream of all haploid-specific genes, SSN6 was also shown to be essential for al-(x2 repression (Keleher et a!., 1992). The DNA-binding specificity of al-x2 will be discussed below.  Binding of al-cL2 to the Haploid-Specific Genes  A DNA sequence potentially recognized by al-cc2 was initially identified by Miller et al. (1985a). They carried out deletion analysis of the HO regulatory region and discovered redundant DNA elements that are similar to sequences found in the regulatory regions of other haploid-specific genes, including MA Tct] and STE5.  Two different versions of this sequence found in the HO promoter were inserted into the  promoter of the CYC]:LacZ reporter gene.  Significant decreases in the expression of the reporter gene  carrying this potential a1-c2 control element were detected in a/n diploids (containing wild-type al and n2), but not in haploids or in a matal/IvL4Tn diploid (both missing al or cL2). This confirmed the in vivo function of this al-cx2 control element (see Figure 2, panel D). Other analyses of M4 Tn also demonstrated the existence of this al-n2 control element. Consistent with the finding of Miller et a!. (1985a), Siliciano and Tatchell (1984) showed that a 14 bp deletion from the intergemc region between ]vL4Ta.2 and MA Tn], abolished diploid-specilic repression. In the wild-type diploids, M4 Tn] transcription is completely repressed and M4 TcL2 transcription is reduced by about 10-fold (Klar et a!., 1981; Nasmyth et a!.. 1981) but deletion of the 14 bp’s in the intergenic region led to constitutive transcription of both genes in a/n cells. Siliciano and Tatchell further analyzed a larger 53 bp intergenic region. Constitutive expression of MA Tn] and MA TcL2 resulted when this 53 bp region was removed.  A 28 bp sequence, corresponding to a putative control element, was able to restore normal  expression of both genes. The similarity between the sequences of the al-cL2 control element and the n2 control element was pointed out by Miller et a!. (1985a). Both eLements contain the sequence motif ATGT.. ..ACAT in inverted orientation (Figure 2). However, the two sequences differ from each other by the spacing between these inverted repeats with the al-n2 element having its two half-sites closer together.  This sequence  arrangement led to the idea that cc2 binds to both half sites under the influence of al (inducing a  11  conformational change) or alternatively, that al and cL2 form a heterodimer, with each subumt occupying a half-site. Goutte and Johnson (1988) demonstrated directly that the interaction of al with (x2 alters the DNA sequence recognized by x2. They found that an al-x2 control element sequence was able to compete with al-dependent x2 binding to an al-x2 control element but did not offset binding to the c2 control element. This work while confirming the influence of al on the DNA-binding specificity of a2, did not show a direct interaction between al and the DNA target. More recently, Dranginis (1990) provided the evidence that al and cL2 bind directly to the a1-o2 element cooperatively as a heterodimer and demonstrated a difference in the pattern of DNA contacts of at and of cL2. Both the al and (x2 proteins as well as a DNA probe carrying the al-ix2 element were found to be necessary to form the ternary complex. Investigation of the base contact pattern of the al-cL2 heterodimer indicated that the 2 protein changes the binding pattern upon interaction with al such that al-oc2 binds to its operator sequence predominantly on a single face of the helix instead of binding to the opposite faces of the double helix as in the case of cL2 binding to the x2-operator (Keleher et a!., 1988; Passmore et a!., 1989). Besides a1-c2 binding to the control element, there are other factors which govern al-cQ repression. In a search for mutants which specifically affect al-cL2 repression, Harashima et a!. (1989) isolated mutants which failed to sporulate (defective in a1-c2 repression) but were still capable of mating as c cells (i.e. have functional cc2 repression). A single complementation group of al-x2 repression defective mutants was identified and designated aarl (1-lpha 2 repression 1). The A.4R1 gene was subsequently cloned and its DNA sequence analyzed. Comparison of nucleotide sequences showed that AARJ is identical to the TUP] gene. As described in Section of this introduction, both TUP 1 and SSN6 are also involved in cL2 repression.  Therefore, it is now clear that the cc2 repressor defines its target sites by  interacting with different DNA-binding co-repressors such as al and MCM1. In spite of the differences in protein-protein interactions and in DNA binding, c2 needs the common effector TUP1-SSN6 for repression. This is believed to be a general repressor in yeast (Keleher et a!., 1992).  12  In summary, as illustrated in Figure 2 and discussed in this section (1 1.3.), transcriptional regulation of cell-type specific genes is in a combinatorial fashion. That is, the products of the M4 T loci and MCM1 work in concert to control the expression of cell-type specific genes. 1.1.4.  MCM1 and STE12 in the Pheromone Response Pathway A complete overview of the transcriptional control of mating-types requires a description of the  role of the STEI2 gene product in mating-type control. STE12 plays a part in induced expression and also in setting the basal level of expression of the a- and c-specffic genes (reviewed in Dolan and Fields, 1991). The a cells produce and export a pheromone (the a-factor, a short polypeptide which triggers c cells to respond and proceed to mating). In addition, they contain a receptor for the pheromone secreted by the opposite cell type. Likewise, c cells secrete the o-factor and bear the a-factor receptor. The expression of a-specific genes and ce-specific genes increases by a few fold with the pheromone response. STEJ2 encodes a DNA-binding protein that recognizes the pheromone response element (PRE) (Errede and Ammerer, 1989; Dolan et a!., 1989), the sequence that mediates increased transcription of responsive genes following the treatment of cells with pheromone. The PRE sequence, (A/T)TGAAACA, is found often in multiple copies in all upstream regions of a-specific genes (Hatting et a!., 1986; Kronstad et at., 1987). Like cd and cL2, STE12 is able to cooperate directly with MCM1 for binding the PRE site in a-specific genes (Errede and Amerer, 1989; Primig et at., 1991), Although STE12-MCM1 interaction is necessary for optimal STE12 activity on some genes (e.g. a-specific genes), STE12 alone can act as a pheromone-responsive transcriptional activator without any other protein bound at or adjacent to the PRE sites (Song et a!., 1991; Yuan and Fields, 1991). Likewise, optimal activation of transcription of ft-specific genes upon a-factor induction is also dependent on STE12. Even though there is no clear homology to the PRE consensus sequence at the PQ site, this sequence alone can mediate pheromone-induced transcription (Jarvis et at, 1988; Sengupta and Cochran, 1990), indicating PQ is a distinct element which can mediate pheromone response. Although both MCM1 and cd are responsible for basal activation of ft-specific gene transcription, only ad and STE 12 were shown to be required for the a-factor induced activation (Hwang-Shum et at., 1991; Sengupta and Cochran,  13  1991).  STE12 is a phosphoiylated protein and its level of phosphorylation directly correlates with the  transactivation ability of STE12 and increases following the treatment of cells with pheromone (Song et al., 1991). 1.1.5.  Regulation of the Expression of MAT Locus Transcriptional and posttranscriptional regulation have been reported for the M4 Ta and MA Ta2  alleles. The first to be investigated was the transcriptional regulation of the M4 Ta. locus (Siliciano and Tatchell, 1984; Shore and Nasmyth, 1987; Giesman et a!., 1991; Kurtz and Shore, 1991).  The RAP1  (jepressor activator protein 1) protein was shown to activate M4 Ta. transcription since a 10 bp positive regulatory upstream sequence which coincides with a RAP 1-binding site acts as an upstream activating sequence (Siliciano and Tatchell, 1984; Giesman et at., 1991). The abnormal mating phenotype resulting from point mutations in the intergenic region of the M4 Ta. locus was found to be suppressed by overproduction of RAP1 (Giesman et a!., 1991).  Thus, these results indicate positive regulation of  transcription of the MA Ta. locus by RAP 1. On the other hand, negative auto-regulatory control of M4 Ta. has also been reported.  As  described earlier, A’L4 Ta. is under diploid-specific al-a.2 repression in which M4 Ta.I transcription is totally repressed and !vL4 Ta.2 transcription is reduced approximately 10-fold (Klar et a!., 1981; Nasmyth et at., 1981). These repressions are mediated by the MAT products al and a.2 (see description on the al-a.2 repression above). Although transcriptional regulation of the MA Ta locus has not been established, posttranscriptional processes have been studied (Miller, 1984; Ner and Smith, 1989; Nakazawa eta!., 1991). By Si nuclease mapping, Miller verified the existence of two introns in the MA Ta] transcripts and noticed that the 3 -intron is processed less efficiently. This observation led to the speculation that there may be two different functional al proteins resulting from differential splicing of the introns.  Ner and Smith  investigated the role of the different transcripts (those with both or only one intron spliced) by analyzing the function of M4 Ta] derivatives carrying two, one or none of the introns. The results indicated that the only functional protein is the one produced when both introns are spliced from the mRNA.  Furthermore,  14  consistent with the importance of intron splicing, Nakazawa et a!. isolated a class of a1-c2 repression deficient mutants called aar2. These mutants accumulated unprocessed ]v14 Ta] transcripts, and the mutant phenotype could be partially suppressed by supplementing with the intronless version of the M4 Ta] gene (Nakazawa eta!., 1991). 1.1.6.  Yeast Cell-type Specific Transcription as a Model System Although budding yeast is a unicellular organism, it exhibits processes that are analogous to those  found in the development of multicellular organisms, such as cell-type specification (mating-type control) and cell-cell interaction (pheromone-response pathway). The relative simplicity of yeast and the ease of genetic manipulation of this organism have resulted in an integrated view of how regulatory proteins coordinate with intercellular signaling to determine cell-types. The negative control of transcription of celltype specific genes includes interaction among c2 and al, MCM1, SSN6 and TUP1. These factors contain functional motifs that are common to many other eukaryotic transcriptional factors. For example, ci2 and al belong to the group of DNA-binding homeodomain proteins (McGinnis et a!., 1984; Scott and Wenier, 1984), while MCM1 is similar to other members of the MADS family (Schwarz-Sommer eta!., 1990). [The MADS family was first identified by the similarity between the yeast MCMI, and the plant developmental genes, namely gamous and deficiens, as well as the human serum response factor, SRF (Schwarz-Sommer et a!., 1990).]  SSN6 and TUP1 both contain TPR (tetratricopeptide repeat, Sikorski et a!., 1990), and a  f3-transducin-related repeat motif, respectively (Zhang et a!., 1991).  Therefore, studies on these yeast  factors will help to further the understanding of transcriptional regulation in higher eukaryotes.  1.2.  1.2.1.  TRANSCRIPTIONAL REGULATION IN EUKARYOTES  A Common Theme in Eukaryotic Transcription In eukaryotes the synthesis of mRNA from template DNA is carried out by a group of proteins  whose central component is RNA polymerase II (Pol II). Accurate and regulated transcriptional initiation depends on a battery of auxilliary protein factors (transcriptional factors) which recognize and bind to  15  specific DNA sequences (cis-acting regulatory elements) located within the promoters of eukaryotic genes (reviews in Mitchell and Tjian, 1989; Ptashne, 1988; Struhi, 1989). It is believed that there are common molecular mechanisms of transcriptional regulation among all eukaryotes ranging from humans to yeast. Conservation of the regulatory mechanism is suggested by the existence of functional interchangeable transcriptional factors and cis-acting elements. Yeast and higher eukaryotic cells contain structurally similar and functionally analogous transcriptional factors that recognize essentially identical DNA sequences (jun/AP-1 and yeast GCN4, Bohmann et a!., 1987; Struhi, 1987; Angel et a!., 1988; yAP-i, Harshman et al., 1988; Jones eta!., 1988; mammalian CCAAT recognition factor, CP1 and yeast HAP2IHAP3, Chodosh et a!., 1988; TATA binding factor/TFIID, Buratowski et at., 1988; mammalian SRF and yeast MCM1, Grueneberg et al., 1992). Furthermore, yeast proteins can activate transcription in a wide variety of eukaryotic organisms, including mammalian cells (Kakidani and Ptashne, 1988; Webster et a!,, 1988), Drosophila (Fischer et a!., 1988), and plant cells (Ma et a!., 1988). Vertebrate proteins can also stimulate transcription in yeast cells (Struhl, 1989; and references therein). 1.2.2.  Cis-Elements on Promoters The promoter region of eukaryotic genes is typically composed of two classes of DNA-sequence  elements. The first class of elements constitutes the core promoter and is common to most genes. This promoter region is the recognition site for a subset of transcriptional factors (the general transcriptional factors) that are responsible for recruiting and aligning Pol II at the promoter. As well, this region directs accurate transcriptional initiation carried out by the transcription machinery, albeit at a low and unregulated level (Saltzman and Weimnann, 1989; Weis and Reinberg, 1992). The most important element of this first class is the TATA box, which is found 20-30 bp upstream from the site of transcriptional initiation in higher eukaryotes. However, the TATA element of yeast promoters is located at a greater and more variable distance from the transcriptional initiation site (roughly 40-120 bp). Although the TATA element is essential for proper initiation of transcription, in yeast the actual site of initiation is dependent on the “initiator” element that is located near the site of initiation (Chen and Struhl, 1985; Hahn et at., 1985; Nagawa and Fink, 1985; McNeil and Smith, 1986; Smale and  16  Baltimore, 1989). There are also other TATA equivalent elements, termed “initiator”, which serve as the recognition site for the transcriptional machinery (Weis and Reinberg, 1992). A second region of the promoter is responsible for gene-specffic regulation of transcription initiation. This region lies upstream of the TATA element (or equivalent elements) and is comprised of multiple short DNA-sequence elements (6-12 bp). Commonly found in many higher eukaryotic promoters are the CCAAT boxes (Efstratiadis et a!., 1980), the GC box (with the DNA-sequence GGGCGG; Benoist and Chambon, 1981) and the octamer elements (ATGCAAATNA; Falkner and Zachau, 1984; Parsiow et a!., 1984). Many of these elements (including the CCAAT and GC boxes) are located within a few hundred base pairs of the transcriptional start site. However, the enhancer elements may be positioned in either orientation many kilobases away and function either upstream or downstream from the transcriptional initiation site (Serfling et a!., 1985).  The specificity of the enhancer is apparently derived from the  combination of sequence motifs present within it (Yamamoto, 1985; Schirm eta!., 1987). In yeast, the function of an enhancer is carried out by “upstream activating sequences” (UAS) (Guarente and Hoar, 1984; Struhl, 1984).  Analogous to mammalian enhancer sequences, yeast UAS  elements (typically 10-30 bp in length) also function in both orientations and at long and variable distances (anywhere from 100 to 1500 bp upstream of the initiation site depending on the gene) from other promoter elements and the mP.NA initiation site.  In contrast to enhancers, UAS elements do not activate  transcription when located downstream of the mRNA initiation site (Guarente and Hoar, 1984; Struhi, 1984). The gene-specific regulatory region of a promoter contains positive regulatory cis-elements and negative elements.  Negative regulatory elements have been found in the human n—interferon gene  (Nourbakhsh et a!., 1993), the human cytokine IL-4 gene (Li-Weber et a!., 1992), the long terminal repeat of the human immunodeficiency virus type 1 (Rosen et a!., 1985; Smith and Greene, 1989; Yamamoto et a!., 1991), the mouse renin genes (Ren-id; Barrett et a!., 1992), the chicken lysozyme gene (Baniahmad et a!., 1990) and the Drosophila Ubx promoter (Biggin and Tjian, 1989). In the case of yeast, some promoters contain operator elements that repress the level of transcriptional initiation (Brent, 1985; Struhl, 1989). Operators, like the UAS elements, function bi-directionally and at variable distances upstream of TATA  17  elements. Although repression is generally much more efficient when the operator lies between the upsteam and TATA elements, some can repress transcription when located upstream of UAS elements (Siliciano and Tatchell, 1984; Johnson and Herskowitz, 1985; Struhl, 1985).  An exception to this is the mating-type  “silencer” DNA-sequence which can efficiently repress transcription when located at a distance as far as 2 kb upstream or downstream from the mRNA initiation site (Feldman eta!., 1984; Brand et al., 1985). 1.2.3.  Trans-Acting Factors There are two classes of transcriptional factors acting on the cis-elements. The first class consists  of the general transcriptional factors which bind to the core promoter DNA-elements. Pol II cannot initiate transcription on its own and thus requires a set of general transcription factors which include components of the transcriptional machinery for RNA polymerase II (TFIIA, B, D, E, F and H; reviewed in Zawel and Reinberg, 1992). The association of general factors with the promoter is an ordered process and allows the assembly of a “pre-initiation complex”. Binding of TFIID to the TATA sequence is thought to be the first step in the assembly of a functional transcriptional complex at the promoter.  In higher eukaryotes the  general transcriptional factor TFIID is a large multi-subunit complex consisting of the TATA-binding protein (TBP) and at least seven distinct and tightly associated factors called TAFs (TBP associated factors; Dynlacht et a!., 1991; Tanese et a!., 1991). TAFs are required in addition to promoter-specific activators and the basal transcriptional factors in order to achieve stimulated levels of transcription.  It has been  proposed that the function of these TAFs is to serve as “adaptor” molecules connecting the activation domains of regulatory factors and the basal transcriptional machinery (Pugh and Tjian, 1990; Hoey et a!., 1993). Equivalent transcriptional adaptors in yeast have also been reported (Berger et a!., 1990; Kelleher et a!., 1990; Flanagan et a!., 1991).  However, unlike TAPs, the yeast “adaptor” proteins are not stably  associated with TBP. TBP can be purified from yeast as a free protein (Buratowski eta!., 1988). The second class of the trans-acting factors contains numerous, different, gene-specific transcriptional factors. Most of these factors bind to their cognate recognition sites on DNA and then act, either positively or negatively, to influence the rate at which Pol II initiates transcription at a promoter (Roeder, 1991).  Typically, there are at least two functional surfaces on these sequence-specific DNA-  18  binding proteins.  One surface interacts with the DNA target, while the other (one or more) surface is  involved in protein-protein interactions. These interactions include the oligomerization of DNA-binding proteins, and the interaction of the effector domains (activator or repressor domains) with the transcriptional machinery (in some cases, with co-activator or co-repressor proteins). 1.2.4.  Positive and Negative Regulations In positive transcriptional regulation, the transcriptional activation domain of the gene-specific  factor, once tethered at a promoter or enhancer via the DNA-binding domain, modulates the rate of initiation by Pol II via the general factors. Modulations may occur when activation domains interact directly or indirectly with one or more components of the transcriptional machinery, either recruiting general factors to the promoter, or altering their activity states (Weis and Reinberg, 1992). Recent studies indicate that one possible target for the activation domains is the TAFs, since these polypeptides are required for the stimulatory action of the transcriptional factors, but not for basal-level transcription. Furthermore, TAFs have been shown to associate with transcriptional activators in vivo and in vitro (Hoey eta!., 1993). However, as implied by the discovery of the Dr transcriptional repressors that bind TBP and inhibit recruitment of the rest of the transcriptional machinery (Inostroza et a!., 1992), activation by genespecific factors may result from a removal of this negative control. Alternatively, activators may function in a more indirect manner by altering chromatin structure, thereby clearing a promoter and allowing the transcriptional machinery to assemble (Felsenfeld, 1992). Repressors or negative transcriptional factors can interfere with many and perhaps all the steps required for transcriptional activation. Transcriptional repression mechanisms fall into the following three classes (Renkawitz, 1990): (i) inhibition of DNA binding, (ii) blocking of activation, and (iii) silencing. (i) Inhibition of DNA-binding affects the binding of a particular positive factor to its promoter sites (or the enchancer or UAS) or the binding of the initiation complex to the transcriptional start site or the TATA box of the promoter. Most examples of DNA-binding inhibition involve competition between a negative factor and a positive trans-activating factor.  For instance, the CCAAT displacement protein  interferes with the binding of the CCAAT-binding factor to the sperm-specific histone H2B-1 gene; and the  19  GC-box factor competes with Spi or other positive factors for binding to GC-rich sequences in several genes (Kageyama and Pastan, 1989). The competing negative factors can themselves be trans-activating.  In these cases, a positive  transacting factor for one gene can act as a repressor for another gene by competing for the DNA-binding site.  In Drosophila development, induction of the segment polarity gene, engrailed, (en) by the  homeodomain gene products, zen-related (z2),fushi tarazu Q?z) or paired (prd), is inhibited by the products of the en or eve (even-skipped) genes which are also positive factors for other genes (Levin and Manley, 1989). In another example, a single monomer of a dimeric activating factor can be replaced by another monomer, resulting in an inactive heterodimer. This is the case of the proto-oncogene product Jun-B, which inhibits the activator protein AP-1, a heterodimer of Jun-Fos, by forming an inactive c-Jun-Jun-B dimer and competing out Jun-Fos for binding to DNA (Chiu eta!., 1989; Schütte et al., 1989). Also, inhibitor proteins that are not able to bind DNA can dissociate and inactivate a DNA-bound trans-activating factor.  For  example, the 90 kD heat shock protein forms a cytoplasmic complex with the glucocorticoid receptor, thereby causing it not to bind to its DNA target (Pratt et a!., 1988). (ii) Blocking of activation may occur when a repressor or a negative factor interferes with the transmission of a positive signal from a DNA-bound positive factor to the transcriptional initiation complex. Such interference may involve a negative factor binding to the DNA somewhere between the positive factor and the initiation complex. Another form of interference involves a protein binding directly to the trans activating domain of the positive factor (“quenching”) (Ma and Ptashne, 1988; Levin and Manely, 1989) or an intermediary co-activator protein, (“squelching”) (Ptashne, 1988). A well known example of repression by “quenching” is the GAL8O-mediated repression of trans-activation by the GAL4 protein in yeast. GAL8O, which cannot bind to DNA, binds GAL4 and blocks its trans-activating domains. When a trans activating domain is artificially attached to the GAL8O protein, it no longer inhibits expression (Ma and Ptashne, 1988). The adenovirus protein Ela, likely represses a large battery of different transcriptional factors, including AP-1, VP16 (herpes simplex virus protein), the glucocorticoid and estrogen receptors, and a large number of different trans-activating factors binding to the SV4O enhancer. It is thought that Ela represses via “squelching” by binding to and inactivating an intermediary factor that would normally  20  transmit the positive signal from the activating factor to the initiation complex.  (Rochette-Egly et at.,  1990). (iii)  Silencing, similar to the position- and orientation-independent action of an enhancer,  involves mechanisms that do not depend on a particular arrangement of positive and negative regulatory sequences and does not even require the presence of enhancer elements (Brand, 1985).  Negative trans  acting factors can bind to a “silencer-element” and cause silencing. For example, the RAP1 (repressor and activator protein 1) and ABF1 (a protein binding to the replication origins) bind to the HMR-E “silencer” and repress the expression of the silent mating-type loci. It has been suggested that RAP1 may be involved in chromatin loop formation between two silencers or between a silencer and the promoter of the silent mating-type locus (Hofmann et at., 1989). Such a configuration may lock the gene promoter within the loop, thus making the promoter unavailable to transcriptional activators. Alternatively, the silencer factors may lock the transcriptional initiation complex in a simpler manner (and make it unavailable for activating factors) or disorganize the transcriptional initiation complex. Although the cL2 repressor functions by the “blocking of activation” of MCM1, the repressor also functions as a silencer protein. repression occurs.  When an cL2-operator is placed upstream of a UAS of a test promoter,  This repression may be best explained as a locking of the transcriptional start site  (Keleher et at., 1988). Such locking may also involve recruiting SSN6 and TUP1 (Keleher et at., 1992) and accurate nucleosome positioning or alteration of the chromatin structure (Roth et a!., 1990; Shimizu et a!., 1991; Roth et at., 1992).  Structure-Function of Transcriptional Factors  1.2.5.  The Modular Structure Concept  In spite of the great variety of trans-acting factors, these factors are composed of a limited number of structural and functional motifs. Generally, these functional motifs are combinations of simple secondary structures such as the x-helix and n-sheet. These motifs are capable of folding into separated and stable domains (reviewed in Johnson and McKnight, 1989; Mitchell and Tjian, 1989; Frankel and Kim, 1991;  21  Harrison, 1991; Pabo and Sauer, 1992).  Many factors were initially grouped into different families  according to the similarity of their DNA sequences. Later, the common structures within each class were confirmed by structural studies. The different structural motifs were also associated with unique functions such as sequence-specific DNA-binding and protein-protein interactions.  Combinations of separated  functional domains can give rise to a large variety of regulatory factors.  DNA-Binding Motifs  The following is a brief description of the structural aspects of (i) the homeodomains and related motifs, (ii)  zinc-requiring DNA-binding motifs which include the zinc finger protein and the nuclear  steroid receptors, (iii)  basic-coiled coil (basic-leucine zipper) proteins, and (iv)  the helix-loop-helix  proteins. (i) The homeodomain motif is found in a large family of eukaryotic gene-specific transcriptional regulators. It is a highly conserved region of about 60 amino acids and was originally found in Drosophila homeotic proteins that regulate body patterning during embryo development.  The homeodomain is also  found in a wide range of eukaryotic organisms, including mammals, plant and fungi (reviewed in Scott et a!, 1989; Affolter eta!., 1990). The homeodomain is a distant relative of the bacterial “helix-turn-helix” DNA-binding motif which was the first DNA-binding motif to be defined in a number of bacterial regulators (for example, the Cro protein, k-repressor, and CAP repressor; reviewed in Takeda et a!., 1983). DNA binding by members of this “helix-turn-helix” class is achieved by the binding of a short cL-helix (termed the “recognition helix”) to base pairs in the major groove of the DNA. A second cL-helix, connected to the first by a short turn, sits at a right angle across the major groove, locking the recognition helix in place.  The “helix-turn-helix”  proteins bind as dimers, with each monomer binding in adjacent major grooves on one face of the DNA. The X-ray crystallographic and NIVIR-deriveci structures of homeodomains, including that of the Drosophila engrailed and antennapedia proteins (Kissinger et a!., 1990; Oiling et at., 1990) as well as the yeast (x2 repressor (Wolberger et al., 1991), retain the essential feature of the “helix-turn-helix” structure (described in Section 1.4.2.). However, unlike the “helix-turn-helix” motif isolated homeodomains bind  22  DNA as monomers and rely on additional DNA contacts made by an extended recognition helix and an additional cc-helix. Furthermore, the orientation of the “helix-turn-helix” with respect to the bound-DNA target is very different for the yeast and Drosophila homeodomains as compared to that of the bacterial 2repressor, Structures of several homeodomain-related DNA-binding motifs have been reported. The first group of these is the POU-specific domain which uniquely juxtaposes the POU-homeodomain in the POU DNA-binding domain. The POU-domain was first discovered in three mammalian transcriptional factors and the recent NMR-derived structures show superimposable “helix-turn-helix” motifs when compared with the k-repressor motif except that in the POU-specific domain the first helix and the linker to the second helix of the motif are extended (Assa-Munt et a!., 1993; Dekker et a!., 1993). The second group of the homeodomain-related motifs includes those found in the rat liver transcription factor LFB1 (Finney, 1990) and the yeast STE12 (Yuan and Field, 1991). In contrast to the typical homeodomain which has a turn of only three amino acid between helix 2 and helix 3, these related motifs contain an extend ioop of up to 21 amino acids in between the corresponding helixes. Nevertheless, as in the case of LFB 1, the topology and orientation of the helices is essentially the same as that found in the MATcc2, engrailed and Antennapedia homeodomains (Ceska et a!., 1993). (ii) Zinc-requiring DNA-binding motifs are found in many eukaryotic transcriptional factors and consist of two large families: (a) zinc-fingers and nuclear receptors, and (b) the representative cysteine rich DNA-binding proteins such as yeast GAL4 activator. The zinc finger is a common structural motif used by this second large family of DNA-binding proteins. The motif is characterized by a 30 amino acid repeat unit (typically two or three repeats) that contains a zinc ion co-ordinated by two cysteines and two histidines. This motif was first identified in the Xenopus RNA polymerase III transcription factor, TFIIIA (Miller et a!., 1985b), and has subsequently been found in a multitude of mammalian transcriptional factors. The structure of all three DNA-bound “finger” motifs of Zif268 (Pavietich and Pabo, 1991) has been resolved. Each finger interacts with the DNA in a similar way and shows a compact structure comprised of one cc-helix and two stretches of n-sheet folded  23  around a single atom of zinc.  The x-he1ix penetrates the major groove and defines the specific-base  contacts. The DNA-binding domain of members of the steroid hormone nuclear receptor family, such as the glucocorticoid (Luisi et a!., 1991) and estrogen (Schwabe et a!., 1990) receptors, consists of two zinc ions each attached to four cysteines. The first c-helix of the bi-helical domains sits in the major groove of the DNA and is critical for specific-base contacts, while the second helix makes contacts to the DNA phosphate backbone and provides a surface for dimer formation. The yeast GAL4 zinc-containing DNA-binding domain called a zinc-cluster (Kraulis et al., 1992; Marmorstein et a!,, 1992.) is somewhat different from those described above. The domains bind as a dimer to the DNA. In each motif two zinc atoms are coordinated by six cysteines in a structure comprising two x-helices. (iii) The basic-coiled coil family of transcriptional factors has adjacent dimerization and DNAbinding domains. This class of DNA-binding motifs is also known as the basic region-Zipper (bZip) class. The structure was first proposed according to a comparison of the DNA-binding domains from the yeast factor GCN4,  the oncogenes fos and Jun. and the mammalian transcriptional factor CIEBP  (CCAAT/enhancer binding protein) (Landschulz eta!., 1988a). All domains carry a consensus region of 30 amino acids that contains a repeat of hydrophobic residues (often leucine) at every seventh residue. These heptad repeats of leucine form an oc-helix with a hydrophobic surface.  Adjacent to this dimer-forming  interface is a basic region which is also capable of forming an x-helix. Together, these domains can homo or heterodimenze to form a parallel coiled-coil structure with extensive interactions over the hydrophobic dimer interface (O’Shea et a!., 1991). The basic region, containing about 30 residues, is rich in arginines and lysines, and is primarily responsible for the sequence preferences of the leucine zipper proteins (Agre et a!., 1989). The recent X-ray structure of the homodimeric GCN4 bZip domains-DNA complex shows that the two t\mctionally separated domains form one single uninterrupted ct-helix, with the N-terminal basic helix sitting in the major groove almost perpendicular to the DNA axis (Ellenberger et a!., 1992). However, the basic region alone can bind specifically to its DNA site as long as a disuffide bond is added to allow dimer formation (Talanian et a!., 1990). Furthermore, this region may stay as a disordered structure in  24  solution until bound to DNA when it adopts a helical conformation (Abate et a!., 1990; O’Neil et al, 1990; Pate! et a!., 1990; Weiss et a!., 1990). (iv) The basic “helix-loop-helix” (bHLH) proteins (Murre et a!., 1989a,b) show some similarities with the bZip family. Like the bZip proteins, the bHLH proteins have a basic region that contacts the DNA and a neighbouring region that mediates dimer formation (Voronova and Baltimore, 1990). Amino acid sequence similarity between the basic regions of HLH and Zip classes was first noticed by Prendergast and Ziff (1989) and they proposed that the FILH structure may be very similar to that of the zipper (coiled-coil). As illustrated by the structure of the dimeric Max DNA-binding domain-DNA complex (Ferré-D’Amaré et a!., 1993), the DNA-binding domain consists of two lengthy cL-helices separated by a loop. The N-terminal cL-helix of Max is continuous, and includes residues from the basic and the first amphipathic helix region. The second a-helix, also continuous, is composed of the second amphipathic helix and the additional coiledcoil (zipper) regions. Two monomers of the binding domain form a homodimer of a parallel, left-handed, four-helix bundle.  The two x-helices, containing the basic regions, project from the four-helix bundle  towards the DNA and enter the major groove in opposite directions, in a manner reminiscent of the basic coiled-coil GCN4 (Ellenberger eta!., 1992).  Activation Motifs  The molecular structures of transcriptional activation domains are not as well defined as that of the DNA-binding motifs. However, based on the primary sequences of known activating domains several classes of activating motifs have been formed (reviewed by Mitchell and Tjian, 1989). These motifs were thought to have no similarity in sequence and have no common structure other than an over-representation of acidic residues and other particular residues (for example, 20 to 30% of glutamine or proline). The most extensively studied activating motif is the acid-rich domain, which was first noticed in the activation domains of the yeast GCN4 and GAL4 factor (Hope et a!., 1988; Gill and Ptashne, 1987). The acidic-rich domain is thought to consist of amphipathic cL-helices or unstructured regions that recruit transcriptional factors by virtue of their high density of negative charges (Ptashne, 1988; Sigler, 1988). The strength of activation correlates with the number of negative charges, as observed in the studies of GAL4  25  (Gill and Ptashne, 1987; Gill et a!., 1990), GCN4 (Hope et at., 1988) and VP16 (a herpes simplex virus activator) (Sadowski et a!., 1988; Tnezenberg et a!., 1988; Cress and Triezenberg, 1991). This observation suggested the importance of negative charge. On the other hand, an argument for formation of a 13-sheet structure suggested hydrophobic interface is essential for activation. This paradoxical view was based on recent mutational analysis of the GAL4 acidic-rich domain (Leuther et at., 1993) and biophysical studies on peptides corresponding to the GAL4 and GCN4 activating domains (Van Hoy et at., 1993). However, no structure for VP16 can be associated with the activating domain, so far.  Oligomerization Motif  Oligomerization with members of the same family is often necessary for DNA binding. The surface for multimerization is often closely juxtaposed to the DNA-binding domain, The best characterized oligomerization motif of this kind is the zipper! coiled-coil motif (reviewed in Hu and Sauer, 1992) which was first found in the bZip DNA-binding proteins. Another common oligomerization motif is the “helixloop-helix” structure which contains two amphiphatic c-helices connected by a loop. The second helix is similar to that of the zipper helix and is important for dimer fonnation (Voronova and Baltimore, 1990; Ferré-lYAmard et a!., 1993). Although formation of complexes occurs among factors of the same family, it is not totally restricted as such. Complex formation is also observed between members of families with distinct motifs. For instance, there is genetic and biochemical evidence showing inter-family interactions including those for MADS!homeodomain [particular examples are: cc2IMCM1 (Vershon and Johnson, 1993), SRF!STE12 (Primig et a!., 1991) and MCMI/Phoxl (Grueneberg et a!., 1992)] and TRP/13-transducin-related protein [for example, the SSN6!T(JP1 (Williams et a!., 1991)]. The molecular basis for these interactions is still unknown.  Modular Structures  The concept of ‘modular” structures or functional motifs (Frankel and Kim, 1991) has been useful in characterizing transcriptional factors at the molecular level. Combinations of different motifs, even at the level of a single factor, are evidenced by the discovery of families, such as bZip (Landschultz et at., 1 988a),  26  bHLH-Zip (Vinson and Garcia, 1992), Homeodomain-Zip (Schena and Davis, 1992) and Zinc finger-Zip (Coornaert et at., 1992).  In addition, oligomerization of factors allows sharing of common factors in  different regulatory circuits. In turn, this contributes to coordination of the regulatory networks which account for complex tissue-specific and signalling-dependent transcription in eukaryotic organisms. Combinations of motifs in a single factor and oligomerization of factors do not appear to be a totally random process. This notion is well illustrated, as evolutionary-conserved subunits of factors in yeast and mammalian cells are interchangeable and can functionally substitute for each other (Struhi, 1989). This suggests a conservation in the protein-protein interaction interface.  Therefore, further studies of the  homeodomain-contaimng yeast x2 repressor, its protein-protein interactions, and the modulation of DNAbinding by these interactions, may lend some insight to the structural and functional aspects involved in eukaryotic transcriptional regulation.  1.3.  STRUCTURE-FUNCTION ASPECTS OF THE a.2 PROTEIN Yeast is an excellent eukaryotic model system.  In yeast the c2 repressor is particularly  interesting because of its dual functions in negative transcriptional regulation. Thus far, structure-function studies focusing on c2 repressor have covered the characterization of a broad range of activities including nuclear targeting, selective rapid protein degradation, DNA-binding specificity and protein-protein interactions. 1.3.1.  Overall Strucure of the z2 Protein As predicted from the DNA sequence (Astell et a!., 1981; see Figure 3), c2 is a basic protein of  210 amino acids.  Structurally, it contains two domains separable by proteolytic digestion (Sauer et at.,  1988): The N-terminal domain (residues 1-103) and the C-terminal domain (residues 132-210 which contains the DNA-binding homeodomain) are connected by a linker region which is thought to be flexible and disordered (Vershon and Johnson, 1993),  27  N M K I P I K D L L AAGAAAAAAAGGAAGATAAGCAAGAAAAA ATG AAT AAA ATA CCC AH AAA GAC CTT TTA P G I N T D F E AAT CCA CAA ATC ACA GAT GAG TTT  F L CTC TTT  (10)  S K S I I L N 0 K K (28) AAA TCC AGC ATA CTA GAC ATA AAT AAA AAG  S I C C N K I P I P VT E S T E (46) TCT ATT TGC TGT AAT TTA CCT AAG TTA CCA GAG AGT GTA ACA ACA GAA  E V E E L GAA GAA OTT GAA TTA  R DI G I AGG GAT ATA TrA GGA  F  rrc  L S TTA TCT  A R N K N (64) AGG GCC AAC AAA AAC  .1 R K S 0 E E K L K L a T T s CGT AAG ATT AGT OAT GAG GAG AAG AAG HG TTG CAA ACA ACA TCT  a L T CAA CTC ACT  (82)  T T I T V L L K E M B S I N E 0 R S (100) ACT ACC ATT ACT GTA HA CTC AAA GAA ATG CGC AGC ATA GAA AAC OAT AGA AGT V  N V AAT TAT  Q I CAA CTT  a T K N K S A 0 G L ACA CAG AAA AAT AAA TCG GCG GAT GGG TTG  V F GTA TrT  N V (118) AAT GTG  T a V I 0 M N K S T K P G V B H R F GTA ACT CAA OAT ATG ATA AAC AAA AGT ACT AAA CCT TAC AGA GGA CAC CGG TTF he)ix 1 T E R N V I L S W F E ACA AAA GAA AAT GTC CGA ATA CTA GAA AGT TGG TTT he)ix 2 P V CCA TAT  (136)  A N I E N (154) GCA AAG AAC ATC GAG AAC  I 0 K 0 1 E N I M K N T S L S CTA GAT ACC AAG GGC CTA GAG AAT CAT ATG AAG AAT AAC AGT TAA TCT he(ix 3  (172)  R i a i N K W V S N R B B K E K T II (190) CGC ATT CAA ATC AAA AAC TGG GTT TCG AAT AGA AGA AGA AAA GAA AAA ACA ATA  T I ACA ATC  A P E I A 0 1 S 1 0 E P OCT CCA GAA HA GCG GAC CTC HG AGC GGT GAG CTT  1 A K K (208) CTG GCA AAG AAG  * E K AAA GAA TGA GCCCGAAAAACAAATATGTATATATCTGTGTAGAATATATATATATATATTTCGCAAAAATA  CATAAACAATCAACCCTCTCCTCAGACACTACTAAGATGTTTG  Figure 3. The DNA and Amino acid Sequences ofM4 Tcc2. The DNA and predicted amino acid sequences of M4 TcL2 are indicated (Astell et a!., 1981). The DNA sequence starts from the most 5’ transcriptional initiation site and ends at the most 3’ transcription termination site. Two stable structural domains connected by an unstructured region within the protein are indicated by solid triangles. The single triangle indicates the last residue of the N-terminal domain and the double triangle indicates the first residue of the C-terminal domain (Sauer et at., 1988). The residues in regions of the C-terminal domain corresponding to the helix 1, 2 and 3 of the DNA-binding homeodomain are shown by the boxes of solid line (Wolberger et a!., 1991). In the N-terminal domain, the two dashedline boxes enclose the residues which may constitute the two putative short amphipathic helices involved in the heterodimerization of a1-c2 (investigated in this study).  28  1.3.2.  Structure-Functional Analysis of a2  Nuclear Targeting  Two nuclear targeting signals residing on the N- and C-terminals of c2 have been identified (Hall et al., 1984; Hall et al., 1990).  Various portions of the MA TcL2 gene were fused to the E. coil  (3-galactosidase gene and the hybrid genes were transformed into yeast. The J3-galactosidase activity was localized by cytological immunofluorescence assay following subcellular fractionation. In this way, the first 13 amino acid residues or the C-terminal 141 to 159 residues were shown to be sufficient to direct accumulation of the fusion protein in the nucleus.  This region resembles regions of influenza virus  nucleoprotein (Davey et aL, 1985), Xenopus nucleoplasmin (BUrglin and De Robertis, 1987; Dingwall et a!., 1988), and adenovrnis protein TP 1 and pTP2 (Zhao and Padmanabhan, 1988), all of which have been shown to mediate nuclear accumulation. The short sequence Lys-Ile-Pro-Ile-Lys in the amino terminal region is also found in other yeast nuclear proteins, such as H2B (Wallis et ai., 1980), H2A (Choe et a!., 1982), and H4 (Smith and Andrésson, 1983) and the presumed nuclear protein MATod (Astell et ai., 1981). However, matc2-—galactosidase fusion proteins carrying internal deletions of this region are still localized to the nucleus, leading to the speculation that this sequence may be a signal for rapid uptake into the nucleus but not for steady state accumulation (Hall et at., 1990). The other targeting signal in residues 141-159 was thought to be a more effective signal since internal deletion of this region reduces nuclear localization and causes mutant protein to accumulate at discrete sites on the nuclear envelope. It is noteworthy that this region coincides with a large part of the first helix and its adjacent ioop region (from the fourth residue of helix 1 to the first of helix 2) of the homeodomain (Wolberger, 1991). Hall (1990) pointed out that nuclear localization could be a function of homeodomains in general.  Rapid Selective Protein Degradation  There are two distinct regions within the N-terminal and C-terminal domains that signal the instability of the x2 repressor (Hochstrasser and Varshavsky, 1990). With an in vivo half-life of only 4 to 5  29  minutes, the repressor is among the most short-lived eukaryotic proteins known. By immunoprecipitation of (X2-f3—galactosidase (c2-gal) fusions carrying different portions of o2 with anti-f3 gal antibodies, the amino67 residues and the C-terminal residues from 136 to 210 were demonstrated to destabilize the fusion proteins. Furthermore, fusions with the N-terminal 52 residues or the C-terminal residues after 141 did not show selective degradation of the proteins, suggesting that the essential “signal” for this function must be confined residues 52-67 and 136-141. Degradation of c2 proteins involves two different pathways (Hochstrasser and Varshavsky, 1990). This was indicated by the isolation of two unlinked mutations, doal and doa2, which suppressed the rapid degradation of the amino-67 fusion protein, but not the one carrying the C-terminal degradation signal. A significant fraction of cL2 degradation is dependent on its ubiquitination. Recently, the doa2 gene has been cloned and shown to be the UBC6 gene which encodes one of the many ubiquitin-conjugating (UBC) enzymes (Chen et at., 1993).  These enzymes participate in selective degradation by transferring the  ubiquitin moiety to the substrate protein. This finding is consistent with the earlier observation that c.2 is conjugated to ubiquitin in vivo (Hochstrasser et al., 1991). 1.3.3.  Separate Domains for Dimerization and DNA Binding Sauer and Johnson (1988) demonstrated that c2 is folded into two stable confonnational domains  where the C-terminal domain binds to the DNA target and the N-terminal domain likely mediates homodimerization. The N-terminal domain is thought to be important for repression (Hall and Johnson, 1987). One of its roles is to mediate homodimerization of (x2 and three lines of evidence for this were presented by Sauer and Johnson (1988). First, the close proximity of the N-tenrnnal domains of the homodimer was suggested since intact c2 proteins purified from E. coli or N-terminal chymotryptic fragments are predominantly disuffide-linked dimers. This observation suggests the involvement of the N-terminus in homodimerization. Secondly, as demonstrated by DNase I and hydroxyl-radical protection assays, x2 binds to its operator with a two fold symmetry. Thirdly, the disuffide-linked dimers bind to the natural operator DNA with higher affinity than unlinked monomers or the C-terminal domain alone (20- and 40-fold higher,  30  respectively). Furthermore, only intact c42 proteins which contain the N-terminal domain, show cooperative binding to the DNA. Sauer and Johnson also demonstrated that cx.2 dimers bind with high flexibility. In experiments where the sequences of the wild-type and a variant cc2-operator (carrying a central 13 bp deletion) were used, intact x2 displayed comparable binding affinity to both sequences regardless of the spacing between the two cs2 binding sites (Sauer eta!., 1988; Smith and Johnson, 1992).  Interaction with MCM1 by the Linker Regions of a2  Although the cL2 homodimer alone binds to its two half-sites with great flexibility, MCM1 apparently restricts how cL2 binds to the DNA. By using in vifro DNA-binding gel shift assays, Smith and  Johnson (1992) showed that the c2 dimer, on its own, binds with equal affinity to the wild-type operator and its variants (containing insertions of 4, 14 and up to 100 bp in the central region and containing direct or everted repeats). On the contraiy, only the wild-type operator, not the variants, allowed x2-MCM1 DNA binding. 2-MCM1 binds to the wild-type operator with nearly 1000-fold higher affinity than when cs2 alone binds to the half-site. The flexible linker region of x2 (residues 110-128 between the N-terminal and C-terminal domains) was shown to be responsible for the c42-MCM1 interaction (Vershon and Johnson, 1993) by DNAbinding studies. MCM1 protein was used together with c42 protein expressed and purified from E. coli or  with c2 C-terminal domains (homeodomain) that contained or lacked the preceding 20 linker residues. Cooperative DNA binding of MCM1 was observed only with full length and shorter proteins containing the linker region, but not with the C-terminal domain which lacks the linker. In addition, this region (linker  plus C-terminal domain) is sufficient to confer MCM1 cooperative binding to the Drosophila homeodomain (from the engrailed protein) (Vershon and Johnson, 1993).  Random Mutagenesis on the Homeodomain  Extensive effort has focused on studying the C-terminal DNA-binding domain of c2 because of the similarity it shares with the homeodomain family.  These efforts mainly include structural and  31 functional analyses on the homeodomains DNA-binding activily in cL2-repression (Porter and Smith, 1986; Hall and Johnson, 1987; Phillips et al., 1991; Wolberger et al., 1991; Lorimer et a!., 1992). Hoewever, there have been few studies on the structural aspects of the c2 repressor regarding its al-cL2 repressor function.  Strathern et al. (1988) and Harashima et al. (1989) have independently investigated the c2  function in al-x2 repression by using random chemical mutagenesis and genetic selection or screening. Missense mutations which affect only the repression of diploid-speciflc genes (a1-o2 repression) but not aspecific genes (x2 repression) were isolated in order to define the amino acid residues of the c2 protein that are essential for the al -cL2 repression but not the x2 repression. Harashima et a!. obtained three mutants that each have an amino acid replacement of Tyr or Phe for Cys codon 33 in the x2 cistron of HMLcc or Iv.L4 TcL. They proposed that this Cys residue may be important for al-a.2 interaction, and speculated that it may stabilize dimers by forming a disuffide bridge or by serving as a ligand to form metal-linked dimers as described for the tat protein of human immunodeficiency virus (Frankel et a!., 1988). Strathern et a!. (1988) found another a/ce-specific repression-defective missense mutation with a Leu to Ser replacement at codon 196. This residue, in the region outside the last helix of the homeodomai n DNA-binding motif was thought to be important for a1-x2 interaction, or alternatively for sequencespecific recognition required for ahx-specific repression but not for repression in c cells. Consistent with the first explanation, NMR studies on the interaction between the cL2 homeodomain and al homeodomain suggested a direct interaction through the oL2 “tail” region (residues 196-203) C-terminal to the homeodomain (Phillips, 1992). Therefore, residue Leu196 may be important for maintaining the function of this C-terminal “tail”, presumably by stabilizing the al-x2 dimer. This view is supported by the recent finding that mutant x2 missing residues 188-210 forms a less stable al-x2-DNA complex than does the wild-type c42 protein (Mak and Johnson, 1993).  32  1.4.  1.4.1.  DNA-BINDING HOMEODOMAIN OF TUE a2 REPRESSOR  x2 Homeodomain The homeodomain, encoded by the homeobox, is a DNA-binding domain which resembles the  bacterial “helix-turn-helix” DNA-binding motif The homeobox is a highly conserved segment that was first found in the homeotic genes of Drosophila (McGinnis et al., 1984; Scott and Weiner, 1984) and subsequently in many other eukaryotic organisms, including vertebrates and human beings (reviewed by Scott et a!., 1989). The cv2 homeodomain was first noticed when a comparison of the primary amino acid sequences (Laughon and Scott, 1984; Shepherd et a!., 1984) and the C-terminal portions of the M4 TcL2 and MZ4 Ta] genes showed that they were similar to the homeodomain of the Drosophila homeotic genes. Porter and Smith (1986) carried out semi-random cassette mutagenesis to saturate the homeodomain’s homologous region with missense mutations and tested the in vivo functions of these mutants to repress both a-specific and haploid-specific genes. The results demonstrated that the mutation of invariant residues in the homeodomain genes destroyed both repression of a- and haploid-specific genes. Subsequently, a correlation between DNA-binding to the upstream region of the a-specific gene (BAR 1) and in vivo repressor functions was illustrated by using in vitro translated wild-type and mutant c2 proteins (Porter, 1987). The motif was shown to be essential for repressor activity because of its direct participation in DNA binding. Furthermore, it was pointed out that a subclass of missense mutants affected repression of a-specific genes only and not haploid-specific genes (Porter and Smith, 1986). These mutations affect the region following the homeodomain homology, a region believed to be involved in cL2-12 DNA binding but not a1-c2 binding (Porter and Smith, 1986). Therefore, a slightly different role for the c2 polypeptide in each of its two functions was speculated. By a deletion study using x2-3gal fusion proteins, Hall and Johnson (1987) reached a similar conclusion that the homeodomain-like region in yeast cL2 protein functions as a sequence-specific DNAbinding domain. In addition, they demonstrated that the homeodomain alone is not sufficient for repression,  33  since a mutant carrying the C-terminal amino acids 136-2 10 still exhibits in vitro DNA-binding ability, but does not repress transcription in vivo. A deletion as minimal as the N-terminal residues 4-7 eliminated both repression of a- and haploid-specific genes. Consistent with these findings, Sauer et al. (1988) showed that the C-terminal domain containing the homeodomain can bind alone with the same specificity as judged by DNA-footprinting experiments, but with significantly less affinity (40-fold less). Lorimer et a!. (1992) further characterized the DNA-binding domain by randomizing the homeodomain residues, Ser5O and Arg54, which may be involved in specifying the DNA-binding sequence of the (x2 binding site as implied by the cases of bicoid (Manes and Brent, 1989) and antennapedia (Kissinger et a!., 1990). The results suggested a restriction of amino acid types at the Ser5O position and also unveiled an absolute requirement for Arg54.  These findings are consistent with the published  structural studies on the x2 homeodomain (Wolberger et a!., 1991), in which the residues Ser5O and Arg54 were shown to specifically contact the DNA bases at the recognition site through van der Waals interaction and hydrogen bonding, respectively. 1.4.2.  DNA-Binding Specificity of a2 Homeodomain Although the isolated homeodomain may not truly reflect the DNA-binding specificity of the  native protein in vivo (where interactions with other factors such as MCM1 and al take place), valuable infonnation has been obtained from the structural studies of the cL2 homeodomain.  This structural  information has explained the biochemical and genetic results, and has brought insight to how the homeodomain may determine DNA-binding specificity. The secondary structure of the c2 homeodomain was first determined by using solution ‘H and N NIvIR spectroscopy (Phillips et a!., 1991) of a purified c42 fragment (residues 128-210) expressed in E. 5 ‘ coli. The NMR study revealed the existence of three helical segments in the cL2 fragment. The positions and the length of these helices corresponded well to those of the Drosophlia antennapedia (Antp) and engrailed (en) homeodomains.  These results agreed with the X-ray crystallographic study on the  homeodomain-DNA complex using the same purified x2 (128-210) fragment from E. coli and a 21 bp synthetic oligonucleotide duplex corresponding to the variant x2 operator (which has the central 12 bp  34  deleted and the two x2-binding half-sites from the STE6 operator adjacent to each other) (Wolberger et al., 1991; see Figure 4). The high-resolution 2.7A cL2-homeodomain-DNA co-crystal provided an overall structure which can be visualized as shown in Figure 4 (Wolberger et at., 1991). Helix 1 and helix 2 pack against each other in an antiparallel arrangment.  Helix 3 is roughly perpendicular to the first two helices.  The  hydrophobic face of the extended helix 3 packs against helices 1 and 2 to form the interior of the compact protein domain. The main contacts with the DNA target in the c2 co-crystal are made by residues in helix 3, which fits directly into the major groove, and by residues in the extended N-terminal arm, which fits into the minor groove (Figure 4 and Wolberger et at., 1991). Earlier genetic and biochemical analyses revealed that missense mutants of Asn5 1 and Trp48 or G1n44 were nonfunctional (Porter and Smith, 1986). The X-ray data confirmed the importance of these residues by showing that these residues make specific base pair contacts and contact the phosphodiester backbone. Residues near the middle of helix 3, namely Asn5 1, Arg54 and Ser5O, are closest to the bases and make direct contacts. Asn5 1 and Arg54 form hydrogen bonds both to each other and to base pairs 4 and 5 (see Figure 4, for the numbering scheme of the bases). Ser5O contacts the thymine at base pair 3 presumably by van der Waals interaction, as the distance between the base and the side chain is too long for a hydrogen bond. Therefore, these three residues of helix 3 may define the TGT core of the c2 recognition site (CATGTAATT). This core sequence is very important since mutations in this region resulted in losses in both in vivo repression and in vitro c2-DNA binding (Lorimer et at., 1992; S.S. Ncr and M. Smith unpublished data). In addition, the extended N-terminal arm fits into the minor groove with residue Arg7, making base contact to the two thymines at base pairs 8 and 9. Furthermore, as shown in Figure 5, an extensive set of contacts with the sugar-phosphate backbone was made along the edge of the major groove, on both strands, mainly by the residues in helix 3 and, also to a lesser extent, by residues not in helix 3 (Phe8 in the N-terminal arm and the Tyr25 in the loop connecting helix I and helix 2). Comparison of the x2 co-crystal with another homeodomain-DNA complex structure, the engrailed-DNA co-crystal, revealed remarkable structural similarities between them, despite limited amino  35  Figure 4. The cw2 Homeodomain-DNA Co-Crystal Model (A) The sequence of the DNA duplex used in the cL2 homeodomain-DNA co-crystal structure: the duplex is a variant i2-operator which is derived from the operator sequence of the STE6 gene by deletion of the central 13 bp (Wolberger eta!., 1991). (B) The consensus DNA sequence of the x2-binding half-sites: this panel shows the alignment of ten cL2binding half-sites from the c2-operators of five a-specific genes (adapted from Wolberger eta!., 1991). The dotted-line boxes show the core sequence (TGTA) and the invariant thymidine (T at position 9) of the c2binding half-sites. (C) The ci.2 homeodomain-DNA co-crystal model showing the base-specific contacts: the sketch displaying a side view of the x2-homeodomain (ie. perpendicular to the axis of helix 3) shows the specific base contacts to the TGT core (bp 3-5) in the major groove by helix 3 residues (ie. Ser5O, Arg54, Asn5l); also, contact with base pairs 8 and 9 in the minor groove by the N-terminal arm (Arg7) is shown. (figure adapted from Wolberger et a!., 1991; for the numbering of the homeodomain residues, see Figure 5).  36  A  -1  1  2  3  4  S  6  7  8  9  10  aCATGTAATTc ATTTACACGC GTACATTAAg TAAATGTGCGt 9  8  STE6  CA;TGTA:ATT! C GT G T AlA AlT  8AR1  CG!TGTAATiT C AT G T AA TT!  STE2  CATGTA..CT;T C AlT G T AlA A!T  MFA1  TG:lTGTAATT C AT G T A .A AT 1 CAi,TGTATTl:T C Ai.T G T AlA A!T  MFA2 consensus  T  C AT G T AlA TIT  Figure 4. The cs.2 Homeodomam-DNA Co-Czystal Model  6  S  4  3  2  1  -1. -2  37  acid similarities. The two protein-DNA complexes can be precisely superimposed, suggesting a conserved docking mechanism by which all homeodomains interact with their cognate DNA target (Figure 5; Wolberger et a!., 1991).  In the proposed mechanism, the helix 3 is oriented and aligned by a set of  conserved “side chain-DNA” contacts while the variable “specificity-defining” residues make specific contacts with the bases of the recognition sequence. The conserved residues and the invariant residues form a DNA-binding scaffold. This scaffold provides contacts with the phosphodiester oxygen atoms along the recognition site to orient the helix 3 in the major groove. The invariant Asn5 1 aligns the helix, by making hydrogen bonds with an adenine (Treisman et a!., 1989; Otting et a!., 1990; Percival-Smith et a!., 1990; Hanes and Brent, 1991; Wolberger et a!., 1991). Furthermore, as in the cases of the homeodomain-DNA complexes of x2, engrai!ed and antennapedia, the residues at the N-terminal arm projecting into the minor groove provide an additional source of binding specificity. The homeodomains make only a small number of base-specific contacts which do not totally account for the DNA specificity.  This prompted the question of how other parts of the protein may  modulate or help in defining the DNA-binding activity of the homeodomain. For example, (12 binds to a different DNA target in the presence of the al protein (Goutte and Johnson 1988; Dranginis, 1990); c2 also binds cooperatively with MCM1 with higher affinity to the DNA target than (12 does alone (Keleher et a!., 1988). It is not clear how c2 defines the binding specificity beyond the TGTA core. It is believed that MCM1 overlaps with c2 at the binding sites, so MCM1 might actually dictate the sequence preference outside the TGTA core (Wolberger eta!., 1991). Based on their recent work regarding the function of the short N-terminal extension of the (12 homeodomain, Vershon and Johnson (1993) suggested that an extension-homeodomain arrangement could explain why groups of homeodomain proteins appear to have similar DNA-binding specificities in vitro but quite different target speciflcities in vivo. By deletion studies, they demonstrated that approximately 20 residues (110-128) adjacent to the homeodomain are required for cooperative binding with MCM1 to the (12 operator. This short region between the N-terminal dimerization domain and the C-terminal homeodomain was thought to be flexible and sufficient in mediating a specific interaction with MCM1. When grafted to the Drosophila engrailed homeodomain, this unstructured extension confers cooperative DNA binding with  38  Figure 5. DNA-Binding Specificity of the ct2 Homeodomain. (A) Sequence alignment of x2 with other homeodomains: the homeodomain sequences of a2, engrailed (labelled as en), and al are shown. The amino acids are numbered according to the convention used by Wolberger et a!. (1991). Boxes indicate the three x-helices in both x2 and engrailed. The residues 129-151 and 155-190 of c2 correspond to homeodomain consensus residues 1-23 and 24-5 9, respectively; the c2 homeodomain contains a 3-residue insertion (residues 152-154) which is designated as residues 23a, b and c. The al residues 69-126 are shown in this alignment. Residues that are conserved in all homeodomains are marked by asterisks below the sequences. Four invariant homeodomain residues, Trp48, Phe49, Asn5 1, and Arg53 are indicated by bold-type, italic letters below the amino acid sequence (although, note that c2 has a Va149 rather than a Phe49). (B) A summary of the contacts within the cL2 homeodornain-DNA complex (adapted from Wolberger et a!., 1991): the DNA is represented as a cylindrical projection, and the shading represents the TGT core of the oL2-binding half-site. Phosphates are indicated by open circles, and phosphates which are contacted by the c2 homeodomain are indicated by hatched circles. The conserved or invariant residues (ie. PheS, Tyr25, Gln44, Trp48, and Arg53) form a “DNA-binding scaffold” that contacts the phosphate backbone. Another invariant residue, Asn5 1, contacts an ‘A’ residue to align helix 3 in the major groove. Variable “specificitydefining” residues of the homeodomains are involved in specific base contacts in the major groove. In the case of x2, Ser5O and Arg54 contact the ‘T’ and ‘G’ of the TGT core sequence of the cL2-binding half-sites. In the minor groove, bases 8 and 9 of the half-sites are contacted by Arg7.  01 Ui  W  —  -‘  VI  *  z  rn,o  Crfl  zzz  —,o  rr  rnz  rn  i1e  rrr  ‘ni  zz  .1111n  rrr  >0 Z  Or!1r!1  40  MCM1 to an operator containing both the engrailed recognition sequence and the MCM1 binding site. The extension-homeodomain arrangement was hypothesized to be a general theme by which homeodomain proteins specify cooperative interaction with an additional factor to define new DNA-binding specificity. The rationale for this hypothesis was that: (i) Comparisons of primary amino acid sequences of homeodomains have revealed additional sequence similarities that lie outside the homeodomain, in positions correlating to the rx2 homeodomain extension.  For example, a short sequence, YPWM, is found just  upstream of the homeodomains of a number of Drosophila regulatory proteins including ultrabithorax, antennapedia, abdominal-A, and deformed (Scott et al., 1989). (ii) Interactions between c2 and MCM1 may be in part conserved, since x2 interacts with SRF (human serum response factor) to cooperatively bind DNA (SRF shares great similarity with MCM1 and also recognizes the MCM1 site) (Primig et al., 1991).  1.5.  PUTATiVE COILED COIL DIMERIZATION MOTIFS OF (z2  1.5.1.  Requirement of the N-terminal Domain for Protein-Protein Interactions The functions of c2 in repressing two sets of target genes, namely the a-specific and haploid  specific genes, is dependent on its ability to form both cc2-cQ homodimers and al-c42 heterodimers. Although information obtained from genetic and biochemical analyses clearly indicate the essential role of cL2’s N-terminal domain for dimer formation, the molecular basis of dimerization has still not been explored.  2 Ilomodimers  The existence of functional homodimers was originally proposed by Johnson and Herskowitz (1985) based on DNA-binding studies using c2-f3—ga1actosidase fusion proteins. When the fusion proteins were purified from a cell extract of the yeast transformant expressing the ft2-—galactosidase, four types of tetramers were obtained. There were tetramers of f3—galactosidase with none, one, three, or four complete c2 moieties attached, as a result of the natural selective degradation of the cc2 moiety (Johnson and Herskowitz, 1985; Hochstrasser and Varshavsky, 1990). Purified tetramers of J3—galactosidase with none or one complete (x2 moiety attached did not demonstrate any specific DNA-binding activity to the STE6  41  operator, while that with three or four complete c2 moieties did (Johnson and Herskowitz, 1985). This observation led to the conclusion that x2 may bind to the DNA target as a dimer. Later, more supporting results from DNase I and hydroxyl-radical protection assays (Sauer et at., 1988) indicated that c2 binds to the two ends of the STE6 operator but leaves the center free, thus demonstrating a two-fold symmetry.  a1-c2 Ileterodimers  As described earlier, the al-cL2 heterodimer has a one to one stoichiometry of al and (x2 molecules (Dranginis, 1990). Heterodimers bind to the al-x2 operator cooperatively. This DNA-binding activity requires amino acids 21-62 from the N-terminus of cL2 (Goutte and Johnson, 1988) but not the first 20 N-terminal residues (Goutte and Johnson, 1988; Dranginis, 1990). This indicated that the N-terminal residues 21-61 are required for cL2 to interact with al. By determining the amino acid sequence of cL2 required for the dominant negative phenotype in the a/cL diploids, Hall and Johnson (1987) concluded that the region containing residues 20-140 is essential for c2’s interaction with al. Deletion of crucial information for in viva repressor function (for example, the deletion of the N-terminal residues 4-7 or elimination of a part of the homeodomain) resulted in non functional mutant matc’2 alleles. Over-expression of these mutant alleles in diploid cells, which contain wild-type al-n2, led to a decrease in the normal endogenous al-cL2 repression; thus, these mutants exhibited a dominant negative phenotype. Such trans-inactivation of the wild-type cL2 was attributed to the possibility that this mutant of cL2 may be still capable of interacting with al. However, deletion mutants of x2 lacking amino acids 21-62 or 68-140 do not exhibit the dominant negative phenotype.  Therefore, this mutant  cannot interact with al and an al-cL2 interaction site was proposed to lie between amino acids 20 and 140 of cL2 (Hall and Johnson, 1987). Consistent with the above notion, missense mutations in this region (Tyr or Phe replacements of Cys at position 33) yielded al-cL2 repression-deficient mutants which failed to repress haploid-specific genes but did repress a-specific genes (Harashima et at., 1989).  This Cys residue was postulated to play an  important role in al-cc2 interaction and in the stability of al-x2 dimers. Since there is a Cys residue in the al protein (Astell et at., 1981), Cys33 of cL2 was thought to stabilize the heterodimer by forming a disulfide  42  bridge or by serving as a ligand to form metal-linked dimers as described for the tat protein of the human immunodeficiency virus (Frankel et al., 1988). Regardless of the validity of such a hypothesis, Harashima’s experiment located a mutational “hot spot” on the x2 cistron, as three genetically defined al-x2 repressiondeficient mutations all occur in the Cys codon at position 33. Such an observation prompted a search for critical structures in this region of the N-terminus. 1.5.2.  Putative Coiled-Coil Motifs  3,4-Hydrophobic Heptad Repeats in 2 and al  In this thesis work, a search for structural motifs in the N-terminal dimerization domain was attempted in order to understand the molecular nature of the a1-c2 interaction site on the c2 protein. Examination of the amino acid sequences in the regions of cL2 between residues 20 to 62 and 67 to 103 has  led to the identification of two stretches of 3,4-hydrophobic heptad repeat sequences in the a1-c2 interaction site. These heptad repeats are consistent with the coiled-coil protein folding motif (Sodek et a!., 1972; reviewed by Hodges, 1992). The two repeat motifs are termed x2A and cL2B, and extend between residues 11e22 and Va143, and residues 1le67 and 1le95, respectively (Figure 6). Similar motifs were also found in the b-Zip class of transcriptional factors (Landschulz et at., 1988a; Vinson et at., 1989) and are known to mediate dimerization. The term ‘leucine zipper’ was first used to describe a dimerization motif which involves helices containing leucine residues at every seventh amino acid. Hypothetically, when two ‘leucine zipper’ helices come together on the interface, the leucine side chains projecting out of each helix interdigitate (Landschulz et al., 1988a). This class of proteins has been shown to adopt a two-stranded coiled-coil structure (Oas et a!., 1990; O’Shea et a!., 1991) consisting of two amphipathic right-handed x-helices aligned so that their non-polar faces are continuously adjacent, with residues in the a and d positions of each helix (abcdefg represents the seven residues of one complete turn of the helix) making up the hydrophobic interface (illustrated in Figure 7).  43  A 22  18  29  36  32  39  43  FKSSILDINKKLFSICCNLPKLPESV  cx2A  def g  67  cL2B  25  abc  71  def g  74  abc  78  de  81  f gab c de f g a  85  92  88  95  ISDEEKKLLQTTSQLTTTITVLLKEMRSI d e  f  g a  b  C  d e  fg a b  c d e  B  f  g a  b  cd e  /  g  a b cd  37 P C 33  30  e  34 C  g  , 3 S  35 N  Figure 6. The Putative Dimerization Motifs of cL2. (A) Two 3-4 heptad repeats, termed cL2A and c2B, are identified at the N-terminus of cL2, These heptad repeats can be designated (abcdefg) and have hydrophobic residues at the a and d positions. In this diagram, the amino acid positions in the primary sequence of x2 are labelled above the one letter code sequences. The heptad positions (abcdefg) are shown below them. Each of the boxed sequences may adopt an x-helical conformation and form a coiled-coil structure with a compatible ix-helix of al (see Figure 7). (B). The helical wheel presentation of cL2A (top view): Besides the a and d positions, the e and g positions which are close to the “a, d” hydrophobic interface are also important for the stability of a coiled-coil structure. On the other hand, the b, j and c positions are usually less significant. In this wheel diagram, only the boxed residues 1le22 to Pro37 of the x2A region in part (A) are represented.  44  Figure 7. A Scheme of Two-Stranded Coiled-Coil Structure. The side and end views of a “paint-roll” model of the two-stranded coiled-coil are taken and adapted from O’Shea et at. (1992). A 3,4-hydrophobic heptad repeat can potentially adopt a right-handed cc-helical conformation which has one complete turn for every seven residues (abcdefg). With residues in the a, d positions of each helix making up the hydrophobic interface, two such amphipathic cc-helices align in parallel so that their non-polar faces are continuously adjacent. This interaction results in a left-handed superhelix. A detailed illustration of a GCN4 coiled-coil structure can be seen in Figure 9B.  45  In order for c2 to heterodimerize with al via a coiled-coil interaction, 3,4-hydrophobic heptad repeats in the al protein are necessary. Indeed, examination of the al primary sequence revealed two such overlapping sequences extending between Meti and 11e18 and between Phel6 and 11e37 (Figure 8). These two sequences would give rise to two separate but partially overlapping hydrophobic faces if the entire Meti to 1le37 segment adopts a single ce-helix. Immediately, this observation suggested a model for specific a1-c2 heterodimerization in which the two short putative amphipathic x-helices of cL2 align with the two putative hydrophobic faces of al, each pair in a coiled-coil manner. 1.5.3.  Coiled-Coil (Leucine Zipper-Like) Motifs The coiled-coil is a common stabilizing motif in proteins of all types (reviewed in Cohen and  Parry, 1986). This structure is involved in a wide range of biological functions which include structural functions [as in fibrous proteins such as keratin (Coulombe, 1993) and myosin (Titus, 1993)J, enyzmic function (as in tetramerization of the bacterial J3-galatosidase, Alberti et at., 1991) and physiological activities [as in calcium-binding protiens (Ballinger et a!., 1993) and a secreted DNA-binding protein, nucleobindin (Miura et a!., 1992)]. In recent years, the coiled-coil structure became most familiar as the protein-dimerization motif among different classes of DNA-binding proteins. These include the “helixloop-helix” (Max, Ferré-D’Amarë et at., 1993), zinc DNA-binding motif (GAL4, Marmorstein et a!., 1992), and the basic-region leucine zipper (GCN4, FOS, JUN, AP4 etc., reviewed by Hu and Sauer, 1992). A two-stranded coiled-coil is comprised of two amphipathic, right-handed x-helices which adopt a left-handed supercoil such that their nonpolar faces are continously adjacent. Crick (1953) proposed this folding motif for keratin. Tropomyosin was also found to be made up of a two-stranded x-helical coiled-coil and the 3,4-hydrophobic repeat pattern that is responsible for the formation and stability of coiled-coils was identified (Sodek et a!., 1972). The repeats contain hydrophobic residues at positions a and d of the abcdefg heptad sequence. The first high-resolution crystal structure of coiled-coils was obtained by Alber and co workers (see Figure 9; O’Shea et a!., 1991) for the GCN4 leucine zipper dimerization domain. Recent X-ray structures of DNA-bound transcriptional factor fragments (GCN4IDNA, Ellenberger et at., 1992;  46  A 10  20 a’  d’  30 a’  40 a’  d’  MDDI CS MAENI NRTLFNI LGTEI DEINLNTNNLYNFIMES d  a  d  ab  de  b  alA  e  b  b  e  aiB  C  B  N  alA 22 6E 9 N  q 1 R 7 N 0 3 Y 34 f  NE 31  25  aiB 14 T  21 T 28 L  g  10 N  1)34 N3’  C  Figure 8. Putative Overlapping Coiled-Coil Motifs of al. (A) The primary sequence of al is predicted according to the work of Astell eta!. (1981). The two overlapping arrays of heptad repeats, alA (Metl to Leul8) and aiB (Phel6 to 11e37), are indicated with boxes. Positions a, d, b, and e are shown below the sequence. The hydrophobic positions of alB are noted as a’ and d’ above the sequence. (B) The helical wheel presentation of alA and aiB: the residues on the two nearly orthogonal hydrophobic faces are indicated within the shadowed boxes. (C) A schematic view of the hypothetical al dimerization helix (shown as a cylinder): the two hydrophobic faces of alA and aiB are shown in shadowed and black areas, respectively.  47  A C  E  A V  R H  KNS  Y A K-.  B  Figure 9. The Two-Stranded Parallel Coiled-Coil Structure of GCN4.  48  C  N  C  N  C  5 Leu  12 Leu 9 Va1  6 Asn’  19 Leu  Leu 23 Va1  VaP°  Figure 9. The Two-Stranded Parallel Coiled-Coil Structure of GCN4 (Cont.). (A) The helical wheel presentation of the GCN4 coiled-coil illustrates the major forces involved in coiledcoil stability. The GCN4 sequence from the 250th to the 281st residue NKQjEDKyEELLSKN YNLENEVARLKKLVGER) corresponds to the sequence which forms the coiled-coil. The view is from the N-terminal methionines labelled within the boxes. The next six amino acids of the first repeat are circled. Residues at the heptad positions, a, a’, g and g’ (darkly shadowed shapes), alternate in the interface with residues at d d’, e and e’(lightly shadowed circles) from one layer to the next. Hydrophobic interactions at the interface determine the overall coiled-coil stability. Intra- and interhelical ion pairs also influence the coiled-coil stability. Interhelical ionic interactions between the e and g positions and intrahelical interactions between the c and g positions are indicated by dashed lines. (B) The end view of the X-ray ciystal structure of the GCN4 coiled-coil: this view is along the superhelical axis from the N-terminal end. The main chain is highlighted with a twirled ribbon and the reduced van der Waals surfaces of the side chains at positions a and d are stippled in yellow. The helices are curved with an overall superhelical twist of nearly 90 degrees. (C) The side view of the GCN4 coiled-coil as seen perpendicular to the superhelix axis. The buried side chains of residues in the 3,4-hydrophobic repeat are highlighted in bold while the residues themselves are identified below. The conserved leucines at position d and the alternate’ hydrophobic residues at position a make up the successive layers of the interface. Diagrams in this figure are adapted from Alber (1992) and O’Shea eta!. (1991).  49  GAL4/DNA, Marmorstein eta!., 1992; Max/DNA, Ferré-D’Amaré eta!., 1993) and tropomyosin (Whitby et a!., 1992) revealed coiled-coil conformations that generally agree with Alber’s structure (Figure 9). Dimerization leads to a greater variety of transcriptional regulators than would a few simple monomers.  It is the goal of this thesis to deal with the specificity and mechanism of dimerization.  Therefore, it is worthwhile to review the individual forces contributing to formation and stability of coiled coils. Dimerization specificity of transcriptional factors through the coiled-coil motif is determined by the stability of the coiled-coil and by the compatibility of the residues at the hydrophobic core interface as well as the exterior residues. In the hydrophobic core, the role of hydrophobicity in the formation and stabilization of coiledcoils was apparent in the highly conserved 3,4-hydrophobic repeat sequence of these structures. Experiments using synthetic model peptides corresponding to the tropomyosin coiled-coil show that hydrophobic packing and the hydrophobicity within the interface is the major contributing force to overall stability (Hodges et a!., 1990; Zhou eta!., 1992a, 1992b; Zhu et a!., 1993). The X-ray crystallographic data (O’Shea et a!., 1991) also indicated that in parallel coiled-coils, the side-chains of residues at position a and d are almost totally buried in the dimer interface, suggesting that the hydrophobic effect is a major driving force for dimerization. In this interface, the a and d side-chains of one helix pack next to the a’ and d’ sidechains, respectively, of the other helix, showing in-register a-a’ and d-d’ interactions (Figure 9). With large bulky hydrophobes (such as Leu, lie and Val) at a and d the a-a’ pair is capable of van der Waa!s interactions with d-d’ pairs above and below. Therefore, a continuous hydrophobic core is maintained along the axis of the coiled-coil. Polar and helix-breaking residues implanted in the core are usually disruptive (Hu et a!., 1990).  Replacement of core Leu residues with less hydrophobic Ala was also shown to  destabilize the synthetic coiled-coil (Zhou eta!., 1992a, 1992b). The a and d positions are structurally distinct and contribute differently to the stability.  Ala  replacements of Leu residues at the a position of the synthetic coiled-coil (usually containing Leu residues at both the a and d positions) were more destabilizing than the same replacements at d (Zhou et a!., 1992a,b). Replacement of Val or Ile for Leu in synthetic coiled-coils exhibited a preference for f3-branched amino acids at a, but not at the d position (Zhu et a!., 1993).  In fact, such preferences regarding stability is  50  consistent with frequent occurrences of Val, Tie, and Thr at the a site in native proteins (Hu and Sauer, 1992). Furthermore, the a position is more tolerant of charged and strongly polar replacements than the d position. The difference in tolerance of these positions to hydrophilic replacements is accounted for by the fact that orientations of the side chains differ with respect to the long axis of the coiled-coil at the two sites. At the d position, side chains point directly into the dimer interface, while the same side chains point outward at the a position. This allows the polar residues with lengthy hydrophobic side chains to contribute to core packing while their polar end groups extend to the solvent-exposed exterior. Ionic interactions in coiled-coils were observed and also thought to be important for stability. For example, interchain ionic interactions between e and g side chains (Lys and Glu pairs) were reported in the GCN4 coiled-coil and a synergistic relationship between these salt bridges and the hydrophobic core was presumed (O’Shea et at., 1991). By extending across the edge of the hydrophobic interface, the salt bridges are stabilized by a local environment which is not entirely aqueous. In turn, the salt bridges reduce the solvent accessibility of the hydrophobic core.  As well, the methylene groups of e and g side chains  contribute to core hydrophobicity. O’Shea eta!. (1992) and Schuermann et a!. (1991) have investigated the role of residues in the e and g postions, as well as the hydrophobic core residues, in directing heterodimer specificity in Fos and Jun.  Both groups demonstrated that general electrostatic effects are important in  determining the ability of two charged helices to associate. More stable heterodimers tend to have net charges close to zero for the interfaces (which are the a, d, e, g residues of both helices). The predominantly acidic residues at the e, and g positions of Fos are responsible for preventing Fos-Fos homodimer formation since the homodimers are destabilized at physiological pH (O’Shea et a!., 1992). The basic residues at the g, and e position of Jun are complementary to Fos at that pH and stabilize the Fos-Jun heterodimers by electro attraction. Electro-interactions, therefore, drive the equilibrium in favor of these Fos-Jun heterodimers. Although, the work of O’Shea et at. (1992) suggests that interactions between e and g residues are sufficient to direct heterodimer specificity, there are observations which argue that a and d residues also play a significant role in determining the overall stability of these heterodimers (Schuerman et at., 1991). That is, coiled-coils containing identical numbers of general and specific ionic interactions but differing only in a or d have different stabilities. Conversely, Lovejoy et at. (1993) suggested that interhelical salt bridges are  51  not a dominant driving force in the formation of coiled-coils but rather an indirect consequence of coiledcoil formation. Their conclusion is based mainly on the observation that grouping of like charges between the e and g positions of antiparallel helices is allowed since these charged side chains have sufficient conformational flexibility to move away from each other. Intrastrand ionic interactions (of the type I to 1+3 and i to 1+4) could add additional stability to the coiled-coils. Such solvent-exposed intrastrand salt bridges were reported for the GCN4 X-ray structure (O’Shea et a!., 1991). In model peptides, these ionic interactions stabilize the individual helix (Marqusee and Baldwin, 1987) and thus directly influence coiled-coil stability (Zhou et a!., 1992c). Although it seems logical to assume an important role for intrahelical salt bridges between charged amino acids [as these ionic interactions are a conserved feature of a native coiled-coil protein, troponin C (Sundaralingam et a!., 1985)], the physiological significance of this type of ionic interaction for coiled-coil stability is still unclear. The number of residues in a stable coiled-coil varies from as few as 14 (GAL4-DNA complex, Marmorstein eta!., 1992) to as many as 1086 (myosin, McLachlan and Karn, 1982). The minimal length of the stable model peptide coiled-coils, which is based on the consensus 3,4-hydrophobic repeat sequence of tropomyosin, is about four to five repeats (Hodges et a!., 1988; DeGrado et a!., 1988). Furthermore, ideal 3,4-hydrophobic repeats may give rise to three- or four-stranded coiled-coils, in both parallel or anti-parallel configurations (Banner et a!., 1987; Cusack et a!., 1990; Lovejoy et a!., 1993).  Although hydrophobic  packing in antiparallel coiled-coils has been shown to provide greater stability than in parallel coiled-coils (Adamson et a!., 1993; and reference therein), coiled-coils tend to pack parallel. interactions must direct the orientation of these helices.  Therefore, other  Polar sidechains buried in the hydrophobic  environment of the interface provide a way to achieve such control in the packing orientation. CAP (Weber and Steitz, 1987) and GCN4 (O’Shea eta!., 1991) are parallel coiled-coils known to contain hydrogen-bonds in the interfaces.  The parallel packing minimizes the destabilizing effect of polar side chains in the  hydrophobic core. In summary, hydrophobic interaction is the predominant driving force of coiled-coil formation. Inter- and intrastrand ionic interaction also influence coiled-coil stability. Thus, in general the residues in the a, d or e, g positions (of the interfaces) are most important for stability while those in the b, c and f  52  positions, although showing a higher frequency of charged or hydrophilic residues, usually do not have significance in coiled-coil formation per Se.  1.6.  1.6.1.  OBJECTiVES OF THESIS PROJECT  Characterization of the Dimerization Domain of a.2 The purpose of this thesis project was to understand the structure and function relationships of the  MATcL2 repressor. Since protein-protein interactions play an important role in the function of the repressor, the primary goal of the present study was to define the mechanism by which o2 repressor dimerizes with another transcriptional regulator, al. Some recent advances in the protein-protein interaction aspects of the c.2 repressor have been reported such as the involvement of the linker region of the cL2 protein in MCM1 interaction and the N-terminal domain in homo- and heterodimerization (Vershon and Johnson, 1993). It is still not known how c42 interacts with its partner to regulate cell-type specific genes.  A complete  understanding of the mechanism by which the N-terminus dictates dimerization would require structural analysis of the domain using NMR or X-ray crystallography. At the moment, while these structural data are missing, mutational analysis of the potentially significant protein-protein interaction motifs is instructive. The functional significance of the two potential dimerization motifs, consistent with the coiledcoil or 3,4-hydrophobic repeats on the N-terminal domain of the c2 repressor, in the heterodimerization with al was examined in this study. Oligonucleotide-directed site-specific mutagenesis was used to alter the structure of the putative dimerization motifs. The effects of the mutations on al-cc2 dimer formation was measured directly using a DNA-binding gel mobility-shift assay, co-immunoprecipitation of the in vitro synthesized al and c2 proteins, and by measuring in vivo al-c2 repressor function. Indirectly, the effect of these mutations was also characterized by determining their ability to trans-inactivate wild-type al-x2 repressor function when combining these mutations with a dominant negative mutant cL2. The mutational effects on al-c42 dimerization were measured both in vivo and in vitro.  53  1.6.2.  Study the DNA-Binding Specificity of cc2 Another purpose of this study was to investigate the influence of MCM1, the co-repressor of a2,  in detennimng the DNA-binding specificity. The ability of cL2 to recognize a variant DNA target site in vitro, with the aid of MCM1, was determined by DNA-binding band-shift assays. To facilitate this, cL2 repressor protein was expressed in E. coli. In order to delineate the flexibility of cL2-binding specificity under MCM1’s influence, in vivo functional assays, involving direct measurement of the regulation on a reporter gene’s transcription, were also used.  54  MATERIALS AND METHODS  2.1.  2.1.1.  REAGENTS  Enzymes Restriction enzymes and DNA modifying enzymes were purchased from New England BioLabs,  Promega, Pharmacia, or Bethesda Research Laboratories. RQ1 DNase, SP6 RNA polymerase, T7 RNA polymerase, and RNasin® were supplied by Promega; pancreatic RNase was from Sigma; Zymolase 60,000 was from Seikagaku Kogyo Co Ltd. and Glusulase was from Dupont. 2.1.2.  Nucleotides Deoxyribonucleotide triphosphates, dideoxynbonucleotide triphosphates, polydl.dC, m GpppG 7  and ribonucleotide triphosphates were purchased from Pharmacia; 35 S-cjxrP and 32 P-labelled nucleotides were from DuPont. 2.1.3.  Oligonucleotides Oligonucleotides were either purchased from Pharmacia or synthesised by T. Atkinson or  Oligonucleotide Synthesis Unit of the Biotechnology Laboratory at the University of British Columbia on an Applied Biosystems 380A DNA Synthesizer.  Table I lists the sequence of each oligonucleotide.  Oligonucleotides synthesized on the DNA synthesizer were purified by electrophoresis and C18-Sep-Pak reversed-phrase chromotography as described by Atkinson and Smith (1984), or by butanol extraction as described by Sawadogo and Van Dyke (1991). 2.1.4.  Autogradiography Materials Kodak KRP-1 film was used for all autoradiography and Hyperfilm-ECL (Cat. No.RPN 2104)  from Amersham was used for ECL detection. When needed, a Dupont Cronex, Quanta III or High Speed intensifying screen was used. New England Nuclear (NEN) Enlightening was used for fluorography of 3  proteins.  55 Table L List of Oligonucleotides Name  Application  SN44  5’-ATCCTCCATGTAATTACCGAAAAAGGAAATTTACATGGAGCTGATCC GCCGTCGACTCCTCAATGTAGAAAA GTACATCGAGCAGTAACAGTATAG CAG-3’  Generate a synthetic duplex DNA with the complementary SN45 for DNase I protection assay  SN4S  5’-CTGCTATACTGTTACTGCTCGATGTACTTTTCTACATTGAGGAGTCGA CGGCGGATCAGCTCCATGTAAATTT CCTITITCGGTAATTACATGGAGGA T-3’  Complementary sequence of SN44  SN51  5’-GAAGATAAAGCAGGATCCAGA AAAAATGA-3’  Create BamHI site in the 5’untranslated region of MA TcL2  SN52  5’-GAAAGAATGATCACGAAA-3’  Create BclI site in the 3’untranslated region of MATcL2  OLIGOAa2A  5’-P-ATCACAGATGAGTTTAAATC AAGCTTAACAACAGAAGAAGTT GAATTCAGGGAT-3  Delete cL2A heptad repeats of IvL4 Tcc2  B 2 OLIGO  5’-P-GCCAACAAAAACCGTAAGGA AAACGATAGAAGTAAT-3’  Delete cL2B heptad repeats of MA TcL2  .. 4 OLIGOA 7  5’-P-GATAAGCAGGATCCAGAAAA AATGAATAAAGATCTTTTAAATC CACAAATC-3’  PCR primer to delete N-terminal codons 4-7 ofMA TCL2  OLIGOA4-62  5’-P-AGAAAAAATGAATAAAAAA AACCGTAAGATTAGTGAT-3  Delete N-terminal 4-62 codons of MA To’2  6 OLIGOSP  5’-GGAATTCGATTTAGGTGACACTA TAGAATATAAGCAGGATCCAGAAA AAATGAATAA-3’  Introduce the SP6 promoter sequence to the MA TcL2 gene in PCR reactions  CYH12  5’-ACACAGGAGTATATTATGAATAA AATACCCATTAAAGAC-3’  Delete CCP coding region from the CCP-MATcL2 fusion  CYH17  5’-P-GAAAGAATGATCAAGCTT-3’  Create BclI site inMATx2 on PVT plasmid  a Underlined sequences indicate the newly introduced endonuclease restriction sites. The small case letters separated by “I” indicate the mixed nucleotides at the position.  56 Table I. List of Oligonucleotides (Cont.) Name  Sepuencea  Ayplication  cYrns  5’-GTCTAGTATGCTGGATTrAAAC TC-3’  Extension primer to start extension at 14th codon of MATc2 cDNA  CYH19  5’-CCATTAAAGATCTTTTAAATCC-3’  Create a BgIJJ site and a silent mutation at the 8th codon ofM4 TcL2  CYB2O  5’-ATATGTAGCTTTGGATCCCATGA TTTATCT-3’  Create a BamHI site after the second codon of URA3  CY1126  5’-GTGCATGATATTAAATAGCTTGG CA-3’  Extension primer for CYC]: URA3 fusion genes  CYH27  5’-P-GGTCATATAATGGGATCCGCAA TTAATCAT-3’  Introduce a BamHJ site after the first codon of BAR]  CYH29  5’-P-TGGCATCTAATTACCGTAAAAG GAAATTAGATGGCGA-3’  Mutate the cc2 recognition sequence at the BAR] promoter  L29R  5’-P-AAGCGCTTTTCTATT-3’  Create Arg replacement of Leu29 of M4Tci2  L36R  5-P-TGCTGTAATAGACCTAAGTTA3  Create Arg replacement of Leu36 of MATcw2  C33FIY  5’-P-ATAAAAAGC’ITIITiCTATfl’ tJaCTGTAATTTACCT-3’  Change Cys33 of MATcL2 to Phe or Tyr, and introduce a HindIJJ site  C34F/Y (L291)  5’-P-AATAAAAAGATC’riTi’CTATfl’ GCTt/aTAATflACCTAAG-3’  Change Cys33 of MA TcL2 to Phe or Tyr, and introduce a BglJJ site  167A  5’-P-AACCGTAAGGCTAGTGA TGAG-3’  Ala replacement of 11e67 inM4TcL2  L74A  5’-P-GAGAAGAAGGCTflGCA AACAAC-3’  Leu74  LS1A  5’-P-ATCTCAAGCCACTACT-3’  Leu8l  a Underlined sequences indicate the newly introduced endonuclease restriction sites. The small case letters separated by “I” indicate the mixed nucleotides at the position.  57 Table L List of Oligonucicotides (Cont.) Name  1 Seciuence’  Application  L81A/PstI  5’-P-AAGAAGTTGCTGCAGACAACAT CTCAAGCCACTACTACC-3’  Ala  LS8A  5’-P-ACCATrACrrAGCACT CAAAGAA-3’  Leu88  LS8A(T84N)  5’-P-AACATrACTGTAGCACTCAAAG AA-3  LeuS8  195A  5’-P-ATGCGCAGCGCAGAAAAC-3’  1eu95  L8IR  5’-P-ATCTCAACGCACTACT-3’  Arg  167T  5’-P-CCGTAAGACTAGTGATG-3’  Thr replacement of 11e67 inM4TcL2  TS3S  5’-P-CATCTCAACTCACTAGTACCA3’  Ser replacement of Thr83 in MATc2  5’-CACAGATATATACATATTT GTTTTTCGTGATCAT-3’  PCR primer to generate full length ]vf4Tc2  MATh2-3’ (CYH53)  primer  replacement M4 TcL2  replacement !vL4Tx2  of Leu8l  of Leu8l  a Underlined sequences indicate the newly introduced endonuclease restriction sites. The small case letters separated by “I” indicate the mixed nucleotides at the position.  in  in  58  2.1.5.  Purified a.2 Protein from E. coli  This was purified and provided by Dr. Peter L. Davies according to the procedure described previously (Sauer et a!., 1988).  The final protein concentration was approximately 2.1 mg/mi and was  dissolved in S buffer (50 mM Tris-HC1 at pH 8, 1 mM EDTA, 0.28 mM 3-mercaptoethanol) plus 500 mM NaC1. 2.1.6.  Antibodies Alkaline-phosphatase and horseradish-peroxidase-coupled goat anti-rabbit IgG antisera were from  Bio-Rad. The rabbit anti-ct.2 serum was raised by Dr. Peter L. Davies using the KLH-coupled synthetic polypeptide from the 181st to 210th amino acid of the M4 Tci.2 predicted sequence as antigen. 2.1.7.  In vitro Translation Materials Wheat germ extract, an amino acid mixture without methionine, and a potassium acetate solution  were purchased as a kit from Promega.  S-methionine, at 0.41 MBg/ and 44.0 TBq/mmol specific 35  activity, was obtained from DuPont. 2.1.8.  Media Components Amino acids, vitamins and ampicillin were supplied by Sigma.  The remainder of the media  components was supplied by Difco.  2.2.  BACTERIAL AND YEAST STRAINS E. coli Strain RR1 (F, hsdS2O (rB, mB), ara-14, proA2, lacY], galK2, rpsL2O (Smr), xyl-5, mtl-],  supE44, i)(Bolivar et al., 1977) was used for retrieval of yeast plasmid DNA. RE404 (darn-3, dcm-6, metE], galK2, galT22, lacY], thi-], tonA3], tsx-78, mtl-], supE44) was used to generate umnethylated plasmid DNA for methylation sensitive restriction enyzmes such as BclJ (Brent and Ptashne, 1980).  59  JMIO1 (zl(lac pro), thi, supE, F’traD36 proAB, lacJqZ MiS) was the host strain used for propagating plasmids and for isolating single-stranded DNA for single-stranded DNA sequencing (Messing, 1983). RZ1032 (HfrKL16 P0/45 [lysA(61-52K)], duti, ungi, thil, relAl, Zbd::TnlO, supE44), (Kunkel, 1985) was used for the production of single-stranded uracil-containing DNA. DH5x  ( F-, 080 lacZiiJvf15 zI(1acZYA-argF) U169, deoR, recAl, endL4I, hsdR]7 (rK,mKj  SupE442- thi-1 gyrA96 reM]) was used for plasmid construction and for isolation of double-stranded DNA for d.s. DNA sequencing (Hanahan, 1983). S. cerevisiae The genotypes and sources for all strains used in this study are listed in Table II. Strain S67x66 was constructed as follows: strain S67 was transformed with YEPJ3 (Broach, et at., 1979), a LEU2 containing marker, and crossed with YHE64-transformed S66 (for the description of YHE64, see Section The resulting diploids were selected on SC-Leu-Trp medium. Then the Leu+ Trp+ diploids were streaked on a YPD plate and the Leu colonies identified by replicate plating on SC-Leu or SC+Leu medium. A single Leu colony was picked and propagated in YPD nonselective rich medium. This culture was diluted and spread on a SC+5FOA plate and Ura isolates were obtained as 5FOA resistant colonies.  The final Ura Leu S67x66 diploid was used for the study of the dominant  negative matcs2 mutant.  2.3.  MEDIA AND GROWTH CONDITIONS E. coli Bacterial cells were grown at 37°C in 2xYT (1.6% Bacto-tryptone, 1% Bacto-yeast extract, 0.5%  NaC1) or in M9 (50mM , 4 H 2 Na PO 25 mM 4 PO 8.5 mM NaCI, 1mM MgSO 2 KH , , 0.1 mM CaCl 4 , 10 mM 2 glucose, 0.00 1% thiamine).  100 of ampicillin was added as a supplement when required. Plates  contained 1.5% agar and soft top agar contained 0.7% agar. For strain DH5cç TB broth (1.2% Bacto  60  Table U. S. cerevisiae Strains  Strain  Genotype  Source  23.75 (SiS)  matciX75 a Irpi leu2 ura3 his4  K. Tatchell  1402 (S66)  a ho:LacZ ura3 leu2 trp]  R. Jensen  1420 (S67)  isogenic to 1402, except a  R. Jensen  1402x1420  a/cL ho:LacZ/ ho:LacZ ura3/ura3 leu2/leu2 trp]/trpl  this study  SP2  iimat::CANb ste6:LacZ ura3 leu2 trp] his3 his4  S. Porter  k1107(S85)  a HMRa hmla ho.LacZ ura3-52 ade2 can] met] his3 leu2  K. Nasmyth  (S67x66)  trpl 1160 (S64)  a ste6:lacZ C ura3 leu2 trpl his3 his4  K. Wilson  D311-3A (S17)  a lys2-] his] trp2  J. Boss  1(204 (Sil)  mata2-1/IvL4Ta sirl-1/SIR1 ade6/ADE6 meti/MET] his6/HIS6  K. Nasmyth  rme]/RZv[E1 frp]/frp] ho/ho a This is a transpiacement of mutant matcX75 (Tatchell et al., 1981) into the MAT locus. b  The Amat::CAN allele was created by the gene replacement of the M4 T locus with CAN.  C  f3-galatosidase activity from the ste6:LacZ fusion was less than 10% of that from the plasmid borne CYCI (a]-cL2).LacZ reporter, thus negligible in the present study.  61  tryptone, 2.4% Bacto-yeast extract, 0.4% glycerol, 17mM 4 PO 72mM 4 2 KH , P0 supplemented with 200 2 K ) jag/mi of ampicillin was used. S. cerivisiae The following media were used to maintain the Saccharomyces cerevisiae strains: YPD  1% yeast extract, 2% peptone, 2% glucose  SC  0.67% yeast nitrogen base, 2% glucose, supplemented with the following: 20 mg/i each of adenine, uracil, tryptophan, histidine, arginine, methionine; 30 mg/i each of tyrosine, leucine, isoleucine, lysine; 50 mg/l of phenylalanine; 100 mgIl each of glutamic acid and aspartic acid; 150 mg/i of threonine; 375 mg/l of serine. One or more of the above were excluded where indicated. 0.2% of 5-fluoroorotic acid was added when required.  Spo  1% potassium acetate, 0.1% yeast extract, 0.05% glucose  R-agar  SC with twice the concentration of the amino acid supplement in 1M sorbitoi, 2% YPD and 3% agar.  For plates, 2% agar was added to the above media. Yeast cultures were grown at 3 0°C.  2.4.  2.4.1.  TRANSFORMATION OF E. COLI  Calcium Chloride Method Competent cells were prepared as described by Dagert and Ehrlich (1979). A 25 ml solution of  2xYT was inoculated with 0.25 ml of an overnight culture of E. coli and incubated at 37°C to exponential phase (A 600 of 0.2 to 0,3). The culture was chilled on ice for 5 mm and centrifuged at 3000 x g for 5 mm. The cells were resuspended in 25 ml of ice-cold 50 mM CaC1 2 and incubated on ice for 30 mm.  The  harvested cells were resuspended in 1 ml ice-cold 50 mM CaCl . Plasmid DNA (1 to 100 ng) was added to 2 0.1 ml of competent cells, these were incubated on ice for 30 to 40 mm, and then transferred to a 42°C water bath for 2 mm. After heat shock, the cells were incubated with 1 ml of 2xYT in a 37°C shaker for 30 to 40 mm. Aliquots of 0.1 to 0.2 ml were plated on selective medium and incubated at 37°C overnight.  62  2.4.2.  Electroporation Method Electroporation-competent cells were prepared as follows. 1 litre of 2xYT was inoculated with 10  ml of a fresh overnight culture. Cells were grown at 37°C with vigorous shaking to an 0D 600 of 0.5 to 0.8. The culture was chilled on ice for 15 to 30 minutes, and centrifuged at 4000 x g at 4°C for 15 minutes. The harvested cell pellet was washed once with 1 litre of ice cold water then again with 0.5 litre, and then with 20 ml of ice cold 10% glycerol. The cells were then resuspended in 3 ml of ice cold 10% glycerol. The cell suspension was frozen in aliquots on dry ice, and stored at -70°C. Frozen competent cells were thawed at room temperature and immediately placed on ice. 40 of the cell suspension was mixed with 1 to 2 l (‘..O. 1 .tg) DNA solution, and electro-pulsed in a 0.2 cm Gene Pulser cuvette (Bio-Rad) on a Gene Pulser Transfection Apparatus (Bio-Rad) at 25 .tF, 200 2 and 2.50 Ky. i ml of SOC (2% Bacto-tryptone, 0.5% Bacto yeast extract, 10mM NaCl, 2.5 mM KC1, 10 mM MgC1 , 10 mM MgSO 2 , 20 mM glucose) medium 4 was added to the cuvette.  Then the cell suspension was transferred to a 1.5 ml polypropylene  microcentrifuge tube and incubated at 37°C for 1 hour, 200 d of aliquots were spread on selective medium and incubated at 37°C overnight.  2.5.  TRANSFORMATION OF S. CEREVISL4E Yeast cells were grown in YPD to mid-log phase, centrifuged, and washed in 1/10 volume 1 M  sorbitol. Cells were resuspended in 1/10 volume 1 M sorbitol containing 20 mM DTT and 1% glusulase, and were gently agitated at 30°C for 45-60 minutes. A drop of the cell suspension was mixed with an equal volume of 1% SDS and observed under phase contrast to ensure spheroplast formation.  The cells were  washed once in 1/10 volume 1 M sorbitol and once in 1/10 volume STC (1 M sorbitol, 10 mM Tris-HC1 at pH 8, 10 mM CaCl ) and resuspended in 1/100 volume STC. Approximately 1 ig of DNA was incubated 2 with 50 of cells at room temperature for 10-15 minutes. 40 of 20% PEG 3350, 10 mM Tris-HC1 at pH 8, 10 mM CaC1 2 was then added and the suspension was further incubated for 15 minutes. These cells were then added to approximately 15 ml of R-agar at 55°C and plated on SC selective media. colonies were visible after 2-3 days at 30°C.  Transformed  63  2.6.  2.6.1.  TECHNIQUES FOR DNA MANIPULATION AND ANALYSIS  Restriction Digests and DNA Modification Restriction digestion reactions were usually carried out in 20 tl volumes at 37°C for 1-2 hours.  DNA concentrations were approximately 100-300 ng/.d in buffer as described in Maniatis et at., 1982. 2.6.2.  Gel Electrophoresis For analysis of DNA by polyacrylamide gel electrophoresis, the gel buffer and the running buffer  was 50 mM Tris-borate at pH 8.3, 1 mM EDTA. The ratio of acrylamide to bis-acrylamide in the gel was 29:1. Agarose gels were made and run in the same buffer. 2.6.3.  Isolation of DNA Fragments Fragments used for sequencing by the Maxam and Gilbert method (1977), DNase I protection  footprinting or DNA-binding band-shift assay were isolated by polyacrylamide gel electrophoresis. DNA samples were recovered from the excised polyacrylamide gel by incubation overnight in DNA elution buffer (0.5M ammonium acetate and 2 mM EDTA) at 52 °C. Fragments used for cloning or other purposes were generally isolated from agarose gels using the Gene Clean DNA Purification Kit (BiolOl) or the Sephaglas DNA Purification System (Pharmacia) as described by the manufacturers. 2.6.4.  Ligations Ligations were carried out in One-Phor-All-Plus buffer (Pharmacia) supplemented with 1 mM  ATP or ligation buffer (50 mM Tris-HC1 at pH 7.5, 10 mM Mg Cl , 10 mM DTT, 1 mM spermidine, 1 mM 2 ATP, 0.5 mglml BSA) with approximately 200 ng of linearized vector, and three- to five-fold molar excess of the insert and 4 units of T4 DNA ligase. The reaction was incubated either at room temperature for 4 hours (for blunt-end ligations) or 16°C for 12 to 18 hours (for sticky-end ligations).  When the vector  64  contained compatible ends, the linearized vector was dephosphorylated with Calf Intestinal Phosphatase as described by the manufacturer. 2.6.5.  DNA Sequence Detennination DNA sequence analysis was performed either by the method of base-specific chemical cleavage  (Maxam and Gilbert, 1977; Maniatis et a!., 1982) or by the dideoxyribonucleotide chain-terminator method (Sanger et a!., 1977) which was modified by Ner et a!. (1988). For double-stranded DNA sequencing, the Sequenase kit (US Biochemical Corp.) was used. 2.6.6.  Isolation of Plasmids and Bacterial Phage DNA  Isolation of Plasinid DNA from E.eoli  The isolation of plasmid DNA from small cultures (1 to 1.5 ml) by the alkaline lysis procedure was carried out as described by Maniatis et a!. (1982). For double-stranded DNA sequence analysis, DNA purification was done using the Magic DNA Minipreparation Purification System (Promega).  Isolation of Single-Stranded DNA  Single-stranded DNA of pEMBL plasmids or otherfl origin containing plasmids was isolated as described by Dente et al. (1983). Cells from a single colony of freshly transformed E. coil JM1O1 were used to inoculate 1.5 ml of 2xYT with ampicillin.  The inoculum was incubated with shaking at 37°C for  approximately 1 hour to 1-5x10 8 cells/mi. The cells were then superinfected with R408 helper phage at a multiplicity of infection (m.o.i) of 20:1. The cultures were incubated at 37°C for 6 hours. Phage particles were precipitated from the supernatant with 0.25 volume of 20% PEG-6000, 2.5 M NaCI at room temperature for 15 minutes. After centrifugation, the phage pellets were resuspended in 200 i1 TE and extracted with 200 .tl phenol/chloroform (1:1). The phage DNA was ethanol precipitated and redissolved in 35 l TE.  65  2.6.7.  In vitro Oligodeoxyribonucleotide-Directed Site-Specific Mutagenesis Single-stranded, uracil-contaimng templates for site-directed mutagenesis were prepared, and the  in vitro site-directed mutagenesis carried out as described previously (Ner et at., 1988). Except for the generation of the combined mutation L29R+L36R at the c42A region as well as the mutiple alanine replacement at the x2B region, two or more mutagemc oligonucleotides were hybridized to the uracil containing template (Figure 10, panel B). After annealing, a 30-minute ligation at 16°C allowed the headto-tail oligonucleotides to form a longer mutagenic oligonucleotide, and then extension and ligation was carried out as described by Lorimer et at. (1992). 2.6.8.  Polymerase Chain Reaction Polymerase chain reactions (PCR) were performed in 100 volumes. 50 pmole each of the PCR  primers were mixed with 0.1 .tg of the DNA template in a 100 il volume PCR reaction which contained 50 mM KC1, 10 mM Tris-HC1 at pH 8.3, 200 i.iM of dNTP, 3 mM MgCI , and 2.5 units of Taq DNA 2 polymerase. 80 .tI of mineral oil was added and the PCR reactions performed in a Perkin-Elmer Cetus DNA Thermal Cycler (N801-0150). After a 2-minute initial incubation at 95°C, and 25 cycles of ‘95°C-is sec, 50°C-30 sec, and 72°C-i mm”, the PCR products were purified with the Magic PCR Preps DNA Purification System (Promega). Purified PCR products were dissolved in 50 l water. In order to introduce the SP6 promoter sequence to different MA TcL2 mutants by PCR amplification, the OLIGO 6 primer and MA TcL2 -3’ (CYH53) primer (Table I) were used. p 5 For generation of ,natcc2A4-7 deletions, 25 pmole of each of the deletion primer (CYH48) and the 3’-primer CYH35 (Table I) were mixed with 1 unit of Vent DNA polymerase (New England Biolabs) in a 50 il reaction which contained 10 mM KC1, 10 mM , S0 20 mM Tris-HC1 at pH 8.8, 2 mM MgSO4, 2 ) 4 (NH 0.1% Triton X-I00, 1 mM MgC1 2 100 BSA. The PCR reactions and the product purification were as described above.  66 Figure 10. In vitro Oligonucleotide-Directed Site-Specific Mutagenesis. The mutagenesis procedures for multiple mutations are shown: (A) The introduction of BamHI and BclI recognition sites into the MA Tci.2 gene are shown in this panel. The two mutagenic primers, SN5 I and SN52 (described in Table I), were annealed to the single-stranded pEMBL-MATc DNA template (Porter and Smith,. 1986). The template DNA contained uracil. The annealed DNA mixture was extended (with dNTP) and ligated to give a covalently closed and doublestranded DNA product. The newly synthesized DNA strand containing the BamHI and BclJ sequences was free of uracil. The DNA mixture was then used to transform E. coil JM1O1. This process selected against the uracil-containing DNA strand from the original template. Double-stranded DNA was isolated from individual E. coil transformants. The presence of the mutations was verified by restriction endonuclease mapping and DNA sequencing. (B) This panel shows a different mutagenic scheme in which a set of “head-to-tail’ mutagenic primers were used. After the annealing procedure, ligation of the adjacent primers (p-Mi, p-M2, p-M3) was carried out to generate a longer mutagenic oligonucleotide. Extension, ligation, and transformation of E. coil JM1O 1 or DH5c were performed as before. Desired mutant DNA plasmids were verified by restriction endonuclease mapping and DNA sequencing.  67  A  : Bell  MATh2  +  anneal  BclI Bill  extend, ligate  Bell  + U-pEMBL8  extend, ligate  Bell  veri1j mutants by restriction enzyme mapping and DNA sequencing  isolate d.s. DNA 4  transform JM1O1 4  B Ml  MATcL2  Ml  ligation  anneal P M3  + U-pEMBL8  extend, ligate  M2 Ml  veri1,’ mutants by restriction enzyme mapping or/and DNA sequencing  isolate d.s. DNA  transform E. colt  -.  Figure 10. In vitro Oligonucleotide-Directed Site-Specific Mutagenesis.  M2  M3  68  2.7.  2.7.1.  IN VITRO TRANSCRIPTION AND TRANSLATION  In vitro Transcription 5’ capped transcripts were synthesized in a 50 tl reaction containing 40 mM Tris-HC1 at pH 7.5,  6 mM MgC1 , 2 mM Spermidine, 10 mM NaC1, 10 mM DTT, 100 BSA, 80 units of RNasin® 2 G(5’)ppp(5’)G (Pharmacia), 7 (Promega), 500 j.tM ATP, 500 p.M CTP, 500 p.M UTP, 50 p.M GTP, 500 p.M m 10 p.1 of DNA template and 40 units of T7 or SP6 RNA polymerase (Promega). For IvL4 Ta] transcription, 5 p.g of HinclI linearized E162 plasmid DNA (S.S. Ner and M. Smith, unpubl.) and SP6 DNA polymerase were used. For all the pGEM3Zf(+)-MATcL2 constructs, 5 p.g of the HindilI linearized plasmid DNA and T7 DNA polymerase were used. For pEMBL-MATc constructs, 10 p.1 aliquots of a 50 p.1 solution which contained -5 p.g of SP6-promoter-introduced PCR fragments (see Polymerase Chain Reaction, in this section) were used. After 1 hour at 3 7°C, 10 units of RQ1 RNase-free DNase I (Promega) was added and incubation was continued for another 15 minutes at 3 7°C. The reactions were extracted twice with one volume of chloroform: isoamyl alcohol: phenol (24: 1: 25). The synthetic mRNAs were precipitated with ethanol after adding an equal volume of 5 M amonium acetate at pH 7.8, collected by centrifugation, and dissolved in 20 p.1 of water treated with diethylpyrocarbonate. The reactions typically yielded 1.0 to 5.0 p.g of RNA. 2.7.2.  In vitro Translation In vifro translation reactions were performed as described (Michael et al., 1989).  50 p.1  translation mixtures contained 25 p.1 of wheat germ extract (Promega), 4 p.1 of an amino acid mix (supplied S-MET (1 by the manufacturer), 4 p.1 of 1 M potassium acetate, 40 units of RNasin® (Promega), 2.5 p.1 of 35 mCiI9l p.1, 1190 Cilmmol, Dupont) and 2.5 p.1 of MA Ta] in vitro transcript or/and 8.0 p.1 of MATcc2 transcripts. After 1 hour at 25°C, the mixtures were immediately used for DNA-binding electrophoresis mobility-shift assay or co-immunoprecipitation. 0.5 p.1 of the translation products were subjected to 17.5% SDS-PAGE and fluorography.  69  2.8.  2.8.1.  ISOLATION AND ANALYSIS OF YEAST RNA  Isolation of Yeast RNA The rapid RNA isolation procedure is described as follows: 20 ml of yeast culture (grown in SC  selective medium) with an 0D 600 of 0.5-1.0 was filtered through a nitrocellulose filter. The harvested cells were washed in 0.5 ml of breakage buffer (0.5 M NaCl, 0.2 M Tris-HC1 at pH 7.5, 10 mM EDTA), and then resuspended in 200 ).tl of breakage buffer. The cell suspension was mixed with 200 111 of glass beads (HCi washed) and 200 1 .d of 4:4:1:1 phenollchloroformllo% SDS/3 M sodium acetate at pH 5.2 and vortexed for 5 minutes. This extraction was repeated twice, and then the mixture was extracted once with chloroform. The aqueous phase was transferred and added to 0.4 volume of 5 M ammonium acetate at pH 7.4. The RNA was precipitated by the addition of 2 volumes of ethanol, washed with 70% ethanol, dried, and dissolved in 20 iii water (incubation at 80-90°C for 2 minutes to aid dissolving). The amount of RNA was estimated by 260 and the final concentration was adjusted to 10 g/.tl. A ,  5 of the RNA preparation was used for  primer extension analysis. 2.8.2.  Primer Extension Analysis of Yeast RNA The oligonucleotide primer was radioactively labelled as follows.  1 pmole of oligonucleotide  , 5 mM DTT, 4.5 units of T4 2 primer was incubated with 100 mM Tris-HC1 at pH 8, 10 mM MgC1 polynucleotide kinase and 3 pmole of yP ATP (10 j.tCi/ in a 20 jil reaction, at 37°C for 30 minutes. 32 The labelling reaction was stopped by incubation at 65°C for 5 minutes. The annealing proceeded with the addition of 1 .il of kinase-labelled primer, 5 111 of yeast RNA (10 Ig/Il), and 1 .tl of buffer A (500 mM Tris HC1 at pH 8.5, 1 M KC1, 0.1 M MgC1 ), and the incubation of the mixture at 85°C for 2 minutes before 2 placing on ice for 10 minutes. The extension reaction was then carried out at 42°C for 1 hour after adding 3 tl of 500 .tM dNTP and 10 mM DTT. 10 J.Ll of formamide dye mix (80% formamide, 10 mM EDTA at pH 8.0, 1 mg/mi xylene cyanol FF, 1 mg/ml bromophenol blue) and 1 j.d of 0.2 M NaOH were added to stop the reaction. The final mixture was subjected to denaturing polyacrylamide gel electrophoresis.  70  2.9.  PREPARATION OF CRUDE YEAST EXTRACTS Crude yeast extract was prepared using essentially the method described by Hayes et at. (1988).  1.0 ml of a fresh overnight culture of SP2 or S85 (see Table II for description of genotype) was used to inoculate 1.5 1 of YPD. The inoculum was incubated at 30°C with shaking at 250 rpm to mid-log phase. All steps were carried out at 4°C. Cells were harvested and resuspended in 1 volume of extraction buffer , 1 mM EDTA, 10% glycerol (vollvol), 1 2 (200 mM Tris-HC1 at pH 8.0, 400 mM 4 S0 10 mM MgC1 2 ) (NH , mM DTT, and 0.5 mM PMSF, 0.2 ig/ml chymostatin, 0.2 Ig/m1 aprotinin, 2 ig/ml pepstatin A, 2 .tg/ml leupeptin, and 0.5 mM benzamidine). 1.5 volume of glass beads was added to the suspension. The mixture was vortexed for 1 minute, then cooled on ice for 3 minutes. This procedure was repeated again nine times before 2 volumes of extract buffer were added. The mixture was then incubated on ice for 30 minutes. The supernatant was separated from the cell pellet after centrifugation at 10,000 x g for 10 minutes in a SS-34 rotor. 0.57 volume of ammonium-sulfate-saturated extract buffer was added slowly with stirring over a 30 minute period. The 10-40% ammomum sulfate cut protein pellet was harvested by centrifugation at 20,000 x g for 30 minutes. The pellet was then dissolved in 1/8 volume of dialysis buffer (20 mM HEPES at pH 8.0, 5 mM EDTA, 20% glycerol, 1 mM DTT), and dialyzed against 200 volumes of dialysis buffer twice in 16 hours. The dialysate was stored at -70°C.  2.10. SDS POLYACRYLAMIDE GEL ELECTROPHORESIS (SDS-PAGE) OF PROTEINS SDS polyacrylamide gel electrophoresis (PAGE) of proteins was carried out as described by Maniatis et at., 1982. For 8.3 x 10.2 cm protein “mini-gels”, constant voltage at 100-150 V was applied. The sample buffer for SDS-PAGE contained 50mM Tris-HC1 at pH 6.8, 2% SDS, 0.1% bromophenol blue and 10% (V/V) J3-mercaptoethanol. For detection of disulfide-linked c42 dimers, 3-mercaptoethanol was omitted in the sample buffer.  71  2.10.1. Fluorography The gels were fixed for 30 minutes in 40% methanol! 10% acetic acid, and then soaked in autoradiography enhancer (Enlightening, Du Pont) for 15 minutes. The gel was then dried and exposed to X-ray film.  2.11. YEAST FUNCTIONAL ASSAYS  2.11.1. Mating Assay Quantitative mating tests were performed as follows: Cells were grown in a small culture (5 ml of selective medium) to mid-log phase, and 0.5 ml of sterile water was added to 2x10 6 cells of each the strains to be tested as well as to the tester strain. The mixture was collected on a nitrocellulose filter and placed on a YPD plate for 6 hours at 3 0°C. The cells were then resuspended in 10 ml sterile water and 0.5 ml of the cell suspension was removed, vortexed 1 minute, appropriately diluted (generally 10), and plated in duplicate on plates selecting separately for the two haploid strains (and the diploid) or for the diploid alone. The Efficiency of Mating refers to the number of diploids with respect to that of the haploid of the least concentration. 2.11.2. Sporulation 5 ml of a SC selective medium was inoculated with 0.1 ml of a freshly saturated yeast culture. 600 of 2.5-3.0) was harvested After shaking at 30°C for approximately 7 hours, 0.5 ml of the culture (0D and washed twice with 1.5 ml of sterile water, resuspended in 1 ml of sporulation medium (1% potassium acetate, 0.1% yeast extract, 0.05% dextrose) and incubated at 30°C with shaking at 250 rpm for 2-4 days. A suspension of cells was examined under the microscope, and the percentage of asci with respect to the total number of cells determined.  72  2.11.3. f3-Galactosidase Assays Cell were grown to early-mid log phase and the 3-galactosidase activity was measured after treatment with SDS and chloroform essentially as previously described (Ruby et a!., 1983). The 0D 600 of the cell culture was recorded, 500 iii of cell culture was harvested by centrifugation, and the cell pellet was quick-frozen and stored at -70°C. Cells were resuspended in 250 jil Z Buffer [60 mM 4 PO 40 mM 2 Na , PO 10 mM KCL, 1 mM MgSO 2 NaH , 4 , 50 mM 3-mercaptoethanol, pH 7.0 (Ruby et at., 1983)], then 4 permeabilized with 50 t1 chloroform and 25 l of 0.1% SDS.  After vortexing for 3 seconds, and pre  incubating at 28°C for 5 minutes, 200 p.1 of 4 mg/mi ONPG in Z Buffer was added to the cell suspension. The mixture was then vortexed, incubated at 28°C until a faint yellow color appeared and then the reaction CO The reaction mixture was then centrifuged and 0D 2 Na . 420 of was stopped by adding 500 p.1 of 1 M 3 the clear supernatant was measured after being separated from the cell debris. The unit activity was defined / [reaction time (mm) x volume of cell culture (ml) x 0D 420 as 1000 x 0D ]. 600 2.11.4. Growth Rate Determination Growth conditions were defined as described (this section). In a 250 ml Erlenmeyer flask, 100 ml of liquid medium was inoculated with a fresh late-log phase culture. The growth rate of the yeast was determined by taking 1 ml samples at different time intervals. The samples were mixed with 1 ml of 0.3% formaldehyde to immediately stop growth and the 0D 600 of each mixture was determined.  2.12. IN VITRO DNA-BINDING ASSAYS  2.12.1. DNA-Binding Gel Electrophoresis Mobility-Shift Assay DNA-binding gel electrophoresis mobility-shift assay was carried out as described (Dranginis, 1990).  4.5 p.1 of the in vitro translation product (see description in this section) was mixed with 2 p.g  polydl.dC in 15 p.1 binding buffer (50 mM NaC1, 10 mM Tris buffer at pH 7,4, 1 mM MgC1 , 1 mM EDTA, 2 1 mM dithiothreitol, 5% glycerol). The Dranginis procedure was modified by the addition of 2 p.g of a DNA probe (a 29 base pairs synthetic duplex, 5’-GCTTCCCAATGTAGAAAAGTACATCATAG-3’, 5’-CTATG-  73  ATGTACITIICTACATrGGGAC-3’) instead of 2 ng, and then the binding reaction was incubated for 20 minutes at room temperature.  5 il of gel loading buffer containing 1. 5X binding buffer and 1 mg/mi  bromophenol blue was added and the mixture was loaded onto a 5% nondenaturing polyacrylamide gel (Maniatis, 1982) in lx TBE (45 mM Tris-borate, 1 mM EDTA, pH 8.2). The gel was run at 200 volts for 1 hour and then subjected to fluorography. 2.12.2. DNA-Binding Band-Shift Assay All DNA-binding band-shift assays were performed (mostly as described by Keleher et a!., 1988, Hayes et al., 1988) at room temperature, run using 1X TBE buffer (Maniatis et al., 1982) on a native 4% polyacrylamide gel at 200 volts for 2 hours, dried and autoradiographed.  Each of the final reactions  contained 20 j.ii of 20 mM HEPES at pH 8.0, 5 mM EDTA, 20 % glycerol, 1 mM DTT plus 100 mM NaC1, 100 .tg/m1 of poiy dI-dC, —20 pmole/ml 32 P-labelled DNA probe, 0-2 il of purified c2 from E. coil, and 010 of yeast crude extract.  The c42 protein, the yeast protein extract, and poiy dI-dC were incubated  together for 10 minutes, and the labeled DNA-binding probe was added to the mixture and incubated for an additional 10 minutes before loading the reactions onto the gel. For the “competition study” experiments, approximately 100-fold of unlabelled DNA probe or competitor was added and incubated for 5 minutes before the labelled DNA probe was added. The procedure for labelling DNA probes for binding assays involved using the large fragment of DNA polymerase (Klenow) to “fill-in” the 5’-overhang of restriction fragments with 32 P-dNTP (described by Maniatis et al., 1982). 2.12.3. DNase I-protection Footprinting Assay The method for the DNase I-protection footprinting assay was as described previously (Keleher et a!., 1988). Pancreatic DNase I protections were performed at room temperature in 50 l reactions. The protein, nonspecific DNA, and the labeled DNA-operator fragment were mixed (conditions were identical to 2 were added to a final concentration of 10 mM 2 and MgCl the DNA band-shift assay above, except CaCl and 6 ruM, respectively) and incubated for 10 minutes. Then 0.1 unit of DNase I was added, and the mixture was incubated for another 60 seconds. degradation.  1 i.l of 0.5 M EDTA was added to each reaction to stop  The reactions were extracted once with phenol:cliloroform and then ethanol precipitated.  74  Finally, the nicked DNA fragments were dissolved in 10 il of a sequencing-gel dye mix, and loaded onto a denaturing 8% polyacrylamide electrophoresis gel in 1X TBE (Maniatis et a!., 1982). The gel was run, dried, and autoradiographed.  2.13. IN VITRO IMMUNOLOGICAL ASSAYS  2.13.1. Affinity Purification of Rabbit Anti-MATa2 Serum -Cys-Gly-Gly at the 2 The synthetic peptide, MATa2 181-C containing an additional NH N-terminal of the amino acid sequence from 181 to 210 of MATx2, was coupled to an activated thiol sepharose 4B column.  The rabbit anti-(x2 serum was affinity-purified as described (Harlow and Lane,  1988). The affinity-purified antiserum contained about 1 mg/ml of IgG in PBS (phosphate-buffered saline) with 0.02% sodium azide and 2% BSA. 2.13.2. Co-Immunoprecipitation of in vitro Translated Proteins The MA Tc’2 gene with different mutations were first cloned into pGEM3Zf(+) (Promega). The resulting pGEM3Zf(+)-MAT(x2 plasmid DNA was linearized with HindIJJ and transcribed with T7 RNA polymerase. For a 20 .tl translation reaction, 1.9 p.g of MA Tcw2 RNA and 0.6 ig of IvL4 Ta] RNA was used. 1 p1 of the translation product was mixed with 20 il of SDS-PAGE sample buffer (described above), and 5 .d of this mixture was subjected to SDS-PAGE. 15 p1 of the translation product was mixed with 600 p1 of RIPA-C (10 mM Tris-HCI at pH 8.0, 100 mM NaCl, 1 mM EDTA, 1% NP4O, 20 jig/mI PMSF, 0.5 jig/nil of leupeptin, 15 jig/mI of Aprotimn) plus 2% BSA. 10 p1 of 25X-diluted pre-immune serum was added and the mixture was incubated on ice for 15 minutes. 40 p1 of a 10% Staphyloccus aureus cell suspension in RIPA-C plus 2% BSA was added and mixed end over end on a roller wheel at 4°C for 30 minutes. After microcentrifugation, the pre-adsorbed supernatant was mixed with 7.5 jil of the purified anti-x2 181-C serum (7.5 jig/jil) and put on ice for 20 minutes.  Another 40 jil of the 10% Staphyloccus aureus cell  suspension was added and mixed at 4°C for 60 minutes. The immuno-complex was pelleted and washed  75  with RIPA-C buffer twice. The immunoprecipitate was eluted with 50 .tl SDS-PAGE sample buffer, and then incubated for 15 minutes at 37°C. 5 to 10 of the eluate was examined by SDS-PAGE. 2.13.3. Electrophoresis and Western Blot Analysis A 1.5 ml culture of E. coil transformants expressing the cL2 protein was harvested and cells were resuspended in 200 il of TE (10 mM Tris-HC1, 1 mM EDTA, pH 8.0). 1-10 .d of the cell suspension was mixed with 2-times concentrated SDS-PAGE sample buffer (described above), boiled for 5 minutes, and then immediately loaded onto a 12.5% SDS-polyacrylamide gel for electrophoretic separation. The proteins were electroblotted onto a nitrocellulose filter in 20 mM Tris base, 150 mM glycine, pH 8.3, 20% methanol at 4°C for 12-16 hours at 150 mA or for 2-4 hours at 0,5 A.  The filter was then incubated at room  temperature with Tris-saline (10 mM Tris-HC1 at pH 7.4, 150 mM NaC1) with 5% skim milk powder , 3 (Carnation brand) for 0.5-2 hours. Rabbit antiserum at 1:50 to 1:1000 dilution in Tris-saline, 10 mM NaN 2.5% skim milk was added and incubated at room temperature for 60 minutes. The filter was washed with three changes of Tris-saline, 0.25% n-lauryl sarcosine (or 0.2% Tween 20, if ECL was to be used later), at 3  minutes per wash.  The bound antibodies were detected by either horseradish-peroxidase or alkaline  phosphatase-coupled anti-rabbit IgG antisera (Bio-Rad).  For detection of alkaline-phosphatase-linked  2 (pH 9.8) was added to the filter, then 180 of 50 , 1 mM MgCI 3 antibodies, 30 ml of 0.1 M NaHCO mg/mi NET in 70% DMF and 180 ILl of 25 mg/mi BCIP in DMF was also added. A red colour was usually detected after 5-15 minutes at room temperature. For detection of horseradish-peroxidase-tagged antibodies, the ECL detection system (Amersham) was used.  2.14. PLASMIDS  2.14.1. pE162 The pE162 plasmid is derived from the SP65 plasmid (Promega.). This derivative contains the intronless IvL4 Ta] gene (S.S. Ner and M. Smith, unpubl.) cloned in the same orientation and downstream of  76  the SP6 promoter. A unique Hincil site located in the 3’ untranslated region allows “run on” in vitro transcription. 2.14.2. Plasmid pLG669-Z and the YpSN series The CYC]:LacZ reporter plasmids, YpSN3 1, YpSN32, and YpSN33, were constructed by Dr. Sarbjit S. Ner (unpubi.). Plasmids YpSN3 1, YpSN32 and YpSN33 are derivatives of pLG669-Z (Guarente and Ptashne, 1981). These plasmids carry insertion of the different operator sequences (illustrated in Figure 11) into the downstream XhoI site of the CYC] promoter. YpSN3 1 contains the c42 operator from BAR]. YpSN32 and YpSN33 contain the c2 operator sequence from the M4 Tccl gene, and the pseudo c2 operator (psuedo BAR]), respectively. 2.14.3. CYCJ:URA3 and BAR1:URA3 Reporter Piasmids Plasmid Y11E36 Plasmid YHE36 was constructed as outlined (Figure 12) and used as the parental plasmid for both the CYC]: URA3 and BAR]: URA3 reporter plasmids. YHE36 is a derivative of the plasmid PVT-100-U (Vernet et a!., 1987).  YHE36 contains three important features including the Sall-XhoJ 2.3 kb LEU2  fragment from YEP13 (Broach et al., 1979) as well as the ADH]:MA Tcc2 transcriptional fusion gene and a promoterless ura3 gene. (i) The ADH]:M4Tx2 gene was generated by subclomng the BamHl-BclJ 0.65 kb fragment (containing the entire MA Tcc2 coding region) from pEMBL-MATc-(BamHI-BciI) into the BamHI site of PVT-100-U. A BgIJI and a Bc!! site at the 8th codon and the stop codon ofM4Ta2 were generated by in vitro mutagenesis, respectively (Figure 12).  (ii)  The promoterless ura3 gene was generated by  introducing a BamHl site after the first codon of URA3 (Rose eta!., 1984) and replacing the URA3 promoter with the LEU2 fragment from YEP13 (Broach et a!., 1979). A Sail and a BarnHJ site were regenerated in the ura3 gene to facilitate subsequent insertion of CYC] promoter or BAR] promoter. Plasmids YHE38 to YHE41 Plasmids YHE38, YHE39, YHE4O, and YHE41 contain in-frame CYC]: URA3 translation fusion genes (Figure 13), This series of plasmids was constructed by subcloning the SaiI-BamHl fragment from  77  pLG669Z  URA3  ‘S x /  /  /  /  /  CYCI  /  XhoI  /  LacZ  /  /  I Sail  F I  I  XhoI  pLG669Z  BamH[  Smai  XhoI CAATC CTCc3AG CAGATCC  CAATC  dcl  l cAGATCC  YpSN3 1  (MAToI (a1-c2)operator AATGTAGAAAAGTACATC  aATGTAgaa- ACCGAAAAAGGAA  pseudo BARI (  CATGTAATT-  ‘it  -AaTACATc  YpSN32  2) operator  ACCGAAAAAGGAA-ATTTAcATG  I_________ YpSN33  BAR1 X2) operator  Figure 11. CYCJ :LacZ Reporter Plasmids. The plasmid pLG669Z cariying the CYCI.’LacZ fusion reporter gene is shown (Guarente and Ptashne, 1981). The a1-x2 operator sequence from the haploid-specific gene, MA Tct.], was inserted at the XhoI site to generate plasmid YpSN3 1 (S.S. Ner and M. Smith, unpubi.). Plasmid YpSN32 and YpSN33 contain the pseduo BAR] (or pseudo cc2) and BAR] (x2) operator sequences respectively at the same site as YpSN3 1. The nucleotide changes within the c2 recognition site of the pseduo BAR] sequence (when compared to the BAR] sequence) are indicated by lower case letters, The MCM1 recognition site that was inserted into the middle of the pseudo BAR] sequence is denoted in upper case lettering and underlined.  78  insert BamHI-BcII MATo2 fragment ofMCS  MATx2  BamI-U  cli  Smal  MCS  Hlp  URA3 Sphl  /  I  I  HdIII  BamHI  SphI  Nrul (i) Deletion of MCS(BamHI-HdllJ) (ii) Introduction of BcIl site at the 3end and MATcx2 BgIlI site at codon S onto (iii) Introduction of a BamFil site after the first codon of IJRA3 (iv) Removal of NruI-BamHI URA3 promoter containing region and regeneration of BamHI site by a Sail linker inserton  XhoI EcoRl  Cial  Sail  Smal /  //  Kb XhoI-SaII LEIJ2 fragment from YEPI3  /  Sail ditat ‘.  /  ADH1p-MATh2  BcIl  /  \hgation  Bgiul  Figure 12. Construction of Plasmid YBE36. YHE36 is a derivative of the plasmid PVT-100-U (Vernet et al., 1987). The construction scheme of YHE36 shows the insertion of the M4 TcL2 coding region into the multiple cloning site (MS C) of the PVT-100-U vector which generated theADHl:IvIATct2 transcriptional fusion gene. The intermediate plasmid YHE34-1 was derived from this recombinant plasmid after a BamHI recognition site was created at the second codon of the URA3 gene using oligonucleotide-directed site-specific mutagenesis. Subsequently, the promoter region of this variant URA3 gene was replaced by the LEU2 fragment from the yeast plasmid YEP13 (Broach et al., 1979). This gave rise to the YHE36 plasmid which contains a promoterless URA3 gene.  79a  AHtp2.  Remove Sphl-Smal 2ii -region  Yep74 XbaI + Smal  XbaI  Smal  [1cEN pLG669Z isolate blunt-ended Xbal-SmaI CEN-ARS fragment  blunt-end ligate into SphI site LacZ  CYCI  Sail  ura3 BamHJ IGGA-TCc-AAA-GC ARSI CEN3  isolate Sail-BamHI CYC1 promoter fragment  Sail  Amp  deletion of Bgffl-BeII MATa2 fragment  YHE4I  CYC1:URA3 I  sI  1  ATG-ACCRA-TCcAAA-GCT CYC1  ura3’  YHE38  ):URA3 2 CYC1(al-ct  Y11E39  CYC1(4Ici2)URA3  YHE4O  CYC1(ct2):URA3  Figure 13. Construction of CYC]: UR.43 and BAR]: URA3 Reporter Plasmids. Plasmids YHE38, 39, 40, and 41 contain the different CYC]:URA3 reporter genes. Plasmids YHE64 and 65 cany the BAR]: URA3 and the variant BAR] *: URA3 reporters, respectively (page 79b). The construction procedures are described in the Materials and Methods section. The nucleotide sequences at the fusion junction of these reporter genes are displayed to indicate the “in-frame” fusions.  79b  PZV9 isolate 1.8kb HinDIII-BglII BAR lfraEment  pEMBL1 8+  subcloned into MCS  pEMBL-BAR1  BglII Rlj XbI  Hindill  ARS1  CEN3  RI  Clal  BARIp W Tx2operator AATFiATG ATG-TCT-GCA CAIGTAATr ±1  (SphIISmaI) Clal  OLIGO. CYH2 introduce BamHI site BamHI I3ARIp  W.Ta2operator  ATG-gga-TCc-GCA +1 BamHI OLIGO. CYH29 introduce mutations at operator  isolate ClaI-BamHI fragment  /  SalI-bluntend BamHI  N  BamHJ  /  /isolate ClaI-BamHI fragment mutan2operator-...  irirr  (Sall/Clal)  ATG-gga-TCc-GCA +1 BamHI  BAR1 :URA3  BAR1p  YHE64 W.T. or mutant cL2 operator  AATTAgATG  ATG-gga-TCc-AAA-GëT YHE65  +1  BAR1*:URA3 BAR1 :URA3  Bami-il  Figure 13. Construction of CYCJ: UR.43 and BAR]: UR.43 Reporter Plasmids (Cont.). Plasmids YHE38, 39, 40, and 41 contain the different CYC1: URA3 reporter genes (Page 79 a). Plasmids YHE64 and 65 carry the BAR]: URA3 and the variant BAR] *: UR.43 reporters, respectively. The construction procedures are described in the Materials and Methods section. The nucleotide sequences at the fusion junction of these reporter genes are displayed to indicate the ‘in-frame” fusions.  80  YpSN3 1, YpSN32, YpSN33 and pLG667-Z (Figure 11) into the SalI-BamHJ site of plasmid Y1{E36 (Figure 12). Plasmids YEE43 to YHE46 The plasmids YHE43, YHE44, YHE45, and YHE46 are pEMBL18+ (Dente at aL, 1983) derivatives which contain the SalI-BamHI CYC1 promoter fragments from YpSN3 1, YpSN32, YpSN33, and pLG669-Z subcloned into the multiple cloning site.  These plasmids were generated to facilitate  preparation of the radioactive DNA fragments for DNA-binding band-shift assay. Plasmids YHE62 Plasmid YHE62 is a derivative generated by replacing the yeast 2i sequence with CEN3-ARS1 and removing most of the MATa2 coding regions of plasmid YHE36 (outlined in Figure 13). First, the XbaJ-SmaJ (CEN3-ARS) fragment from Ycp74 (Ngsee, 1987) was used to replace the 2i.t sequence of YHE36. This gave rise to plasmid YHE48. Secondly, the BglJI-BcU (M4 Tcx2) fragment was released from YHE48 to generate YHE62 (Figure 13). 2. 14.3.5. Plasmids YHE64 and Y11E65 Plasmids YHE64 and YHE65 contain in-frame BAR]: URA3 translation fusion genes (Figure 13). YHE64 was constructed by subcloning the ClaI-BamHI (BAR 1* mutant promoter) fragment from pEMBL BAR1* into the Sail (blunt ended)-BamHI-cut YHE62 vector. YHE65 was generated similarly, except the ClaJ-BamHl wild-type BAR] promoter fragment (from pEMBL-BAR1 BamHI site containing derivative) was subcloned instead of the mutant promoter. pEMBL-BAR1 and pEMBLBAR1* were constructed from pEMBL18+ as shown in Figure 13. pEMBL-BAR1 contain the HindlII-Bgil (BAR]) fragment from plasmid PZV9 (this plasmid contains the entire 2.7 kb a-specific BAR] gene in pUC13, Kronstad et a!., 1987) subcloned into the HindJlI-BamHI site in the multiple cloning region. A BamHI site was introduced at the second codon of BAR] (MacKay et al., 1988) (illustrated in Figure 13). In addition, the x2 recognition site in the BAR] operator in pEMBL-BAR1 was mutated using in vitro mutagenesis to generate pEMBLBAR1* (Figure 13).  81  2.14.4. Construction of YRP-MATx2 Plasmid and its Derivatives Plasmids YRP-MATx2, YRP-matcL2M-64 and YP.P-matc2-1 were constructed from the plasmid YRP7 (Struhl eta!., 1979). The 4.2 kb HindlIl M4 Tc, fragment was isolated from plasmid pEMBL-MATx (BamHJ-BclI), This MATcL fragment containing the wild type MATc,2 gene was ligated to the linearized YRP7 plasmid which had been partially digested with HindilI. The recombinant plasmid with the MA Tcc fragment cloned into the HindIJJ site was verified by restriction endonuclease mapping. Likewise, plasmids YRP-matc2A4-64 and YRP-mat(x2-1 were constructed by insertion of the varied MATci. fragments carrying the matcj2A4-64 and matcL2-1 mutations, respectively. The matci2M-64 mutant matcL2 gene contains the N-terminal deletion of residues 4-64. This deletion was created by in vitro site-specific mutagenesis using pEMBL-MATcw.-(BamHI-BclJ) and the mutagemc oligonucleotide 64 .. (Table I). 4 OLIGOA  The matcw2-1  mutant carries an ochre mutation at codon 96 of the MA T(x2 gene obtained from the plasmid YEP-mat(x (Porter and Smith, 1986). 2.14.5. E. coil Expression Plasmid pLacMATh2 The BamHJ-BclJ MA Ta.2 fragment from pEMBL-Yr-MATx was cloned into the plasmid pLacCCP (the E. coli expression vector for yeast Cytochrome C peroxidase, (CCP) Goodin et a!., 1991) between the unique BamHI and BclJ site in the CCP coding region. The resulting recombinant plasmid pLacCCP-MATx2 was subjected to in vifro mutagenesis using the deletion CYH12 oligonucleotide (Table I).  Deletion of the CCP coding sequence from the CCP-MA Tc2 fusion gene gave rise to the plasmid  pLacMATc2. The plasmid was first screened by restriction enzyme mapping, then confirmed by DNA sequence analysis. 2.14.6. Construction of pEMBL-MATh-(BamHI-BciI) pEMBL-MATcL-(BamHI-BclJ) was derived from pEMBL-MATc (Porter and Smith, 1986). The parental plasmid was subjected to in vitro mutagenesis using two mutagenic oligonucleotides, SN5 1 and SN52 (Table I), in one round of in vitro site-directed mutagenesis (see Figure 10). The desired BamHI- and BclJ-contaimng mutant plasmid was screened by restriction enzyme digestion. Introduction of BamHI and  82  Bell sites onto the 5’ and 3’ end of the IvL4Tci2 do not affect its expression and repressor functions. This version of the IvL4 Tx2 gene was used as the wild type in this study (Figure 14). 2.14.7. Construction of pEMBL-Yr-MATa For expression of wild-type or mutant matcL2 alleles, the yeast E. coil shuttle vector pEMBL-Yr25 (Baldari and Cesareni, 1985) which contained the TRPJ marker was used.  The 4.2 Kb HindiJI ]vL4 TcL  fragment from pEMBL-MATcL-(BamHJ-BciJ) contains the entire MA TcL sequence.  This fragment was  subcloned into the BamHJ site of pEMBL-Yr25 at the mutiple cloning site by blunt-end ligating the the 4.2 Kb HindiJI ]vIATcL fragment with BamHI-linearized vector fragments. The resulting recombinant plasmid allowed packaging of the noncoding strand of the MA Tcj.2 gene (Figure 14). 2.14.8. Construction of pGEM3Zf(+)-MATcL2 For the purpose of in vitro transcription, the BamHJ-BciJ fragment containing the coding region of the MA Tci2 gene was released from pEMBL-Yr-MATx by restriction enzyme digestion and subcloned into the BamHI site of pGEM3Zf(+). The MA TcL2 gene was cloned in the same orientation as the T7 promoter, and the resulting recombinant allows a high efficiency of in vitro transcription when using T7 P.NA polymerase.  2.15. CONSTRUCTION OF MA Tcjr2 MUTANTS  2.15.1. Mutations Generated by in vitro Oligonucleotide-Directed Site-Specific Mutagenesis All matcL2 mutations were generated by in vitro oligonucleotide-directed site-specific mutagenesis (described above). Except for the multiple replacement mutants, c2B-1 and 167A+L74A+L81A, and the deletion mutants of x2A (&L2A) and c42B ( which were derived from pEMBL-Yr-MATx (Figure 14), all mutants in this study were derived from pEMBL-MATcw-(BamHJ-BciJ) (Figure 14).  83  pEMBL- MATh MATa2  introduce BaniHI, Bell sites pEMBL-MATx--(BHI-BcII) Xbal BolL  BamHl  1-lindlll  Hindlll  —I  B,mHl  pEMBL  ISedifi  MATa2 fl on  Hindlil digest blunt-end  isolate 4.2kb  on  BamHl digest  MATn.2 fragment  ARSI  XbaI  Amp Hthdlll  ARS1  Figure 14. Plasmids pEMBL-MATh-(BamHI-BclI) and pEMBL-Yr-MATct. The construction of the MATct2 expression plasmid, the pEMBL-Yr-MATct, is shown. A 4.2 kb HindIll MA Ta. fragment was obtained from the pEMBL-MATcc-(BamHI-BclI) plasmid which contains the M4 Tct2 gene with the BamHI and Bell sites at the 5’ and 3’ ends of its coding region. This 4.2 kb DNA fragment was subcloned into the BamHI site of the pEMBL-Yr-25 (Baldari and Cesarerri, 1985) to generate pEMBL-Yr-MATa..  84 Generation of Single or Multiple Replacements For all other single mutantions of the matci2 (except for those has been described above), the corresponding mutagenic oligonucleotides were used in the mutagenesis. Multiple replacement mutations were generated using head-to-tail mutagenic oligonucleotides which hybridized to adjacent sites as shown in Figure 10. The L8lIPstJ oligonucleotide was the mutagemc oligonucleotide for the single x2B mutant L81A. L88A(+T84N) was used in conjunction with oligonucleotides 167A, L74A and L81A for the synthesis of the cL2B-1 mutant. The following describes mutations generated using mixed oligonucleotides. Screening of C33F and C33Y matc,2 Mutants C33F and C33Y mutants were generated by site-directed mutagenesis using the mixed CYH63 oligonucleotide with the C33F/Y changes which contained mixed sequences corresponding to Phe/Tyr, TTIAC, at the 33rd codon (Table I). Mutants were first screened for the presence of an extra Hindill site, then by double-stranded DNA sequencing. Screening of C34F/L291 and C34F matcv2 Mutants For C34F1L29I and C34F mutants, the mutagenic oligonucleotide used was C34F/Y(+L291) in which the L291 mutation also introduced an extra BglII site. The C34F1L291 double mutant was screened for the acquisition of a BglJJ restriction site, and then verificated by DNA sequencing. The C34F single mutant was selected first by obtaining a mixed population of DNA plasmids containing both the single and double mutants from a single colony of the E. coil transformants.  Thus, the mix population contained  plasmids both with and without an extra BgiJJ site. This mixed population of transformants was resolved by restreaking on YT medium plates containing ampicillin so that each resulting single colony contained only one type of plasmid. The plasmid without a Bg1JI site was further analysed by DNA sequencing for the sole presence of the C34F mutation.  85  2.15.2. Mutations Generated by PCR Amplification The matci2A4-7 mutation was combined with the series of heptad mutations by PCR amplification of the mutant matcL2 genes using the deletion primer, CYH48, and the 3’-primer, CYH35 (Table I). The PCR products were digested completely with a 10-fold excess of BamHJ and Bell, then purified by the Magic PCR Preps DNA Purffication System (Promega) and ligated to the BamHl-BclI-cut pEMBL-Yr-MATx vector fragment which had been treated with Calf Intestine Phosphatase. After transformation of the DH5c cells, positive clones were first screened by BgIII restriction enzyme mapping, then confirmed by double stranded DNA sequencing.  86  RESULTS AND DISCUSSION (CHAPTER 1)  3.1.  3.1.1.  MUTATIONAL ANALYSIS OF THE MATa1-a2 DIMERIZATION DOMAIN  Identification of Two Hydrophobic Heptad Repeat Motifs: a2A and c2B The amino acid sequence of the x2 N-terminal domain was examined to identi1y any potential  dimerization motifs. Two 3,4-hydrophobic heptad repeat sequences which are consistent with the coiledcoil protein folding motif (Sodek et a!., 1972) were discovered. As described in the Introduction, the coiledcoil structure consists of two amphipathic right-handed x-helices aligned such that their non-polar faces are continuously adjacent, with residues in the a and d positions of each helix making up the hydrophobic interface (Alber, 1992; Adamson et a!., 1993). In the c2 N-terminal domain, the first repeat, x2A, extends between residues 1le22 and Val43, while the second, x2B, lies between 11e67 and 11e95 (Figure 6). To test the significance of these two repeat sequences in MATa1-c2 dimerization, mutations in these regions were generated and their effect on al-x2 dimerization was measured using different functional assays (described below). 3.1.2.  Functional Assays for al-ot2 Dimerization Four functional assays were employed to investigate the effect of mutations in the c2A and cc2B  regions. These assays will be discussed individually, as follows: (i) In vivo assays on the al-(x2 repressor function. The al-cc2 repressor activities of different 2 mutant proteins were measured by determining 3-galactosidase activity expressed from a LacZ heterologous reporter gene. Both the CYC](al-ci2):LacZ and HO:LacZ reporter genes were used for this purpose. The CYC] (a1-c2).LacZ reporter system was used for the initial studies on al-x2 dimerization (see Section 3.1.3.). The CYC] gene encoding the abundant iso-i cyctochrome c is not usually regulated by mating-type control (Smith et at., 1979). However, insertion of the al-cL2 operator (taken from the haploid specific gene promoter for MA Ta.!) onto the CYCI promoter causes al-ct2 repression (S. S. Ner and M. Smith, unpublished data; also see Miller et a!., 1985a). Therefore, the f3-galactosidase activity expressed  87  from the CYCJ (al-cL2).LacZ reporter carrying the al -c2 operator reflects repressor functions and hence al-a.2 heterodimerization capabilities. To maximize the sensitivity of the reporter system, the high-copy number 2.t-based YEP13 plasmid containing the CYCJ(al-cL2):LacZ reporter gene was used (Figure 11). Mutant c2 proteins were co-expressed individually with the CYC1(al-cL2):LacZ chimeric reporter genes in a cells, 1160 (for genotype, see Table Ii), via co-transformation of plasmids carrying the appropriate genes. Relative 3-galactosidase activities expressed from CYC] (al-ci2):LacZ were measured (as in Materials and Methods) and compared with that of the wild-type cr2. An alternative method to measure these activities for comparison with the wild-type ci.2 was used in later studies (Sections 3.1.4 and 3.1.5.).  This method employed an integrated single-copy HO:LacZ  reporter gene in the a cells (1402, genotype in Table II). HO is a haploid-specific gene that codes for the 110 endonuclease involved in mating-type switching in homothaflic haploid yeast strains (Russell et a!., 1986  and references therein).  Although the HO:LacZ reporter system  is weaker than the  CYCJ(al-ci.2):LacZ reporter, it is preferred as a reporter of al-c42 repression since using an authentic haploid-specific gene promoter eliminates potential artifacts which may be associated with a heterologous promoter. (ii) In vivo trans-inactivation assay. An indirect method of measuring the effect of different x2 mutants on al-(x2 heterodimerization involves the use of an in vivo trans-inactivation assay. In this assay, mutations of interest are made to the dominant negative mutant c2A4-7 which contains a deletion of residues 4 to 7 and normally heterodimerizes with al but yields a complex incapable of al-cL2 repression (Hall and Johnson, 1987). The dominant negative mutant, cc2M-7, competes out normal c42 binding to al and thus disrupts the normal control of gene expression, often by formation of non-functional regulatory oligomers or protein-DNA complexes (Herskowitz, 1987).  Loss of dimer stability due to mutations is  therefore reflected in terms of inability to increase expression of f3-galactosidase activity when the HO:LacZ reporter system in aJc diploids (1402x1420, genotype described in Table II) is used. The dominant negative mutant alone caused an approximately 23-fold increase in 3-galactosidase activity (Table V).  However,  mutations which diminish the affinity of c2A4-7 for al would allow the wild-type al -cL2 to repress the 3galactosidase activity (Sections to  88  (iii) In vitro DNA-binding electrophoresis mobility-shift assays. An assay used in this thesis to measure the binding of the al-cL2 heterodimer to its DNA target is essentially the same as that described by Dranginis (1990). Mutant cL2 and al proteins were produced by in vitro transcription and translation (see Materials and Methods). Because both c2 and al protein sequences contain four residues of Met, this allowed incorporation of radioactive methionine into the translated protein products. When non-denaturing gel electrophoresis is applied, free 35 S-labelled al-cL2 protein, which has a p1 (isoelectric pH) higher than the pH used in electrophoresis and thus exists mainly as non-charged or positively charged protein, would stop on top of the gel (For an example, see Figure 16). However, labelled a1-c2 protein acquires a higher mobility when bound to the negatively charged, unlabelled DNA probe (a 29 bp synthetic-oligonucleotide corresponding to the al-o2 operator sequence). Therefore, the protein-DNA complex shows up as a distinct band near the bottom of the gel after fluorography (Figure 16). The amount of a1-c2 binding to the DNA probe for the mutants is compared to that for the wild-type. (iv)  Co-immunoprecipitation assay.  A more direct in vitro method for measuring a1-c2  heterodimerization than the DNA-binding assays is an immunochemical one, In order to test the capability of c2 mutants to interact with al using such a method, rabbit anti-MATct2 antibodies were raised against a synthetic C-terminal fragment (residues 181 to 210). The anti-serum precipitates all of the cL2 mutants tested (in the absence of al) as well as wild-type x2 and cL2M-7. This serum also efficiently precipitated al-cL2, showing significant recovery of al after precipitation (approximately 50% recovery). The degree of complex formation between al and oc2 mutants was determined by SD S-PAGE estimation of the amount of al co-precipitated with mutant c2 after exposure of the mixture to x2-antibody (Section 3.1.3.  Construction and Analysis of the Internal Deletion Mutants of MATa2  (1) Construction of the z2A and x2B Deletion Mutants of MA Tcir.2 To determine the significance of the two heptad repeat sequences, internal deletion mutations, Aix2A and Ac2B, which lack the entire cL2A region (from 1le22 to Val43) and x2B region (from 1le67 to 11e95) respectively, were constructed by in vitro site-directed mutagenesis. The 4.2 Kb HindIll fragment of the MA TCL locus from plasmid pEMBL-MATc-(BamHI-Bc1I) (see Materials and Methods, Figure 14) was  89  subcloned into the yeast-E. coil shuttle vector, pEMBL-Yr25 (Figure 14).  The deletion mutants were  created using oligonucleotides OLIGO A and 2 2 OLIGO& B . (ii) Analysis of the Effect of the Internal Deletions The effects of &L2A and AcL2B on dimerization were measured indirectly by quantitating 3-galactosidase activity expressed from the CYC] (al-cc2):LacZ reporter gene (discussed in Section 3 1.2.). The 3-galactosidase activity expressed from this reporter carrying the al-cQ operator reflects repressor functions and hence al-x2 heterodimerization capabilities. Table III reports the relative 3-galactosidase expressions from CYC](al-cL2):LacZ.  Both Ac2A and AcL2B showed decreases in the a1-c2 repressor  fimction (approximately 70% and 40% of the wild-type function respectively; Table III).  These results  indicate that both (x2A and c2B are required for maximal al-cc2 repression. 3.1.4.  Production and Functional Analyses of Multiple Replacements in z2A and z2B  (i) Construction of Mutiple Arginine or Alanine Replacements Since large deletions in a polypeptide may drastically affect the overall folding properties, introducing missense mutations in (x2A or (x2B is probably a better strategy for demonstrating any bonafide involvements in dimerization. Therefore, c2 mutants L29R+L36R carrying a double Arg replacement at the a positions in a2A, as well as c2B-1 and x2B-2 mutants having four and five replacements respectively at the d positions with Arg or Ala residues (Figure 15), were prepared to examine the effect of altering the heptad repeat sequences on a1-c2 repressor function. replacements is as follows. 1990).  The rationale for multiple arginine and alanine  (1) Arginine and alanine are helix-favoring residues (O’Neil and DeGrado,  Therefore, potential cL-helical propensity in cL2A and OL2B carrying these multiple replacements  would not decrease. (2) Introducing charged residues or decreasing the hydrophobicity on the interface should destabilize the coiled-coil structure of the “Leucine-Zipper” motif (Schuermann et at., 1989; Turner and Tjian, 1989; Zhou eta!., 1992a,b).  90  Table ifi. The Effect of Ax2A and Act2B Mutations on a1-2 Repressor Function  CYC] (al-ci2):LacZ  al-cL2 Repression  Rel. 13-Gal. Activity.  (%)  Wild type  1.0  100  Ac2A  5.2  70  Ao2B  9.6  40  Vector  17.1  0  MA Tcj.2 allele  The ability of mutants &QA (deletion of 11e22 to Val43) and i\c2B (deletion of 1le67 to 11e95) to carry out a1-c2 repression was measured in the a cells, strain 1160. The plasmid-borne CYCI(aI-cv2):LacZ fusion gene was used as a reporter. The relative 13-galactosidase activities (Rel. 13-Gal. Act.) of yeast transformants expressing the (x2 mutants (versus that expressing the wild type) are reported. The extent of al-ft2 repression in percentage is estimated by the following calculation: [1 (Rel. 13-Gal. Act. of mutant! of vector)] x 100%; 100% repression has been arbitrarily assigned to the wild type. The average values of the absolute activities for the wild type and the vector controls are 0.58 and 9.99 units, respectively (see Materials and Methods for the definition of the unit). -  91 A 22  25  OL2A  ILDINK  \V.T.  a  d  j—;i L a  Mutant L29R L36R  1  36  def  a  32  39  43  SICCNLPKLPESV d  a  R R  L29R+L36R C33F C33Y  R  R F  C34F C34F/L291  F F  I  B 67  71  74  78  81  85  88  92  95  caB  ISDEEK  LLQTTSQLTTTITVLLKEMRSI  \V.T.  d  d  a  a  d  a  d  a  d  Mutant  167A L74A  L8 1A L88A 195A 167A+L74A L74A+L81A L81A+L88A L88A+195A I67A+L74A+L81A L74A+L81A+L88A L81A+L88A+195A cz2B-1 a2B-2  A  A A A A A A A  A A A  A A  R  A  Figure 15. The Identities of Various x2A and c2B Mutants. a2A and x2B mutations are shown in panels A and B, respectively. The a, d and e, fheptad positions are indicated below the amino acid sequence. Oblong boxes mark the mutations at the heptad repeats for the various mutants.  92  (ii) The Effect of Multiple Replacements on al-x2 Repression The al-x2 repressor activities of the three mutants, L29R+L36R, (x2B-1 and x2B-2, were measured in terms of -galactosidase activity expressed from an integrated single-copy HO:LacZ reporter gene (discussed in Section 3.1.2.). The oL2B-1 and x2B-2 mutants showed 45% and 60% of the wild-type al-0L2 repressor activity, respectively (Table IV). This indicated that extensive mutation of the putative coiled-coil region reduces activity to levels similar to those obtained with &x2B, the x2 mutant lacking this region altogether. Also, the mutant L29R+L36R, having two Arg replacements in the a positions, revealed a decrease in activity (to 40% of the wild-type level; Table IV), demonstrating an effect similar to that observed for the z\cL2A mutant which lacked the entire ix2A region. Studies on the Function of a2A and x2B in al-x2 Heterodimerization  3.1.5.  Construction of 2A and a2B Mutant a2 Genes  The results described above indicate the importance of the two cL2 heptad regions in repressor functions. In order to find out whether these two regions are actually involved in dimerization, a collection of mutations in ‘x2A and 2B was constructed. These mutations were targeted to the hypothetically most intolerant positions of the repeat regions. The mutants were constructed by oligonucleotide-directed site-specific mutagenesis as described in the Materials and Methods section (listed in Figure 15). Multiple head-to-tail primers specifying the substitutions were used to allow rapid generation of this series of mutations. In addition, some of the heptad mutations were individually combined with the dominant negative matcL2A4-7 allele using the polymerase chain reaction (PCR). A mutagenic DNA primer, which carried the deletion of the fourth to seventh codons, was used to amplify the mutant c42 genes using PCR. The resulting amplified c2 mutant genes contained both the zS.4-7 deletion and the heptad mutations of mterest.  The  abilities of the combined mutants in al-x2 repression and trans-inactivation of the wild-type al-cc2 system were tested and are described in sections and  93  Table 1V. The Effect of Multiple Heptad Mutations on a1-c2 Repressor Function  HO:LacZ  al-x2 Repression  Re!. n-Gal. Activity  (%)  Wild type  1.0  100  L29R+L36R  8.1  40  cL2B-1  7.6  45  cL2B-2  5.6  60  Vector  13.8  0  M4 TcL2 allele  The relative -galactosidase activities (Re!. 3-Ga1. Act.) of yeast transformants expressing the cL2 mutants (versus that expressing the wild-type) are reported. The extent of al-cL2 repression in percentage is estimated by the following calculation: [1 (Rel. 13-Gal. Act. of mutant] of vector)] x 100%; 100% repression has been arbitrarily assigned to the wild type. a1-c2 repression caused by multiple replacement mutants L29R+L36R, cL2B-1 and cL2B-2 is shown (for description of these mutants, see Figure 15). The relative 13-Gal. activities from the chromosomal HO:LacZ reporter were measured in a cells, strain 1402, which expressed these x2 mutants. The average values of the absolute activities for the wild type and the vector control are 0.01 and 0.14 unit, respectively. -  94  Determining the Effect of Ileptad Mutations on the Dominant Negative matcr2A4-7 Allele  The first group of mutants tested was cL2A(L29R+L36R), (x2B-1 and (x2B-2 because they showed significant decreases in cc2 repressor function (above). In order to measure their abilities in interacting with al, an in vivo trans-inactivation assay was employed (discussed in Section 3.1.2.). In this assay, mutations were made to the dominant negative mutant x2A4-7. Mutations which diminish a1-c2 stability would not permit cQA4-7 to increases expression of 3-galactosidase activity when the HO:LacZ reporter system in aJc diploids was used. While the dominant negative mutant alone caused an approximately 23-fold increase in 3-galactosidase activity (Table V, line number 9), none of the three mutations were capable of significantly increasing the j3-galactosidase activity (Table V, line numbers 7, 8; Table VI line number 6), demonstrating the poor affinity of these mutants for al and thus allowing the wild-type al-(x2 to repress the J3-galactosidase activity. The decreased heterodimer stability involving these multiple-site mutants was also shown, albeit qualitatively, using DNA-binding electrophoresis mobility-shift assays (Table V. Figure 16), in which the extent of binding of 35 S-labelled al-x2 to a synthetic-oligonucleotide promoter site is observed directly (discussed in Section 3.1.2.). As expected, x2B-1 and cL2A4-7/c.2B-l failed to show any binding (Figure 16). The loss of heterodimerization capacity caused by these mutations was also demonstrated using a co-inmiunoprecipitation assay.  The degree of complex formation between al and the (x2 mutants was  determined by the amount of al co-precipitated with mutant cL2 (discussed in Section 3.1.2.). Examination of both c2B-1 and (x2A4-7/x2B-1 in this manner indicated a more than two-fold decrease in al band intensity from the immunoprecipitate as compared to the same experiments using the wild-type c2 (Figure 17).  Mutational Analysis on the d Positions of z2B  In order to determine the relative contribution of residues at the 3,4-hydrophobic heptad repeat positions of 2B to the stability of the a1-2 complex, systematic replacements were made at the d sites and the integrity of the resulting heterodimers measured indirectly. Loss of hydrophobicity at these positions  95  Table V. Summary on Functional Analysis of cL2B Mutations  NO.  MATu2  HO:LacZ  alleles a  Rd. 13-Gal.  al-cL2 repression  C  Trans-  al-cL2 DNA  inactivation in  binding activity e  Activity b  (%)  a/cL diploids d  1  L81R  3.9  70  n.d.  +1-  2  167A+L74A  2.5  80  5.3  +++  3  L74A+L81A  4.3  70  0.8  +1-  4  L81A+L88A  5.6  60  0.7  +1-  5  L88A+195A  2.9  80  1.1  ++  6  L74A+L81A+L88A  5.2  60  0.7  +1-  7  x2B-1  7.6  45  1.1  8  cL2B-2  5.6  60  1.7  +1-  9  M-7  7.2  50  23  -I-H-  10  wild type  1.0  100  1.0  11  vector  14  0  1.3  -  n.d.  a  The cc2 mutations in the putative coiled-coils are described in Figure 15.  b  The relative f3-galactosidase activity of a HO:LacZ reporter in a cells (strain 1402) containing pEMBL-Yr-MATcL mutants versus the wild-type allele. Standard error of the mean was less than 40%.  C  The extent of al-cL2 repression in percentage is estimated by the following calculation: [1 (Rel. 13-Gal. Act. of mutant! of vector)] x 100%; 100% repression has been arbitrarily assigned to the wild type. -  d The relative 13-galactosidase activity in a/cL diploids (1402x 1420) of each x2 mutation combined with a second mutation, the dominant negative M-7 mutation (deletion of the N-terminal residues 4-7) is shown in line numbers 1 to 6. The ability of these combined mutants to trans-inactivate the wild type al-cL2 repression is reported as an increase in the relative 13-galactosidase activity in the a/cL diploids (1402x1420). The notation “n.dmeans not determined. e  The al-cL2 DNA binding activity as qualitatively determined from fluorographs (Figure 18, 19): +++, wild type activity; ++, less than WT; +, weak binding, +1-, very weak binding; -, no binding.  96  A 123456789 I  I  ‘p  1  th’ recr’r ccb,1.  ,i <  al  —  3i i Z  ————+++++ -‘  B  123456789  Figure 16. Effect of cL2B-1 Mutation on a1-c2 DNA-Binding. (A) The al and cL2 encoding plasmids (W.T., cL2B-1, M-7, or M-7Ax2B-1) were transcribed, and the S-Met in vitro. The translated products of wildRNAs were co-translated and the products labelled with 35 type and M-7 mutant showed a1-c2 dependent DNA-binding activity (Lanes 1 to 4, and Lanes 5, 7 of the indicated electrophoretic gel-mobility shift assay), while that of mutants c2B-1 and M-7/cL2B-1 failed to bind DNA with al (Lane 6 and 8 of the same fluorograph). (B) The in vitro translated products in (A) were resolved by 17.5% SDS-PAGE to ensure efficient translation of both a 1 and c42. One tenth of the amount of translation products used in panel A was loaded in each corresponding lane. The identity of the contents in each lane are the same as those indicated in the caption above the fluorograph in panel A. Lanes 1 to 4 do not contain al, and Lanes 5 to 9 contain the same amounts of al. The numbers on the left indicate the relative molecular weight of the BRL prestained low -molecular-weight markers in kiloDaltons.  97  246 135  7  8  9  10 12 11  x2 ‘-  r*;  al  —  o  thr  ++ + ++  — — e  ial Translation  Immuno precipitation  Figure 17. Co-Immunoprecipitation of al and x2. The al and ft2 proteins were co-translated and radioactively-labelled with 35 S-Met in vitro. The content of each lane is shown in the box above the fluorograph. The “+‘ and -‘ indicate co-translation with or without the al protein, respectively. Lanes 1 to 6 show the respective in vitro co-translated products. Lanes 7 to 12 show the products that were iinmunoprecipitated with the rabbit anti-MATcL2 serum.  98  due to replacements with Ala (or the polar residue Mg in the case of x2B-2) is expected to be destabilizing to the coiled-coil formation (Schuermann et a!,, 1989; Turner and Tjian, 1989; Zhou et a!., 1992a,b). Mutants incorporating all possible single, double and triple replacements at adjacent heptad positions, as well as mutants containing four and five replacements (Figure 15), were generated. DNA-binding assays of the cL2B single, double and triple mutants (Figure 18) generally showed a decrease in a1-x2 stability as the number of d position replacements increased. As in the case of mutants x2B-1 and x2B-2, comparable results were obtained with the repressor activity assays (Figure 18, Table V). Mi mutants carrying the L81A replacement showed significant losses in repressor activity, suggesting an important role for this residue which is located in the middle of the putative coiled-coil sequence. L8 1R, which carries a polar replacement at this central location, also showed reduced DNA-binding activity. Single and double mutants carrying replacements in the N-terminal region of x2B (167A, L74A, and 167A+L74A) displayed wild-type levels of DNA-binding, indicating a greater tolerance for diminished hydrophobicity in this region of x2B. It is interesting to note that the addition of the L81A replacement to the fully-active 167A+L74A double mutant (thus giving the 167A+L74A+L81A triple mutant) completely abolished activity. Results of the in vivo assays suggested a positional dependence for replacement sensitivity. First, only the 167A+L74A mutant, which showed wild-type DNA-binding activity, demonstrated a low level of trans-inactivation in the a/ct. diploid when combined with the ct.2A4-7 mutation, while the other Ala replacement mutants tested all conferred recessiveness (ie. did not interfere with wild-type al-cc2 repression) (Table V).  Secondly, the three Ala-replacement mutants that included the L8 1A mutation, namely,  L74A+L8 1A, L8 1A+L88A and L74A+L8 1A+L88A, all showed greater decreases in the in vivo al -cc2 repressor function than the double mutants lacking the L8 IA mutation. For example, both L74A+L8 1A and L81A+L88A (4.3-fold and 5.6-fold higher 3-galactosidase activity relative to wild-type x2, respectively, Table V) demonstrated less al-cc2 repressor activity than 167A+L74A or L88A+195A (2.5-fold and 2.9-fold activity, respectively), indicating the greater destabilizing effect of the former mutations, These quantitative results can be interpreted, along with the qualitative results from the DNA-binding assays, to suggest an  99  A a,  2468 1357 I  I  I  I  I  I  •  I  P  •  I  I  •  I.II•  ..l  I  I  I  I  10 12 141618 9 1113 15 1719  I  I  I  <  I  I  I  I<III<<Ic  I  I  I  •cC<.cc<IXc  I ,4I 1<  I  I  I.I<I I  I  •  —w-  I-  B 2468 1357  1012141618 91113151719  -H  a1  Figure 18. The al-cL2 DNA Binding Assay of the a2B Mutants (A) The DNA-binding electrophoretic gel-mobility shift assay of the in vitro co-translated al and cL2B mutants is shown. The isoleucine and leucine residues in the d positions of cx2B are indicated in the lefthand caption above the fluorographed gel. The letters, “A” and “R”, in the boxes above the fluorographs indicate alanine and arginine replacements at positions 67, 74, 81, 88 and 95. A dash in the boxes indicates no change at this position. All lanes contain co-translated al. (B) A 17.5 % SDS PAGE of the same co-translated products used in panel A is shown. One twentieth of the amount used in the DNA-binding assays was loaded onto each corresponding lane. The relative molecular weights of the BRL prestained low molecular weight markers are shown in kiloDaltons between lanes 8 and 9.  100  increasing order of importance of c2B residues at the d position to al-x2 complex stabilization: Ile67<Ile95Leu74<Leu88<Leu8 1 (Table V).  Mutational Analysis of 2A  (1) Investigation of Functional Significance of Cys33 and Cys 34 Phe or Tyr replacements of Cys33 in cL2 are known to yield al-cc2 repression deficient mutants which fail to repress haploid-specific-genes but not a-specific genes (Harashima et al., 1989), implying that Cys33 is essential for a1-c2 heterodimerization. This observation suggested that Cys33 might serve as a metal liganding site for the formation of metal-linked a1-c2 dimers, in a manner similar to that found in the  tat protein dimers of 11W (Frankel et a!., 1988; Harashima et at,, 1989). This view implies that the adjacent Cys34 should be equally essential,  C33F, C33Y and C34F mutants of c2 were therefore prepared to  determine the roles of these residues in a1-c2 heterodimer formation and activity. In accordance with Harashima’s findings (1989), Phe or Tyr replacement of Cys33 greatly impaired the DNA binding ability of al-x2 (Figure 19). The C33F mutant is significantly less functional than the wild-type for al-c42 repression as reported by HO:LacZ gene expression in the a/cL diploids, and when combined with the dominant negative x2M-7 mutation, C33F confers recessiveness. Phe replacement of Cys34, on the other hand, was far less detrimental to DNA binding and repressor activity (Figure 19, Table VI). C34F is fully functional for repression and also does not confer recessiveness to the cc2M-7 mutant. (ii) The Role of Leucine Residues in a Positions of 2A Two a position Leu residues in cL2A were mutated to Arg in order to disrupt the putative coiledcoil structure of this region. A greater tolerance for this replacement at Leu36 was shown by both in vivo and in vitro assays. Lower 3-galactosidase activity is permitted by the L36R mutant in the a1-o2 repression assay (Table VI). L29R lost all DNA-binding ability, while a low level was retained by L36R. The fact that the double mutant L29R+L36R did not further negate the repressor functions (Figure 19, Table VI), suggests that L29R alone can cause the maximal effect detectable under the present experimental conditions.  101  A  123456789 c1 -.-.-  +  OOOO..i..J Z  al  B  123456789  29:O 18.4 —-  -ic2 4 al  Figure 19. The al -c2 DNA-Binding Assay of the ft2A Mutants. (A) An electrophoretic gel-mobility-shift assay of co-translated al and cL2A mutants is shown. Different ci2A mutants (Lane 1 to 7) and wild-type x2 (Lane 8) were co-translated and radioactively-labelled with S-Met in vitro along with the a 1 protein. The identity of the c2A mutations are indicated above the 35 fluorograph. (B) A 17.5% SDS-PAGE gel of the co-translated products assayed in panel A was fluorographed. The relative molecular weight standards are shown.  102  Table VI. Summary on Functional Analysis of cc2A Mutations  NO,  MATcL2  HO:LacZ  al-o2  Trans-  al-ci2 DNA  alleles a  Rel. 13-Gal. Activity b  repression C  inactivation in  binding activity e  (%)  a/cL diploids d  1  C33F  2.2  85  0.7  +1-  2  C34F  1.3  100  9.0  +++  3  C34F/L291  3.2  80  0.8  +1-  4  L29R  8,1  40  1.0  5  L36R  5.0  60  1.5  6  L29R+L36R  8.1  40  0.8  7  i4-7  7.2  50  23  8  wild type  1.0  100  1.0  9  vector  14  0  1.3  -  +  -  ±++  n.d.  a The x2 mutations in the putative coiled-coils are described in Figure 15. b  The relative f3-galactosidase activity of a HO:LacZ reporter in a cells (strain 1402) containing pEMBL-Yr-MATcL mutants versus the wild-type allele. Standard error of the mean was less than 40%.  C  The extent of al-cL2 repression in percentage is estimated by the following calculation: [1 (Rel. 13-Gal. Act. of mutantl of vector)] x 100%; 100% repression has been arbitrarily assigned to the wild type. -  d The relative 13-galactosidase activity in a/cL diploids (1402x1420) of each c2 mutation combined with a second mutation, the dominant negative M-7 mutation (deletion of the N-terminal residues 4-7) is shown in line numbers 1 to 6. The ability of these combined mutants to trans-inactivate the wild type al-cr2 repression is reported as an increase in the relative J3-galactosithse activity in the a/cL diploids (1402x1420). e  The al-cL2 DNA binding activity as qualitatively determined from fluorographs (Figure 18, 19): +++, wild type activity; ++, less than WT; +, weak binding, +1-, very weak binding; -, no binding. The notation “n.d”means not determined.  103  Interestingly, a relatively conservative replacement of Leu29 to Tie in the highly functional C34F mutant generated a double mutant (L291+C34F) with significantly reduced activities (Table VI), thereby suggesting the remarkably sensitive nature of the Leu29 site. This result is not completely unexpected, however, since Zhou et al. (1993) have recently demonstrated that the packing and stabilizing influence of Ile differs significantly from that of Leu in the hydrophobic interface of coiled-coils.  3.2.  DISCUSSION ON THE SIGNIFICANCE OF a2A AND a2B IN DIMERIZATION The functional significance of the two 3,4-hydrophobic heptad repeats, namely (x2A and x2B, on  the c2 N-terminal dimerization domain has been studied by mutational analysis. Several lines of evidence suggest the importance of both ft2A and x2B in the formation of al-cL2 heterodimers. First, consistent with Hall and Johnson’s proposal that the al-cL2 interaction site lies between residues 20 and 140 (Hall and Johnson, 1987), deletion mutations of either x2A or c42B demonstrated detrimental effects on al-cL2 repression. Decreased repressor function was likely caused by the removal of functional elements rather than by perturbation of the overall protein folding. This idea is supported by the characterization of double Arg replacements in cL2A as well as the multiple Ala replacements in cL2B. Both multiple replacement mutants showed comparable decreases in function to the internal &2A and Ac2B mutants. Secondly, several experiments demonstrated that non-functional x2A or x2B mutations impaired the ability of the dominant negative matci2li4-7 mutant to interact with al. Results from three different approaches including in vitro DNA-binding assay, co-immunoprecipitation and in vivo trans-inactivation assay, all indicated that stability of al-x2 heterodimers requires both intact x2A and x2B regions. The action of these mutations on the dominant negative gene can be explained by the functional role of the two heptad repeats in interaction with al (illustrated schematically in Figure 20).  The replacements, in  particular cL2A(L29R+L36R), cL2B-1 and x2B-2, decrease the dimerization ability of a2A4-7, thus destroying DNA-binding and the capacity to interfere with al-x2 repression. In addition, the partial in vivo al-oc2 repressor activities associated with the single deletion mutant of either the x2A or x2B region,  104  N  N  N  DNA recognition site  Figure 20. A Schematic Model of the al-cc2 Heterodimer. The possible arrangements for the putative coiled-coil helices of the al-ct2 heterodimers are presented here. Cylindrical structures represent ct-helices. Ellipses indicate the homeodomains which bind to the DNA target (shown by the hatched bar). The left panel shows the N-terminus of the al helix interacting in a parallel fashion with ct2A through its alA region, and the C-terminus of al interacting in an anti-parallelmanner with oc2B via the aiB region. The right panel shows a different model in which the alA and alB helical regions interact, in parallel, with a2A and ct2B, respectively.  105  suggest that multiple elements are involved in al and x2 oligomerization.  Thus, this observation also  supports the present model that two coiled-coil forming motifs exist in the N-terminus of (x2. Further mutational analysis of x2A and x28 agrees with the premise that the 3,4-hydrophobic heptad repeats participate in coiled-coil like interactions with al (Figure 20), in which x2A and x2B fold as -helices and mesh onto a putative compatible helix in al. As predicted by the characteristics of coiledcoils (Hu and Sauer, 1992), a decrease of hydrophobicity in the a positions of x2A correlates with the loss of the ability to dimerize. Furthermore, the effect of systematic Ala replacements in d position sites suggested an  increasing  order  of  importance  Ile67<Ile95Leu74<Leu88<Leu8 1.  of  the  d  position  to  a1-c2  complex  stabilization:  The lesser contribution toward the end of the putative 0L2B helix is  consistent with the finding that hydrophobic interactions are less important at the flexible ends of the twostranded coiled-coil synthetic peptides (Zhou et a!., 1992 a,b), even though replacements in the terminal regions of integral coiled-coil proteins are often not tolerated (Kanaan et a!., 1992; Dowell et a!., 1992) because packing forces exerted by non-coiled-coil regions restrict flexibility at the end of the coiled-coil, Similar to that of x2B, an unequal significance among residues on the putative hydrophobic interface was also detected in cx2A. Leu29 in the a position of c2A is remarkably sensitive to mutations when compared with Leu36 in the preceeding a position, since Arg replacements showed a maximally detectable effect for Leu29 but not for Leu36 (Table IV, Figure 19). In addition, when tested on the C34F functional mutation, replacement of Leu29 with Ile, which differs from Leu by the hydrophobic side chain only, drastically reduced functional activities. However, such a position-dependent effect is not restricted to the hydrophobic interface of the putative coiled-coil. While neither of the two adjacent cysteines in x2 are known to contribute to disuffide bonds in the al-c42 heterodimer, the dimer stability was highly sensitive to Cys33 mutations but tolerant of Cys34 mutations. In a coiled-coil projection of the cc2A sequence (Figure 6), Cys33 and Cys34 are in the e and f positions, respectively; thus the results are compatible with the previous observation that residues in the f positions do not contribute significantly to coiled-coil stability, while residues in the a, d, e and g positions do (Hu and Sauer, 1992; Adamson et a!., 1993). In addition, the relative importance of Cys33 and Cys34 certainly indicates that the former is in a more sensitive environment within the protein. The coiled-coil projection reveals that Leu29 is in much closer proximity to  106  Cys33 (and thus to the sensitive environment) than is Leu36.  Again, this aspect of the model is in  agreement with the greater intolerance observed at Leu29. At this point, the action of the mutational effect is unclear, because the loss of function can be due to the absence of a functional Cys residue, or alternatively to an increase in the hydrophobicfty in the e position.  Nevertheless, the different effects for the two  positions argue against the speculation put forward by Harashima et al., (1989). If Cys33 serves as a ligand for formation of metal-linked a1-c2 dimers, then Cys34 should be equally important.  3.3.  3.3.1.  IMPLICATION OF THE MUTATIONAL STUDIES ON THE DIMERIZATION DOMAIN  Models of the al-cL2 Heterodimers In order for cc2 to heterodimerize with al via a coiled-coil interaction, the al protein must also  contain 3,4-hydrophobic heptad repeats. In accord with such a prediction, al contains two overlapping sequences, extending between Meti and 11e18 (called alA), and between Phel6 and 11e37 (called aiB) (Figure 8). These two sequences would give rise to two separate but partially overlapping hydrophobic faces if the entire Meti to 11e37 segment adopts a single cL-helix.  Therefore, this putative al helix allows  hydrophobic packing with both the putative c2A and (x2B helices. Although no al mutants were prepared in this study, Drangims (1990) has found that the N-terminal four amino acids are required for al-cc2 dimer formation. This tetrapeptide sequence contains a residue which is included in one of the proposed coiledcoil-forming regions of al. Since neither of the two putative hydrophobic interfaces on a 1 is large enough to accommothte either cL2A or x2B in their full length, x2A and c2B must be shorter than originally proposed. The length of the putative c2A helix is tentatively restricted to 11e22 to Pro37. Because many short coiled-coils are disrupted by the presence of proline (Hu et a!., 1990: Wild et at., 1992; Dobrzanski et at., 1993), the two Pro residues preceding Leu36 likely disrupt the helix.  Furthermore, as the mutational analysis of cL2B  suggests, the putative cL2B helix includes at least Leu8l and Leu88, as well as Leu74 or 11e95, while 11e67 may be excluded.  107  Apparently, hydrophobic interaction is the main force determining overall stability of the proposed al-cL2 coiled-coils. No obvious interstrand ionic interaction can be proposed. Figure 21 shows the two possible arrangments of the three putative helices. These arrangments are selected by the following criteria:  (i)  a maximum number of pairings at hydrophobic interfaces,  (ii) a maximum number of  favorable residues according to position; ie. unbranched hydrophobic residues at the d position and bulky branched hydrophobic residues at the a position (see Introduction), and (iii) a minimum steric hindrance between c2A and x2B putative helices in al-c2 heterodimer. Both arrangments for the heterodimer have x2A (11e22 to Leu36) forming an anti-parallel coiled-coil with alA (11e4 to 11e18), but they differ from each other by the sequence of x2B and by its hydrophobic packing with the aiB region. The putative c2B helix sequences Leu74 to Met92 and Thr78 to 11e95 may interact with the aiB sequences Phel6 to Leu33 and Leul9 to 11e37 respectively, in parallel configurations. The length of all proposed sequences of the putative coiled-coil helices in the models is shorter than the usual length in the b-Zip class coiled-coils (four to five heptad repeats). However, the models are feasible because the minimum length of the stable synthetic coiled-coil proteins is equally short (DeGrado et a!., 1989). Moreover, the GAL4 DNA-binding domain contains a short coiled-coil dimerizing structure which has as few as 14 residues (Marmorstein eta!., 1992). 3.3.2.  Mechanism to Define the al-z2 Heterodimerization Specificity  The model for al-c2 heterodimer formation utilizes hydrophobic faces of putative helical regions in both al and c2, suggesting that either protein might have the ability to homodimerize via coiled-coils involving the same hydrophobic faces. However, al-al homodimers have not been shown to form either in vivo or in vitro, and al protein alone does not exhibit DNA-binding properties (Dranginis, 1990; also demonstrated in this study, Figure 16).  These observations can be reconciled in terms of the present model.  The two  hydrophobic faces of al are roughly orthogonal when presented on the surface of the putative long helix (Figure 22). Although this arrangement may be ideally suited for association with the two separate and complementary hydrophobic faces of x2, both al faces cannot be fully paired in an al-al complex. For example, the pairing of the most N-terminal alA repeat sequences between two al proteins would leave the C-terminal aiB sequences unpaired, and likely exposed to the aqueous environment Thermodynamically,  108  Parallel a2B 78  74  2 NH  [1  88  85  92  L---T--[L[--I--L [78  85  81  OOH  95  92  COOlI  —jT--L---I--Lf.--M--I 2 NH  alA 1  1  V1a’ 30  20  10  —MDII CSMAENI NRTLFNI LGTEI D 2 NH [d  36  HOOC  -  -I d  -  -L  de  a:b  29  32  La  d  a  25 -  -  -  a  I d  -  -  221 2 I—NH  a,  b  LNTNNLYNFI—COOH li  b  e  aiB  a]  Antiparallel c2A  Figure 21. The Possible Arrangements of the Putative Coiled-Coil Helices in the a 1-ix2 Dimer. The two most likely arrangements of the putative coiled-coil helices in the a1-c2 heterodimer are presented here. The alA and aiB heptads are respectively aligned with the c2A and (x2B repeats to show the possible side chain interactions at the proposed hydrophobic interface (a, d positions). The amino acid sequence of al is displayed and the heptad position of alA and aiB are marked with lower case letters (a-d and a’-d’, respectively). The positions of the amino acids are also indicated by the numbers above the al sequence. Only the a, d residues in the cL2A and x2B repeats are shown while the other heptad residues are indicated by dashes because they do not interact with al at the hydrophobic interface. Hypothetically, only the a, d residues of the al repeats can make contact with the aligned ft2 hydrophobic residues shown below or above the al sequence. The putative alA helix interacts simply with the antiparallel x2A whereas the aiB helix could interact with the parallel cc2B in one of two possible arrangements. The shadowed rectangle encloses the d residue at the center of the two possible interacting regions of x2B (namely Leu8 1 or Leu88) with the 11e26 of the aiB region it potentially contacts. ,  109  A  (2)  (1)  (3)  N  Fa1A alA  aiB  C  B  destablized  destablized N  aiB C  stablized  Figure 22. Hypothetical Coiled-Coil Interactions of al Homodimers. In panel A, a schematic representation of the al helices is displayed. The destabilizing mechanism which may prevent al homodimerization is illustrated. A side view of the al helix (drawing 1) shows the two nearly orthogonal hydrophobic faces of alA (shadowed) and aiB (solid black). The top views of two different arrangements for al-al homodimerization are illustrated in drawings 2 and 3. In panel B, the side view of a variant al helix is shown. On this variant helix, the alA and aiB hydrophobic faces are aligned. Such an alignment can be achieved by deleting one amino acid such as Phel6 or by adding six amino acids at the alA-alB overlapping region. Alignment of the two hydrophobic faces would significantly stabilize the homodimers since both faces can be paired up in such a novel al-al dimer (shown in the right side of panel B).  110  these unsheltered hydrophobic faces would greatly destabilize the hypothetical al-al homodimer (Figure 22, panel A). This destabilization of the al-al homodimers can serve as the underlying mechanism for the control of dimerization specificity between al and c42.  This mechanism differs substantially from that  which directs heterodimerization specificity between the oncogenic gene products, Fos and Jun (Schuermann et at., 1991; O’Shea et a!., 1992). In the Fos-Jun case, complementary net charges between the coiled-coil-forming regions of these proteins have been shown to promote heterodimenzation while discouraging homodimerization. Functional Relationship of Putative Coiled-Coil Motifs and Homeodomain Proteins  3.3.3.  Evidence for Functional Coiled-Coil Motifs in Bomeodomain Proteins  Both al and x2 contain C-terminal homeodomains for DNA binding (see Introduction).  The  homeodomain is comprised of three (x—helices connected by short peptides and DNA-binding specificity is mainly defined by the third helix which makes specific contacts with nucleotides in the major groove of the target DNA duplex. DNA sequence analysis of the homeobox genes, Athb and HAT, isolated from the plant Arabidopsis thaliana, has revealed that these proteins contain a leucine zipper motif located Cterminal to the homeodomain (Ruberti et a!., 1991; Schena and Davis, 1992). However, the target genes regulated by these plant proteins, and the nature of their interactions with other proteins, is unknown. The human immunoglobulin octamer binding protein, Oct-2, also contains both the leucine zipper motif (Clerc, et at., 1988; Lupas et at., 1991) and a functional homeodomain, although the leucine zipper was shown to be non-essential for cooperative DNA binding (LeBowitz et a!,, 1989), The putative coiled-coil-forming sequences in al and c2 may therefore signil the first evidence of functional homeodomain-coiled-coil proteins.  Position of Leucine-Zipper-Like Motifs with Repect to Homeodomains  The location of the coiled-coil with respect to DNA-recognition domain may not be as restricted in the homeodomain-coiled-coil class of proteins as it is in the b-Zip class (Kerpolla and Curran, 1991; Hu and Sauer, 1992) and in the plant HD-ZIP class (Sessa et al,, 1993), since the proposed dimerization  111  domains of al and cL2 are linked to the homeodomains by peptides of slightly different lengths. In x2, 42 residues separate the last 3,4-hydrophobic repeat residue (1le95) and the first residue of the first helix in the homeodomain (Lys138, Wolberger, 1991). The analogous residues in al (11e37 and Pro79) are separated by 41 residues. In contrast, the C-terminal hydrophobic repeat and the homeodomain of BAT (Schena and Davis, 1992) have a short spacing of seven residues which is identical to the distance between the basic DNA-binding domain and the dimerization domain in the b-Zip family of proteins (Landschulz et a!., 1988a,b; Agre eta!., 1989; Struhl, 1989; Jones, 1990). This spacing, presumably part of a continuous helix as in GCN4 (Ellenberger et a!., 1992), has been shown in many cases to be critical for proper orientation of the DNA-binding domain with respect to the site of dimerization.  Also Sessa et a!. (1993) recently  demonstrated that the correct spacing between the putative homeodomain and the C-terminal “zipper” helix is critical for the in vitro specific DNA binding. On the other hand, the long spacer in c2, which apparently adopts a random extended-coil structure (Vershon et a!., 1993), suggests that these two domains could be less interdependent than in the b-Zip class, and that the coiled-coil motifs could serve only as dimerization domains and exert little influence on the orientation of the homeodomains to the DNA. Such flexibility may be necessary to accommodate interaction of x2 with MCM1.  3.4.  3.4.1.  FUTURE PROSPECTWES  Evaluation of the al-z2 Model The experimental data in this study have demonstrated that the two 3,4-hydrophobic heptad  repeats of x2, x2A and x2B, are important in dimerizing with al. The importance of hydrophobicity at the interfaces of the proposed c2A and x2B helices, in addition to the discovery of compatible overlapping heptad sequences in al, led to a tentative model for al-cc2 heterodimers. Such a model serves as an useful guideline for future investigations on protein-protein interactions among homeodomain proteins. Before a crystallographic or NMR 3-D structural model of the a1-c2 heterodimer (or the al, cL2 monomers) is available, further in vitro and in vivo studies using mutational analyses would help evaluate the accuracy of the present model.  112  First of all, the involvement of the putative al helix should be determined. The simultaneous binding of both (x2A and c2B on the al helix would strictly require the native orthogonal and spatial orientation of the two 3,4-hydrophobic heptad repeats on the al helix, namely alA (Ile4 to 11e18) and aiB (Phel6 to 1le37). alA and aiB sequences are partial overlapping and are out of phase by one amino acid, yet their hydrophobic faces are nearly orthogonal. Hypothetically, one can manipulate the orientation of these two faces by deleting or adding amino acids in the overlapping region. The effect on al-x2 dimer functions can be analyzed by approaches similar to those in this thesis.  Specifically, the in vivo al-c2  repressor function of such al mutants can be measured by quantification of f3-galactosidase activity expressed from the HO.LacZ reporter gene in M4 Tcc cells when different versions of the al proteins are individually expressed in the host cells. In vitro DNA-binding assays and co-immunoprecipitation assays can also be used to test this model. Deletion of Phel6 of al should bring the two hydrophobic faces into alignment. Aligning the two faces may prohibit hetero-dimerization activities because a direct overlap of the interfaces prevents x2A and cL2B from binding at the same time (Figure 21).  On the other hand, such mutations may promote  homodimerization of al (Figure 22). The de novo DNA-binding activity may be assayed with a newly designed DNA probe which contains a pair of symmetric al-binding half-sites instead of the a1-c2 recognition site. If such a novel al-al homodimer occurs, the present model would be justified. The importance of the putative al helix can also be demonstrated by replacement with helixdisrupting residues, Pro or Gly, as well as multiple Ala replacements (which presumably destroy the amphipathy of the helix) along the two arrays of a and d sites on the al helix. Approaches similar to those of the mutational analyses employed in this thesis will also be useful in defining the extent of the two putative al coiled-coil regions. 3.4.2.  Further Study on a2A and u2B of 2 Although the a, d positions which make up the hydrophobic interface of coiled-coil helices are  usually the most important in determining the overall stability of coiled-coils, non-a, d positions could contribute to intrastrand or interstrand ionic interactions and thus could also contribute to stability  113  (Adamson et a!., 1993). In this thesis, in an initial attempt to test the validity of coiled-coil formation in x2, mutations were directed to the most obvious residues (Leu or Tie) at the a positions for cL2A, and the d sites for cL2B. Such an approach has proven useful in obtaining informative data. Nevertheless, a few mutations were made to non-a,d positions. These included Phe replacements in the e (Cys33) andf(Cys34) positions for x2A, and the conservative replacement of Thr83 by Ser in thef position of cL2B (characterization of this mutation is reported in Appendix A). However, one would prefer to further characterize the two regions by introducing more mutations in non-a and non-d sites of x2A and x2B, respectively. In this regard, one can create alanine replacements scanning through each position in the entire helical region of interest. Such an approach was previously utilized for studies on the interaction of the human growth hormone with its hormone receptor (Cunningham and Wells, 1989). More systematic and exhaustive multiple alamne replacements (Fisher et a!., 1993) throughout the helical regions would also be important for a thorough study of the x2A and x2B putative helices.  Besides the helix-favouring  characteristics of Ala, Ala replacement is advantagous because such a mutation is not liable to any complication of interpretation due to introduction of basic, acidic, bulky-hydrophobic or long-aliphatic residues. For instance, this approach will help clarify the role of Cys33 in cr2A. Because of the bulkiness of the Phe or Tyr side chains, it was not possible to determine whether the C33F or C33Y phenotype was caused by the replacment of the Cys residue or simply by an intolerance for a bulky group in that position. Ala replacement at this site would eliminate the latter possibility. It would also be desirable to determine the end residues of the two putative cL2 helices. In the proposed model, the length of both cL2A and x2B were estimated according to the two hydrophobic repeats on the putative al helix. In regard to the first repeat of cL2, the following should be considered: (i) The proline residues at position 37 and 39 may be important for specifying the end of the c2A helix. It would be interesting to know if replacements of these prolines with other helix-promoting amino acids (such as alanine) would affect the dimerizing ability. Eliminating the proline residues may allow continuation of the helical propensity of x2A and this may destabilize the putative coiled-coil because new hydrophobic packing of a longer mutant c2A helix could conceivably introduce steric hindrances to the ends of z2B and the mutant cL2A helices. Furthermore, redefining the end may also shorten the linker sequence between the two  114  helices and the helices may be constrainted in how they orient. In both cases, a negative effect is expected. (ii) Since 11e22 is the first possible hydrophobic residue (as the sequence before it is not required for dimer formation), studying the dimerization ability of mutants carrying an Arg replacement at this putative end residue (or an addition of Pro or Gly to the C-terminal side of this position) may help justif’ 11e22 as the end. The two parallel arrangements of cL2B and alB sequences (where either Leu8l or Leu88 of ct2B interacts with 11e26 of al) were proposed in the al-x2 models (Figure 21). Because data from this work did not distinguish which of the two end residues (Leu74 or fle95 of cL2B) proposed for the different arrangements is correct, further experiments should be done. To determine which of these arrangements is in fact the case, one may cross-link the two faces of al and cL2. If one or the other of the x2B-alB models is valid, then specific d-d residues could be cross-linked via replacement with a pair of cysteine residues on the different interfaces. For example, if Cys replacement of Leu88 shows a higher rate of cross linking with al (containing a Cys replacement of 11e26) when compared with the replacement at Leu8 1, then the x2B (Leu78-11e95) and alB (Phel6-Leu33) arrangement would be more appropriate. In any case, the correct arrangement would be more stable since studies in synthetic coiled-coils indicate that disuffides at the d positions are not significantly destabilizing (while ones at the a positions are) [Zhou et al,, 1993]; and thus, the preferred arrangement would be more readily cross-linked by the disulfides (due to the close proximity brought about by the coiled-coil formation). Competition Study on al-i2 Heterodimer  3.4.3.  The co-immunoprecipitation experiment on al-u2 in this study and the trans-inactivation ability of the dominant negative mutant lacking a functional homeodomain (Hall and Johnson, 1987) indicated that the heterodimerization activity does not require DNA-binding. Hence, the N-terminus canying intact c2A and a2B suffices for al-cL2 dimer formation. This premise can be tested by a competition study using a similar DNA-binding assay to the ones described in this thesis. One could co-translate, in vitro, a truncated version of the wild-type or mutant cL2 bearing x2A and u2B mutations, along with the full-length wild-type al and c2.  These wild type or mutant truncated protein domains might be able to compete for al  115  dimerization sites and block the full-length al-cL2 DNA-binding. This activity can then be measured by DNA-binding assays. The N-terminal domain of the truncated proteins which include intact (x2A and c2B (or even the N-terminal domain of the al helix) should be able to compete the DNA-binding of the wild type a1-x2 dimers, whereas a mutant N-terminus might not. Alternatively, stable synthetic peptides corresponding to the x2A, cc2B or al helices can be used as competitors for DNA-binding. The peptide helices may compete by binding with the N-terminus of wildtype al or c2 before they can dimerize and bind DNA. A similar approach to this has been successfully applied by Schmidt-Don et a!. (1991) for the study of the dimerization motif of the oncoprotein Jun.  116  RESULTS AND DISCUSSION (CHAPTER 2)  4.1.  CHARACTERIZATION OF o2-MCM1 DNA-BINDING SPECIFICITY  Although both the al-oc2 and cL2 DNA-binding half-sites (Figure 11 and 26) contain a consensus core-dyad sequence [ATGT  TACA] (Miller et al., 1985a), the a1-c2 operator differs from the c2  operator in two ways. First, the 9 bp a1-x2 half sites (from MA TcL1) contain as little as 3 or 4 bp differences just outside the core-dyad compared with the cL2-binding half-sites (from BAR]) (Figure 26, bottom panel). Secondly, the al-c42 operator lacks the MCM 1-recognition sequence which is located in the middle of the c2 operator. Since x2 alone is unable to recognize the a1-c2 operator, at least one of the differences stated above must contribute to its inability to bind the al-cL2 operator. It is expected that insertion of the MCM1binding site into the center of the al-oc2-operator sequence would create a functional chimeric operator for cL2-MCM1, provided that the x2-MCM1 DNA-binding can tolerate the small sequence differences between the al-cc2 and c2 half-sites.  In this part of the thesis, studies on cL2-MCM1 DNA-binding to such a chimeric operator are reported, A variant x2 operator (pseudo BAR], Figure 11 and 26) was chosen to test the DNA-binding specificity of x2-MCMI.  The pseudo BAR] operator was created by inserting an MCM1-recogmtion  sequence into the centre of the a1-c2 operator which was taken from the MA TCL] gene (S. S. Ner unpublished data). Thus, this chimeric operator resembles the wild-type x2-operator from the BAR] gene [the c2 (BAR]) operator, Figure 11 and 26]. Two separate approaches were taken to examine the functions of this pseudo BAR] operator. First, in vitro cL2 and MCM1 cooperative DNA-binding to this operator was determined. Second, in vivo operator functions for both c42- and a1-c2-repression were analyzed.  4.1.1.  In vitro 2-MCM1 DNA-Binding To facilitate characterization of the c2-MCM1 DNA-binding activity, the x2 protein was  expressed and purified from E. coil (gift of Peter L. Davies). DNA-binding band-shift assay and DNase I  117 footprinting were used to ensure that the c42 and MCM1 bound cooperatively to the wild-type c2 (BAR]) operator in vitro.  E. coli-Expressed x2 Proteins  For over-expressing cL2 proteins in E. coil, the M4 Tci2 gene was placed downstream of the Lac promoter on an E. coil expression plasmid (see Method and Materials, Figure 23). As expected, expression of the J.v14 To’.2 gene from this recombinant plasmid, pLacMATcc2, was under Lac promoter control and the c2 proteins were produced only when the Lac repressor was inactivated by IPTG (Figure 23). Moreover, the inducibly-expressed products were recognized by the rabbit anti-MATc2 serum and exhibited the correct relative molecular weight (26,000) for the a2 protein (Figure 23).  The proportion of the disullide-linked dimeric and the unlinked monomeric x2 proteins in the E. coil expression system was examined. The reason for examining the proportion of these two forms of ct2 proteins is that the c2 dimer binds to the DNA target at about 100-fold higher affinity than the monomer does, thus existing disuffide-linked dimers could complicate interpretation of the in vitro DNA-binding results. E. coil lysates containing the cL2 proteins, with and without the reducing agent f3-mercaptoethanol which can destroy the disuffide-linkage, were resolved by SD S-PAGE. After Western analyses, the results showed mainly monomeric x2 proteins in the E. coli lysates, although a small portion of the isolated c2 proteins from E. coil were detected as disulfide-linked dimers (Figure 23).  DNA-Binding Band-Shift Assay of the cL2 Proteins  The purified, E. coil-expressed c2 protein was tested for its ability to bind to DNA cooperatively with MCM1. DNA-binding band-shift assays were carried out using the CYC] promoter fragment which contained a wild-type x2 (BAR]) operator sequence (Figure 24). Restriction enzyme digests of plasmid YHE45 released a 280 bp DNA fragment which corresponded to the x2-operator-contaimng sequence. This fragment was radioactively labelled and was used as the DNA-binding probe in the band-shift assays. After mixing this probe with various protein samples (e.g. the x2 protein or the MCM1-containing yeast extract  118  A  MATa2  cZp pLacZ-MATu27 ATG  1  2  BcII  XbaI  43.0 29.0  x2  4  18.4 14.3 I PTG  B  Whole  cell  lysate  isolated  a2  40  cx2—ct2 1  I 43.0—  -29.0 %‘  —18.4  .+_/  1 4.3  H  -  +  +  —mercaptoethanoI  -14.3  I  -  +  +  I  Figure 23. The Expression of the c2 protein in E. coil and Dimer Analysis of the Expressed Protein. (A) IPTG induction of c2 protein expression in E. coil: E. coil JM1O 1 harbounng the MA TcL2 expression plasmid pLac-MATu2 (shown on the top of this panel) was grown at 37°C and IPTG was added to induce the LacZ promoter. After a 2 hour induction, the E. coil cells were lysed and the whole cell lysate was subjected to 12.5% SDS-PAGE. Western blot analysis was carried out using the rabbit anti-MATc2 serum and the alkaline-phosphatase-coupled anti-rabbit IgG. Lane 1 is the control without IPTG induction which is indicated with the notation. Lane 2 is the IPTG induced sample which is noted by the ‘+‘ symbol. The numbers on the left indicate the BRL low molecular weight standards in kiloDaltons. ‘-‘  (B) Disuffide-linked dimer analysis of E. coil expressed c2 protein: The presence of disulfide-linked ce2 homodimers in the E. coil whole cell lysate and in the purified c2 proteins was tested. The tested samples were run on a 12,5% SDS-PAGE gel in the presence (lanes 3, 4, 6, and 7) or absence (lanes 1, 2, and 5) of f3-mercaptoethanol and then Western-blotted on a nitrocellulose membrane. The cL2 monomers and dimers were detected using the ECL system (Amersham). The unlinked monomers and disulfide-linked dimers are indicated by solid arrows. The dashed arrow marks the expected position of the linked dimer. Lanes 1 to 4 are E. coil whole cell lysates prepared from the JM1O1 transformants carrying the plasmid pLacZ-MATcL2 (lanes 1 and 3) or the control plasmid pEMBL18+ (lanes 2 and 4). 100 ng of isolated E. coil x2 proteins were loaded onto lanes 5 and 6. Lane 7 is a “no protein” control. The molecular weight standards in kD are shown on the left and right sides of the two Western blots.  119  1  2  3  4  5  6  7  cz2-MCM1 -DNA  WWhiiIi  MCM1  -  +  -  9 1011  P b  +  +  +  +  +  +  +  +  -  +  +  +  +  +  +  +  +  +  V flC)C)flV <•<< -<me  CD  C)  o 3 CD  w  Rb  free DNA  a2  8  MCM1 -DNA  * -  -  -  ‘-‘O -  DO  0 -  Figure 24. The Competition Analysis of the cL2-MCM1 DNA Binding. The competition study for specific (x2-MCM1 binding was carried out using DNA-binding band-shift assays. P-labelling of the 280 bp MIuI-BamHJ fragment from the The DNA-binding probe was prepared by 32 plasmid YHE45 (see Materials and Methods). This DNA probe carrys a part of the CYC] promoter as well as the complete BAR] c2-operator. The absence or presence of 2 and a MCM1 -containing yeast extract in the reactions is indicated by the ‘-‘ and ‘+‘ signs below each lane. Unlabelled-competitor DNA fragments were added in lanes 4 to 10. Their identities are as follows: lane 4 contains the native CYC] promoter fragment without any insert; lanes 5-7 contain the CYC] promoter fragment carrying the al-cL2-operator; the pseudo x2-operator and the x2-operator, respectively; lanes 8-10 contain pEMBL18+, the synthetic ste6 operator DNA-duplex, and the variant ste6* operator duplex, respectively.  120 alone, or both of them together), the mixtures were subjected to non-denaturing polyacrylamide gel electrophoresis. The protein-bound DNA probes ran at a lower mobility than the protein-free form did. Therefore, these “shifted’ DNA bands indicated protein binding to the DNA binding probes.  The following three observations were made from the band-shift assays. First, both the purified E. co/i c2 and the crude yeast extract containing MCM1 were necessary for forming the x2-MCM 1-DNA complex in the band-shift assay (Figure 24, lane 1, 2, 3, and 11). The crude MCM1-contaimng yeast extract was prepared from MA Ta cells (S85, Table II) that lack endogenous cL2 proteins.  Secondly, this DNA-  binding was specific to the BAR] ct2-operator. Only the CYC] promoter fragment containing the BAR] c2operator competed for DNA binding, while a similar fragment containing the a1-c2 operator or the pseudo BAR] operator did not (Figure 24, lane 4, 5, 6, and 7). Thirdly, an intact MCM1-recognition site on the c2operator sequence is essential for cc2-MCM1 binding to the DNA.  The synthetic DNA duplex  corresponding to the wild-type x2 operator sequence (from the STE6 gene), which contains an intact MCM1-recognition site, was capable of competing off the binding (Figure 24, lane 9). On the other hand, a similar duplex carrying a non-functional MCM 1-recognition site (called the ste6* mutant c42-operator) was not able to compete (Figure 24, lane 10). In this ste6* operator, the critical GG residues of the two MCM1binding half sites (ATTAGG) were changed to CC residues, therefore destroying its function. Thus, this observation led to a conclusion that the MCM1 proteins in the crude yeast extract must be a component of the protein-DNA complex.  DNase I-Protection Footprinting with MCM1 and 2  To confirm the observed sequence-specific binding by the E. coil-expressed cL2 protein, DNase I protection footprinting assays were also used to examine the x2-MCM1 binding to the c2-operator. A 99 bp synthetic DNA duplex, SN44/5N45 (Table I) which corresponds to an unrelated DNA sequence carrying the x2-operator was subcloned into the pEMBL18+ plasmid. The 117 bp BamHI-PstI restriction fragment from this resulting plasmid was 32 P-labelled on the 3’ end of the SN54 strand. When this x2-operator-containing fragment was subjected to DNase I digestion, the resultant DNA-ladder exhibited a continuous pattern.  121  However, when this labelled DNA fragment was bound by the x2-MCM1 complex, the DNase I digestion showed a gap in the DNA-ladder covering the protein-protected region. The exact sequence was determined by running the Maxam and Gilbert sequencing ladders of the DNA probe along with the footprinting samples.  The footprinting results showed that the purified proteins protected the x2-operator sequence when supplied with a crude yeast extract containing MCM1 (Figure 25, lane 4), thus verifring the sequencespecific binding observed in the band-shift assays described above. In addition, the cL2 proteins or the MCM1-contaimng yeast extract alone failed to protect the c42-operator sequence (Figure 25, lane 3, 6) when similar experimental DNase-footprinting conditions were applied.  4.1.3.  x2-MCM1 DNA Binding to the pseudo BARJ Operator  Using the purified E. coil cL2 protein, the MCM1 -containing yeast extract, and the DNA probe which corresponded to the CYC] promoter fragment containing the pseudo BAR] operator (see Materials and Methods, Figure 26), in vitro DNA-binding band-shift assays were carried out to determine the c2MCM1 binding to the pseudo BAR] operator. As shown in lanes 6 and 7 of Figure 26  ,  the pseudo BAR]  operator was still capable of x2-MCM1 binding. Nevertheless, the binding to the pseudo BAR] operator was significantly less than that for the wild-type BAR] cL2-operator (Figure 26, lane 5 to 10 compared with lane 1 to 4).  Consistent with the decreased binding affinity to the pseudo BAR] operator, this DNA  fragment also failed to compete off the binding to the wild-type operator (Figure 24, lane 6).  4.1.4.  In vivo Function of the pseudo BAR1 Operator  To determine whether the pseudo BAR] operator causes c2 repression or not, a sensitive in vivo reporter system was constructed in order to measure the operator function. This in vivo reporter system consisted of two parts. The first part involved a high-expression of the MA TcL2 gene and the second part consisted of the heterologous CYCI: URA3 reporter gene which carried the pseudo BAR] operator sequence (as well as the similar reporter genes carrying the al-x2- or oc2-operators).  122  51  27  C T C C A T G T A A A T T T C C T T T T T C G  a2r- +++MCM1 competitor  -  -  — —  —  -  + ÷ 1234567 —  —  base ‘4  J4  1  117  80 65  4  50  4  35  ::..  Figure 25, The DNase I Footprinting Assay of the BAR] a2-Operalor. The 99 bp synthetic duplex SN541SN55 (described in Table I) carrying the BAR] x2-operator was cloned into the HinclI site of the plasmid pEMBL18+. The cloned 117 bp DNA fragment was released by PstI-BamHJ.restriction digest, 3-end labelled with 32 P-dGTP, isolated and used as the DNA-binding probe for DNase I footprinting experiments. 3 ng of labelled DNA probe was used in each footprinting assay. Additions of the isolated x2 protein, the MCM1 -containing extract and/or the competitor are indicated on top of the autoradiogram by a ‘+“ sign (“-“ indicates exclusion of the component in the assay). The competitor was the unlabelled 117 bp PstJ-BarnHf DNA fragment. The BAR] operator sequence on the DNA probe is shown, with the cL2 recognition half-sites boxed in and the MCM1-binding site indicated by the solid black bar.  123  Figure 26. Analysis of (x2-MCM1 Binding to the pseudo BAR] operator. Top panel: The results of the DNA-binding band-shift assays on the pseudo BAR] operator are shown. A CYCJ promoter fragment carrying either the wild-type BAR) cL2-operator (W.T. BAR 1) or the pseudo BAR] operator (y BAR!) was used as the DNA-binding probe in each lane. The caption below also shows the composition of each lane. The ‘+ and ‘-‘ indicate the presence or absence of the x2 protein or the MCM1containing extract. The ‘++‘ sign means that twice as much yeast extract was added to the reaction as compared to that marked with a ‘+‘. The bands in the autoradiogram corresponding to the free DNA probe, the MCM1-DNA complex and the i2-MCM1-DNA ternary complex are indicated by lines and arrows. Bottom panel: The sequence of the pseudo BAR] (y x2) operator is compared with the BAR] ct2-operator and the MATcd a1-c2-operator sequences which are responsible for mating-type control. The c2 recognition half-sites are noted with the shadowed boxes and the nucleotides are numbered according to Wolberger et a!. (1991). The proposed c2 recognition site in the pseudo BAR] (w c2) operator is marked with clear boxes and also numbered correspondingly. The nucleotides different from the wild-type BAR] sequence are indicated with small letters. The central MCM1-bmding sites of both the wild type and the pseudo BAR] operators are shown with a solid black bar.  124  1  2  4  3  5  6  7  8  9  10  !I!LE::;: 1 a2-MCM1-DNA  a2-MCM1-DNA  DNA  +  -  MCM1-  +  -  ++++-  -  W..T. BAR1  MATcd al-a2  -  H  +  +  -  ++++÷+-  BAR1  IAATGT AGAP1AAGT ACAT C T T AC AT CTT T T CAT GT AG  pseudoBARl  TACAT  ‘V  TGGCTTTTTCCTT 10 9’  1  BAR1  2  3  4  5  6  7  8  9  8  7’  6’  5’  4’  3’  2’  1  10  ICAT6TAAT11ACCGAAAAAGGAATTT ACAT Gj GTACATTAATGGCTTTTTCCTTTAAATGTAC  Figure 26. Analysis of cL2-MCM1 Binding to the pseudo BAR] operator.  125  In the first part of this system, high expression of the MA Tc2 gene was achieved by expressing M4 TcL2 under the ADHI (Alcohol Dehydrogenase 1) promoter in the yeast expression plasmid YHE36 (see Materials and Methods, Figure 12). The ADHJ promoter was chosen because of its constitutiveness and high promoter strength (Ammerer, 1983). In addition, this yeast expression plasmid has a 2.t-replication origin which allows a high copy number, thus also increasing the ADH]:IvL4Tci2 gene copy in vivo. As shown by the primer-extension analyses, the zlmat strain which harbours the ADH]:MA TcL2 expression vectors, YHE36, gave higher transcription of the ]vL4 Tci2 gene than did the native MA Tc strain (nearly 10-fold higher; Figure 27).  The in vivo c2 functions resulting from the ADHJ:IvIATo2 expression were also analyzed by determining the mating and sporulation efficiencies, and by measuring repression on the LacZ reporters. The mating efficiency was determined for the matciX7S transformants (which lack the endogenous M4 Tci2 gene, Table II) habouring the ADH1:MATc2 expression vectors (Table VII, number 5).  The mating  efficiency for similar transformants harbouring MA Tci2 control was also determined (Table VII, number 2) using the YEP-MATc plasmids. The results of this physiological test indicated comparable in vivo c2 function was obtained from the ADH]:MA TcL2 expression. Moreover, in agreement with the results of these mating tests, the sporulation efficiency caused by the ADHJ:MA Tci.2 expression was shown to be similar to that by the MAToc2 control (Table VII, number 4, and 2). The c2 repression on the STE6:LacZ reporter was measured in the AMA T strain, and al-cL2 repression on the HO:LacZ reporter was measured in the MA Ta strain (Table VIII, line 3), After transforming these host strains with the expression vectors which carried either the ADH1:IvL4 TcL2 or the native MA Tc2, the 3-galactosidase activities were measured. The decreases in the enzymatic activities due to the presence of the ADH]:MA Tci2 gene were comparable to that due to the presence of the native MA TcL2 gene (Table VIII line 3, and 2), indicating that the ADH1:MA TcL2 gene exhibited a normal level of both c42- and a1-c2-repression on the LacZ reporters.  In the second part of the in vivo reporter system, a series of CYCI: URA3 reporter fusions was constructed using plasmid YHE36 (Figure 12).  The URA3 gene, a non-cell-type specific gene which  126  Table VII. Physiological Functions of the ADHJ:MA Ta2  NO.  M4 Ta2 alleles a  Mating Efficiency b  Spomlation Efficiency  (%)  (%)  0  0  1  YEP13  2  YEP-MATh  100  10  3  YEP-mato2-1  0  3  4  ADH1:]vL4Tcj2 (21i)  90  8  C  a matcL2-deficient strains were transformed with the high copy 2j.t-YEP13 plasmid (Broach et at., 1979) or its derivatives carrying different MA TcL2 alleles. The YEP-MATcL plasmid is the YEP 13 derivative with the entire IvL4 Tc fragment inserted. The YEP-matc2-1 plasmid (Porter, 1987) is the derivative containing an ochre mutant of cx2 which serves as a non-functional mata2 control here. YFIE36 (Figure 12) is the 2u yeast plasmid carrying the ADH].IvL4 TcL2 fusion gene (number 4). b  The different MA Tct2 alleles were transformed into the SP2 matoi.2-deficient strain and tested for their mating efficiency by crossing with a tester strain D311-3A (S 17). The mating tests were performed in duplicate.  C  The diploid matx2-deficient strain, K204 (Sil), is also deficient in sporulation. This strain was transformed with each of the plasmids carrying the different MA TcL2 alleles. The sporulation efficiencies of these transformants were determined in dupicate.  127  Table Vifi. In vivo Repressor Functions of the ADHJ:MATa2 Fusion  NO.  MATo2 alleles a  zI!vL4T  M4Ta  ste6:LacZ b  Ho:LacZ C  1  YEP13  9.0±(0.1)  12.6±(0.2)  2  YEP-MATcL  3.5±(0.5)  0.43 ±(0.3)  3  ADH1:MATx2 (2i)  2.0 ± (0.6)  0.36 ± (0.02)  a  The YHE36 plasmid (Figure 12) containing the ADH1 :]vL4 TcL2 fusion gene was used to test for c42 and al-(x2 repression. The YEP13 plasmid (Broach et at., 1979) and its IvL4 T(x—contaimng derivative, the YEP-MATc plasmid, were also used in these tests as the negative and positive controls, respectively.  b  The clwomosomal ste6:LacZ fusion was used as the a-specific reporter gene. The -galactosidase activities of the /Jmatcc, ste6.LacZ (SP2) strain harbouring the different MA Tcw.2 alleles were measured. The mean value for each of the three transformants is reported and the standard deviation is included in parenthesis.  C  The chromosomal HO:LacZ fusion was used as the haploid-specific reporter gene. The f3-galactosidase activities of the MA Ta, HO:LacZ (S85) strain harbouring the different MA Tc42 alleles are shown.  128  -a  > -I 4  > “1 N  4  primer  Figure 27. Primer Extension Analysis of theADifi MA Ta2 Expression. P-labelled at the 5’ end. Total yeast RNAs The MATcL2-speciflc extension primer CYH 18 (Table I) was 32 from three different yeast strains were mixed individually with the extension primer, annealed, and extended with reverse transcriptase as described in the Materials and Methods section. The extended products were analyzed by denaturing 6% polyacrylamide gel electrophoresis. The results for the iJinat strains (SP2, Table II) which habour the PVT- 100-L control plasmid or the ADHJ :M4 TCL2 expression plasmid (YHE36) are shown in lanes 1 and 2, respectively. The result for the M4 To’. strain (S25, Table II) which produces endogenous MA To’.2 transcripts is displayed in lane 3. The extension-product bands resulting from the ADH1:M4Tx2 or the MATcL2 transcripts are indicated by the arrows.  129 encodes the OMP-decarboxylase (Rose et at., 1984), was made into a reporter gene for the mating-type control. The promoter of this URA3 gene was deleted, then the CYC] promoter fragments carrying different operator sequences for mating-type control were placed upstream of the gene URA3 (see Materials and Methods, Figure 13).  These different operator sequences included the pseudo BAR], the c2- and the  al-cL2-operators, and the resulting CYC]. URA3 reporter genes were named according to the operator each of them carried: ie. the CYC](ycL2): URA3, CYCJ(cw2): URA3 and CYC1(al-cL2): UR43 genes, respectively (Table IX, Figure 28). The x2 repressor functions necessary for the mating-type control were provided by the ADH1:M4TcL2 gene in the same plasmid. The mating-type control on this series of CYC]: URA3 genes was studied by primer-extension analyses as described below.  In order to measure the cx2 repression on these CYC]: UR43 reporter genes, the plasmids carrying them were transformed into the zimat strain (SP2, Table II). These plasmids were also transformed into the MA Ta strain (S85, Table II) for the measurement of al-cc2 repression. The total yeast RNA was isolated from each transformant. A 32 P-labeled DNA primer, which hybridized specifically to the URA3 transcripts (described in Table I) was used to detect the different CYC1: URA3 transcript levels in these transformants. The DNA primer was annealed to the total RNA and then was extended towards the 5’ end of the CYC]: URA3 transcripts by reverse transcriptase.  Although both the zimat and 1v14 Ta host strains gave  endogenous URA3 transcripts from the ura3-52 allele (which has a transposon element inserted in the URA3 coding region; Rose and Winston, 1984), the primer-extension products of the CYC]: URA3 transcripts were readily distinguishable from those of the endogenous ura3-52 ones, by their longer lengths Figure 28 shows the CYC]: URA3 dependent extension products in lane 3 but not in lane 2). The results of these primerextension analyses were summarized qualitatively (Table IX).  The wild-type a1-c2- and cs2-operators  conferred the expected mating-type control on the CYCI: URA3 reporters (Table IX, numbers 4 and 6). That is, transcription of the CYCJ (al-cL2): URA3 gene in the Amat strain is not repressed but it is repressed in the M4 Ta strain, thus demonstrating the al-dependent repression.  On the other hand, transcription of the  CYC1(ct.2): URA3 gene was repressed in both the 4mat and M4Ta host strains, indicating that c42 expression  130  Table IX. The Mating-Type Control of the CYCJ:URA3 Reporters  NO.  URA3 reporter fusions a  1  control  2  Ap-ura3  3  CYC1: URA3  4  CYC1(aI-ci2):URA3  5  CYC](i.ycL2):URA3  6  CYC](ci2):URA3  SP2 b  S85  zlmat  M4Ta  -  -  -  -  +++  -  ±  +  -  -  C  The results of the primer extension analyses on the CYC1: URA 3 reporters are summarized. A qualitative representation of the results is shown using these notations: undetectable extension level; ‘±‘, marginally detectable; detectable; ‘+++‘, maximal extension level. See Figure 28 for actual autoradiograms. ‘-‘,  ‘+‘,  a  Total yeast RNA samples in the primer extension experiments are derived from the transformants harbouring the URA3 reporters listed in this table. YHE36 (Figurel2 and 13) was used as the promoterless ura3 reporter plasmid (line number 2). Its derivatives, YBE41,38,39 and 40 are the plasmids (Figures 11 and 13) which carry the different CYC]: URA3 reporters listed from line number 3 to 6. The ADH]:MA TcL2 fusion gene is also present on these plasmids to express x2 for the repression studies. The ‘control’ is the parental strain without plasmid transformation.  b  The x2 repression on each URA3 reporter gene was tested using the matx2-deficient strain, SP2.  C  The al-cc2 repression on these URA3 reporter genes was also tested using the MA Ta strain, S85.  131 Figure 28. The c2 and al-x2 Repressions on the CYCI: URA3 Reporter Genes. Primer extension analysis of the CYC]: URA3 reporters was carried out using the zimat (SP2, Table II) and the MA Ta (S85, Table II) strains. The total RNAs were prepared from the transformants of these host P-labelled URA3-specffic primer strains. Equal amounts of the total RNAs were reverse-transcribed using 32 (CYH26, Table I) or the labelled LEU2-speciflc primer (Table I).  Top Panel: The results for the URA 3-specific transcripts are shown in the top panels. The different UR.43 reporters carried by the zlmat or the IvL4Ta transformants in this experiment are identified as follows: lane 1 has the untransformed parental strain as the ‘no URA3 gene” control; lane 2, the promoterless ura3 gene (YHE36, Figurel2); lane 3, the CYC]: UR.43 reporter (YHE41); lane 4, the CYC](al-cx2): URA3 reporter (YHE3 8); lane5, the CYC1 (y cL2): URA3 (YNE39) reporter; Lane 6, the CYCI (cL2): URA3 reporter (YHE4O). The CYC]: URA3 fusion transcript products are marked according to McNeil et al. (1986) and indicated by double-headed arrows. The bands on the bottom of the autoradiograms caused by the endogenous ura3-52 transcripts are not marked. Bottom Panel: The results of the LEU2-specific primer extension assays are displayed in the bottom panel. Equal amounts of the same total RNAs were analyzed as described above and shown in the corresponding lanes.  132  mat  MATa  123456  123456 —  -69  •—-61 •  -46  t  —  •—-38 -  ——  ..  123456  123456  —  Amat  MATa  Figure 28. The x2 and al-cc2 Repressions on the CYCJ: URA3 Reporter Genes.  133 from the ADH]:MA TcL2 gene is sufficient for cL2 repression. These observations supported the validity of this in vivo reporter system in reflecting the mating-type control.  Consistent with the findings of the in vitro DNA-binding assays (described earlier), the primerextension product of the CYC] (cx2): URA3 gene in the zlmat strain was barely detectable (Table IX, number 5; also see Figure 28 left panel, lane 5), while that of the CYC1(c2): URA3 gene was totally undetectable (Figure 28, left panel lane 6). This result indicated that the pseudo BAR] operator mediated weaker but significant (x2-repression when compared with the wild-type BAR] c2-operator. Interestingly, an increase of primer-extension products in the M4 Ta strain was detected for the CYC] (c2): URA3 gene but not for the CYC] (cs2): URA3  gene,  suggesting that the  presence  of al  in  this  strain  derepressed  the  cL2-repression mediated by the pseudo BAR] operator but not that mediated by the wild-type BAR] x2-operator (Table IX, number 5 and 6; also see Figure 27, top right panel, lanes 5 and 6). The possible molecular mechanisms for this derepression will be discussed later.  4.2.  DISCUSSION OF z2-MCM1 COOPERATWE DNA-BINDING  The pseudo BAR] operator, which contains two al-oL2 recognition half-sites flanking the MCM1binding site, was characterized by DNA-binding gel-shift assays, and its in vivo function in mating-type control was determined using the CYC]: URA3 reporters. This variant operator differs from the wild-type BAR] oL2-operator in the regions beyond the presumed MCM1-recognition site (CCGAAAAGG) and the core-consensus of the flanking cL2-recogmtion half-sites (TGT. .ACA; Figure 26). .  If the c2 operator  function requires only the presumed MCM1-recogrntion site and the consensus core x2-binding sequence which has been shown to base-contact with the cQ-homeodomain (Wolberger et al., 1991), then this pseudo BAR] operator should function just as well as the wild-type BAR] x2-operator. However, the differences between the pseudo BAR] and the wild-type operators caused significant decreases both in the in vitro DNA-binding by x2-MCM1 and in the in vivo x2-repression (Figure 26 and Table IX). These observations  134 suggested that the DNA-sequence outside the core consensus sequences of the o2- and MCMI-binding sites are required for optimal c2-operator function.  The in vitro DNA-binding band-shift assays (Figure 26) showed that the pseudo BAR] operator formed an x2-MCM1-DNA ternary complex, although at a lower affinity than that of the wild-type BAR] operator. The pseudo BAR] operator was unable to compete out x2-MCM1 binding to the BAR] operator. DNA fragments carrying different operator sequences, namely the (x2-, a1-c2-, and pseudo BAR] operators, were used as competitors for (x2-MCM1 binding to the BAR] operator. In these studies, only the fragment containing the wild-type BAR] ot2-operator was able to compete off the binding. In addition, the DNaseI protection footprinting assay showed that protection of this BAR] c2-operator sequence is x2-MCM1dependent. Similarly, the pseudo BAR] operator may be recognized and bound in the same (x2-MCMIdependent way.  However, the present in vitro DNA-binding results do not distinguish between cooperative x2MCM1 binding to the pseudo BAR] operator (both proteins binding synergistically to the DNA as in the wild-type operator case) and c42-enhanced MCM1-binding (cL2 not bound to DNA). In the former case, the ct.2 proteins bind to the separated a1-c2 half-sites on the pseudo BAR] operator, while in the latter case, the c2 proteins simply enhance MCM1-binding to the DNA via protein-protein interactions.  This type of  enhanced MCM 1-DNA binding has been proposed for the human homeobox protein, Phoxi (paired like human homeobox 1; Grueneberg et at., 1992) which has a binding site identical to the MCM1-site. The Phoxi protein binds to this site and enhances MCM1-DNA binding by catalyzing the formation of the MCM1-DNA complexes (Phoxl itself dissociates from the DNA and is replaced by MCM1). However, in cQ-MCM1 binding, cr2 binds to its own recognition site stays bound as a component in the DNA-protein complex despite the fact that MCM1 is also bound to the flanking DNA-site. The presence of the core cr2-recognition sequence (TGT. .ACA) in the pseudo BAR] operator suggests that cr2 is likely bound to the . .  DNA and thus there is no involvement of Phoxi-lilce enhancement of MCM1-binding. Although there are “A to T” and “T to A” changes in the 8 and 8’ position respectively of the cr2-half-sites of the pseudo BAR] operator, these replacements may not cause any decrease in the cr2-binding.  As shown by the X-ray  135 crystallography-model, the Arg7 residue of the c2-homeodomain’s N-terminal arm contacts the T-base in the 8 or 8’ positions via the minor groove. It is also believed that in the minor groove, this Arg7 residue does not discriminate between an A or T residue in these positions (Wolberger et al., 1991).  On the other hand, the invariant 9 position of the c2 consensus half-site (Figure 4 and 26), normally a T residue, is replaced by an A residue in the pseudo BAR] operator. The complementary T residue changed to A in this invariant position is presumably contacted by Arg7 of the c.2-homeodomain given the fact that Arg7 normally contacts this position prior to the change. Also, since Arg7 of c2 is unable to discriminate between an A or T residue via a minor groove contact, and the c2-protected footprinting sequences overlap with the MCM 1-protection, it is believed that MCM1 may be the protein recognizing this residue and thus determining the preference for the T residue in this position (Wolberger et a!., 1991). Therefore, the replacement of T with an “A” residue in the invariant 9 position may cause a decrease in the MCM1 binding. Consequently, this can lead to the decrease observed in x2-MCM1 cooperative binding and the in viva x2-repression mediated by the pseudo BAR] operator.  Although disuffide-linked cL2 homodimers which bind to the DNA with higher affinity than the monomers, were detected in the isolated n2 protein, the binding of x2-MCM1 to the pseudo BAR] target was not likely an artifact caused by the presence of these dimers. The isolated c2 proteins alone were incapable of binding to the pseudo BAR] operator, implying that MCM1’s participation was the major determining factor in the binding. Also, the cL2 proteins in the E. coil lysate are not disulfide-linked before isolation, suggesting that the disullide-linkage may not exist in vivo. Despite the presence of ci2 disulfide linked homodimers, the fact remains that the pseudo BAR] operator, in agreement with the in vifro DNAbinding data, mediates the partial x2-repression in vivo.  As illustrated by the primer-extension analyses of the CYCJ(\ycL2): URA3 reporter, the pseudo BAR] operator conferred partial  x2-repression in vivo.  Interestingly, the repression on this  CYC1(cv2): URA3 reporter was significantly less in the MATa cells than in the zimat cells and this derepression was not detected for the similar reporter CYC](cL2): URA3 which carries the wild-type BAR]  136 x2-operator. Although there are many possible explanations which can account for this observation, the derepression in the MA Ta cells can readily be explained in the following two ways:  In the first possibility, the a1-c2 heterodimers in MA Ta cells may not cause any al-cc2 repression on the CYC] (cl2): URA3 reporter because the dimers do not recognize the split al-cQ site in the pseudo BAR] operator. In addition, the fonnation of these al-x2 dimers depletes the cL2 molecules available for repressing the reporter. Thus, in effect, this lowers the intracellular c2protein concentration and leads to the observed derepression.  However, derepression was not  detected for the CYCJ(ci2): URA3 reporter in !vL4 Ta cells even though the effective c42 concentration remains the same. These observations indicate that the pseudo BAR] operator may confer a weaker x2-repression than would the wild-type operator. Therefore, this correlates with the lower in vitro cL2-MCM1-binding affinity to the pseudo BAR] operator which was previously observed.  Alternatively, the a1-c2 dimers may actually bind to the split target site in the pseudo BAR] operator but may still fail to conduct normal repression. Therefore, the al-u2 dimer would compete with c42-MCM1 for DNA-binding and cause the derepression of the CYC] (a2): URA3 reporter.  Furthermore, the al-x2 dimer does not affect the repression of the CYCJ (c2): URA3  reporter because it fails to recognize the wild-type oc2-operator and thus, does not compete out x2MCM1 repression. Another possibility is that the al-cc2 dimer does bind the wild-type x2-operator, but because c2-MCMl binds with a higher affinity to it than to the pseudo BAR] operator, the al-cL2 dimer fails to compete for DNA-binding in the wild-type case. Therefore, the normal cc2-repression mediated by the wild-type ft2-operator is not derepressed.  Finally, it is of note that a remarkable sensitivity to subtle changes in the c2 repressor functions can be demonstrated and directly measured by determining the transcriptional level of the CYC1: URA3 reporters. Particularly, the partial x2-repression detected in the present study was previously undetectable using the enzymatic assays of the LacZ reporter genes (S. S. Ner and M. Smith, unpublished data), where the CYC1:LacZ reporters carrying a similar set of operators were studied and where the c2 repressors were  137 expressed from the native single-copy M4 Tcs2 gene.  Although the increased M4 Tcs2 transcript level  resulting from the ADH]:M4Tcw2 gene in the present reporter system may contribute to the sensitivity in detecting the partial cs2 repression (mediated by the pseudo BAR] operator), caution should be taken in evaluating how much the increase in the MA TcL2 transcript level actually contributes to such sensitivity. This increase in transcript level was not reflected in the physiological functions tested, namely the mating and sporulation efficiencies. Conceivably, any possible effects to these physiological functions from the increase in transcript level may be offset by the rapid turnover rate of the x2 repressor proteins in vivo (Hochstrasser and Varshavsky, 1990).  4.3.  CONCLUSIONS AND FUTURE PROSPECT1VES  The present studies on the pseudo BAR] operator indicate that x2 and MCM1 bind cooperatively to this variant operator in vitro. The binding affinity is, however, less than that to the wild-type x2-operator from the BAR] gene. Furthermore, the pseudo BAR] operator mediates a partial x2 repression in vivo, indicating that the repressor function correlates with the binding affinity. The observations from the work done in this thesis is consistent with the notion that transcriptional regulation involving homeodomain proteins requires both a combinatorial mechanism and a competition mechanism (Hoey and Levine, 1988; Han eta!., 1989).  In the combinatorial mechanism, each homeodomain protein recognizes a qualitatively different DNA regulatory sequence. Whether a particular target gene is ‘on’ or ‘oil’ depends on what combination of homeodomain proteins is bound to the cis-element on that gene.  On the other extreme of this is the  competition mechanism where all homeodomain proteins bind to the same limited set of DNA-recognition sequences but each of them binds with different affinities. In this case, therefore, different homeodomain proteins “compete” for binding to a specific cis-element. Whether it is ‘on’ or ‘oil’ for a target gene depends on which homeodomain protein has the highest affinity as well as on the relative intracellular concentrations of the different proteins. In the case of cL2 repression mediated by the variant operator  138 (pseudo BAR]), MCM1 cooperates with x2 and modulates its DNA-binding affinity to this operator (combinatorial mechanism). In turn, this change in binding affinity influences the regulatory effect on the variant pseudo BAR] operator.  When both a1-c2 and x2-MCM1 are present, x2-MCM1 may need to  compete with the al-(x2 heterodimer for binding to the variant operator so that x2-MCM1 can carry out (x2-repression (competition mechanism).  Furthermore, how MCM1 modulates the c2 DNA-binding specificity can also be studied using a similar approach to that reported in this thesis.  One can generate mutant operators which carry an  unchanged MCM1 site and randomly varied mutant x2-binding sites. Using the E. coil-expressed cL2 and the crude MCM 1-containing yeast extract, the mutant operators which are still recognized by c2-MCM1 can be selected by the DNA-binding plate assay developed by Lorimer et al. (1992). B. coil transformants expressing the cL2 proteins can be lysed, and the x2 proteins immobilized onto a nitrocellulose filter. The MCM1-containing yeast extract and the pool of random mutant-operator DNA are applied to the immobilized (x2 proteins. Mutant operators which are not recognized by both MCM1 and ct2 would be washed ofl while the functional mutant operators are immobilized by binding to the x2-MCM1 on the filter, and are thus retained. After eluting with high salt buffer, these operator-DNA fragments can be cloned and analyzed by DNA sequencing. Their in vivo function can also be tested by determining their abilities to confer mating-type control on the CYCI(x2): URA3 reporter. Such experiments may offer further insights into the modulation of homeodomain DNA-binding specificity by the MADS proteins.  139  APPENDIX A  5.1.  5.1.1.  IN VITRO a1-2 DNA-BINDING OF MISSENSE SUBSTITUTED a.2 MUTANTS  Missense Mutations in the ct2B Coiled-Coil Dimerization Motif  Three missense mutants of cL2, namely 167T, T83S, and L196S, have been tested for their in vitro a1-c2 DNA-binding activities. Mutants 167T and T83S were originally created for a random mutagenesis experiment involving the introduction of a Spel restriction site at codons 67 and 83. Coincidentally, the changes in the DNA sequence also gave rise to serine replacements in the d and f positions of the putative coiled-coil dimenzation motif (cL2B, see Figure 6). Since extensive alamne replacements at the hydrophobic d positions of the c2B heptad repeat caused significant decreases in both a1-c2 interaction and DNAbinding by al-0L2 (see Results and Discussion, Table VI), the mutational effects of 167T and T83 S on a1-c2 dimerization were also characterized by the al-c2 DNA-binding assay.  The mutant c2 proteins were  individually co-translated with al and radioactively labelled in vitro. The protein mixture was combined with the al-cQ target-DNA fragments and binding was tested in the gel mobility-shift assays. The results suggest that the respective threomne and serine replacements of both 11e67 (d-position) and of Thr83 (Jposition) are tolerated for al-x2 dimerization since these changes did not affect the DNA binding (Figure 29). Although conservative threonine to serine change in thefposition is expected to be tolerated, replacing the d-residues with less hydrophobic or polar amino acids is usually detrimental. Nevertheless, the observed tolerance for the Thr67 replacement at the d position is consistent with the early finding that a single alamne replacement of 11e67 did not affect the dimerization function and suggests that 11e67 is not an important part of the putative coiled-coil motif.  140  1234567  1234567 I  I_. 43.0 —  oc2 N> al  — —  —  29.0 184  —14.3  Figure 29. Effect of the c2 Amino Acid Replacements on al-ot2 DNA-Binding. The replacement mutants of c2, 167S, T83S and L196S, were co-translated and radioactively-labelled with the al protein. The a1-c2 DNA-binding ability of these mutants was tested by gel-mobility-shift assays. The results are shown in the top panel. For comparison, the results for the (x2B mutants, L81A and L81R, are also included. The different (12 proteins used in lanes 1 to 7 are as follows: (1) L8 1A, (2) 16 iT, (3) L81R, (4) T83S, (5) wild-type a2, (6) a “no-protein’ control, and (7) the L196S mutant. In the bottom panel, the fluorograph of a 17.5% SD S-PAGE contains one tenth the amount of the same co-translated products used in the binding assays. The relative molecular weight standards are indicated in kiloDaltons.  141 5.1.2.  Serine Replacement of Leu196 of a2  The DNA-binding activity of the L196S mutant was also investigated. This a1-c2 repressiondefective mutant was first isolated by classical genetics (Strathern et a!., 1988). Since the L196S mutation does not affect c2 repression, it is believed that it affects a1-x2 DNA-binding but not c2 DNA-binding. In order to test this premise, this mutation was constructed by oligonucleotide-directed site-specific mutagenesis using the pEMBL-MATc-(BamHI-BclJ) plasmid and the L196S oligonucleotide (Materials and Methods). In vitro co-translation and al-c2 DNA-binding mobility-shift assays were carried out.  The  results show that the L196S mutation caused a moderate decrease in DNA-binding by the a1-c2 dimer (Figure 29), therefore lending support to Strathen and colleagues’ assumption about the effects of the L196S mutant on DNA binding. Moreover, this observation is consistent with Mak and Johnson’s (1993) recent finding that deletion of residues 188-210 of c2 reduces al-cc2 DNA-binding but not x2-MCM1 DNAbinding.  There are at least two possible causes for the decreased al-cL2 DNA-binding of the L196S mutatnt protein. First, the leucine residue at position 196 may be responsible for optimal protein-DNA interaction. For example, replacement of this leucine residue may destroy the essential structure which makes direct contacts to the DNA. Secondly, Leu 196 may be required for optimal stability of the a1-x2 heterodimers. Although the residues from position 140 to the C-terminal was shown to be dispensable for the negative dominant phenotype in aJc diploids, thus indicating that this region is not essential for al-cL2 heterodimerization (Hall and Johnson, 1987), NIvIR studies of the interaction between the al and ct2 homeodomains suggest that the “tail” (residues 196-203) preceding the c2 homeodomain may still interact significantly with the al homeodomain and therefore further stabilize the al-cL2 heterodimers (Phillips, 1992). Consistent with this, Mak and Johnson (1993) showed that deletion of the “tail” region (residues 188-210) destabilizes the al-c42-DNA complex but does not disturb the DNA footprinting pattern, thus indicating that residue 196 is not essential for making direct base contacts with the DNA.  142  However, L196S mutation is less detrimental than the alanine replacement mutant x2B(L81A) and the argimne replacement mutant c2B(L8 1R) when judged by the in vitro DNA-binding mobility-shift assays (Figure 29, lanes 1, 3 and 7). These results suggest that the “tail” region may contribute less to the al-x2 dimer’s stability when compared with the x2B putative coiled-coil region. This idea can be tested by co-immunoprecipitation assays by determining if L196S brings down more al protein than do the cL2B(L81R) or the c2B(L81R)fL196S mutants.  143  APPENDIX B  5.2.  BARJ:URA3 REPORTER SYSTEM FOR SELECTING (12 FUNCTION  To develop an efficient system which allows in vivo selection of the x2-repressor function, particularly (x2-repression, a reporter system was constructed using a fusion reporter which is comprised of the promoter of the a-specific gene, BAR], and the coding region of the URA3 gene. This reporter gene was used because the absence of URA3 function is a selectable trait.  URA3 encodes the OMP decarboxylase  (Bach et at., 1979) and is involved in the UMP biosynthesis pathway (Bach et at., 1979; and see Figure 30 for a schematic presentation). When the orotate analog, 5-fluoroorotic acid (5FOA), is supplied, normal URA3 expression leads to the production of the toxic 5F-UIvIP which in turn affects both the syntheses of pyrimidine deoxyribonucleotides and ribonucleotides. One of the known toxic metabolites of 5F-UMP, SF dUMP, inhibits thymidylate synthetase, a key enzyme for DNA synthesis, and causes cell death (Pagolotti and Santi, 1977). However as long as uracil is supplied to bypass this UMP biosynthesis pathway, a ura3 null mutation allows growth on 5FOA medium by blocking the toxic action of 5FOA which relies on this pathway. Therefore, when BAR]: URA3 is put under tight c2 repression, 5FOA resistance should result, and thus permiting a selection of x2-repression using the 5FOA. When BAR]: URA3 is repressed by cL2, OMP decarboxylase is not expressed, 5F-UIvIP is not produced from 5FOA and cannot inhibit the syntheses of pyrimidine deoxyribonucleotides and ribonucleotides, thus allowing normal DNA or RNA synthesis and the cell to grow as a selectable 5FOA-resistant colony.  The feasibility of using this BAR]: URA3 fusion reporter system to select for x2-repression has been tested using an x2-deficient yeast strain, S15 matc,X75, (which has a small deletion at the transcriptional termination site) (see Table II for geneotype; Tatchell et at., 1981).  This S15 matcwX75  recipient strain was co-transformed with a low-copy LEU2-plasmid carrying the BAR]: URA3 gene (plasmid YHE65, Figure 13) and a high-copy plasmid carrying different MATa2 alleles and the TRP] selectable marker.  The transformant habouring both plasmids, and thus complemented for the leu2 and tip]  144 2 CO  orote  o  I  >  I  d  OMP decafboxylase  5FOA Resistance URA3 5FOA  I 5F  I  5F d4P  NO  [  ura3  :4>.  5FOA  (tJracil Supplement)  ).i(  YES  BARL:URA3 SFOA I  IIIIIIr.. BAR1 :URA3  5FOA  —  NO  XI -__  (Uracil Supplement)  a2  YES  Repression  Figure 30. 5FOA Selection of (x2 Repression. The three panels show the rationale of 5FOA selection for c2 repression. In the top panel, OMP is synthesized from orotate. The URA3 gene product, OMP decarboxylase, then catalyzes the conversion of OMP to UMP, thus allowing production of dUMP. In the middle panel, the orotate analog, 5FOA, if supplied, will be incorporated into OMP. Decarboxylation of this product leads to generation of the toxic 5F-UMP which affects both the syntheses of pyrimidine deoxyribonucleotides and ribonucleotides. One of the known toxic metabolites of 5F-UMP, 5F-dUMP (shadowed boxes), inhibits a key enzyme for DNA synthesis and causes cell death. When the decarboxylation reaction is blocked (indicated by the cross) due to a ura3 null mutation, production of the toxic 5F-dUMP as well as other toxic metabolites is inhibited. Thus, this allows cell growth on 5FOA medium with uracil supplement. In order to maintain cell growth, uracil supplementation is required to bypass these reactions and provide UMP for DNA and RNA synthesis. This 5FOA resistance can also be acquired if the expression of the URA3 gene is under strong repression (illustrated in the bottom panel). c2 repression on the BAR]: URA3 fusion gene can efficiently block 5F—UIvIP production, thus conferring 5FOA resistance. This provides a means for direct selection of (x2 repression on 5FOA medium if uracil is supplemented.  145 mutations, was first selected by its capability to grow without leucine and tryptophan supplement and with uracil. Its resistance to 5FOA was then determined by measuring the growth rate in liquid 5FOA medium The results indicated that x2-repression on the BAR]: URA3  or the extent of growth on 5FOA plates.  reporter confers 5FOA resistance (Figure 31). This 5FOA resistance depends on the expression of wild-type MATcL2 and an intact BAR] promoter, 5FOA inhibited the growth of transformants expressing the non functional mata2 alleles, namely the ochre mutant matci2-] or the deletion mutant M-64 matc42, and of those carrying the BAR] *: URA3 gene (plasmid YHE64) in which the critical core c2-binding half sites “TGT  ACA” has been changed to “TCT. .AGA” (Figure 31). . .  Although transformants harbouring the BAR]: URA3 reporter genes exhibited 5FOA resistance which depended on cL2-repression, application of this reporter system for selecting cL2 repression  was  hindered by a technical problem. That is, after transformation with the BAR]: URA3 reporter system, a lag period is required for developing the c2 repression, therefore, it is impossible to do transformation and selection on 5FOA medium at the same time. Also, a high transformation efficiency is usually essential for a selection system because a large number of transformants is normally needed in order to examine a random pool for cc2 repression or some other phenotype. However, the highly efficient “spheroplast transformation” procedure requires regeneration of normal cells from transformed spheroplasts in agar. Thus, plate replicates cannot be made using this procedure while some other less efficient transformation protocols would allow this (e.g. transformation using lithium ions).  To overcome such technical problems, a simple but crucial modification was introduced to modify the conventional spheroplast transformation procedure. After a spheroplast transformation, transformants containing the BAR]: URA3 reporter gene are allowed to regenerate in top agar which is poured in molten form on a nitrocellulose filter sitting on top of agar medium that does not contain 5FOA. After 24 to 36 hours of regeneration, this top agar is lifted and transferred onto a 5FOA-contaimng agar medium for selection.  This modified spheroplast transformation method was used to select for 5FOA-resistant  transformants which had been co-transformed with plasmid YHE65 (carrying the BAR].’ URA3 reporter) and a mixed population of the test plasmids, namely the MA Tx2-contaiing plasmid (YRP-MATo.2) and the  146  Figure 31. The Growth Curves of the BAR 1: URA3 Reporter Strains on 5FOA Medium. Graphs (a) to (d) indicate the BAR] and M4 TcL2 dependent 5FOA resistance. The matx2-deficient host yeast strain, matc,X75 (Table II), which is also ura3, 1eu2 and trp], was co-transformed with different BAR]: URA3 reporter plasmids and the MATci2 expression plasmids. Two LEU2-CEN plasmids,YHE65 and YFIE64, canying the BAR]: URA3 and the variant BAR] *: UR.43 reporters, respectively, were used in the transformations. The TRP1-ARS plasmids, YRP-MATc2, YRP-matcc2-1, and YRP-matoL2z\4-64, carry and express the wild-type MA T2 gene, the ochre mutant gene, and the deletion mutant gene, respectively. The growth curves of the co-transformants harbouring these BAR]: URA3 reporter plasmids and the MATci2 expression plasmids are shown. The notations are as follows: A dotted line and open circles indicate the growth on SC-TRP-LEU (+TJRA) medium; a solid line and triangles indicate growth on SC-TRP LEU+O.2% 5FOA; a dotted-dashed line and open squares indicate growth on SC-LEU-TRP-URA medium. Graphs (a), (b), and (c) are growth curves of transformants carrying the wild-type BAR]: URA3 reporter in addition to the wild-type MA TcL2 gene, the matcL2-] gene, and the matcL2M-64 gene, respectively. Graph (d) shows the growth curves of the transformant carrying the BAR] *: URA3 variant reporter and the wild type MA Tx2 gene.  147  14  +URA  12  BAR1:URA3 I MATa2  C  . -  10  0  0  •  ‘  o  I• 2  •I— 4  6  8  I 12  10  I  14  16  18  20  22  Growth Time / hr (a)  14-  ÷URA  12  -  C  BAR1:URA3 / mata2-1  10-  -y .7  L) 2  5FOA A—  0 0  2  4  6  8  10  12  14  16  18  20  22  Growth Time I hr (b)  Figure 31. The Growth Curves of the BARJ:URA3 Reporter Strains on 5FOA Medium.  148  18 16  +URA  BAR1:URA3 /matcz2 A4-62  14  -  -  12  o  /  /  -  .-URA  / /  ag-.  0  6  .“ /  —  2 5FOA  A  () 0  2  4  6  8  10  12  Growth  14  16  20  18  22  24  Time / hr  (c)  14  +URA  12  BAR1*:{JRA3 /MATx2  o  \C1o.  -  2  5FOA A  --A A —----1  0• 0  2  4  6  8  10  12  I 14  16  18  20  22  Growth Time / hr (d)  Figure 31. The Growth Curves of the BARI:URA3 Reporter Strains on 5FOA Medium.  149  NO. MATa2/Vector Colonies Ratio on ilate  1 2 3 4 5 6 7 8  2  0:2 2:0 1:1 1:10 1:20 1:50 1:100 0:0  0 800 900 520 280 1 50 1 00 0  3  5  7  4  6  8  Figure 32. 5FOA Selection of cx2 Repression. The rnatc2-deficient yeast strain, rnatciX75 (SiS, see Table II), was co-transformed with the YHE65 plasmid (canying the wild-type BAR]: URA 3 reporter) and a mixed population of the (x2-expressing YRP-MAT(x2 plasmid and the non-expressing YRP7 plasmid (the parental vector of YRP-MATcL2). The co-transformants were allowed to regenerate in SC-LEU-TRP ‘top agar” for 36 hours and then were transferred on top of 0.2% 5FOA-selective medium. The original ratios of YRP-MATx2 to YRP7 plasmids in the different transformations are indicated on the top panel and identified with numbers from 1 to 8. The total number of colonies resulting from each of the eight transformations are reported. Sections from the eight “top agars” in this 5FOA selection experiment are shown.  150 parental YRP7 vector (negative control). The results show that the number of 5FOA-resistant colonies correlated with the 1v14 Tc2/vector ratio, thus indicating the usefulness of this simple top agar lifting procedure (Figure 32). More importantly, this system is applicable to selecting cc2 repression and will facilitate future studies on x2 structure-function relationships.  151  REFERENCES Abate, C., Luk, D., Gentz, R., Ravscher, F.J,, 3rd. and Curran, T. 1990. Proc. Nat!. A cad. Sd. USA 87: 1032-1036. Adamson, J.G., Zhou, N.E., Hodges, R.S. 1993. Curr. Opin. Biotech. 4: 428-437. Affolter, M., Schier, A. and Gehring, W.J. 1990. Curr. Opin. Cell Biol. 2: 485-495. Agre, D., Johnson, P.F. and McKnight, S.L. 1989. Science 246: 922-925. Alber, T. 1992. Curr. Opin. Genes. Dev. 2: 205-210. Alberti, S., Oehler, S., von Wilcken-Bergmann, B., Kramer, H. and Muller-Hill, B. 1991. New Biologist 3: 57-62. Ammerer, G. 1983. Meth. Enzym. 101: 192-201. Ammerer, G. 1990. Genes. Dev. 4: 299-312. Angel, P., Allegretto, E.A., Okino, S.T., Hattori, K., Boyle, W.J., Hunter, T. and Karin, M. 1988. Nature 332: 166-17 1. Assa-Munt, N., Mortishire-Smith, R.J., Aurora, R., Herr, W. and Wright, P.E. 1993. Cell 73: 193-205. Astell, C.R., Ahlstrom-Johnasson, L., Smith, M., Tatchell, K., Nasmyth, K.A. and Hall, B.D. 1981. Cell 27: 15-23. Atkinson,T and Smtih, M. 1984. pp. 82 in M. J. Gait, ed. Oligonucleotide Synthesis: A Pratical 35 Approach. IRL Press Ltd., Oxford. Bach, M., LaCroute, F. and Botstein, D. 1979. Proc. Nat!. Acad. Sci. USA 76: 386-391. Baldari, C. and Cesareni, G. 1985. Gene 35: 27-32. Ballinger, D.G., Xue, N. and Harshman, K.D. 1993. Proc. Nat!. Acad. Sci. USA 90: 1536-1540. Baniahmad, A,, Steiner, C., KOhne, A.C. and Renkawitz, R. 1990. Cell 61: 505-5 14. Banner, D.W., Kokkinidis, M. and Tsernoglou, D. 1987. 1 Mo!. Biol. 196: 657-675.  152  Barrett, G., Horiuchi, M., Paul, M., Pratt, R.E., Nakamura, N. and Dzau, V.J. 1992. Proc. NatI. Acad. Sd. USA 89: 885-889. Berger. S.L., Cress, W.D., Cress, A., Trienzenberg, S.J. and Guarente, L. 1990. Cell 61: 1199-1208. Bender, A. and Sprague, G.F., Jr. 1987. Cell 50: 681-691. Benoist, C. and Chambon, P. 1981. Nature 290: 304-3 10. Biggin, M.D. and Tjian, R. 1989. Cell 58: 433-440. Bohmann, D., Bos, T.J., Admon, A., Nishimura, T., Vogt, P.K. and Tjian, R. 1987. Science 238: 1386-1392. Bolivar, F., Rodriguez, R.L., Greene, P.J., Betlach, M.C., Heyneker, H.L. and Boyer, H.W. 1977. Gene 2: 95-113. Brent, R. and Ptashne, M. 1980. Proc. Nat. A cad. Sd. USA 77: 1932-1936. Brand, A.H. 1985. Cell 41: 41-48. Brand, A.H., Breeden, L., Abraham, J., Sternglanz, R. and Nasmyth, K. 1985. Cell 41: 41-48. Brent, R. 1985. Cell 42: 3-4. Brewer, B.J. and Fangman, W.L. 1987. Cell 51: 463-47 1. Broach, J.R., Strathern, J.N. and Hicks, J.B. 1979. Gene 8: 121-133. Buratowski, S., Hahn, S., Sharp, P.A. and Guarente, L. 1988. Nature 334: 37-42. Burglin,T.R. and De Robertis, E.M. 1987. EMBO J. 6: 26 17-2625. Carison, M., Osmond, B.C., Neigeborn, L. and Bostein, D. 1984. Genetics 107: 19-32. Ceska, TA., Lamers, M., Monaci, P. Nicosia, A. Cortese, R. and Suck, D. 1993. EMBO 1 12: 1805-1810. Chen, P., Johnson, P., Sommer, T., Jentsch, S. and Hochstrasser, M. 1993. Cell 74: 357-369. Chen, W. and Struhi, K. 1985. EMBOJ. 4: 3273-3280.  153  Chiu, R., Angel, P. and Karin, M. 1989. Cell 59: 979-986. Chodosh, L.A., Olesen, J., Hahn, S., Baldwin, A.S., Guarente, L. and Sharp, P.A. 1988. Cell 53: 25-35. Choe, J., Kolodrubetz, D. and Gmnstein, M. 1982. Proc. Nati. Acad. Sci. USA 79: 1484-1487. Clerc, R.G., Corcoran, L.M., LeBowitz, J.H., Baltimore, D. and Sharp, P.A. 1988. Genes Dev. 2: 1570-1581. Cohen, C. and Pany,D.A.D. 1986. Trends Biochem. Sci. 11:245-248. Coornaert, D., Vissers, S., Andre, B. and Grenson, M. 1992. Curr, Genet. 1992. 21: 301-307. Coulombe, P.A. 1993. Curr. Opin. Cell. Biol. 5: 17-29. Cress, W.D. and Triezenberg, S.J. 1991. Science 251: 87-90. Crick, F.H.C. 1953. Acta Cryst. 6: 689-697. Cunningham, B.C. and Wells, J.A. 1989. Science 244: 1081-1095 Cusack, S., Berthet-Colominas, C., Härtlein, M., Nassar, N. and Lebennan, R. 1990. Nature 347: 249-255. Dagert, M. and Ehrlic, S.D. 1979. Gene 6: 23-28. Davey, J., Dimmock, N.J. and Colman, A. 1985. Cell 40: 667-75. DeGrado, W.F., Wassenuan, Z.R. and Lear, J.D. 1989. Science 243: 622-628. Dekker, N., Cox, M., Boelens, R., Verrijzer, C.P., van der Vilet, P.C. and Kaptein, R. 1993. Nature 362: 852-855. Dente, L., Cesareni, G. and Cortese, R. 1983. Nuci. Acids Res. 11: 1645-1655. Dingwall, C., Robbins, J., Dilworth, S.M., Roberts, B. and Richardson, W.D. 1988. J Cell. Biol. 107: 84 1-849. Dobrzanski, P., Ryseck, R.-P. and Bravo, R. 1993. Mol. Cell Biol. 13: 1572-1582.  154  Dolan, J.W., Kirkman, C. and Fields, S. 1989. Proc. NatI. Acad. Sd. USA 86: 5703-5707. Dolan, J.W. and Fields, S. 1991. Biochim. Biophys. Acta 1088: 155-169. Dowell, S.J., Tsang, J.S.H, and Mellor, J. 1992. NucleicAcidsRes. 20: 4229-4236. Dranginis, A.M. 1990. Nature 347: 682-685. Dynlacht, B.D., Hoey, T. and Tjian, R. 1991. Cell 66: 563-576. Efstradiatis, A., Posakony, J.W., Maniatis, T., Lawn, R.M., O’Connell, C., Spiritz, R.A., DeRiel, J.K,, Forget, B.G., Slighton, L., Blechl, A.E., Smithies, 0., Baralle, F.E., Shoulders, C.C. and Proudfoot, N.J. 1980. Ce1121:653-668. Ellenberger, T.E., Brandl, C.J., Struhl, K. and Harrison, S.C. 1992. Cell 71: 1223-1237. Errede, B. and Ammerer, G. 1989. Genes Dev. 3: 1349-136 1. Falkner, F.G. and Zachau, H.G. 1984. Nature 310: 71-74. Feldman, J.B., Hicks, J.B. and Broach, J.R. 1984. 1 Mo!. Blot. 178: 8 15-834. Felsenfeld, G. 1992. Nature 355: 219-224. Ferré-D’Amaré, A.R., Prendergast, G.C., Zifl E.B. and Burley, S.K. 1993. Nature 363: 38-45. Finney, M. 1990. Cell 60: 5-6. Fischer, J.A., Giniger, B., Maniatis, T. and Ptashne, M. 1988. Nature 332: 853-856. Fisher, D.E., Parent, L.A. and Sharp, P.A. 1993. Cell 72: 467-476. Flanagan, P.M., Kelleher, R.J., III, Sayre, M.H., Tschochner, H. and Kornberg, R.D. 1991. Nature 350: 436-438. Frankel, A.D., Bredt, D.S. and Pabo, C.0. 1988. Science 240: 70-73. Frankel, AD. and Kim, P.S. 1991. Cell 65: 717-719. Gerlach, W.L. 1974. Heredity 32: 241-249. Giesman, D., Best, L. and Tatchell, K. 1991. Mol. Cell. Blot. 11: 1069-1079.  155  Gill, G. and Ptashne, M. 1987. Cell 51: 12 1-126. Gill, G., Sadowski, I. and Ptashne, M. 1990. Proc. NatI. Acad. Sd. USA 87: 2127-213 1. Goodin, D.B., Davidson, M.G., Roe, J.A., Mauk A.G. and Smith M. 1991. Biochemistry 30:  495 3-4962. Goutte, C. and Johnson, A.D. 1988. Cell 52: 875-882. Grayhack, E.J. 1992. Mol. Cell. Biol. 12: 3573-3582. Grueneberg, D.A., Natesan, S., Alexandre, C. and Gilman, M.Z. 1992. Science 257: 1089-1095. Guarente, L. and Ptashne, M. 1981. Proc. NatI. Acad. Sd. USA 78: 2 199-2003. Guarente, L. and Hoar, E. 1984. Proc. Nat!. Acad. Sd. USA 81: 7860-7864. Hagen, D.C., McCaffrey, G. and Sprague, G.F., Jr. 1986. Proc. Nat!. Acad. Sci. USA 83: 1418-1422. Hahn, S., Hoar, E. and Guarente, L. 1985. Proc. Nat!. Acad. Sci. USA 82: 8562-8566. Hall, M.N., Hereford, L. and Herskowitz, I. 1984. Cell 36: 1057-1065. Hall, M.N. and Johnson, A.D. 1987. Science 237: 1007-1012. Hall, M,N., Craik, C. and Hiraoka, Y. 1990. Proc. Natl. A cad. Sci, USA 87: 6954-6958. Han, K., Levine, M.S. and Manley, J.L. 1989. Cell 56: 573-583. Hanahan, D. 1983. 1 Mo!. Biol. 166: 557-580. Hanes, S.D. and Brent, R. 1989. Cell 57: 1275-1283. Hanes, S.D. and Brent, R. 1991. Science 251: 426-430. Harashima, S., Miller, A.M., Tanaka, K., Kusumoto, K.-I., Tanaka, K.-I., Mukai, Y., Nasmyth, K. and Oshima, Y. 1989. Mo!. Ce!!. Biol. 9: 4523-4530. Harlow, E. and Lane, D. 1988. Pp. 313-315 in Antibodies: A Laboratory Manual. Cold Spring Harbor Laboratory, New York, USA  156  Harrison, S.C. 1991. Nature 353: 715-719. Harshman, K.D., Moye-Rowley, W.S. and Parker, C.S. 1988. Cell 53: 32 1-330. Hailing, A., Holly, J., Saari, G. and MacKay, V.L. 1986. Mol. Cell. Biol. 6: 2 106-2114. Hayes, T.E., Sengupta, P., and Cochran, B.H. 1988. Genes Dev. 2: 1713-1722. Herskowitz, 1. 1987. Nature 329: 219-222. Hicks, J., Strathern, J.N. and Kiar, A.J.S. 1979, Nature 282: 478-483. Hicks, J.B. and Herskowitz, I. 1976. Genetics 83: 245-258. Hochstrasser, M. and Varshavsky, A. 1990. Cell 61: 697-708. Hochstrasser, M., Ellison, M.J., Chau, V. and Varshavsky, A. 1991. Mol. Cell. Biol. 88: 4606-46 10. Hodges, R.S., Semchuk, P.D., Taneja, A.K., Kay, C.M., Parker, J.M.R. and Mant, C.T. 1988. Pept. Res. 1: 19-30. Hodges, R.S. 1992. Curr. BioL 2: 122-125. Hodges, R.S., Zhou, N.E., Kay, CM. and Semchuk, P.D. 1990. Peptide Res. 3: 123-137. Hoey, T. and Levine, M. 1988. Nature 332: 858-86 1. Hoey, T., Weinzierl, R.O.J., Gill, G., Chen, J.-L., Dynlacht, B.D. and Tjian, R. 1993. Cell 72: 247-260. Hofinann, J.F.-X., Laroche, T., Brand, A.H. and Gasser, S.M. 1989. Cell 57: 725-737. Hope, l.A. Mahadevan, S. and Struhi, K. 1988. Nature 333: 635-640. ,  Hopper, AK. and Hall, B.D. 1975. Genetics 80: 41-59. Ru, J.C., O’Shea, E.K., Kim, P.S. and Sauer, R.T. 1990. Science 250: 1400-1430. Ru, J.C. and Sauer, R.T. 1992. NucI. Acids Mol. Biol. 6: 82-101. Huberman, J.A., Spotila, L.D,, Nawotka, K.A., El-Assouli, SM. and Davis, L.R. 1987. Cell 51: 473-481.  157  Hwang-Shum, J.-J., Hagen, D.C., Jarvis, E.E., Westby, C.A. and Sprague, G.F., Jr. 1991. Mol. Gen. Genet. 227: 197-204. Inostroza, J.A., Mermeistein, F.H., Ha, I., Lane, W.S. and Reinberg, D. 1992. Cell 70: 477-489. Jarvis, E.E., Hagen, D.C. and Sprague, G.F., Jr. 1988. Mol. Cell. Biol. 8: 309-20. Jarvis, E.E., Clark, K.L. and Sprague, G.F., Jr. 1989. Genes Dev. 3: 936-945. Jensen, R., Sprague, G.F., Jr. and Herskowitz, I. 1983. Proc. Natl. A cad. Sd. USA 80: 3035-3039. Johnson, A.D. and Herskowitz. 1985. Cell 42: 23 7-247. Johnson, P.F. and McKnight, S.L. 1989. Annu. Rev. Biochem. 58: 799-83 9. Jones, N. 1990. Cell6l:9-11. Jones, R.H., Moreno, S., Nurse, P. and Jones, N.C. 1988. Cell 53: 659-667. Kageyama, R. and Pastan, I. 1989. Cell 59: 815-825. Kakidani, H. and Ptashne, M. 1988. Cell 52: 161-167. Kanaan, M.N., Fu, Y.-H. and Marz1uf G.A. 1992. Biochemistry 31: 3197-3203. Kassir, Y. and Simchen, G. 1976. Genetics 82: 187-206. Keleher, C.A., Goutte, C. and Johnson, A.D. 1988. Cell 53: 927-936. Keleher, C.A., Passmore, S. and Johnson, A.D. 1989. Mol. Cell. Biol. 9: 5228-5230, Keleher, C.A., Redd, M.J., Schultz, J., Carison, M. and Johnson, A.D. 1992. Cell 68: 709-7 19. Kelleher, R.J., III, Flanagan, P.M. and Kornberg, R.D. 1990. Cell 61: 1209-1215. Kerpolla, T.K. and Curran, T. 1991. Curr. Opin. Sfruct. Biol. 1: 71-79. Kissinger, C.R., Liu, B., Martin-Binaco, E., Komberg, T.B. and Pabo, C.O. 1990. Cell 63: 579-590. Kiar, A.J.S., Fogel, S. and Radin, D.N. 1979. Genetics 92: 759-776. Kiar, A.J.S., Strathern, J.N., Broach, J.R. and Hicks, J.B. 1981. Nature 289: 239-244.  158  Kraulis, P.3., Paine, A.R.C., Gadhavi, P.L. and Laue, E.D. 1992. Nature 356: 448-450. Kronstad, J.W., Holly, l.A. and MacKay, V.L. 1987. Cell 50: 369-377. Kunkel, T.A. 1985. Proc. Nati, Acad. Sd. USA 82: 488-492. Kurjan, J. and Herskowitz, I. 1982. Cell 30: 933-943. Kurtz, S. and Shore, D. 1991. Genes Dev. 5: 616-628, Landschulz, W.H., Johnson, P.F. and McKnight, S.L. 1988a. Science 240: 1759-1764. Landschulz, W.H., Johnson, P.F., Adashi, E.Y. Graves, B.J. and McKnight, S.L. 1988b Genes Dev. 2: 786-800. Laughon, A. and Scott, M.P. 1984. Nature 310: 25-3 1. LeBowitz, J.H., Clerc, R.G., Brenowitz, M. and Sharp, P.A. 1989. Genes Dev. 3: 1625-1638. Lemontt, J.F., Fugit, D.R. and Mackay, V.L. 1980. Genetics 94: 899-920. Leuther, K.K., Salmeron, J.M. and Johnston, S.A. 1993. Cell 72: 575-585. Levine, M. and Manley, J.L. 1989. Cell 59: 405-408. Li-Weber, M., Eder, A., Krafft-Czepa, H. and Krammer, PH. 1992. J Immunology 148: 19 13-1918. Lorimer, I.A.J., Ho, C.-Y. and Smith, M. 1992. BioTechniques, 12: 536-543. Lovejoy, B., Choe, S., Cascio, D., McRorie, D.K., DeGrado, W.F. and Eisenberg, D. 1993. Science 259: 1288-1293. Luisi, B.F., Xu, W.C., Otwinowski, Z., Freedman, L.P., Yamamoto, K.R. and Sigler, P.B. 1991. Nature 352: 497-505. Lupas, A., Van Dyke, M., and Stock, 1. 1991. Science 252: 1162-1164. Ma, I. and Ptashne, M. 1988. Cell 55: 443 -446. Ma, J., Przibilla, E., Hu, J., Bogorad, L. and Ptashne, M. 1988. Nature 334: 63 1-633. MacKay, V.L. and Manney, T.R. 1 974a. Genetics 76: 255-271.  159  MacKay, V.L. and Manney, T.R. 1974b. Genetics 76: 273-288. Mackay, V. L., Welch, S.K., Insley, M. Y., Manney, T,R., Holly, J., Saari, G.C., and Parker, M. L. 1988. Proc. Nat!. Acad. Sd. USA 85: 55-95. Maine, G.T., Sinha, P. and Tye, B,-K. 1984. Genetics 106: 365-385. Mak, A. and Johnson, A.D. 1993. Genes Dev. 7:1862-1870. Maniatis, T., Fritsch, E.F., and Sambrook, J. 1982. in Molecular cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, USA Manney, T.R. 1983. 1 Bacterioh 155: 291-301. Marmorstein, R., Carey, M., Ptashne, M. and Harrison, S.C. 1992. Nature 356: 408-414. Marqusee, S. and Baldwin, R.L. 1987. Proc. Nat!. Acad. Sd. USA 84: 8898-8902. Maxam, A. and Gilbert, W. 1977. Proc. Nat!. Acad. Sci. USA 74: 560-564. McGinnis, W., Levine, M.S., Hafen, E., Kuroiwa, A. and Gehring, W.J. 1984. Nature 308: 428-433. McLachian, A.D. and Karn, J. 1982. Nature 299: 226-230. McNeil, J.B. and Smith, M. 1986. 1 Mol. Biol. 187: 363-378. Messing, J. 1983. Meth. Enzym. 101: 20-78. Michael, C., Kakidam, H., Leatherwood, J., Mostashari, F. and Ptashne, M. 1989. 1 Mo!. Biol. 209: 423-432. Miller, A.M. 1984. EMBOJ. 3: 1061-1065. Miller, A.M., McLachlan, A,D. and King, A. 1985b. EMBO 1 4: 1600-1611. Miller, J., Mackay, V.L. and Nasmyth, K.A. 1985a. Nature 314: 598-603. Mitchell, A.P. and Herskowitz, I. 1986. Nature 319: 738-742. Mitchell, P.J. and Tjian, R. 1989. Science 245: 37 1-378.  160  Miura, K., Titani, L., Kurosawa, Y. and Kanai, Y. 1992. Biochem. Biophys. Res. Commun. 187: 375-380. Mortimer, R.K. and Hawthorne, D.C. 1969. pp. 385-460 in Rose, A.H. and Harrison, J.S., eds. The Yeasts, Vol. 1. Academic Press, New York, USA Mukai, Y., Harashima, S. and Oshima, Y. 1991. Mo!. Cell. Biol. 11: 3773-3779. Murre, C., McCaw, P.S. and Baltimore, D. 1989a. Cell 56: 777-783. Murre, C., McCaw, P.S., Vaessin, H., Caudy, M., Jan, L.Y., Jan, Y.N., Cabrera, C.V., Buskin, J.N., Hauschka, S.D., Lassar, A.B., Weintraub, H. and Baltimore, D. 1989b. Cell 58: 537-544. Nagawa, F. and Fink, G.R. 1985. Proc. Nat!. Acad. Sci. USA 82: 8557-856 1. Nakayama, N., Miyajima, A. and Arai, K. 1985. EMBO 1 4: 2643-2648. Nakazawa, N., Harashima, S. and Oshima, Y. 1991. Mo!. Cell. Biol. 11: 5693-5700. Nasmyth, K.A. and Tatchell, K. 1980, Cell 19: 753-764. Nasmyth, K.A.,Tatchell, K., Hall, B.D., Astell, C.R. and Smith, M. 1981. Nature 289: 244-250, Ner, S.S. and Smith, M. 1989. Mo!. Cell. Biol. 9: 4613-4620. Ner S.S., Goodin, D.B., Pielak, G.J., and Smith, M. 1988. BioTechniques 6: 408-412. Ngsee, J.K. 1987. PhD. thesis, University of British Columbia. Nourbakhsh, M., Hoffmann, K. and Hauser, H. 1993. EIvIBO 1 12: 45 1-459. O’Neil, KT. and DeGrado, W.F. 1990. Science 250: 646-65 1. O’Neil, KT., Hoess, R.H. and DeGrado, W.F. 1990. Science 249: 774-778. O’Shea, E.K., Klemm, J.D., Kim, P.S. and Alber, T. 1991. Science 254: 539-544. O’Shea, E.K., Rutkowski, R. and Kim, P.S. 1992. Cell 68: 699-708. Oas, T.G., McIntosh, L.P., O’Shea, E.K., Dahlquist, F.W. and Kim, P.S. 1990. Biochemistry 29: 2891-2894.  161  Otting, G., Qian, Y.Q., Billeter, M., Muller, M, Affolter, M., Gehring, W.J. and Wüthrich, K. 1990. EMBOJ 9: 3085-3092. Pabo, C.O. and Sauer, R.T. 1992. Annu. Rev. Biochem. 61: 1053-1095. Pagolotti, A.L., Jr. and Santi, D.V. 1977. Bioorg. Chein. 1: 277-311. Parsiow, T.G., Blair, D.L. Murphy, W.J. and Granner, D.K. 1984. Proc. Nat!. Acad. Sci. USA 81: 2650-2654. Passmore, S., Maine, G.T., Elbie, R., Christ, C. and Tye, B.-K. 1988. 1 Mo!. Bio!. 204: 593-606. Passmore, S., Elbie, R. and Tye, B.-K. 1989. Genes Dev. 3: 921-935. Patel, L., Abate, C. and Curran, T. 1990. Nature 347: 572-575. Pavietich, N.P. and Pabo, CO. 1991. Science 252: 809-8 17. Percival-Smith, A., Muller, M., Affolter, M. and Gehring, W.J. 1990. EMBO 1 9: 3967-3974, Phillips, C.L., Vershon, A.K., Johnson, A.D. and Dahiquist, F.W. 1991. Genes Dev. 5: 764-722. Phillips, C.L. 1992. p.p. 49 in Santa Cruz Summer Conferences in Biology: Protein Structure and Fuction. University of California, Santa Cruz. (Abs) Porter, S.D. and Smith, M. 1986. Nature 320: 766-768. Porter, S.D. 1987. Ph.D. thesis, University of British Columbia. Pratt, W.B., Jolly, D.J., Pratt, D.V., Hollenberg, S.M., Giguere, V., Cadepond, F.M., Schweizer-Groyer, G., Catelli, M.-G., Evans, R.M. and Baulieu, E.-E. 1988. 1 Bio!. Chem. 263: 267-273. Prendergast, G.C. and Zifl E.B. 1989. Nature 341: 392. Primig, M., Winkler, H., and Ammerer, G. 1991. EMBOI 10: 4209-4218. Ptashne, M. 1988. Nature 335: 683-689. Pugh,B.F.andTjian,R. 1990. Ce1161: 1187-1197.  162  Renkawitz, R. 1990. Trends Genet. 6: 193-196. Rochette-Egly, C., Fromental, C. and Chambon, P. 1990. Genes Dev. 4: 137-150. Roeder, R,G. 1991. Trends Biochem. Sd. 16: 402-408. Rose, M., Grisafi, P. and Botstein, D. 1984. Gene 29: 113-124. Rosen, C.A., Sodroski, J.G. and Haseltine, W.A. 1985. Cell 41:813-823. Roth, S.Y., Dean, A. and Simpson, R.T. 1990. Mo!. Cell. Blot. 10: 2247-2260. Roth, S.Y., Shimizu, M., Johnson, L., Grunstein, M. and Simpson, R.T. 1992. Genes Dev. 6: 411-425. Ruby, S.W., Szostak, J.W. and Murray, A.W. 1983. Meth. Enzynz. 101: 253-269. Ruberti, I., Sessa, G., Lucchetti, S. and Morelli, G. 1991 EMBO 1 10: 1781-1791. Russell, D.W., Jensen, R., Zoller, M.J., Burke, J., Errede, B., Smith, M. and Herskowitz, I. 1986. Mo!. Cell. Blot. 6: 4281-4294. Sadowski, I., Ma, J., Triezenberg, S.J. and Ptashne, M. 1988. Nature 335: 563-564. Saltzman, A.G. and Weimnann, R. 1989. FASEBI 3: 1723-1733. Sanger, F., Nicklen, S. and Coulson, A.R. 1977. Proc. Natt. Acad. Sd. USA 74: 5463-5467. Sauer, R.T., Smith, D.L. and Johnson, AD. 1988. Genes Dev. 2: 807-816. Sawadogo, M. and Van Dyke, M.W. 1991. Nuct. Acids Res. 3: 674 Schena, M. and Davis, R.W. 1992. Proc. Nat!. Acad. Sd. USA 89: 3894-3898. Schirm, S., Jiricny, J and Schaffner, W. 1987. Genes Dev. 1: 65-74. Schmidt-DOff, T., Oertel-Buchheit, P., Pernelle, C., Bracco, L., Schnarr, M. and Granger-Schnarr, M. 1991. Biochemistry 30: 9657-9664. Schuermann, M., Neuberg, M., Hunter, J.B., Jenuwein, T., Ryseck, R.-P., Bravo, R. and Muller, R. 1989. Cell 56: 507-516.  163  Schuermann, M., Hunter, J.B., Hennig, G. and Muller, R. 1991. Nucl. Acids Res. 19: 739-746. Schultz, J. and Carison, M. 1987. Mo!. Cell. Biol. 7: 3637-3645. Schultz, J., Marshall-Carison, L. and Carison, M. 1990. Mo!. Cell. Biol. 10: 4744-4756. SchUtte, I., Viallet, J., Nan, M., Segal, S., Fedorko, J. and Minna, J. 1989. Cell 59: 987-997. Schwabe, J.W.R., Neuhaus, D. and Rhodes, D. 1990. Nature 348: 458-46 1. Schwarz-Sommer, Z., Huijser, P., Nacken, W., Saedler, H. and Sonimer, H. 1990. Science: 250: 93 1-936. Scott, M.P., Tamkun, J.W. and Hartzell, G.W., III 1989. BBA Rev. Cancer 989: 25-48. Scott, M.P. and Wenier, A.J. 1984. Proc. Nat!. Acad. Sci. USA 81: 4115-4119. Sengupta, P. and Cochran, B.H. 1990. Mol. Cell. Biol. 10: 6800-6812. Sengupta, P. and Cochran, B.H. 1991. Genes Dev. 5: 1924-1934. Serfling, E., Jasin, M. and Schaffner, W. 1985. Trends Genet. 1: 224-230. Sessa G., Morelli, G. and Ruberti, 1. 1993 EMBOJ. 12: 3507-3517. Shepherd, J.C.W., McGinnis, W., Carrasco, A.E., De Robertis, E.M. and Gehring, W.J. 1984. Nature 310: 70-71. Shimizu, M., Roth, S.Y., Szent-Gyorgyi, C. and Simpson, R.T. 1991. EIv1BOI 10: 3033-3041. Shore, D. and Nasmyth, K. 1987. Cell 51: 72 1-732. Sigler, P.B. 1988. Nature 333: 210-212. Sikorski, R.S., Boguski, M.S., Goebi, M. and Hieter, P. 1990. Cell 60: 307-3 17. Siliciano, P.G. and Tatchell, K. 1984. Cell 37: 969-978. Simpson, R.T. 1990. Nature 343: 387-389. Smale, S.T. and Baltimore, D. 1989. Cell 57: 103-113.  164  Smith, D.L. and Johnson, A.D. 1992. Cell 68: 133-142. Smith, M., Leung, D.W., Gillam, S., Astell, C.R., Montgomery, D.L. and Hall, B.D. 1979. Cell 16: 753-761. Smith, M.M. and Andrésson, O.S. 1983. J. Mo!. Blot. 169: 663-690. Smith, M.R. and Greene, W.C. 1989. Proc. Nat!. Acad. Sd. USA 86: 8526-8530. Sodek, 3., Hodges, R.S., Smillie, L.B. and Jurasek, L. 1972. Proc. Natl. Acad. Sci. USA 69: 3800-3804. Song, D., Dolan, J.W., Yuan, Y.O. and Fields, S. 1991. Genes Dev. 5: 741-50. Sprague, G.F., Jr. and Herskowitz, I. 1981. J. Mo!. Blot. 153: 305-321. Sprague, G.F., Jr., Jensen, R. and Herskowitz, I. 1983. Cell 32: 409-415. Strathern, J., Hicks, J. and Herskowitz, I. 1981. J Mol. Biol. 144: 357-372. Strathern, 3., Shafer, B., Hicks, 3. and McGill, C. 1988, Genetics 120: 75-81. Struhl, K., Stinchcomb, D.T., Scherer, S. and Davis, R.W. 1979. Proc. Nati. A cad. Sd. USA 76: 1035-1039. Struhi, K. 1984. Proc. Nath Acad. Sd. USA 81: 7865-7869. Struhi, K. 1985. Nature 317: 822-824. Struhi, K. 1987. Cell 50: 84 1-846. Struhi, K. 1989. Ann. Rev. Biochem. 58: 105 1-1077. Sundaralingam, M., Drendel, W. and Greaser, M. 1985. Proc. Nat!. Acad. Sci. USA 82: 7944-7947. Takeda, Y., Ohlendort D.H., Anderson, W.F. and Matthews, B.W. 1983. Science 221: 1020-1026. Talanian, R.V., McKmght, C.J., and Kim, P.S. 1990. Science 249: 769-771. Tan, S., Ammerer, G. and Richmond, T.J. 1988. EMBO J 7: 4255-4264. Tan, S. and Richmond, T.J. 1990. Cell 62: 367-377.  165  Tanese, N., Pugh, B.F. and Tjian, R. 1991. Genes Dev. 5: 2212-2224. Tatchell, K., Nasmyth, K.A. and Hall, B.D., Astell, C. and Smith, M. 1981. Cell 27: 25-35. Titus, M.A. 1993. Curr. Opin. Cell. Biol. 5: 77-81. Treisman, J., GOnczy, P., Vashishtha, M., Harris, E. and Desplan, C. 1989, Cell 59: 553-562. Triezenberg, S.J. Kingsbury, R.C. and McKnight, S.L. 1988. Genes Dev. 2: 718-729. Trumbly, R.J. 1986. f Bacteriol. 166: 123-1127. Turner, R. and Tjian, R. 1989. Science 243: 1689-1694. Van Hoy, M., Leuther, K.K., Kodadek, T. and Johnston, S.A. 1993. Cell 72: 587-594. Vernet, T., Dignard, D. and Thomas, D.Y. 1987, Gene 52: 225-233. Vershon, A.K. and Johnson, A.D. 1993. Cell 72: 105-112. Vinson, C.R., Sigler, P.B. and McKnight, S.L. 1989. Science 246: 911-916. Vinson, C.R. and Garcia, K.C. 1992. New biol. 1992. 4: 396-403. Voronova, A. and Baltimore, D. 1990. Proc. NatI. A cad. Sci. USA 87: 4722-4726. Wallis, J.W., Hereford, L. and Grunstein, M. 1980. Cell 22: 799-805. Weber, I.T. and Steitz, T.A. 1987. 1 Mol. Biol. 198: 3 11-326. Webster, N., Jin, J.R., Green, S., Hollis, M. and Chambon, P. 1988. Cell 52: 169-178. Weis,L. and Reinberg,D. 1992. FASEBJ. 6: 3300-3309. Weiss, M.A., Ellenberger, T., Wobbe, C.R., Lee, J.P., Harrison, S.C. and Struhl, K. 1990. Nature 347: 575-578. Whitby, F.G., Kent, H., Stewart, F., Stewart, M., Xie, X., Hatch, V., Cohen, C. and Phillips, G.N., Jr. 1992. 1 Mo!. Biol. 227: 441-452. Wild, C., Oas, T., McDanal, C., Bolognesi, D. and Matthews, T. 1992. Proc. Nat!. Acad. Sci. USA 89: 10537-10541.  166  Williams, F.E. and Trumbly, R.J. 1990. Mo!. Cell. Biol. 10: 6500-6511. Williams, F.E. Varanasi, U. and Trumbly, R.J. 1991. Mo!. Cell. Biol. 11: 3307-33 16. Wilson, K.L. and Herskowitz, I. 1984. Mo!. Cell. Biol. 4: 2420-2427. Wolberger, C., Vershon, A.K., Liu, B., Johnson, A.D. and Pabo, CO. 1991. Cell 67: 5 17-528. Yamamoto, K. 1985. Annu. Rev. Genet. 19: 209-252. Yamamoto, K., Mori, S., Okamoto, T., Shimotohno, K. and Kyogoku, Y. 1991. NucleicAcidsRes. 19: 6107-12. Yamamoto, K. and McKnight, S. 1992. eds. in Transcriptional Regulation. Cold Spring Harbor Laboratory, New York, USA Yuan,Y.O.andFields,S. 1991. Mo!. Cel. Blot. 11: 5910-5918. Zawel, L.and Reinberg, D. 1992. and Curr. Opin. Cell Blot. 4: 488-495. Zhang, M., Rosenblum-Vos, L.S., Lowry, C.V., Boakye, K.A. and Zitomer, R.S. 1991. Gene 97: 153-161. Zhao, L.-J. and Padmanabhan, R. 1988. Ce!! 55: 1005-1015. Zhou, N.E., Kay, C.M. and Hodges, R.S. 1992a. I Bio!. Chem. 267: 2664-2670. Zhou, N.E., Kay, C.M. and Hodges, R.S. 1992b. Biochemistry 31: 5739-5746. Zhou, N.E., Zhu, B-Y., Kay, C.M. and Hodges, R.S. 1992c. Biopotymers 32: 419-426. Zhou, N.E., Kay, CM. and Hodges, R.S. 1993. Biochemistry 32: 3 178-3187. Zhu, B.-Y., Zhou, N.E., Kay, C.M. and Hodges, R.S. 1993. Protein Sd. 2: 383-394.  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items