Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Identification and characterization of chromosome instability mutants in the yeast Saccharomyces cerevisiae… Yuen, Wing Yee Karen 2007

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

Item Metadata

Download

Media
831-ubc_2007-268278.pdf [ 45.37MB ]
Metadata
JSON: 831-1.0100361.json
JSON-LD: 831-1.0100361-ld.json
RDF/XML (Pretty): 831-1.0100361-rdf.xml
RDF/JSON: 831-1.0100361-rdf.json
Turtle: 831-1.0100361-turtle.txt
N-Triples: 831-1.0100361-rdf-ntriples.txt
Original Record: 831-1.0100361-source.json
Full Text
831-1.0100361-fulltext.txt
Citation
831-1.0100361.ris

Full Text

IDENTIFICATION AND CHARACTERIZATION OF CHROMOSOME INSTABILITY MUTANTS IN THE YEAST SACCHAROMYCES CEREVISIAE AND IMPLICATIONS TO HUMAN CANCER by WING Y E E K A R E N Y U E N B.Sc. (Hon), Simon Fraser University, 2001 A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of D O C T O R OF P H I L O S O P H Y in The Faculty of Graduate Studies (Medical Genetics) The University of British Columbia January 2007 © Wing Yee Karen Yuen, 2007 A B S T R A C T Chromosome instability (CLN) is a hallmark of cancers and may contribute to tumorigenesis. Many genes involved in maintaining chromosome stability are conserved in eukaryotes, and some are mutated in cancers. The goal of this thesis is to use Saccharomyces cerevisiae as a model to identify and characterize genes important for chromosome maintenance, investigate the relevance of C L N to cancer, and develop a strategy to identify candidate therapeutic target genes for selective ki l l ing o f cancer cells. To systematically identify genes important for chromosome stability, non-essential gene deletion yeast mutants were examined using 3 complementary C I N assays. The chromosome transmission fidelity assay monitors loss of an artificial chromosome. The bimater assay monitors loss of heterozygosity at the mating type locus in homozygous diploid deletion mutants. The a-like faker assay detects loss of the MAT a mating type locus in haploid deletion mutants. 293 C I N mutants were identified, including genes functioning in the chromosome or cell cycle, and genes not clearly implicated in chromosome maintenance, such as MMS22, MMS1, RTT101 and RTT107. Phenotypic, genetic and biochemical analyses of these 4 gene products indicate that they function in double strand break repair. They may form a ubiquitin ligase complex that regulates the level of some proteins, including Mms22p itself, during D N A damage response. Human homologues of 10 yeast C I N genes identified were previously shown to be mutated in cancers, suggesting that other human homologues are candidate cancer genes. 101 human homologues o f yeast C L N genes were sequenced in a panel of colorectal cancers, identifying 20 somatic mutations in 8 genes. In particular, 17 mutations were found in 5 genes involved in sister chromatid cohesion. Further functional studies should reveal whether mutations in cohesion genes contribute to C I N in cancers. While C I N mutations may contribute to cancer, C I N cancer cells may become inviable when combined with another non-essential mutation, providing the basis for n cancer ce l l -speci f ic therapy. Muta t ions i n CTF4, CTF18, and DCC1 i n yeast cause synthetic lethali ty w h e n combined w i t h mutations i n various C I N genes whose human homologues are mutated i n cancers. S u c h analyses i n yeast can propose potential drug targets i n human for cancer therapy. i n T A B L E O F C O N T E N T S A B S T R A C T i i T A B L E O F C O N T E N T S iv LIST O F T A B L E S ix LIST O F F I G U R E S x A C K N O W L E D G M E N T S x i i C O - A U T H O R S H I P S T A T E M E N T . xv C H A P T E R 1 I n t r o d u c t i o n : M a i n t e n a n c e o f C h r o m o s o m e S t a b i l i t y i n E u k a r y o t e s a n d the R e l a t i o n s h i p w i t h C a n c e r 1 1.1 Maintenance of chromosome stability in eukaryotes 2 1.1.1 The cell and chromosome cycles in eukaryotes 2 1.1.2 Budding yeast as a model organism to study the cell and chromosome cycle.. 4 1.1.3 Biological processes that affect chromosome stability 6 1.1.3.1 Kinetochores mediate the attachment with mitotic spindles 6 1.1.3.2 Mitotic spindle checkpoint 10 1.1.3.3 Sister chromatid cohesion 13 1.2 Chromosome instability and cancer 15 1.2.1 Aneuploidy is a hallmark of cancer 15 1.2.2 Relationship between a state o f aneuploidy and an increased rate o f chromosome instability (CLN) 17 1.2.3 C I N occurs at early stage of cancer, and can be a driving force in tumorigenesis 17 1.2.4 Genetic basis of C I N in cancer 18 1.2.4.1 Cancer-prone syndromes 18 1.2.4.2 Mutations in mitotic spindle checkpoint 19 1.2.4.3 Misregulation of kinetochore proteins 22 iv 1.2.4.4 Additional examples of mutations in genes involved in chromosome segregation 24 1.2.4.5 Therapeutic implications 25 1.3 Overview of thesis 29 CHAPTER 2 Identification of Chromosome Instablity Mutants in the Budding Yeast Saccharomyces cerevisiae and the Implication to Human Cancer 43 2.1 Introduction 44 2.2 Materials and methods 47 2.2.1 Genome-wide screens 47 2.2.1.1 C T F screen 47 2.2.1.2 Bimater screen 48 2.2.1.3 a-like faker screen 49 2.2.2 Strain verification 50 2.2.3 Bioinformatic analysis 51 2.2.3.1 Functional analysis '. 51 2.2.3.2 B L A S T analysis 52 2.2.3.3 Protein and synthetic lethal interaction network 52 2.2.4 Electrophoretic karyotype of a-like fakers 52 2.3 Results 53 2.3.1 Genome-wide marker loss screens identify 130 yeast deletion mutants 53 2.3.2 Functional distribution of yeast C I N genes 56 2.3.3 Integration of genome-wide phenotypic screen with genetic screens reveals functions of uncharacterized genes in chromosome stability maintenance 58 2.3.4 Chromosome loss is the major mechanism of M 4 r i o s s in a-like fakers 59 2.3.5 Many yeast C I N genes are conserved 60 2.3.6 A strategy for cancer therapy: synthetic lethality and selective cancer cell ki l l ing 61 2.4 Discussion 62 v CHAPTER 3 Identification of Somatic Mutations in Cohesion Genes in Colorectal Cancers with Chromosome Instability 79 3.1 Introduction 80 3.2 Materials and Methods 82 3.2.1 Gene identification 82 3.2.2 Sequencing 82 3.2.3 Yeast SMC1 mutants construction and characterization 82 3.3 Results 84 3.3.1 20 somatic mutations were found in 8 C I N genes 84 3.3.2 Mutation frequency in comparison to prevalence of mutations 84 3.3.3 A conserved missense mutation in yeast SMC1 causes mild C L N 85 3.4 Discussion 87 3.4.1 SMC1L1, CSPG6, NIPBL, STAG2 and STAG3 87 3.4.2 A L M 87 3A.3RNF20 88 3.4.4 UTX 89 CHAPTER 4 Characterization of MMS22, MMS1, RTT101 and RTT107 in the Maintenance of Genome Integrity 103 4.1 Introduction 104 4.1.1 RTT107 105 4A.2RTT10P 107 4.1.3 MMS22 mdMMSl 108 4.2 Materials and Methods : ". 110 4.2.1 Yeast strains and media 110 4.2.2 Quantification of chromosome transmission fidelity (ctf) 110 4.2.3 Genome-wide yeast-two-hybrid screens 110 4.2.4 Co-immunoprecipitation 110 vi 4.2.5 Mass spectrometry 111 4.2.6 Survival assay in HO-induced double strand breaks I l l 4.2.7 Microscopy I l l 4.3 Results 113 4.3.1 mms22A, mmslA, rttlOlA and rttl07A exhibit chromosome instability 113 4.3.2 mms22A exhibits defects in cell cycle progression 113 4.3.3 mms22A has reduced survival rate with the introduction of D S B s 114 4.3.4 Mms22p interacts with replication initiation and D N A repair proteins that may constitute a novel repair pathway 116 4.3.4.1 Mass spectrometry analysis 116 4.3.4.2 Yeast-two-hybrid analysis - 117 4.3.4.3 Genetic interaction analysis 118 4.3.5 R t t lO lp regulates Mms22p 119 4.4 Discussion 122 4.4.1 Conservation of the R t t lO lp complex? 122 4.4.2 Dia2p may play a redundant role with R t t lO lp complex in replication regulation 124 4.4.3 Identifying targets for R t t lO lp ubiquitin ligase 124 C H A P T E R 5 C o n c l u s i o n s a n d F u t u r e D i r e c t i o n s 151 5.1 Conclusions 152 5.2 Future Directions 154 R E F E R E N C E S 156 A P P E N D I C E S 183 Appendix 1 High confidence yeast C L N genes identified by the 3 marker loss screens 183 Appendix 2 Lower confidence yeast C I N genes identified by the 3 marker loss screens 195 v i i Appendix 3 Gene ontology (GO) cellular component annotation enrichment among non-essential yeast C L N genes 206 Appendix 4 Gene ontology (GO) biological process annotation enrichment among non-essential yeast C L N genes 217 Appendix 5 Homologues of budding yeast C I N genes in human, mouse, worm, fly, and fission yeast 222 Appendix 6 A subset of yeast C I N genes identify human homologues that are mutated in cancer or are associated with other human diseases 239 v i i i L I S T O F T A B L E S C H A P T E R 1: Table 1.1 Germline mutations of C L N and M I N genes causing cancer-prone syndromes 31 Table 1.2 Kinetochore and spindle checkpoint gene mutation or misregulation associated with cancer 33 C H A P T E R 2: Table 2.1 24 ctf mutants cloned to date 66 Table 2.2 List of yeast strains used in chapter II 67 Table 2.3 Human proteins homologous to yeast C I N genes are mutated in cancer 68 C H A P T E R 3: Table 3.1 Relationship of 100 human candidate C I N genes used in (Wang et al., 2004b) with yeast genes : 91 Table 3.2 List o f yeast strains used in chapter III 94 Table 3.3 101 candidate C I N genes analyzed in this study 95 Table 3.4 Somatic mutations identified in candidate C I N genes in C L N colorectal cancer cells 98 C H A P T E R 4: Table 4.1 Types D N A lesions generated by various D N A damaging agents 127 Table 4.2 List of yeast strains used in chapter IV 128 Table 4.3 Quantification of chromosome loss (CL) , non-disjunction (NDJ) and chromosome gain (CG) by half-sectored assay 131 i x L I S T O F F I G U R E S C H A P T E R 1: Figure 1.1 The budding yeast cell cycle and chromosome cycle 34 Figure 1.2 Organization of centromere 35 Figure 1.3 The process of achieving bipolar attachment 36 Figure 1.4 Types of kinetochore-microtubule attachments 37 Figure 1.5 Structure of cohesin and a possible mechanism by which it might hold sister chromatids together 38 Figure 1.6 The stages of mitosis 39 Figure 1.7 Cellular processes involved in replication and segregation of chromosomes during mitosis 40 Figure 1.8 Multiple roads to aneuploidy 41 Figure 1.9 Flowchart of therapeutic strategy based on candidate C I N gene identification 42 C H A P T E R 2 : Figure 2.1 Three screen methods 69 Figure 2.2 Three marker loss screens 73 Figure 2.3 130 high confidence nonessential yeast C I N genes 74 Figure 2.4 Functional groups of 293 C L N genes 75 Figure 2.5 92 protein interactions among 102 C I N proteins 76 Figure 2.6 A - l i k e fakers result from whole chromosome loss, gross chromosomal rearrangement, and gene conversion 77 Figure 2.7. Common synthetic lethal interactions among yeast C I N genes that have human homologs mutated in cancer 78 x C H A P T E R 3: Figure 3.1 Mutations in human SMC1L1 in colorectal cancers and analogous mutations in yeast SMC1 99 Figure 3.2 Synthetic lethal interactions of yeast C I N genes whose human homologs were found mutated in colorectal cancers 102 C H A P T E R 4: Figure 4.1 D N A damage and repair mechanisms 132 Figure 4.2 Two-dimensional hierarchical clustering of drugs and yeast deletion mutants...J : 133 Figure 4.3 Ce l l cycle and morphology defects of mms22A 137 Figure 4.4 Defects of mms22A in double-strand breaks 138 Figure 4.5 Kinetics of D S B repair 139 Figure 4.6 Mms22p co-immunoprecipitates with R t t lO lp and Rtt l07p 141 Figure 4.7 Physical interaction of R t t lO lp with M m s l p 142 Figure 4.8 Yeast-two-hybrid interactions using bait protein: (A) Mms22p, (B) R t t lO lp , and (C) Rtt l07p 143 Figure 4.9 Physical interactions of Mms22p with M m s l p 144 Figure 4.10 Genetic interactions of mms22A mutants 145 Figure 4.11 Interaction network of MMS22, MMS1, RTT101 and RTT107 146 Figure 4.12 Mms22p expression is regulated by R t t lO lp 148 Figure 4.13 Ce l l cycle expression of R t t lO lp 150 x i ACKNOWLEDGMENTS The completion of this thesis would not be possible without the help and support of many people, who I would like to take this opportunity to thank and acknowledge. First, I would like to thank my supervisor Phi l Hieter for his guidance and advices throughout the course of my PhD study. Phi l has provided a very supportive lab environment with many great scientists and resourceful colleagues. Phi l also has extensive networking with scientists in the yeast community and worldwide. Phi l has introduced me to many experienced scientists, and he has been instrumental in setting up productive collaborations. I am also thankful for Phi l ' s openness, positivity and generosity. Second, I would like to thank members of the Hieter lab for a constructive and fun working and learning atmosphere. I want to thank the many helpful post-doctoral fellows/visiting professors in the lab, including Viv ien Measday, Kris t in Batez, Melanie Mayer, Daniel Kornitzer, Shay Ben-Aroya, K i r k McManus, A m i r Aharoni, and Giora Simchen, who all never hesitated to offer guidance and assistance. I am grateful to grow scientifically with fellow students in the lab. Thanks Andy Page for introducing me to the basics of yeast biology and genetics. Thanks Isabelle Pot for teaching me many molecular and biochemical techniques. I would not survive these years without Isabelle's consistent encouragement and our daily energizing 'cookie club'. Thanks Ben Cheng, Elaine Law, Ben Montpetit, and Jan Stopel for insightful discussion and interactions (and a little bit of pressure). I am appreciative of the technical assistance by Teresa K w o k . A n d thanks Irene Barrett and Dave Thomson for keeping the lab in place. Next, I like to thank collaborators who have contributed to this work. I am indebted to Forrest Spencer at Johns Hopkins University for enormous guidance and support. Her enthusiasm, compassion and understanding always brighten my days. Thanks also go to Cheryl Warren (from Forrest Spencer's lab) who performed the a-like faker screen and subsequent analysis (Chapter 2); Mark Flory (from Rudi Aebersold's lab at the Institute of Systems Biology) who performed the mass spectrometry analysis x i i (Chapter 4); Tony Hazbun (from Stan Fields' lab at the Yeast Resource Centre) who performed the yeast-two-hybrid analysis (Chapter 4); Shira Goldstein (from Martin Kupiec's lab at Tel A v i v University) who performed the Southern blot and P C R analysis for monitoring the repair kinetics of double-strand D N A breaks (Chapter 4). Tom Barber, Marcelo Reis (in Victor Velculescu and Bert Vogelstein group at Johns Hopkins University) and Christoph Lengauer have been instrumental to the cancer mutation testing project (Chapter 3). In addition, I need to thank members of my advisory committee, Carolyn Brown, A n n Rose, and L i z Conibear, for guidance, directions and support throughout the years, and invaluable feedbacks on my thesis. I also want to thank Miche l Roberge and M i k e Kobor for stimulating discussion and Joanne Fox (in U B C Bioinformatics Centre) for bioinformatic assistance (Chapter 2). Besides, I would like to thank teachers and mentors who have inspired me to pursue a scientific endeavor. M y high school physics teacher in Hong Kong, M r . Y . S . A u , has exposed me to the excitements in doing experiments. Dennis Jewell demonstrated his passion in teaching and supported me tremendously during my first year in Canada. Working with Jean St. Pierre at Ballard Power Systems in my first co-op experience has exposed me to a new perspective in applied research. M y honor thesis supervisor Lynne Quarmby at Simon Fraser University has demonstrated to me her creativity and enthusiasm in studying cellular biology. Finally, I am thankful to my parents, Mandy and John Yuen, for their unconditional love, continuous trust, support and an unlimited supply of heart-warming soups! I could not imagine how my years of graduate life would be like without the love, back-up, encouragement and understanding of my soul mate Simon Chiu. I need to thank my brother Ken, aunt Pamela Cheng, very good friends Rosa Tchao, Helen Leung, Vernice Y u , Ivy N g , Daphne Chow and WaiTak Tsun for their long-distance support as well . I also want to say thanks to the brothers and sisters at Pacific Grace M B Church, especially Oliver Chan and Rev. Issac Chang, for their prayers and support. A n d thank God for all the blessings and challenges. x m Throughout this thesis w o r k , I was f inanc ia l ly supported b y the N a t i o n a l Science and Eng inee r ing Research C o u n c i l ( N S E R C ) postgraduate scholarships and the U n i v e r s i t y o f B r i t i s h C o l u m b i a graduate fe l lowship , w h o I w o u l d l i ke to thank here. xiv CO-AUTHORSHIP STATEMENT Chapter 2 of this thesis was co-written by Cheryl D . Warren. I designed and performed all the work described in this chapter except for the following: - a-like faker screen and retest - sequencing of unique tags of mutants - electrophoretic karyotyping of a-like fakers These experiments were performed by Cheryl Warren and Ou Chen in Forrest Spencer's laboratory. xv CHAPTER 1 Introduction: Maintenance of Chromosome Stability in Eukaryotes and the Relationship with Cancer Part of this chapter has been published. Karen W Y Yuen*, Ben Montpetit* and Phi l Hieter (These authors contributed equally to this work). (2005) The Kinetochore and Cancer: What's the Connection? Current Opinion in Ce l l Biology. 17:1-7. 1 1.1 Maintenance of chromosome stability in eukaryotes 1.1.1 The cell and chromosome cycles in eukaryotes Over 100 years ago, Walter Flemrning first described mitosis (1874), and Theodor Boveri showed the dramatic synchronous separation of chromosomes during the first mitotic division of fertilized sea urchin eggs (1902) (reviewed in (Manchester, 1995; Paweletz, 2001)). The maintenance of an individual organism requires that each daughter cell receives a full and exact complement of genetic information from its mother cell. To ensure the conservation of euploidy (normal number of chromosomes) in eukaryotic cells, genetic information must be accurately copied and transmitted to each daughter cell during every mitotic division cycle. Errors in chromosome segregation (including chromosome non-disjunction and chromosome loss) result in aneuploidy (abnormal number of chromosomes). Phenotypic consequences of these imbalances in chromosome number could be profound and dire. Boveri later postulated that unequal segregation of chromosomes might be a cause for tumor development and birth defects (1914) (reviewed in (Manchester, 1995)). Indeed, aneuploidy is a hallmark o f cancer, and the relationship between chromosome missegregation and cancer w i l l be discussed in section 1.2. First, the progression of a normal mitotic cell cycle is reviewed. The mitotic/somatic cell cycle is divided into 4 phases: G l (growth/gap), S ( D N A synthesis), G2 and M (mitosis). The M phase is subdivided into: prophase, prometaphase, metaphase, anaphase, and telophase. In prophase of metazoans, sister chromatids condense, and the nuclear envelope breaks downs. During prometaphase, sister kinetochores undergo the process of establishing bi-polar orientation with opposite spindle poles. Once bi-polar orientation has been achieved by all kinetochores, the cell enters metaphase when all chromosomes congress to a central position called the metaphase plate. Anaphase then begins and is composed of 2 steps: anaphase A during which sister chromatids separate, and move away from each other toward spindle poles; and anaphase B when the spindle poles separate by moving in opposite directions. In telophase, chromosomes decondense and a new nuclear envelope forms. Finally, cytokinesis occurs when cytoplasm divides and 2 daughter cells are formed. Each step in 2 the cell cycle has to be executed with high fidelity and coordinated temporally and spatially in order to maintain genetic integrity. Cel l cycle progression is regulated mainly through stage-specific phosphorylation of proteins by cyclin-dependent kinases (CDKs) . C D K activity is controlled by both positive and negative regulatory subunits called cyclins, and C D K inhibitors (CKI) (e.g. SIC1), respectively. Cyclins are targeted for ubiquitin-mediated degradation at specific stages of the cell cycle. In addition, key proteins are degraded in a cell cycle-specific manner to prevent events such as D N A re-replication and centrosome re-duplication. Otherwise, polyploidy (multiple sets of the normal number of chromosomes) or aneuploidy could result at an unacceptably high level. Proteins targeted for degradation are first ubiquitylated. Ubiquitin (Ub) is an essential 76-amino acid protein that is conserved in all eukaryotes. The polyubiquitylation reaction requires enzymes E l - 3 . The ubiquitin-activating enzyme ( E l ) activates Ub by forming unstable thioester bonds with Ub. The ubiqui tin-conjugating enzyme (E2) then transfers U b covalently to the substrate. The ubiquitin-ligase (E3) determines the specificity of the reaction by binding with the substrate and E2. One large class of E3 is cullin-dependent ubiquitin ligase ( C D L ) , which contains a catalytic core that is composed of a cullin and a R I N G finger protein, and substrate recognition modules. The cullin 'culls ' or sorts different substrates for ubiquitylation, and the R I N G finger protein stabilizes the E2-cull in interaction. Two C D L s crucial for the cell cycle progression are the Skplp-cull in-F-box protein (SCF) complex or the anaphase promoting-complex/cyclosome ( A P C / C ) . Covalent attachment of a polyubiquitin chain on lysine residues of the substrate mediates its recognition and subsequent degradation by the 26S proteasome. At various points of the cell cycle (e.g. S phase, metaphase), checkpoints exist and serve as surveillance mechanisms to ensure sequential execution of events within the cell cycle, such that the execution of a later event is dependent upon the completion of a prior event (Hartwell and Weinert, 1989). In the event of a spontaneous error or a failure to complete a step, activation of a checkpoint causes transient arrest of cell cycle 3 progression until the earlier event has been successfully completed, giving the cell more time to correct the error. Similar to the mitotic cell cycle, the fidelity of D N A transfer in the germ-line during meiosis has to be precise for the maintenance of a species. Meiosis I involves recombination (exchange of D N A ) between homologous chromosomes and their segregation, whereas meiosis II, like mitosis, involves segregation of sister chromatids. Defects in meiosis have devastating effects like miscarriage or birth defects, but a detailed discussion of this topic is beyond the scope of this thesis and w i l l not be elaborated upon further. 1.1.2 Budding yeast as a model organism to study the cell and chromosome cycle Since the cell cycle and chromosome cycle are basic and fundamental cellular processes, the mechanisms and genes involved are highly conserved among eukaryotes (reviewed in (Chan et al., 2005; Kitagawa and Hieter, 2001)). Therefore, studies in model organisms greatly facilitate the understanding of normal human biology and mechanisms of human diseases. For instance, the baker's yeast Saccharomyces cerevisiae has multiple experimental advantages, including its short life cycle and ease of genetic manipulation as either haploids or diploids, and the availability of a battery of powerful molecular and biochemical techniques. In addition, the cell cycle of S. cerevisiae can be followed by cellular morphology because it divides by budding. Therefore, the size of the daughter bud and the location of nuclear D N A allow assessment of the cell cycle stage within a population of cells. For example, cells in G l are unbudded; cells in S phase are small budded; and those in G 2 / M are large budded (Figure 1.1). Indeed, Leland Hartwell, the 2001 Nobel prize laureate in Physiology and Medicine, identified key regulators of the cell cycle in the cell division cycle (cdc) mutant collection by isolating mutants that arrest at particular stages of the cell cycle (Hartwell et al., 1974; Hartwell et al., 1970). His studies laid the foundation for our understanding of the eukaryotic cell cycle. However, due to the small size of budding yeast chromosomes, microscopic examination of chromosome behaviors has traditionally been hindered by poor resolution. Cytological 4 studies in larger eukaryotic cells have provided descriptions of spindle dynamics and chromosome movements (Rieder and Salmon, 1994). More recently, elegant molecular genetic methods such as tagging chromosomes with green fluorescent protein (GFP) fused to a repressor, which binds to operator arrays integrated at a specific location in the genome, have allowed direct observation of chromosome dynamics in wild-type and mutant yeast strains (Straight et al., 1996). Studies in different organisms thus complement each other and often reveal common, conserved cellular mechanisms. Phenotype screening based on marker stability in budding yeast has provided a powerful approach for detecting and analyzing mutants in genes that act to preserve genome structure. Several collections of yeast mutants were isolated in the last 2 decades by forward genetics (proceedings from phenotype to genotype) with the primary criterion of chromosome or plasmid loss, including the smc, mem, chl, cin and ctf collections (Hegemann et al., 1999; Hoyt et al., 1990; Kouprina et al., 1988; Larionov et a l , 1985; Larionov et al., 1987; Maine et al., 1984; Spencer et a l , 1990). MIF and CST genes are wild-type loci that induce chromosome instability when overexpressed (Meeks-Wagner et al., 1986; Ouspenski et al., 1999; Sarafan-Vasseur et al., 2002). Assays for gross chromosomal rearrangements (GCRs) have also been developed (Huang et al., 2003; Myung et al., 2001a; Myung et al., 2001b). Many of the ede mutants also exhibit increased chromosome loss and/or mitotic recombination (Hartwell and Smith, 1985). Not surprisingly, different genetic screens have led to the identification of different yet overlapping gene sets important for various steps in the chromosome cycle, including proteins that function at the kinetochores, telomeres, origins of replication, and in microtubule dynamics, sister chromatid cohesion, D N A replication, D N A repair, D N A condensation and cell cycle checkpoints. Genes identified by this strategy have often supported successful identification of functional homologues in other eukaryotes. In 1996, S. cerevisiae became the first eukaryote to have its genome completely sequenced (Bassett et al., 1996; Goffeau et al., 1996), and has subsequently served as a test-bed for the development of genomic, proteomic, bioinformatic and systems biology tools. These advances have greatly facilitated and accelerated the identification and 5 characterization of genes important for chromosome maintenance. In the following sections, key cellular components and mechanisms pertinent to chromosome segregation in the budding yeast w i l l be discussed, and major differences with other eukaryotes w i l l be highlighted. 1.1.3 Biological processes that affect chromosome stability 1.1.3.1 Kinetochores mediate the attachment with mitotic spindles Centromere is the region of D N A on a chromosome where the multiprotein kinetochore complex binds, and mediates chromosome-microtubule attachment. Interestingly, the C E N D N A size and composition vary greatly in eukaryotes. The budding yeast C E N D N A consists of only 125bp, with 3 conserved elements - the 8bp non-essential C D E I , the 78-86bp AT- r i ch C D E I I and the 25bp essential CDEIII (Fitzgerald-Hayes et al., 1982) (Figure 1.2B). The C D E elements are flanked by highly phased nucleosome arrays for >2 kb (Bloom and Carbon, 1982). In contrast, the fission yeast S. pombe C E N D N A is more similar to higher eukaryotes C E N D N A in terms of size and organization. Fission yeast C E N D N A is 35-100kb, consisting of a 4-7kb central core of non-repetitive sequence (cni) flanked by innermost repeats (imr) and outer repeats (ptr) (Figure 1.2B). Interestingly, some plants, insects and the nematode Caenorhabditis elegans contain holocentric chromosomes, where the kinetochores assemble all along the entire length of the chromosome. Mammalian C E N D N A spans 2-4Mb, and is composed of highly repeated a-satellite (171bp) D N A arrays (reviewed in (Chan et a l , 2005; Kitagawa and Hieter, 2001; Pidoux and Allshire, 2005; Sharp and Kaufman, 2003; Yanagida, 2005)) (Figure 1.2B). The difference in C E N D N A sizes maybe related to the difference in chromosome size: budding yeast chromosomes are - 1 M b , whereas human chromosomes are ~150Mb. The increase in chromosome size in mammals may require larger forces for chromosome movements. Indeed, one kinetochore of budding yeast binds only one microtubule, whereas one kinetochore of fission yeast binds 2-4 microtubules, and one kinetochore of higher eukaryotes binds 10-45 microtubules. Despite the differences in C E N D N A size and the number of microtubules binding to a kinetochore in eukaryotes, 6 many kinetochore proteins and spindle checkpoint components are conserved (see below). This raises the possibility that metazoan kinetochores are assembled from repeated subunits, where each repeat might resemble the unit module of the yeast kinetochore. With advances in experimental techniques, the list o f kinetochore-associated proteins in model organisms and human exploded in recent years (reviewed in (Chan et a l , 2005; Fukagawa, 2004; Houben and Schubert, 2003; M c A i n s h et al., 2003; Pidoux and Allshire, 2004; Yanagida, 2005)). To date, over 65 S. cerevisiae kinetochore proteins have been identified (McAinsh et al., 2003), while the number of the mammalian kinetochore proteins is predicted to be over 100 (Fukagawa, 2004). Kinetochore proteins are classified as structural or regulatory. Structural components physically bridge C E N D N A to spindle M T s (McAinsh et al., 2003). Structural kinetochore components are further classified as inner, central, and outer kinetochore proteins based on their proximity to the C E N D N A (reviewed in (Cheeseman et al., 2002b; M c A i n s h et al., 2003)). Inner kinetochore proteins interact with centromeric chromatin (e.g. Cse4p (yeast)/CENP-A (human) (see below)), while outer kinetochore mediates interaction with microtubules (e.g. the D a m l p / D A S H complex in yeast). Central kinetochore complexes (including the conserved, essential Ndc80p complex in yeast and mammalian cells (Janke et al., 2001; Wigge and Kilmartin, 2001)) link the inner and outer layers. Regulatory proteins, including motor proteins, MT-associated proteins, regulatory proteins such as Ip l l (yeast)/Aurora B (human) kinase, and spindle checkpoint components, function to regulate kinetochore-MT attachment and to co-ordinate events within the cell cycle (Biggins and Walczak, 2003; M c A i n s h et al., 2003). Centromeres of higher eukaryotes are visualized as primary constrictions in metaphase (Figure 1.2 A ) . O f particular note is the continued discovery of the conservation of individual kinetochore proteins and the overall organization of protein complexes between higher eukaryotes and yeast. These findings support the concept that the basic building blocks of kinetochores in these organisms may not be as different as first suspected based on the differences in underlying D N A sequence and size. 7 Despite that C E N D N A sequences vary among eukaryotes, C E N chromatin organization is conserved. D N A of eukaryotic chromosomes is packaged into chromatin. The most basic level of packaging involves 146bp of D N A wrapping in 1.75 turns around a nucleosome, which is composed of an octamer of core histones (2 of each of H 2 A , H 2 B , H3 and H4). A l l eukaryotes contain a centromere-specific nucleosomal structure, and the inner kinetochores contain a specialized histone H 3 , Cse4p (yeast)/CENP-A (mammals) (Stoler et al., 1995). Centromeres of fission yeast and higher eukaryotes contain transcriptionally inactive heterochromatin and involve epigenetic control. Centromeric silencing in fission yeast and higher eukaryotes depends on the R N A interference machinery (Hall et al., 2002; Volpe et al., 2002), requires the histone methyltransferase C L R 4 / S U ( V A R ) 3 - 9 , which methlyates lysine 9 of H3 (Lehnertz et al., 2003), and the heterochromatin-binding protein SWI6/HP1, which binds to the trimethylated lysine 9 of H3 (Bannister et al., 2001; Lachner et al., 2001). Mutation of either gene leads to chromosome instability (CIN) (Wang et al., 2000a). Mutation of histone deacetylase (David et al., 2003) or of its target H3 also results in improper establishment of pericentric heterochromatin and leads to aneuploidy (Wei, 1999). Outer kinetochore proteins include microtubule-associated proteins ( M A P ) (e.g. D a m l p (yeast), B i k l p (yeast)/CLIP-170 (human), B i m l p (yeast)/EBl (human)) and motor proteins (e.g. C E N P - E (human), Kip3 (yeast)ZMCAK (human), Cin8p (yeast)/BIMC (human), dynein), all of which interact with microtubules (Heald, 2000; Hoyt and Geiser, 1996). Microtubules are hollow cylindrical tubes consisting of a heterodimer of a and P tubulins. Microtubules are polar molecules, with a dynamic 'plus' end and a 'minus' end at the microtubule organizing centre ( M T O C ) . The plus end, which w i l l capture kinetochores, undergoes rapid growth and shrinkage by polymerization and depolymerization, respectively, switching between these two states in events called "catastrophes" and "rescues" (Maddox et al., 2000). The minus side of microtubules nucleates at the M T O C , which is called spindle pole body (SPB) in yeast and centrosome in higher eukaryotes. The centrosome is made up of 2 barrel-shaped centrioles surrounded by a matrix of pericentriolar material. The S P B is a disk-shaped 8 structure made up of three plaques. Besides kinetochore microtubules, there are 2 other types of microtubules: (1) interpolar microtubules which project towards the spindle midzone and interact with chromosome arms or overlap with microtubules emanating from the other pole, through the interaction of micro tubule-associated factors and motors; and (2) astral (cytoplasmic) microtubules which project towards the cortex and are instrumental in spindle orientation and positioning (Wittmann et al., 2001). Spindle disassembly is necessary for cytokinesis, and it is thought to occur by depolymerization of the interpolar microtubules from their plus ends. Unlike other eukaryotes, the yeast nuclear membrane does not break down during mitosis, so the SPBs remain embedded within the nuclear membrane (Hoyt and Geiser, 1996). The nuclear and cytoplasmic faces of the SPB are linked by a central plaque embedded in the nuclear envelope. During S phase, the SPB duplicates, while in prometaphase and metaphase, the SPBs separate (Winey and O'Toole, 2001). Initially, yeast sister kinetochores are both attached to one SPB, the "o ld" S P B , and this type of kinetochore-microtubule attachment is called 'syntelic attachment' (Tanaka, 2002). The yeast kinetochores are positioned near the SPBs throughout the cell cycle (Jin et al., 2000). On the other hand, kinetochore-microtubule interactions in mammalian cells only take place during mitosis after the nuclear envelope breaks down. A t the onset of mitosis, rapidly growing and shrinking microtubules probe the cytoplasm for kinetochores, in a 'search and capture' mechanism that is stochastic and error-prone in nature (Figure 1.3). During the early stage of chromosome orientation, usually only one sister is attached to a pole, and this kind of attachment is called 'monotelic attachment' (Figure 1.4C). Recently, mono-oriented chromosomes in mammalian cells were shown to laterally interact with kinetochore microtubules of bi-oriented chromosomes, which serve as tracks to help the mono-oriented chromosomes to 'hitch a hike' to the spindle equator (Kapoor et al., 2006). This interaction and chromosome movement is dependent on the mammalian kinesin-7 family member C E N P - E . This cooperative process increases the likelihood that mono-oriented chromosomes w i l l achieve bi-orientation because the middle of the spindle is rich in microtubules extending from the opposite spindle poles. If 9 both sister kinetochores attach to the same spindle pole (syntelic attachment, Figure 1.4B), the Ipllp/Aurora B kinase facilitates re-orientation by phosphorylating kinetochore targets (see below). The spindle checkpoint signal is maintained until sister kinetochores of all chromosomes bi-orient to opposite spindle poles, and this manner of attachment is called 'amphitelic attachment' (FigureT.4A). When bipolar attachment is achieved, tension generated at the kinetochore by forces from opposite spindle poles and cohesin (see below) has a stabilizing effect on kinetochore microtubules (Ault and Nicklas, 1989; K i n g and Nicklas, 2000). Another type of attachment error occurs when a single kinetochore becomes attached to microtubules from both spindle poles, which is called 'merotelic attachment' (Figure 1.4D). However, this defect is not detected by the spindle checkpoint (Cimini et al., 2001). Nevertheless, merotelic attachment rarely cause chromosome missegregation in mammals because the kinetochores usually make enough bipolar attachments to pull the sister chromatids to opposite poles. Interestingly, budding yeast chromosomes do not undergo congression to the metaphase plate (O'Toole et a l , 1999), but form two lobes that lie on either side of the spindle midzone (Goshima and Yanagida, 2000; He et al., 2000). Both yeast and mammalian centromeric chromatin undergo transient separation before cohesion degradation at anaphase (Goshima and Yanagida, 2000; He et a l , 2000; Shelby et al., 1996; Tanaka et al., 2000). 1.1.3.2 M i t o t i c s p i n d l e c h e c k p o i n t Spindle checkpoint components present at the kinetochore in turn monitor M T attachment and/or tension and sense the completion of metaphase, when bi-polar attachment of all chromosomes has been achieved (Lew and Burke, 2003; Tanaka, 2002). BUB1 and 3 (budding uninhibited by benzimidazole) and MAD1, 2, and 3 (mitotic arrest deficient) are checkpoint genes first identified in yeast in genetic screens that looked for mutants that fail to detect kinetochore-microtubule attachment errors caused by microtubule-depolymerizing drugs. A s a.result, these mutants do not arrest before anaphase despite the presence of chromosomes not properly attached to the spindle, which leads to increased chromosome missegregation and increased sensitivity to 10 microtubule-depolymerizing drugs (Hoyt et al., 1991; L i and Murray, 1991; Weiss and Winey, 1996). These findings lead to the description of a spindle checkpoint pathway that detects kinetochores that are not attached to microtubules or are not under tension (Yu, 2002). Even a single unattached kinetochore can delay segregation of already aligned chromosomes (Rieder et a l , 1995). Mammalian and yeast checkpoint proteins were shown to localize to kinetochores that have not yet attached to the mitotic spindle (reviewed in (Cleveland et al., 2003)). The exact sequence of spindle checkpoint sensing and signaling is not completely understood, but probably involves amplification of diffusible signals. BTJB1 is a kinase that is known to phosphorylate M A D 1 and B U B 3 , and B U B 3 in turn binds to and activates B U B 1 . B U B R 1 is the mammalian homolog of yeast Mad3p, but it has evolved to contain a kinase domain that is not present in Mad3p. Localization of B U B R 1 to the kinetochore is dependent on its interaction with B U B 3 , and B U B R 1 is postulated to act as a mechanosensor. Interaction of B U B R 1 with the kinesin-like protein C E N P - E stimulates B U B R 1 kinase activity is stimulated (Chan et al., 1999; Mao et a l , 2003). C E N P - E is thought to act as a tension sensor and increases the efficiency of microtubule capture; it is able to activate the spindle checkpoint in the presence of mono-oriented chromosomes (reviewed in (Compton, 2006)). Yeast M p s l p (monopolar spindle) was originally identified to be involved in SPB duplication, but was later found to have a role in the spindle checkpoint by phosphorylating M a d l p and recruiting other checkpoint components to unattached kinetochores (Weiss and Winey, 1996 (Winey and Huneycutt, 2002). M A D 1 binds to and recruits M A D 2 to the kinetochore (Chen et a l , 1999). M A D 2 binds to C D C 2 0 / S L P 1 / F I Z Z Y / P 5 5 , the substrate specificity factor of the A P C / C , inhibiting its ubiquitin ligase activity (Yu, 2002). Interestingly, M A D 1 and C D C 2 0 contain a similar domain to interact with M A D 2 , so their interaction with M A D 2 is mutually exclusive (Luo et al., 2002). M A D 1 - M A D 2 binding may catalyze a conformational change in M A D 2 so that it is compatible for C D C 2 0 binding. M A D 1 hyperphosphorylation may be required to dissociate M A D 2 from M A D 1 for C D C 2 0 binding. B U B R 1 also directly binds C D C 2 0 and A P C / C components (Chan et al., 1999). 11 In addition, B U B R 1 forms a stoichiometric mitotic checkpoint complex ( M C C ) with B U B 3 , M A D 2 and C D C 2 0 (Sudakin et al., 2001). Vertebrate M A D 1 and M A D 2 are displaced from kinetochores with proper M T attachments. M A D 2 phosphorylation maybe involved in silencing of the checkpoint (Wassmann et al., 2003). The microtubule motor dynein has been implicated in the highly dynamic turnover of checkpoint components and in checkpoint silencing. Checkpoint proteins like M A D 2 and B U B R l a r e thought to be released from the kinetochore through dynein-dependent transport via spindle M T s , and also via direct release of proteins ((Howell et al., 2001); reviewed in (Chan et al., 2005)). Z w l O , Zwi l ch and Rod which were first identified in Drosophila are also found in higher eukaryotes but not in yeast. They form the R Z Z complex that is required for dynein localization ((Wojcik et a l , 2001); reviewed in (Karess, 2005)). Even when kinetochore-microtubule connections are intact, a lack of tension at the kinetochore can activate the checkpoint (Stern and Murray, 2001). The Aurora B/IPL1 kinase works with INCENP/SLI15 as a tension sensor to promote turnover of syntelic attachments; it works by destabilizing kinetochore-microtubule attachments through phosphorylation of the microtubule-destabilizing mitotic centromere-associated kinesin ( M C A K ) in vertebrate, analogous to D a m l p in yeast (Andrews et al., 2004; Cheeseman et al., 2002a; He et al., 2001; Kang et al., 2001; Lan et al., 2004; Stern and Murray, 2001; Tanaka, 2002). Aurora B may also phosphorylate B U B R 1 and M P S 1 for checkpoint signaling (Biggins and Murray, 2001). Aurora B / I P L 1 , INCENP/SLI15 and SURVrvrN/BIRl are chromosomal passenger proteins that dynamically appear first in the inner centromere region between sister kinetochores, then move onto the elongating spindle in anaphase, and finally concentrate at the spindle midzone. Some structural kinetochore proteins (e.g. the N D C 8 0 complex) are also required for a functional checkpoint, which may first require the assembly of a functional kinetochore (Gardner et a l , 2001; He et al., 2001; Janke et al., 2001). 12 1.1.3.3 S i s t e r c h r o m a t i d cohes ion A t the end of M phase in fission yeast and metazoan cells, or during late G l in budding yeast, the cohesin complex, the "molecular glue" that holds sister chromatids together is loaded onto unreplicated D N A by the loading complex (SCC2, SCC4) (Ciosk et al., 2000). Cohesin is composed of 4 subunits: S C C l ' / M C D 1 / R A D 2 1 , SCC3/IRR1 (SAI and SA2 variants in human), S M C 1 and S M C 3 (structural maintenance of chromosomes). S M C 1 and S M C 3 contain globular ends with a hinge dimerization domain and a head ABC- type ATPase domain, and a coiled-coil domain (Losada et al., 1998; Michaelis et a l , 1997) (Figure 1.5A). They form intra-molecular coiled coils by folding back on themselves, forming rod shaped proteins with the globular ATPase head at one end and the heterodimerization domain at the other (Haering et al., 2002). S M C 1 and S M C 3 dimerize through the hinge domain. The C-terminal and N-terminal ends of SCC1 bind to the head region of S M C 1 and S M C 3 , respectively, and S C C 3 binds to the complex through SCC1 (Haering et al., 2002; Haering et al., 2004). A T P hydrolysis is needed for cohesin loading onto D N A . , Cohesion is first established while sister chromatids are replicated in S phase and is maintained until anaphase. A working model for cohesin is that it forms a ring structure that wraps around the sister chromatids in a topological association (Figure 1.5B). Cohesion is established along the whole length of chromosomes, but is concentrated at the pericentromeric regions, spanning 50-60 kb, and at convergent transcription sites (intergenic AT- r i ch region) (Glynn et al., 2004). Kinetochores stimulate the recruitment of cohesin, but this ability is not necessarily dependent on the centromere sequence per se (Megee et al., 1999; Weber et al., 2004). Cohesin recruited by the kinetochore may move to flanking regions, or kinetochores may influence surrounding chromatin to recruit cohesin. In S. pombe, the enrichment of cohesin at peri-centromeric regions depends on the binding of the H P 1-like protein SWI6 to histone H3 that is trimethylated on lysine 9 by C L R 4 (Bernard et al., 2001; Nonaka et al., 2002). In contrast, the nucleosome-remodeling complex R S C has been implicated in the establishment of chromatid arm cohesion only (Huang et a l , 2004; Huang and Laurent, 2004). Cohesin may also be redistributed to different places during transcription. 13 The establishment of cohesion is not completely understood, but is thought to require the acetyltransferase E S C 0 1 / C T F 7 (Skibbens et al., 1999; Toth et a l , 1999), a variant replication factor C (RFC-CTF18 , C T F 8 , D C C 1 ) (Mayer et al., 2001; Mayer et al., 2004), a polymerase a-interacting protein C T F 4 (Hanna et al., 2001; Petronczki et a l , 2004), M R E 1 1 (Warren et al., 2004a), and the helicase C H L 1 (S, 2000; Skibbens, 2004). Additionally, PDS5 is required to maintain cohesion at centromere proximal and distal sequences (Hartman et al., 2000). Cohesion sterically forces a back-to-back orientation to sister centromeres and promotes bi-orientation (Tanaka et al., 2000). Cohesion resists the force exerted by spindle microtubules emanating from opposite spindle poles on sister kinetochores, thereby generating tension (He et al., 2000). Cohesin is also recruited to double strand break (DSB) sites in S / G 2 / M , and this recruitment requires S C C 2 (Strom et al., 2004; Unal et al., 2004). Damage-induced cohesion may be important for D S B repair by holding broken ends close to homologous sequences, thereby facilitating homologous recombination. In budding yeast, cohesin remains associated with whole chromosomes until anaphase, whereas in mammalian cells, cohesins dissociate from chromosome arms in prophase in a Polo-like kinase 1 (PLKl)-dependent manner (Hauf et al., 2005). P L K 1 is activated as C D K levels rise at the onset of M phase, and P L K 1 promotes arm cohesin dissociation through phosphorylation of the SCC3-l ike subunits, SA1 and S A 2 . However, cohesin at centromeres persists until anaphase. This retention is dependent on shugosin (SG01/MEI-S332 in Drosophila) (Hauf et al., 2005). Protein phosphatase 2 A (PP2A) associates with S G O l and is required for protection of centromeric cohesion by dephosphorylation of cohesin (Kitajima et al., 2006; Riedel et al., 2006). Before the onset of anaphase, an inhibitory chaperone, securin/PDSl, binds to the separase/ESPl, thereby inhibiting it but also priming its activity, possibly by promoting its nuclear localization or protecting it from degradation (Ciosk et al., 1998). When all chromosomes align at the metaphase plate, A P C / C C D C 2 ° targets securin for degradation (Cohen-Fix et al., 1996). Separase is then released, and it cleaves the cohesin subunit S C C 1 , leading to the 14 breakdown of cohesion and the beginning of anaphase where sister chromatids move to opposite spindle poles (Michaelis et al., 1997; Uhlmann et al., 1999; Uhlmann et al., 2000) (Figure 1.6). P L K 1 phosphorylation of SCC1 enhances its cleavage by separase. A P C / C also degrades cyclins, lowering C D K activity and promoting exit from M phase. In meiosis, sister chromatids segregate to the same pole during meiosis I and b i -orientation of sister chromatids is suppressed by monopolin ( M A M 1 , C S M 1 , and LRS4) (Toth et al., 2000). The SCC1 subunit of the meiotic cohesin complex is replaced by R E C 8 . The paired homologous chromosomes are held together at chiasmata that are formed during recombination. In meiosis I, R E C 8 present on chromosomal arms is cleaved, thereby resolving chiasmata, while centromeric R E C 8 is protected during meiosis I by S G O l . In anaphase I, homologous chromosomes segregate to opposite poles (Buonomo et al., 2000; K l e i n et al., 1999). Meiosis II resembles mitosis with sister chromatids segregating to opposite poles. 1.2 Chromosome instability (CIN) and cancer 1.2.1 Aneuploidy is a hallmark of cancer Two types of genetic instability are observed in cancers: (1) instability at the nucleotide level, especially at microsatellite repeats ( M I N , microsatellite instability) and (2) instability involving whole chromosomes or large portions of chromosomes (CIN, chromosomal instability). The majority of solid tumors exhibit genomic instability at the chromosomal level (Rajagopalan et al., 2003) (e.g. 85% colon cancers exhibit C L N and 15% exhibit M I N ) , and the occurrence of M I N and C L N usually does not overlap. M I N tumors exhibit a 1000-fold increase in point mutation rate, in particular accumulation of length alterations in simple repeated sequences (units of l-3bp), whereas C I N tumors exhibit increased rates of chromosome missegregation, leading to the generation of aneuploid cells. Changes in whole chromosome number or structural rearrangement of chromosomes are commonly observed in tumors (Cahill et al., 1998; Rajagopalan et al., 2003). Large-scale chromosomal gains or losses can be detected by flow cytometry in a 15 cell population, and chromosomal rearrangements (>1 Mb) in an individual cell can be revealed by comparative genomic hybridization (CGH) , multiplex fluorescence in situ hybridization (M-FISH) or spectral karyotyping ( S K Y ) . Aneuploidy and chromosomal rearrangements may play a role in tumor progression by causing an imbalance in the dosage of many genes at once. For instance, chromosome loss or partial chromosomal deletion results in loss of heterozygosity (LOH) , which can lead to reduced expression of tumor suppressor genes located in the region, or uncover recessive mutations in the remaining allele. Chromosome gain or partial amplification can amplify oncogenes within the region. Overexpression of oncogenes and/or reduced expression of tumor suppressor genes would create a growth advantage through increased proliferation or reduced cell death, and result in clonal expansion. This scenario would repeat for each new growth-promoting mutation, and constitutes the basis of the theory of multi-step carcinogenesis (Boland and Ricciardiello, 1999). Identifying recurrent chromosomal aberrations at specific loci in cancer cells may provide clues for the identification of oncogenes and tumor suppressor genes. A n average cancer of the colon, breast, pancreas or prostate loses 25% of its alleles on a chromosome (Lengauer et al., 1998). Some primary breast cancers exhibit >20 regions with L O H when analyzed with microsatellite markers (reviewed in (Loeb, 2001)). Interestingly, analysis of polymorphic markers in 5 chromosomes in colorectal cancer cell lines indicated that mechanisms underlying L O H were chromosome-specific. Partial losses were predominant for some chromosomes, while whole chromosome losses were responsible for others. For partial loss, gross chromosomal rearrangement (GCR) , not mitotic recombination, was the predominant mechanism. For whole chromosome loss, mitotic nondisjunction was responsible, and reduplication of the remaining chromosome was followed in some cases. L O H occurs at different frequency at different regions of each chromosome, implying that L O H is coupled with clonal selection for loss of tumor suppressor genes (Thiagalingam et al., 2001). 16 1.2.2 Relationship between a state of aneuploidy and an increased rate of chromosome instability (CIN) Observation of a state o f aneuploidy in cancer cells does not directly imply an increased rate o f C I N , because aneuploidy may be caused by factors other than C I N . For instance, aneuploidy can be caused by chromosome missegregation in a single cell division (at a normal rate), followed by clonal expansion of the aneuploid cell due to some selective advantage; or, the survival of an aneuploid cell can result from a defect in the apoptotic pathway. However, an analysis of 98 aneuploid gastric tumors by F I S H and flow cytometry showed intratumoral variations in chromosome copy number; this population heterogeneity suggests that aneuploidy is associated with C I N (Furuya et al., 2000). In another study, 16 out of 25 pancreatic carcinomas showed karyotypically related clones, signifying monoclonal origin and evolutionary variation (Gorunova et al., 1998). Similarly, F I S H analysis of aneuploid colorectal cancer cell lines for 6-7 generations showed that losses or gains of chromosomes occurred at >10"2 per chromosome per generation, which is 10-100 times more often than in diploid cancers of the same histological subtype (Lengauer et al., 1997). These observations are consistent with the hypothesis that aneuploidy in cancers is caused by C I N . To further delineate the relationship between C L N and aneuploidy, Lengauer et al. introduced an extra chromosome into a diploid cell line and fused two diploid lines to artificially create aneuploid cell lines. These lines, unlike natural C L N tumor lines, did not display C I N , suggesting aneuploidy per se does not cause C I N (Lengauer et al., 1997). 1.2.3 CIN occurs at early stage of cancer, and can be a driving force in tumorigenesis The timing of C L N occurrence during tumorigenesis, and the role of C L N in tumorigenesis have been highly debated. One hypothesis postulates that for a cancer cell to accumulate the 6-10 genetic alterations required for its proliferation and survival, it must be genetically unstable, thereby suggesting that genetic instability occurs at the early stage of cancer, and represents an important step in the initiation and/or progression 17 of tumorigenesis (Davies et al., 2002; Hartwell et al., 1997; Parsons et al., 2005). In support of this hypothesis, aneuploidy has been observed in small benign colorectal tumors and uterine leiomyomas (El-Rifai et al., 1998), and >90% of early colorectal adenomas studied (1-3 mm in size) have allelic imbalance (Bardi et al., 1997; Bomme et al., 1998; Lengauer et al., 1998; Shih et al., 2001). The prevalence of aneuploidy in benign colorectal tumors is less than that in cancers, but the deviations from a normal karyotype increase as the tumors enlarge in size (Bardi et al., 1997; Shih et al., 2001). Aneuploidy is associated with poor prognosis and correlates with the severity of the disease (Rajagopalan and Lengauer, 2004a). C I N may serve as an engine of both tumor progression and heterogeneity (Jallepalli and Lengauer, 2001; Vogelstein and Kinzler, 2004). 1.2.4 G e n e t i c basis o f C I N i n cance r 1.2.4.1 C a n c e r - p r o n e s y n d r o m e s Germline mutations causing genomic instability, particularly in genes involved in D N A damage recognition and repair, are now recognized as being important predisposing conditions for cancer (reviewed in (Hoeijmakers, 2001; Levitt and Hickson, 2002; Vogelstein and Kinzler, 2004)). For instance, the less common M I N phenotype in colorectal cancer was first described in 1992. The similarity o f phenotype in MUST tumor cells and D N A mismatch repair ( M M R ) mutants in yeast and E. coli rapidly led to the identification of mutations in M M R genes (based on a candidate gene approach) in hereditary nonpolyposis colon cancer ( H N P C C ) , which account for 3% of colon cancer (Fishel et al., 1993; Leach et a l , 1993; Papadopoulos et al., 1994; Strand et al., 1993). Another example is provided by xeroderma pigmentosum (XP) patients whose cells have defects in the nucleotide excision repair (NER) pathway and high mutation rates due to pyrimidine dimers; these patients develop skin cancers at high rates. Table 1.1 summarizes germline mutations in genes involved in maintaining genomic integrity that are known to underlie cancer-prone syndromes, and lists the function/pathway of the encoded protein, evolutionary conservation between yeast and human genes, and the 18 mode of inheritance of the diseases. These "caretaker" genes, unlike conventional oncogenes and tumor suppressor genes which directly control cell birth and death, affect the integrity of the genome and control the mutation rate. Interestingly, despite the ubiquitous expression of these genome maintenance proteins, mutations in these genes lead to tissue-specific tumor predispositions. In addition, somatic mutations in these same genes may not occur in sporadic tumors of the same type (Sieber et a l , 2003). In fact, although tumor types from one specific organ have a tendency to share mutations in certain genes or in different genes within a single pathway, they rarely have uniform genetic alterations, demonstrating the heterogeneous nature of cancer (Boland and Ricciardiello, 1999). In contrast to M I N , the genetic basis of the commonly observed C L N in sporadic cancers is not well understood. Cytologically, many cancer cells exhibit aberrant cell architecture, including abnormal centrosomes, multipolar spindles, and breakage-fusion-bridge cycles (Gisselsson, 2003; Saunders et al., 2000). Intuitively, C I N , and therefore aneuploidy, can be caused by errors in chromosome segregation. Many cellular mechanisms are responsible for proper chromosome transmission, such as D N A replication, sister chromatid cohesion, centrosome duplication and segregation, kinetochore-microtubule attachment, mitotic spindle checkpoint, D N A condensation, D N A repair and cytokinesis (Figure 1.7). One approach to determine the genetic basis of CLN in tumors is to identify mutations in genes known to be important for chromosome segregation in human cells, or in human homologues of C I N genes discovered in model organisms, which serve as cross-species candidate C L N genes. 1.2.4.2 M u t a t i o n s i n m i t o t i c s p i n d l e c h e c k p o i n t Many C L N genes were originally identified and studied in model organisms such as yeast, and later found to have conserved functions and cancer relevance, including the genes listed in Table 1.1 and mitotic spindle checkpoint components. One important class of cancer relevant genes first discovered in yeast are the spindle checkpoint proteins, which monitor kinetochore-MT attachment and alert the cell to potential chromosome 19 segregation errors by specifically binding to kinetochores that have not attached to M T s . BUB1 and BUB IB (encoding BTJBR1) are mutated in colorectal tumors and several other cancer types at a low frequency (Cahill et al., 1998; Gemma et al., 2000; Ohshima et al., 2000; R u et al., 2002; Shichiri et a l , 2002) (Table 1.2). Epigenetic silencing through promoter hypermethylation of BUB1 and BUB IB has also been found in aneuploid colon carcinoma (Shichiri et al., 2002). The recent report that germline biallelic mutations in the spindle checkpoint gene, BUB1B, is associated with mosaic variegated aneuploidy ( M V A ) and inherited cancer predispositions strongly supports a causal link between C L N and cancer development (Hanks et al., 2004). Human homologues of other yeast mitotic checkpoint proteins (MAD1, 2, 3 and BUB1, 3) then became candidate C I N genes and were subsequently tested for mutations in tumors. MAD2 is mutated in gastric cancers ( K i m et al., 2005), and downregulated in cancer cell lines ( L i and Benezra, 1996; Miche l et al., 2001; Wang et al., 2002b). However, no mutation in other spindle checkpoint genes was found (Cahill et al., 1999), suggesting they could be altered by misregulation or that other C L N genes could be affected. For example, MAD1 binds to the Tax oncoprotein from the human T-cell leukaemia virus type 1 ( H T L V - 1 ) and prevents MAD2 activation (Jin et al., 1998). L A T S 1 / W A R T S (large tumor suppressor homologue 1), a paralog of B U B 1 , is a mitosis-specific serine/threonine kinase that interacts with M O B 1 (Mpsl -One binder) and may play a role in the mitotic exit network, cytokinesis, and coordination between cell proliferation and apoptosis. Downregulation of L A T S 1 has been found to contribute to tumor formation (Bothos et al., 2005; Hergovich et al., 2006; L a i et al., 2005; Yang et al., 2004). The recent survey of C L N colorectal tumors for mutations in 100 human homologues of C I N genes identified in yeast and flies, including 6 kinetochore/spindle checkpoint proteins, represents a stunning proof of principle: Wang et al. identified.5 new C L N cancer genes, including the kinetochore/spindle checkpoint genes Rod, ZwlO, and Zwilch (Table 1.2), which together account for - 2 % of the mutational spectrum in colorectal cancers (Wang et al., 2004b). These proteins function together as the R Z Z complex to recruit the dynein-dynactin complex and M A D 1 - M A D 2 to the kinetochore. The R Z Z complex is thought to have a role in spindle checkpoint 20 activation and inactivation (reviewed in (Karess, 2005)). The infrequent mutation rate in spindle checkpoint genes raises the possibility that C L N in cancer cells could be caused by mutation of any one of many genes involved in chromosome segregation, including other kinetochore proteins. Because of the large number of candidate genes that could be mutated to give a C L N phenotype, the frequency of a particular mutation may be low, as is observed for the spindle checkpoint genes. Interestingly, analysis of the mitotic index of cancer cell lines in response to microtubule-disrupting reagents showed that the mitotic spindle checkpoint is often impaired, but not completely absent (Gascoyne et al., 2003; Saeki et al., 2002; Takahashi et al., 1999). Absence of the checkpoint proteins M A D 2 , B U B 3 , or B U B R 1 in mice and C. elegans yields early embryonic lethality (Babu et al., 2003; Baker et al., 2004; Dai et al., 2004; Kalitsis et al., 2000; Kitagawa and Rose, 1999; Kops et al., 2004; Miche l et al., 2001). Mouse models of defective checkpoints, where a checkpoint component is reduced in concentration, result in a small increase in cancer susceptibility. For example, mice heterozygous for BUB IB or BUB 3 are more prone to colorectal or lung tumors after challenge with carcinogen (Babu et al., 2003; Dai et al., 2004). R A E 1 , which has homology to B U B 3 , mediates nuclear export of m R N A through nuclear pores during interphase and binds to B U B 1 at kinetochores during mitosis. Heterozygous RAE1 mice have increased aneuploidy and develop lung tumors at an increased rate (Babu et al., 2003). 28% of heterozygous MAD2 mice develop lung tumors at high rates after long latencies (Babu et al., 2003; Baker et al., 2004; Da i et al., 2004; Kitagawa and Rose, 1999; Kops et al., 2004; Miche l et al., 2001). Additionally, some tumor suppressor genes affect the levels of checkpoint components at the transcript level. For example, B R C A 1 regulates MAD2 transcript levels directly by binding to its promoter, and mouse cells that express mutant BRCA1 have decreased expression of M A D 2 , B U B 1 , B U B R 1 and Z w l O (Wang et al., 2004a). A single nucleotide polymorphism in MAD1 that affects M A D 2 binding and recruitment of M A D 2 to kinetochores has recently been found in a breast cancer cells (Iwanaga et al., 2002). These results suggest that biallelic expression of 21 checkpoint components is important for their function, and a weakened checkpoint might facilitate tumorigenesis. 1.2.4.3 M i s r e g u l a t i o n o f k i n e t o c h o r e p ro t e in s Mutations in genes encoding structural kinetochore proteins have not yet been identified in cancer cells, possibly because most have not been examined. Since 5 out of 8 (BUB1, BUBR1, Rod, ZwlO, Zwilch, CDC4, MRE11A, and Ding) C I N genes known to be mutated in C L N colon cancers encode kinetochore or spindle checkpoint proteins (Cahill et al., 1998; Rajagopalan et al., 2004; Wang et al., 2004b), the kinetochore offers a logical choice for mutational testing. Furthermore, the -100 predicted human genes that encode kinetochore components comprise a large mutational target that could be mutable to a C L N phenotype (Fukagawa, 2004). For example, kinetochore proteins constitute a significant portion of the collection of chromosome transmission fidelity (ctf) mutants identified in a classical genetic screen in yeast (9 out of the 24 C T F genes cloned and characterized to date; see Table 2.1) (Spencer et al., 1990). Systematic mutational analysis of kinetochore genes in various cancers would shed light on the frequency of specific mutations in kinetochore genes and their potential role in tumorigenesis. On the other hand, expression studies have suggested a correlation between overexpression of several kinetochore proteins and cancer (Table 1.2). C E N P - A is overexpressed and mistargeted in colorectal cancer tissues (Tomonaga et al., 2003). Overexpressed C E N P - A localizes to the entire chromosome and dissociates from native centromeres. This causes a subset of kinetochore proteins to be recruited to non-centromeric chromatin, leading to ectopic formation of pre-kinetochore complexes, potentially depleting some kinetochore components, thereby disrupting the native centromere-kinetochore complex and causing C I N (Van Hooser et al., 2001). Another inner kinetochore protein, C E N P - H , which is important for kinetochore organization, is also upregulated in colorectal cancer tissues (Tomonaga et al., 2005). Transfection of a C E N P - H expression plasmid into diploid cell lines induces aneuploidy and increases the incidence of aberrant micronuclei, suggesting that upregulation of C E N P - H can lead to a 22 CLN phenotype. In addition, Aurora B (ALM-1) and I N C E N P , two chromosome passenger proteins that localize to the kinetochore from prophase to metaphase and to the mitotic spindle in cytokinesis, are upregulated in tumor cell lines (Adams et al., 2001; Sorrentino et a l , 2005; Tatsuka et al., 1998). Aurora B phosphorylation is required for chromosome condensation, controlling M T dynamics including destabilizing syntelic M T attachments to kinetochores, and regulation of cytokinesis (reviewed in (Giet et al., 2005)). Aurora B-overexpressing cells exhibit C L N and contain multinuclei, and injection of these cells into nude mice induces tumor growth (Ota et al., 2002; Sorrentino et al., 2005). Conversely, a block of Aurora B expression increases the latency period and reduces the growth of thyroid anaplastic carcinoma cells (Sorrentino et al., 2005), supporting a causative link between Aurora B expression and cancer initiation or progression. Similarly, overexpression of C E N P - F (mitosin) correlates with tumor proliferation and metastasis; hence, C E N P - F is suggested to be a potentially valuable proliferation marker for diagnosis and prognosis (Clark et al., 1997; de la Guardia et al., 2001; Erlanson et al., 1999; Esguerra et al., 2004; L i u et al., 1998; Shigeishi et al., 2005). C E N P - F is a cell cycle-regulated protein that associates with the outer kinetochore in M phase and is rapidly degraded upon completion of mitosis. It associates preferentially with kinetochores of unaligned chromosomes, and may play a role in the spindle checkpoint (Chan et al., 1998; Yang et al., 2005; Yang et al., 2003). The evidence above suggests that overexpression of kinetochore components may contribute to tumor progression by driving CLN. Stoichiometric expression of kinetochore components may be important for functional kinetochore assembly and the dosage may be crucial for spindle checkpoint signalling. However, it is possible that overexpression is a consequence rather than a cause of dysfunctional cell cycle regulation in carcinogenesis. To delineate the causal relationship between kinetochore protein mutation/misregulation and cancer development, further functional studies must be performed in diploid cell lines or mouse models to investigate whether kinetochore mutation/misregulation leads to C L N or cellular transformation. 23 1.2.4.4 Additional examples of mutations in genes involved in chromosome segregation Systematic mutation testing of candidate C I N genes in colorectal cancer also identified somatic mutations in CDC4, MRE11A, and Ding (Rajagopalan and Lengauer, 2004b; Wang et a l , 2004b). Known mutations together account for only - 2 0 % of the CLN mutational spectrum of colon cancer. C D C 4 is a conserved F-box protein that functions in the S C F E3 ubiquitin ligase, involved in regulating the G l - S cell cycle checkpoint. Cyc l in E , an oncoprotein and a known target of C D C 4 in mammalian cells, is overexpressed when C D C 4 is defective (Rajagopalan et al., 2004; Strohmaier et al., 2001). M R E 1 1 A is involved in sister chromatid cohesion and D S B repair. Germline mutations in M R E 1 1 A are responsible for ataxia telangiectasia-like syndrome (see Table 1.1). Ding is uncharacterised, but its C-terminus is homologous with the yeast securin, PDS1 (Wang et al., 2004b). The human securin, also known as pituitary tumor transforming gene 1 (PTTG1), is overexpressed in some cancers and its expression level is correlated with the invasiveness (Pei and Melmed, 1997; Zhang et al., 1999; Zou et al., 1999). Aurora A kinase ( S T K 1 5 / B T A K ) at the centrosome is amplified and overexpressed in cancers (Zhou et al., 1998), and is associated with centrosome amplification, tetraploidization and aneuploidy. Indeed, approximately 80% of invasive tumors show centrosome abnormalities in size and number, and a significant proportion of solid tumors are tetraploid, such as in Barrett's oesophagus and ulcerative colitis (Rajagopalan and Lengauer, 2004a). Familial polyposis coli (FAP) patients and over 85% of colorectal tumors have somatic mutations of adenopolyposis coli (APC) (see Table 1.1), and this is the earliest event in sporadic colorectal tumor. Most APC mutations lead to loss of the C-terminal domain that interacts with microtubules (and binds components of checkpoint), failure to degrade beta-catenin, and have been postulated to contribute to C L N (Fodde et al., 2001a; Fodde et al., 2001b; Green and Kaplan, 2003; Kaplan et al., 2001). However, some cells with APC mutations undergo polyploidization in whole-genome increments instead of 24 aneuploidy, and some M I N cell lines with APC mutations remain diploid, so the exact significance of APC mutation in C L N is still unclear (Fodde et al., 2001b). The genetic basis for C L N is just beginning to be understood (Figure 1.8). The daunting task of screening the remaining hundreds of candidate C I N genes lies ahead. Systematic mutation screening in candidate genes in signalling pathways have yielded success (Davies et al., 2002; Parsons et al., 2005). However, mutations in C I N genes could be functional (leading to CLN) or merely "passenger" mutations that accompany tumorigenesis. The prevalence of point mutations in sporadic C I N colorectal cancers was determined to be approximately one nonsynonymous somatic change per M b of tumor D N A , which is consistent with a rate of mutation in normal cells (Wang et al., 2002a). These data suggested that most sporadic C L N colorectal cancers do not display M I N or instability at the nucleotide level (Wang et al., 2002a). These results have significant implications for the interpretation of somatic mutation observations in candidate tumor-suppressor genes, suggesting these are likely to be of functional relevance. 1.2.4.5 T h e r a p e u t i c i m p l i c a t i o n s Designing effective therapeutics for cancer w i l l rely first on understanding the genetic basis of cancer, including the cause o f C L N and its contribution to human cancers. This w i l l involve identifying the mutational spectrum and analyzing expression profiling of candidate C L N genes, and determining their functional consequence. Such knowledge could have several important practical applications. First, it would allow sub-classification of tumors based on the specific C L N gene mutation or misregulation, which could have implications for improved diagnostics, prognosis, or predictions of response to therapy. For example, overexpression of either Aurora A or B kinases causes CLN. Inhibition of aurora kinases results in a 98% reduction in tumor volume in nude mice injected with human leukemia cells (Harrington et al., 2004). One complication in studying cancer is that cancer is a heterogeneous disease, with many different genes mutated at low frequencies in different tumors and in sub-population of cells within individual tumors. Genetic instability is expected to contribute to heterogeneity. 25 However, i f a defined subset of C I N genes represents the major C L N mutational targets in cancer, they may provide a rationale for therapeutic design to selectively k i l l tumor cells carrying C L N mutations (Hartwell et al., 1997; Hartwell and Weinert, 1989). While C I N may be important in the development of a tumor, understanding the genetic and phenotypic differences between C I N tumor cells and normal cells may define an "Achilles heel" in C L N tumors (relative to adjacent normal tissue), allowing selective ki l l ing of tumor cells (Hartwell et al., 1997; Hartwell and Weinert, 1989). One approach is to identify drug targets that are specifically present and essential for the viability of cancer cells, but are not present in normal cells. For instance, fusion oncoproteins are generated by cancer-associated chromosomal translocations, such as the . fusion of the breakpoint cluster region (BCR) with Ableson murine leukemia viral oncogene homologue ( A B L ) in chronic myelogenous leukaemia ( C M L ) . However, it is difficult to identify drugs that can discriminate a protein between its normal and pathogenic state. Imatinib mesylate (produced by Glivec) inhibits both B C R - A B L and A B L , and several other kinases. On the other hand, a drug screening strategy aimed at restoring the function of tumor suppressor genes and defective apoptotic pathways, though genetically different in tumor and cancer cells, might turn out to be a suboptimal approach, because it w i l l be unlikely to identify drugs that can reactivate genes/proteins to restore normal protein function (Sager and Lengauer, 2003). Context-driven therapeutics depend on the identification of conditions in which the requirement for a particular target is enhanced in the context of cancer cells compared with normal cells, which can be due to intrinsic (e.g. genetic or epigenetic) or extrinsic (microenvironmental) changes, or both. Most anticancer drugs in use today affect targets present in both normal and cancer cells (Kaelin, 2005). One scenario of differential requirements that can be exploited is the phenomenon of synthetic lethality (SL). Synthetic lethality occurs when mutations in two different genes, while non-lethal as mutations, become lethal when combined in a cell as a double mutant. B y targeting a specific gene in a cancer cell containing another known mutation could lead to synthetic lethality and selective kil l ing. In this regard, an on-going effort in model organisms such 26 as yeast has been to construct a comprehensive synthetic lethal genetic interaction map. B y definition, these second-site loss-of-function mutations (which are otherwise non-lethal in the CIN-gene wild-type cells) define proteins that, when reduced in activity, cause lethality in the reference C I N mutant. I f the synthetic lethal interactions are conserved in humans, these second-site genes may suggest cross-species candidate proteins in humans that when inhibited (e.g., by a drug) would specifically k i l l tumor cells relative to normal cells. Synthetic lethal interactors that are common to multiple CLN mutants may suggest candidate drug targets that are effective for selective ki l l ing C I N cancers with different C I N gene mutations. I f kinetochore proteins, for example, turn out to represent a significant fraction of the C I N mutational spectrum in cancer, it is conceivable that second-site genes w i l l exist that are synthetically lethal in combination with an entire set of the kinetochore gene mutations, and therefore provide common drug targets for ki l l ing a broad spectrum of C I N cancers. While the relevance of kinetochore dysfunction to cancer still needs to be verified, defects in human mismatch-repair genes, MLH1, MSH2, and PMS2, are known to confer predisposition to colon cancer. Synthetic lethality data in yeast show that they are lethal in combination with mutations in D N A polymerase 5 and s that are otherwise viable. These latter enzymes catalyze D N A replication, and in the process they proofread the growing strand of D N A for errors. The results in yeast reveal the possibility of selectively ki l l ing M I N cancer cells by interfering with D N A polymerases (reviewed in (Friend and Oliff, 1998)). Recently, R N A interference screens have been applied in mammalian cells to decipher synthetic lethality relationships and identify novel targets (Ngo et al., 2006; Will ingham et al., 2004) (reviewed in (Brummelkamp and Bernards, 2003)). The selective ki l l ing concept can be expanded to synthetic dosage lethality where one loss-of-function mutation causes lethality when combined with overexpression of another protein. In therapeutics, loss of function of the second gene can be caused by drug inhibition. One example is that inactivation of retinoblastoma protein (RB) in cancer leads to an increase in E2F activation, which in turn activates various genes involved in S-phase entry, including topoisomerase 2. Topoisomearse II inhibitors such as etoposide 27 bind to topoisomerase II, causing D N A strand breaks and apoptosis. Therefore, RB-pathway mutations sensitize cells to topoisomerase II inhibitors. In a similar way, phosphatase and tensin homologue (PTEN) tumor-suppressor protein negatively regulates the phosphatidylinosital 3-kinase (PI3K) pathway, and a mammalian target of the rapamycin (mTOR) pathway. PTEN~'~ cells are reported be more sensitive to the antiproliferative effects of m T O R inhibitors than wild-type cells (Neshat et al., 2001). Defective mechanisms to maintain genomic stability render most cancers more vulnerable to genotoxic challenges. For example, XP mutations cause sensitivity to U V radiation, and mutations in ATM and BRCA2 cause sensitivity to ionizing radiation (IR). In addition, caffeine inhibits A T R , and can induce S-phase cells to undergo premature chromosomal condensation. Cells lacking P53 in G I control are found to be more susceptible to caffeine treatment. To identify the target pathway of anticancer drugs, the differential sensitivity of isogenic yeast mutants, each defective in a particular D N A repair or cell cycle checkpoint function, to Food and Drug Administration ( F D A ) -approved cytotoxic anticancer agents, such as cisplatin, camptothecin, and hydroxyurea, were tested (Dunstan et al,,2002; Hartwell et al., 1997; L u m et a l , 2004; Simon et al., 2000). Similar drug sensitivity assays can be set up in matched pairs of cell lines (which differ in one genetic alteration) and in tumor cell lines (Dolma et al., 2003; Hartwell et al., 1997; Torrance et a l , 2001). Another avenue of cancer therapeutics is to activate the cell cycle checkpoint, arrest cells and induce apoptosis. Taxanes (stabilizing microtubules) and vinca alkaloids (inhibiting microtubule assembly) are used to treat breast and ovarian cancer patients, and they reduce tension or produce unattached kinetochores in mitosis by altering microtubule dynamics and cause long-term mitotic arrest. It is suggested that cells exit long-term mitotic arrest (through adaptation), but then apoptosis is induced i n ' G l (Tao, 2005). A n inhibitor of K S P / E G 5 , a kinesin MT-dependent motor, is under phase I and II clinical trials and has the advantage that it only affects dividing cells because K S P only functions in mitosis. However, cells with a weakened checkpoint are less sensitive to microtubule-targeted drugs. 28 Paradoxically, complete inhibition of the mitotic checkpoint can also be effective in cancer therapeutics. Reducing M A D 2 or B U B R 1 to <10% level in various tumor cell lines causes complete inactivation of the mitotic checkpoint and results in massive chromosome misdistributions, and lethality results in 2-6 cell division (Kops et al., 2004; Miche l et al., 2004). This is probably because the rate of chromosome missegregation is elevated to such high level that is incompatible with cell viability. A s mentioned above, MAD and BUB genes were indeed first identified as mutants defective in triggering cell cycle arrest in response to microtubule inhibitors. If drugs could be used as tools to identify genes involved in a related process, we should be able to use C I N genes to identify new anticancer drugs and better understand their modes of action. Whether genomic instability reflects cause or effect of altered cell physiology during tumorigenesis, a comprehensive identification of genes whose mutation leads to chromosome instability is an important* but daunting, goal yet to be achieved. Understanding the etiology and tolerance of genome instability in viable cells is fundamental to understanding the development and survival of cancers, and may be instrumental in the design of therapeutic approaches that take advantage o f specific vulnerabilities exhibited by cancer cells. 1.3 O v e r v i e w o f thesis The goal of my thesis was to systemically identify functional determinants required for mitotic chromosome transmission in yeast, and extend the investigation to human cells based on cross-species protein sequence comparison. This analysis provided candidate C I N genes for cancer mutation testing in cancer patients. The results w i l l be directly relevant to understanding of cancer development, and may be useful in developing strategies for cancer therapy and for sub-classification o f tumors based on their CLN mutational spectrum (Figure 1.9). Chapter 2 describes the systematic identification of non-essential yeast genes important for the maintenance of chromosome stability using multiple assays. The comprehensive C I N gene set includes expected and unexpected genes, providing not only 29 a rich resource for the study of mechanisms required for accurate chromosome transmission, but also a list of candidate human C I N genes based on protein sequence similarities. Examples of known C L N gene mutations in cancers are shown, suggesting others are also candidate cancer genes needing to be tested. A C I N cancer cell-selective ki l l ing concept, utilizing synthetic lethality interactions between a C I N gene mutation and a second-site mutation, is discussed. Chapter 3 describes mutation testing of the list o f candidate human C I N genes generated from Chapter 2 in a panel of colorectal cancer patients. The significance of novel mutations, including genes involved in sister chromatid cohesion is discussed with regards to the mutation frequency and prevalence of mutations in C L N tumors. Functional analysis of several of the mutations found in colon cancers was performed in yeast by introduction of the mutations at the corresponding sites in the yeast SMC1 gene and scoring the C I N phenotype. Chapter 4 describes the characterization of 4 C L N genes identified in the genome-wide screens in Chapter 2. These 4 genes were only preliminarily characterized at the time the screens were completed. They potentially form an ubiquitin ligase complex. Phenotypic, genetic, and protein interaction data pertaining to this complex are analyzed and discussed. Chapter 5 draws conclusions from the above chapters, discusses the future directions of research in C I N , and show how yeast research can benefit cancer research. 30 T a b l e 1.1 G e r m l i n e mutations o f C I N and M I N genes causing cancer-prone syndromes Human gene Yeast gene Protein Function Associated disease/syndrome Major tumor types Mode of Inheritance Reference MSH2, MLHl, MSH6, PMS2 MSH2, MLHl Mismatch repair (MMR) Hereditary nonpolyposis colon cancer (HNPCC) (accounting for 3-5% of colorectal cancer) Colon, uterus, endometrium, ovary Autosomal dominant (Fishel et al., 1993) MYH /MUTYH mutY (E.coli) Base excision repair (BER) MYH-associated polyposis (MAP) Colon Autosomal recessive (Al-Tassan et al., 2002) XPA-G RAD1-4, RAD14 Nucleotide excision repair (NER) Xeroderma pigmentosum (XP) Skin Autosomal recessive O M I M BRCAl, BRCA2 Double strand break (DSB) repair Hereditary breast cancer Breast, ovary Autosomal dominant (Tutt et al., 1999; Weaver et al., 2002) NBS1 XRS2 Double strand break (DSB) repair Nijmegen breakage syndrome (NBS) Lymphoma, brain Autosomal recessive (Varon et al., 1998) MRE11A MRE11 Double strand break (DSB) repair Ataxia Telangiectasia-like (ATL) Myelodysplasia, acute myeloid leukemia Autosomal recessive (Stewart et al., 1999) BLM /RECQL3 SGS1 D N A helicase Bloom Syndrome Leukemia, lymphoma, skin Autosomal recessive (Mohaghegh and Hickson, 2001) WRN /RECQL2 SGS1 D N A helicase Werner syndrome Bone, skin Autosomal recessive (Mohaghegh and Hickson, 2001) RECQL4 SGS1 D N A helicase Rothmund-Thomson syndrome (RTS) Bone, skin Autosomal recessive (Mohaghegh and Hickson, 2001) ATM MEC1, TEL1 D N A damage checkpoint Ataxia Telangiectasia (AT) Seckel syndrome Leukemia, lymphoma, medulloblastomas and gliomas Autosomal recessive (Savitsky et al., 1995) Human gene Yeast gene Protein Function Associated disease/syndrome Major tumor types Mode of Inheritance Reference P53, CHK2 RAD53 DNA damage checkpoint Li-Fraumeni syndrome (LFS) Soft tissue sarcomas and osteosarcomas, breast cancer, brain tumors, leukemia, and adrenocortical carcinoma Autosomal dominant (Varley et al., 1997) (Bell et al., 1999) BUB1B MAD3 Mitotic spindle checkpoint Mosaic variegated aneuploidy (MVA) Rhabdomyosarcoma, Wilms tumor, and leukemia Autosomal recessive (Hanks et al., 2004) APC Wnt signaling inhibition; chromosome segregation? Familial adenomatous polyposis (FAP) Colon, thyroid, stomach, intestine Autosomal dominant (Green and Kaplan, 2003) FANCA,B,C,D I,D2,E,F,G,I,J ,L,M repair of DNA interstrand cross-links Fanconi anemia (FA) Leukemia Autosomal recessive & X-linked (Niedernhof er et al., 2005) T a b l e 1.2 Kine tochore and spindle checkpoint gene mutat ion or mis regula t ion associated w i t h cancer. (* shown as No. of positive patients or cell lines over the total No. tested) Gene Mutation/misregulation Frequency* Tumor type Reference BUB1 Dominant negative heterozygous deletion and missense mutation 2/19 Colorectal cancer (Cahill et al., 1998) Heterozygous missense mutation 1/30 Lung tumor (Gemma et al., 2000) Heterozygous missense mutation 1/10 Acute T-cell lymphoblastic leukemia (Ohshima et al., 2000) Dominant negative heterozygous deletion in kinetochore localization domain 1/2 Acute lymphoblastic leukemia (Ru et al., 2002) Deletion in kinetochore localization domain 2/2 Hodgkin's lymphoma (Ru etal., 2002) Overexpressed 30/36 Gastric cancer (Grabsch et al., 2003) BUB1B One heterozygous deletion, and one missense mutation 2/19 Colorectal cancer (Cahill et al., 1998) One heterozygous and one homozygous missense mutation, one homozygous deletion 3/10 Acute T-cell lymphoblastic leukemia (Ohshima et al., 2000) Downregulated (10 fold) 3/109 Colorectal cancer and others (Shichiri et al., 2002) Overexpressed 19/28 Gastric cancer (Grabsch et al., 2003) BUB3 Overexpressed 26/34 Gastric cancer (Grabsch et al., 2003) MAD2 Missense mutation 22/49 Gastric cancer (Kim et al., 2005) Downregulated 1/1 Breast cancer cell line (Li and Benezra, 1996) Downregulated 2/5 Nasopharyngeal cancer cell lines (Wang etal., 2000b) Downregulated 3/7 Ovarian cancer cell lines (Wang et al., 2002b) Rod Homozygous missense mutation 1/192 Colorectal cancer (Wang et al., 2004b) Zw10 Heterozygous missense mutation 2/192 Colorectal cancer (Wang et al., 2004b) Zwilch Heterozygous premature truncation 1/192 Colorectal cancer (Wang et al., 2004b) CENP-A Overexpressed (1.5-32.5 fold) 11/11 Colorectal cancer (Tomonaga et al., 2003) CENP-H Overexpressed (1.7-9.6 fold) 15/15 Colorectal cancer (Tomonaga et al., 2005) CENP-F (mitosin) Amplified (1.6-2.5 fold) Overexpressed (2.1-4.2 fold) 7/72 25/72 Head and neck squamous eel carcinomas (de la Guardia et al.,2001) Overexpressed 25/26 Salivary gland tumor (Ota etal., 2002) HEC1 (highly expressed in cancer; NDC80) Overexpressed 9/9 Cervical, acute lymphocytic leukemia, breast and colorectal cancer lines (Chen et al., 1997) Aurora-B (AIM1) Overexpressed 12/12 Thyroid cancer lines (Sorrentino et al., 2005) Overexpressed 7/7 Colorectal cancer (Tatsuka et al., 1998) INCENP Overexpressed (2.4-4.7 fold) 4/4 Colorectal cancer cell lines (Adams et al.,2001) 33 F i g u r e 1.1 T h e budd ing yeast c e l l c y c l e and chromosome cyc le (adapted from Pot , 2004) A. Yeast cells reproduce by budding. The cell cycle is divided into four stages [Gl, S (DNA replication), G2 and M (mitosis)]. The size of the bud gives an approximate indication of cell cycle stage. The nucleus is shown in red; in yeast, the nuclear membrane does not break down during mitosis. During cell division, chromosomes undergo a replication and segregation cycle that is synchronized with the cell cycle B. To follow the chromosome cycle in yeast, DNA content can be analyzed by flow cytometry of cells in which DNA has been stained with a fluorescent dye such as propidium iodide. A typical histogram showing the fluorescence distribution of a population of cycling cells is shown. Haploid cells in G l phase have a IN DNA content, while cells that have replicated their DNA and are undergoing mitosis (G2/M) have a 2N DNA content. 34 F i g u r e 1.2 Organ iza t ion o f centromere (reprinted from (C leve l and et a l . , 2003) Centromeres and kinetochores: from epigenetics to mi to t ic checkpoint s igna l ing , Cell, 112, 407 -421 , C o p y r i g h t 2003 , w i t h pe rmiss ion from Elsev ie r ) A. Overall organization of the centromere. A mitotic chromosome has been sectioned along the plane of the spindle axis, revealing the symmetric bipolar organization of a chromosome fully engaged on the spindle. (Right) Key elements have been pseudo colored. (Violet) The inner centromere, a heterochromatin domain that is a focus for cohesins and regulatory proteins such as Aurora B. (Red) The inner kinetochore, a region of distinctive chromatin composition attached to the primary constriction. (Yellow) The outer kinetochore, the site of microtubule binding, is comprised of a diverse group of microtubule motor proteins, regulatory kinases, microtubule binding proteins, and mitotic checkpoint proteins. B. Schematic illustration of centromere loci. Organization of centromeric DNA sequences from the four example organisms. (Top) Budding yeast with a 125 bp centromere comprised of three sequence domains (pink, red, yellow). Fission yeast centromeres show an organized structure, with a nonconserved central core (red), flanking inner repeats (pink arrows) at which the CENP-A-containing nucleosomes assemble, and conserved outer repeats (stippled purple). The Drosophila centromere spans -v.400 kb (red) embedded in constitutive heterochromatin (purple). (Bottom) Human centromeres have sizes approaching 10 Mb and are comprised of OJ-I satellite DNA (red) and a more divergent, less regular o i l satellite (pink), flanked by heterochromatin (purple). 35 F i g u r e 1 . 3 T h e process o f ach iev ing b ipo la r attachment (reprinted from ( P i n s k y and B i g g i n s , 2005) The spindle checkpoint : tension versus attachment, Trends Cell Biol, 15, 486-493 , C o p y r i g h t 2005 , w i t h permiss ion from Elsev ie r ) Micro" utXites Sp ind le po le Sister chromatid (a) Unattached kinetochores (b) 'Search and Capture': Single kinetochore captured by the side of a probing spindle microtubule (c) Sister chromatids transported to the pole where additional microtubules bind (d) Sister kinetochore interacts with microtubule from the opposite pole (e) Chromosomes congress to the spindle center and kinetochores come under tension TRENDS in Cell Boiogy 36 Figure 1.4 Types of kinetochore-microtubule attachments (reprinted from (Pinsky and Biggins, 2005) The spindle checkpoint: tension versus attachment, Trends Cell Biol, 15, 486-493, Copyright 2005, with permission from Elsevier) (a) Amphitelic: either bipolar or bioriented attachment. Sister kinetochores face opposite poles and bind only microtubules arising from the adjacent pole. (b) Syntelic: sister kinetochores face the same pole and attach to microtubules emanating from that pole. (c) Monotelic: sister kinetochores face opposite poles but only one kinetochore binds microtubules, leaving an unattached kinetochore. (d) Merotelic: sister kinetochores face opposite poles but one (or both) kinetochore(s) interact with microtubules from both poles. Amphitelic Syntelic Monotelic TRENDS in Cell Biology 37 F i g u r e 1.5 Structure of cohesin and a possible mechanism by which it might hold sister chromatids together (reprinted from (Nasmyth, 2002) Segregating sister genomes: the molecular biology of chromosome separation, Science, 297, 559-565, Copyright 2002, with permission from AAAS) (A) Smcl (red) and Smc3 (blue) form intramolecular antiparallel coiled coils, which are organized by binge or junction domains (triangles). Smcl/3 heterodimers are formed through heterotypic interactions between the Smcl and Smc3 junction domains. The COOH terminus of Sec 1 (green) binds to Smcl's ABC-like ATPase head, whereas its NH2 terminus binds to Smc3's head, creating a closed ring. Scc3 (yellow) binds to Sccl's COOH-terminal half and does not make any direct stable contact with the Smcl/3 heterodimer. Sccl's separase cleavage sites are marked by arrows. Cleavage at either site is sufficient to destroy cohesion. By analogy with bacterial SMC proteins, it is expected that ATP binds both the Smcl and Smc3 heads, alters their conformation, and possibly brings them into close proximity. By altering Sccl's association with Smc heads, ATP binding and/or hydrolysis could have a role in opening and/or closing cohesin's ring. (B) Cohesin could hold sister DNA molecules together by trapping them both within the same ring. Cleavage of Sec 1 by separase would open the ring, destroy coentrapment of sister DNAs, and cause dissociation of cohesin from chromatin. A SMC3 Separase 3 8 F i g u r e 1.6 The stages o f mitosis (reprinted from (Weaver and Cleveland, 2005) Decoding the links between mitosis, cancer, and chemotherapy: The mitotic checkpoint, adaptation, and cell death, Cancer Cell, 8, 7-12, Copyright 2005, with permission from Elsevier) Chromosomes enter mitosis as pairs of replicated sister chromatids that are linked by proteins known as cohesins. A: The chromatids condense during prophase and are released into the cytoplasm by nuclear envelope breakdown, which marks the transition into prometaphase and also represents the first irreversible transition into mitosis. B: During prometaphase, the initially unattached chromatids make connections to the microtubules of the mitotic spindle and the mitotic checkpoint is active, which means that the kinetochores assembled at the centromeres of unattached chromosomes generate a diffusible "wait anaphase" inhibitor. Antimitotic drugs delay cells in prometaphase by producing unattached kinetochores. C: At metaphase, every chromosome has made proper attachments to the mitotic spindle and has congressed to a central position. Production of the diffusible "wait anaphase" inhibitor has been silenced by stable kinetochore-microtubule interactions. As the checkpoint inhibitors decay, the anaphase promoting complex (APC), an E3 ubiquitin ligase, becomes active and recognizes securin and cyclin B, provoking their degradation. D: Loss of securin activates the protease, separase, that cleaves the cohesins, triggering sister chromatid separation and chromosome segregation during anaphase A. E: At anaphase B, the spindle elongates. F: At telophase, the now segregated chromosomes begin decondensing and the nuclear envelopes reform. G: Cytokinesis separates the nuclei into two daughter cells that re-enter interphase. g P R O M E T A P H A S E mitotic checkpoint OH ^ P R O P H A S E First stage ol mitosis INTERPHASE F T E L O P H A S E final stage of mitosis microtubule centrosome ^ chromosome nuclear envelope O c h e c k p o i n t *Ji2B complexes j-g g C d e 2 0 activated Anaphase Promoting Complex j b i q - M i r chain inactive separase active soparaso o c M E T A P H A S E mrtooc checkpoint OFF Cdc20 ACTIVATION of A P C Ubiqutiration and degradation of securin ^ A N A P H A S E A sister chromatid separation E A N A P H A S E B spindle elongation 39 F i g u r e 1.7 C e l l u l a r processes i n v o l v e d i n repl ica t ion and segregation o f chromosomes dur ing mi tos i s (reprinted from (Lengauer et a l . , 1998) Genet ic instabil i t ies i n human cancers, Nature, 396 , 643-649, Copyr igh t 1998, w i t h permiss ion from M a c m i l l a n Publ ishers L td ) Processes involved include chromosome condensation, cohesion of sister chromatids, and centrosome/rnicrotubule formation and dynamics. Checkpoints that are required in chromosome replication and segregation include the mitotic spindle checkpoint, which ensures that chromosomes are aligned correctly before anaphase, and the DNA-damage checkpoint, which prevents cells with DNA damage from entering prophase. Aberrations in these processes and checkpoints could give rise to the CIN phenotype. 40 F i g u r e 1.8 M u l t i p l e roads to aneuploidy (reprinted f rom (Rajagopalan and Lengauer , 2004a) A n e u p l o i d y and cancer, Nature, 432 , 338-341 , Copyr igh t 2004, w i t h pe rmiss ion f rom M a c m i l l a n Publ ishers L td ) The schematic illustrates a simplified cell cycle, highlighting processes that have been implicated in the advent of aneuploidy. Several pathways within the cell cycle (indicated in red) can be disrupted. Genes (indicated in green) associated with these processes and structures have been found to be mutated or functionally altered in aneuploid cancers. G1/S checkpoint 4 1 F i g u r e 1.9 F lowcha r t o f deve lop ing therapeutic strategy based o n candidate C I N gene ident i f icat ion. Identify C I N genes i n m o d e l organisms * Screen for mutations i n candidate C I N genes i n tumors Ar F u n c t i o n a l studies o f cancer gene mutants i n m o d e l organisms and h u m a n cel l s A> Synthetic lethal screen i n yeast us ing C I N muta t ion Determine analogous secondary target i n human V a l i d a t e synthetic le thal i ty i n human Ar Determine pharmaco log ica l feasibi l i ty Ar D r u g screen for secondary target 42 CHAPTER 2 Identification of Chromosome Instability Mutants in the Budding Yeast Saccharomyces cerevisiae and the Implication to Human Cancer A modified version of this chapter has been accepted for publication. Karen W . Y . Yuen*, Cheryl D . Warren*, Ou Chen, Teresa Kwok , Phi l Hieter, and Forrest A . Spencer (*These authors contributed equally to this work). Systematic Genome Instability Screens in Yeast and Their Potential Relevance to Cancer. Proceedings of the National Academy of Sciences of the United States of America. 43 2.1 I n t r o d u c t i o n Genome instability is a hallmark of cancer and falls into 2 classes: M I N (microsatellite instability, reflecting an increased mutation rate) or C I N (chromosome instability, reflecting an increased chromosome missegregation rate). While much is known about the spectrum of germline and somatic mutations causing M I N in cancer cells, little is known about the spectrum of somatic gene mutations causing C I N in cancer cells (as described in Chapter 1). Recently, a cross-species candidate gene approach has been used to define - 2 0 % of the C I N mutational spectrum in colon cancer (Cahill et al., 1998; Rajagopalan and Lengauer, 2004b);(Wang et al., 2004b). To comprehensively identify additional C I N gene mutations in cancers, one approach would be to identify all genes mutable to a C L N phenotype in a model organism, and to systematically test human homologues of the model organism C I N gene set for somatic mutations in tumors. Such a cross-species candidate gene approach, previously termed 'homologue probing' (Bassett et al., 1997), should in theory significantly expand our understanding of the C L N mutational spectrum in cancer. Comprehensive identification of genes whose mutation leads to C L N is an important, but daunting, goal yet to be achieved. Phenotype screening based on marker stability in budding yeast by random mutagenesis has provided several gene collections (e.g. chromosome transmission fidelity (ctf), chromosome loss (chl), minichromosomes maintenance (mem), and chromosome instability (tin)), and these genes are often functionally conserved in other eukaryotes (Hoyt et al., 1990; Kouprina et al., 1988; Maine et al., 1984; Meeks-Wagner et a l , 1986; Ouspenski et al., 1999; Spencer et al., 1990). Not surprisingly, different genetic screens have led to identification of gene sets important for various steps in the chromosome cycle, including those functioning at kinetochores, telomeres, and origins of replication, or in microtubule dynamics, sister chromatid cohesion, D N A replication, D N A repair, D N A condensation and cell cycle checkpoints. A l l these processes must be executed at high fidelity and coordinated temporally and spatially within the cell cycle to maintain genetic integrity. 44 For instance, the chromosome transmission fidelity (ctf) mutant collection was generated by Spencer et al. (Spencer et al., 1990) in the Hieter laboratory through mild E M S mutagenesis (0% kil l ing, 10-fold increase in canavanine resistant colonies), followed by screening of -600,000 yeast colonies for an elevated colony sectoring phenotype, which reflects loss of an artificial chromosome fragment. In total, 136 mutant strains were isolated. Based on complementation tests, this collection represents - 50 genes that could encode any of the many components necessary for the chromosome cycle to proceed with high fidelity. Specific secondary screens have been applied to the ctf collection with the aim of identifying mutants defective in a particular structure or process (Doheny et al., 1993). To date, about half (24) of the genes represented in the ctf collection have been cloned and characterized (Table 2.1). Among these, 9 genes encode kinetochore proteins, 10 encode proteins important for sister chromatid cohesion, and 5 encode other functions in D N A / R N A metabolism. Both essential and non-essential genes were identified from the screen. Interestingly, the top 7 complementation groups altogether contain 75 alleles, representing over half of the total number of isolates. These 7 yeast genes are the most highly mutable to C I N in this particular assay. Despite the ease of random mutagenesis and the possible recovery of hypomorphs of essential genes, random mutagenesis approaches rarely achieve screen saturation, because mutability varies among genes due to differences in size, base composition, and the frequency of mutable sites that can lead to viable cells with a detectable phenotype. However, the use of the S. cerevisiae gene knockout collection supports new and powerful strategies based on direct phenotyping of the null mutants. The -4,700 non-essential gene-deletion mutants represent >70% of yeast genes, but over 30% of mutants remain functionally unclassified (Giaever et al., 2002; Winzeler et al., 1999) (Saccharomyces Genome Database, www.yeastgenome.org). In this study, I have used the gene knockout set to carry out 3 systematic screens that follow marker inheritance in different chromosomal contexts to identify genes important for maintaining genome stability in yeast (i.e. non-essential yeast C I N genes). In addition to extending the catalog of genes known to affect genome stability, several 45 themes emerge from the analysis of the screen results. Because all mutants characterized are null, phenotype strength reflects the magnitude of the role played by each gene in genome stability. Thus direct comparisons are meaningful, between different mutants in a given assay system or between different assay systems for a given mutant. Some mutants exhibit phenotypes that are screen-specific, suggesting that chromosomal contexts determine what pathways predominate in protecting against genomic change. Protein similarity was used to identify candidate C I N homologues in other species, in particular human genes with relevance to cancer. For yeast C I N genes whose human homologues are mutated in cancers, yeast genetic interaction data were analyzed to identify common synthetic lethal interactors. Human homologues of these common synthetic lethal interactors may be useful as drug targets with broad spectrum applicability for selective elimination of C I N cancer cells. 46 2.2 M a t e r i a l s a n d me thods 2.2.1 G e n o m e - w i d e screens 2.2.1.1 C T F sc reen The synthetic genetic array (SGA) selection scheme (Tong et al., 2001a) was used to introduce the ade2-101 ochre mutation and an artificial chromosome fragment (CF) into MATa deletion mutants obtained from Research Genetics (www.resgen.com). To construct donor strains used to mate with the MATa deletion mutant array, Y2454 (see Table 2.2 for strain list) was first co-transformed with 2 P C R products. One P C R product contained an ade2-101 allele and terminated with an adjoining 20-bp o f TEF promoter. The other P C R product contained the TEF promoter, the natMX cassette, the TEF terminator, and 40-bp overlap with genomic ADE2 downstream sequence. Transformants were selected on medium containing lOOmg/L c lonNAT (Werner B io Agents, Germany). Cointegration oiade2-l01 and natMX at the ADE2 gene locus was confirmed by P C R . The resulting strain (YPH1724) was crossed with either YPH255 or YPH1124 , which contained CFVII (RAD2.d) or CFIII (CEN3.L) , respectively (Spencer et al., 1990). The 2 CFs were derived from different yeast chromosome arms and w i l l cover recessive mutations in that chromosomal region. The resulting diploids were sporulated, and spore progeny with the appropriate markers were recovered as the donor strains YPH1725 and YPH1726. Two S G A analyses, each using one of the donor strains, were performed as described previously (Tong et al., 2004) with the following modifications. The MATa yeast deletion mutant set (MATa ura3 his3 ykoA::kanMX) was arrayed at a density of 768 colonies/omni tray for robotic pinning. Each deletion mutant strain was represented in duplicate in each S G A analysis. The donor strain (MATa ade2-101::natMX ura3 his3 can! A mfalA::MFAlpr-HIS3), containing a £/iL43-marked C F , was mated with yeast deletion mutants on rich medium for 1 day. A l l plates were incubated at 25°C. Then diploids were selected twice (2 days and 1 day) on synthetic complete medium (SC) containing 200mg/L G418 (frivitrogen) and lacking uracil. Diploids were then pinned onto sporulation medium for 9 days, and MATa spore progeny with a C F were selected 47 on haploid selection medium (SC medium lacking histidine, uracil, and arginine but. containing 50mg/L-canavanine (Sigma)) for 2 days. Finally, MATa deletion mutants with a C F and the ade2-101 mutation were selected in 2 successive rounds (2 days each time) on the haploid selection medium with G418 and c lonNAT. A l l strains from the final selection plate were streaked to single colonies on SC medium with 20% of the standard adenine concentration. Plates were incubated at 25°C for 6-7 days, then at 4°C for 5-7 days to enhance the development of red pigment. A n instability o f the C F was indicated by a colony color-sectoring phenotype as in (Spencer et al., 1990). Briefly, red color in yeast cells is caused by accumulation of pigment due to a block in adenine production caused by the ade2-101 (ochre) mutation. This block is relieved in the presence of the SUP 11 gene located on the telocentric arm of the C F , encoding an ochre-suppressing t R N A T y r . Cells that contain the C F are therefore unpigmented, whereas cells that do not develop red color (Gerring et al., 1990; Hegemann et al., 1988; Warren et al., 2002). Colonies exhibiting unstable inheritance o f this C F develop red sectors, whereas wi ld-type strains form mostly white colonies. The severity of the phenotype was scored qualitatively by eye as mild, intermediate and severe (indicated as 1,2, and 3, respectively in Appendix 1 and Appendix 2). Figure 2.1.a shows the scheme of the C T F screen, and examples of severe and mild sectoring colonies. A l l deletion mutants that displayed a sectoring phenotype, or were identified in at least 1 of the other screens ( B i M or A L F ) were retested for sectoring phenotype. A miniarray was constructed and S G A analyses were undertaken as described above. Deletion mutants showing a sectoring phenotype in at least 2 out of 8 isolates (either in the original genomic screen or in the retest) were scored as having a C T F phenotype. 2 . 2 . 1 . 2 B i m a t e r screen The homozygous diploid deletion set obtained from Open Biosystems (www.openbiosystems.com) in 96-array format was grown on Y P D agar medium containing 200mg/L G418 for 3 days at 25°C. MATa and MATa mating tester lawns (YPH315 and Y P H 3 1 6 respectively) were generated by spreading 2 m l of saturated 48 culture on solid medium and grown at 25°C for 2 days. Each plate containing 96 deletion mutants was replica plated onto 4 test plates: an Y P D plate with 200mg/L G418 (a positive control), a synthetic complete medium plate lacking histidine, uracil, adenine, lysine, tryptophan and leucine (SC-6, a negative control), and 2 Y P D pre-mating plates. MATa and MATa mating tester lawns were each replica plated to the Y P D plate containing freshly replica plated deletion mutants. These were incubated at 25°C for 2 days for mating. The 2 Y P D plates were each replica plated to synthetic complete medium lacking histidine, uracil, lysine, adenine, tryptophan, and leucine (SC-6) and incubated at 25°C for 3 days to select mated products. These SC-6 plates were visually inspected, and densitometry measurements were obtained for each deletion mutant using the QuantityOne program (BioRad). Deletion mutants that exhibited elevated mating rates with both MATa and MATa mating testers were identified as candidate bimaters for further study (Figure 2.1.b). For confirmation, all bimater candidates were retested for bimater phenotype along with all positive mutants from the other two screens ( A L F and C T F ) . For the retest, 4 independent isolates per mutant were patched in 1cm squares. Each plate also contained negative and positive control patches: wild-type diploid (YPH1738) and Ml A/Ml A, respectively. After selection, the'number of colonies in each patch was estimated. To minimize the effect of early or late events during population growth, the median number of colonies was used to calculate fold-change (mutant/wild-type ratio). Homozygous deletion mutants with an average of > 1.5-fold increase in mating tests with both MATa and MATa testers were identified as bimaters. The severity of each mutant phenotype was recorded as an estimate of 2- to >5-fold increase over wild-type frequency after rounding (see Appendix 1 and 4). 2.2.1.3 a - l ike f a k e r screen The MATa haploid deletion collection (MATaykoA::kanMX) obtained from Research Genetics were manually arrayed in 1cm squares on Y P D plates. Each plate contained 3 controls: wild-type MATa (BY4742), wild-type MATa (BY4741), and the 49 MATa biml A:: kanMX from the MATa deletion collection. A lawn of 5 x 10 MATahisl mating tester cells (YPH316) freshly spread and dried onto solid rich medium. The presence of a-type mating cells in MATa mutant populations was detected as previously described (Warren et al., 2004a). ykoA patches were transferred onto the mating tester lawn by replica plating, followed by incubation at 30°C for 20-24 hr. Mated patches were then replica plated to SC-6, incubated for 2 days at 30°C to select mated products. Results from the primary screen were scored by comparing the number of colonies per patch to the wild-type MATa control patch for that plate (Figure 2.1 .c). A l l positive mutants were retested as described above (along with all positive mutants from bimater and ctf screens), but using 4 independent isolates for each mutant. After selection, mated products were counted and a fold-change calculation was generated (mutant median colonies per patch/wild-type median colonies per patch). Mutants were scored as positive i f they exhibited > 2-fold increase over wild-type frequency. 2 . 2 . 2 S t r a i n v e r i f i c a t i o n To evaluate the mutant identity of the 96-well content from each collection, cells from frozen stocks were patched on Y P D plates containing G418. A barely visible clump of cells (~105) was added directly to 20 ul lysis buffer (25mM T r i s - H C l pH8.0, 0.005% sodium dodecyl sulfate, 14 m M (3-mercaptoethanol, 0.5 m M E D T A ) , and incubated at 85°C for 10 min. P C R was performed in 96 well format using primers D I and K a n C for downtags (Shoemaker et al., 1996). P C R products were purified using Qiagen Qiaquick P C R purification kit or Macherey Nagel NucleoFast 96 P C R kit. D N A yield was determined using PicoGreen (Molecular Probes), and the concentration of each well was adjusted to 10-50 ng/ml. The downtag sequence was determined using oligonucleotide 5'-catctgcccagatgcgaagttaag-3' as primer. For mutants lacking a downtag, or when the downtag sequence analysis was ambiguous, the uptag sequence was obtained from P C R products generated using primers U l and K a n B , followed by sequencing obtained using a K a n B l (Shoemaker et al., 1996) sequencing primer. 50 The sequence o f the tag P C R products was determined b y d i d e o x y terminat ion at the J H M I Sequencing C o r e F a c i l i t y or SeqWr igh t Incorporated (www.seqwr igh t . com) . T h e results were ana lyzed us ing B L A S T N against a tag database, and al ignments w i t h expected tags h a v i n g e-values <10" 1 0 were considered evidence o f w e l l va l ida t ion . Sequence traces for B L A S T n outcomes >10" 1 0 were read manua l ly . A m o n g these were c lean traces ind ica t ing tag mutations present i n yeast that prevented a l ignment w i t h the correct tag sequences w i t h significant e-value (Eason et a l . , 2004) , or traces w i t h two or more peaks at m a n y posi t ions. Often w h e n two tags were present, the ident i ty o f a contaminat ing mutant i n a g iven w e l l c o u l d be determined. In homozygous d ip lo ids , different but 'correct ' tag alleles were often noted (e.g. where one tag was a frameshifted ve rs ion o f the expected sequence) (Eason et a l . , 2004) . Fa i lu re i n mutant ver i f ica t ion b y tag sequencing is c lass i f ied as " w r o n g " (incorrect strain(s) present), " con tamina t ion" (correct strain present but a contaminat ing strain was evident), or " n d " (not determined because the sequence obtained was unreadable, or that delet ion co l l ec t ion contained no yeast to validate) . Subsets o f C L N mutants were freshly generated b y transformation and phenotyped. TheykoA: .kanMXcassette was P C R from the delet ion set mutant w i t h gene f l ank ing pr imers located ~300bp upstream and downstream o f the gene, and the P C R product was transformed i n the respective parental wi ld - type . Transformants were conf i rmed b y pr imers f l ank ing the P C R product to con f i rm integration o f the dele t ion al lele and ensure r e m o v a l o f the w i ld - type locus. 2.2.3 Bioinformatic analysis 2.2.3.1 Functional analysis Over-representat ion o f G O b i o l o g i c a l process and ce l lu la r component annotation i n the yeast C I N gene list , compared to a l l yeast genes, was determined us ing G O T e r m F i n d e r as o f M a y 3, 2006 (db .yeas tgenome.org /cg i -b in /GO/goTermFinder .p l ) . 51 2.2.3.2 BLAST analysis Pro te in sequences o f yeast C I N genes were used as queries i n a B L A S T p al ignment search against protein sequence downloads for Homo sapiens ( R e f S e q pro te in database, as o f June 2004) , Mus musculus (RefSeq protein database June 2004) , C . elegans (Wormbase June 2004), D. melanogaster ( F l y B a s e release 3.2.0), and S. pombe (Sanger Institute, pompep June 2004) . H u m a n proteins f rom R e f S e q w i t h B L A S T p al ignments to yeast C I N prote in queries (e-value <10" 1 0 , J u l y 2004) were searched against O M I M ( w w w . n c b i . n l m . n i h . g o v / o m i m ) and cancer census (Futreal et a l . , 2004) prote in datasets for disease, especia l ly cancer, associat ion. 2.2.3.3 Protein and synthetic lethal interaction network Protein-prote in interactions and genetic interactions w i t h yeast C I N genes were obtained f rom the G R I D database, and the interaction networks were v i s u a l i z e d through the O S P R E Y program ( v l . 2 . 0 ) (Brei tkreutz et a l . , 2003a; Bre i tkreutz et a l . , 2003b) . 2.2.4 Electrophoretic karyotype of a-like fakers F o r the electrophoretic karyotype analysis , a MATa his5A::kanMX tester strain was used. T h i s strain contains a chromosome III length p o l y m o r p h i s m that dist inguishes it from chromosome III o f the delet ion co l l ec t ion background. M a t i n g between MATa strains w i l l occur w h e n the MATa locus from either parental genotype is lost. D u r i n g the A L F screen, the basal rate o f loss for the MATa hisl mat ing tester ( Y P H 3 1 5 ) was observed to be m u c h lower than that o f B Y 4 7 4 2 (parental strain to the MATa de le t ion co l lec t ion) . Thus , near ly a l l events detected were due to genome ins tabi l i ty i n the dele t ion mutant be ing characterized. Sample preparation, pulsed f ie ld ge l analysis , and in -ge l hybr id iza t ions were performed as descr ibed i n (War ren et a l . , 2004a). 52 2.3 Results 2.3.1 Genome-wide marker loss screens identify 130 yeast deletion mutants Three complementary marker loss assays were performed us ing the non-essential gene-delet ion mutant set. In the first screen ( C T F , for chromosome t ransmiss ion f idel i ty , F igu re 2.2a), inheritance o f an ar t i f ic ia l ch romosome fragment ( C F ) was moni to red us ing a c o l o n y co lo r marker. I mod i f i ed the Synthetic Genet ic A r r a y ( S G A ) me thodo logy ( T o n g et a l . , 2001a) to construct hap lo id delet ion strains ca r ry ing a C F , and performed a c o l o n y color-sec tor ing assay as an indicator o f chromosome ins tabi l i ty (Hie ter et a l . , 1985; Spencer et a l . , 1990). L i n e a r ar t i f ic ia l C F s serve as sensit ive indicators because their presence or absence does not affect v i ab i l i t y , and they resemble natural chromosomes i n their structure and s tabi l i ty (wi th 1.7-7 loss events per 1 0 4 d iv i s ions ) (Ger r ing et a l , 1990; H e g e m a n n et a l . , 1988; K o s h l a n d and Hieter , 1987; Shero et a l . , 1991; W a r r e n et a l . , 2002) . S ince the markers o n the C F are surrounded b y sequences that have no s imi l a r i t y to the yeast genome, loss o f markers p r i m a r i l y represent loss o f the w h o l e C F . In the second and th i rd screens, an endogenous locus (the ma t ing type locus MAT o n ch romosome III) was explo i ted as a marker. A bimater screen (designated B i M , F igu re 2.2.b) f o l l o w e d inheritance o f the MATa and MATaloci i n h o m o z y g o u s d i p l o i d dele t ion mutants. D i p l o i d cel ls heterozygous at MAT do not mate due to codominant suppression o f haploid-speci f ic c e l l differentiation pathways. L o s s o f either the MATa or the MATa a l le le results i n ma t ing competence, where the ma t ing type is determined b y the r ema in ing al lele . R e c i p r o c a l mat ing tests w i t h M 4 7 a or MATa ma t ing testers were performed o n the homozygous d i p l o i d delet ion set to ident ify c e l l populat ions w h i c h fo rm mated products w i t h both MATa and MATa at h igh rates. T h e endogenous rate o f loss o f either MAT al lele i n wi ld - type cel ls is 2-4 events i n 10 5 d iv i s ions (L i ras et a l . , 1978; Spencer et a l . , 1990), where the predominant mechan i sm is mi to t ic recombina t ion between homologues . T h i s loss o f heterozygosi ty can also be due to ch romosome loss, c h r o m o s o m a l rearrangement (deletions or translocations w i t h loss), or gene convers ion (al lele replacement). 53 In the th i rd screen (designated A L F , for a- l ike faker, F igure 2.2.c), the MATa locus inheri tance was s i m i l a r l y f o l l o w e d i n the MATa hap lo id delet ion set b y a ma t ing test. T h e MATa locus encodes transcript ion factors that suppress a-specific and promote a - spec i f i c gene expression (Strathern et a l . , 1981). L o s s o f the MATa locus leads to the default ma t ing type i n yeast, w h i c h is the a-type differentiation state. Thus , MATa ce l l s that lose the MAriocus w i l l mate as a-type cel ls , and are ca l led ' a - l ike fakers ' (Strathern et a l . , 1981). The frequency o f a- l ike faker cel ls i n a popula t ion is detected b y prototrophic select ion o f mated products. In wi ld - type yeast, A L F mi to t ic segregants are generated at a rate o f ~10" 6 ( (Herskowi tz , 1988b); C D W and F A S unpubl ished) . M e c h a n i s m s leading to MATa locus loss i n MATa ce l l s are s imi l a r to L O H i n d i p l o i d ce l ls , except that i n haploids there is no h o m o l o g for mi to t ic recombinat ion . H o w e v e r , the silent ma t ing type locus HMRa can m i t o t i c a l l y recombine w i t h the MATa locus (see be low) . The mutants ident i f ied i n the 3 assays were subjected to addi t ional va l ida t ions . Fi rs t , mutants from each p r imary screen were retested b y a l l 3 assays to ensure phenotype reproducib i l i ty . 310 knockou t strains were ident i f ied after secondary screening (84 C T F , 130 B i M , and 247 A L F ) . N e x t , the effect o f c ross -wel l contaminat ion was evaluated b y de termining the ident i ty o f the delet ion mutations present i n each o f the 310 w e l l locat ions i n each o f the 3 delet ion arrays (see A p p e n d i x 1 and 2 for details). T h i s was accompl i shed b y sequencing the ol igonucleot ide ' t ag ' unique to each dele t ion al lele (Giaever et a l . , 2002) . T h e presence o f >1 tag sequence or an incorrect tag sequence was evidence o f contaminat ing or w r o n g delet ion strains, and the phenotypes o f these locat ions were discarded. The 310 w e l l posi t ions exhibi ted 2 2 % , 9%, and 1 4 % error i n the MATa, MATa, and homozygous d i p l o i d sets, respect ively. Af t e r adjustment, 293 knockou t strains were ver i f ied as exh ib i t ing C I N i n at least 1 o f the 3 assays. O f these, 210 (72%) were uncontaminated i n a l l 3 sets. T o investigate the overa l l error rate i n each dele t ion set, w e sequenced strains from 60 r andomly chosen w e l l addresses and found 12%, 3%, and 3 % contaminated we l l s i n MATa, MATa, and homozygous d i p l o i d sets, respect ively. The higher error rate among yeast C I N mutants, relat ive to a r a n d o m l y 54 chosen set, m a y reflect a higher representation o f s l o w - g r o w i n g yeast strains among the C L N gene set that are read i ly replaced b y faster g r o w i n g contaminants. These tag sequence analyses suggest that the false negative frequency was between 3 and 2 2 % (i.e. phenotype detection cannot be performed due to the absence or contaminat ion o f mutant from the appropriate pos i t i on i n a col lec t ion) . Several different error or ig ins were observed i n each co l lec t ion . These inc luded ne ighbor ing-we l l sp i l lover (20-50% o f cases), plate-to-plate carryover (mutants from a conserved w e l l pos i t ion but from a different plate, 5 -15% o f cases), c o m m o n substitution b y a s ingle recurrent strain (that c o u l d occur from m e d i a contaminat ion, ~10%) , and i n d i v i d u a l events w i t h no apparent p h y s i c a l pattern (30-50% o f cases). F i n a l l y , an addi t ional source o f artifact i n delet ion co l l ec t ion phenotyp ing is the occas iona l presence o f undesired ' secondary ' mutations that cause the phenotype be ing screened (i.e. pos i t ive phenotypes caused b y mutations that are not at the site o f the knockou t al lele) . Ga i eve r et al. estimated the presence o f lethal or s l ow-g rowth phenotypes caused b y mutations i n genes that do not segregate w i t h a knockou t al lele to occur at a frequency o f 6 .5% (Giaever et a l . , 2002) . S u c h "col la tera l damage" w o u l d be expected to occur at an even higher frequency i n non-essential genes. T o ve r i fy that the C I N phenotype is actual ly due to the knockou t al lele , subsets o f mutants were regenerated b y independent transformation and phenotyped. Mutan t s w i t h phenotypes i n at least 2 screens were reconf i rmed as C I N mutants i n new transformants at a h i g h rate (13/13 C T F , 9/10 B i M , 9/11 A L F ) . O n the other hand, mutants ident i f ied w i t h phenotypes i n o n l y a s ingle assay were reconf i rmed i n new transformants at l ower rates, - 4 3 % for the hap lo id col lec t ions (2/6 o f mutants exh ib i t ing C T F on ly , 4/8 o f mutants w i t h A L F phenotype on ly , and - 7 5 % for the d i p l o i d co l l ec t ion (3/4 o f mutants exh ib i t ing B i M phenotype on ly ) . These data indicate a s ignif icant frequency o f secondary muta t ion effects i n the assay-specific subsets o f C I N mutants ident i f ied i n the p r imary screens, and emphasize the va l ida t ion inherent i n per forming screens i n mul t ip le col lec t ions . T h e higher reconf i rmat ion rate o f B i M from the homozygous dele t ion mutants is consistent w i t h the presence o f secondary mutations, w h i c h w o u l d often be covered b y the w i ld - type 55 al lele dur ing the construct ion o f d ip lo ids w h e n independent hap lo id segregants were mated. In total, 130 mutants are o f h i g h confidence (the 115 delet ion strains ident i f ied i n more than 1 assay, together w i t h 15 mutants reconf i rmed independently to have a pos i t ive phenotype i n o n l y 1 assay). These 130 genes are l is ted i n F igure 2.3 and A p p e n d i x 1, w h i c h reflect the current data status i nc lud ing a l l confirmations performed to date. T h e remain ing 163 mutants ident i f ied i n o n l y 1 screen are l is ted i n A p p e n d i x 2, and are regarded w i t h l ower confidence (wi th - 4 3 % and - 7 5 % true pos i t ive frequencies a m o n g the assay-specific subsets ident i f ied i n the hap lo id and d i p l o i d mutant screens, respect ively) . A p p e n d i x 1 and 2 inc lude measures o f phenotype severity, as w e l l as annotations for w e l l contaminat ion i n any o f the 3 delet ion sets and changes due to independent knockou t evaluat ion. 2.3.2 Functional distribution of yeast CIN genes C o m p a r i n g the gene on to logy ( G O ) annotations o f the 293 C L N genes to that o f the entire yeast genome (Harr is et a l . , 2004) indicated that the C I N gene set has an over-representation i n numerous expected ce l lu la r components: nucleus, chromosome, kinetochore, micro tubule , cytoskeleton, spindle, nuclear pore, spindle po le body , rep l ica t ion fork, and chromat in ( A p p e n d i x 3). F o r G O b i o l o g i c a l processes, the C L N gene list is enr iched i n genes i n v o l v e d i n c e l l cyc le , c e l l prol i ferat ion, response to D N A damage response, and nuclear d i v i s i o n ( A p p e n d i x 4). U s i n g G O b i o l o g i c a l process annotation, the 293 C I N genes fa l l i n broad functional groups, i n c l u d i n g - 4 0 % funct ioning i n D N A metabol i sm, chromosome, or c e l l cyc le , - 4 0 % funct ioning i n processes not o b v i o u s l y impl ica ted i n marker loss, and - 2 0 % w i t h u n k n o w n funct ion (Figure 2.4). These G O annotations reflect the current knowledge o f studied genes, ind ica t ing that these screens ident i f ied genes k n o w n to be funct ioning i n genome maintenance. Interestingly, genes not p rev ious ly k n o w n to contribute to s tabi l i ty were also ident i f ied. F o r example , 7 yeast mutants i n the adenine biosynthet ic pa thway {adel, ade2, ade4, ade5/7, ade6, ade8, adel 7) gave rise to elevated a- l ike fakers at frequencies 56 ranging from 2- to 31- fo ld above wi ld - type ( A p p e n d i x 1 and 4). Three o f these (adel, ade6 and adel 7, s h o w n i n F igure 2.3) were tested i n fresh transformants, and a l l 3 were val idated. T h i s indicates that adenine, adenine pa thway intermediates or der ivat ive metaboli tes are important for genome stabil i ty, and that compensatory mechanisms used b y ce l l s w h e n de novo adenine synthesis is b l o c k e d are not fu l ly sufficient. U s i n g the G R I D database and O S P R E Y network v i sua l i za t ion program, 92 p h y s i c a l interactions were found among 103 o f 293 C I N proteins, i n c l u d i n g prote in complexes , networks and pathways that are k n o w n to be important for ma in ta in ing ch romosome stabi l i ty (Brei tkreutz et a l . , 2003a; Bre i tkreutz et a l . , 2003b) ( F i g . 2.7). Inspect ion o f ne twork interactions should reveal n o v e l hypotheses regarding functions o f uncharacter ized genes. F o r example , msb2A was ident i f ied as a bimater, and M s b 2 p p h y s i c a l l y interacts w i t h M a d 2 p and M a d 3 p i n the spindle checkpoint pa thway and N d c 8 0 p at the kinetochore. W h i l e M s b 2 p is k n o w n to be an osmosensor prote in and is required to establish c e l l polar i ty , its role i n main ta in ing chromosome stabi l i ty has not been explored . YBP2, ident i f ied i n the bimater test, is impl ica ted i n ox ida t ive stress response and the gene product interacts w i t h N u p l 4 5 p , w h i c h is essential. N u p l 4 5 p interacts w i t h N u p l 2 0 p and N u p 8 4 p , w h i c h were found i n our screens a long w i t h another nuc leopor in N u p l 3 3 p , suggesting that nucleopor ins p l a y important roles i n main ta in ing ch romosome stabil i ty. A subcomplex o f nucleoporins conta in ing N u p 5 3 p , N u p l 7 0 p , and N u p l 5 7 p are associated w i t h the spindle checkpoint proteins M a d l p and M a d 2 p . Interestingly, ybp2A is synthetic lethal w i t h mad2A, suggesting Y b p 2 p m a y p l a y a ro le i n genome stabi l i ty i n associat ion w i t h nucleoporins and the spindle checkpoint . Indeed, Y b p 2 p was recent ly found to interact w i t h mul t ip le kinetochore proteins, and was found to spec i f i ca l ly interact w i t h centromere D N A sequences b y chromat in immunoprec ip i t a t ion (Kentaro O h k u n i and K a t s u m i K i t a g a w a , personal communica t ion) . In addi t ion, several C I N genes ident i f ied i n the screens were recent ly characterized to p l a y a ro le i n genomic stabil i ty. F o r example , D i a 2 p , a F - b o x prote in i n the S C F ( S k p l p -C d c 5 3 p / C u l l i n - F - b o x ) E 3 ub iqu i t in ligase complex , was recently s h o w n to be i n v o l v e d i n regulat ing D N A repl ica t ion and important for stable passage o f rep l ica t ion forks through 57 regions o f damaged D N A and natural fragile regions ( B l a k e et a l . , 2006; K o e p p et a l . , 2006) . N c e 4 p is i n v o l v e d i n media t ing S g s l p - T o p 3 p helicase-topoisomerase c o m p l e x ( C h a n g et a l . , 2005; M u l l e n et a l . , 2005) , and M m s 2 2 p and M m s l p are i n a n o v e l D N A damage repair pa thway ( B a l d w i n et a l . , 2005) (see Chapter 4) . Integrating the C L N gene catalog w i t h other phenotypic , genetic and p h y s i c a l interaction data proves to be a fruitful avenue to further our understanding o f mechanisms that main ta in genomic stabil i ty. 2.3.3 Integration of genome-wide phenotypic screen with genetic screens reveals functions of uncharacterized genes in chromosome stability maintenance U s i n g 14 hypomorph ic and 3 hypermorphic kinetochore alleles as queries, genome-wide synthetic lethal ( S L ) and synthetic dosage lethal i ty ( S D L ) screens were performed o n the non-essential yeast delet ion mutant set (Measday et a l . , 2005) . S L interactions occur between genes i n v o l v e d i n the same, para l le l or redundant b i o l o g i c a l pa thway. S D L interact ion occurs w h e n overexpress ion o f a prote in i n wi ld - type ce l l s remains v iab le , but causes lethal i ty i n a mutant. S D L can occur between genes w i t h i n the same c o m p l e x where their s to ichiometry is important. Overexpress ion o f the query prote in m a y titrate out another prote in that is required for c e l l v i a b i l i t y i n the knockou t strain. I f the funct ion o f the query prote in is to regulate the mutant protein, then overexpress ion o f the regulatory factor c o u l d be detrimental to the knockou t mutant. Conve r se ly , i f the no rma l funct ion o f the mutant prote in is to regulate the query protein, then overexpress ing the query prote in i n a strain defective for its regulatory factor c o u l d be lethal ( M e a s d a y and Hieter , 2002) . T h e kinetochore S L and S D L screens ident i f ied 211 non-essential dele t ion mutants i n total, but surpr is ingly, o n l y 14 gene mutants were ident i f ied i n both S L and S D L screens. H o w e v e r , these over lapping mutants were enr iched for ch romosome t ransmiss ion f ide l i ty (ctf) defects (8/14 mutants from both S L and S D L screens vs . 20/197 mutants from either S L or S D L screens d i sp layed a ctf phenotype) ( rev iewed i n (Baetz et a l . , 2006; Eisenste in , 2005)). O n e gene ident i f ied i n this over lapp ing set and the ctf screen was RCS1IAFT1, an iron-regulated t ranscr ipt ion factor (Rutherford and B i r d , 2004) . Indeed, R c s l p co- loca l izes w i t h a k inetochore prote in 58 by indirect immunofluorescence analysis on chromosome spreads, and has both genetic and physical interactions with the inner kinetochore protein Cbflp (Measday et al., 2005). Such an example illustrates the power of complementing genome-wide SL, SDL and phenotypic screens to uncover hidden relationships and predict functions of genes. 2.3.4 Chromosome loss is the major mechanism of M47a loss in a-like fakers The CTF and B i M phenotypes have been widely used to study genome instability. However, the A L F phenotype has been only rarely used (Lemoine et al., 2005; Liras et al., 1978; Warren et al., 2004a), and has not been as well characterized. The electrophoretic karyotype of mated colonies obtained after selection was analyzed to infer the mechanism of MATa locus loss. Possible events include loss of the entire chromosome III, deletion or translocation removing the MAT locus, or gene conversion from MATa to MATa by recombination with HMRa. The chromosome III in the mating tester strain was larger than that in the knockout strains, and could be visually differentiated by pulsed-field gel electrophoresis (Figure 2.6a). In-gel hybridizations with a probe that hybridizes to 3 distant sites on chromosome III (the MAT locus, and the silent mating type loci located distally on each arm) allowed detection of the 2 parental chromosome III bands as well as aberrant chromosome III derivatives. Aberrant chromosomes III were observed in a variety of sizes, including a 200 kb product likely to represent homologous recombination between MATa and the silent locus HMRa, known to generate an active MATa locus concomitant with a large deletion on the right arm of chromosome III (reviewed in (Herskowitz, 1988a)). Electrophoretic karyotyping of mated products from wild-type indicated that 68% of events were due to whole chromosome loss, 20% to chromosomal rearrangement, and 12%> to gene conversion (Figure 2.6b). Karyotype analysis of 13 high frequency A L F mutants showed that in 11 A L F mutants, loss of whole chromosome III was the predominant mechanism, similar to wild-type cells. In 2 A L F mutants, different predominant mechanisms were observed in a statistically significant manner. rad27A showed predominantly chromosome rearrangement, whereas most sovlA had an intact 59 chromosome III. RAD27 encodes an endonuclease that promotes O k a z a k i fragment maturat ion dur ing D N A repl icat ion. The A L F associated rearrangements are consistent w i t h previous characterizat ion o f RAD2'7 as a gene that protects against gross ch romosoma l rearrangements ( C h e n and K o l o d n e r , 1999). SO VI has been impl i ca ted i n respirat ion based o n its loca l i za t ion to the mi tochondr ia ( S G D ) . T o further define the events g i v i n g r ise to a- l ike fakers i n the sovlA mutant, P C R was used to detect the presence o f MATa and MATa l o c i i n the mated products ( H u x l e y et a l . , 1990). Interestingly, a l l mated products f rom the sovlA mutant contained both MATa and MATa l o c i , ind ica t ing in t roduct ion o f the MATa al lele into MAT'by gene convers ion . T h i s was not the general pattern observed i n wi ld - type or i n other mutants, where o n l y 3 % (1/39) or 6% (24/386) o f isolates tested were o f this type, respect ively. Interestingly, some h i g h frequency a-l ike fakers that showed w h o l e ch romosome III loss fa i led to exhibi t a sectoring phenotype i n the C T F screen. O f 13 frequent A L F mutants ana lyzed i n F igu re 2.6, o n l y 5 were ident i f ied b y C T F phenotype i n the h i g h -throughput screen: 3 w i t h strong (kar3A, siclA, and dia2A) and 2 w i t h weak (rad27A and nce4A) phenotypes. T o con f i rm the presence o f assay difference, 5 frequent A L F mutants were d i rec t ly retested for the C T F phenotype i n fresh transformants. T w o o f these (karZA, sicJA) exhib i ted a strong C T F phenotype as expected, and 3 showed m i l d sector ing (esc2A, rad50A, xrs2A, F igure 2.6c). Thus , frequent A L F p roduc t ion does not s t r ic t ly correlate w i t h frequent C F loss. T h i s c o u l d indicate that different factors inf luence the inheri tance o f endogenous chromosome III and the C F . O n e explanat ion is that the telocentric structure o f the C F m a y enhance ins tabi l i ty i n some mutants. A n o t h e r is that the presence o f a part ia l homologous chromosome p rov ided b y the C F m a y suppress instabi l i ty . Further w o r k w i l l be required to determine the under ly ing b i o l o g i c a l mechanisms that exp la in these uncorrelated phenotypes. 2.3.5 Many yeast CIN genes are conserved Current understanding o f mechanisms that contribute to genome stabi l i ty has been la rge ly fueled b y w o r k from m o d e l systems. T h i s approach has been informat ive for 60 human biology because of the remarkable functional conservation within the chromosome cycle. To evaluate conservation of yeast C I N genes identified in the screens, B L A S T p searches using yeast amino acid sequences against proteomes from S. pombe, C. elegans, D. melanogaster, M. musculus, and H. sapiens were performed. Among the 293 yeast C L N genes, 103 (35%) have homologues with e-values <10"1 0 in all 5 organism proteomes searched (see Appendix 5, which contains alignment results and functional summaries). Previous work showed that - 4 0 % of yeast proteins are conserved through eukaryotic evolution (Rubin et al., 2000), and 30% of known genes involved in human diseases have yeast homologues (Bassett et al., 1997). In agreement, 124 (42%) of the yeast C I N genes identified in this study have homologues in human, with e-values <10"10. Human homologues of yeast C L N genes represent candidates that may cause a C I N phenotype in human cells when mutated. Genetic perturbation causing a C L N phenotype can be a predisposing condition for cancer initiation or progression. Among the 130 high confidence C I N gene list, 10 'top hit' human homologues (with e-values <10"10) (Table 2.3 and Appendix 6) have been previously shown to exhibit somatic mutations in cancer. 2.3.6 A strategy for cancer therapy: synthetic lethality and selective cancer cell killing While C L N mutations can contribute to tumorigenesis, the altered genotype of a cancer cell may define a genetic "Achil les heel" that supports the selective ki l l ing of tumor cells relative to adjacent normal cells. Genetic interactions resulting in cell lethality hold promise for the design of therapeutic approaches in cancer. One kind of genetic interaction with properties useful for this strategy is synthetic lethality, observed when two mutations individually capable of supporting viability cause cell death when present together. Synthetic lethal mutant pairs identify genes that function in parallel or related pathways that cannot be simultaneously lost (Ooi et al., 2006). Fol lowing this logic, cancer cells with a specific C I N mutation can be killed through loss of function of a synthetic lethal partner, while sparing normal cells (Hartwell et al., 1997; Kael in, 2005). Systematic, large scale synthetic lethality analysis in yeast provides a means for identifying such second-site loss-of-function mutations (Eason et al., 2004; Harris et al., 61 2004; Pan et al., 2006; Tong et al., 2004). These budding yeast studies provide candidate human proteins whose inhibition (e.g., by a drug) may specifically k i l l tumor cells relative to normal cells. In this regard, gene deletions that exhibit synthetic lethality with multiple different C I N gene mutants are particularly attractive, as they might define broad-spectrum therapeutic targets. To address this concept, an analysis all known synthetic lethal interactions available for the yeast C I N genes that have cancer gene homologues (shown in Table 2.3) was performed (8 of the 10 have published synthetic lethal data). These 8 mutants are connected to 250 partners by 371 synthetic lethal interactions based on B i o G r i d (Stark et al., 2006) (data not shown). Among the 250 partners, 61 bridge at least 2 yeast cancer homologues (Figure 2.7). Notably, 3 mutants (ctf4A, ctfl8A, dcclA) exhibit synthetic lethality with at least 6 cancer gene homologues. Interestingly, these 3 yeast genes share a role in sister chromatid cohesion (Mayer et al., 2004; Warren et al., 2004a). The 'hub' position of these 3 mutants in the interaction network implies that different C L N gene mutants share a common genetic vulnerability, and these common synthetic lethal interactors can serve as broad spectrum targets. The existant synthetic lethal dataset in budding yeast, although incomplete, is continuously expanding (Tong et al., 2004). Therefore, more 'hubs/common nodes' maybe identified. A comprehensive synthetic lethal network, together with an increased understanding of the mutation spectrum in cancers, could provide insights pertinent to the design of therapeutic approaches in which human cancer cells are efficiently targeted for death by clinical intervention. Integration of knowledge among emerging high throughput datasets in model organisms w i l l stimulate new research directions and applications in combating human diseases. 2.4 Discussion This work identified an extensive catalog of genome instability mutants, based on phenotypic testing of haploid and diploid yeast knockout collections for chromosome transmission fidelity (CTF), bimater behavior ( B i M ) , and a-like faker formation ( A L F ) . This study characterized all non-essential yeast genes due to their accessibility for 62 phenotyping. Because many essential genes are known to contribute to genome stability from traditional approaches, a similar systematic screening effort for essential genes would be of great interest, but w i l l first require the development of a comprehensive hypomorphic mutation resource. A n extensive catalog is useful for the understanding of mechanisms that maintain genome stability, for the identification of new pathways important for genome maintenance, and for the organization of functional networks. Systematic screening of arrayed non-essential mutants avoids the sampling problem in traditional mutagenesis methods, and supports the direct comparison of phenotypes observed because all alleles are null. Differences in both phenotype severity and assay specificity were observed. The relative contributions of specific gene products to genome maintenance are revealed directly by phenotype strength. For example, rad27A, dial A, nce4A, and xrs2A exhibited the strongest A L F phenotypes (> 56-fold above wi ld type) among the high confidence yeast C I N genes, whereas well-studied damage response genes such as mec3A, mrclA, ddclA, and rad9A showed milder phenotypes (~11-fold). Apparently, under the growth conditions used in the screen, damage caused by the absence of Rad27p, Dia2p, Nce4p, or Xrs2p proteins exceeds that resulting from checkpoint loss. In addition to phenotype comparisons within a given assay, results from different assays can be compared. For example, xrs2A exhibited one of the highest A L F frequencies but a mi ld or absent C T F phenotype, indicating that the damage associated with xrs2A is more relevant to the maintenance o f a haploid chromosome III than to the artificial C F . In general terms, different chromosome marker stability assays (CTF, A L F , and B i M ) defined both distinct and overlapping gene sets. The screen specificities could be due to sensitivity differences, but likely reflect mechanistic differences revealed by the assay systems. For instance, screening in both haploids and diploids may give insight into how ploidy affects the maintenance of chromosomes. The results demonstrate the importance of using complementary assays to comprehensively identify genome maintenance determinants. The error observed in the deletion arrays for non-essential genes underscore the importance of mutant validation. It is widely known that mutant arrays are "evolving 63 resources" that accumulate changes due to manipulation and selective pressure (i.e. cross well contamination, aneuploidy, second site mutation, etc.). For example, 8% of deletion mutants exhibit chromosome-wide expression biases indicative of aneuploidy for whole chromosomes or chromosomal segments (Hughes et al., 2000). Lethal or slow-growth phenotypes caused by mutations in genes that do not segregate with a knockout allele occur at a frequency of 6.5% (Giaever et al., 2002). However, parameters indicative of array quality are usually not reported in studies using deletion sets. This issue becomes increasingly important as phenotypic data derived from distantly related replicates of the deletion resource are compared and integrated. In this study, tag sequence analysis of CLN mutant strains suggests false negative observations from well contamination were between 3 and 22% in different screens (see Appendix 1 and 4 for details). This phenomenon is l ikely to be observed in other copies of the deletion sets. Because the data were derived from 3 different array sets obtained from commercial distribution sources, and involved 2 laboratories both with experience in handling large strain collections, it is unlikely that well address errors were due to laboratory specific manipulation of the sets. A n empirical measure agrees: the C T F screen of the knockout collection identified 12 of 15 non-essential ctf mutants found previously in a traditional mutagenesis (Spencer et al., 1990). Two out of the 3 missed mutants were due to incorrect strains at the well positions of the array plates obtained from the commercial distribution source when checked by P C R . The false positive frequency due to secondary mutations or aneuploidy in the deletion collection strains can also be estimated. For the haploid collections, the false positive rates were relatively high: l-{26X33%+(10+19)X91%+32}/86=18% for the MATa collection; l-{126X50%+(54+19)X91%+32}/231=31% for the MATa collection; and for the diploid collection, the false positive rate was lower: l-{26X75%+(10+54)X91%+32}/122=5%. These frequencies are consistent with the frequency of unlinked recessive lethal mutations segregating independently of the deletion mutations that was observed during construction of the mutant resource (Giaever et al., 2002). In this study, false positive observations were rare among genes identified 64 in >1 chromosome marker loss assay (i.e. in >1 deletion resource). The results were therefore partitioned into 130 high confidence genes (115 genes identified by >1 screen, plus 15 genes confirmed in new transformants), and 163 lower confidence genes identified by single screens only. A full catalog of yeast C I N genes w i l l provide a rich resource for ongoing studies of genomic instability in many organisms, including human. Additional screens of non-essential yeast mutants (such as G C R screens in (Huang and Koshland, 2003; Smith et al., 2004)) and systematic incorporation of essential mutants w i l l enhance the utility of the yeast model system. Yeast C I N genes define cross-species candidate genes in humans that could contribute to C I N during tumorigenesis. A recent survey of C I N colorectal tumors (Wang et al., 2004b) provides a proof of principle. One hundred human candidate genes (chosen for similarity to model organism C I N genes) were screened for mutations in tumor samples, yielding 5 new C I N human genes mutated in colorectal cancer (MRE11, ZwlO, Zwilch, Rod, and Ding, in addition to the 2 previously known (CDC4, and BUB1) (reviewed in (Yuen et al., 2005)). These 7 C I N cancer genes account for <20% of the C L N mutational spectrum in colon cancer, and many other candidate C I N genes remain untested. Systematic analysis of the mutational spectrum leading to a C I N phenotype in a model eukaryotic organism such as yeast w i l l therefore help to define the mutational spectrum leading to a C I N phenotype in human cancer, and may accelerate the identification of protein targets for selective ki l l ing of cancer cells. 65 T a b l e 2.1 24 ctf mutants cloned to date A n additional 27 single member ctf isolates have not been cloned. ctf # alleles Gene name Essential? Function Cohesion Kinetochore DNA/RNA metabolism 1 30 CTF1/CHL1/LPA9 Cohesion X 2 11 (not cloned) 3 11 CTF3/CHL3 Central kinetochore X 4 8 CTF4/POB1 /CH L15 Cohesion X 5 5 CTF5/MCM21 Central kinetochore X 6 5 CTF6/RAD61 Cohesion X 7 5 CTF7/EC01 X Cohesion (establishment) X 8 3 CTF 8 Cohesion (alternative RFC) X 9 3 (not cloned) 10 3 CTF10/CDC6 X DNA replication X 11 3 PDS5/SP027 Cohesion/condensati on X 12 3 CTF 12/SCC2/AMC3 X Chromosome condensation X 13 1 CTF13/CBF3C X Inner kinetochore (CBF3) X 14 1 CTF14/NDC10 X Inner kinetochore (CBF3) X 15 1 CTF15/RPB4/SEX3 Subunit of RNA polymerase II X 16 1 (not cloned) 17 2 CTF 17/CHL4/MCM1 Central kinetochore X 18 3 CTF18/CHL12 Cohesion (alternative RFC) X 19 2 CTF 19/MCM18/LPB Central kinetochore X s3 BIM 1 /HSN9/YEB1 Microtubule-binding at SPB/kinetochore X si 27 1 SIC1 Cell cycle regulator X sl38 1 SPT4 Chromatin structure/ transcription X S141 1 NUP170/NLE3 Nucleoporin X S143 MAD1 Kinetochore protein /spindle checkpoint X sl55 1 MCM16 Central kinetochore X si 65 1 SCC3/IRR1 X Cohesion X sl66 1 SMC1/CHL10 X Cohesion/condensati on X 27 109 Total 7 10 9 5 66 T a b l e 2 . 2 L i s t o f yeast strains used i n Chapter 2 The genotypes and origins of strains used in this study are shown. Strain Genotype Reference BY4741 MATa (Brachmann et al., 1998) BY4742 MATa (Brachmann et al., 1998) Y2454 MATa mfalA::MFAlpr-HIS3 canlA ura3A0 leu2A0 his3Al lys2A0 (Tong et al., 2001b) YPH1724 MATa ade2-101::natMX mfalA::MFAlpr-HIS3 canJA ura3A0 leu2A0 his3Al lys2A0 This study YPH255 MATa ade2-Wl his3-A200 ura3-52 lys2-801 trpl-A63 Ieu2-Al CFVII(RAD2.d)::URA3 SUP11 Hieter lab YPH1124 MATa ade2-101 his3-A200 ura3-52 lys2-801 trpJ-A63 Ieu2-Al CFIII(CEN3.L)::URA3 SUPU (Pot et al., 2003) YPH1725 MATa ade2-101:: natMX his3 ura3 lys2 canlA mfalAr.MFAlpr-HIS3 CFVII(RAD2.d)::URA3.SUP11 This study YPH1726 MATa ade2-101:: natMX his3 ura3 lys2 can]A mfalA::MFAlpr-HIS3 CFIII(CEN3.L)::URA3 SUP11 This study YPH315 MATa hisl (Spencer et al., 1990) YPH316 MATa hisl (Spencer et al., 1990) YPH1738 MATa/MATa ura3A0/ura3A0 leu2A0/leu2A0 his3Al/ his3Al LYS2/lys2A::kanMX6 MET15/metl5A::kanMX6 This study 67 T a b l e 2.3 H u m a n proteins homologous to yeast C I N genes are mutated i n cancer The protein sequences corresponding to 130 high confidence yeast CIN genes were used as queries in a BLASTP search against the human RefSeq protein database. Online Mendelian Inheritance in Man (OMIM, www.ncbi.nlm.nih.gov/omim) and cancer census (Futreal et al., 2004) databases were used to identify cancer associated mutations in 'top hit' human genes. Yeast Gene Top Human Hit E-value Disease Description, MIM#(disease) MIM# (gene) Reference ADE17 ATIC 0 Anaplastic large cell lymphoma Cancer census RAD54 RAD54L 1E-164 Non-Hodgkin lymphoma; Breast cancer, invasive intraductal; Colon adenocarcinoma 603615 OMIM RAD51 RAD51 1E-122 susceptibility to Breast cancer, 114480 179617 OMIM RDH54 RAD54B 1E-121 Non-Hodgkin lymphoma; Colon adenocarcinoma 604289 OMIM SGS1 BLM 1E-115 Bloom syndrome, 210900 604610 OMIM, Cancer census MRE11 MRE11A 1E-108 Ataxia-telangiectasia-like disorder, 604391; Colorectal cancer 600814 OMIM, (Wang et al., 2004b) DUN1 CHEK2 6E-55 Li-Fraumeni syndrome, 151623; Osteosarcoma, somatic, 259500; susceptibility to Breast cancer, 114480; Prostate cancer, familial, 176807; susceptibility to colorectal cancer 604373 OMIM BUB1 BUB1 1E-41 Colorectal cancer with chromosomal instability 602452 OMIM MAD1 MAD1 5E-12 Lymphoma, somatic; Prostate cancer, somatic, 176807 602686 OMIM CDC73 HRPT2 9E-12 Hyperparathyroidism-jaw tumor syndrome, 145001; Hyperparathyroidism, familial primary, 145000; Parathyroid adenoma with cystic changes, 145001 607393 OMIM 68 F i g u r e 2.1 Three screen methods a) C T F screen method The genotypes of the donor strain and the MA Ta deletion mutants are shown within the schematic yeast cell respectively. The gray line with a gray circle depicts the CF. Each filled circle represents a selectable marker used in the SGA scheme, whereas an open circle in the corresponding color represents the unmarked allele. The ' X ' indicates mating between the 2 strains. The arrows indicates the selection procedures. Yeast cells enclosed by the dashed rectangles represent cells that were selected. The dashed arrow indicates the loss of a CF. Examples of severe and mild sectoring colonies, as well as wild-type colonies are shown. b) B i M screen method An example of an agar plate containing homozygous deletion mutants grown in 96-array format is shown. The yellow rectangles represents agar plates containing the specified media, and the indicated yeast strains were grown on them. The arrows indicates replica plating, pointing from the source plates to fresh destination plates. The blue blocked arrows indicate scanning, visual inspection and densitormetry reading of the Sc-6 plates, which selected for mated products. An example of a set of plates mated with MATa and MATa mating testers, respectively, is shown. Yeast cells enclosed by yellow squares had high mating with both MATa and MATa mating testers; whereas yeast cells enclosed by red and green squares had high mating with MATa and MATa mating testers, respectively, and these were not included in further analysis. The bottom shows an example of a retested plate containing 3 homozygous deletion mutants, each represented by 4 independent colonies, together with the wild-type diploid (YPH1738) and chllA/chllA. On the left is the SC-6 plate replica plated from the YPD plate with MA Ta mating tester (X MATa), and on the right is the SC-6 plate replica plated from the YPD plate with MA Ta mating tester (X MA Ta). c) A L F screen method The scheme of a-like faker selection is shown in the box, with the genotypes of the starting and selected strains indicated. The actual patching and replica plating procedures are shown below. An example of SC-6 plate with mated products derived from mating of the MA Ta mating tester with 12 deletion mutants, positive and negative controls on the same plate, is shown. bub3A, hst3A and yor024wA showed elevated mating frequencies with the MATa mating tester. The retest plate of these 3 mutants, each represented by 4 independent patches, is shown. 69 F i g u r e 2.1a CTF screen method Donor strain (YPH1725 or YPH1726) MA Ta deletion mutants Mating on YPD (1 day) (25°C for all incubation) X Diploid selection by SC-URA+G413 (twice: 2 days & 1day) Loss of CF Sporulation (9 days) Examples of spore genotypes Select MA Ta haploids with CF using Sc-HIS-UFtA-ARG+can (2 days) Select MATa haploids with CF, ade2-101 &ykoA by Sc-HIS-URA-ARG+can+G418+clonNAT (twice: 2 days each) Streak on media non-selective for CF (Sc with 20% ADE)to examine sectoring phenotype (6-7 days at 25°C &5-7 days at4°C) Severe sectoring Mild sectoring Wild-type 70 F i g u r e 2.1b B i M screen method Replica plating MA Ji HIS3 hisl mat rig tester KYPH315) lawn MAT*H)S3 Ms1 mat rig tester tYPH 316) lawri. homozygous diploid deletion mutants in 96-arrayfonrat on YPD+G418 (3 days) imTa vkoAhis3 msi Mf\Tay*oAhis3 HSS1 YPD+G418 Sc 6 (background control) Ljjvlating on YPD (2 days) Mated productj selection on Sc-6 (3 days) floating on YPD (2 days) Mated product] selection on Sc-6 (3 days) Scanning, visual inspection & densitometry reading Example of a 96-array on SC-6 media selectim for mated products • • High mathg rate with MATa only High mating rate wM MATa AND MATa High mathg rate with MATaonfy 1 i i—i I—i . rare Retest of candidates using 4 independent colonies: example X MATa X MATa 71 F i g u r e 2.1c A L F screen method MATaykoA his3 HIS1 Q"jD Loss of MATa locus | [null] ykoA his3 HIS1 (a-likefaker cell) \^J^ 1 C JO MATa HIS3 hi si (mating tester) ( JD MATa YKO HIS3 hisi (His+ mated product) MATaykoA his3HIS1 P a t c h e s ( Y P D overn igh t ) R e p l i c a onto l a w n of MATahisI ma t ing tes ter ( Y P D overn ight ) R e - p h e n o t y p e c a n d i d a t e s us ing 4 i n d e p e n d e n t c o l o n i e s I Fold Change Calculation mutant m e d i a n \4 p a t c h e s l wt m e d i a n [all w t p a t c h e s ] 72 Figure 2.2 Three marker loss screens A. Haploid yeast knockout mutants (ykoA) containing ade2-101 and a chromosome fragment (CF, blue line whose centromere is depicted as a circle) were generated. The CF contains the SUP I1 gene (blue rectangle) on the telocentric arm. ade2-l0l cells develop red color, but the SUP 11 gene on the CF suppress the pigment formation (Gerring et al., 1990; Hegemann et al., 1988; Warren et al., 2002). Cells that contain the CF are therefore unpigmented, whereas cells without the CF are red. Mutants that inherit the CF unstabliy form colonies with red sectors. An example was shown. The chromosomal context of the tested strains was shown at the bottom. Loss of SUP11 is caused by loss of the whole CF. B. Homozygous diploid yeast knockout mutants were tested for 'bimater' phenotype. For example, loss of the MA Ta allele (depicted in gray) causes the development of an a-type mating cell, which is detected by its ability to mate (depicted by 'X ' ) with a MA Ta tester strain containing complementing auxotrophy to support selection of mated diploids. Mutant strains exhibiting unstable inheritance of the MA T locus will lose either allele in individual cell and exhibit a 'bimater' phenotype in a population. The mutants in the squares show elevated formation of mated cells after exposure to either MA Ta or MA Ta testers (Spencer et al., 1990). The heterozygous MAT locus in the homozygous diploid knockout strains was shown. Loss of heterozygosity of the MAT locus can be caused by various mechanisms indicated C. MA Ta haploid yeast knockout mutants were tested for elevated frequency of a-like faker cells. Loss of the MA Ta locus (depicted in gray) in haploids results in dedifferentiation to a-mating type. The presence of these cells is detected by selection for mated products after exposure to a MA Ta tester strain. Examples of mated products formed from a wild-type and an alf mutant were shown CF Chrlll 5 UP11 MATa Chr III MATa Chr I MATa MATa C h r o m o s o m e transmission Diploid Bimater (BiM) fidelity (CTF) Chromosome Loss Chromosome Loss Rearrangement Gene Conversion Mitotic Recombination A-L ike Faker (ALF) Chromosome Loss Rearrangement Gene Conversion 73 F i g u r e 2.3 130 high confidence non-essential yeast C I N genes The Venn diagram shows the distribution of 293 mutants identified initially across the 3 screens. The numbers in parenthesis denote single-assay knockouts confirmed in independent transformants, which are included in the detailed diagram to the right. The detailed diagram to the right summarizes the 130 high confidence genes described in Appendix 1 (For the other 163 genes, see Appendix 2). Al l gene names are connected to 1 or more of the 3 screen nodes (CTF, BiM, or ALF), indicating the screen phenotypes the knockout mutants exhibit. Gene names in black typeface are those fully validated in all 3 deletion arrays: i.e. tag sequencing indicated the presence of only the correct mutation. For these mutant, a present or absent CIN phenotype observed in any screen is meaningful. Gene names in blue italic typeface failed tag validation in at least 1 of the deletion collections, and therefore phenotype information is missing from at least 1 screen. These partially characterized genes are placed to indicate positive phenotypes known from validated deletion sets (Appendix 1 and Appendix 2 contain details of the validation status). The node colors indicate biological process. Genes associated with more than one functional group is represented by the one highest in the color key for simplicity. CNN1^ P S 6 4 J ' IG1 126(11) ALF (122 total) (231 total) Biological Process Chromosome segregation 41 Cell cycle • DNA repair, DNA damage response, DNA recombination 41 DNA replication s j j k DNA metabolism, ™ chromosome organization 4) Transcription 41 Transport 3 41 Protein metabolism 4> Metabolism 4) Oxidative stress response 41 Unknown CIN8 YKL0S3' j i ^ o y y j K A R 3 GR064W' CTF19 L010W-A • U010C-©' 'LUOC PL017C IUP133 TAT P A 1 2 E S C 2 L32 1^RM3 JBR113W MRE11 ®£* WWSS1 • A D E 6 ( ^ I F T 1 74 F i g u r e 2.4 Functional groups o f 293 C L N genes GO biological process terms for the 293 marker loss genes were obtained using Osprey (vl.2.0) functional groupings (Breitkreutz et al., 2003b). For genes associated with more than one GO biological process, a single GO process was assigned according to the priority shown in the Figure 2.3 color key for simplicity. 75 F i g u r e 2.5 92 protein interactions among 102 CIN proteins Different types of protein interactions were shown among CIN proteins using GRID database and OSPREY program, indicated by arrows pointing from a bait/query. Color of the proteins represents the GO biological process. Known complexes or pathways were circled and annotated in red. Abbreviations: chkpt. (checkpoint); rec. (recombination); Ctr l , (control); mod. (modification). Microtubule Q Con Or oanutatwn and Braqenesls CeS Growth and or mamtenanca Prefoldin complex «wn»a)«»i 0 (tubulin folding) # mmmm Nuclear 5iLipi2<a<yp84j*jpi33 "jpore y complex YOR066^ I^C1 Cell & cycle ctrl. f NUP120NUP84 NI AlternativeSl« RFC (cohesion) Mitoj — Protein Protein toosytinnests sphotase^ complex igsomal £ Protein Degradation Transcription II DKA Replication DNA damage* / _ _ \ BRL3 ft SPT2 ISM6 . LSM1 Transcription IEA1 KGD1 JGR058W JTNL047C GRXS YDR279W YLR154C JCROSSg YPUH7W Post-replication Ubiquitination repair /""jJ6R1 / W ^ i RADlS\ ADE2 . SEO DNA ^ ^ 3 , • ^ — * 4 , 5 damage chkpt " RRM3 RA06S s Dtunap» nrj RNA splicing 0 ProteeaanrtnoecMldiowhoiyWion DHADamaoeRewonM 0 pr Meet transport 0 Transport % RNALocabauon Sponuuin s§l nuprricetiina 0 R*OMrn« Bionimesis 0 Carbonydrate uatabMism DSB repair MMS22 ~ RTT101 Y L R 3 2 0 t ^ ESC4 at / R A M I MFT1 T H P i T H I ' (3 CI con iplex _ y » c B R E i n L n n l x P e r l f n e n t a l Sys^ms MUS81 «riM1^ Vr5R100wV ' "Rec repair' .2 RAD59 YI ~3P ,§c. repair CDC73 ASF1 ITIOd. • MWenYPreCKTMION AT fBTITY CHROMATOGRAPHY PUNFIED COMPLEX I M M M 76 F i g u r e 2.6 A- l i ke fakers result from whole chromosome loss, gross chromosomal rearrangement, and gene conversion A. Individual colonies selected after mating were characterized using pulsed-field gel electrophoresis (top, ethidium bromide stained gel) and in-gel hybridization with a radiolabeled probe (bottom, autoradiogram) that hybridizes chromosomal bands containing the mating type locus and/or silent mating type loci located distally on each arm. Chromosomes III from the mating tester and deletion mutant were of distinct size (top and bottom chr III bands, respectively). In some strains, a less intense signal for rearrangement chromosomes reflects poor mitotic transmission or hybridization only to HMRa which has imperfect homology to the radiolabeled probe. B. Summary of electrophoretic karyotypes from 13 ALF mutants. ALF frequency is shown as fold over wild type. Event percentages (chr III loss, gross chromosome rearrangement (GCR), or retention of normal structure) are calculated from independent wild-type or mutant mated products (n=40 and n>10, respectively). The outcome distributions for sovlA and rad27A are significantly different from wild type (chi square, p < 0.01). C. Discordant CTF sectoring phenotypes are observed in knockout mutations with similar ALF frequencies. Ill * m m Fraction Fraction Fraction ykoA ykoA ykoA ALF Fold Chrlll Chrlll Chrlll Strain Change Loss GCR Retained Wild type 1 0.68 0.20 0.12 rad27A 63 0.30 0.60 0.10 dia2A 60 0.70 0.30 0.00 ybrll3wA 56 1.00 0.00 0.00 nce4A 56 0.70 0.20 0.10 xrs2A 48 0.86 0.14 0.00 esc2A 43 0.50 0.30 0.20 top3A 42 0.60 0.40 0.00 kar3A 40 1.00 0.00 0.00 rad50A 36 0.64 0.29 0.07 siclA 31 0.70 0.30 0.00 adelA 31 0.70 0.20 0.10 omalA 30 0.80 0.10 0.10 sovlA 8 0.00 0.00 1.00 77 F i g u r e 2.7. Common synthetic lethal interactions among yeast C I N genes that have human homologues mutated in cancer Eight yeast CIN genes with top hit human homologues mutated in cancer (e-value <10"10) are also found in the public interaction database BioGrid (Stark et al., 2006). They are placed peripherally, and shown in black. 61 interactors that have at least 2 synthetic lethal connections with the yeast cancer homologues are shown. The arrows point from a query to a target gene hit in the synthetic lethal screens. The 27 genes in blue fonts are themselves high confidence CIN genes identified by the screens, and 12 genes in purple fonts are in the lower confidence CIN gene list. The 3 genes that have 6 common synthetic lethal interactions 78 CHAPTER 3 Identification of Somatic Mutations in Cohesion Genes in Colorectal Cancers with Chromosome Instability A modified version of this chapter has been prepared for publication. Karen W . Y . Yuen*, Tom Barber*, Marcelo Reis*, K i r k McManus, Forrest Spencer, Bert Vogelstein, Victor Velculescu, Phi l Hieter, and Christoph Lengauer (*These authors contributed equally to this work). Identification of Somatic Mutations in Cohesion Genes in Colorectal Cancers with Chromosome Instability. 79 3.1 I n t r o d u c t i o n While the majority of colorectal cancers exhibit C I N , the molecular and genetic basis for this phenotype is not well characterized. Over the last decade, only a handful of genes known to be important for maintaining chromosome stability ( C I N genes) have been systemically tested and identified to have mutations in colorectal cancers, including BUB1, BUB1B (Cahill et al., 1998), CDC4 (Rajagopalan and Lengauer, 2004b), MRE11A, ZwlO, Zwilch, Rod and Ding (Wang et al., 2004b). A l l these genes were first identified based on chromosome segregation or cell cycle phenotypes in model organisms such as yeast and fly. These cross-species connections are examples of the high degree of evolutionary conservation in basic cellular mechanisms, and how basic biology studies in model organisms can be applied efficiently to gain an understanding of human disease. Recently, germline biallelic mutations in BUB IB have been associated with mosaic variegated aneuploidy and inherited predispositions to cancer, strongly supporting a causal link between C I N and cancer development (Hanks et al., 2004). However, each of the 8 genes mentioned above accounted for only a small fraction of the somatic mutation spectrum in colon cancer (1-10%), suggesting that more C I N genes could each be mutated to cause C L N in cancer. Systematic mutational analysis of C I N genes in colorectal cancers would therefore be useful to determine the complete mutational spectrum and mutation frequency leading to C I N . Indeed, the Lengauer/Vogelstein groups sequenced 100 candidate C I N genes (based on their similarity to yeast and fly genes) and identified 5 of the 8 genes mentioned above, suggesting that expansion of this kind of study would lead to the identification of additional relevant C I N genes (Wang et a l , 2004b). Among the 100 candidate human C I N genes pursued in the study, the best yeast homologue o f 30 human CLN genes yields the same human genes by reciprocally searching the best human homologue by B L A S T p (Table 3.1), outlining the number of yeast genes that have been used previously to identify human C I N gene mutation in cancer. In this study, the potential role of C I N genes in a panel of colorectal cancers was systematically analyzed by sequencing 101 candidate human C I N genes based on their 80 similarity to yeast homologues. 20 somatic mutations were identified genes that function in 4 functional groups. Seventeen o f the mutations were found in 5 genes that are directly involved in sister chromatid cohesion (SMC1L1, CSPG6, NIPBL, STAG2, and STAG3). Furthermore, single somatic mutations were identified in each of these 3 genes: BLM, the Bloom syndrome gene; RPN20, a E3 ubiquitin ligase; and UTX, a transcription factor. This study broadens the mutational spectrum of C I N genes in colorectal cancer. The results are consistent with a genetic basis for CLN, and with C L N having a role in tumorigenesis. Phenotypic analysis of a conserved missense mutation in yeast SMC1 revealed a modest recessive C L N phenotype in yeast cells. Further functional studies of the somatic mutations found w i l l enhance our understanding on whether these mutations cause C L N in cancer cells. 81 3.2 Materials and Methods 3.2.1 Gene identification 293 non-essential yeast C I N genes identified from recent comprehensive genome-wide screens (Yuen et al., submitted; see Chapter 2) were searched for human homologues by B L A S T p using Refseq database, and 88 human genes were selected based on the extent of sequence similarity (63 had a e-value <1E-10) and strength of the CLN phenotype in yeast. 64 were derived from the high confidence list (including 2 homologues of CHL1) and 24 were from the low confidence list. 2 non-essential CTF genes identified by traditional random mutagenesis (Spencer et al., 1990) but missed in the high-throughput screens (NUP170, RPB4) were also included; as wel l as 11 human genes homologous to essential yeast genes involved in chromosome transmission fidelity (Spencer et al., 1990) and cohesion. A total of 101 candidate C I N genes were analyzed. 3.2.2 Sequencing P C R primer design, amplification, sequencing, and sequence analysis were performed as previously described in (Sjoblom et al., 2006; Wang et al., 2004b). 3.2.3 Yeast smcl mutants construction and characterization Mutations were introduced into the yeast SMC1 gene by performing two rounds of P C R . One set of primary P C R product was amplified from wild-type (WT) genomic D N A using a sense primer ~500b upstream of the mutation, and an antisense primer containing the mutation in the middle; and a second primary P C R product was amplified with a sense primer containing the mutation in the middle and an antisense primer ~200b downstream of SMC7 that included a 20b homology to the T E F promoter. The primary P C R products were gel purified and used as template for a secondary round of P C R using the sense primer ~500b upstream of the mutation and the antisense primer ~200b downstream of SMC7 that included the 20b homology to the T E F promoter. The secondary P C R product was gel purified, cloned in Topo2.1 vector and sequenced. The insert was cut with flanking restriction enzymes and co-transformed with a P C R product 82 containing k a n M X , with flanking T E F promoter and terminator sequences, into a WT diploid strain (YPH986) containing CFIII (CEN3.L) . Transformants were selected on G418 and checked by colony P C R . The mutation was sequenced in the heterozygotes by P C R amplification with a primer upstream of the mutation and a kanMX-specific antisense primer. The heterozygotes were sporulated, and dissected to isolate haploid spore clones. The chromosome transmission fidelity (ctf) phenotype of both the heterozygous diploids and the haploids were checked at 25°C, 30°C, and 37°C (Spencer et al., 1990). Quantitative ctf assays were performed by counting the frequency of half-sectored colonies in haploids at 37°C on SC media with 20% adenine concentration as in (Shero et al., 1991). Yeast strains used in this study are listed in Table 3.2. A t least 3000 cells were plated, and the experiment was done in duplicate. 83 3.3 Results 3.3.1 20 somatic mutations were found in 8 CIN genes Based on recent comprehensive genome-wide screens of the yeast non-essential gene deletion set for C L N mutants (Yuen et al.; see Chapter 2), and previously identified essential yeast C L N genes identified by traditional random mutagenesis (Spencer et al., 1990), a list o f yeast C L N genes was complied and used to identify their human homologues based on protein sequence similarity. The list was prioritized based on the extent of yeast/human similarity and phenotype strength in yeast, and 101 human candidate C I N genes were selected for somatic mutation detection in a panel of 36 colorectal cancers (Table 3.3). In total, 1066 exons encode these 101 candidate genes, and corresponding primer pairs were used to P C R amplify these exons. After excluding known polymorphisms present in the human genomic database, all novel variants were resequenced using matched normal D N A from the patient to distinguish true somatic mutations from pre-existing variants. Somatic mutations in 8 genes (SMC7LI, CSPG6, NIPBL, STAG2, STAG3, BLM, UTX, and RNF20) were identified, 5 of which are involved in sister chromatid cohesion. To assess the frequency at which the genes are mutated in colon cancer, mutational analysis was expanded to an additional 96 colorectal cancer samples for 4 of the 8 genes (Table 3.4). This revealed that three o f the cohesion genes, SMC 1 LI, CSPG6 and NIPBL, have a mutation frequency o f - 3 . 8 % in colon cancers. 3.3.2 Mutation frequency in comparison to prevalence of mutations The identified somatic mutations could represent either 'passenger' mutations that occur as a consequence of tumorigenesis, or 'functional' mutations that underlie tumorigenesis. Mutations in genes with functional relevance are expected to occur at a frequency higher than random chance. Assuming that there is ~1.5kb of coding sequence per gene, ~5.4Mb was sequenced in the initial screening of 101 genes in -36 cancers (1.5kb X 101 genes X 36 tumors). A study by Wang et al. indicated that -1 nonsynonymous somatic change accumulates per M b of C L N tumor D N A , suggesting that 84 the mutation rate in C I N tumor cells is similar to that in normal cells and that most sporadic colorectal cancers do not display a mutator phenotype at the nucleotide level (Wang et al., 2002a). Another study found 151 mutations in 250Mb of D N A for a mutation rate of 0.6 changes per M b (Bert Vogelstein, unpublished). These data have significant implications for the interpretation of somatic mutations in candidate tumor-suppressor genes. Based on this estimation, ~3-5 mutations are expected from the first pass sequencing. Indeed, 8 somatic mutations were identified, suggesting that at least some (or all) of the mutations identified could be passenger mutations. In order to accurately determine the mutation frequency, the sequencing of 4 genes identified in the initial round was scaled up to an additional 96 tumors. Among the 4 genes, SMC 1 LI, CSPG6, and NIPBL are involved in cohesion, and one (BLM) functions in D N A repair. For each of these 4 genes, ~200kb genomic D N A was sequenced (1.5kb X 132 tumors). Based on the 1 mutation/Mb estimation, 0.2 mutations are anticipated for each gene. 5 mutations were identified in each o f the 3 cohesion genes, which is 19 times higher than expected. Such non-randomness in mutation pattern suggests that the mutations in cohesion genes are of functional relevance rather than "passenger" changes. However, no additional mutation was identified for BLM. 3 . 3 . 3 A conserved missense mutation in yeast SMC1 causes mild CIN In order to elucidate whether the mutations found can lead to C I N , human SMC1L1 was aligned with yeast SMC1 (Fig. 3.1a), and the analogous mutations found in human SMC 1 LI were constructed in yeast SMC1 (Fig. 3.1a,b). The smcl mutants were then assayed for chromosome transmission fidelity (CTF) by monitoring the loss of an artificial chromosome fragment in heterozygous diploid and haploid strains. The 1877 nonsense mutation, which results in truncation of the C-terminal region, led to lethality in a haploid background. VI1871, one of the 3 missense mutations is in the conserved ATPase domain at the C-terminal region. In haploids, this conserved mutation caused a 2-fold increase in chromosome loss at 37°C (Fig. 3.1c). Antisense inhibition of SMC1L1 in human fibroblast cells has been shown to lead to aneuploidy and chromosome 85 aberrations, as well as an increased frequency of micronuclei formation and apoptotic cells in long-term cultures (Musio et al., 2003). 86 3 . 4 D i s c u s s i o n 3 . 4 . 1 SMC1L1, CSPG6, NIPBL, STAG2 a n d STAG3 SMC 1 LI, CSPG6 and NIPBL (Delangin/SCC2) were each found to be mutated in 5 of the 132 tumor samples analyzed (i.e. a frequency o f -3 .8% for these genes), and STAG2 and STAG3 each had a single mutation in 36 tumors sequenced. SMC1L1 and CSPG6 (SMC3), together with SCC1 (MCD1IRAD21) and SCC3 (STAG1, STAG2 and STAG3 isoforms in vertebrates), form the essential cohesin complex, which is required for cohesion of sister chromatids and for accurate chromosome segregation (Tanaka et al., 2000). NIPBL, together with SCC4, forms an essential complex that loads cohesin to replicated sister chromatids during D N A replication (Ciosk et al., 2000). Recently, NIPBL and SMC1L1 were found to be mutated in Cornelia de Lange (CdL) syndrome, characterized by facial dysmorphisms, upper limb abnormalities, growth delay and cognitive retardation (Krantz et al., 2004; Musio et al., 2006; Tonkin et a l , 2004). NIPBL is expressed ubiquitously, but with variable tissue-specific expression. NIPBL not only has a role in cohesion, but also functions in developmental regulation by affecting gene expression (Rollins et al., 2004); whether these functions are independent from each other, and how much NIPBL expression is required for each function remains unknown. Recently, CTF7IESC02, an acetyltransferase required for cohesion establishment in S phase, was found to be mutated in Roberts syndrome and SC phocomelia, which has several phenotypes overlapping with Roberts syndrome (Schule et al., 2005; Vega et al., 2005). Precocious sister chromatid separation has been described in C d L syndrome, Roberts syndrome, and mosaic variegated aneuploidy, and various cancers (Kaur et al., 2005). However, patients of C d L and Roberts syndromes do not have cancer predisposition. The present study is the first report identifying mutations in genes functioning in cohesion in human cancers. 3 . 4 . 2 BLM BLM was found to be mutated in one out of 132 tumor samples analyzed. BLM, together with WRN and RECQL4, are homologous to yeast SGS1 in the RecQ helicase 87 family, and all 3 genes are mutated in cancer prone syndromes (see Table 1.1). BLM is a D N A structure-specific helicase, which plays a role in the resolution of D N A structures that arise during the process of homologous recombination repair, by catalyzing Holliday-j unction branch migration and annealing of complementary single-stranded D N A molecules (reviewed in (Cheok et al., 2005)). In the absence of BLM, cells show genomic instability and a high incidence of sister-chromatid exchanges. Gruber et al. (Gruber et al., 2002) determined that carriers of the BLMiAsh) founder mutation (causing frameshift and truncation) have an increased risk of colorectal cancer, and they also observed a low frequency of the BLM(Ash) mutation in lymphoma, breast, ovarian, and uterine cancers. M i c e heterozygous for a null mutation of BLM develop lymphoma earlier than wild-type littermates in response to challenge with murine leukemia virus, suggesting that BLM haploinsufficiency is associated with tumor predisposition (Goss et al., 2002). However, further functional analysis is required to determine whether the heterozygous missense mutation found in this study causes a C L N phenotype. 3 . 4 .3 RNF20 One heterozygous missense mutation was found in RNF20, which encodes a E3 ubiquitin ligase. RNF20 forms a complex with RNF40, interacts with an ubiquitin E2-conjugating enzyme UBCH6, and establishes H2B lysine 120 monoubiquitylation, which is associated with transcriptional activity (Pavri et al., 2006; Zhu et al., 2005). This modification subsequently regulates H2B methylation and expression of homeobox ( H O X ) genes, which are required for proper development. In yeast, H2B ubiquitylation by RAD6(E2)-BRE1(E3) has been shown to be required for the D N A damage checkpoint response (Giannattasio et al., 2005). Interestingly, yeast-two-hybrid analysis of the yeast homologue of RNF20, BRE1, indicated that it interacts with the coiled-coil region of 3 proteins in the structural maintenance of chromosomes family (SMC1, SMC2, SMC3) and some kinetochore components (SLK19, NUF2) (Newman et al., 2000), and may play a direct role in chromosome maintenance. 88 3.4.4 UTX One heterozygous missense mutation was found in UTX. UTX is a transcription factor that contains the tetratricopeptide repeat (TPR) motif. The yeast homologue SSN6ICYC8, together with TUP I, is involved in histone deacetylation, which is associated with transcriptional repression. SSN6 functions as a negative regulator of the expression of a broad spectrum of genes (reviewed in (Smith and Johnson, 2000)), which explains why the phenotype of ssn6 mutant is pleiotropic. ssn6 mutants exhibits a modest effect on the maintenance of minichromosomes (Schultz et al., 1990). While it is evident that the role of cohesion genes and BLM in chromosome maintenance is wel l conserved, it is still unclear whether RNP20 and LVTXplay a direct role in chromosome maintenance in yeast or in human. Further functional studies in model organisms and in mammalian cells w i l l delineate the roles o f these candidate C I N genes, reveal whether the mutations cause C I N , and elucidate the degree of functional conservation. This study initiated the characterization of the cancer somatic mutations in SMC 1 LI by introducing the corresponding mutations in yeast SMCL The conserved mutation caused a mild C I N phenotype in haploids. Although the heterozygous diploids of this conserved mutation or the truncation do not exhibit a detectable C I N phenotype in yeast, it is possible that human cells may be more sensitive to perturbations in C L N genes, due to larger chromosome and genome size. A precedent for this hypothesis is that mitotic checkpoint components are dispensable in yeast, but are essential in higher eukaryotes (Babu et al., 2003; Baker et al., 2004; Dai et al., 2004; Kitagawa and Rose, 1999; Kops et al., 2004; Miche l et al., 2001). The results presented here broaden the mutational spectrum of colorectal cancers, and are consistent with previous observations that each C I N gene is mutated at a low frequency (-3.8% for each of the 3 cohesin related genes). Therefore, a variety of C I N genes could each be responsible for a small proportion of cancers. C L N is a hallmark of most solid tumors, so it w i l l be of interest to compare the mutational spectrum for 89 colorectal cancers to that of other types of solid cancers. The technical information gained in colorectal cancer w i l l be useful for similar analysis in other tumor types. Mutational spectra may also allow classification of tumors, which could have implications for improved diagnosis, prognosis, or predictions of response to therapy. For example, the Vogelstein group recently pursued another large-scale mutational analysis of 13,023 genes in 11 breast and 11 colorectal cancers (Sjoblom et al., 2006). This unbiased study provided an estimate of the total number of nonsynonymous mutations that arise in a typical cancer, thereby allowing statistical differentiation between passenger mutations and mutations with functional implications. Their study revealed 189 genes that were mutated at significant frequency. These cancer genes encode a wide range of cellular functions, including genes that have been shown to be somatically mutated or implicated in tumorigenesis in expression studies, but also many genes that were not previously suspected to contribute to the pathogenesis of cancer. Interestingly, that study also identified substantial differences in the mutational spectra in different tumor types. Knowing the mutational spectra in cancers would be useful for therapeutic design. A s described in Chapters I and II, a second-site loss-of-function mutation that causes synthetic lethality with a C L N mutation can lead to selective ki l l ing of tumor cells. Combining synthetic lethal data available in yeast (Pan et al., 2006; Tong et al., 2004) with mutational spectrum of colorectal cancers found in studies like the one presented here may help to highlight potential targets for this therapeutic strategy. Indeed, mutations of 4 non-essential cohesion genes [ctf4A, ctf8A, ctfl8A, and dcclA) are synthetically lethal with mutations of 5 different C I N gene homologues which are mutated in colorectal cancer (Fig. 3.2). 3 of the 4 same cohesion gene mutations (ctf4A, ctf 18 A, and dcclA) are also synthetically lethal with mutations of C L N gene homologues which are mutated in other cancer types (Fig 2.9). Such analyses suggested that these common synthetic lethal interactors may be attractive drug targets that are effective to a broad spectrum o f C L N tumors with C I N gene mutations. 90 T a b l e 3.1 Re la t ionsh ip o f 100 human candidate C I N genes used i n ( W a n g et a l . , 2004b) w i t h yeast genes The 100 human candidate CIN genes were searched for the best yeast homologue by BLASTp according to the proteome database (www.proteome.com), and the e-value is indicated. The list is sorted by the human gene name in ascending order. 19 genes contain only hCT# but not Genbank accession #, so the gene identities of these human genes are not known (highlighted in gray). 3 genes (ATR, POLE, and RAD51C) have 2 different hCT#, representing different isoforms (highlighted in orange). 13 human genes have no BLASTp hit in yeast (highlighted in purple, or blue, which have putative homologs with an e-value >lE-3). 66 human genes yield a yeast hit by BLASTp with an e-value <lE-3. The best yeast hits for these 66 genes were reciprocally searched for their best human homologue by BLASTp. Among these, 30 of them were reciprocal best hits (highlighted in yellow). 15 of the yeast best hits were found in the CIN screens described in Chapter 2 (highlighted in red), and 10 of the 15 are reciprocal best hits; 3 of them correspond to family members of yeast DUN1. 91 Table 3.1 Page 1 of 2 No. Celera accession Genbank accession Human Gene Other gene name Top yeast hit E-value Reciprocal top human hit E-value Yeast hit found in CIN screens? 5 hCT12678 NM 005883 APCL APC2 VAC8 5E-04 ARMC3 1E-23 VAC8 90 hCT7133 NM 014840 ARK5 SNF1 2E-60 PRKAA2 1E-113 42 hCT1826039 NM 001184 ATR . MEC1 1E-108 ATR 1E-108 43 hCT1826040 NM 001184 ATR MEC1 1E-108 ATR 1E-108 1 hCT10388 NM 016374 BCAA ARID4B; BRCAA1 N/A 80 hCT31470 NM 006768 BRAP YHL010C 7E-54 BRAP 7E-54 9 hCT14094 NM 020439 CAMK1G VWS1 CMK2 1E-59 CMK1D 1E-60 99 hCT9356 NM 172080 CAMK2B CMK2 1E-52 CMK1D 1E-60 19 hCT1643963 NM 001254 CDC6 CDC6 8E-33 CDC6 8E-33 CDC6 -37 hCT1816212 NM 001813 CENPE KIP3 2E-54 KIF18A 1E-81 46 hCT18305 NM 022909 CENPH N / A * .*.. . V . 72 hCT30161 NM 001274 CHK1 CHK1 9E-45 CHEK1 9E-45 82 hCT32245 NM 007194 CHK2 DUN1 3E-50 DCAMKL1 2E-53 DUN1 -21 hCT1646711 NM 001340 CYLC2 YFR016C 4E-09 NEF2 1E-10 11 hCT14628 NM 001348 DAPK3 CMK1 1E-48 CMK1D 9E-63 61 hCT23387 NM 004734 DCAMKL1 DUN1 2E-53 DCAMKL1 2E-53 DUN1 54 hCT20446 NM 015070 DING N/A. . PDS1 ? 85 hCT32971 NM 007068 DMC1 DMC1 1E-99 DMC1 1E-99 55 hCT20952 NM 000123 ERCC5 RAD2 7E-47 ERCC5 7E-47 81 hCT32115 NM 004111 FEN1 RAD27 RAD27 1E-104 FEN1 1E-104 RAD27 _ . 51 hCT1961597 NM 017975 FLJ10036 N/A 3 hCT11790 NM 012415 FSBP RAD54B RDH54 1E-122 RAD54B 1E-122 RDH54 31 hCT17786 NM 014635 GCC185 NUM1/PAC12 9E-08 RSN 3E-09 96 hCT8974 HCA127 N/Aft-K'./:^ • 86 hCT401149 NM 014586 HUNK SNF1 1E-51 PRKAA2 1E-113 16 hCT16364 NM 014915 KIAA1074 AKR2 9E-09 ZDHHC17 1E-48 70 hCT29475 NM 032430 KIAA1811 BRSK1 SNF1 2E-76 PRKAA2 1E-113 58 hCT2308143 NM 014708 KNTC1 ROD N/A 35 hCT1788172 LATS1 CBK1 1E-100 STK38I 1E-131 32 hCT1783089 NM 003550 MAD1L1 N/A MAD1 ? MAD1 57 hCT22552 NM 014791 MELK SNF1 2E-65 PRKAA2 1E-113 98 hCT9098 NM 152619 MGC45428 DUN1 4E-52 DCAMKL1 2E-53 DUN1 13 hCT14856 NM 016195 MPH0SPH1 CIN8 6E-37 KIF11 1E-48 CIN8 . . . . . . . 59 hCT2334792 NM 005591 MRE11A MRE11 MRE11 1E-104 MRE11 1E-104 MRE11 .. 65 hCT24254 NM 002485 NBS1 N/A XRS2 ? 84 hCT32914 NM 021076 NEFH CHS5 6E-13 NEFH 6E-13 64 hCT23665 NM 004153 0RC1L ORC1 2E-39 ORCL1 2E-39 77 hCT30866 NM 177990 PAK7 STE20 5E-80 PAK1 1E-111 74 hCT30362 NM 002592 PCNA POL30 1E-52 PCNA 1E-52 88 hCT6664 PIK3C2A VPS34 3E-48 PIK3C3 1E-139 91 hCT7448 NM 002646 PIK3C2B VPS34 3E-48 PIK3C3 1E-139 7 hCT13660 NM 002647 PIK3C3 VPS34 VPS34 1E-139 PIK3C3 1E-139 89 hCT7084 NM 006219 PIK3CB VPS34 2E-47 PIK3C3 1E-139 34 hCT1787138 NM 005026 PIK3CD VPS34 2E-57 PIK3C3 1E-139 92 hCT7976 NM 002649 PIK3CG VPS34 7E-51 PIK3C3 1E-139 8 hCT14027 NM 002691 P0LD1 CDC2 0E+00 POLD1 0E+00 63 hCT23655 NM 006231 R O L E _ . - POL2 0E+00 POLE 0E+00 95 hCT87415 NM 006231 ROLE . . . . . . POL2 0E+00 POLE 0E+00 100 hCT9836 NM 006904 PRKDC TOR1 1E-34 FRAP1 0E+00 68 hCT28965 NM 133377 RAD1 : N/At^.- . :„ , . ; 66 hCT28290 NM 133338 RAD17 RAD24 4E-17 RAD17 4E-17 RAD24 1 •' .. 14 hCT15239 NM 005732 RAD50 RAD50 2E-59 RAD50 2E-59 RAD50 73 hCT30207 NM 133487 RAD51 RAD51 1E-112 RAD51 1E-112 RAD51 25 hCT1686635 NM 058216 RAD51C " DMC1 3E-18 DMC1 1E-99 28 hCT1767458 NM 058216 RAD51C DMC1 3E-18 DMC1 1E-99 49 hCT18816 NM 133627 RAD51L3 DMC1 1E-13 DMC1 1E-99 24 hCT1686440 NM 134422 RAD52 RAD52 5E-41 RAD52 5E-41 RAD52 6 hCT13183 NM 003579 RAD54L RAD54 1E-160 RAD54L 1Er160 RAD54 71 hCT29790 NM 002913 RFC1 RFC1 1E-115 RFC1 1E-115 97 hCT9089 NM 002914 RFC2 RFC4 1E-109 RFC2 1E-109 60 hCT23382 NM 002915 RFC3 RFC5 2E-67 RFC3 ' 2E-67 40 hCT1823014 NM 002916 RFC4' RFC3 5E-42 RFC5 5E-80 78 hCT30904 NM 007370 RFC5 RFC3 5E-80 RFC5 5E-80 52 hCT19876 NM 002945 RPA1 RFA1 1E-92 RPA1 • . 1E-92 62 hCT23494 NM 012238 SIRT1 SIR2 1E-56 SIRT1 1E-56 2 hCT11285 SIRT2 HST2 2E-59 SIRT3 4E-60 69 hCT29050 NM 012239 SIRT3 HST2 4E-60 SIRT3 4E-60 50 hCT18916 NM 176827 SIRT4 HST2 7E-14 SIRT3 4E-60 92 Table 3.1 Page 2 of 2 67 hCT28652 NM 012241 SIRT5 SIR2 9E-16 SIRT1 1E-56 12 hCT14647 NM 016539 SIRT6 HST2 3E-12 SIRT3 4E-60 33 hCT1786284 NM 016538 SIRT7 HST1 3E-10 SIRT1 5E-48 56 hCT21449 NM 018225 SMU-1 PFS2 7E-17 WRD33 9E-70 36 hCT17934 AA447812 SNRK SNF1 3E-53 PRKAA2 1E-113 87 hCT6634 NM 007027 T0PBP1 N/A 83 hCT32452 NM 003292 TPR AGA1 2E-04 MUC17 3E-25 48 hCT18373 NM 004628 XPC RAD4 3E-26 XPC 3E-26 75 hCT30596 NM 006297 XRCC1 N/A 23 hCT16627 NM 005432 XRCC3 DMC1 2E-19 DMC1 1E-99 76 hCT30844 NM 004724 ZW10 N/A , 79 hCT31391 NM 007057 ZWINT N/A . 39 hCT1817729 NM 012291 ESPL1/SEPARASE ESP1 1E-36 ESPL1 1E-36 4 hCT12352 10 hCT14327 15 HCT15320 17 hCT1642589 18 KCT1643619 20 hCT1644019 22 hCT1657158 26 hCT173001 27 hCT1766645 29 hCT1770914 30 hCT1775724 38 hCT1817706 41 hCT1824077 44 hCT1829493 45 hCT1829782 47 hCT1834200 53 hCT201497 93 hCT87379 94 hCT87385 93 T a b l e 3.2 List o f yeast strains used in Chapter 3 Strain Genotype YPH986 MATa/MATaura3-52/ura3-52 trplA-63/trplA-63 his3A-200/his3A-200 leu2A-l/leu2A-l ade2-101/ade2-101 Iys2-801/lys2-801 CFIII(CEN3.L)-HIS3 SUP 11 YKY1038 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 smcl-Q449W::kanMX CFIII(CEN3.L)-HIS3 SUP 11 YKY1042 MATaura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 smcl-Q449W::kanMX CFIII(CEN3.L)-HIS3 SUP 11 YKY1053 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 smcl-L574M::kanMX CFIII(CEN3.L)-HIS3 SUP 11 YKY1051 MATaura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 smcl-L574M::kanMX CFIII(CEN3.L)-HIS3 SUP 11 YKY1034 MATaura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 smcl- VI 187I::kanMX CFIII(CEN3.L)-HIS3 SUP 11 YKY1031 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 smcl- VI 1871::kanMX CF11I(CEN3.L)-HIS3 SUP 11 YKY1011 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 SMCl::kanMX CFIII(CEN3.L)-HIS3 SUP 11 YKY1023 MATaura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 SMClr.kanMX CFIII(CEN3.L)-HIS3 SUP 11 94 T a b l e 3.3 101 candidate C I N genes ana lyzed i n this study 88 human genes were selected from the non-essential yeast CEN genes based on the extent of yeast/human similarity (63 had a e-value <1E-10) and CIN phenotype strength in yeast. 2 human genes were selected based on their similarity to 2 non-essential CTF genes identified by traditional random mutagenesis (Spencer et al., 1990) but missed in the high-throughput screens. Another 11 human genes were selected because they have sequence similarity to essential yeast CTF genes identified by traditional random mutagenesis (Spencer et al., 1990). Most of these human genes correspond to the top hits of yeast CIN genes, except for 5 human genes which are second or third hits (families are indicated in orange). In total, 101 candidates were analyzed, 74 of which had a e-value <1E-10. The list is sorted in ascending order by the yeast gene name. 15 candidate genes are involved in cohesion (indicated in yellow). 95 Table 3.3 Page 1 of 2 No. Yeast ORF Yeast gene Essential? ctf bim alf Top human hit E-value Protein Accession mRNA Accession 1 YAR015W ADE1 wrong 0 31 phosphoribosylaminoimidazole c 6E-08 NP 006443 NM 006452.2 2 YGL234W ADE5.7 0 0 12 phosphoribosylglycinamide form 0E+00 NP 000810 NM 000819.3 3 YGR061C ADE6 0 0 16 . phosphoribosytformylQlycinamidi 1E-176 NP 036525 NM 012393.1 4 YBR231C AOR1 0 4 0 craniofacial development proteir 5E-10 NP 006315 NM 006324.1 5 YLR085C ARP6 0 3 0 ARP6 actin-related protein 6 hor 2E-44 NP 071941 NM 022496.2 6 YJL115W ASF1 wrong 3 22 ASF1 anti-silencing function 1 h< 8E-51 NP 054753 NM 014034.1 7 YER016W BIM1/CTFS3 3 5 16 microtubule-associated protein. 2E-28 NP 036457 NM 012325.1 8 YDL074C BRE1 0 0 11 RPN20, ring finger protein 20; he 5E-26 NP 062538 NM 019592.5 9 YOR026W BUB3 2 5 24 BUB3 budding uninhibited bv be 2E-24 NP 004716 NM 004725.1 10 YJL194W CDC6/CTF10 Ess 3 #N/A #N/A CDC6 homolog; CDC18 (ceil div 2E-32 NP 001245 NM 001254.2 11 YLR418C CDC73 1 0 2 parafibromin; chromosome 1 op< 9E-12 NP 078805 NM 024529.3 12 YGL003C CDH1 3 3 0 Fzr1 protein; fizzy-related proteii 1E-92 NP 057347 NM 016263.2 13 YPL008W CHL1/CTF1 3 5 7 . OEAD/H (Asp-Glu-Ala-Asp/His) 1E-112 NP 085911 NM 030653.2 14 YPL008W CHL1/CTF1 3 5 . 7 CHLR2/DDX12, DEAD box prott 4E-94 AAB06963.1 U33834 15 YMR198W CIK1 3 wrong 0 golgi autoantigen, golgin subfarr 4E-07 NP 002069 NM 002078.3 16 YEL061C CIN8 3 4 0 kinesin family member 11; Eg5; 2E-67 NP 004514 NM 004523.2 17 YPR120C CLB5 0 3 4 evelin B1; G2/mitotic-specific cy( 6E-39 NP 114172 NM 031966.2 18 YMR048W CSM3 3 5 12 timeless-interacting protein; tipin 8E-08 NP 060328 NM 017858.1 19 YMR078C CTF 18 3 0 29 CTF18, chromosome transmissi 3E-36 NP 071375 NM 022092.1 20 YLR381W CTF3 3 4 0 LRPR1 (CENPI), follicle-stimulat see Measday V, 20C NP 006724 NM 008733.2 21 YPR135W CTF4 3 5 27 WD repeat and HMG-box DNA t 1E-18 NP 009017 NM 007086.1 22 YHR191C CTF8 3 6 18 hCTF8, hypothetical protein MG see Mayer M, 2001 NP 00103523 NM 001039690 23 YJL006C CTK2 3 wrong contam cyclin K (Homo sapiens] 2E-11 NP 003849 NM 003858.2 24 YCL016C DCC1 0 5 18 hypothetical protein MGC5528 [I 8E-11 NP 076999 NM 024094.1 25 YIR004W DJP1 0 0 13 DnaJ (Hsp40) homolog, subfam 4E-18 NP 061854 NM 018981.1 26 YGL240W DOC1 3 nd 3 anaphase-promoting complex si 5E-22 NP 055700 NM 014885.1 27 YFR027W EC01/CTF7 Ess 3 #N/A #N/A establishment factor-like protein 5E-11 NP 443143 NM 052911.1 28 YFR027W EC01/CTF7 Ess 3 #N/A #N/A ESC02, Establishment of cohes 3E-07 NP 001017420.1 NM_001017420 29 YOR144C ELG1 1 4 12. hypothetical protein FLJ12735 [f 2E-05 NP 079133 NM 024857.3 30 YBR026C ETR1 wronq 5 5 nuclear receptor-binding factor 1 5E-47 NP 057095 NM 016011.1 31 YEL003W GIM4 0 contam 19 prefoldin 2 [Homo sapiens] 1E-15 NP 036526 NM 012394.2 32 YCR065W HCM1 0 4. 2 forkhead box 12; Blepharophimc 1E-20 NP 075555 NM 023067.2 33 YBR009C HHF1 0 4 6 histone 2, H4; H4 histone, family 2E-37 NP 003539 NM 003548.2 34 YPR067W ISA2 contam 3 3 HESB like domain containing 1 [ 2E-05 NP 919255 NM 194279.1 35 YPR141C KAR3 3 nd . 40 kinesin family member C1 (Hom< 2E-69 XP 371813 XM 371813.1 36 YDR532C KRE28 0 4 . 30 retinobiastoma-binding protein 1 3E-06 NP 002883 NM 002892.2 37 YDR378C LSM6 1 2 0 Sm protein F (Homo sapiens] 5E-09 NP 009011 NM 007080.1 38 YJL030W MAD2 2 . 3 0 MAD2-like 1; MAD2 (mitotic arre 5E-38 NP 002349 NM 002358.2 39 YPR046W MCM16/CTFS155 3 4 . . 4 , high density lipoprotein binding r. 2E-03 NP 976221 NM 203346.1 40 YDR318W MCM21/CTF5 3 ... 5 3 SCC1/MCD1, RAD21 homolog; 3E-02 NP 006256 NM 006265.1 41 YFL016C MDJ1 0 0 19 . DnaJ (Hsp40) homolog, subfam 2E-31 NP 005138 NM 005147.3 42 YOL064C MET22 0 3 . wrong inositoKmyoM (or 4)-monophosr 4E-05 NP 005527 NM 005536.2 43 YOR241W MET7 0 wrong 13 .'. folylpolyglutamate synthase; foly 1E-11 NP 004948 NM 004957.2 44 YDR386W MUS81 0 4 . . , 9 MUS81 endonuclease homolog 1E-18 NP 079404 NM 025128.3 45 YBL079W NUP170/CTFS141 3 •' #N/A #N/A nucleoporin 155kDa isoform 1; r 2E-30 NP 705618 NM 153485.1 46 YDL116W NUP84 0 4 contam nuclear pore complex protein [H 5E-16 NP 065134 NM 020401.1 47 YKL055C OAR1 wrong wrong B . . . . DKFZP566O084 protein [Homo 2E-13 NP 056325 NM 015510.3 48 YKR087C OMA1 wrong 2 •' 30 metalloprotease related protein 6E-29 NP 660286 NM 145243.2 49 YGR078C PAC10 0 2 3 von Hippel-Lindau binding prote 4E-30 NP 003363 NM 003372.3 50 YHR064C PDR13 1 2 3' • heat shock 70kDa protein 8 isofc 2E-56 NP 006588 NM 006597.3 51 YMR076C PDS5/CTF11 Ess 3 #N/A #N/A androgen-induced prostate proli 1E-29 NP 055847 NM 015032.1 52 YLR273C PIG1 3 0 0 protein phosphatase 1, regulato 2E-06 NP 005389 NM 005398.3 53 YOL054W PSH1 0 4 11 tripartite motif-containing 25; Zin 1E-07 NP 005073 NM 005082.3 54 YPL022W RAD1 0 contam 5 excision repair cross-compleme 1E-109 NP 005227 NM 005236.1 55 YML095C RADIO 0 3 2 excision repair cross-compleme 1E-12 NP 001974 NM 001983.2 56 YCR066W RAD18 wrong 5 19 postreplication repair protein hR, 9E-20 NP 064550 NM 020165.2 57 YER173W RAD24 0 4 6 RAD17 homolog isoform 2; Rad 8E-17 NP 579917 NM 133339.1 58 YLR032W RAD5 0 2 5 SWI/SNF-related matrix-associa 5E-70 NP 620636 NM 139048.1 59 YER095W RAD51 wrong 4 7 RAD51 homolog protein isoform 1E-122 NP 002866 NM 002875.2 60 YML032C RAD52 2 5 22 RAD52 homolog isoform alpha; 2E-40 NP 002870 NM 002879.2 61 YDR076W RAD55 0 4 18 RAD51-like 3 isoform 1; recomb 1E-05 NP 002869 NM 002878.2 62 YDR004W RAD57 0 5 15 RAD51-like 1 isoform 3; RecA-li 4E-19 NP 598193 NM 133509.2 63 YDL059C RAD59 0 4 2 RAD52 homolog isoform alpha; 8E-09 NP 002870 NM 002879.2 64 YGL058W RAD6 1 nd 25 ubiquitin-conjugating enzyme E2 7E-61 NP 003327 NM 003336.2 65 YDR014W RAD61/CTF6 3 4 5 hypothetical protein LOC57821 4E-03 NP 067002 NM 021179.1 66 YDR217C RAD9 0 3 10 dentin sialophosphoprotein prep 1E-07 NP 055023 NM 014208.1 67 YNL072W RNH35 0 3 6 ribonuclease HI, large subunit [H 7E-46 NP 006388 NM 006397.2 68 YJR063W RPA12 0 2 5 zinc ribbon domain containing. 1 1E-14 NP 740753 NM 170783.1 96 Table 3.3 Page 2 of 2 69 YJL140W RPB4/CTF15 3 #N/A #N/A DNA directed RNA polymerase I 3E-12 NP 004796 NM 004805.2 70 YIL018W RPL2B wronq 0 17 ribosomal protein L6; 60S ribosc 1E-102 NP 150644 NM 033301.1 71 YDL204W RTN2 0 0 16- reticulon 2 isoform A: NSP-like p 5E-08 NP 005610 NM 005619.3 72 YOR014W RTS1 wronq 0 19 delta isoform of regulatory subur 1E-148 NP 006236 NM 006245.2 73 YJL047C RTT101 wronq wronci 15. cullin 2 [Homo sapiens] 9E-06 NP 003582 NM 003591.2 74 YDR289C RTT103 2 0 0 chromosome 20 open reading fr 3E-14 NP 067038 NM 021215.2 75 YDR159W SAC3 contam wronq 5 minichromosome maintenance c 4E-26 NP 003897 NM 003906.3 76 YDR180W SCC2/CTF12/CHL8 Ess 3 • #N/A #N/A NIPBL. IDN3 protein isoform A [I 3E-19 NP 597677 NM 015384.3 77 YIL026C SCC3/IRR1/CTFs16! Ess 3 #N/A #N/A STAG1, stromal antigen 1; nuclc 2E-21 NP 005853 NM 005862.1 78 YIL026C SCC37IRR1/CTFS163 Ess 3 . , #N/A #N/A STAG3, Stromal antigen 3 (stror 3E-13 NP 038579.2 NMJM2447 79 YIL026C SCC3/lRR1/CTFs16d Ess 3 #N/A #N/A STAG2, Stromal antigen 2. a me 2E-11 NP 006594.3 NM_O066O3 60 YMR190C SGS1 0 4. 14 . Bloom syndrome protein [Homo 1E-115 NP 000048 NM 000057.1 81 YLR058C SHM2 0 4 2 serine hydroxymethyltransferasr 1E-148 NP 004160 NM 004169.3 82 YBL058W SHP1 2 4 2 p47 protein isoform a THorno sat 6E-34 NP 057227 NM 016143.3 83 YLR079W S1C1/CTFS127 3 6 . 31 hypothetical gene supported bv 7E-02 NP 963859 NM 201565.1 84 YER116C SLX8 1 ' , 5 wronq . ring finger protein 10 [Homo sap 7E-08 NP 055683 NM 014868.3 85 YFL008W SMC1/CTFS166. . Ess 3 #N/A #N/A SMC1 structural maintenance of 1E-149 NP 006297 NM 006306.2 86 YFL008W SMC1/CTFS166 Ess 3 #N/A #N/A SMC1L2, Protein with strong sirr1 SE-41 NP 683515.3 NM_148674 87 YJL074C SMC3 Ess #N/A #N/A #N/A CSPG6, Chondroitin sulfate prot 1E-45 NP 005436.1 NM_0O5445.3 88 YOR308C SNU66 2 0 0 squamous cell carcinoma antige 2E-07 NP 005137 NM 005146.3 89 YGR063C SPT4/CTFS138 1 0 0 suppressor of Ty 4 homolog 1 rh 1E-18 NP 003159 NM 003168.1 90 YPR032W SR07 wronq 2 3 tomosvn-like [Homo sapiensl 4E-14 XP 045911 XM 045911.8 91 YBR112C SSN6/CYC8 0 nd 80 UTX, ubiquitously transcribed te 9E-44 NP 066963 NM 021140.1 92 YOL072W THP1 wronq 4 19 hypothetical protein FLJ11305 [>• 6E-07 NP 060856 NM 018386.1 93 YNL273W TOF1 3 5 10 timeless homolog [Homo sapien 5E-07 NP 003911 NM 003920.1 94 YLR234W TOP3 0 5 - : 42 topoisomerase (DNA) III alpha; t 1E-124 NP 004609 NM 004618.2 95 YAL016W TPD3 nd nd 75 beta isoform of regulatory subun 1E-133 NP 859050 NM 181699.1 96 YML028W TSA1 0 4 8 peroxiredoxin 2 isoform a; thiore 7E-71 NP 005800 NM 005809.4 97 YGR184C UBR1 3. . 3 0 ubiquitin ligase E3 alpha-ll; likely 9E-28 NP 056070 NM 015255.1 98 YDL156W YDL156W 0 2 2 hypothetical protein FLJ12973 [h 3E-19 NP 079184 NM 024908.1 99 YLR193C YLR193C 0 2 2 similar to Px19-like protein (25 k 2E-22 XP 371496 XM 371496.2 100 YGR270W YTA7 wi-ohq 4 6 two AAA domain containing prot 1E-131 NP 054828 NM 014109.2 101 YGR285C ZU01 0 4 wronq similar to M-phase phosphoproK 1E-48 XP 379909 XM 379909.1 97 T a b l e 3.4 Somat ic mutations identif ied i n candidate C I N genes i n C I N colorectal cancer cel ls 20 mutations were identified in 8 genes. The number of mutations found and the number of tumors sequenced are indicated. The nucleotide mutation and corresponding amino acid mutation are indicated by the position based on the coding sequence (CDS), followed by the wild-type sequence>tumor sequence. Thus, 2 different nucleotides in the tumor sequence represent a heterozygous mutation, and a single nucleotide in the tumor sequence represents a homozygous or hemizygous mutation. The corresponding amino acid change is indicated in the same manner. The corresponding tumor, exon number and primer used are indicated. The yeast gene used as query and the e-value are also indicated. Gene No. Human Gene Name Yeast Gene Name e-value • Human RNA ID Human Protein ID No; mutations . No. tumors sequenced PRIMER Exon Somatic Mutation Tumor 1 SMC1L1 SMC1 1E-149 NM 006306.2 NP 006297.2 5 132 hCT9553-7 7 1186T>CT: 396F>UF C094 hCT9553-8 8 1300C>T: 434R>W HX8 hCT9553-10 10 1680C>CG: 560I>I/M CX3 hCT9553-16 16 2562 2563het insA HX171 hCT9553-24 24 3556G>AG: 1186V>IA/ HX129 2 CSPG6 SMG3 1E-45 NM 005445.3 NP 005436.1 5 130 CSPG6 23 . 23 2635C>CT:879R>R/X MX13 f'/.-,';C.SPGfr7/, i 7 415G>AG:139V>IA/ ; HX155 . • *. . CSPG6 8 8 512G>AGi171R>Q/R HX171 r4AeSPG6l:12., ' 12 100C>CT:334L>L/F HX152 1 CSPG6_21 r 21 ' 2321G>A:774R>K HX133 3 NIPBL SCC2 3E-19 NM 015384.3 NP 597677.2 5 132 hCT2293447 9 3 8 1435C>CT:479R>R/X HX7 hCT2293447 9 4 9 2967 2968het insT C071 hCT2293447 10 1 9 1660C>CT:554Q>Q/X HX168 YC03C04F 28 5378T>TA: 1793M>M/K MX24 hCT2293447 40 39 6893G>AG: 2298R>H/R HX171 4 STAG2 SCC3 2E-11 NM 006603 NP 006594.3 1 34 STAG2 24 24 2456C>CT:819S>S/F HX147 5 STAG3 SCC3 3E-13 NM 012447.2 NP 036579.2 1 34 STAG3 32 31 33963G>AG HX110 : 6 BLM SGS1 1E-115 NM 000057.1 NP 000048.1' 1 132 YC08C06B 15 3128C>AC: 1043A>D/A HX63 7 UTX SSN6/CYC8 9E-44 NM 021140 NP 066963.1 1 36 YC14C06D 17 2380A>AC:794Y>Y/S HX68 8 RNF20 BRE1" 5E-26 NM 019592 NP 062538.5 1 36- YC16C06G 3 370C>CT:124R>R/X HX88 Figure 3.1 Mutations in human SMC1L1 in colorectal cancers and analogous mutations in yeast SMC1 A. Protein sequence alignment of human SMC J LI with yeast SMC] reveals 27.8% identity and 42.1% similarity. Known domains are highlighted: the P-loop containing ATPase domain at N - and C-terminus are in blue; the SMC hinge domain is in orange. Somatic mutations found in human cancers and the analogous yeast amino acids are shown in red boxes. A nucleotide insertion leading to frameshift and truncation is indicated by a red arrow and stop sign. B. The table indicates the corresponding mutations constructed. C. Frequency of chromosome fragment loss per division is indicated for WTmd missense mutant haploids. The lower panel shows the sectoring phenotypes of the VI1871 mutant and control WTcells. 99 Figure 3.1 A . 1 1 1 i ? i R * | s s rs; R 3 7f J ? r 3 a - r i p : 3 ; ; 5 •• y~*i-;: A : ; r v ~ E M ; S ; s V 7 T l § 3 i : | 3 |r f 3 R r A | N R A rv ( I ) M 5 R i V 5 § J s s r K S Y § o % - r ^ ; r ; » S K F i § : : ; ? % - ; s i . » . s \ ^ ; i ; 3 r v - 5 \ ^ s s i R S S i J r £ . : | s u | v | x c E N | = : v : 2 i ( l teSMClLl ( 1 ) ! K S M C I ( 1 ) : cow*.* (!) as i : E : N F R S Y X S is rTA333?:.-3S3Es>;iK3A3srvi3 R S I S I K I I I X •62; 82 90 J O Q 110 ,120 J30 ,140 150 ^ " l E ! taSMCUJ (U) »wnt» tE--S*ESR?r*M:vj- --«IE'. R I S S E V Y J L A •• E E : . ; -AM:. ? ~ K S M C I (82) E3iASSN'PCS«XVKiri(0KSSKLVElK!fl:SaX3^ $YKIC5KTV5 5 ( |qJxS:riESEW : - IE^f iVr4JIVE3:* i05F ( M W M ( « 2 ) • s » x : > s Y K : it '•' D Y $ L E i i i R U ( * f i > v r « 0 v e U F Section J d«) J S us m m m m m ao ?« R S S M C I I I (149) f .Lf*A | r irlr?, : - E I A 3 £ T | R S » : ! : E 1 Y K A E E : T j r : : r K P J j | s : Al£RK£AECEREEAlRYS|tK3£vv-RA3V3i3i§KlY s c S M C l (1*3) VEliRlrEEis-:-! :CYKKEV|ELEEIt|EltL5KSME5:R}flU:!!l£LEtvr.EC'INRri|EVi«:L:KKS£l:Kr3%L|cLV Consensus (163) E 3 1EEEI3 3 EY3 K I R R R I U It R 3 Y R 5 0 i r i ! " Section 4 (2441 244 250 260 270 280 290 300 310 324 teSMCui (230) i n r E ^ r i F i N E T I J B K l ^ ^ K 5 M C ] (244) Kl£33KSElTrKlfllSS£3S31£S0j3;XSttXS13K3KSSrVK£5AVISKSRSIC§E Y 3 ERlEEXlVSISlRi: KV?3 3 A A G E Concensus (244) BE E l 1AA K EX E K I I R E I D R R 1 11 "" 1111 1 11 111 " 11 11 — • Section 3 (32S) 325 330 3*0 _ 3 M _ 370 _ 3 W 390 405 h s S M C l U (311) • E A A F . E S | 3 N ' A : ? . K Y R I R K 3 3 P J I E 1 S K E I I £ ^ K S M C I (32S) • S H : E X S | E 5 1 3 R 3 3>^3R?YBR> E - 3 | R " " ^ f l k ? ~ E £ E E | s c s | R S ^ Consensus (325) E I E I OR K E XS E 1 V KAR E S S 3 A R E ! U : S e t t O t> (406) 406 420 430 440 450 460 470 486 h s S M C i n (3»2) :§EiirF^ci;.FH|RjK^ X S M C I (406) Kli^|i};!;;Rl:::lELiRrKE«i3 : | R K | : ? S E 3 ; : • JEEi3qjps3;,r>-3|ycK;iAiKTS8iH£iKKi3sHESA»sa£Y C o n u n u (406) 3 S 3 E 03 I D RR S RI 1 S I it X E 1 SIE % SKts4- 7 (487) f » MP SJfi SO. SO m 5S0 _ 567 fcsSMCli: (472) ••KKEiS3^*E3#;:«s : : scE»sp];3§RaE:iE«lr.aL | ? i svv 35«::i :«?T3RRtO»*4p§i-iR:"Ji3pJ:'>'§sEKE K S M C I (487) HsEE13£r§YK§3:i5A^RE§xlE?JlREX'|^^ Consensus (487) 3 IN 1 1 3 3 B S I % E I 3ER ttS V 3 1 313 ? RY IAVS ILSRM 3A3 3 V 3 - S e c t i o n s n s S M C l l ] (553) K S M C I (567) Consensus (568) (553) l i l ilS^i'lll. 1^ I ' l l ' ' ^ 1 11 I1 I ' I . ' ^ I 1 S^— | - | , | | , i ( ) §3l^:i.B'-R;sl3Tiirlj3IT|ETE3?3 35 3p|3 33V:iS:KlI3VEf-EYE?:ii;VV.: '. igS'.3 3|x 3 A| 3|*V* R 3 ( 5 6 8 ' ~ 3:: rTr. S R A sp3Fi3 3E L D 13 3 Y E T E U O I 33 J U I C I A R 3 E (649) m m m m gso TOO 710 (634; !" : «feS • - - S V I E ~ 31R&KX»XW3ERAV3RlRERRflkTEE|RE0MR|jMRE-XElRSV05CAM33fHRlR (647) T£ JF,lV3Bc,ii3»Rl<4k«kw h s S M C t L J (634) 3RKRI RE-AElF3V3SCAK3l|H;|R K S M C l C c n M K u s (649) R R v 330 1 RAOLISOO S R*3 S x t t o n 10 (730) 730 ,740 / 5 0 ,760 770 730 7SC 800 810 i W S M C t l l (714) H g l 3 p E « 7 K 7 R i f i ^ K 5 M C 1 (726) SlRl3#33SKS513£SSl£3r.YKs|l3£XE3 3r|jrBE|E|331£^^^^ C c n s v r a u i (730) S 1 0 R 1 R 3 PRI 31 RR I E 1 t IR Ef IS I RE EE -Sxtnrll ( 8 1 1 ) 8_M 820 8 3 0 8 4 0 8 5 0 6 6 0 8 7 0 | f e 8 9 1 h s S M C l t l (795) E R V x I i i f T E l K l c R T E ^ ^ K S M C l (806) H : 3 E l K B 3 j i r , £ 1 3 C l « K 3 3 1 T l £ : ; E 1 3 r £ 3 3 R 3 S 3 3 3 R | Y £ R A 3 R 3 l E X i 3 V E l R S l E E 0 £ Y A 3 E M R : 3 5 I E | r . L E l . » r : . C m m u (811) » R O l i r i l O R - £ 1 1 1 : 0 A 1 2 R S S « t » w i 12 (092) 8 9 2 3 0 0 _ _ 310 3 2 0 8 3 0 3 4 0 9 6 0 3 6 0 M J hsSMClLl (876) 3 . K L & R E $ E V K : E | S E | E E I R R K 1 3 - 3 & K K S K « ^ S t S M C l (887) K 1 3 £ 1 3 R R ? V t r J | 5 E l K S S £ 3 I l £ 3 K S ' i K l 3 V l R i t R 3 | : K E 3 l £ r ? 3 l l a v T i l R S - R l s N 3 K l p l 4 3 E C o n k m u s (892) K K E l 1 S 1 1 R E A 3 ZZ B i t 1 3 X 1 3 3 7 1 3 T C S Set t l o r 13 (973) 9 7 3 3 6 0 <gR JPJB WHO 1 0 2 0 1 0 3 0 1 0 4 0 1 0 6 3 teSMCIU (957) * l 5 E 3 p » ^ 3 = : * : S S I Y A R E A L S E 3 3 V 3 3 i : S 3 1 R : A J A E E E | R ; E | ; : ? i i : K l K 5 i ; 5 V i : » 3 A A - - : K t ^ ? | l f SCSMC1 (968) 3|EA : l J s>;5 3 S 3 K Y R . J l f K R Y K E S S - T 3 S A R K s | s : E | a s | E E I i : : £ l 3 ? S A § A § E | Y | S A E O > Consensus (973) S i : s ! : £ A E 3 C I 1 I K E S R A 1 E R 3 K Section 14 ( 1 0 M ) 10U JC6C ,!37C ,1080 1OS0 _ .1100 ,1110 ,1120 1134 teSMCIU (1033 T 3 E T . ? 3 I f E A A l R R A R R A I t E F E I I R R E R r f R f S A : " * S V f T ::I' E ; . f i ; WS; d | .«prrt E 8 I | I I | U « S C S M C K 1 0 3 2 ) E e v I N ! t E T t 0 3 A > E E K r : i l ^ r L R : R R K R R f L r E K ? 4 Y v f : H | r f c l 4 £ l ^ Consensus (1054) E E E R K K K T Z R R U S E E 3 V A : 3 I Y R 1 S R S A A K A 1 E DE»T O I • — — — — S e c t i o n I S (1135) 1 t 3 5 ,1140 ,1150 ,1160 ,HH> 1180 .1180 1200 1215 h s S M c n i (im) i r s H A F i r i i S f M i x i s J T l ^ ^ K S M C l (1113) R Y K A T J f L K ? r | 3 x | Y l J J S E R T V A A l A l i r A : . V 5 Y 3 f | ? * r v i 3 £ l : A A L 3 I T T H I I H A A Y l | » K R : ; F 3 1 3 E q j J 3 3 1 R N 7 Con»»raui (113S) Y t R R T R « 3 1 3 3 3 E R T V A A 1 A 1 1 E A 3 S Y » A ? E r . , 1 3 E I I A A 1 3 T S X R 3 A Y I R 0 I V I S L E Section 16 (1216) 12W .1230 1240 1256 h s S M C K l (1193) I I I MM I I I lM ' ' ! » t l i I I M K S M C l (1194) X | E R B i l | G V Y R 3 - 3 - 0 £ S S 3 R 9 r i 3 3 l ^ Y A | Consensus (1216) T R A 3 A 1 3 5 V Y . S K I I T 3 1 3 Y 3 100 B Human SMC 1 LI amino acid substitution Type of SMC1L1 mutation Yeast SMC1 amino acid substitution F396L Homo/Hemi L380 (no mutation made) R434W Hetero Q449W I560M Hetero L574M Insertion of A between coding sequence 2562 & 2563, leading to amino acid change starting from 855 and termination at amino acid 864 Hetero I877Z VI1861 Hetero V11871 c Frequency of half-sectored & red colonies 0 . 0 0 8 0 . 0 0 7 0 . 0 0 6 « 0 . 0 0 5 a 0 . 0 0 4 I" 0 . 0 0 3 0 . 0 0 2 0 . 0 0 1 0 4> Strain (haploid) V1187I 7 m -• • \ • • * 1*0 » * • aft* * *• ft JfV 101 Figure 3.2 Synthetic lethal interactions o f yeast C I N genes whose human homologues were found to be mutated in colorectal cancers (which are placed on the r im and depicted in blue fonts). A. Only genes synthetic lethal with more than 1 of these CIN genes are shown. B. Only genes synthetic lethal with at least 5 CIN genes are shown. 102 CHAPTER 4 Characterization of MMS22, MMS1, RTT101 and RTT107 in the Maintenance of Genome Integrity 103 4.1 I n t r o d u c t i o n One of the goals of performing genome-wide chromosome instability screens is to identify novel genes and characterize their functions. Among the 293 genes identified in the C I N screens described in Chapter 2, 46 (16%) were uncharacterized (see Appendix 1 and 2; by G O Sl im Mapper on S G D , http://db.yeastgenome.org/cgi-bin/GO/goTermMapper). To prioritize genes for further study, I first examined the 34 genes that were identified in all 3 CLN screens. Many of the genes within this subset are known to be important for maintaining genome integrity, such as those that function at the kinetochore, in cell cycle checkpoints and in D N A repair pathways. A t the time the screens were completed, only a few of the 34 genes were largely uncharacterized, including NCE4, MMS1 and MMS22. However, I was able to gain insights into the functions of these genes by integrating data derived from large scale phenotypic screening (such as the CLN, G C R and drug sensitivity screens (see below)), as well as genetic and physical interaction analyses (Figure 2.5). Mass spectrometry (MS) analysis of immunoprecipitates of overexpressed tagged Mms22p identified Rttl01p/Cul8p, a cullin, and Rttl07p/Esc4p as interacting proteins (Ho et al., 2002), all o f which were also identified in the C L N screens. In addition, R t t lO lp and Mmslp /Rt t l08p /Kim3p were identified as protein-protein interactors with R o c l p / H r t l p , a R I N G finger protein known to bind to cullins in E3 ubiquitin ligases (Ho et al., 2002). These physical interactions suggested that these proteins may function together to maintain genome integrity. Indeed, RTT (retrotransposition) genes were originally identified in a screen for mutants that increase the transposition rate of T y l , a long terminal repeat (LTR) retrotransposon (Scholes et al., 2001). Many of the rtt mutants, including rttlOlA, rttl07A and mmslAlrttl08A, were found to have elevated rates of gross chromosomal rearrangement (GCR) (Kanellis et al., 2003; Luke et al., 2006; Rouse, 2004). On the other hand, MMS22 was identified in another screen set up to look for mutants leading to reduced levels of T y l retrotransposition, and it was shown that the level of T y l c D N A was not affected in mms22A mutants, suggesting that Mms22p affects steps that occur 104 after D N A replication, possibly in the repair of chromosomal D N A damage at integration sites (Griffith et al., 2003). Many C I N mutants are sensitive to D N A damaging agents. Different D N A damaging agents cause different types of D N A lesions and sometimes one agent generates pleiotropic lesions (Table 4.1 and Figure 4.1). Genes whose proteins function in the same or parallel pathway are expected to display similar genetic interaction profiles and phenotypes such as C L N and drug sensitivity (Parsons et al., 2004; Tong et al., 2004). For example, base excision repair (BER) mutants are sensitive to base damaging agents, but not to U V or LR. Nucleotide excision repair (NER) mutants are sensitive to U V and 4 N Q O , but only moderately sensitive to M M S and LR. Post-replication repair mutants are sensitive to M M S , U V and LR. Homologous recombination (HR) mutants are extremely sensitive to IR, but are only moderately sensitive to U V (Figure 4.1). Indeed, MMS genes, including MMS22 and MMS1, were originally identified in a screen for mutants that are sensitive to the M M S (Prakash and Prakash, 1977). In addition, both mms22A and mmslA mutants are sensitive to many other D N A damaging agents including H U , C P T , and moderately sensitive to U V (Baldwin et al., 2005; Bennett et al., 2001; Chang et al., 2002; Hanway et al., 2002; Parsons et al., 2004; Prakash and Prakash, 1977);(Hryciw et al., 2002). Similarly, rttlOlA and rttlOJA mutants are also sensitive to M M S , H U , and C P T (Bennett et al., 2001; Chang et al., 2002; Hanway et al., 2002). The drug sensitivity profiles of these 4 mutants were most similar to that of post-replication repair gene mutants, but not identical (see below), suggesting they may constitute a novel pathway involved in D N A repair. During the last few years, the role of Rttl07p, R t t lO lp , M m s l p and Mms22p in D N A damage response has been characterized in more detail as summarized below. 4.1.1 RTT107 Rttl07p contains 6 B R C T domains, which are usually found in proteins involved in signaling, repairing o f D N A damage or cell cycle regulation (e.g. B R C A 1 , X R C C 1 , and 53BP1). rtt!07A causes delays in S phase, accumulation in G 2 / M , and an increased 105 fraction of cells with checkpoint protein Ddc2p foci, which suggest a higher level of spontaneous D N A damage (Roberts et al., 2006). However, double mutant analysis for M M S sensitivity showed that Rtt l07p is not involved in N E R , H R or cell cycle control (Hanway et a l , 2002). Indeed, Rtt l07p is not required at the time of damage, and rttlOJA mutants are competent for activation of the intra-S-phase checkpoint, which is indicated by Rad53p phosphorylation. In response to D N A damage occurring in S phase, Rtt l07p is phosphorylated by the checkpoint protein M e c l p , and Rtt l07p is important for recovery from D N A damage by promoting restart of stalled replication forks (Rouse, 2004). Slx4p and S l x l p , which form a structure-specific D N A endonuclease required for resolving replication intermediates specifically in the r D N A (Fricke and B r i l l , 2003), physically interact with Rtt l07p. The Slx4p-Rttl07p interaction is independent of D N A damage, and requires the B R C T domains of Rtt l07p. The interaction is required for Meclp-mediated phosphorylation of Rtt l07p (Roberts et al., 2006). In another study, yeast-two-hybrid (Y2H) analysis of the N-terminal region of Rtt l07p (containing 4 B R C T domains) identified Rad55p, Mms22p, Tof lp and Sgslp (Chin et a l , 2006). Like the interaction with Slx4p, the physical interaction between Rtt l07p and Rad55p does not depend on D N A damage (Chin et al., 2006). Like Rtt l07p, Rad55p is also phosphorylated by M e c l p in response to D N A damage, and it forms a heterodimer with Rad57p, which together orchestrates the assembly of the Rad51p filament on replication protein A (RPA)-coated s sDNA. Taken together, these physical interactions suggest that Rtt l07p may associate with s s D N A of stalled replication forks to modulate repair and reinitiation of D N A synthesis. Subcellular localization of Rtt l07p is in agreement with its putative function at stalled replication forks. Rtt l07p displays diffuse nuclear localization in G I and G 2 / M , which is unaffected by M M S treatment (Chin et al., 2006). However, half o f S phase cells have Rtt l07p foci at the edge of the nucleus, and treatment with M M S increases the fraction of cells containing such foci (Chin et al., 2006). Pax2 transactivation domain interacting protein (PTIP), the mammalian protein with highest sequence similarity to 106 Rttl07p, was recently shown to form foci after D N A damage (Manke et al., 2003). The number of cells with Rtt l07p foci also increase in mrclA or toflA mutants, which contain s s D N A accumulated at stalled replication forks (Chin et al., 2006). Interestingly, Rtt l07p foci partially overlap with r D N A repeats. These results support that Rtt l07p may bind stalled replication forks that accumulate s sDNA, and may be involved in the repair of replication forks that collapse within the r D N A repeats. 4.1.2 RTT101 Rt t lOlp is one of the 3 cullins in S. cerevisiae, with demonstrable ubiquitin ligase activity in vitro, but as yet no known substrate in vivo (Michel et al., 2003). rttlOlA mutants accumulate with a short spindle and nucleus positioned at the bud neck (Luke et al., 2006; Miche l et al., 2003). The anaphase onset in rttlOlA mutants is delayed, and this is dependent on the intra-S-phase checkpoint (Mec lp and Rad9p) (Luke et al., 2006; Miche l et al., 2003). rttlOlA mutants display several phenotypes resembling rttlOlA mutants. For example, rttlOlA mutants have increased numbers of D d c l p and Rad52p foci, indicating an increase in spontaneous D N A damage (Luke et al., 2006). rttlOlA mutants are also competent in Rad53p checkpoint activation in response to H U treatment (Luke et al., 2006). Based on double mutant analysis for M M S sensitivity, R t t l O l p is not involved in non-homologous end joining (NHEI) or H R (Luke et al., 2006; Miche l et al., 2003). Like Rtt l07p, R t t lO lp may play a role in replication fork reinitiation or progression. In rttlOlA mutants, replication forks arrested at natural pause sites (e.g. r D N A barriers, centromeres) are more unstable, as indicated by the increased formation of extrachromosomal r D N A circles (Luke et a l , 2006). rttlOlA mutants cannot complete D N A replication during recovery from M M S , which induces fork arrest. O n the other hand, rttlOlA mutants can recover from H U , which causes fork pausing. These results suggest that R t t lO lp promotes fork progression through alkylated D N A (Luke et al., 2006). 107 4.1.3 MMS22 a n d MMS1 Phenotypically, both mms22A and mmslA mutants exhibit slow growth and abnormal cell morphology including large, round and elongated cells (Araki et al., 2003; Hryciw et a l , 2002). To determine whether MMS22 and MMS1 function in known repair pathway, double mutant analysis for M M S sensitivity was performed and suggested that both MMS22 and MMS1 are not involved in N H E J , and N E R (Araki et al., 2003; Hryciw et al., 2002). However, mms22A is synthetically lethal with rad6A (post-replication repair) and rad52A (homologous recombination repair) (Araki et al., 2003; Hryciw et al., 2002) , and mmslA is also synthetically lethal with rad52A in some strain backgrounds (Araki et al., 2003; Hryciw et al., 2002), suggesting that MMS22 and MMS1 may function in a pathway redundant to RAD52 and RAD6. A n mms22A mmslA double mutant exhibits M M S , H U , U V sensitivity that is similar to the mms22A mutant, indicating that mms22A is epistatic to mmslA (Araki et al., 2003). mms22A is also epistatic to rttlOlA and rttl07A (Baldwin et al., 2005). Indeed, sensitivity of mmslA to D N A damaging agents is suppressed by overexpression of MMS22, but not vice versa, suggesting that MMS1 acts upstream of MMS22 in a novel repair pathway. Besides genetically interacting with D N A repair genes, MMS22 and MMS1 also genetically and physically interacts with some essential genes involved in replication initiation. First, high-throughput Y 2 H studies showed that Mms22p (as prey) interacts with Psf lp and Psf2p, 2 of the 4 essential subunits of the GINS complex (Hazbun et al., 2003) . The GINS complex binds to D N A replication origins and facilitates assembly of the D N A replication machinery (Hazbun et al., 2003; Takayama et al., 2003). Second, mutations in MMS22 and MMS1 were both identified in a screen for mutations synthetically lethal with mcml0-l (Araki et al., 2003). Mcml0p/Dna43p is essential for replication initiation and the disassembly of pre-replication complex (pre-RC) after initiation (Araki et al., 2003). Therefore, M c m l O p is required for the smooth passage of replication forks through obstacles such as those created by pre-RCs assembled at active or inactive replication origins. mcml0-l causes replication fork pausing at active and silent origins (Araki et al., 2003). Interestingly, mcml0-l, mms2A and mmslA are all 108 synthetically lethal with dna2-2 (Budd et al., 2005). DNA2 encodes an essential D N A replication protein that contains helicase and single-stranded nuclease activities, and is involved in the processing of Okazaki fragments and in D N A repair. Collectively, MMS22 and MMS1 may constitute a novel D N A repair pathway that is specific for replication-dependent D N A damage (Araki et al., 2003), in agreement with Rtt l07p and Rt t lOlp having roles in the restart of replication upon DSBs . To further elucidate the role of MMS2'2 in D S B repair, a series o f phenotypic, genetic and physical analyses were performed. The survival rate after exposure to a D S B and the kinetics of D S B repair were determined in mms22A mutants and compared to wild-type and known D S B repair mutants. Global identification of Mms22p, R t t lO lp and Rttl07p physical interactors was performed by mass spectrometry (MS) and yeast-two-hybrid (Y2H) analyses. Since physical interactors of R t t lO lp are potential substrates of this putative ubiquitin ligase, the protein expression levels of some interactors were tested. 109 4.2 Materials and Methods 4.2.1 Yeast strains and media Yeast strains used in this study are listed in Table 4.1. Media for growth and sporulation were described previously in Rose et al., 1990. Epitope tagging and gene deletions were made directly at the endogenous loci (Longtine et al., 1998). Yeast transformations were performed as in (Gietz et al., 1995). 4.2.2 Quantification of chromosome transmission fidelity (ctf) Quantification of the ctf phenotype was performed in homozygous diploid strains containing a chromosome fragment (CF) as in (Shero et al., 1991). Briefly, diploid cells with one C F form pink colonies. Diplo id cells that lose the C F form red colonies, whereas those that contain 2 CFs generate white colonies. Chromosome missegregation in the first cell division after plating generates a half-sectored colony, and the frequencies of half-sectored colonies reflect the rates o f chromosome loss and non-disjunction (Shero et a l , 1991). 4.2.3 Genome-wide yeast-two-hybrid screens MMS22, RTT101 and RTT107 were cloned into p O B D 2 as described in (Cagney et al., 2000). The Mms22p-Gal4p-DNA binding domain fusion protein was functional as determined by rescuing sensitivity of mms22A to 0 .2M H U , 10p:g/ml camptothecin and 0.01% M M S (data not shown). Genome-wide two-hybrid screens were performed as described in (Uetz et al., 2000). Briefly, each screen was performed in duplicate, and positives that were identified twice were put into a mini-array for retest. Some reproducible positives were observed in many different screens with baits of unrelated function. These were considered common false positives and were excluded from further analyses. 4.2.4 Co-immunoprecipitation Co-immunoprecipitations were performed as described in (Measday et al., 2002). In brief, yeast extracts were generated using glass beads lysis. The protein concentration 110 of extracts was measured by Bradford assay, and equal amounts (2-5 mg) of extracts were incubated with a n t i - M Y C - or anti-HA- conjugated beads (Covance) for -24 hrs at 4°C. Beads were washed in extract buffer for a minimum of 3 times, and immunoprecipitates were eluted with S D S - P A G E loading buffer. 4.2.5 Mass spectrometry Protein eluates from immunoprecipitation were diluted in 20 m M Tris p H 8.3, 5 m M E D T A so that the final SDS concentration was no greater than 0.05%. 20 ng/ul of sequencing-grade trypsin (Fisher) was added, and digestion was allowed to proceed at 37°C overnight. Samples were then purified using C18 ZipTips (Millipore) according to the manufacturer's instructions. One-dimensional reversed-phase liquid chromatography with on-line mass spectrometry on an ion trap mass spectrometer (Model L C Q , ThermoElectron, San Jose, C A ) was performed as described in (Lee et al., 2004), employing a 90 min. binary gradient from 5%-80% solvent B during which each mass spectrum (MS) scan was followed by three M S / M S scans. Experimental mass spectra were compared with theoretical spectra generated from sequences from the Saccharomyces cerevisiae genome by using the S E Q U E S T algorithm (Yates et al., 1995). Data were displayed and filtered by using the I N T E R A C T software (Han et al., 2001). 4.2.6 Survival assay in HO-induced double strand breaks pJH132 ( p G A L - H O - T R P l ) (Lisby et a l , 2001) was transformed into wild-type, mms22A and rad52A strains. pRS414-TRPl was used as an empty vector control. Equal amounts of cells were plated on S C - T R P (Galactose) and S C - T R P (Glucose). Survival rate was calculated by the number of colonies formed on S C - T R P (Galactose) over that formed on S C - T R P (Glucose) after 8 days. 4.2.7 Microscopy Strains used for microscopy were grown in F P M (Synthetic complete medium supplemented with adenine and 6.5 g/L sodium citrate) in order to reduce auto-111 fluorescence. D A P I (300 ng/ml) was added to live cells for visualization of D N A as described previously (Connelly and Hieter, 1996). Stacks of microscopy images were taken with a Zeiss Axioplan II operated with Metamorph software (Universal Imaging). The presence of Rad52p foci was examined in WT ( Y K Y 8 0 7 ) and mms22A (YTK1364) strains as described in (Lisby et al., 2003). 112 4.3 Results 4.3.1 mms22A, mmslA, rttlOlA and rtt!07A exhibit sensitivity to DNA damaging agents and chromosome instability I analyzed the genome-wide drug sensitivity screen results by 2-dimensional hierarchical clustering, and overlaid them with the C I N screen results (Bennett et al., 2001; Birrel l et al., 2001; Chang et al., 2002; Giaever et al., 1999; Hanway et al., 2002; Parsons et al., 2004) (Figure 4.2). Two clusters are enriched for C L N mutants; the cluster showing sensitivity to benomyl is enriched for genes functioning at the kinetochore and spindle, whereas the cluster exhibiting sensitivity to D N A damaging agents is enriched for genes involved in D N A repair and D N A damage checkpoints, including MMS22, MMS I, RTT101 and RTT107. Interestingly, these 4 genes cluster with the M R X (MRE11, RAD50, and XRS2) complex, which is involved in D S B repair, sister chromatid cohesion and telomere maintenance. Both mms22A and mmslA were identified in all 3 C L N screens, whereas rttlOlA and rttl07A were identified in at least 1 C I N screen (see Chapter 2). In addition, the G C R rates of the rtt mutants are elevated (Kanellis et al., 2003; Luke et al., 2006; Rouse, 2004). When the frequency of chromosome transmission fidelity (ctf) was quantified by half-sectoring assay (Shero et al., 1991), all these mutants exhibited C I N at various levels (Table 4.3). 4.3.2 mms22A exhibits defects in cell cycle progression Based on the C L N phenotype of mms2'2'A, Mms22p may function in a cell cycle step that is crucial for chromosome integrity, or it may be important for repairing certain D N A lesions. Indeed, mms22A cells accumulate at G 2 / M phase (Bennett et al., 2001) and display increased cell size (Jorgensen et al., 2002). The absence of Mms22p could result in a delay in certain steps of the cell cycle, or it could induce spontaneous D N A damage, which would activate the cell cycle checkpoint and lead to cell cycle arrest. Fluorescence activated cell sorting ( F A C S ) analysis of logarithmically growing mms22A cells, compared to wi ld type cells, revealed a larger 2 N peak, which is consistent with 113 abnormalities in cell cycle regulation and aneuploidy (Figure 4.3A). Budding index analysis also revealed that an increased proportion of mms22A cells exhibit large buds and D N A in the bud necks. To investigate whether the G 2 / M accumulation is caused by checkpoint activation, mms22A was combined with several checkpoint mutations. F A C S analysis of these double mutants showed that this G 2 / M delay is not dependent on the spindle checkpoint (mad2A), the D N A damage checkpoint (rad9A, mec3A), or the replication checkpoint (mrclA) (Figure 4.3A). These results suggest that mms22A cells might be slow to enter or exit mitosis, leading to the G 2 / M accumulation. Indeed, mms22A is characterized by slow growth (Araki et a l , 2003). J^Tand mms22A cells were synchronized by arresting them in G l with cc-factor and their cell cycle were followed after release from the block (Figure 4.3C). mms22A cells appear to enter S phase and finish D N A replication at similar times compared to WT; however, mms22A cells enter the next G l at a much later time than WT cells. 4.3.3 mms22A has reduced survival rate with the introduction of DSBs mms22A cells are sensitive to several D N A damaging agents that cause DSBs . To directly test whether mms22A is impaired in D S B repair, the survival rate o f mms22A cells was monitored following the introduction of an H O endonuclease-induced D S B (Lisby et al., 2001) (Figure 4.4A). mms22A mutants have a lower survival rate (30-50%) than WT (80-90%), but are not completely inviable as observed in rad52A mutants. Consistent with the reduced survival rate, when the size of the colonies in the presence of a D S B was examined, mms22A cells formed much smaller colonies compared to WT (Figure 4.4B). This result is consistent with a lower sensitivity o f mms22A mutants to M M S compared to rad52A mutants, suggesting a less important role for Mms22p in D S B repair, compared to Rad52p. It is possible that Mms22p is only responsible for a subset of DSBs , such as D S B s that occur during replication, or that the Mms22p pathway is not the major repair route chosen by the cells. 114 Another way to monitor D S B repair is to examine the formation of D N A repair centres in cells, in which Rad52p aggregates multiple D S B sites together and recruits other homologous recombination proteins for repair in S and G2 phases (Lisby et al., 2003). The number of Rad52p foci is not directly proportional to the number of D S B s , suggesting that each focus likely represents the repair centre of multiple D N A lesions (Lisby et al., 2001). Interestingly, mms22A is synthetically lethal with rad52A (Araki et al., 2003), suggesting that Mms22p and Rad52p may act in parallel pathways with overlapping functions. To assess the level of spontaneous D N A damage and monitor the dynamics of Rad52p-dependent D S B repair in mms22A mutants, the percentage of mms22A and WT cells with Rad52p foci in the absence and presence o f a single D S B induced by H O or I-Scel, or in 0.1% M M S was recorded (Figure 4.4D). To make sure that Rad52p foci were observed with high confidence, another marker was observed in parallel. Lisby observed that 94% of Rad52p foci colocalize with the D S B site (Lisby et al., 2003). A n example of Rad52p foci and D S B site colocalization is shown in Figure 4.4C. 25-50% of budded JfT cells, but only 5-10% of budded mms22A cells, exhibited Rad52p foci in the presence of D N A damage. In general, the frequencies obtained were lower than those reported by Lisby et al., who observed that 22% budded WT cells form spontaneous Rad52p foci, and 62% budded WT cells form foci after l h r exposure to 0.1% M M S (Lisby et al., 2003). This could be attributed to technical variation (e.g. subjective definition of a Rad52p focus). In WT, the proportion of cells with Rad52p foci increases in the presence of D N A damage. In mms22A, the proportion of cells containing Rad52p foci does not differ in the absence and presence of D N A damage. This result is different from many C I N mutants, in which the percentage of cells with spontaneous Rad52p foci increases, as inpol2-100, meclA (Lisby et al., 2001), top3A (71%), sgslA (41%) (Shor et a l , 2005), nupl33A (30%) and rad27A (32%) (Loeillet et al., 2005). On the other hand, in rfalA, Rad52p foci do not form efficiently, revealing the hierarchy in the repair process (Lisby et al., 2004). This result suggests that Mms22p may act early in the repair process, and may be required for the formation of Rad52p foci. 115 In order to determine whether the lower survival rate of mms22A in D S B is related to slower repair kinetics, the presence of a D S B and the completion of repair in mms22A were monitored by southern blot and P C R analyses as in (Aylon et al., 2003) (personal communication with Martin Kupiec; Figure 4.5). While the D S B is resolved by 4 hours in WT cells, the D S B persisted up to 7 hours in mms22A cells. The level of gene conversion was also lower in mms22A cells (60%) compared to WT cells (90%). Interestingly, the survival level as observed by colony formation was lower than the repair rate based on P C R . This discrepancy implies that some cells repair the D S B , but are still unable to survive. A similar phenomenon has been seen in mutants compromised for checkpoint functions (e.g. rad24A, meclA) (Aylon et al., 2003). Although mms22A mutants have been shown to be competent for Rad53p activation (Araki et al., 2003), it w i l l be of interest to elucidate whether mms22A mutants are defective in some aspect of checkpoint function. 4.3.4 Mms22p interacts with replication initiation and DNA repair proteins that may constitute a novel repair pathway Mass spectrometry (MS) analysis of overexpressed tagged Mms22p immunoprecipitates provided preliminary evidence that Mms22p physically interacts with Rttl01p/Cul8p and Rttl07p/Esc4p (Ho et a l , 2002). To confirm these interactions, I performed reciprocal co-immunoprecipitation (co-IP) o f endogenously expressed tagged proteins (Figure 4 . 6 A & B ) . 4.3.4.1 Mass spectrometry analysis To systematically identify additional protein interactors o f this potential novel complex, immunoprecipitation followed by M S was performed using endogenously expressed Mms22p, R t t lO lp and Rttl07p. M S analysis for Mms22p immunoprecipitates did not yield any putative interaction partners, including Mms22p itself, possibly because Mms22p is expressed at a low level (data not shown). M S of Rtt l07p immunoprecipitates identified only Rtt l07p itself but no other protein. Interestingly, M S analysis of R t t lO lp 116 immunoprecipitates identified M m s l p , a protein proposed to function upstream of Mms22p (Figure 4.7A). To verify this interaction, reciprocal co-IP was performed using lysates extracted from logarithmically growing cells that contain no tag, M m s l p - M Y C only, R t t l O l p - H A only, and both M m s l p - M Y C and R t t l O l p - H A (Figure 4.7B). In the anti-HA IP, M m s l p - M Y C is only detected when both M m s l p - M Y C and R t t l O l p - H A are expressed. Reciprocally, in the an t i -MYC EP, R t t l O l p - H A is only detected in strains with M m s l p - M Y C and R t t l O l p - H A . 4.3.4.2 Y e a s t - t w o - h y b r i d ana lys i s Given the common occurrence of false-positives and false-negatives in genome-wide assays and screens, combining results using various methods often yield more informative results. Therefore, yeast-two-hybrid (Y2H) screening was performed using Mms22p, R t t lO lp and Rttl07p as baits. The Y 2 H study using Mms22p as bait identified Rt t lOlp , M m s l p , M e m 1 Op, Ctf4p and several other proteins as interacting proteins (Figure 4 . 8 A & D ) . To confirm the physical interactions between Mms22p and M m s l p , reciprocal co-IP experiments were performed (Figure 4.9). Endogenously expressed M m s 2 2 p - M Y C co-immunoprecipitated with endogenously expressed M m s l - H A in the H A - I P only when both tagged proteins were expressed. Reciprocally, endogenously expressed M m s l - H A co-immunoprecipitated with endogenously expressed Mms22p-M Y C in the M Y C - L P (Figure 4.9B). However, there was some background immunoprecipitation of M m s l p - H A in the H A - I P in the absence of M m s 2 2 p - M Y C . Therefore, reciprocal tagging was tried to avoid the background. H A - M m s 2 2 p expressed from the G A L 1 promoter was used instead. Endogenously expressed M m s l p - M Y C co-immunoprecipitated with overexpressed HA-Mms22p in the H A - L P (Figure 4.9A): However, in the reciprocal M Y C - I P , HA-Mms22p did not co-immunoprecipitate with M m s l p - M Y C . It is possible that the overexpression of Mms22p disrupts the localization of proteins required for its interaction with M m s l p , or it may change the stoichiometry of its physiological protein-protein interactions. Taken together, these results strongly suggest that these two proteins not only interact genetically as reported by Arak i et al. 117 (2003), but also physically, and likely together with R t t lO lp . It w i l l be of interest to investigate whether the pairwise interactions among these 3 proteins are dependent on the third protein. The Y 2 H interaction of McmlOp , a replication initiation protein, with Mms22p is also intriguing, since mcmlO-1 and mms22A are synthetically lethal (Araki et al., 2003). Ctf4p, a Y 2 H interactor of Mms22p, also functions in replication and cohesion, and is important for chromosome transmission fidelity (Mayer et al., 2004; Petronczki et al., 2004; Warren et al., 2004b). In addition, high-throughput Y 2 H studies showed that Mms22p (as prey) interacts with Psf lp and Psf2p, 2 of the 4 subunits of the GINS complex, which is required for D N A replication initiation and progression of D N A replication forks (Gambus et al., 2006; Hazbun et al., 2003; Takayama et al., 2003). I have been unable to confirm the physical interactions of Mms22p with these replication proteins by co-immunoprecipitation of endogenously tagged proteins in logarithmic growth condition (data not shown). It is possible that these interactions represent false-positives identified in Y 2 H screens and do not occur in physiological conditions, but it is also possible that the interactions are transient, occurring only at specific cell cycle stages, or only in a very small fraction of the total protein pool. However, because the Y 2 H interactors of Mms22p are enriched for replication proteins (p-value = 1.84E-5, by G O Term Finder on S G D , http://db.yeastgenome.org/cgi-bin/GO/goTermFinder), and these replication proteins are not seen as common interactors with many other proteins (false-positives), these interactions may be real and functional. The Y 2 H results are also in agreement with the observation that mms22A mutants are sensitive to D N A damaging agents that cause replication-dependent D S B s , such as CPT . 4.3.4.3 G e n e t i c i n t e r a c t i o n ana lys i s Genetically, mms22A interacts with mutations in replication initiation, H R and post-replication repair genes (Pan et al., 2006; Tong et al., 2004), suggesting it may have overlapping functions in these pathways. Unexpectedly, genome-wide S L screens using kinetochore mutants as queries revealed that mms22A also genetically interacts with 118 spc24-9 and spc24-10, temperature sensitive alleles of a gene encoding a central kinetochore protein, by lowering their permissive temperatures (Figure 4.1 OA). A t semi-permissive temperatures, both spc24-9 and spc24-10 mutants have elongated spindles and unequal distribution of chromosomal D N A . spc24-9 is also sensitive to H U (personal communication with Viv ien Measday). It is unlikely that Mms22p functions at the kinetochore, but the combined defects in mms22A and kinetochore mutants may sensitize cells to chromosome missegregation. Pan et al. reported that MMS22, MMS1, RTT101 and RTT107 forms a functional module or minipathway based on high congruence in genome-wide synthetic fitness/lethal (SFL) interaction profiles together with the H R and iL4Z)6~-dependent repair pathways (Pan et al., 2006). Mutations in any of the 4 genes cause similar sensitivity to DNA-damaging treatments, and do not exhibit S F L interaction with one another, except that Pan et al. observed a synthetic fitness defect in the rttlOlA rttl07A mutants (Pan et al., 2006). I did not observe a synthetic fitness defect in the rttlOlA rttl07A mutants in unperturbed condition, but did observe synergistic sensitivity to M M S and H U (Figure 4.1 OB). This study and others large scale physical and genetic interaction studies (Ho et al., 2002; Pan et al., 2006; Tong et al., 2004) have generated enomorous amount of interaction data for MMS22, MMS1, RTT101 and RTT107, which are invaluable to understanding the biological pathways of these genes. These interactions are summarized in a network diagram (Figure 4.11). 4.3.5 R t t l O l p regula tes M m s 2 2 p Miche l et al. (Michel et al., 2003) showed that R t t lO lp has sequence homology with cullins and contains in vitro ubiquitin ligase activity, but there is as yet no known in vivo substrate. Based on the physical interactions between Mms22p and M m s l p with R t t lO lp (Figure 4.6 and 4.7), I hypothesized that Mms22p and M m s l p could be substrates of the R t t lO lp E3 ubiquitin ligase complex. Therefore, I analyzed the steady state protein level of M m s l p and Mms22p in rttlOlA mutants. Expression of M m s l p is 119 not affected by rttlOlA (data not shown). On the contrary, Mms22p is expressed at a much higher level in rttlOlA mutants when compared to WT (Figure 4.12A). This is consistent with the hypothesis that Mms22p is a substrate of R t t lO lp , whereas M m s l p could regulate R t t lO lp activity. Indeed, the Mms22p expression level is similar in rttlOlA and rttlOlA mmslA (Figure 4.12A). It w i l l be of interest to also examine whether Mms22p expression level is affected in mmslA. To further investigate whether R t t lO lp regulates the expression level of Mms22p through its ubiquitin ligase activity, I attempted to analyze Mms22p expression level in an rttlOl mutant that affects its ubiquitin ligase activity. R t t lO lp , like other cullins, is modified by R u b l p at a conserved lysine K791. However, K791 is also the site of R u b l -independent modifications (Michel et al., 2003). The K 7 9 1 A mutation of R t t lO lp was observed to reduce its in vitro ubiquitin ligase activity by 50% (Michel et al., 2003); however, another study reported that the K791R mutation can still complement for R t t lO lp function in a transposition assay, showing that the modification at K791 does not completely disrupt R t t lO lp function (Laplaza et al., 2004). I compared the expression level of Mms22p in an rttlOlA mutant containing a 2JJ, plasmid expressing either wi ld-type RTT101 or rttl01-K791R under control of the Gal promoter. In both cases, the expression level of Mms22p was intermediate, between that observed in WT and rttlOlA, suggesting that both constructs partially complement the lack of R t t l O l p (Figure 4.12A). Due to the ambiguity regarding the function of the K791 modification, it is still difficult to conclude with certainty that Mms22p's expression level is affected by R t t l O l ' s ubiquitin ligase activity. To address this ambiguity, it w i l l be useful to asses the effect of a different mutant of R t t lO lp . The interaction between R t t l O l p and R o c l p is essential for its ubiquitin ligase activity, since deleting the conserved Roclp-interacting domain in R t t lO lp results in complete loss of in vitro ubiquitin ligase activity (Michel et al., 2003). Therefore, comparing the expression level of Mms22p in an rttlOl A mutant containing a plasmid expressing either wild-type RTT101 or an rttlOl mutant lacking the R o c l p -interacting domain (rttlOl-ARocT) would delineate whether the ubiquitin ligase activity of RTT101 is required for regulating Mms22p. 120 While the above experiments looked at the steady state expression level of Mms22p in logarithmic growth, it was of interest to examine the kinetics of Mms22p degradation in the presence or absence of Rt t lOlp . Mms22p was expressed from the galactose promoter in medium containing galactose, and the expression was then shut off by growth in glucose medium. The level of Mms22p was monitored at 20 min intervals for 100 min (Figure 4.12B). Interestingly, Mms22p levels increased to a higher level during the induction period in rttlOl A mutants. However, the degradation rate of Mms22p was not reduced in rttlOl A. It is possible that Mms22p is also degraded by an RttlOlp-independent pathway. In this experimental condition, however, the Mms22p could still be translated from residual m R N A transcripts after promoter shut off. Therefore, in future experiments, cycloheximide, a drug that inhibits protein translation, should be added when the culture is released into glucose. R t t lO lp could regulate the degradation of Mms22p in a cell-cycle dependent manner or in response to D N A damage. I therefore analyzed whether Mms22p and R t t lO lp are induced or modified under various conditions. Microarray analysis revealed that MMS22 m R N A expression is induced 5 minutes after 0.02% M M S addition and 20 minutes after heat shock (Gasch et al., 2001). Western blot analysis of Mms22p in different cell cycle stages or 0.01% M M S for 15 min at different cell cycle stages showed similar Mms22p expression level (data not shown). However, R t t lO lp showed a slower-migrating band in the presence of H U and nocodazole (Figure 4.13), suggesting it may be modified in a cell cycle-specific manner. Further experiments are required to distinguish whether these modifications are cell cycle specific, or whether they are side effects related to drug treatment. 121 4.4 Discussion B y integrating phenotypic, genetic and physical interaction data from the literature and from this study, I confirmed that mms22A is defective in cell cycle progression and D N A D S B repair. Co-immunoprecipitation experiments indicate that Mms22p, R t t lO lp and M m s l p physically interact with each other. These data support that Mms22p functions with M m s l p and Rt t lOlp in an E3 ubiquitin ligase. Indeed, the expression of Mms22p is regulated by Rt t lOlp , and it is possible that Mms22p is a substrate of the R t t l O l E3 ligase. Since M m s l p expression level is not affected by Rt t lOlp , M m s l p may serve as a specificity factor the E3 ubiquitin ligase (see below). This work leads to the proposal of a model in which R t t lO lp may regulate Mms22p and other protein levels in response to D N A damage. 4.4.1 Conservation of the RttlOlp complex? While there is as yet no confirmed substrate for the R t t l O l p ubiquitin ligase, clues regarding its function may be gained from knowledge about other cullins. Cu l l in serves a scaffolding function: it interacts through its N-terminal domain with a substrate specificity factor, and through its conserved globular C-terminal domain (called cullin homology domain) with the R I N G finger protein to form the catalytic core. In addition to Cul lp /Cdc53p/CulAp in SCF, budding yeast has 2 additional cullins: Cul3p/CulBp and Rtt l01p/Cul8p/CulCp, whereas humans have 4 additional cullins: C U L 2 , C U L 3 , C U L 4 A , C U L 4 B , and C U L 5 . R t t lO lp displays protein sequence similarity to all of the human cullins. However, it is unknown whether R t t lO lp is the functional ortholog to any of the known human cullins. Arak i et al. (Araki et al., 2003) claimed that M m s l p has weak similarity to Radl7p and D d c l p . Recently, M m s l p was found to have homology to human D D B 1 , the adaptor of C U L 4 A (Mathias Peter, personal communication). The damaged-DNA binding proteins, D D B 1 and D D B 2 , recognize damaged D N A and are important for global genome repair (GGR) , one pathway in N E R that repairs the D N A damage across the entire genome. X P patients (in the X P E complementation group) have mutations in 122 DDB2. D D B 2 , through its binding to D D B 1 , interacts with C U L 4 A and R O C 1 . In response to U V , C U L 4 A is post-translationally modified, which stimulates the ubiquitin ligase activity of the D D B complex to ubiquitylates X P C (Matsuda et al., 2005). Such ubiquitin modification has non-proteolytic function, but instead signals the cell to a specific D N A repair pathway. Genetic analysis suggests that the R t t l O l p complex functions downstream of P C N A and controls the function of the translesion D N A synthesis (TLS)-polymerase zeta, allowing the replication bypass of damaged templates during D N A replication. Mms22p exhibits weak homology to S. pombe T a z l , a telomere-binding protein that is required for efficient replication fork progression through the telomere (Mil ler et al., 2006). tazl A mutants have stalled replication forks at telomeres and telomere sequences placed internally on a chromosome. T a z l may recruit helicases to facilitate unwinding of the G-rich telomere repeats. T a z l is required to protect telomeres from NHEJ-mediated telomere fusions, and to prevent chromosomal entanglements and missegregation at cold temperatures (Mil ler et al., 2005). Human TRF1 and T R F 2 are putative orthologues of T a z l , and may also orchestrate fork passage through human telomeres. However, no orthologue of T a z l has been identified in S. cerevisiae. It would be of interest to investigate i f Mms22p has a role to facilitate replication fork progression through G-rich regions or other barriers. Interestingly, R t t l O l p and Rttl07p also have a role in facilitating replication fork restart through alkylated and r D N A regions, respectively (Chin et al., 2006; Luke et al., 2006). On the other hand, Arak i et al. (Araki et al., 2003) found that Mms22p has weak similarity to Rad50p, a component of the M R X complex. Mutants of the M R X complex and M M S 2 2 display similar drug sensitivity profiles, but further studies are required to determine whether Mms22p functions in a similar way as Rad50p. 123 4 . 4 . 2 Dia2p may play a redundant role with the RttlOlp complex in replication regulation Like rttlOl A, dial A mutants accumulate in S / G 2 / M , exhibit constitutive activation of Rad53p, increased foci of D N A repair proteins, elevated G C R and C L N as found in our screens (Chapter 2), and are unable to overcome MMS- induced replicative stress. Dia2p, a F-box protein in the S C F , is required for stable passage of replication forks through regions of damaged D N A and natural fragile regions, particularly the replication fork barrier (RFB) o f r D N A repeat loci (Blake et al., 2006). The synthetic lethal interaction profile of dia2A mutants clusters with mutants in D N A replication and repair (rad51A, rad52A, rad54A, rad57A, hpr5Alsrs2A), the replication checkpoint (csm3A, toflA, mrclA), the alternative R F C (dcclA, ctf8A, ctfl8A), the M R X complex, post-replicative repair (rad5A, rad!8A), and rttlOl A, rttlOl A and mmslA (Blake et al., 2006). S C F D , a 2 may modify or degrade protein substrates that would otherwise impede the replication fork in problematic regions of the genome. Interestingly, Dia2p binds to replication origins after origin firing, possibly to reset them for use in the next S-phase (Koepp et al., 2006). It is possible that Dia2p acts in a redundant fashion with R t t lO lp ubiquitin ligase to modify or eliminate substrates at the replication fork. 4 . 4 . 3 Identifying targets for RttlOlp ubiquitin ligase Although Rt t lOlp has in vitro ubiquitin ligase activity, and interacts with the R I N G finger protein R o c l p and the E2, Cdc34p, no in vivo substrate has been confirmed. The next important goal in characterizing the function of the R t t l O l p complex is to identify its target substrates. Ubiquitin modification of targets by the R t t l O l p complex may lead to cell-cycle or D N A damage specific proteolysis, or may determine the D N A repair pathway used by the cell (reviewed in (Huang and D'Andrea, 2006)). This study suggests that Mms22p is a component of the E3 complex, but it could also be a substrate. Autocatalytic degradation has been described for other ubiquitin ligases. For instance, the B R C A l - B A R D I complex can autopolyubiquitylate in response to D N A damage, and this autoubiquitylation stimulates its E3 ligase activity to ubiquitylate histone proteins (Huang 124 and D'Andrea, 2006). In addition, D D B 2 is also ubiquitylated by the D D B - C U L 4 A complex in response to U V (Matsuda et al., 2005). Physical interactors with Mms22p identified from genome-wide methods, in particular the replication proteins such as McmlOp , Ctf4p, Psf lp and Psf2p, are candidate substrates of the R t t lO lp ubiquitin ligase. The GINS complex, including Psf lp and Psf2p, is required for D N A replication initiation and progression of D N A replication forks (Gambus et al., 2006; Hazbun et al., 2003; Takayama et al., 2003). The GINS complex allows the M C M complex to interact with the replisome progression complexes (RPCs), which include Ctf4p, among other replication proteins. Interestingly, R P C s also interact with McmlOp (Gambus et al., 2006). Since mms22A mutants exhibit aneuploidy and some cells accumulate >2N D N A contents, it is possible that Mms22p is involved in the proteolysis of some replication proteins help to ensure that D N A is not re-replicated. It is known that budding yeast employ multiple regulatory mechanisms, including proteolysis of important factors, to serve this function. The replication licensing factor Cdc6p is known to be degraded through S C F C d c 4 in S phase. In addition, Orc2p and Orc6p, components of the origin recognition complex, are phosphorylated by S phase c y c l i n / C D K to inhibit pre-RC reassembly. Furthermore, another replication licensing factor C d t l and the M C M complex are exported from the nucleus (Guardavaccaro and Pagano, 2004; Pintard et al., 2004). Similarly, human M c m l O p is phosphorylated and degraded in a cell cycle-dependent manner (Izumi et al., 2001). In human cells, overexpression of C d t l leads to re-replication and polyploidy, and have been described in many cancers (Feng and Kipreos, 2003). It w i l l be of interest to investigate whether the R t t lOlp complex and their targets are involved in such function. Recently, both Mms22p and R t t lO lp were found to physically interact with H3 (Hhtlp) and H4 (Hhflp) (Krogan et al., 2006) (Figure 4.11). In addition, Mms22p interacts with H 2 B (Htb2p), while R t t lO lp interacts with H 2 A (Hta2p) (Krogan et al., 2006), suggesting the core histones may be potential substrates. Dephosphorylation of H 2 A is necessary for efficient removal of the cell cycle checkpoint (Keogh et al., 2006), but H 2 A may also be regulated by degradation upon completion of D N A repair. 125 In addition to the candidate approach, unbiased target screening methods that have been described for substrate specificity factors should also be applicable for cullins, though the number of substrates for cullins may be greater than that for specificity factors. A preliminary screen was performed using a system described by Deanna Koepp (personal communication), in which an ADE3-gene. fusion plasmid library was transformed in a strain lacking the gene of interest in the ubiquitin machinery (e.g. RTT101). The color of the yeast cells depends on the stability of the fusion protein. A t the same time, a plasmid containing RTT101 was transformed, but with no selection. This leads to loss of the RTT101 plasmid in some cells during colony formation. The generation of sectored colonies indicates that the stability of the fusion protein is affected by the presence or absence of Rt t lOlp . Similarly, in a microscopic screening system described by David Toczyski (personal communication), a GFP-fusion protein signal was compared between WT and strains lacking the gene of interest. This method has successfully identified and confirmed substrates for the F-box protein G r r l p (David Toczyski, unpublished). Since the cullins in yeast do not seem to be functionally redundant based on their differences in phenotypes, substrate screening for R t t lO lp w i l l shed insight to its biological functions. 126 Table 4.1 Types D N A lesions generated b y var ious D N A damaging agents DNA damaging agent DNA lesion(s) Methyl methanesulfonate (MMS) produces predominately 7-mehtylguanine and 3-methyladenine, which block DNA replication; as well as a small percentage of 06-methylguanine and 04-methylthymine, both of which cause base mispairing Hydroxyurea (HU) a ribonucleotide reductase inhibitor, inhibits DNA replication by depleting dNTPs Camptothecin (CPT) traps topoisomerase I (Topi) in the cleavage complex, causing single-stranded DNA (ssDNA) nicks that inhibit DNA replication and can be converted into double strand breaks (DSBs) by the advancing replication fork Ultra-violet (UV) radiation induces primarily cyclobutane pyrimidine dimers and photoproducts, which are efficiently targeted by the nucleotide excision repair (NER) pathway 4-nitroquinoline-l-oxide (4NQO) a UV mimetic agent, introduces bulky DNA adducts that are also mainly removed by NER Ionizing radiation (IR) induces DSBs that are replication-independent 127 T a b l e 4.2 1 l^ ist of yeast strains used in Chapter 4 Strain Genotype Reference YKY90 MATa/MATa ura3-52/ura3-52 trplA-63/trplA-63 his3A-200/his3A-200 leu2A-l/leu2A-I ade2-10l/ade2-Wl lys2-801/lys2-801 CFIII(CEN3.L)-URA3 SUP 11 mms22A::HIS3/ mms22A::HIS3 This study YKY570 MATa/MATa ura3-52/ura3-52 trplA-63/trplA-63 his3A-200/his3A-200 leu2A-l/leu2A-l ade2-l01/ade2-101 lys2-801/lys2-801 CFIII(CEN3.L)-URA3 SUPU rttWlAr.TRPU rttlOlAr.TRPl This study YKY332 MATa/MATa ura3-52/ura3-52 trplA-63/trplA-63 his3A-200/his3A-200 leu2A-l/leu2A-l ade2-101/ ade2-101 lys2-801/ lys2-801 CFIII(CEN3.L)-URA3 SUP11 rttl07A::TRPl/rttl07A::TRPl This study YPH499 MATa ura3-52 trplA-63 his3A-200 leu2A-J ade2-10J lys2-801 Hieter lab YKY62 MATaura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 CFII1(CEN3.L)-URA3 SUPU mms22A::HIS3 This study YKY64 MATa ura3-52 trplA-63 his3A-200 leu2A-J ade2-101 lys2-801 mms22A::HIS3 This study YKY104 MATaura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 CFIII(CEN3.L)-URA3 SUPU mms22A::HlS3 mad2A::HIS3 This study YKY108 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 CFIII(CEN3.L)-URA3 SUPU mms22A::HIS3 rad9A::LEU2 This study YKY210 MATa ura3 trplA-63 his3 leu2A-l mms22A::HIS3 mec3A::kanMX This study YKY248 MATa ura3 trplA-63 his3 leu2A-l mms22A::HIS3 mrclAr.kanMX This study YKY249 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-l0l Iys2-80J pRS414-TRPl This study YKY253 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 Iys2-80J rad52A::LEU2 pRS414-TRPl This study YKY256 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 mms22A::HIS3 pRS4l4-TRPl This study YKY260 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-10l lys2-801 pJH132(pGAL-HO)-TRPl This study YKY264 MATa ura3-52 trplA-63 his3A-200 leu2A-J ade2-l01 lys2-801 rad52A::LEU2pJH132(pGAL-HO)-TRPl This study YKY269 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 mms22A::HIS3pJH132(pGAL-HO)-TRPl This study YKY807 MATa ade2-l barl::LEU2 trpl-l LYS2 RAD5 RAD52-CFP ura3::3xURA3-tetOxll21-Sce(ura3-1) his3-U,15::YFP-LacI-his3-x leu2-3,l 12::LacO-LEU2HO-iYCL018W0eu2-3,l 12) TetR-RFP(iYGLl 19W) pJH1320(pGAL-lSceI)-ADE2 URA3 (Lisby et al., 2003) YTK1364 MATa ade2-l barl.:LEU2 trpl-l LYS2 RAD5 RAD52-CFP ura3::3xURA3-tetOxll2 I-Sce{ura3-1) his3-ll,15::YFP-LacI-his3-x leu2-3,l 12::LacO-LEU2HO-iYCL018W0eu2-3,U2) TetR-RFP(iYGLl 19W) mms22A::kanMX pJH1320(pGAL-lSceI)-ADE2 URA3 This study YKY754 /MK203 MATa-inc ade2 ade3::GALHO ura3::HOcs leu2-3,112 his3-ll,13 trpl-l lys2::ura3::HOcs-inc(RB) (Aylon et al., 2003) YKY755 MATa-inc ade2 ade3::GALHO ura3::H0cs leu2-3,112 his3-11,13 trpl-l lys2:: ura3: :HOcs-inc(RB) rad52A: :LEU2 (Aylon et al., 2003) YKY848 MATa-inc ade2 ade3::GALHO ura3::H0cs Ieu2-3,U2 his3-ll,13 trpl-l lys2::ura3::HOcs-inc(RB) mms22A::kanMX This study YKY713 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 MMS22-13MYC::HIS3 RTT101-3HA::TRP1 This study 128 Y K Y 4 3 5 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 RTT101-3HA::TRP1 This study YKY721 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 MMS22-13MYC:HIS3 RTT107-3HA::TRP1 This study Y K Y 4 6 1 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 RTT107-3HA::TRP1 This study Y K Y 4 1 3 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 RTT101-13MYC:: TRP1 This study Y K Y 4 4 7 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 RTT107-13MYC:: TRP1 This study Y K Y 6 9 0 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-80l MMS22-13MYC:: HIS3 This study YTK1168 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-Wl lys2-801 MMS1-3HA::kanMXRTT101-13MYC:: TRP] This study YTK1132 MATaura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 MMS1-3HA::kanMX This study Y K Y 5 2 7 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 HIS3-pGAL-3HA-MMS22 This study YTK1140 MATaura3-52 trplA-63 his3A-200 leu2A-l ade2-10l lys2-801 MMS1-13MYC::kanMX This study YTK1345 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 MMS1-13MYC::kanMX HIS3-pGAL-3HA-MMS22 This study YTK1375 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 MMS1-3HA::kanMXMMS22-13MYC::HIS3 This study Y K Y 8 2 0 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-80l spc24-8::kanMX V . Measday YKY821 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-Wl Iys2-80J spc24-9::kanMX V . Measday Y K Y 8 2 2 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-10l lys2-801 spc24-Wr.kanMX V . Measday Y K Y 8 2 4 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-10l lys2-801 spc24-8: .kanMX mms22A::HIS3 This study YKY831 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 spc24-9::kanMX mms22A::HIS3 This study Y K Y 8 3 6 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-l0l lys2-801 spc24-Wr.kanMXmms22A::HIS3 This study Y K Y 2 9 7 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 rttl01A::TRPl This study Y K Y 3 2 5 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 rttW7A::TRPl This study Y K Y 6 5 7 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 rttlOl A: :TRP1 mms22A::HIS3 This study Y K Y 6 4 2 MATa ura3-S2 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 rttl07A::TRPl mms22A::HIS3 This study Y K Y 6 4 8 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 rttl07A::TRPl rttlOl A:: kanMX This study Y K Y 7 6 7 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 MMS22-13MYC::HIS3 rttlOl A:: TRP 1 This study Y K Y 9 5 6 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 MMS22- This study 129 13MYC::HIS3 rttlOlAr.TRPl pYES-RTT101-URA3 YKY956 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 lys2-801 MMS22-13MYC::HIS3 rttlOlA::TRP1 pYES-rttl01-K791R-URA3 • This study YKY782 MATa ura3-52 trplA-63 his3A-200 leu2A-l ade2-101 Iys2-80J HIS3-pGAL-3HA-MMS22 rtt!01A::TRPl This study 130 T a b l e 4.3 Quant i f ica t ion o f chromosome loss (CL) , non-dis junct ion (NDJ) and chromosome ga in (CG) b y half-sectored assay Strain Total no. colonies counted No. pink-red colonies C L freq. (fold over WT) No. white-red colonies IS ».l lrc(|. (fold over WT) No. pink-white colonies CGfreq. WT diploid N/A N/A 8.7E-5 (Shero et al., 1991) N/A 8.7E-5 (Shero et al., 1991) N/A N/A '. mms22A mms22A 11890 36 3.0E-3 (35X) 4 3.4E-4 (4X) 3 2.5E-4 rttlOlA rttlOl A 14600 40 2.7E-3 (3IX) 32 2.2E-3 (25X) 93 6.4E-3 rttl07A rttl07A 17730 40 2.7E-3 (26X) i6 ; 1 9.0E-4 (10X) 52 2.9E-3 131 Figure 4.1 DNA damage and repair mechanisms (reprinted from (Hoeijmakers, 2001) Genome maintenance mechanisms for preventing cancer, Nature, 411, 366-374, Copyright 2001, with permission from Macmillan Publishers Ltd) Common DNA damaging agents (top); examples of DNA lesions induced by these agents (middle); and most relevant DNA repair mechanism responsible for the removal of the lesions (bottom). (6-4)PP and CPT, 6-4 photoproduct and cyclobutane pyrimidine dimer, respectively (both induced by UV light); EJ, end joining. lamaqinq agent t X-ray* Oxygen radicals AlKylaung agent* Spontaneous reactions UV kgm Potycydic aromatic hydiooaibons X rays AMi tumour aycnls los-Pt, MMC) Rcpbc anion erors Uracil Abasit site 8 Oxoguanic Single strand ore-alt J6-4JPP Bulky adduct CPD imefstrand crosslink Double-strand txeak t Baae-excison repair IB£ R) Nucleotide-excisian repair |N£ RJ A-G Mbmatch T-C Mismatch Insertion Deltion Recomrjinat»nal repair NR. EJ) t Mismatch repair 132 F i g u r e 4.2 T w o - d i m e n s i o n a l h ierarchica l c luster ing o f drugs (horizontal) and yeast de le t ion mutants (vertical) that show sensi t ivi ty to at least 1 drug based o n genome-wide drug sens i t iv i ty screens, and over la id w i t h the C I N screen results. T h e c lus ter ing was performed us ing the p rogram Clus te r 3.0 and d i sp layed i n Java Tree V i e w (vers ion 1.0.8) ( E i s e n e t a l . , 1998). A. An overview of all deletion mutants and drugs. The severity of drug sensitivity is indicated by color, with bright red indicating high sensitivity, darker red showing milder sensitivity, and black representing the same sensitivity as wild-type. A positive phenotype in the CTF, B iM and ALF screen is indicated by orange, yellow and green, respectively. The benomyl sensitive cluster is highlighted in lime color, and the DNA damaging sensitive cluster is highlighted in light orange. B. Magnification of the benomyl sensitive cluster. C. Magnification of the DNA damaging sensitive cluster. MMS22, MMS1, RTT101 and RTT107 are highlighted in purple, and the MRX complex components are highlighted in green. 133 ' ofo* c -t re G E N E O R F EAP1 YKL204W FYV6 YNL133C P0P2 YNR052C PAC10 YGR078C NPL6 YMR091C ASK10 YGR097W BUB3 YOR026W| BUD22 YMR014W BUD31 YCR063W CDH1 YGL003C | CIN1 YOR349W CIN2 YPL241C CIN4 YMR138W CTK1 YKL139W ESBP6 YNL125C GIM3 YNL153C GUP2 YPL189W HIR3 YJR140C H0P2 YGL033W H0S2 YGL194C INP52 YNL106C MAD1 YGL086W | MAD2 YJL030W MFA1 YDR461W YPL077C YPL077C N0P13 YNL175C YKE2 YLR200W YJR084W YJR084W YIL102C YIL102C PAC2 YER007W YDR219C YDR219C YBR032W YBR032W RPS22A YJL190C TUB3 YML124C SWR1 YDR334W SET6 YPL165C SHS1 YDL225W SRV2 YNL138W SNT309 YPR101W PFD1 YJL179W PH088 YBR106W GIM5 YML094W KRE28 YDR532C YOR073W YOR073W YML094C-. YML094C-A YDR442W YDR442W YCL005W YCL005W B i M YGR078C YGR078C ?YOR026W YOR026W [YGL003C YOR349W YPL241C (YGL086W YJL030W YIL102C YER007W YML124C YDR334W YJL179W YML094W YDR532C YDR532C YOR073W YOR073W 135 9£T 3 F i g u r e 4.3 Ce l l cycle and morphology defects of mms22A A. FACS analysis of WT haploid, mms22A, mms22A mad2A, mms22A rad9A, mms22A mec3A, and mms22A mrclA cells in logarithmic growth. IN and 2N DNA contents, corresponding to unreplicated and replicated DNA amount in haploid cells, respectively, are indicated. B. Budding index of logarithmic growing WT haploid and mms22A cells, n equals the number of cells counted. C. FACS analysis of WT and mms22A cells in logarithmic growth, synchronized by a-factor block (Gl arrest) and release. After 80 min release from alpha-factor (as indicated by the arrow), a majority of both WT and mms22A cells have undergone replication. After 120 min release from alpha-factor (as indicated by the dashed arrow), a portion of WT cells have gone through mitosis and return to G l , whereas most of mms22A cells still accumulate at G2M with 2N DNA content. A . Qm iWThapoid WT haploid 1000 1N 2N mms22A madlA mms22A mrclA mms22A o looo F U - H mms22A radQA I 0 sooo FU-H mms22A mec3A 0 1000 F13-H 75 $ mm IT A J il m B . G1/S M Others WT haploid 38% 23% 5% 3% 5% | 27% 1% 0% 220 mms22A 15% 18% 8% 18% 11 % | 22% 4% 5% 473 137 Figure 4.4 Defects o f mms22A in double-strand breaks A. Survival rate of WT, mms22A, and rad52A in HO-induced DSB. B. Colony size and number of WT, mms22A in HO-induced DSB, compared to empty vector control (pRS414) on galactose (after 6 days at 30°C). C. A schematics of chromosome IV (the black line with the centromere represented by a circle) in the tested strains was shown. Adjacent to the I-Scel cut site (I-Scelcs, the black triangle), an array of 336 tetO (336xtetO, the red rectangles) was inserted. The tested strains also contain a plasmid encoding a galactose inducible I-Scel endonuclease, a Rad52-CFP fusion protein, and tet-repressor fused to RFP (tetR-RFP) which binds to the tetO array, indicating the cut site. An example image of Rad52p-CFP foci (top left panel) and tetR-RFP (top right panel) colocalizing (bottom left panel, arrows) was shown. The corresponding differential interference contrast (DIC) image was shown as well. D. Percentage of cells with Rad52p foci in WT, mms22A in unperturbed condition, I-Scel-induced DSB, and 0.1% MMS. Sc-TRP Sc-TRP l-Scelcs 336xtetO » _ I I I J • ^ • ^ ^ ^ • u 138 F i g u r e 4.5 Kinetics o f D S B repair A. Schematic diagram of experimental set up reproduced from (Aylon et al., 2003). Open rectangles represent the ura3 alleles on chromosome II and V. A black box represents the HOcs; a gray box depicts the inactive HOcs-inc flanked by the BamHI (B) and £coRI (R) restriction sites. These polymorphisms are used to monitor the transfer of information between the chromosomes. The HO gene is under transcriptional control of the GAL1 promoter, and induction results in gene conversion. B. Southern blot analysis of DNA extracted at different times after transfer to galactose-containing medium. The DNA was digested with Clal and probed with a fragment of chromosome V carrying the URA3 gene. The % of DSB is quantified and normalized with the standard. C. Equal amounts of PCR product of the chromosome V region was digested with BamHI and subjected to gel electrophoresis. The extent of gene conversion (GC) is measured by the relative amounts of intact chromosome V containing the BamHI restriction site. 139 Figure 4.5 A . B O a - s C h r . V t K = a H — C h r . / / 0 = i • K U _ H O c s - i n c Transfer to galactose J ^) HO endonuckase H E Gene conversion B ' standard Intact chrV DSB Time in Gal (hr): 0 0.5 2 3 4 -mk2rj3mms22 -nt203 0 05 1 15 2 2 5 3 35 i 45 5 55 6 6 5 7 75 8 85 tine ihri MK203 mms22A Time in Gal (hr): 0 2.5 3 4 5 6 MK203 WT 100 90 3D i 60 o o 93 * )0 30 20 10 0 -mk203mm522 -mk203 7 ^ i i i i i i i i i 0 05 1 15 2 25 3 35 i (5 5 5.5 6 65 7 7.5 t i l l ! 140 Figure 4.6 Mms22p co-immunoprecipitates with R t t lO lp and Rtt l07p A. Anti-MYC immunoprecipitation using 9E10 affinity matrix (Covance) and anti-HA immunoprecipitation using HA. 11 affinity matrix (Covance) were performed in untagged strain and strains containing Mms22p-MYC only, RttlOlp-HA only, and both tagged proteins. The tags were fused to the C-terminus of the proteins and expressed from endogenous promoters. The strains were grown to log phase and lysed. Whole-cell lysates (total) and equal amounts of immunoprecipitates from the 4 strains were loaded on SDS-PAGE gels. Anti-MYC antibodies (9E10, Convance) and anti-HA antibodies (12CA5, Boehringer Mannheim) were used for Western blot analysis. B. Anti-MYC and anti-HA immunoprecipitations were performed as in A in untagged strain and strains containing Mms22p-MYC only, Rttl07p-HA only, and both tagged proteins. M M S 2 2 - 1 3 M Y C R T T 1 0 1 - 3 H A 2 5 0 k D a -1 5 0 k D a -T o t a l M Y C - I P HA-IP + + - M m s 2 2 ( a - M Y Q 1 0 0 k 0 a -Rt t101 (a-HA) B . M M S 2 2 - 1 3 M Y C R T T 1 0 7 - 3 H A 2 5 0 k D a -I S O k D a T o t a l M Y C - I P HA-IP + + - M m s 2 2 (ct-MYC) - Rtt107 (a-HA) 141 Figure 4.7 Physical interaction o f R t t lO lp with M m s l p A. MS analysis of RttlOlp irnmunoprecipitates. The experiment was performed in duplicates, and only proteins identified reproducibly in both analyses are shown. However, the peptide sequence, the number of peptides, and the peptide score of identified protein may vary in the two independent experiments. The peptides and score shown are obtained from one of the two experiments. Common false positives are not shown. B. Anti-MYC and anti-HA immunoprecipitations were performed as in figure 4.6, in untagged strain and strains containing Mmslp-MYC only, RttlOlp-HA only, and both tagged proteins. A. Prey ORF Prey Name Different Peptides Identified Peptide Prophet Score Description YPR164W MMS1 K.IELQALEEIQQK.H K.IELQALEEIQQK.H R.LGINQSNTESSLIFATDAVSNNR.I Y.NAVALDKPIQDISYDPAVQTLY.V K.SISPLPSNPINLDSR.S R.LSPYNAVALDKPIQDISYDPAVQTLY.V 1 0.96 1 0.4736 0.9972 0.2537 Protein likely involved in protection against replication-dependent DNA damage YJL047C RTT101 R.DIDNTYSINESFKPDMK.K K.DLALVLK.S K.YLNENLPILR.L R.LFDEVVQLANVDHLK.I 0.9917 0.922 0.9214 0.9998 Cullin subunit of a Roc lp-dependent E3 ubiquitin ligase complex YDR028C REG1 R.IVNNTPSPAEVGASDVAIEGYFSPR.N 0.9999 Regulatory subunit of type 1 protein phosphatase Glc7p B. Total HA-IP MYC-IP MMS1-13MYC RTT101-3HA 250kDa -150kDa -100kDa -75kDa -- Mms1 (a-MYQ -Rtt101 (rx-HA) 142 Figure 4.8 Yeast-two-hybrid interactions using bait protein: (A) Mms22p, (B) RttlOlp, and (C) Rttl07p. Out of the 2 genome-wide screens and the retest, only genes identified at least 2 times are shown. (D) Examples of miniarrays in retest. Each strain contains a different pOAD-fusion protein. The interactors are indicated in yellow. A strain with just the pOAD is used as a negative control. MIG1 is a common false positive. A. pOBD2-MMS22 No. times id. ORF Name GO Biological Process 3 YER180C ISC 10 sporulation 3 YER127W LCP5 rRNA modification 3 YCL032W STE50 signal transduction during conjugation with cellular fusion 3 YKL075C unknown 3 YIL150C MCM10 DNA replication initiation 3 YJL047C RTT101 ubiquitin-dependent protein catabolism 3 YDR026C unknown 3 YPR164W MMS1 DNA repair 3 YLR320W MMS22 double-strand break repair 2 YER029C SMB1 nuclear mRNA splicing, via spliceosome 2 YOL091W SP021 meiosis 2 YPR135W CTF4 DNA repair B. pOBD2-RTT101 No. times id. ORF Name GO Biological Process 2 YIL105C LIT2 actin cytoskeleton organization and biogenesis 1 (at retest) YLR320W MMS22 double-strand break repair C. pOBD2-RTT107 No. times id. ORF Name GO Biological Process 3 YLR320W MMS22 double-strand break repair 2 YLR135W SLX4 DNA replication 143 Figure 4.9 Physical interactions o f Mms22p with M m s l p A. Anti-MYC and anti-HA immunoprecipitations were performed as in figure 4.6, in untagged strain and strains containing endogenously expressed Mmslp-MYC only, HA-Mms22p expressed from the GAL1 promoter only, and both tagged proteins. Lysates were prepared from cultures in log phase grown in galactose-containing media. B. Anti-MYC and anti-HA immunoprecipitations were performed as in figure 4.6, in untagged strain and strains containing endogenously expressed Mms22p-MYC only, endogenously expressed Mmslp-HA only, and both tagged proteins. A . p G A L - 3 H A - M M S 2 2 + M M S 1 - 1 3 M Y C -Total MYC-IP HA-IP - Mms1 ( a - M Y Q - M m s 2 2 (a-HA) Total MYC-IP HA-IP M M S 1 - 3 H A - • M M S 2 2 - 1 3 M Y C + -+ . + + + - Mms1 (a-HA) - M m s 2 2 ( a - M Y Q 144 Figure 4.10 Genetic interactions of mms22A mutants A. Synthetic lethal interactions of mms22A with spc24-9 and spc24-10 at 33°C. Three temperature-sensitive alleles of SPC24 (non-permissive temperature is 37°C), mms22A, and 2 isolates of each double mutants were streaked on YPD plates and inoculated at 33°C. B. MMS and HU sensitivity of single, double and triple mutants of MMS22, RTT101 and RTT107 were analyzed by serial dilutions on YPD, 0.01% MMS and 0.05M HU plates. meclA is a positive control. 145 Figure 4.11 Interaction network of MMS22, MMS1, RTT101 and RTT107 A. Physical interactions obtained from the literature and this study are displayed using OSPREY. Only interactions with the 4 genes (shown in blue fonts) are shown. The color of the nodes indicates the GO biological process, and the color of the edges represents the type of interaction with the arrow pointing from the bait to the prey. B. Genetic interactions shown as in A. 146 Figure 4.11 YAR010C " * MCM10 VMMS1JDH2 * * G 0 H 2 3 YDR026C YKL07SC LSW1 CDC* PATWAD5»CDC73 MTOLGEglMOTJCVLROSIW y G f t 1 5 0 c ^ " E 1 .fl*8** Mdkj2vPS«5*» JAF7 CTKlJpOAJJCKI * OKI b e W c g j ^ p B ? : ^ J r f - J w o BEM2HDM31 SPC24 CHI Oraerwation and eugmsis ^ RNA pioceufcej Sponeatton 9 OMAiecotnbination ONA Oamew Response 0 DNA Replication Stress Response Proteei amino acid depiusphorylation # UNKNCRM4 # DNAmetaboewn ^ Iiansctdtion 0 Carbohydrate tdetarjatisni 0 Protein transport 0 Transport Signal transduction £ Protein biosynthesis if? RMA Localization Experimental Systems) I TWOHYBRJD H SYNTHETIC GROWTH DETECT | SYNTHETIC II THAT ITY | AFHMTO CAPTURE-eJS • DOSAGE GROWTH DEFECT SYNTHETIC RESCUE | BWCHEleTCAl ACTIVITY • AEFMtTY CAPTURE-WESTERN 147 Figure 4.12 Mms22p expression is regulated by R t t lO lp A. The level of endogenously expressed Mms22p-MYC was analyzed in WT, rttlOlA and rttlOlA mmslA in logarithmic growth. Equal amounts of lysates were loaded on SDS-PAGE gel, and anti-MYC antibodies (9E10) were used in Western blot to detect Mms22p-MYC. Two different exposure times are shown. Polyclonal anti-NDCIO antibodies (from Benjamin Cheng and Phil Hieter, unpublished data) were used to detect Ndc 1 Op, a loading control. B. Gal shut-off chase experiment of pGall-HA-Mms22p in WTand M l 01 A. Cultures were grown to log phase in media containing 2% raffinose. 2% galactose was then added to the cultures. After 3 hours in galactose, the cultures were washed and release into media with 2% glucose. Western blot was performed as in A using anti-HA (12CA5) and anti-NDCIO. NdclOp expression level was used as a loading control to normalize Mms22p level. The normalized Mms22p level (Mms22p/Ndcl0p) in the 2 strains is plotted against time and a logarithmic trendline representing the degradation rate is shown. 148 Figure 4.12 < 2 2 H o C O C O I ss e a cu a. < < < < o o o o e e e « i l « tt Mms22 j ( a - M Y C ) -Mms22 (higher exposure) NddO (loading control) B . 3.5 3 0.5 0 Raftlnose Galactose (3 hrs) Rtt101 Glucose (20 m time point) rtt101A Glu (m) Glu (m) £ O 20 40 60 80 100 £ £ 20 40 60 80 100 pGAL-3HA-MMS22 NddO (loading control) N o r m a l i z e d M m s 2 2 E x p r e s s i o n y = -1.001Ln(t) +1.6697 = -1.6275Ln(t)+3.2596 1 n. n ^ ^ 1^ ^ ft ft ft I, wt 1 « r ft trioi § ft I i ft s t r a i n 149 Figure 4.13 Ce l l cycle expression of R t t l O l p Cells are arrested in G l phase by a-factor, S phase by HU, and M phase by nocodazole, and the corresponding FACS analyses are shown. Protein expression of RttlOlp-HA is analyzed, using NdclOp as a loading control. A slower migrating form of RttlOlp-HA was shown in HU and nocodazole arrested cells (indicated by *). 150 CHAPTER 5 Conclusions and Future Directions 151 5.1 Conclusions During each cell cycle, accurate transmission of chromosomes to daughter cells is crucial to the maintenance of genetic information in an organism. Failure to do so can be detrimental. Chromosome instability is commonly observed in cancers, and has been proposed to underlie tumorigenesis. To better understand the cellular mechanisms used to maintain chromosome stability, it is necessary to identify genes required for the various processes involved. Since genes involved in basic cellular mechanisms are often conserved throughout eukaryotes, model organisms have been effectively utilized to study these processes. In Chapter 2,1 presented a systematic examination of all non-essential gene deletion mutants in the budding yeast Saccharomyces cerevisiae to screen for mutants with a C I N phenotype using 3 complementary chromosome marker loss assays. The chromosome transmission fidelity (CTF) assay monitors loss of an artificial chromosome fragment by a colony color-sectoring readout. The bimater ( B i M ) assay monitors loss of heterozygosity at the mating type locus in homozygous diploid deletion mutants using a mating test. The a-like faker ( A L F ) assay detects loss of the MATa mating type locus in haploid deletion mutants by a mating test. The 3 screens identified an overlapping and unique set of genes. In total, 293 C I N mutants were identified, including genes already known to function in the maintenance of chromosome integrity, and genes not previously known to be important for chromosome maintenance. A further application of this study was the ability to provide a list o f candidate human C L N genes based on their sequence similarity to the yeast C L N genes identified. Review of the literature indicated that some human C L N genes, including 10 homologous to the yeast C L N genes identified in our screens, were mutated in cancers. B y definition, the remaining human C L N genes represent candidate genes that may be somatically mutated in cancers. In Chapter 3,1 described the somatic mutation analysis of 101 candidate human C L N genes in a panel of colorectal cancer patients. Nove l mutations were identified, including mutations in genes functioning in sister chromatid cohesion with statistically significant frequencies. Knowing the mutational spectrum in different 152 types of cancers is useful for classification of cancers based on their differential genetic and cellular characteristics. In Chapter 2 and 3,1 discussed the strategy of targeting the genetic vulnerability of C I N cancers containing a C I N gene mutation by synthetic lethality. Selective ki l l ing of several C I N cancer cells may be possible by inactivating a common protein that is required for the viability in cells with various C L N mutations. The common protein can be identified first using genome-wide synthetic lethal interaction analysis in yeast, and then the synthetic lethal interactions can be tested and verified in human cells. The genome-wide C I N gene screening in yeast also provided a rich source of genes for the study of cellular processes important for chromosome maintenance. MMS22, MMS1, RTT101 and RTT107, identified in my C L N screens, were not well characterized at the time of identification. A s described in Chapter 4,1 performed a battery of molecular, genetic, and biochemical analyses to gain insight into the functions of these 4 genes. Additionally, ongoing more detailed studies of these genes by other groups were published during the course of my study. Taken together, these results demonstrate that these genes are required for the recovery from D S B s and may form a ubiquitin ligase complex that regulates protein stability, possibly including Mms22p, during the response to D N A damage. A s highlighted by this study, research in yeast has significant implications to the development of a cancer therapeutic strategy based on candidate C L N gene identification (as described in Figure 1.9). First, the genome-wide identification of C L N genes in yeast has provided a systematic source of candidate human C I N gene that may be relevant to tumorigenesis. Second, synthetic lethal interactions identified in yeast serve as prototypes for testing analogous interactions in human. Third, detailed characterization of yeast C L N genes may shed insights on the functions of orthologous human genes. Elucidating the conservation in gene function and genetic interactions between human and model organisms is the key for success in such translational studies. 153 5.2 Future Directions The construction of the non-essential gene deletion mutant set in yeast, containing - 8 5 % of all the genes, has allowed systematic assessment and comparison of the many phenotypes among mutants, such as C I N , cohesion defects (Marston et al., 2004), morphological defects (Ohya et a l , 2005), telomere maintenance (Gatbonton et al., 2006), and sensitivity to a range of agents (Bennett et al., 2001; Birre l l et al., 2001; Chang et a l , 2002; Giaever et al., 1999; Hanway et al., 2002; Parsons et al., 2004). In order to gain a complete understanding of basic cellular mechanisms such as the maintenance of chromosomes, essential gene mutants w i l l also need to be assessed. In this aspect, several resources have been or are being developed, including a tetracycline-inducible gene collection ( Y u et al., 2006), a set of genes that are'fused to a heat-inducible-degron cassette which targets the fused protein for proteolysis at 37°C (Dohmen et al., 1994), and a collection of temperature-sensitive or hypomorphic alleles for essential genes (Shay Ben-Aroya and Phi l Hieter, personal communication). Phenotypes in hypermorphs have also been explored systematically by overexpressing each protein in yeast (Sopko et al., 2006). With advances in resources developed for reverse genetics approaches, such as R N A interference, screens for genome instability phenotypes have been pursued in other organisms such as C. elegans, which serves as an excellent multi-cellular model (Shima et al., 2003; van Haaften et al., 2004). With the rapid and on-going generation of high-throughput data sets, advances in bioinformatics should allow the integration of data in an organized, interpretable way that w i l l be useful to biologists (Kelley and Ideker, 2005). To delineate the relationship between C L N and cancers, systematic mutation, expression, and D N A modification (e.g. methylation) analyses should be applied to a comprehensive set of candidate human C L N genes derived. The recent large-scale mutational testing project of over 13,000 genes in 2 cancer types by the Vogelstein group has set a baseline for mutation prevalence in a typical C L N cancer, which is important for the interpretation of the significance in any mutation study. Importantly, examining the C L N phenotypes in cells with a mutated or misregulated gene, and understanding how the 154 underlying mutations cause these phenotypes, w i l l be a critical next step to distinguish between passenger mutations and mutations that perturb function. A s mentioned in Chapter 1, the genetic vulnerability of C I N cancer cells can be explored beyond synthetic lethal interactions to synthetic dosage lethality. In addition, since many cancers are polyploid, identifying genes that are essential only in polyploid cells but not in diploid cells may lead to the discovery of novel drug targets that are specific to cancer cells (Storchova et al., 2006). For example, Pellman found that wi ld-type tetraploid yeast cells have a high incidence of defective kinetochore-microtubule attachments, which may be related to scaling defects in SPBs, spindles and kinetochores (reviewed in (Storchova and Pellman, 2004)). Often, the understanding of human cellular biology can benefit from studies in model organisms; but the opposite is also true. Although the Rt t l01p-Mms22p-Mmslp complex, a putative ubiquitin ligase complex, has no substrate identified yet, the homology of individual yeast proteins to proteins involved in D N A damage response and chromosome maintenance in other organisms (e.g. C U L 4 A (H.s), T a z l (S.p), andDDBl (H.s), respectively) has shed light on the function of the complex in regulating protein levels during D N A damage response. To further characterize the role of this complex, two approaches w i l l be necessary. First, it w i l l be essential to directly test candidate substrates hypothesized based on physical and genetic interactions. In addition, unbiased genome-wide screening for the substrates w i l l also be informative in characterizing the function of the complex. Complementing basic research in model organisms such as yeast with clinical findings of cancer patients w i l l advance our understanding in the genetic basis of cancer. Such knowledge w i l l allow better diagnosis, classification of tumors, and prognosis. More importantly, this understanding w i l l facilitate the development of therapeutics that is selective to cancer cells. 155 R E F E R E N C E S Adams, R.R. , Eckley, D . M . , Vagnarelli, P., Wheatley, S.P., Gerloff, D . L . , Mackay, A . M . , Svingen, P .A . , Kaufmann, S.H. and Earnshaw, W . C . (2001) Human L N C E N P colocalizes with the Aurora -B/AIRK2 kinase on chromosomes and is overexpressed in tumour cells. Chromosoma, 1 1 0 , 65-74. Al-Tassan, N . , Chmiel, N . H . , Maynard, J., Fleming, N . , Livingston, A . L . , Wil l iams, G.T., Hodges, A . K . , Davies, D.R. , David, S.S., Sampson, J.R. and Cheadle, J.P. (2002) Inherited variants of M Y H associated with somatic G : C ~ > T : A mutations in colorectal tumors. Nat Genet, 3 0 , 227-232. Andrews, P .D. , Ovechkina, Y . , Mortice, N . , Wagenbach, M . , Duncan, K . , Wordeman, L . and Swedlow, J.R. (2004) Aurora B regulates M C A K at the mitotic centromere. Dev Cell, 6 , 253-268. Araki , Y . , Kawasaki, Y . , Sasanuma, H . , Tye, B . K . and Sugino, A . (2003) Budding yeast mcml0/dna43 mutant requires a novel repair pathway for viability. Genes Cells, 8, 465-480. Ault , J .G. and Nicklas, R . B . (1989) Tension, microtubule rearrangements, and the proper distribution of chromosomes in mitosis. Chromosoma, 98, 33-39. Aylon , Y . , Liefshitz, B . , Bitan-Banin, G . and Kupiec, M . (2003) Molecular dissection of mitotic recombination in the yeast Saccharomyces cerevisiae. Mol Cell Biol, 2 3 , 1403-1417. Babu, J.R., Jeganathan, K . B . , Baker, D.J . , Wu , X . , Kang-Decker, N . and van Deursen, J . M . (2003) Rae l is an essential mitotic checkpoint regulator that cooperates with Bub3 to prevent chromosome missegregation. J Cell Biol, 1 6 0 , 341-353. Baetz, K . , Measday, V . and Andrews, B . (2006) Revealing hidden relationships among yeast genes involved in chromosome segregation using systematic synthetic lethal and synthetic dosage lethal screens. Cell Cycle, 5, 592-595. Baker, D.J . , Jeganathan, K . B . , Cameron, J.D., Thompson, M . , Juneja, S., Kopecka, A . , Kumar, R., Jenkins, R . B . , de Groen, P .C. , Roche, P. and van Deursen, J . M . (2004) B u b R l insufficiency causes early onset of aging-associated phenotypes and infertility in mice. Nat Genet, 3 6 , 744-749. Baldwin, E . L . , Berger, A . C . , Corbett, A . H . and Osheroff, N . (2005) Mms22p protects Saccharomyces cerevisiae from D N A damage induced by topoisomerase II. Nucleic Acids Res, 3 3 , 1021-1030. Bannister, A . J . , Zegerman, P., Partridge, J.F., Miska , E . A . , Thomas, J.O., Allshire, R . C . and Kouzarides, T. (2001) Selective recognition of methylated lysine 9 on histone H3 by the H P 1 chromo domain. Nature, 4 1 0 , 120-124. Bardi, G . , Parada, L . A . , Bomme, L . , Pandis, N . , Wil len, R., Johansson, B . , Jeppsson, B . , Beroukas, K . , Heim, S. and Mitelman, F. (1997) Cytogenetic comparisons of synchronous carcinomas and polyps in patients with colorectal cancer. Br J Cancer, 7 6 , 765-769. Bassett, D .E . , Jr., Basrai, M . A . , Connelly, C , Hyland, K . M . , Kitagawa, K . , Mayer, M . L . , Morrow, D . M . , Page, A . M . , Resto, V . A . , Skibbens, R . V . and Hieter, P. (1996) 156 Exploiting the complete yeast genome sequence. Curr Opin Genet Dev, 6 , 763-766. Bassett, D .E . , Jr., Boguski, M . S . , Spencer, F., Reeves, R., K i m , S., Weaver, T. and Hieter, P. (1997) Genome cross-referencing and X R E F d b : implications for the identification and analysis of genes mutated in human disease. Nat Genet, 1 5 , 339-344. Be l l , D .W. , Varley, J . M . , Szydlo, T .E. , Kang, D . H . , Wahrer, D . C . , Shannon, K . E . , Lubratovich, M . , Verselis, S J . , Isselbacher, K . J . , Fraumeni, J.F., Birch, J . M . , L i , F.P., Garber, J.E. and Haber, D . A . (1999) Heterozygous germ line h C H K 2 mutations in Li-Fraumeni syndrome. Science, 2 8 6 , 2528-2531. Bennett, C . B . , Lewis, L . K . , Karthikeyan, G . , Lobachev, K . S . , Jin, Y . H . , Sterling, J.F., Snipe, J.R. and Resnick, M . A . (2001) Genes required for ionizing radiation resistance in yeast. Nat Genet, 2 9 , 426-434. Bernard, P., Maure, J.F., Partridge, J.F., Genier, S., Javerzat, J.P. and Allshire, R . C . (2001) Requirement of heterochromatin for cohesion at centromeres. Science, 2 9 4 , 2539-2542. Biggins, S. and Murray, A . W . (2001) The budding yeast protein kinase Ipl l /Aurora allows the absence of tension to activate the spindle checkpoint. Genes Dev, 1 5 , 3118-3129. Biggins, S. and Walczak, C . E . (2003) Captivating capture: how microtubules attach to kinetochores. Curr Biol, 1 3 , R449-460. Birrel l , G.W. , Giaever, G . , Chu, A . M . , Davis, R . W . and Brown, J . M . (2001) A genome-wide screen in Saccharomyces cerevisiae for genes affecting U V radiation sensitivity. Proc Natl Acad Sci USA, 9 8 , 12608-12613. Blake, D . , Luke, B . , Kanellis, P., Jorgensen, P., Goh, T., Penfold, S., Breitkreutz, B . J . , Durocher, D . , Peter, M . and Tyers, M . (2006) The F-box protein Dia2 overcomes replication impedance to promote genome stability in Saccharomyces cerevisiae. Genetics. Bloom, K . S . and Carbon, J. (1982) Yeast centromere D N A is in a unique and highly ordered structure in chromosomes and small circular minichromo'somes. Cell, 29, 305-317. Boland, C R . and Ricciardiello, L . (1999) How many mutations does it take to make a tumor? Proc Natl Acad Sci USA, 9 6 , 14675-14677. Bomme, L . , Bardi, G . , Pandis, N . , Fenger, C , Kronborg, O. and Heim, S. (1998) Cytogenetic analysis of colorectal adenomas: karyotypic comparisons of synchronous tumors. Cancer Genet Cytogenet, 1 0 6 , 66-71. Bothos, J., Tuttle, R . L . , Ottey, M . , Luca, F .C . and Halazonetis, T .D. (2005) Human L A T S 1 is a mitotic exit network kinase. Cancer Res, 6 5 , 6568-6575. Brachmann, C . B . , Davies, A . , Cost, G.J. , Caputo, E . , L i , J., Hieter, P. and Boeke, J.D. (1998) Designer deletion strains derived from Saccharomyces cerevisiae S288C: a useful set of strains and plasmids for PCR-mediated gene disruption and other applications. Yeast, 1 4 , 115-132. Breitkreutz, B . J . , Stark, C. and Tyers, M . (2003a) The G R I D : the General Repository for Interaction Datasets. Genome Biol, 4 , R23. 157 Breitkxeutz, B . J . , Stark, C. and Tyers, M . (2003b) Osprey: a network visualization system. Genome Biol, 4, R22. Brummelkamp, T.R. and Bernards, R. (2003) New tools for functional mammalian cancer genetics. Nat Rev Cancer, 3, 781-789. Budd, M . E . , Tong, A . H . , Polaczek, P., Peng, X . , Boone, C. and Campbell, J .L. (2005) A Network of Multi-Tasking Proteins at the D N A Replication Fork Preserves Genome Stability. PLoS Genet, 1, e61. Buonomo, S.B., Clyne, R . K . , Fuchs, J., Loid l , J., Uhlmann, F. and Nasmyth, K . (2000) Disjunction of homologous chromosomes in meiosis I depends on proteolytic cleavage of the meiotic cohesin Rec8 by separin. Cell, 103, 387-398. Cagney, G . , Uetz, P. and Fields, S. (2000) High-throughput screening for protein-protein interactions using two-hybrid assay. Methods Enzymol, 328, 3-14. Cahil l , D.P. , da Costa, L .T . , Carson-Walter, E . B . , Kinzler, K . W . , Vogelstein, B . and Lengauer, C. (1999) Characterization of M A D 2 B and other mitotic spindle checkpoint genes. Genomics, 58, 181-187. Cahil l , D.P. , Lengauer, C , Y u , J., Riggins, G.J . , Wil lson, J .K. , Markowitz, S.D., Kinzler, K . W . and Vogelstein, B . (1998) Mutations of mitotic checkpoint genes in human cancers. Nature, 392, 300-303. Chan, G . K . , Jablonski, S.A., Sudakin, V . , Hittle, J .C. and Yen , T.J. (1999) Human B U B R 1 is a mitotic checkpoint kinase that monitors C E N P - E functions at kinetochores and binds the cyclosome/APC. J Cell Biol, 146, 941-954. Chan, G . K . , L i u , S.T. and Yen , T.J. (2005) Kinetochore structure and function. Trends Cell Biol, 15, 589-598. Chan, G . K . , Schaar, B .T . and Yen, T.J. (1998) Characterization of the kinetochore binding domain of C E N P - E reveals interactions with the kinetochore proteins C E N P - F and h B U B R l . J Cell Biol, 143, 49-63. Chang, M . , Bellaoui, M . , Boone, C. and Brown, G . W . (2002) A genome-wide screen for methyl methanesulfonate-sensitive mutants reveals genes required for S phase progression in the presence of D N A damage. Proc Natl Acad Sci USA, 99, 16934-16939. Chang, M . , Bellaoui, M . , Zhang, C , Desai, R., Morozov, P., Delgado-Cruzata, L . , Rothstein, R., Freyer, G . A . , Boone, C. and Brown, G .W. (2005) R M I 1 / N C E 4 , a suppressor of genome instability, encodes a member of the RecQ helicase/Topo III complex. Embo J, 24, 2024-2033. Cheeseman, I . M . , Anderson, S., Jwa, M . , Green, E . M . , Kang, J., Yates, J.R., 3rd, Chan, C.S. , Drubin, D . G . and Barnes, G . (2002a) Phospho-regulation of kinetochore-microtubule attachments by the Aurora kinase Ip l lp . Cell, 111, 163-172. Cheeseman, I . M . , Drubin, D . G . and Barnes, G . (2002b) Simple centromere, complex kinetochore: linking spindle microtubules and centromeric D N A in budding yeast. J Cell Biol, 157, 199-203. Chen, C. and Kolodner, R . D . (1999) Gross chromosomal rearrangements in Saccharomyces cerevisiae replication and recombination defective mutants. Nat Genet, 23, 81-85. 158 Chen, R . H . , Brady, D . M . , Smith, D . , Murray, A . W . and Hardwick, K . G . (1999) The spindle checkpoint of budding yeast depends on a tight complex between the M a d l and Mad2 proteins. Mol Biol Cell, 1 0 , 2607-2618. Chen, Y . , Riley, D.J . , Chen, P .L . and Lee, W . H . (1997) H E C , a novel nuclear protein rich in leucine heptad repeats specifically involved in mitosis. Mol Cell Biol, 1 7 , 6049-6056. Cheok, C.F. , Bachrati, C .Z . , Chan, K . L . , Ralf, C , Wu , L . and Hickson, I.D. (2005) Roles of the Bloom's syndrome helicase in the maintenance of genome stability. Biochem Soc Trans, 3 3 , 1456-1459. Chin, J .K. , Bashkirov, V . I . , Heyer, W . D . and Romesberg, F .E . (2006) Esc4/Rttl07 and the control of recombination during replication. DNA Repair (Amst), 5 , 618-628. Cimin i , D . , Howell , B . , Maddox, P., Khodjakov, A . , Degrassi, F. and Salmon, E . D . (2001) Merotelic kinetochore orientation is a major mechanism o f aneuploidy in mitotic mammalian tissue cells. J Cell Biol, 1 5 3 , 517-527. Ciosk, R., Shirayama, M . , Shevchenko, A . , Tanaka, T., Toth, A . and Nasmyth, K . (2000) Cohesin's binding to chromosomes depends on a separate complex'consisting of Scc2 and Scc4 proteins. Mol Cell, 5 , 243-254. Ciosk, R., Zachariae, W. , Michaelis, C , Shevchenko, A . , Mann, M . and Nasmyth, K . (1998) A n ESP1/PDS1 complex regulates loss of sister chromatid cohesion at the metaphase to anaphase transition in yeast. Cell, 9 3 , 1067-1076. Clark, G . M . , Al l red, D . C . , Hilsenbeck, S.G., Chamness, G .C . , Osborne, C . K . , Jones, D . and Lee, W . H . (1997) Mitosin (a new proliferation marker) correlates with clinical outcome in node-negative breast cancer. Cancer Res, 5 7 , 5505-5508. Cleveland, D .W. , Mao, Y . and Sullivan, K . F . (2003) Centromeres and kinetochores: from epigenetics to mitotic checkpoint signaling. Cell, 1 1 2 , 407-421. Cohen-Fix, O., Peters, J . M . , Kirschner, M . W . and Koshland, D . (1996) Anaphase initiation in Saccharomyces cerevisiae is controlled by the APC-dependent degradation of the anaphase inhibitor Pds lp . Genes Dev, 1 0 , 3081-3093. Compton, D . A . (2006) Chromosomes walk the line. Nat Cell Biol, 8 , 308-310. Connelly, C. an