....yu. . 5:. a .... , 4.3%.: L. V . aways} 1:3...m; O I‘l‘ . o h" 25.! l I . In.“ ru.?.3 $1.an . :1 fr; ...:. h. . .53 .. v1.5. . ¢ .4 .. Luv»! .1. 59:. 4.. 2 E .4 luff. .... ; .antll. f? V ’1‘ 3......1 V . ti. LIBRARIES MICHIGAN STATE UNIVERSITY EAST LANSING, MICH 48824-1048 \ “N 998 3051'” This is to certify that the dissertation entitled FUNCTIONS OF RA052 IN DNA DAMAGE REPAIR AND TELOMERE MAINTENANCE IN SACCHAROMYCES CEREVIS/AE presented by LIANJIE Ll has been accepted towards fulfillment of the requirements for the PhD. degree in MICROBIOLOGY AND MOLECULAR GENETICS WM. /W Major Q’rofessor's Signature M/OZ/sz Date MS U is an Affirmative Action/Equal Opportunity Institution -’ —'— -- ' _ h PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 6/01 c/CIRC/DateDue.p65-p.15 FUNCTIONS OF RAD52 IN DNA DAMAGE REPAIR AND TELOMERE MAINTENANCE IN SACCHAROM Y CES CEREVISIAE BY LIANJIE LI A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Microbiology and Molecular Genetics 2004 ABSTRACT FUNCTIONS OF RAD52 IN DNA DAMAGE REPAIR AND TELOMERE MAINTENANCE IN SACCHAROMYCES CEREVISIAE Lianjie Li Homologous recombination is the major and the most efficient pathway to repair double-strand breaks in Sacc/zaronn’ces cerevisiae. This pathway depends on the RAD52 epistasis group, including RAD50, RAD5I, RAD52, RAD54, RAD55, RAD57, RAD59, RDH54. MREII and XRS2. Among these genes, RAD52 plays a central role as it is required in virtually all forms of homologous recombination. Homologous recombination also plays important roles in telomere maintenance in telomerase-negative yeast cells. Two distinct homologous recombination pathways operate in telomerase-negative yeast cells. The type I pathway, which generates type I survivors, depends on RAD52, RAD51. RAD54. RAD55 and RAD5 7; the type II pathway, which generates type II survivors, requires RAD50, RADSL’, RAD59. MREI 1 and XRS2. In this thesis, I have examined the effects of RAD52 overexpression on DNA damage repair. I demonstrated that overexpression of Rad52 has a strong negative effect on DNA damage repair induced by the DNA-damaging agent methylmethanesulfonate. This effect is mediated by sequestration of RadSl by excess amounts of Rad52. I found that Rad52 overexpression also has a negative effect on a RAD5l-independent DNA damage repair pathway(s). However, this additional effect appears to be nonspecific. In addition, I found that among the five in-frame 5’-terminal ATGs in the RAD52 gene, the third, fifth, and possibly the fourth, can be used as a translation initiation codon in vivo. Rad52 translated from the fifth ATG is as competent as that translated from the third ATG in DNA damage repair induced by methylmethanesulfonate. I have also studied how Rad52 differentially participates in the two homologous recombination pathways that generate survivors in telomerase-negative yeast cells. I screened a library of randomly mutagenized rad52 and identified fifty-seven rad52 alleles that have negative effects on survivor pathways. I studied four alleles for their functions in telomerase-independent telomere maintenance and homologous recombination. I found that rad52R7OG. rad52K15QE, rad52Rl 715 have defects in the type II pathway, and they seem to be able to carry out the type I pathway normally. In contrast, rad520164G has defects in both type I and type II pathways. I also demonstrated a correlation between the two survivor pathways and different homologous recombination events: a mutant defective in the type I pathway also has defects in interchromosomal recombination; a mutant defective in the type II pathway also has defects in direct-repeat recombination. This correlation supports the proposed model of interchromosomal recombination for type I survivors. The results also argue that telomeres in type II survivors are most likely maintained by telomere looping back to copy telomere repeats intrachromosomally. To my husband, Xiaoyu, and my daughter, Evelyn ACKNOWLEDGEMENTS I would like to thank my guidance committee, Dr. Susan Conrad, Dr. Walter Esselman, Dr. Michele Fluck and Dr. Min-Hao Kuo for their understanding, support, collective thoughts and invaluable expertise. Especially, I give my sincere thanks to Dr. Walter Esselman and Dr. Min-Hao Kuo who have taken the responsibility to be my mentors and made my education during graduate school a success. I would like to thank my former mentor Dr. He Wang for his guidance and help. I would also like to thank all my friends in the Department of Microbiology and Molecular Genetics, who have offered me enormous friendship and support. I would like to thank Xiaoyu, my husband. for his understanding and support through these years. Finally, I would like to thank Evelyn, my daughter. who gives me so much joy and happiness. TABLE OF CONTENTS LIST OF TABLES .................................................................................. vii LIST OF FIGURES .............................................................................. viii CHAPTERI: LITERATURE REVIEW ........................................................................... 1 INTRODUCTION ............................................................................................ l HOMOLOGOUS RECOMBINATION PATHWAYS FOR DOUBLE-STRAND BREAKS REPAIR ......................................................... 2 RECOMBINATION PROTEINS IN S. CEREVISIAE ..................................... 5 STRUCTURE, FUNCTION AND MAINTENANCE OF S. CEREVISIAE TELOMERES ................................................................................................. 12 SUMMARY .................................................................................................... 26 APPENDIX I: TABLES AND FIGURES FOR CHAPTER I ...................... 28 REFERENCE .................................................................................................. 45 CHAPTER 2: STRONG NEGATIVE EFFECTS OF RAD52 OVEREXPRESSION ON DNA DAMAGE REPAIR IN S. CEREVIAISE ....................................... 61 ABSTACT ...................................................................................................... 61 INTRODUCTION .......................................................................................... 62 MATERIALS AND METHODS .................................................................... 63 RESULTS ....................................................................................................... 67 DISCUSSION ................................................................................................. 74 SUMMARY .................................................................................................... 79 APPENDIX 2: FIGURES AND TABLES FOR CHAPTER 2 ...................... 81 REFERENCES ............................................................................................. 125 CHAPTER 3: NOVAL ALLELES OF RAD52 THAT DEFERENTIALLY AFFECT THE TWO SURVIVOR PATHWAYS IN S. CEREVISIAE ........................ 128 ABSTRACT .................................................................................................. 128 INTRODUCTION ........................................................................................ I29 MATERIALS AND METHODS .................................................................. 132 RESULTS ..................................................................................................... I38 DISCUSSION ............................................................................................... 145 SUMMARY .................................................................................................. 150 APPENDIX 3: FIGURES AND TABLES FOR CHAPTER 3 .................... 152 REFERENCES ............................................................................................. 190 v i Table 1-1 Table 2—1 Table 2-2 Table 3-1 Table 3-2 Table 3—3 Table 3-4 LIST OF TABLES Homologous recombination proteins and their functions ......................... 39 S. cerevisiae strains used in this study ...................................................... 82 Constructs used in this study ............................................................... 83, 84 S. cerevisiae strains used in this study .................................................... 153 Plasmids used in this study ..................................................................... 154 rad52 alleles identified in this study ............................................... 161, I62 Differential utilization ofthe two survivor pathways by rad52 mutants .................................................................................... 173 vii LIST OF FIGURES Figure I-1 Double-strand break repair model ............................................................... 31 Figure 1-2 Synthesis-dependent strand annealing ......................................................... 33 Figure 1-3 Single-strand annealing ................................................................................ 35 Figure 1-4 Break-induced replication ............................................................................ 37 Figure 1-5 Model for RAD5l-cata1yzed homologous pairing and strand exchange ..... 39 Figure 1-6 Detection oftelomeres in type I and type 11 survivors by Southern blotting analysis ................................................................................. 41 Figure 1-7 Two survivor pathways that maintain telomeres in the absence of telomerase ......................................................................................................... 43 Figure 2-1 Analysis ofthe MMS sensitivity of racl52A cells expressing RAD52 variants containing a single translation initiation site and controlled by RAD52 genomic promoter ......................................................... 85, 87, 89, 91, 93 Figure 2-2 Analysis ofthe MMS sensitivity ofrad52A cells expressing Different RA D52 constructs controlled by the galactose— inducible GAIJ promoter ............................................................................ 95, 97 Figure 2-3 Analysis of RAD52 expression level in rad52A cells expressing different RAD52 constructs controlled by the GA Ll promoter ............................................................................ 99. 101, 103. 105, 107 Figure 24 Analysis ofthe MMS sensitivity of )‘(ld52A cells co-overexpressing R.»I[)5._7 with RAID” 1 or R4 D5 9 ......................... 109, l l l. I I3 viii Figure 2-5 Analysis ofthe MMS sensitivity of l‘(l(/52A cells overexpressing a mutant Rad52 with a defect in interaction with Rad51 ............................... l 15, l 17, 1 19 Figure 3-1 Construction ofrad52 library .................................................................... 155 Figure 3-2 Screening scheme for rad52 mutants ......................................................... 157 Figure 3-3 Growth phenotype ofJPlbéSlO expressing rad5] alleles ........................ 159 Figure 3-4 Distribution ofmutations ........................................................................... 163 Figure 3-5 MMS sensitivity assay ofrad52 mutants ................................................... 165 Figure 3-6 Steady state protein levels of rad52 alleles ................................................ 167 Figure 3-7 Growth potential ofrad52 mutants through senescence and recovery process ....................................................................................... 169 Figure 3-8 Southern blot analysis oftelomeric DNA in survivors obtained by liquid assay .................................................................................. 171 Figure 3-9 Southern blot analysis oftelomeric DNA in survivors obtained by single colony assay ...................................................................... 174 Figure 3-10 Growth potential ofrad52 mutants in a rad5 l A strain through senescence and recovery process ....................................................... 176 Figure 3-1 1 Growth potential ofrad52 mutants in a rad5921 strain through senescence and recovery process ....................................................... 178 Figure 3-12 Effect of Rad52 mutations on interchromosomal recombination and direct-repeat recombination ............................................................................ 180 Figure 3-13 Effects of Rad52 mutations on survivor pathways and homologous recombination ...................................................................... 188 CHAPTER 1 LITERATURE REVIEW INTRODUCTION Homologous recombination refers to the exchange or transfer of information between homologous partners. The primary function of homologous recombination in mitotic cells is to repair double-strand breaks resulting from replication fork collapse, from spontaneous damage, and from exposure to DNA-damaging agents (reviewed in [1]). In Sacclzaromyces cerevisiae, homologous recombination depends on the RAD52 epistasis group, including RAD50, RAD5I, RAD52, RAD54, RDH54, RAD55. RAD57. RAD59, MREII, and XRS2 (reviewed in [1, 2]). RAD52 is the central component (reviewed in [1, 2]). It is required in virtually all homologous recombination pathways. Mutations of RAD52 lead to the most severe defects in homologous recombination and highest sensitivity to DNA-damaging agents. Telomeres are. specialized protein-DNA structures at the ends of eukaryotic chromosomes. Their specialized structure caps chromosome ends to provide protection against degradation and end-to-end fusion, as well as to prevent chromosome ends from being recognized as double-strand breaks [3]. In most cases, telomeres are replenished by telomerase, the reverse transcriptase which adds telomere repeats to chromosome ends [3]. In the absence of telomerase, telomeres continue to shorten with each cell division until eventually, they lose their capping function. The chromosome ends are then recognized as double-strand breaks and induce cell cycle arrest at Gg/M phase [4]. Most telomerase-negative S. cwevisiae cells cease division in 50-100 generations. However, a small population of cells can survive and proliferate at a rate similar to telomerase- positive cells [5]. In these survivors, telomeres are maintained by homologous recombination between telomere repeats or subtelomeric regions [5-7]. Since RAD52 is the central player in homologous recombination, it is not surprising that it also plays key roles in telomere maintenance in telomerase-negative cells. Double knockout mutants of telomerase and RAD52 can not generate any survivors [5]. This thesis will focus on the functions of Rad52 in DNA damage repair and telomerase-independent telomere maintenance. Therefore, this literature review will address the following topics: (1) homologous recombination pathways for double-strand breaks repair; (2) recombination proteins in S. cerevisiae, with an emphasis on Rad52 and Rad51; and (3) structure, function and maintenance of S. cerevisiae telomeres HOMOLOGOUS RECOMBINATION PATHWAYS FOR DOUBLE-STRAND BREAKS REPAIR Double-strand breaks (DSB) are generally considered the most severe DNA damage in mitotic cells. Homologous recombination is the major and the most efficient repair pathway in S. cerevisiae. This pathway utilizes a sequence homologous to the damaged DNA for repair. There are several mechanisms of homologous recombination by which yeast cells repair DSB, including gene conversion, single-strand annealing, and break-induced replication (reviewed in [1, 2]). I‘d Gene conversion Gene conversion is defined as a nonreciprocal transfer of genetic information between two homologous partners. Two models proposed for gene conversion have been widely accepted: double-strand break repair and synthesis-dependent strand annealing. Double-strand break repair model (DSBR) In DSBR model (Figure 1-1), the 5’ ends ofa DSB are resected to form 3’ single- stranded tails that can invade an intact homologous template. Following strand invasion, the 3’ end acts as a primer for new DNA synthesis. The noninvading 3’ single-stranded tail on the other side ofthe DSB will pair with the D—loop formed by strand invasion, and initiate DNA synthesis. This process leads to the fomtation of a double-Holiday-junction. Alternate resolution of the two Holiday junctions will yield crossover or noncrossover products. If the resolution is random, an equal number of crossover or noncrossover products should be expected. However. only 10-20% of mitotic gene conversion events are associated with crossing over (reviewed in [2]). Synthesis-dependent strand annealingLSDSA) SDSA model was proposed to account for the low frequency of gene conversion associated with crossing over. In this model, one or both resected 3’ ends invades the homologous duplex and initiates DNA synthesis. For a two-ended invasion, both newly synthesized strands will be unwound from their templates and anneal to each other (Figure l-2A) [1, 2]. For a one-end invasion, three possible subsequent events have been proposed [2]. The newly synthesized strand is displaced and annealed to the other side of the D88 (Figure l-ZB). Alternatively, the second 3’ end anneals with the D-loop (Figure 1-2C). In the third scenario, which is termed repair replication fork capture, the invasion of one 3’ end establishes a modified replication fork. DNA synthesis will continue until the repair replication fork is “captured” by the other side of the D88 (Figure l-2D). It should be pointed out that the annealing with D-loop and replication fork capture mechanisms can produce crossover products. Single-strand annealing(SSA) SSA is an efficient repair pathway when a D88 occurs between direct repeats [2]. In SSA model, 5’ to 3’ resection of a D88 produces 3’ single-stranded tails. The resection will continue until homologous sequences are revealed. The homologous single-stranded DNA then anneals, and the nonhomologous tails are removed, resulting in deletion ofthe intervening sequence and one ofthe repeats [1, 2, 8](Figure I-3). Break-induced replication (BIR) Double-stranded breaks sometimes produce only one end. Collapsed replication forks generate only one end. The chromosome ends will also be recognized as one-ended double-strand breaks when telomeres are uncapped and eroded. These damages can not be repaired by gene conversion, which requires a second end. Instead, these DSBs are repaired by break-induced replication, in which the broken ends invade homologous sequences and initiate DNA synthesis to copy all the donor sequences to the chromosome ends (Figure 1-4) [1, 2]. This process is thought to act to maintain telomeres in telomerase-negative cells [7, 9]. RECOMBINATION PROTEINS IN S. CEREVISIAE In S. cerevisiae, homologous recombination depends on the RAD52 epistasis group, including RAD52, R5105], RADS4, RAD55, RAD57, RAD59, RDH54, RAD50. MREII and XRSZ (Table 1-1). Most of these genes were identified by their increased sensitivity to DNA-damaging agents, such as ionizing radiation (reviewed in [1]). These genes can be further grouped into two subgroups. One consists of RAD50, MR5]! and XRS2 and the other consists of RAD52, RAD5I, RAD54, RAD55, RAD57, RAD59 and RDH54. Within the RAD52 subgroup, RAD52 stands alone as it is essential for all forms of homologous recombination during mitotic growth. RAD51. RAD54, RAD55 and RAD57 are required for gene conversion and break-induced replication. whereas RAD5 9 is involved in break-induced replication and single-strand annealing. R4050, MR5]! and XRS2 function at the early stage of DSB repair. They are involved in processing DSBs to form 3’ single-stranded DNA tails that can invade homologous donor and initiate DNA synthesis (Figure 1—5) (reviewed in [10]). The 3’ single-stranded tails are bound by Rad51, the recombinase that catalyzes homologous pairing and strand exchange [11]. The functions of Rad51 are stimulated by Rad52, RadS4, Rth4, Rad55 and RadS7 at different phases during recombination (Figure 1-5 and see below). RadSO/Mre11/Xr52 Rad50, Mrell and er2 form a complex (MRX) which probably contains two molecules of Rad50, two molecules of Mrell and one molecule of er2 [12, 13]. Mrell (~80 kDa) has manganese-dependent nuclease activities [14-16], including a 3’ to 5’ exonulease activity on both double- and single-stranded DNA, an endonulcease activity on single-stranded DNA, and a structure-specific endonulease activity that cleaves the 3’ single-stranded overhangs at the single-/double-stranded junction. The exonuclease activity and the structure-specific endonuclease activity of Mrell is enhanced by interacting with Rad50 [15, 16] and er2 [l7]. er2 (~96 kDa) binds to both single- and double-stranded DNA. However, its preferred substrate is tailed duplexes, indicating that er2 recognizes the junction between the double- and the single—stranded regions of DNA molecules [17]. Rad50 (~150 kDa) has an ATP-dependent double-stranded DNA binding activity [18]. The conserved nucleotide-binding motifs Walker A and Walker B are located at the N-terminus and the C-terminus, respectively. The ATP-binding motifs are indispensable for Rad50 functions [19]. MRX complexes are involved in processing DSBs into 3’ single-stranded tails that initiate strand invasion [10, 20]. However, Mrell is a 3’ to 5’ exonuclease. Several models have been proposed to solve the directionality conflict (review in [10]). MRX may cooperate with a heliease to unwind DNA duplex, Mrell could then process the ends by its endonuclease activity. Alternatively, MRX is responsible for the initial processing of DSBs using its structure-specific endonuclease activity, after which other nucleases further process the ends to generate the 3’ single-stranded tail. Indeed, a DNA unwinding activity has been observed for MRX [17]. Rad51 RAD51 encodes a 400-amino acid protein (43 kDa) with significant homology to RecA [21, 22]. The highest homology is located at the central portion ofthe two proteins, including the Walker A and Walker B motifs for nucleotide binding and/or hydrolysis. Like RecA, Rad51 catalyzes homologous pairing and strand exchange [11]. The strand exchange reaction catalyzed by Rad51 occurs from 3’ to 5’ relative to the single strand [23. 24]. This polarity is opposite to that observed for RecA. However, it has been shown that Rad51 can catalyze strand exchange bidirectionally when the ends of a double- stranded DNA exist as overhanging structures [25]. Rad51 binds to both ssDNA and dsDNA to form right-handed helical filaments similar to that formed by RecA [23. 26]. However, dsDNA with single-stranded tails is the preferred binding substrates of Rad51 [27].While ssDNA binding is greatly enhanced by the presence of ATP, dsDNA binding completely depends on ATP [21]. ssDNA- Rad51 filaments are active in strand exchange, whereas dsDNA-Rad51 filaments in fact inhibit the reaction [23]. Rad51 has a DNA-dependent ATPase activity [I 1]. ssDNA is more effective in activating ATP hydrolysis. When a conserved lysine residue (K191) in the Walker A-box is mutated to arginine, the ATPase activity is abolished [24]. However, the mutant protein can still bind to DNA in a ATP-dependent manner, and catalyze homologous pairing and strand exchange. Consistent with this result, wild type Rad51 is able to catalyze homologous pairing and strand exchange in the presence ofnonhydrolizable ATP analogs [24]. Furthermore, rad51K191R can complement the MMS sensitivity of rad5/A strains [24]. These results argue that nucleotide binding is sufficient for Rad51 biological functions. Rad51 self-associates through its N-terminal domain [28]. The importance ofthis interaction is demonstrated by the dominant negative phenotype of rad5 l K 1 91A . rad5/K 159A is inactive in strand exchange [24]. When expressed from a high copy plasmid, rad5/K159A negatively affects DNA repair in wild type cells, but it has no additional effect in rad51£l strains, suggesting that rad5/[(191.4 exerts its negative effect by associating with wild type Rad51 [28]. In addition, the crystal structure of Rad51 nucleoprotein filaments suggests that the functional unit of Rad51 is a dimer [29]. Since dsDNA-Rad51 nucleoprotein filaments inhibit strand exchange [23], secondary structures within single stranded DNA substrates must be eliminated. This function is fulfilled by replication protein A (RPA) [30], a heterotrimeric single-stranded DNA binding protein (Figure 1-5). However, the stimulatory effect of RPA on strand exchange can only be observed when it is incorporated into reactions after Rad51 has bound to ssDNA [31—33]. If RPA is introduced in the nucleation phase of Rad51, it inhibits the subsequent reaction by competing for ssDNA binding [31, 34]. RPA is abundant and present during ssDNA—Rad51 filament formation in viva. Cells overcome the inhibitory effect of RPA by employing a set of mediator proteins, including Rad52 and Rad55/Rad57 heterodimer (discussed below). Rad54 and Rth4 RAD54 encodes a 898-amino acid protein (102 kDa) with a dsDNA-dependent ATPase activity [35]. Rad54 topologically unwinds dsDNA [36, 37]. It directly interacts with Rad51 as demonstrated by two-hybrid and co-immunoprecipitation experiments [38, 39]. In vitro, Rad54 stimulates Rad51-mediated strand exchange reactions [35, 36, 40- 42]. This stimulatory effect is not due to Rad54 facilitating Rad51 nucleoprotein filament formation, but rather is mediated by direct interactions of Rad54 with preassembled filaments [37]. This interaction also enhances the ATPase and DNA unwinding activity of Rad54 [37], which in turn promotes Rad51-catalyzed homologous pairing [36]. In addition, Rad54 can stabilize Rad51 nucleoprotein filaments [43], as well as stimulate heteroduplex extension of joint molecules [40]. Therefore, Rad54 functions in both synaptic and postsynaptic phases during strand exchange (Figure 1-5). Rad54 has also been shown to displace Rad51 from dsDNA, which may be important for Rad51 tumover [10, 41]. Rth4 is a Rad54 homologue. Like Rad54, it possesses a dsDNA-dependent ATPase activity and promotes a conformational change of circular dsDNA [44]. It also displays a similar stimulatory effect on Rad51-catalyzed strand exchange reactions [44]. Rad55/Rad57 Rad55 (406 aa, 46 kDa) and Rad57 (460aa, 52 kDa) are referred to as Rad51 paralogs since they share sequence similarity with Rad51, especially at the putative nucleotide binding motifs [45, 46]. Mutation of the conserved lysine residue within the Walker A-type box to arginine or alanine in Rad55 (Lys49) results in a severe defect in DNA damage repair induced by 7 radiation. However, an analogous mutation in Rad57 (Lysl3 1) has no effect on 'y-ray resistance [47]. Unlike Rad51, neither Rad55 nor Rad57 exhibits self-interaction [47]. Instead, Rad55 and Rad57 form a stable heterodimer [34, 47, 48]. Rad55 also interacts with Rad51 in two-hybrid systems [34, 47, 48]. Rad55/Rad57 promotes Rad51-catalyzed strand exchange by facilitating the displacement of RPA from ssDNA (Figure 1-5) [34]. Consistent with this observation, a set of rad5] alleles with an increased DNA binding activity can partially bypass the requirement of Rad55/Rad57 [49]. Rad55/Rad57 only interacts with Rad51, but not with RPA [47, 48], suggesting that the mediator function of Rad55/Rad57 is different from that of Rad52. which interacts with both Rad51 and RPA (see below). Rad52 RAD52 encodes a protein of 471 amino acids with a molecular weight of 52.4 kDa. Rad52 has multiple functional domains. The partially overlapping DNA binding and self-association regions are located at its N-terminus [50-52]. These regions are highly conserved throughout eukaryotes. The C-terminal two-thirds of the protein interacts with RPA and Rad51 [50, 53-55]. This region is less conserved. In fact, the Rad52-Rad51 interaction is species-specific [54]. Rad52 is expressed constitutively throughout the cell cycle [56]. It forms discrete foci during S phase [57]. Rad52 expression is induced by 9-fold early in meiosis [58]. Surprisingly, DNA damaging agents only have a moderate effect at very high dosage [56]. Consistent with this observation, Rad52 forms multiple foci (~15/nucleus) in meiotic cells, whereas only 1-2 foci per cell when cells are treated with y—rays [57]. Rad52 performs multiple functions essential for homologous recombination. Rad52 can bind to both ssDNA and dsDNA with a slight preference for ssDNA [50]. Human Rad52 has also been shown to specifically bind to ssDNA termini and tailed duplex DNA [59]. Both yeast and human Rad52 form multimeric ring structures [51, 60, 61]. Human Rad52 rings have been shown to further assemble into higher order multimers [51]. While the ring formation between monomers is mediated by the conserved N-terminal self-association domain, the assembly of the higher structures requires the C-terminus. Rad52 by itself can efficiently promote annealing between short oligonucleotides [50, 60]. However, it needs RPA to efficiently anneal longer ssDNA. Since RPA has little effect on the annealing of longer DNA free of secondary structures, such as poly(dT), its primary role is to eliminate secondary structures in DNA molecules 10 [62]. Rad52 also interacts with Rad59, which has been suggested to augment Rad52’s activity in strand annealing [63]. Rad52 acts as a mediator between RPA and Rad51 to stimulate Rad51-meidated DNA strand exchange by facilitating Rad51 nucleating on ssDNA substrates (Figure 1-5) [31, 33, 64, 65]. The physical interaction between Rad52 and Rad51 is required for this activity since a rat/52 mutant unable to interact with Rad51 also fails to perform its mediator function [55]. The C-temtinal one-third of Rad52 is both necessary and sufficient for the Rad52-Rad51 interaction [54, 55]. The relative ratio of Rad52 to Rad51 is important for Rad52’s mediator function [32]. A maximal mediator function is achieved when Rad52 is about one-tenth of the amount of Rad51. The physical interaction between Rad52 and RPA has been demonstrated in a yeast two- hybrid system as well as in co-immunoprecipitation experiments [53, 60], and is suggested to be important for Rad52’s mediator function [53, 64]. Rad52 exhibits a different stimulatory function probably by stabilizing Rad51 presynaptic filament through its interaction with Rad51when RPA is present at subsaturating Ievel[66]. However, unlike its mediator role, the physical interaction of Rad52-RPA is not required in this C'dSC. Rad59 RAD59 encodes a 238 amino acid protein (26 kDa) with a significant homology to the N-terminal half of Rad52 [67]. Rad59 shares several biochemical activities with Rad52. It binds to DNA, with a higher affinity for ssDNA. It self-associates to form multimers [68]. Rad59 also anneals complementary ssDNA in vitro [63, 69]. However, unlike Rad52, Rad59’s single strand annealing activity is not promoted by RPA [69]. In fact, Rad59 can not displace RPA from ssDNA [63]. The physical interaction between Rad52 and Rad59 has been demonstrated using a two-hybrid system and by co- immunoprecipitation experiments [63]. A complex containing Rad52, Rad51 and Rad59 can be immunoprecipitated, but Rad51 and Rad59 fail to interact in the absence of Rad52 [1], suggesting that Rad59 and Rad51 bind to different interaction interfaces on Rad52. STRUCTURE, FUNCTION AND MAINTENANCE OF S. CEREVISIAE TELOMERES Telomere biology began with the pioneering work done by Herman Muller and Barbara McClintock back in the 19305. Muller analyzed chromosome rearrangement following X-irridiation in Drosophila [70]. He never recovered a chromosome with terminal deletion. McClintock discovered that broken chromosome ends in maize could fuse with each other to form an unstable dicentric chromosome [71]. These pioneering studies sugvest that telomeres are essential parts of eukaryotic chromosomes, and that C they have special structures and functions to prevent chromosome fusion. Telomeric DNA and telomere-associated sequences Telomeres are specialized DNA and protein structures at the ends of eukaryotic chromosomes. Telomeric DNA consists of tandem arrays of short repeats. These repeats are polarized with G-rich repeats oriented from 5’ to 3’ (centromere to telomere), therefore called G-strand for convenience (and the complementary strand C-strand). However, the two strands do not contain significantly more G or C residues in some species [72]. Some organisms, such as tetrahymena and some fungi, have homogenous arrays of repeat sequences, while others have heterogeneous repeats (reviewed in [73, 74]). The sequence of telomeric DNA in S. cerevisiae is (TG)i-(,TG3-3 [75]. The length of yeast telomeric DNA at individual telomeres varies with an average length of 300i75 base pairs [76]. Telomeres do not terminate with blunt ends ([77] and reviewed in [73]). The G strand extends to form a 3’ overhang. Long single-stranded G-tails (~50-100 bases) can be detected at late S phase in yeast [78-80]. In addition to the simple repeats, many organisms also have middle repetitive DNA sequences located immediately internal to the short repeats (reviewed in [73, 81]). These sequences are referred to as telomere-associated sequences (TAS). Yeast has two classes of TAS, X and Y’ [82]. X, ranging from 0.3 to 3.75 kb. consists of at least five species of ~45-~560 bp in size, not all of which are present in every X [83]. However, the longest species, also called core X (~560 bp), is present at most telomeres [83]. Y’ has two major variants, the 6.7 kb Y’ long and the 5.2 kb Y’ short. Y’ is found on only a subset of chromosomes [84]. Some chromosomes have up to 4 copies of Y’ elements [82]. If both X and Y’ are present on the same chromosome, X is located internal to Y’ [83]. Telomere chromatin strfiucture Depending on the type of telomeres examined, telomeres may assume different chromatin structures (reviewed in [73]). In yeast, the entire terminal Cl-3A/TG.-3 duplex is packed in a nonnucleosomal chromatin structure called telosome [76]. The telosomes do not seem to build around the nucleosomal cores since the stability of protein-DNA interactions in telosomes is different from those in nucleosomes [85], and deletions or duplications of histone genes which lead to changes in histone stoichiometry do not disrupt telomere chromatin structures [86]. Furthemtore, the telosome contains twice as much DNA as a nucleosome because the entire telomere duplex is protected by telosome [76]. Both X and Y’ elements are assembled in nucleosomes. However, subtelomeric nucleosomes are less accessible to dam methylase [87]. In addition, the amino terminal tails of histones H3 and H4 are hypoacetylated [88]. These results suggest that subtelomeric regions are more compact than elsewhere in the genome. Functions of Telomeres and TAS Early work of B. McClintock suggested that telomeres are important for chromosome stability [71]. Indeed, yeast cells with defects in EST], which lead to progressive shortening of telomeres, exhibit increased chromosome loss and cell death [89]. Greider group examined the mutation rate of the CAN] locus as telomeres progressively shorten in est] strains [90]. Early passages of est] cells exhibit a mutation rate similar to that of wild type strain. The mutation rate increases about 10 -fold at the peak of senescence when telomeres are critically short, then drops to wild type level as survivors emerge (survivors are discussed in detail in the section of Telomere Maintenance in Yeast). By monitoring the fate of CAN] and the ADE2 gene which was placed telomeric to CAN], Greider and co-worker also showed that telomere shortening leads to terminal deletions. Therefore, the increased mutation rate is caused by gross chromosomal rearrangements, rather than small deletions or point mutations. In another study, the fate ofa nonessential test chromosome in which the entire telomere tract can be eliminated in a controlled manner was monitored [91]. The loss rate ofthis chromosome increases ~10 -fold after cleavage ofthe telomere tract. In cells that the test chromosome is maintained, the telomere-less chromosome is frequently (~7()%) healed by RAD52- mediated homologous recombination or de novo telomere addition. However, unhealed chromosomes can be replicated and segregated for four to ten cell divisions before being lost. These results suggest that while telomeres are essential for stable maintenance of yeast chromosomes, they are probably not required for cells to maintain a chromosome for a given cell cycle [91]. Telomeres help to distinguish chromosome ends from DNA breaks. Double- strand DNA breaks induce RAD9-dependent G3 /M arrest. The loss of a single telomere on a nonessential chromosome leads to temperate RAD9-mediated Gg/M arrest [91]. The Gg/M arrest is more permanent if multiple dysfunctional telomeres are present [92]. Telomeres also serve as substrates for telomerase, the specialized reverse transcriptase that elongates telomeres. Conventional DNA polymerases need RNA primers to synthesize DNA. Removal of the most distal RNA primer leads to incomplete replication of the lagging strand. Without a mechanism to compensate for the sequence loss, telomeres will continuously shorten with each cell division. Telomerase allows the complete replication of the ends of the chromosomes by utilizing its RNA subunit as the template for telomere replication (Also discussed in the section of Telomere Maintenance in Yeast). In addition, telomeres affect the transcription of adjacent genes. The transcription of a gene is repressed when it is placed near a telomere [93]. This phenomenon is referred to as telomere position effect (TPE). In general, TPE is reversely proportional to the distance from telomeres [94]. Longer telomeres have greater silencing effect [95]. However, iftelomere lengthening is accompanied by the loss of certain telomere binding proteins, TPE could be reduced [96]. One possible function of TAS is to act as a buffer zone to prevent TPE from repressing essential genes located near telomeres [3]. It should be noted that not every natural telomere displays TPE [97], and the density of ORFs near telomeres is lower than elsewhere in the genome (reviewed in [3]). In telomerase-negative yeast cells, TAS is also important for telomere maintenance. This function will be discussed in the section of Telomere Maintenance in Yeast. Telomere binding proteins Telomeres provide binding sites for proteins that are important for maintenance of telomere length and structures. These proteins can be divided into four groups based on the sites they bind to: proteins that bind to the single-stranded tails ofG strands, including C dc l 3, Stn 1 , Ten] and Estl; proteins that bind to the border between the double-stranded and the single-stranded region of telomeres, including yKu70 and yKu80; proteins that bind to the double-stranded region of telomeres, including Rapl, Rifl, Rif2, and Sir proteins; and proteins that bind to subtelomeric regions, such as belp [98-100]. CDC/3 encodes an essential protein (924 aa, 104 kDa) that specifically binds to single-stranded telomeric DNA [101-104]. It contributes to telomere maintenance in two ways: by protecting telomere ends and by controlling the access of telomerase [105]. Cdcl3, along with Stnl and Tenl, forms a complex to “cap” telomeres. Single-stranded G-tails generated at late S phase are normal intermediates of telomere replication [78]. They are generated by degradation ofC-strands [78, 80]. A temperature sensitive mutant, Cdc/34, accumulates single-stranded DNA at chromosome ends and arrests at G2 in a RAD9-dependent manner when grown at restrictive temperature (37”C) [103, 106]. The single-stranded DNA, could be as long as 30 kb, contains TAS and telomeres l6 corresponding to the G strands [103]. These data suggest that Cdc13 limits C-strand degradation at the late S phase and helps to shield chromosome ends from DNA damage checkpoint. Cdc13 regulates telomerase activity both positively and negatively. Cdc13 is responsible for recruiting telomerase through the interaction between Cdcl3 and Estl, a subunit of telomerase holoenzyme [107]. The length of telomere repeats is increased about 0.9 kb in ("(1613-1 mutant, suggesting that Cdcl3 also serves as a negative regulator oftelomere replication [108]. STN] was isolated in a screen looking for suppressors for ('(/('/3-/. Like Cdcl3-l, a temperature sensitive mutant Suzi-13 accumulates single-stranded DNA at the chromosome ends and displays elongated telomeres [108]. Physical interaction between Cdcl3 and Stnl has been demonstrated by a two-hybrid system [108]. A fusion protein consists ofthe DNA-binding domain of Cdc13 and Stnl is able to rescue the lethality of cat-132] strain, indicating that Stnl is the primary participant in chromosome end protection, and Cdc13 serves as a delivery vehicle [107]. TNE] (160 aa) was isolated as a suppressor of temperature sensitive sm] mutants [109]. It interacts with both Cdc13 and Stnl [109]. Like ca’cl3-1 and Still-13, ten] mutants also accumulate long ssDNA at telomeric region which induces a RAIN-dependent G3 arrest [109]. While overexpression of Tenl can not complement ale] 3-] , rescue of ('dC] 3-1 by Stnl can be improved by co- overexpression of Stnl and Tenl [109]. In addition, Tenl-Stnl fusion protein rescues inviability of sin/A cells and ten] A cells. These data suggest that Tenl participates in chromosome end protection and telomere length regulation in association with Stnl and Cdc l 3, and together these proteins form a protective cap to shield telomeres [1 10]. Estl is a subunit of telomerase holoenzyme. It will be discussed in the section of Telomere Maintenance in Yeast. yKu70/de1 and yKu80/de2 form a heterodimer to bind to the junction between the single- and double-stranded regions of telomeres. They play key roles in telomere structures. Mutations in either protein result in the increase in the single-stranded G tails throughout cell cycle [111, 112]. yKu70/yKu80 also positively regulates telomerase activity. A Cdcl3-yKu7O fusion protein results in longer than wild type length telomeres [1 13]. Although yKu70/yKu80 functionally interacts with Cdcl3, they show no association in viva. It appears that yK70/th180 fulfill their function in telomerase regulation by interacting with TLC] RNA, a subunit oftelomerase [l 14]. Rapl (827 aa, 120 kDa) binds to the double-stranded region of telomeres to regulate telomere length and TPE [3, 100, 115]. Rapl interacts with Rifl [l 16] and RifZ [1 17] to form a negative regulator for telomere addition. Deletion mutations of RIF] and R1F2, as well as a C-terminal truncation mutation of Rapl result in dramatic telomere elongation [116, 117]. It has been proposed that telomere length is regulated by a negative feedback mechanism in which the number of Rapl molecules bound to telomeres is counted [1 18]. Rapl also acts as a positive regulator for telomere elongation [119, 120]. It appears that Rapl helps to recruit telomerase and increase the activity of telomerase [1 19]. It has been suggested that the balance between intemal Rapl promoting telomerase activity and Rapl binding to the more terminal region oftelomeres controlling telomerase access maintain telomeres at a constant length [1 19]. Rap] interacts with Sir3 and Sir4 to form complexes to organize heterochromatin formation at telomeres and other transcription silencing loci [100, 121-124]. Sir2 is involved in those complexes through the interaction with Sir4 [125]. Telomere maintenance in yeast Telomere maintenance by telomerase In wild type yeast, telomere replication occurs in late S phase [126]. There are three activities participating in telomere replication in yeast. The bulk oftelomeric DNA is replicated by conventional DNA polymerases. Telomerase binds to the single-stranded tails of G-strands and elongates telomeres. C-strands can then be replicated by conventional DNA polymerase using G-strands as the templates. Telomerase holoenzyme consists of four subunits: a RNA subunit encoded by TLC] and a catalytic protein subunit encoded by EST2 form a catalytic core; two accessory subunits encoded by EST] and EST 3 regulate in viva telomerase activity [127- 130]. In vitro, TLC] RNA and Est2 alone can catalyze telomere addition, since cell extracts from est/A or (4821 strains display telomerase activity [131]. However, both Estl and Est3 are necessary in viva. Deletion of either gene leads to progressive telomere shortening and senescence, the same phenotype shown by {lo/A or (as/2A mutants [127, 128]. The 1.3 kb TLC] RNA contains a sequence of 5’-CACCACACCCACACAC-3’ serving as the template for telomere replication [132]. It also serves as a scaffold for the assembly of telomerase holoenzyme [133-135]. Binding of Est2 to TLC] RNA requires nt. 101-138, and nt. 728-864, binding ofEstl to TLC] RNA needs nt. 553-707 [136]. Est2 is a 804-amino acid, 103 kDa protein with a reverse transcriptase activity [127, 137, 138]. The reverse transcriptase domain lies between amino acids 420-740 [137]. There are three invariant aspartic acid residues within that motif among different reverse transcriptases. Point mutations of these residues in Est2 (Asp530Ala or Asp530G1u; Asp670Ala; Asp67l Ala) abolish its reverse transcriptase activity and lead to telomere shortening and senescence [137] [138]. EST] encodes a 82 kDa single-stranded telomere binding protein [89, 128, 139]. Efficient binding of Estl to chromosome ends requires Cdc13-Est1 interaction [129]. Estl directly interacts with TLC] RNA as demonstrated by co-immunoprecipitation experiments [140, 141]. It has been proposed that Estlrecruits telomerase to the single- stranded chromosome termini as an adaptor between Cdcl3 and the catalytic core of telomerase [107, 130]. Indeed, fusion protein of Est2-Cdc13 rescues the senescence phenotype of estI/J cells [130]. However, recent studies reveal that Estl and Est2 telomeric binding is uncoupled [142], and Est2 associates with telomeres in the G1 phase ofthe cell cycle when telomeres are not replicated [143]. These findings lead to a second model of Estl action in which Est2/ TLC] RNA associates with telomeres nonproductively by binding to the more internal regions of telomeres, and Cdcl3-Estl will translocate Est2/TLC] RNA to telomere termini [129]. Indeed, telomeric binding of Estl requires a free 3’ terminus [139]. Est3 is a stable component of telomerase since it is co-immunoprecipitated with TLC] RNA and telomerase activity [144]. Its association with telomerase complex requires an intact catalytic core [144]. A recent study suggests that the N-temtinal domain of Est2 is required for Est3 binding [145]. The precise function of Est3 is still unknown. Telomere maintenance in telomerase-negative survivors Survival through senescence 20 Yeast cells with TLC] or any of the EST genes deleted display gradual telomere shortening accompanied by a progressive decline in growth potential. termed cellular senescence [5, 89, 127, 132]. Most cells cease division after 50-100 generations. However, a subpopulation outgrow senescence and become survivors [5, 6]. Survivor generation in a telomerase-deficient strain is not an isolated event. In fact, it occurs with a high frequency as demonstrated by the appearance of survivors in all of the more than 100 est/A strains examined in one study [5]. Survivors display dramatic changes in subtelomeric and telomeric regions [5, 6]. Based on those changes, the survivors are grouped into two types. Type I survivors have amplified Y’ elements (70-fold on average) followed by short tracts of telomeric repeats. Type 11 survivors maintain long and heterogeneous telomeric repeats with little Y’ amplification. These two types of survivors can be distinguished by the pattern of telomeric X1201 fragments using Southern blot analysis (Figure 1-6) [5, 6]. There is a single X1101 site located at ~900 bp from the 3’ end of Y’ elements. Type I survivors yield three major X1201 fragments detected by a 3’ Y’ probe or a poly(dG-dT) probe (Figure 1- 6A & B). The sizes of these bands are ~1.3 kb, 5.2 kb and 6.7 kb. The ~1.3 kb fragment is the terminal fragment consisting of the distal portion of the terminal Y’ and telomeric repeats. The strong signals at 5.2 and 6.7 kb are due to amplified Y’-short and Y’-long, respectively. In contrast, type 11 survivors have many X1101 fragments with different sizes, which hybridize to a 3’ Y’ probe and a poly(dG-dT) probe (Figure 1-6A & B). These fragments can not be detected by probes that hybridize to other portions of Y’ (Figure 1- 6A), indicating that they are terminal fragments that contain telomere repeats. Although survivors are healthy cells that have recovered from senescence, continued streakouts for single colonies reveal variable growth patterns in both types of survivors [5]. Some survivors display stable growth rates that are comparable to wild type cells for extended periods, while others show a gradual decline in growth rate and senesce again. For the latter group, survivors can reappear readily. It appears that all type 1 survivors undergo senescence repeatedly, whereas only a subset of type 11 survivors display similar re-senescence phenotype. In addition, type 11 survivors grow faster than type I survivors [6]. Amplification of Y’ results in about 10% increase in genome size. The burden of replicating such increased genome might contribute to the growth disadvantage of type I survivors [146]. Type I survivors are not stable. They can convert to type 11 during outgrowth [6]. In contrast, type 11 survivors are stable. The type 11 pattern of telomeric X1201 fragments can be maintained for at least 250 generations [6]. However, the individual X1201 fragment shortens overtime. When a single telomere was marked, the rate of telomere shortening was measured at ~3bp/cell division [6]. Survimrs and 12022201090218 ree0mbination The appearance of survivors from telomerase-deficient strains is mediated by homologous recombination since no survivor can be recovered from strains lacking both telomerase and Rad52, which is essential in virtually all forms of homologous recombination [5, 7, 147]. The roles of the RADS2 epistasis group in telomerase- independent telomere maintenance have been studied in detail. While RAD52 is indispensable for both type I and type 11 survivors, RAD5], RAD54, Rad55 and presumably RAD5 7 are essential to generate type I survivors [7, 9], and RAD50, Mre] 1, er2 and RAD59 are required to generate type 11 survivors [7, 9]. Double mutants of tlelA and genes involved in the generation of either type I or type 11 survivors (but not both) do not affect survivor generation, whereas triple mutants of field rad5 1A rac]5()/1 and 110121 rad5/A l‘(l(/59A completely block survivor generation [7, 147]. The genetic requirements of the two survivor pathways and the structures of survivor telomeres have led to the proposal oftwo distinct genetic pathways that function in telomerase-deficient yeast cells to maintenance telomeres (Figure 1-7) [7]. The type I pathway, which generates type I survivors, is mediated by recombination between Y’ elements on different chromosomes. The type II pathway, which generates type 11 survivors, is mediated by recombination between telomere repeats on the same or different chromosomes. Although telomere shortening results in senescence, senescence is not strictly correlated with telomere length. Double mutants of tie/.4 20215221. tlel/J rad5 1A, tie/A rat/54A and tlelzl rad5721 display an accelerated decline in growth potential compared to tlelA single mutants [5, 147]. However, single mutants with deletion mutations of these recombination genes have telomere length similar to that in wild type cells [147]. In addition. the rate of telomere shortening in tlelzf mic/52A mutants is similar to that in tlcrlzf mutants [7]. These observations also suggest that recombination starts to contribute to telomere functions in the initial phase oftelomere shortening. It might be expected that rare survivors arising from telomerase-deficient strains are hyper-recombination mutants. In fact, recombination rates in survivors and wild type cells are statistically indistinguishable [5]. However, when a recombination reporter is placed in the subtelomeric region of one telomere, it is found that recombination rate is increased by up tolOOO-fold in telomerase-deficient strains [148]. Therefore, survivors display hyper-recombination phenotype in a telomere-specific manner. Four homologous recombination-based mechanisms have been proposed for telomere maintenance in telomerase-deficient cells [146]: (1) break-induced replication, (2) integration of extrachromosomal DNA, (3) rolling circle replication, and (4) elongation via t-loop. Break-induced replication (BIR): is a one-ended nonreciprocal recombination process in which a broken chromosome end invades into homologous sequences on an intact chromosome and copy the donor sequence all the way to telomeres. There are a RAD51-dependent BIR pathway, as well as a RAD51-independent, RAD50/RAD59- dependent BIR pathway [149-153]. This suggests that type I survivors are generatedvia RAD5I-dependent BIR, whereas type 11 survivors arise through RAD51-independent BI R. It seems that the degree of homology between Y’ elements or that between telomere repeats could be one of the factors that determine which pathway to employ. Rad51 is very sensitive to the mismatches in the homologous region during strand exchange [154]. The efficiency of strand exchange is only 20% of the wild type level when a 6 bp nonhomologous insertion exists in a duplex substrate [155]. Indeed, Y’ elements are highly conserved with about 1% divergence within a strain [156]. F urthemtore, although Y’ long and Y’ short share more than 5 kb homology, most Y’-Y’ recombination occurs between elements of the same size [157]. In addition, the RAD51-dependent pathway needs at least ~100 bp of homology to initiate strand invasion, whereas the RAD51- independent pathway requires only ~30 bp [150]. Integration of extrachromosomal DNA and rolling circle replication: These two models provide altemative mechanisms to explain the sudden changes in the size of telomeric and subtelomeric repeats that can not be readily explained by BIR. Telomeres in type 11 survivors continue to shorten at a rate of ~3 bp/cell division. This gradual shortening is interspersed with episodes of sudden telomere elongation, increasing the size of telomeres by 1 to 2 kb [9]. This one-step of telomere elongation has been proposed to be mediated by integration of multiple extrachromosomal telomeric DNA [146]. Alternatively, the 3’ tail of a G-strand could invade an extrachromosomal telomeric circle and prime DNA synthesis [9, 146]. Elongation of t-loop: This model provides an alternative to the rolling circle replication model. Instead of invading an extrachromosomal telomeric circle, the 3’ tail of a G-strand invades the internal duplex telomeric region and forms an intramolecular loop. This structure, called t—Ioop, has been observed in evolutionarily unrelated organisms [158, 159], suggesting they are a conserved feature of eukaryotic telomeres. Although t-loops have not been observed in yeast telomeres, similar structures have been proposed to mediate telomere length regulation and transcriptional regulation of genes placed in subtelomeric region [122. 160-162]. In addition, the telomere binding protein Rapl can promote association of single-stranded telomeric sequence with its homologous duplex sequence [163]. Telomerase-independent telomere maintenance in human cells: ALT Most of human tumor samples and immortalized human cell lines exhibit telomerase activity. However, a subset oftumor cells and cell lines maintain telomeres in the absence of telomerase ([164] reviewed in [165]). These telomeres are maintained by so-called Alternative Lengthening of Telomere (ALT) pathway. Rapid elongation of telomeres following gradual shortening has been observed in human telomerase-negative cells [166]. The long and heterogeneous telomeres observed in ALT cells are similar to that in type II yeast survivors, suggesting that human ALT is mediated by a recombination process similar to that occurs in type II yeast survivors [6, 146]. Indeed, DNA sequences can be copied from telomere to telomere [167]. ALT human cells contain ALT-associated PML bodies (APB), which are novel promyelocytic leukemia (PML) bodies [168]. APBs contain extrachromosomal telomeric DNA, telomere-specific binding proteins, and proteins involved in DNA replication and recombination. Noticeably, the appearance of APB coincides with the activation of ALT. The existence of ALT poses a new question for tumor therapy. For ALT tumors, treatment with telomerase inhibitors will not be effective. For telomerase-positive tumors, telomerase inhibition can induce apoptosis and senescence [169-171]. However, such treatment may provide a selective advantage to cells that activate ALT. It may be important to develop inhibitors of ALT. It seems that normal cells and some telomerase- positive immortal cells contain repressors for the ALT telomere phenotype [172]. SUMMARY This literature review focuses on homologous recombination repair of D885 and mechanisms that contribute to telomere maintenance in Saar-120272222)'ees eerew’siae, including telomerase-dependent pathway and telomerase-independent, homologous recombination-mediated survivor pathway. Although the functions of RAD52 in homologous recombination and telomere maintenance have been studied in great detail, it remains controversial as to how cells respond to changes in Rad52 concentration. In addition, it is still unknown whether and how Rad52 differentially participates in telomere maintenance in different types of survivors. This study investigates the effects of Rad52 overexpression on DNA damage repair and demonstrates that the Rad52 cellular level needs to be tightly controlled to fulfill its functions. This study also investigates the functions of Rad52 in the two survivor pathways by charactering four novel RAD52 alleles identified in a genetic screen. APPENDIX 1: TABLES AND FIGURES FOR CHAPTER 1 Table 1-1 Homologous recombination proteins and their functions In S. cerevisiae, homologous recombination depends on the RAD52 epistasis group, including RAD52, RAD51. RAD54, RAD55, RAD57. RAD59, RD1154, RAD50, MREll and XRSZ. These genes can be further grouped into two subgroups. One group consists of RADSO, MREII and XRSZ. These genes are required for processing DNA ends. The other group consists of RAD52. RAD51. RAD54, RAD55, RAD57, RAD59 and RD1154. Within the RAD52 subgroup, RAD52 stands alone as it is essential for all forms of homologous recombination during mitotic growth. RAD51. RAD54, RADSS and RAD57 are required for gene conversion and break-induced replication. RAD59 is involved in Rad51-independent break-induced replication and single-strand annealing. ‘3 () Table l-1 Homologous recombination proteins and their functions Recombination Functions in homologous Homologous protein recombination recombination pathways Rad52 Promoting annealing between Single strand annealing; complementary single-stranded DNA ; Gene conversion; Stimulating Rad51-mediated Break—induced replication homologous pairing and strand [2] exchange [31, 66] Rad51 Catalyzing homologous pairing and Gene conversion; strand exchange [1 1] Break-induced replication [2, 153] Rad54 Stimulating Rad51-mediated strand Gene conversion; exchange [36, 40-42] Break-induced replication [2, 153] Rth4 Stimulating Rad51—mediated strand Gene conversion; exchange [44] Break-induced replication [2, 153] Rad55/Rad57 Stimulating Rad51-mediated strand Gene conversion; exchange [34] Break-induced replication [2, 153] Rad59 Promoting annealing between Single strand annealing; complementary single-stranded DNA [63, 69] Break-induced replication [2, 153] Rad50/Mre1 l/erZ Processing DSB ends [10, 20] 3t.) Figure 1-1 Double-strand break repair model After a D88 is created, 5’ to 3’ resection generates 3’ single-stranded tails. The resulting 3’ ends invade a homologous template to initiate DNA synthesis. Two Holiday junctions formed are resolved independently to generate crossover or noncrossover products. DSB 5’ to 3’ resection ——>————— / Strand invasion -—-—————>» X x New DNA synthesis V V t> m <1 Holiday junction resolution A A Non-crossover Crossover Figure 1-1 Taken from PAQUES, F and I. E. HABER Microbiol Mol Biol Rev 63(2): 394-404 Figure 1-2 Synthesis-dependent strand annealing After a D88 is created, 5’ to 3’ resection generates 3’ single-stranded tails. Both (A) or one (B, C & D) of the resulting 3’ ends invade a homologous template to initiate DNA synthesis. For a two-ended invasion. both newly synthesized strands are displayed and annealed to each other (A). For a one-end invasion, the noninvading 3’ end anneals with the displayed newly synthesized stand (B) or the D-loop (C). Alternatively, a repair replication fork can be established following strand invasion (D). 33 Non-crossover l l Non-crossover Crossover B 5’ to 3’ resection 5’ to 3’ resection Strand invasion _—_—_—+ xi:— Strand invasion New DNA synthesis :T/JC—_ New DNA synthesis Strand annealing :3“; Strand annealing Non-crossover D 5’ to 3’ resection S’ to 3’ resection Strand invasion Gk Strand invasion New DNA synthesis :5: New DNA synthesis Strand annealing —/«X Strand annealing Non-crossover Crossover Figure 1-2 Taken from PAQUES. F and J. E. IIABER Microbiol Mol Biol Rev 63(2): 394-404 34 Figure 1-3 Single-strand annealing A DSB made between direct repeats is subjected to 5’10 3’rescction. When complementary sequences are revealed, the single-stranded DNA anneals resulting in deletion ofthe intervening sequence and one ofthe repeats. DSB : l 5’ to 3’ resection {—H — ‘ w l Annealing of complementary ssDNA /_ / l 3’ end trimming ligation _ I I 1 J Deletion of intervening sequence and one copy of the repeats Figure 1-3 Taken from PAQUES, r and 1. E. HABER Microbiol Mol Biol Rev 63(2): 394-404 36 Figure 1-4 Break-induced replication When only one end of a D88 is available for homologous recombination, or a telomere becomes uncapped, the broken end can invades a homologous sequence and initiate DNA synthesis. Replication will proceed to the end ofthe chromosome. 37 Telomere uncapping Loss of the other end and degradation of a D88 protection factors, Loss of telomere degradation of 5’ strandl l 5’ to 3’ resection AL /\ CD Strand invasion, replication to end of chromosome Figure 1-4 Taken from Krogh, B. O. and L. S. Symington Annu Rev Genet 38: 233—271 38 Figure 1-5 Model for RAD51-catalyzed homologous pairing and strand exchange When a D88 is created (only one side of the D88 is shown), (1) MRX and/or other exonucleases process the ends to generate 3’ single-stranded tails. (2) RPA binds to the single—stranded tails to remove secondary structures. (3) Rad52 recruits Rad51 to the RPA-bound single-stranded DNA and facilitates the initial displacement of RPA. (4) Rad55/Rad57 facilitates Rad51 nucleoprotein filament extension. (5) The Rad51 nucleoprotein filament searches and locates homologous sequence. (6) RadS4 promotes DNA unwinding and strand annealing. 2 r. SW? strait 83 SW” 1 4 V Figure 1-5 Taken from Krogh, B. O. and L. S. Symington Annu Rev Genet 38: 233-271 40 Figure 1-6 Detection of telomeres in type I and type 11 survivors by Southern blotting analysis A. Telomeric and subtelomeric region of S. ('erevt’siae. The CHA/ TG.-3 DNA is shown in black. The open rectangle and the striped rectangle represent the Y’ and X element, respectively. The solid lines indicate probes that can be used to detect telomeres for Southern blotting analysis. X1101 restriction site is also indicated. B. Telomeres in wild type and telomerase-negative survivor yeast cells. Genomic DNA from wild type cells. type I survivors and type 11 survivors was digested with X1101, resolved in 1% agarose gel. The southern blot was hybridized to a poly(dG-dT) probe. 41 X1201 /'m I probe — Y’S’ Y’ middle Y’3’ E. E -1 -1 B :1 E S 6 (D 0 n: "' = 2m- a ' .n I I Al 6kb— I ' 5kb— "f g '.‘-.£‘| lkb— _ Figure 1-6 (TG 1-3) H Telomere shortening l I 1 l I I I 1 Type I survivors RAD5I RAD.” / .— R4052 —p RADSS RAD57 I I l I 1 Continued Telomere shortening I I Figure 1-7 44 T} pe 11 survivors \ R4050 RAD59 MREII XML? I I .' I I I .. ”Lu .. 10 Modification of the model proposed by Greider and colleagues. MCB 21(5)1819- 1827 Figure 1-7 Two survivor pathways that maintain telomeres in the absence of telomerase A telomere containing two copies of Y’ (gray boxes) and TG1-3 repeats (small white boxes) is shown at the top. Telomere shortening occurs in the absence of telomerase. Survivors can be generated via two pathways, both of which require RAD52. A. The type I survivor pathway also depends on RAD51. RAD54, RAD55 and RAD5 7. Telomere shortening exposes Y’. 3’ single-stranded tail initiate recombination between Y’ on different chromosomes. B. The type II survivor pathway also depends on RAD50, MREII, XRSZ and R405 9. Recombination is initiated between telomere repeats on different chromosomes, or the 3’ single-stranded tail pairs with the duplex region of the telomere and primes DNA synthesis. 43 REFERENCE Symington, L.S., Role ofRn'I D52 epistasis group genes in homologous recombination and double-strand break repair. Microbiol Mol Biol Rev, 2002. 66(4): p. 630-70, table of contents. Paques, F. and J.E. Haber, Multiple pathways 0frecombination induced by double-strand breaks in Saccharom}ices cerevisiae. M icrobiol Mol Biol Rev, 1999. 63(2): p. 349-404. Zakian, V.A., Structure,function, and replication of Saccharmn)'ces cerevisiae telomeres. Annu Rev Genet, 1996. 30: p. 141-72. Johnson, F.B., et al., The Saccharomyces cerevisiae WRN homo/0g Sgs] p participates in telomere maintenance in cells lacking telomerase. Embo .1, 2001. 20(4): p. 905-13. Lundblad, V. and EH. Blackburn, An alternative pathway/0r yeast telomere maintenance rescues est] - senescence. Cell, 1993. 73(2): p. 347-60. Teng, SC. and VA. Zakian, Telomere-telomere recombination is an Ljficient bypass pathwayfor telomere 222aintenance in Sacc'haromyces cerevisiae. Mol Cell Biol, 1999. 19(12): p. 8083-93. Chen, Q., A. I jpma, and CW. Greider, Two survivor pathways that allow growth in the absence of telomerase are generated by distinct telomere recombination events. Mol Cell Biol, 2001. 21(5): p. 1819-27. Sugawara, N. and J.E. Haber, Characterization ofdouble-strmul break-induced recombination: homologv requirements and single-stramled DNA formation. Mol Cell Biol, 1992. 12(2): p. 563-75. Teng, SC, et al., Telomerase-independent lengthening oilyeast telomeres occurs by an abrupt Rad50p-dependent, Riflinhibited recombinationa] process. M01 Cell, 2000. 6(4): p. 947-52. Krogh, 8.0. and LS. Symington, Rec-0mbination Proteins in 1”ast. Annu Rev Genet, 2004. 45 12. 15. I6. 17. 18. 19. 20. 21. 22. Sung, P., Catalysis of A TP-dependent homologous DNA pairing and strand exchange by yeast RAD51 protein. Science, 1994. 265(5176): p. 1241-3. Anderson, D.E., et al., Structure of the Rad50 x Mrell DNA repair complex from Saccharomyces cerevisiae by electron microscopy. J Biol Chem, 2001. 276(40): p. 37027-33. Chen, L., et al., Promotion of Dnl4-catalvzed DNA end-joining by the Rad50/Mre1 l/erZ and del/Hdfl complexes. Mol Cell, 2001. 8(5): p. 1 105-15. Usui, T., et al., Complex formation andfunctional versatility of'Mrell ofbudding yeast in recombination. Cell, 1998. 95(5): p. 705-16. Trujillo, K.M. and P. Sung, DNA structure-speci/ic nuclease activities in the Saccharomyces cerevisiae Rad50*Mre1 ] complex. J Biol Chem, 2001. 276(38): p. 35458-64. Paull, T.T. and M. Gellert, The 3 ' t0 5' exonuclease activity of Mre ] ] facilitates repair of DNA double-strand breaks. Mol Cell, 1998. 1(7): p. 969-79. Trujillo, K.M., et al., Yeast .rrs2 binds DNA and helps target rad50 and mrel] to DNA ends. J Biol Chem, 2003. 278(49): p. 48957-64. Raymond, W.E. and N. Kleckner, RA D50 protein of S.cerevisiae exhibits A T P- dependent DNA binding. Nucleic Acids Res, 1993. 21(16): p. 3851-6. Alani, E., R. Padmore, and N. Kleckner, Analysis ofwild-type and rad50 mutants of yeast suggests an intimate relationship between meiotic chromosome synapsis and recombination. Cell, 1990. 61(3): p. 419-36. Haber, J.E., The many interfaces of'Mrel]. Cell, 1998. 95(5): p. 583—6. Shinohara, A., H. Ogawa, and T. Ogawa, Rad5] protein involved in repair and recombination in S. cerevisiae is a RecA-like protein. Cell, 1992. 69(3): p. 457- 70. Aboussekhra, A., et al., Semidominant suppressors of 52s.? helicase mutations of Saccharomyces cerevisiae map in the RAD5 ] gene, whose sequence predicts a 46 25. 26. 27. 30. 32. protein with similarities to procatyotic RecA proteins. Mol Cell Biol, 1992. 12(7): p. 3224-34. Sung, P. and D.L. Robberson. DNA strand exchange mediated by a RAD51- ssDNA nucleoprotein/ilament with polarity opposite to that ofRecA. Cell, 1995. 82(3): p. 453-61. Sung, P. and SA. Stratton, Yeast Rad5] recombinase mediates polar DNA strand exchange in the absence o/‘A T P hydrolysis. .1 Biol Chem, 1996. 271(45): p. 27983-6. Namsaraev, E. and P. Berg, Characterization of strand exchange activity Q/yeast Rad5] protein. Mol Cell Biol, 1997. 17(9): p. 5359-68. Ogawa, T., et al., Similarity ofthe yeast RAD5 ] filament t0 the bacterial RecA filament. Science, 1993. 259(5103): p. 1896-9. M azin, A.V., et al., Tailed duplex DNA is the preferred substratefor Rad5 ] protein-mediated homologous pairing. Embo .I, 2000. 19(5): p. 1 148-56. Donovan, .1 .W., G.T. Milne, and D.T. Weaver, Homotypic and heterotipic protein associations control Rad5] function in double-strand break repair. Genes Dev, 1994. 8(21): p. 2552—62. Conway, A.B., et al., C rvstal structure ofa Rad5] filament. Nat Struct Mol Biol, 2004. 11(8): p. 791-6. Sugiyama, T., E.M. Zaitseva, and SC. Kowalczykowski, A single-stranded DNA- binding protein is needed for efficient presynaptic complex formation by the Saccharomyces cerevisiae Rad5] protein. J Biol Chem, 1997. 272(12): p. 7940-5. Sugiyama, T. and SC. Kowalczykowski, Rad52 protein associates with replication protein A (RPA)single-stranded DNA to accelerate Rad51-mediated displacement ofRPA and presynaptic complexformt'ttion. J Biol Chem, 2002. 277(35): p. 31663-72. Song, B. and P. Sung, Functional interactions among ,t'east Rad5] recomln'nase, Rad52 mediator, and replication protein A in DNA strand exchange. J Biol Chem, 2000. 275(21): p. 15895-904. 47 33. 34. 35. 36. 37. 38. 40. 41. 43. Shinohara, A. and T. Ogawa, Stimulation by Rad52 o/iveast Rad51-mediated recombination. Nature, 1998. 391(6665): p. 404-7. Sung, P., Yeast Rad55 and Rad57 proteins/02722 a heterodimer that functions with replication protein A to promote DNA strand exchange by Rad5 l recombinase. Genes Dev, 1997. 11(9): p. 1111-21. ' Petukhova, G., S. Stratton, and P. Sung, Catalysis of homologous DNA pairing by yeast Rad5] and Rad54 proteins. Nature, 1998. 393(6680): p. 91-4. Petukhova, G., et al., Yeast Rad54 promotes Rad5l-dependent homologous DNA pairing via A T P hydrolysis-driven change in DNA double helix con/orination. .1 Biol Chem, 1999. 274(41): p. 29453-62. Mazin, A.V., et al., Rad54 protein is targeted to pairing loci by the Rad5] nucleoproteinfilament. Mol Cell, 2000. 6(3): p. 583-92. Clever, B., et al., Recombinational repair in yeast: functional interactions between Rad5] and Rad54 proteins. Embo J, 1997. 16(9): p. 2535-44. Jiang, H., et al., Direct association between the yeast Rad5] and Rad54 recombination proteins. .1 Biol Chem, 1996. 271(52): p. 33181-6. Solinger, J.A., et al., Rad54 protein stimulates heteroduplex DNA formation in the synaptic phase of DNA strand exchange via specific interactions with the presynaptic Rad5] nucleoproteinfilament. J Mol Biol, 2001. 307(5): p. 1207-21. Solinger, J.A., K. Kiianitsa, and W.D. Heyer, Rad54. a SwiZ/Sn/2-like recombinational repair protein. disassemblcs Rad5 l .‘dsDNA filaments. Mol Cell, 2002. 10(5): p. 1175-88. Solinger, .1 .A. and W.D. Heyer, Rad54 protein stimulates the postsynaptic phase ofRad51 protein-mediated DNA strand exchange. Proc Natl Acad Sci U S A, 2001. 98(15): p. 8447-53. Mazin, A.V., A.A. Alexeev, and SC. Kowalczykowski, A nave/function of Rad54 protein. Stabilization of the Rad5] nucleoproteinfilament. J Biol Chem, 2003. 278(16): p. 14029-36. 48 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. Petukhova, G., P. Sung, and H. Klein, Promotion of Rad5 1 -dependent D-loop formation by yeast recombination factor Rdh54/Tid]. Genes Dev, 2000. 14(17): p. 2206-15. Lovett, S.T., Sequence of the RAD55 gene of Saccharomyces cerevisiae: similarity of RAD55 to prokaryotic RecA and other RecA-like proteins. Gene, 1994. 142(1): p. 103-6. Kans, J.A. and R.K. Mortimer, Nucleotide sequence of the RAD5 7 gene of Saccharomyces cerevisiae. Gene, 1991. 105(1): p. 139-40. Johnson, RD. and LS. Symington, Functional differences and interactions among the putative RecA homologs Rad5], Rad55, and Rad5 7. Mol Cell Biol, 1995. 15(9): p. 4843-50. Hays, S.L., A.A. F irmenich, and P. Berg, Complexformation in yeast double- strand break repair: participation ofRad51, Rad52, Rad55, and Rad5 7 proteins. Proc Natl Acad Sci U S A, 1995. 92(15): p. 6925-9. F ortin, GS. and LS. Symington, Mutations in yeast Rad5] that partially bypass the requirementfor Rad55 and Rad5 7 in DNA repair by increasing the stability of Rad51-DNA complexes. Embo J, 2002. 21(12): p. 3160-70. Mortensen, U.H., et al., DNA strand annealing is promoted by the yeast Rad52 protein. Proc Natl Acad Sci U S A, 1996. 93(20): p. 10729-34. Ranatunga, W., et al., Human RAD52 exhibits two modes (vised-association. J Biol Chem, 2001. 276(19): p. 15876-80. Kagawa, W., et al., Crystal structure of the homologous-pairing domain from the human Rad52 recombinase in the undecamericform. Mol Cell, 2002. 10(2): p. 359-7]. Hays, S.L., et al., Studies of the interaction between Rad52 protein and the yeast single—stranded DNA binding protein RPA. Mol Cell Biol, 1998. 18(7): p. 4400-6. Milne, G.T. and D.T. Weaver, Dominant negative alleles o/‘RAD52 reveal a DNA repair/recombination complex including Rad5] and Rad52. Genes Dev, 1993. 7(9): p. 1755-65. 49 55. 56. 57. 58. 59. 60. 61. 63. 64. 65. Krejci, L., et al., Interaction with Rad5] is indispensablefor recombination mediatorfitnction ofRad51 J Biol Chem, 2002. 277(42): p. 40132-41. Cole, G.M., et al., Regulation ofRAD54— and RAD52-lacZ gene fusions in Saccharomyces cerevisiae in response to DNA damage. Mol Cell Biol, 1987. 7(3): p. 1078-84. Lisby, M., R. Rothstein, and DH. Mortensen, Rad52f0rms DNA repair and recombination centers during S phase. Proc Natl Acad Sci U S A, 2001. 98(15): p. 8276-82. Cole, G.M., D. Schild, and R.K. Mortimer, Two DNA repair and recombination genes in Saccharomyces cerevisiae, RAD52 and RAD54, are induced during meiosis. Mol Cell Biol, 1989. 9(7): p. 3101-4. Parsons, C .A., et al., Precise binding of'single-sti'andea’ DNA termini by human RAD52 protein. Embo J, 2000. 19(15): p. 4175-81. Shinohara, A., et al., Rad52 forms ring structures and co-operates with RPA in single-strand DNA annealing. Genes Cells, 1998. 3(3): p. 145-56. Stasiak, A.Z., et al., The human Rad52 protein exists as a heptameric ring. Curr Biol, 2000. 10(6): p. 337-40. Sugiyama, T., J .H. New, and SC. Kowalczykowski, DNA annealing by RAD52 protein is stimulated by specific interaction with the complex of replication protein A and single-stranded DNA. Proc Natl Acad Sci U S A, 1998. 95(1 1): p. 6049-54. Davis, AP. and LS. Symington, The yeast recombinational repair protein Rad5 9 interacts wit/2 Rad52 and stimulates single-strand annealing. Genetics, 2001. 159(2): p. 515-25. New, J.H., et al., Rad52 protein stimulates DNA strand exchange by Rad5] and replication protein A. Nature, 1998. 391(6665): p. 407-10. Sung, P., Function of avast Rad52 protein as a mediatorbetween replication protein A and the Rad5] recombimtse. J Biol Chem, 1997. 272(45): p. 28194-7. 50 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. New, J .H. and SC. Kowalczykowski, Rad52 protein has a second stimulatory role in DNA strand exchange that complements replication protein-A function. J Biol Chem, 2002. 277(29): p. 26171-6. Bai, Y. and LS. Symington, A Rad52 homo/0g is required/or RAD51- independent mitotic recombination in Saccharomyces cerevisiae. Genes Dev, 1996. 10(16): p. 2025-37. Davis, AP. and LS. Symington, The Rad52-Rad59 complex interacts wit/2 Rad5 l and replication protein A. DNA Repair (Amst), 2003. 2(10): p. 1 127-34. Petukhova, G., S.A. Stratton, and P. Sung, Single strand DNA binding and annealing activities in the yeast recombination factor Rad5 9. J Biol Chem, 1999. 274(48): p. 33839-42. Muller, H.J., The remarking ofchromosomes. The Collecting Net, 1938. 13: p. 181-195, 198. M CC lintock, B., The behavior in successive nuclear divisions ofa chromosome broken at meiosis. Proc. Natl. Acad. Sci., 1939. 25: p. 405-416. McEachem, MJ. and EH. Blackburn, A conserved sequence motifwithin the exceptionallydiverse telomeric sequences of'budding yeasts. Proc Natl Acad Sci U S A, 1994. 91(8): p. 3453-7. Henderson, E., Telomere DNA structure. Cold Spring Harbor monograph series, 1995. 29: p. 11-34. McEachem, M.J., A. Krauskopf, and EH. Blackburn, Telomeres and their control. Annu Rev Genet, 2000. 34: p. 331-358. Shampay, J ., J.W. Szostak, and EH. Blackburn, DNA sequences of telomeres maintained in yeast. Nature, 1984. 310(5973): p. 154-7. Wright, J.H., D.E. Gottschling, and V.A. Zakian, Saccharomyces telomeres assume a non-22ucleosomal chromatin structure. Genes Dev, 1992. 6(2): p. 197— 210. ‘ 51 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. Henderson, ER. and EH. Blackburn, An overhanging 3 ' terminus is a conserved feature oftelomeres. Mol Cell Biol, 1989. 9(1): p. 345—8. Wellinger, R.J., et al., Evidence/or a new step in telomere 222aintenance. Cell, 1996. 85(3): p. 423-33. Wellinger, R.J., A]. Wolf, and V.A. Zakian, Saccharomyces telomeres acquire single-strand TGl-3 tails late in S phase. Cell, 1993. 72(1): p. 51-60. Dionne, I. and RJ. Wellinger, Cell cycle-regulated generation ofsingle-stranded G-rich DNA in the absence oftelomerase. Proc Natl Acad Sci U S A, 1996. 93(24): p. 13902-7. Louis, E.J., The chromosome ends of'Saccharom}'ces cerevisiae. Yeast, 1995. 11(16): p. 1553-73. Chan, CS. and BK. Tye, Organization of'DNA sequences and replication origins at yeast telomeres. Cell, 1983. 33(2): p. 563-73. Louis, E.J., et al., The chromosome end in yeast: its mosaic nature and influence on recombinational dvnanu'cs. Genetics, 1994. 136(3): p. 789-802. Zakian, V.A. and HM. Blanton, Distribution oftelomere-associated sequences on natural chromosomes in Saccharomyces cerevisiae. Mol Cell Biol, 1988. 8(5): p. 2257-60. Wright, J.H. and V.A. Zakian, Protein-DNA interactions in soluble telosomes from Saccharomyces cerevisiae. Nucleic Acids Res, 1995. 23(9): p. 1454-60. Norris, D., B. Dunn, and M.A. Osley, The effect of histone gene deletions on chromatin structure in Saccharomyces cerevisiae. Science, 1988. 242(4879): p. 759-61. Gottschling, D.E., T elomere-proximal DNA in Saccharomyces cerevisiae is refractory to methyltraIts/erase activity in vivo. Proc Natl Acad Sci U S A, 1992. 89(9): p. 4062-5. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. Braunstein, M., et al., Transcriptional silencing in yeast is associated with reduced nucleosome acetylation. Genes Dev, 1993. 7(4): p. 592-604. Lundblad, V. and J .W. Szostak, A mutant wit/2 a defect in telomere elongation leads to senescence in yeast. Cell, 1989. 57(4): p. 633-43. Hackett, J.A., D.M. F eldser, and CW. Greider, T elomere dysfunction increases mutation rate and genomic instability. Cell, 2001. 106(3): p. 275-86. Sandell, LL. and V.A. Zakian, Loss ofa yeast telomere: arrest, recovety. and chromosome loss. Cell, 1993. 75(4): p. 729-39. AS, 1.1. and CW. Greider, Short telomeres induce a DNA damage response in Saccharomyces cerevisiae. Mol Biol Cell, 2003. 14(3): p. 987-1001. Gottschling, D.E., et al., Position effect at S. cerevisiae telomeres: reversible repression of Pol II transcription. Cell, 1990. 63(4): p. 751-62. Renauld, H., et al., Silent domains are assembled continuouslyfrom the telomere and are defined by promoter distance and strength, and by SIR3 dosage. Genes Dev, 1993. 7(7A): p. 1133-45. Kyrion, G., et al., RAP] and telomere structure regulate telomere position effects in Saccharomyces cerevisiae. Genes Dev, 1993. 7(7A): p. 1 146-59. Wiley, EA. and V.A. Zakian, Extra telomeres, but not internal tracts oftelomeric DNA, reduce transcriptional repression at Saccharomyces telomeres. Genetics, 1995. 139(1): p. 67-79. Pryde, F .E. and El. Louis, Limitations ofsilencing at native yeast telomeres. Embo J, 1999. 18(9): p. 2538-50. Koering, C.E., et al., Identification of high affinity T bf] p-binding sites within the budding yeast genome. Nucleic Acids Res, 2000. 28(13): p. 2519-26. Bilaud, T., et al., The telobox. a Myb-related telomeric DNA binding motif found in proteins from yeast, plants and human. Nucleic Acids Res, 1996. 24(7): p. 1294-303. 53 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. Dmitriev, P.V., A.V. Petrov, and O.A. Dontsova, Yeast telosome complex: components and theirfimctions. Biochemistry (Mosc), 2003. 68(7): p. 718-34. Lin, 1.1. and V.A. Zakian, The Saccharomyces CDC 1 3 protein is a single-strand TGl-3 telomeric DNA-binding protein in vitro that affects telomere behavior in vivo. Proc Natl Acad Sci U S A, 1996. 93(24): p. 13760-5. Lin, Y.C., et al., Specific binding ofsingle-stranded telomeric DNA by C dc13p of Saccharomyces cerevisiae. J Biol Chem, 2001. 276(27): p. 24588-93. Garvik, B., M. Carson, and L. Hartwell, Single-stranded DNA arising at telomeres in cdc13 mutants may constitute a specific signal for the RAD9 checkpoint. Mol Cell Biol, 1995. 15(1 1): p. 6128-38. Boums, B.D., et al., Sir proteins, Rifproteins, and Cdc/3]) bind Saccharomt‘ces telomeres in vivo. Mol Cell Biol, 1998. 18(9): p. 5600-8. Lustig, A.J., Cdc/3 subcomplexes regulate multiple telomerefunctions. Nat Struct Biol, 2001. 8(4): p. 297-9. Weinert, TA. and L.H. Hartwell, Cell cycle arrest of'cdc mutants and specificity ofthe RAD9 checkpoint. Genetics, 1993. 134(1): p. 63-80. Pennock, E., K. Buckley, and V. Lundblad, Cdc/3 delivers separate complexes to the telomerefor end protection and replication. Cell, 2001. 104(3): p. 387-96. Grandin, N., 3.1. Reed, and M. Charbonneau, Stnl, a new Saccharonn'ces cerevisiae protein. is implicated in telomere size regulation in association with Cdc/3. Genes Dev, 1997. 11(4): p. 512-27. Grandin, N., C. Damon, and M. Charbonneau, Ten] functions in telomere end protection and length regulation in association wit/2 Stnl and Cdc/3. Embo J, 2001. 20(5): p. 1173-83. Enomoto, S., et al., Telomere cap components influence the rate. ofsencscence in telonterase-deficientyeast cells. Mol Cell Biol, 2004. 24(2): p. 837-45. 54 111. 114. 115. 116. 117. 118. 119. 120. Polotnianka, R.M., J. Li, and A.J. Lustig, The yeast Ku heterodimer is essential for protection of the telomere against nucleolvtic and recombinational activities Curr Biol, 1998. 8(14): p. 831-4. Gravel, S., et al., Yeast Ku as a regulator of'chromosomal DNA end structure. Science, 1998. 280(5364): p. 741-4. Grandin, N., C. Damon, and M. Charbonneau, C dc] 3 cooperates wit/2 the yeast Ku proteins and Stnl to regulate telomerase recruitment. Mol Cell Biol, 2000. 20(22): p. 8397-408. Peterson, S.E., et al., Thefimction ofa stem-loop in telomerase RNA is linked to the DNA repair protein Ku. Nat Genet, 2001. 27(1): p. 64-7. Gilson, E., et al., Distortion of‘the DNA double helix by RAP] at silencers and multiple telomeric binding sites. J Mol Biol, 1993. 231(2): p. 293-310. Hardy, OF, L. Sussel, and D. Shore, A RAP] -interacting protein involved in transcriptional silencing and telomere length regulation. Genes Dev, 1992. 6(5): p. 801-14. Wotton, D. and D. Shore, A novel Raplp-interactingfactor, RifL’p, cooperates wit/2 Riflp to regulate telomere length in Saccharomyces cerevisiae. Genes Dev, 1997. 11(6): p. 748-60. Marcand, S., et al., Rap] p and telomere length regulation in yeast. C iba Found Symp, 1997. 211: p. 76-93; discussion 93-103. Ray, A. and K.W. Runge, The C terminus ofthe major yeast telomere binding protein Rap] p enhances telomereformation. Mol Cell Biol, 1998. 18(3): p. 1284- 95. Krauskopf, A. and EH. Blackburn, Control oftelomere growth by interactions of RAP] with the most distal telomeric repeats. Nature, 1996. 383(6598): p. 354-7. C ockell, M., et al., The carboxy termini of'Sir4 and Rap] affect Sir3 localization: evidencefor a multicomponent complex required for yeast telomeric silencing. J Cell Biol, 1995. 129(4): p. 909-24. 55 123. 124. 125. 127. I29. 130. 131. 132. Grunstein, M., Molecular modelfor telomeric heteroc/u'omatin in yeast. Curr Opin Cell Biol, 1997. 9(3): p. 383-7. Moretti, P., et al., Evidence that a complex of SIR proteins interacts with the silencer and telomere-binding protein RAP]. Genes Dev, 1994. 8(19): p. 2257-69. Pina, B., et al., The different (.s‘222‘1face.s of'Rap ] p. Mol Genet Genomics, 2003. 268(6): p. 791-8. Strahl-Bolsinger, S., et al., SIR] and SlR4 interactions differ in core and extended telomeric heterochromatin in yeast. Genes Dev, 1997. 11(1): p. 83-93. Wellinger, R.J., A.J. Wolf, and V.A. Zakian, Origin activation amlfimnatimi of single-strand T G ] -3 tails occur sequentially in late S phase on a yeast linear plasmid. Mol Cell Biol, 1993. 13(7): p. 4057-65. Lendvay, T.S., et al., Senescence mutants of'Saccharomy'ces cerevisiae with a defect in telomere replication identify three additional EST genes. Genetics, 1996. 144(4): p. 1399-412. Hughes, T.R., et al., The role ofthe ES T genes in yeast telomere replication. C iba Found Symp, 1997. 211: p. 41-7; discussion 47-52, 71-5. Bianchi, A., S. Negrini, and D. Shore, Delivetjv of‘yeast telomerase to a DNA break depends on the recruitmentfunctions odecl3 and Est]. Mol Cell, 2004. 16(1): p. 139-46. Evans, SK. and V. Lundblad, Est] and C ch3 as comediators of telomerase access. Science, 1999. 286(5437): p. 117-20. Lingner, J., et al., Three Ever Shorter T elomere (ES T ) genes are dispensablefor in vitro yeast telomerase activity. Proc Natl Acad Sci U S A, 1997. 94(21): p. 1 1 190-5. Singer, MS. and DE. Gottschling, TLC]: template RNA component of Saccharomyces cerevisiae telomerase. Science, 1994. 266( 5184): p. 404-9. 56 133. 135. 136. 137. 139. 140. 141. 142. 143. Zappulla, DC. and TR. Cech, Yeast telomerase RNA: aflexible scaffold/0r protein subunits. Proc Natl Acad Sci U S A, 2004. 101(27): p. 10024-9. Lin, J., et al., A zmiversal telomerase RNA core structure includes structured motifs required for binding the telomerase reverse transcriptase protein. Proc Natl Acad Sci U S A, 2004. 101(41): p. 14713-8. C happell, AS. and V. Lundblad, Structural elements required/0r association of the Saccharomyces cerevisiae telomerase RNA with the Est2 reverse transcriptase. Mol Cell Biol, 2004. 24(17): p. 7720-36. Livengood, A.J., A.J. Zaug, and TR. Cech, Essential regions of Saccharomyces cerevisiae telomerase RNA: separate elements/or Estl p and Est2p interaction. Mol Cell Biol, 2002. 22(7): p. 2366-74. Lingner, J., et al., Reverse transcriptase motifs in the catalytic subunit of telomerase. Science, 1997. 276(5312): p. 561-7. Counter, C.M., et al., The catalytic subunit ofyeast telomerase. Proc Natl Acad Sci U S A. 1997. 94(17): p. 9202-7. Virta-Pearlman, V., D.K. Morris, and V. Lundblad, Est] has the properties ofa single-stranded telomere end-binding protein. Genes Dev, 1996. 10(24): p. 3094- 104. Steiner, B.R., K. Hidaka, and B. Futcher, Association of the Est] protein with telomerase activity in yeast. Proc Natl Acad Sci U S A, 1996. 93(7): p. 2817-21. Zhou, J ., K. Hidaka, and B. F utcher, The Est] subunit ofyeast telomerase binds the Tlcl telomerase RNA. Mol Cell Biol. 2000. 20(6): p. 1947-55. Taggart, A.K., S.C. Teng, and V.A. Zakian, Est] p as a cell cycle-regulated activator of telomere-bound telomerase. Science, 2002. 297(5583): p. 1023-6. Marcand, S., et al., Cell cycle restriction oftelomere elongation. Curr Biol, 2000. 10(8): p. 487-90. 57 144. 145. I46. 147. I48. 149. 151. 154. Hughes, T.R., et al., The Est3 protein is a subunit of yeast telomerase. Curr Biol, 2000. 10(13): p. 809-12. Friedman, K.L., et al., N—terminal domain ofyeast telomerase reverse transcriptase: recruitment of Est3p to the telomerase complex. Mol Biol Cell, 2003. 14(1): p. 1-13. Lundblad, V., T elomere maintenance without telomerase. Oncogene, 2002. 21(4): p. 522-31. Le, S., et al., RAD50 and RA D5 ] define two pathways that collaborate to maintain telomeres in the absence oftelomerase. Genetics, 1999. 152(1): p. 143- 52. McEachem, M .J . and S. Iyer, Short telt‘mteres in ,i-‘east are highly recombinogenic. Mol Cell, 2001. 7(4): p. 695-704. Signon, L., et al., Genetic requirements for RAD5 ]- and RAD54-independent break-induced replication repair of a chromosomal double-strand break. Mol Cell Biol, 2001. 21(6): p. 2048-56. Ira, G. and J .E. Haber, Characterization 0fRAD5l—indq2endent break-induced replication that acts preferentially with short homologous sequences. Mol Cell Biol, 2002. 22(18): p. 6384-92. Malkova, A., E.L. Ivanov, and J.E. Haber, Double-strand break repair in the absence of RA D5 ] in yeast: 0 possible rolefor break-induced DNA replication. Proc Natl Acad Sci U S A, 1996. 93(14): p. 7131-6. Davis, AP. and LS. Symington, RAD5l-dt72endent break-imluced replication in yeast. Mol Cell Biol, 2004. 24(6): p. 2344-51. Kraus, E., W.Y. Leung, and J .E. Haber, Break-induced replication: 0 review and an example in budding yeast. Proc Natl Acad Sci U S A, 2001. 98(15): p. 8255- 62. C hen, W. and S. Jinks-Robertson, The role ofthe mismatch repair machinery in regulating mitotic and meiotic recombination between diverged sequences in yeast. Genetics, 1999. 151(4): p. 1299-313. 58 155. 156. 158. 159. 160. 161. 162. 163. 164. 165. Holmes, V.F., et al., Bypass of heterology during strand transfer by Saccharomyces cerevisiae Rad5] protein. Nucleic Acids Res, 2001. 29(24): p. 5052-7. Louis. E]. and J.E. Haber, The structure and evolution of'subtelomeric Y ' repeats in Saccharomyces cerevisiae. Genetics, 1992. 131(3): p. 559-74. Louis, E]. and J.E. Haber, Mitotic recombination among subtelomeric Y ' repeats in Saccharomyces cerevisiae. Genetics, 1990. 124(3): p. 547-59. Munoz-Jordan, J .L., et al., t-loops at I]:1’p(]11().8‘()111c’ telomeres. Embo J, 2001. 20(3): p. 579-88. Griffith, J .D., et al., Mammalian telomeres end in a large duplex loop. Cell, 1999. 97(4): p. 503-14. Li, B. and A.J. Lustig, A novel mechanismfor telomere size control in Saccharonn'ces cerevisiae. Genes Dev, 1996. 10(11): p. 1310-26. de Bruin, D., et al., T elomere looping permits gene activation by a downstream UAS in yeast. Nature, 2001. 409(6816): p. 109-13. de Bruin, D., et al., T elomerefolding is required for the stable maintenance of telomere position effects in yeast. Mol Cell Biol, 2000. 20(21): p. 7991-8000. Gilson, E., et al., RAP] stimulates single- to drntble-strand association ofyeast telomeric DNA .' implications/or telomere-telomere interactions. Nucleic Acids Res, 1994. 22(24): p. 5310-20. Bryan, T.M., et al., Evidence/or an alternative mechanism/or maintaining telomere length in human tumors and tumor—derived cell lines. Nat Med, 1997. 3(11): p. 1271-4. Henson, J.D., et al., Alternative lengthening oftelomeres in mammalian cells. Oncogene, 2002. 21(4): p. 598-610. 59 166. 167. 168. 169. 170. 171. 172. Reddel, R.R., T.M. Bryan, and JP. Mumane, Immortalized cells with no detectable telomerase activity. A review. Biochemistry (Mose), 1997. 62(1 1): p. 1254-62. Dunham, M.A., et al., Telomere maintenance by recrmzbination in human cells. Nat Genet, 2000. 26(4): p. 447-50. Yeager, T.R., et al., T elomerase-negative immortalized human cells contain a novel type ofpromyelocytic leukemia (PML) body. Cancer Res, 1999. 59(17): p. 4175-9. Saretzki, G., et al., Ribozyme-mediated telomerase inhibition induces immediate cell loss but not telomere shortening in ovarian cancer cells. Cancer Gene Ther, 2001. 8(10): p. 827-34. Zhang, X., et al., T elomere shortening and apoptosis in telomerase-in]:ibited human tumor cells. Genes Dev, 1999. 13(18): p. 2388-99. Hahn, W.C., et al., Inhibition of telomerase limits the growth of'human cancer cells. Nat Med, 1999. 5(10): p. 1164-70. Perrem, K., et al., Repression of an alternative mechanismfor lengthening of telomeres in somatic cell hybrids. Oncogene, 1999. 18(22): p. 3383-90. 61.) CHAPTER 2 STRONG NEGATIVE EFFECTS OF RAD52 OVEREXPRESSION ON DNA DAMAGE REPAIR IN S. CEREVIAISE ABSTACT Rad52 is an essential multifunctional component of the homologous recombination machinery in S. cerevisiae. In this work, I examined the effect of RAD52 overexpression on DNA damage repair. 1 demonstrated that among the five 5’ ATG triplets, the third, fourth and fifth can be used as translation initiation codons. Rad52 protein translated from the fifth ATG is as competent as that translated from the third ATG in DNA damage repair. The 99 bp sequence between the first ATG and the third initiation ATG has a strong influence on the level of Rad52 expression controlled by a heterologous GAL] promoter. When overexpressed, Rad52 has a strong negative effect on DNA damage repair. Overexpression of Rad51 completely suppresses the negative effect of Rad52 overexpression. Overexpression ofa mutant Rad52, which is defective in the Rad51 interaction, has a greatly reduced negative effect on DNA damage repair. These data suggest that the negative effect of Rad52 overexpression results mainly from the sequestration of Rad51 from other essential functions. 61 INTRODUCTION Homologous recombination is an essential pathway for DNA damage repair in eukaryotic cells. In S. cerevisiae, the major components of the homologous recombination machinery are the protein products of the RAD52 epistasis gene group, including RAD52, RAD5], RAD54, RAD55, RAD57, RAD59, RDH54/TID], MREll, RAD50, and XRS2. Rad52, Rad51, Rad54, Rad55, and Rad57 are required for gene conversion and break-induced replication [1-3]. Rad52 and Rad59 are required for additional types of homologous recombination, including RAD5I-independent break- induced replication and single-strand annealing [2, 3]. Consistent with the essential roles of Rad52 in all types of homologous recombination, rad52 null mutants have the most severe defect in mitotic recombination and are most sensitive to DNA damaging agents, such as y-rays and methylmethanesulfonate (MMS) [4, 5]. In vitro studies suggest that Rad52 is required at a level substoichiometric to Rad51 to achieve an optimal recombination activity [7]. Consistent with this observation, Rad52 is of lower cellular abundance than Rad51 [7, 8]. While both genetic and biochemical studies suggest that cells are sensitive to the changes in Rad52 cellular level, especially under DNA damaging conditions [6-8], previous studies found no appreciable effect of Rad52 overexpression on DNA damage repair [9, 10]. However, the exact levels of Rad52 protein in those studies are unknown. There are five in-frame ATG triplets at the 5’ end of the RAD52 genomic sequence. In the previous studies on the effect of Rad52 overexpression, the entire ORF starting from the first ATG was placed downstream of a heterologous promoter (ENOI, GAL], ADH) for the purpose of overexpressing Rad52 [9, 10]. Here 1 demonstrate that the third and fifth, possibly the fourth ATG, but not the first and the second ATG can initiate protein translation in vivo. The sequence between the first ATG and the third ATG has a great negative influence on Rad52 protein expression controlled by the GAL] promoter. Removal of this sequence leads to a 40-fold increase in Rad52 protein level. Overexpression of RAD52 has a strong negative effect on DNA damage repair. This effect is specific for DNA damage repair since cells overexpressing Rad52 show no apparent growth defect in the absence of a DNA damaging agent. Furthermore, overexpression of Rad51 completely suppresses the negative effect of Rad52 overexpression. MATERIALS AND METHODS Strains Yeast strains used in this study are listed in table 2—1. JP166 was generously provided by Dr. John Prescott (University of Califomia, San Francisco). BY4735 (ATCC 200897) was obtained from ATCC. All other strains were derived from these two strains. Disruption of specific genes was carried out as previously described [1 1, 12]. S. cerevisiae strains were propagated at 30°C in dropout media lacking the amino acids required for plasmid selection. Yeast transformation was performed according to Agatep, R. et al. [13]. Plasmids at_td site-directed mutagenesis To create pRS415RAD52, the Sall genomic fragment containing the RAD52 coding sequence and its promoter [14] was cloned into pRS415, a CEN vector with a LEU2 marker [15]. To create mutant rad52 containing a single ATG initiation codon, all 63 other ATGs were mutated to ATC by site-directed mutagenesis in the pRS415RAD52 construct. Site directed mutagenesis was performed following the Quick-Change protocol using Pfu-Turbo polymerase (Stratagene). Vector pRSG415 (generously provided by Dr. John Prescott, University of California, San Francisco) for galactose-inducible overexpression contains the GAL] promoter and the CYC] terminator on a HindIII-Sacl fragment in pRS415 backbone. A BamHI site and an Spel site were previously engineered between the GAL] promoter and the C YC ] terminator for cloning purposes. To create pRSG414 vector, The HindIII-Sacl fragment from pRSG415 was cloned into pRS414, a C EN vector with a TRP] marker [15]. To clone RAD52 into the pRSG415 vector, the single BamHI site within the RAD52 ORF was removed by a silent mutation via site-directed mutagenesis (5’- CGACAGAGAAGGACCCCGTTGTAGJ’). The RAD52 coding sequence starting from the first (GALl-RAD52F1), the third (GALI-RAD52F3) or the fifth ATG (GALI- RAD52F5) to the stop codon was amplified by PCR to introduce a BamHI site at the 5’ and an Spel site at the 3’, and subcloned into pRSG415, placing the coding sequence downstream of the GAL] promoter but upstream of the C YC ] terminator. To create GALl-RAD52F1ATG1, the second, third, fourth and fifth ATG in GALl-RAD52F1 were mutated to ATC by site-directed mutagenesis. To create GALl-RAD52F3ATG3, the fourth and fifth ATG in GALl-RAD52F3 were mutated to ATC by site-directed mutagenesis. For selection with the TRP] marker, the Hindlll-Sacl fragments containing RAD52, the GAL] promoter, and the C YC ] terminator were inserted into the pRSG4l4 vector to generate GAle-RAD52F1, GAle-RAD52F3 and GAle-RAD52F5. 64 For overexpression of Rad51 and Rad59, the RAD5] and RAD59 genes were amplified from genomic DNA and cloned into pRSG415 between the BamHI and the Spel sites, placed downstream of the GAL] promoter but upstream of the CYC] terminator. Western blotting Yeast proteins were prepared following a procedure from the laboratory of Steven Hahn (www.flicrcorg”lab‘hahn). Equal amounts of proteins, determined by Bradford method (Pirece), were separated on 10% SDS-PAGE and blotted onto polyvinylidene difluoride membrane (Amersham Pharmacia Biotech). Rad52 was detected with a goat anti-Rad52 antibody RAD52 yC-17 (Santa Cruz Biotech) followed by a rabbit anti-goat IgG HRP (Sigma). The membrane was stripped and probed with a rabbit anti-G-6-PDH antibody (Sigma) followed by a goat anti-rabbit IgG HRP (Sigma) as a loading control. MMS sensitivity assav The strains harboring galactose-inducible gene constructs were cultivated in appropriate dropout liquid medium containing 2% raffinose as the sole carbon source to mid-log phase. 10-fold serial dilutions were prepared and 10 to 10" cells were spotted on appropriate solid medium containing 2% raffinose and 0.05% galactose with or without MMS at concentrations specified in figure legends. The plates were then incubated at 30°C for 3 or 4 days and photographed. The strains harboring gene constructs controlled by the genomic promoter were evaluated in the same way except that the liquid and solid culture medium contained 2% glucose as the carbon source. 65 RNA dot blotting Total RNA was prepared by extraction with hot acidic phenol following the protocol described in Short Protocols In Molecular Biology, Fourth Edition. Total RNA (4 ug or 0.4 ug) was dot-blotted onto nylon membranes (Amersham Pharmacia Biotech) and UV-crosslinked. Blotted membranes were hybridized with a 32P-radiolabeled RAD52 DNA probe or a TDH4 DNA probe. Hybridization probes were labeled with [oz-32P] dCTP using random primer DNA labeling kit (Invitrogen). Hybridization was performed in 0.5M N3H3P04/N83HPO4 pH7.2, 7% SDS, 10 mM EDTA, at 60°C, according to Church and Gilbert [16]. The membrane was exposed to phosphorescent screens and the images were scanned with Phosphorimager (Molecular Dynamics). The signals were quantified using ImageQuant software (Molecular Dynamics). BBQ. cDNA was synthesized using Superscript II reverse transcriptase (Invitrogen). Total RNA (5 ug) was mixed with 12.5 pmole of a RAD52 specific primer (5’- TTTTCACCAGGTTCTTCGTCG-3’) and 20 nmole of dNTPs. The mixture was incubated at 65°C for 5 min and quickly chilled on ice. Following addition of First Strand buffer (Invitrogen) and DTT (Invitrogen). the mixture was incubated at 42°C for 2 min. Reverse transcriptase was added to half of the mixture, lul of DEPC-treated H30 was added to the remaining half as the control. cDNA was synthesized at 42°C for 50 min. The cDNA was then amplified using a pair of RAD52 specific primers (5’- GAGAAGAAGCCCGTTTTC-3’ and 5’-CGGGTATTGTTGTTGTTC-3”) and Taq polymerase (Promega). 66 Primer extension An oligonucleotide probe (5’-TTACTCTCCAACCTTCG-3’) was labeled with ['y-izP] dATP using T4 polynucleotide kinase (New England Biolabs). Total RNA (12 ug) was hybridized to the radiolabeled probe in 150 mM KCI, 10 mM Tris-HCI pH 8.3, 1 mM EDTA at 65°C for 90 min. Extension reaction was carried out in reaction buffer containing 20 mM Tris-HCI pH 8.3, 10 mM MgCl:, 5 mM DTT, 0.15 mg/ml actinomycin D, 0.15 mM dNTPs. 5 U of AMV reverse transcriptase (Invitrogen). The reaction was performed at 42°C for 60 min. Reaction mix was then digested with RNase and extracted with phenol/chloroform and precipitated with ethanol. The reaction products were resolved in a 6% PAGE/7 M Urea gel. The gel was dried on a vacuum gel drier at 80°C for 1 hour and exposed to phosphorescent screens and scanned with Phosphorimager (Molecular Dynamics). RESULTS Determining the translation initiation site(s) of the Rad52 protein There are five in-frame ATG triplets at the 5’ end of the RAD52 genomic sequence. Recent studies suggest that the third ATG (99 bp downstream from the first ATG) is likely to be the translation initiation codon [5, 14]. I examined the possible translation initiation sites using a more systematic approach. 1 mutated four of the five ATG triplets to ATC in different combinations, keeping a single ATG for translation initiation. These RAD52 variants were named RAD52ATG], 2, 3. 5, indicating the presence of the corresponding ATG (Figure 2-1A). Low copy plasmids carrying these variants of RAD52 controlled by the endogenous promoter were tested for their ability to 67 complement a raa’52A strain using sensitivity to the DNA damaging agent MMS as an assay (Figure 2-lB). Compared to the wild type strain (JP166) expressing the endogenous RAD52, the rad52A strain, JP166L1, is at least 10,000 times more sensitive. Neither RAD52ATG] nor RAD52ATG2 could complement rad52/J. In contrast, the plasmid carrying wild type RAD52, RAD52ATG3, or RAD52ATG5 fully complemented the sensitive phenotype of rat/52A. To confimt that the Rad52 protein initiated from the fifth ATG is functional, we tested rad52d]03-]20, a rat/52 mutant lacking the six amino acids downstream of the third ATG to the fifth ATG (Figure 2-1A). This deletion mutant indeed complemented the MMS sensitive phenotype of rad52/1 cells (Figure 2-1C). I also examined the level of Rad52 protein expressed by these variants. Consistent with the results of the complementation test, Rad52 protein in RAD52A TG3 and RAD52A TG5 containing cells was expressed at a level similar to that of endogenous Rad52 in the wild type strain JP166, while no Rad52 protein was detected in RAD52A TG] and RAD52A TG2 expressing cells (Figure 2-1D). To examine whether the ATG to ATC mutations in RAD52A TG] or RAD52/1 TG2 impaired transcription of RAD52, RT-PCR was performed. RAD52 transcripts existed in both RAD52ATG] and RAD52ATG2 expressing cells (Figure 2-lE). Next I measured the RAD52 mRNA level. RAD52 mRNA was transcribed at similar levels from all the variants tested as shown by RNA dot blotting analysis (Figure 2-1F), indicating that the absence of Rad52 in RAD52A TG] and RAD52A TG2 expressing cells is likely caused by a posttranscriptional defect. Taken together. these results indicate that efficient complementation ofrad52A can be achieved by exogenous expression of the RAD52 gene expressed under the control of the endogenous promoter. However, it appears that only the third or fifth, and possibly the 68 fourth ATG, can be used as translation initiation sites in vivo. The first and the second ATG do not serve as translation initiation sites. The failure of RAD52ATG] and RAD52A T G2 to complement the MMS sensitivity of rad52A is due to a lack of Rad52 protein expression as a result of posttranscriptional defects. Overexpression of RAD52 has a strong negative effect on DNA damage repair induced by MMS Both in vivo and in vitro data suggest that cells are sensitive to changes in the level of Rad52, especially under DNA damaging conditions [6-8]. To directly examine the effect of RAD52 overexpression on DNA damage repair, I cloned the coding sequence downstream of the galactose-inducible GAL] promoter, which results in high level of expression. Since Rad52 proteins initiated from the third and fifth ATG were both found to be functional, 1 engineered constructs starting from the third or the fifth ATG into the pRSG415 vector, designated as GALI-RAD52F3 and GALl-RAD52F5, respectively (Figure 2-2A). Previous studies on the effect of Rad52 overexpression utilized constructs containing RAD52 starting from the first ATG [9, 10]. Thus, I also tested a similar construct, GALl-RAD52F1 (Figure 2-2A), as a control and to compare my results with the previous studies. These three constructs were tested for their ability to complement the MMS sensitive phenotype of rad52A by spot assays. Surprisingly, RAD52 expressed from the GAL] promoter induced phenotypes different from those of genes expressed from the endogenous promoter. Expression of GALl-RAD52F3 or GALl-RAD52F5 did not complement the MMS sensitivity of rad5221, while that of GALl-RAD52F1 restored the MMS resistance ofrad52A cells to a level similar to that of the wild type strain JP166 expressing the chromosomal copy of RAD52 (Figure 2-28). 69 Since the GALl-RAD52F3 construct contains three candidates for translation initiation codon, ATG 3, 4 and 5, I mutated the forth and fifth ATG to ATC, and tested this new construct named GALI-RAD52F3ATG3 (Figure 2-2A). Similar to the result with GALI- RAD52F3, this construct did not complement the MMS sensitive phenotype of rad52/I (Figure 2-2C). In the absence of MMS, cells expressing these constructs showed no apparent defect in growth compared to wild type cells (Figure 2-2B). The observed phenotypic differences between different RAD52 constructs prompted me to investigate the expression level of RAD52. The RAD52 RNA level was increased to similar high levels (37-89 fold) in GALl-RAD52F1, GALl-RAD52F3 or GALl-RAD52F5 expressing cells compared to that of wild type cells expressing the genomic RADSZ (Figure 2-3A). The level of Rad52 protein expressed by GALI- RAD52F3 or GALl-RAD52F5 was also increased significantly compared to that of the wild type strain (Figure 2-3B). In contrast, Rad52 protein in GALl-RAD52F1 expressing cells was increased less than 5 fold (Figure 2-3B). A quantitative comparison revealed a 4()—fold increase in Rad52 level expressed from GALl-RAD52F3 compared to that expressed from GALl-RAD52F1 (Figure 2-3C). To investigate the possible cause for the relatively low level of Rad52 protein expressed from GALl-RAD52F1, a primer extension experiment was performed to examine the RAD52 transcripts. Most of the overexpressed transcripts were initiated upstream of the first ATG (Figure 2-3D). Only a minor species was transcribed from the natural initiation site, i.e. downstream of the second but upstream ofthe third ATG, for the endogenous RAD52 transcripts in the wild type cells (Figure 2-3D). This result suggests that the relatively low level of Rad52 protein expression by GALl-RAD52F1 is not due to defects in transcription initiation or 70 decreased transcript stability. It appears that the additional sequences at the 5’ of RAD52 mRNA transcribed from GALl-RAD52F1 are deleterious for protein accumulation. The existence of the longer RAD52 transcripts in GALl-RAD52F1 suggests possible utilization of the first ATG for translation initiation. I constructed GALI- RAD52F1ATG1 by mutating all but the first ATG to ATC in GALl-RAD52F1 (Figure 2-2A). Unexpectedly, GALI-RADSZFIATGI failed to complement rad52A (Figure 2- 3E), and no Rad52 protein could be detected (Figure 2-3F). These results suggest that the first ATG can not initiate translation, and Rad52 protein expressed by GALl-RAD52F1 is translated from a downstream ATG. In summary, removal of the sequence between the first and the third ATG greatly increases Rad52 protein expression level controlled by the GAL] promoter. A high level of Rad52 expression has a strong negative effect on DNA damage repair induced by MMS. Overexpression of RADSI suppresses the negative effect of RAD52 overexpression Since Rad52 physically interacts with other components of the homologous recombination repair pathway, including Rad51, Rad59 and RPA, the negative effect of Rad52 overexpression may be due to the sequestration of these interaction partners. To test this hypothesis, I co-overexpressed Rad52 with Rad51 or Rad59 to examine whether the negative effect of Rad52 overexpression could be at least partially rescued (Figure 2- 4A). The fragments containing the GAL] promoter, RAD52 and the C YC ] terminator from GALl-RAD52FI, GALl-RAD52F3 and GALl-RAD52F5 were subcloned into pRS414 to create GAle-RAD52F1, GAle-RAD52F3 and GAle-RAD52F5, respectively. Overexpression of Rad51 or Rad59 by themselves could not rescue the 71 MMS sensitive phenotype ofrad5221 cells, indicating the essential role of Rad52 in DNA damage repair. Overexpression of Rad51 or Rad59 showed no effect in GAle- RAD52F1 expressing cells, suggesting that Rad51 or Rad59 overexpression has no negative effect on DNA damage repair. Overexpression of Rad51 almost completely rescued the MMS sensitive phenotype of cells expressing GAle-RAD52F3 or GAle- RAD52F5. In contrast, overexpression of Rad59 had no effect. This is not due to a lack of Rad59 expression or malfunction, since the same Rad59 construct can complement the MMS sensitive phenotype ofa rat/59A strain (Figure 2-4B). I also confirmed high level of Rad52 expression by GAle-RAD52F3 and GAle-RAD52F5 in Rad51 overexpressing cells (Figure 2-4C). Thus, the above results suggest that the negative effect of Rad52 overexpression is likely due to the sequestration of Rad5 1. Overexpressiog of a rad52 mutant with a specific defect in interaction with Rad_1_ has a greatly reduced defect in DNA damage repair. The ability of overexpressed Rad51 to rescue the defect in DNA damage repair as a result of Rad52 overexpression suggests that the Rad51-dependent pathway of homologous recombination is affected in GALl-RAD52F3 or GALl-RAD52F5 expressing cells. I tested this possibility in a more direct approach by examining the effect of overexpressing rad52d409-4l2, a rad5] mutant that has a specific defect in interaction with Rad51 [17]. In contrast to the highly sensitive phenotype of GALl- RAD52F3, rad5221 cells expressing GALl-RAD52F3d409-412 showed only a slightly increased MMS sensitivity (figure 2-5A). Cells containing GALl-RAD52F3 or GALl- RAD52F3d409-412 express similar level of Rad52 (figureZ-SB). Therefore, overexpression of a Rad52 mutant protein with a specific defect in Rad52-Rad51 interaction has a greatly reduced negative effect on DNA damage repair. RAD52 overexpression affects a RADSI-indepengent DNA damage remair pathway(s) I also examined whether other DNA repair pathways are affected by Rad52 overexpression. If a Rad51-dependent pathway is the major repair pathway for the damages induced by MMS, the effects on other pathways might not be readily detected in the presence of Rad51. Therefore, I examined GALl-RAD52F3 and GALl-RAD52F5 in a rad5121 strain, JP166W1. As shown in Figure 2-6A, rat/51A cells were highly sensitive to 0.005% MMS treatment (the same concentration used in earlier experiments). GALI- RAD52F3 and GALl-RAD52F5 did not significantly alter the MMS sensitivity of rad51A cells. Therefore, we reduced the MMS concentration to 0.001%, under which condition the sensitivity of the rad51A cells was significantly reduced (figure 2-6B), whereas rad52A cells were still highly sensitive (data not shown). Under this condition, rad5lA cells expressing GALl-RAD52F3 or GALI-RAD52F5 did show a significantly higher MMS sensitivity than cells containing the control vector. Thus, these results suggest that Rad52 overexpression has a negative effect on a RAD51-independent repair pathway(s). Since recombination repair pathways in yeast consist of the Rad51- and the Rad59-dependent pathways, we examined whether the negative effect observed in ram/5121 cells was due to an impaired Rad59-dependent pathway. Rad52 and Rad59 were co-overexpressed in a rad52zf rad5 12] strain BY20031 (Figure 2-6D). Overexpression of Rad59 alone did not alter cells’ ability to repair DNA damages induced by MMS. Overexpression of Rad52 alone from GAle-RAD52F3 or GAle-RAD52F5 resulted in 73 a MMS sensitive phenotype similar to that shown in Figure 2—6B. Co-overexpression of Rad52 with Rad59 did not restore the MMS resistance (Figure 2-6D). Thus, the additional negative effect of Rad52 overexpression observed in rad5 [A cells is not likely caused by a defective Rad59-dependent pathway. It is possible that other DNA damage repair pathways are affected when Rad52 is overexpressed. DISCUSSION In this study, 1 demonstrated that overexpression of Rad52 has a specific negative effect on DNA damage repair induced in response to MMS. First, cell growth under DNA damaging conditions is negatively affected by Rad52 overexpression. Second, this negative effect can be suppressed by Rad51 overexpression, but not Rad59 overexpression. This study provides in vivo evidence that the ratio between Rad52 and Rad51 is critical in homologous recombination mediated DNA damage repair. In addition, our results also suggest that Rad52 overexpression affects a RAD5]- independent DNA damage repair pathway(s). I also tested possible translation initiation sites of Rad52. Among the five in-frame ATGs at the 5’ end of the RAD52 sequence, the third, fourth and fifth ATG can be used as a translation start codon. I also demonstrated that Rad52 translated from the fifth ATG is as competent as that translated from the third ATG in DNA damage repair induced by MMS. Translation initiation site(s) of Rad_5_2 The existence of five in-frame ATG triplets at the 5’ ofthe RAD52 sequence has caused confusion as to which of these serves as in vivo translation initiation site(s). Previous studies have suggested that the first and the second ATG are not used to initiate 74 translation in vivo [5, 14]. However, which of the remaining three ATGS is in fact used for translation initiation is unclear. By mutating four of the five ATGS in different combinations and keeping only one intact, I examined Rad52 translated from a certain ATG under the control of its own promoter. When only the third or the fifth ATG is available to initiate translation, Rad52 protein can be expressed at a level similar to the endogenous protein in wild type cells (Figure 2-1D). More importantly, Rad52 translated from the third and the fifth ATG have same ability to fully complement the MMS sensitivity of rat/52A (Figure 2-lB). However, when only the first or the second ATG is intact, no Rad52 protein can be detected (Figure 2-1D). This is consistent with the result of an SI nuclease protection analysis, which reveals that the 5’ end of a major RAD52 transcript is located between the second and the third ATG [14]. A number of reports studied Rad52 by placing the coding sequence from the first ATG under the control of a heterologous promoter. It is unclear whether the RAD52 mRNA start site was altered under those conditions, hence producing Rad52 with additional 5’ sequences. When the entire RAD52 coding sequence is placed under the control of the GAL] promoter, the majority of RAD52 mRNA is transcribed from upstream ofthe first ATG (Figure 2-3D). However, the first ATG is not likely to initiate translation as demonstrated by the lack of Rad52 expression from GALI- RAD52F1ATG1 (Figure 2-3F), which only has the first ATG available for translation initiation. Therefore, although RAD52 mRNA can be transcribed from upstream of the first ATG, Rad52 protein translation is still initiated from a downstream ATG. Furthermore, it appears that the additional sequences at the 5’ end of RAD52 mRNA are deleterious for protein accumulation. Although expression from GALl-RAD52F1 results 75 in a ~90-fold increase in RAD52 mRNA compared to the endogenous level (Figure2-3A), there is a less than 5—fold increase in Rad52 protein level (Figure 2—3B). In contrast, removal of the sequence between the first and the third ATG leads to at least a 40-fold increase in Rad52 protein level (Figure 2-3C). It is likely that the addition sequences transcribed from GALl-RAD52F1 block the protein translation machinery. Effects of RAD52 overexpression on DNA damage repair Rad52 is the central player in homologous recombination. It is involved in multiple direct protein-protein interactions, including self-association and interactions with RPA, Rad51, and Rad59 [18-21]. Changes in cellular concentration of components involved in hetero-multimeric complexes will result in imbalance among the components. Such imbalance often leads to distinct phenotypes [22]. Consistent with this idea, Rad52 should be kept at a level ~l/10 of the amount of Rad51 to achieve its maximal mediator function in strand exchange reactions [7]. In vivo, Rad52 protein level is tightly controlled by both transcriptional and posttranslational regulation [6]. When overproduced, Rad52 exerts a negative effect on DNA damage repair (Figure 2-ZB). In fact, a similar negative effect has been observed for Rad54 overexpression [23]. The negative effect of Rad52 overexpression results mainly from sequestration of Rad51, since overexpression of Rad51 can completely suppress such effect (Figure 2-4A). Indeed, overexpression of rad52d409-412, which is defective in interaction with Rad51, displays a greatly reduced effect on DNA damage repair (Figure 2-5A). It is possible that when overexpressed, rad52a’409-4l2 may have residual Rad51-binding activity which enables it to mediate Rad51 function. This could account for its relative resistance to MMS treatment. The negative effect of overexpression of Rad52 observed in our study is not caused by malfunction of the Rad52 protein. Overexpression of Rad51 can not rescue the MMS sensitive phenotype of rad52A cells (Figure 2-4A). Only when Rad52 and Rad51 are overexrepssed simultaneously, is MMS sensitivity of rat/5221 rescued (Figure 2-4A), indicating that Rad52 expressed from GALl-RAD52F3 or GALl-RAD52F5 is functional. The strong negative effect of RAD52 overexpression demonstrated by our results does not agree with the previous observation that overexpression of RAD52 has no appreciable effect on DNA damage repair [10]. In the previous study, the RAD52 overexpression construct was made by placing RADSZ gene from the first ATG under the control of the END] promoter. Although the authors showed that the RNA level of RAD52 was lO-fold higher than that of wild type cells, they did not examine the steady state protein level. In fact, the sequence between the first ATG and the third ATG seems to block protein translation (discussed above). In addition, this sequence appears to contain a competent promoter. GALl-RAD52F1 can complement the MMS sensitivity of rat/52A in non-inducing medium (Figure 2-7). pRS415RADSZB, a RAD52 construct with neither a heterologous promoter nor the genomic sequence upstream of the first ATG, can also complement the MMS sensitivity of rad52A cells (Figure 2-8). Therefore, it is possible that the expression of Rad52 by GALl-RAD52F1 is driven by a potential control element within the sequence between the first and the third ATG. Indeed, the primer extension study revealed a RAD52 mRNA species with a 5’ end identical to that of the endogenous RAD52 transcript (Figure 2-3D). It is highly likely that Rad52 77 translated from this species accounts for the MMS resistance of the rad52A cells expressing GAL l -RAD52F 1. Significance of the ratio between Rad52 and Rad51 in DNA damage repair in viva Recent studies on Rad51 and its partner protein Brca2 suggest that the site in Rad51 involved in Rad52 binding is the same site for Rad51-Rad51 interaction. Human Rad51 contains a conserved motif (85-G_ETT_ATE-91, a comparable region of ScRad51 has a sequence of 143-GEV_T_AAD-l49) which serves as the interface for Rad51 oligomerization [24, 25]. Mutation of this sequence prevents DNA damage-induced Rad51 foci formation [24]. This motif was initially recognized in the BRC repeats of Brca2, a Rad51 interacting protein. X-ray structure study and mutational analysis have shown that such a motif in Brca2 is indeed responsible for Rad51 binding [24, 25]. Interestingly, the C-terminus of ScRad52 which interacts with Rad51 contains a similar sequence 315-TEVEKA-32l [25]. Indeed, a rad52 mutant (rad52K353E), which is defective in Rad51 interaction, has a single amino acid substitution in this motif[26]. The sharing of a single binding site of Rad51 for two different protein interactions would explain the inhibition of Rad51-mediated DNA damage repair by excessive Rad52 and provide a plausible mechanism for Rad51 action. The rate limiting step of Rad51 nucleoprotein filament assembly is the nucleation step, after which the filament elongates rapidly [27]. The Rad52-Rad51 interaction, as well as BrcaZ-RadSl interaction, recruits Rad51 to DNA damage sites [25, 28]. Rad52 then facilitates initial RPA displacement by Rad51 [27]. After nucleation, the free form of Rad51 can interact with Rad51 and with DNA to form Rad51-DNA nucleoprotein filament. However, excess amount of Rad51 interacting protein would exclude Rad51 from oligomerization during filament formation. 78 Consistent with this model, Rad51 is of higher cellular abundance than Rad52. Since they form stable stoichiometric complexes as demonstrated in co-immunoprecipitation experiments, most of the cellular Rad51 is free from interaction with Rad52 [7, 8]. Effects of RADS2 overexpression on RADSI-indepengent DNA damage repair Qathways Rad51-dependent and Rad59-dependent pathways are the major pathways in homologous recombination repair in S. cerevisiae (reviewed in [2]. rad5lA rad59A double mutants display similar phenotypes as rad52A mutants [2, 29]. However, the additional DNA damage repair defect observed in rad51A strain from Rad52 overexpression is not likely due to an impaired Rad59-dependent pathway, since overexpression of RAD59 can not rescue the defect (Figure 2-6D). Nucleotide excision repair and base pair excision repair can repair DNA damages induced by MMS [30]. It is possible that the additional negative effect is mediated by nonspecific association of overexpressed Rad52 with protein components involved in other DNA damage repair pathways. Alternatively, DNA lesions caused by MMS are channeled to repair pathways which normally do not function in repairing these damages when Rad51-dependent pathway is functional. Therefore, this effect may only be observed in the absence of Rad51. SUMMARY In summary, the present study has demonstrated that overexpression of Rad52 has a strong negative effect on DNA damage repair induced by MMS. This effect is caused mainly by sequestration of Rad51 by excess amount of Rad52. Overexpression of Rad52 79 also has a negative effect on a Rad51-independent DNA damage repair pathway(s). This effect appears to be nonspecific. In addition, the experiments reveal that the third, fifth, and possibly the fourth ATG at the 5’ end of RAD52 can be used as a translation initiation codon in vivo. The sequences between the first and the third ATG appears to contain a promoter, and these sequences have a strong negative effect on Rad52 protein expression controlled by a heterologous GAL] promoter. 80 APPENDIX 2: FIGURES AND TABLES FOR CHAPTER 2 81 Table 2-1. S. cerevisiae strains used in this study strain Genotype JP166” MATa his321 leuZA ura3A tic/A pRS316TLC] JP166L1° MA Ta his321 leu221 ura321 rad52AssHlS3 tlclA pRS316TLCl JP166W17’ MATa his321 leu2A ura321 rad5/21::HIS3 tlclA pRS3l6TLC] JP166W2” MA Ta his3A [£21222] W34 rad59AssHlS3 22cm pRS316TLC] BY4735“ MATa ade221:.'hisG his321200 leu2AO met/5210 trplA63 Ura3A0 MA Ta ade2AsshisG his3/1200 leu2AO met/5210 trp1A63 Ura3A0 rad522lssHlS3 BY4735L1 " MA Ta adeZA.'.'hisG his32120l) 1eu2210 met] 5210 trplA63 Ura3A (1 raa’52A.°:HIS3 rad5]21::KANMX6 tlclA:.'Met]5 pRS316TLC] BY20031 " " From Dr. John Prescott (University of California, San Francisco) b Derivative ofJPl66; this study " Obtained from ATCC " Derivative of BY4735; this study Table 2-2. Constructs used in this study Plasmid Description pRS415 A CEN shuttle vector with a LEU2 marker pRS414 A CEN shuttle vector with a TRP] marker pRSG415 GAL] promoter and C YC ] terminator are inserted into pRS415 pRSG414 GAL] promoter and C YC ] terminator are inserted into pRS4 l4 pRSG415RAD52F1 RAD52 gene from the first ATG to stop codon is cloned into (GALl-RAD52F1) pRSG415, placed under the control ofthe GAL] promoter pRSG415RAD52 The second, third, fourth and fifth ATG were mutated to ATC FlATGl (GALl- in GALl-RAD52F1 RAD52F1ATG1) pRSG415RAD52F 3 RAD52 gene from the third ATG to stop codon is cloned into (GALl-RAD52F 3) pRSG415, placed under the control of the GAL] promoter pRSG415RAD52F3M The fourth and the fifth ATG in GALI-RAD52F3 are mutated 38-401 (GALl- to ATC RAD52F3ATG3) pRSG415RAD52F5 RAD52 gene from the fifth ATG to stop codon is cloned into (GALl-RAD52F5) pRSG415, placed under the control of the GAL] promoter pRSG414RAD52 Fl (GAle-RADSZFI) RAD52 gene from the first ATG to stop codon is cloned into pRSG414, placed under the control of the GAL] promoter pRSG414RAD52F 3 RAD52 gene from the third ATG to stop codon is cloned into (GAle-RAD52F3) pRSG414, placed under the control ofthe GAL] promoter pRSG414RAD52F5 RAD52 gene from the fifth ATG to stop codon is cloned into (GAle-RAD52F5) pRSG414, placed under the control of the GAL] promoter 83 Table 2-2. Constructs used in this study (Continued) pRSG415RAD51 RAD5] gene is placed under the control of GAL] promoter (GALl-RADS l) pRSG415RAD59 RAD59 gene is placed under the control of GAL] promoter (GALl-RAD59) pRS415RAD528 RAD52 gene from the first ATG to stop codon followed by C YC ] terminator is cloned into pRS415 pRS415RAD52 The Sall fragment containing genomic copy ofRAD52 gene (RAD52) (from ~lkb upstream to ~08 kb downstream) { Adzuma, 1984 #2} is cloned into pRS415 pRS415RAD52d103- The 6 amino acids from downstream ofthe third ATG to the 120 fifth ATG are deleted from pRS415RAD52 (RAD52le3-l20) pRS415RAD52ATGl pRS415RAD52 with only the first ATG kept intact, all others (RAD52ATG1) are mutated to ATC pRS415RAD52ATG3 pRS415RAD52 with only the third ATG kept intact, all others (RAD52ATG3) are mutated to ATC pRS415RAD52ATG5 pRS415RAD52 with only the fifth ATG kept intact, all others (RAD52ATG5) are mutated to ATC 84 Figure 2-1 Analysis ofthe MMS sensitivity of rad52A cells expressing RAD52 variants containing a single translation initiation site and controlled by RAD52 genomic promoter A. Schematic representation of RAD52 variants Open circles represent ATG triplets. Closed circles represent ATC triplets. The Sall genomic fragment containing RAD52 coding sequence and its own promoter was cloned into the pRS415 vector. Four out of the five 5’-terminal in-frame ATG triplets were mutated to ATC to leave a single ATG for translation initiation. The variants were named RAD52/1 TG]. 2, 3 or 5, corresponding to the individual ATG triplets. rad52d]03- 120 was constructed by deleting the six amino acids downstream of the third ATG to the fifth ATG. The deletion is indicated as a dashed line. The arrows indicate the transcription start sites based on previous reports [5, 14]. B. MMS sensitivity of rad52zf cells expressing RAD52, RAD52ATG], RAD52ATG2, RAD52A TG3, 0r RA D52A TG5 The wild type strain JP166 expressing genomic RAD52 and the rad52A strain JP166L1 expressing RAD52 variants or the control vector were cultivated in SC-Ura-Leu medium containing 2% glucose to mid-log phase. 10-fold serial dilutions containing 10° to 10 cells were spotted on SC-Ura-Leu solid medium containing 2% glucose with or without 0.005% MMS. _-~ 2:“: .\ A an; I". ‘ .. W J p.35: 3:... .530.) Be Show: Gs 53$ M m Ga 535‘ m w w n Gs 5.3.x: r. / . Raw: x... . r .\ m2: :2 m2: 3mg... m w,” c c_c c RTQENQE C I I I I wUk—xmmavfi I I C 1 I Mbkvdudvfi I I I C 1 NDkVNWQ—xz 95.40:. 0 a c c_u c Gs aware UP223— . _ i E5 - luv .11. o o o O b b0 94 r «I «I b« a V hm, Am; a»; a, 299/ so 0 0.» A». a s 0 iv .v G r p xv re AQNMEEV Sew—3:. lOO B. Comparison of the steady state protein level of Rad52 in I'(I(/52A cells expressing GALl-RADSZFI. GALl-RAD52F3 and GALl-RADSZFS Proteins were extracted from cells cultivated in medium containing 2% raffinose and 0.05% galactose. JP166 was cultivated in media containing 2% glucose. The Western blot was probed with an anti-Rad52 antibody. after which the membrane was stripped and probed with an anti-G-6-PDH antibody. l()l J Pl 66L! (rad52A ) 2min 5‘ «73 t 0 5&7? 5' Q, o 8 80%? ~17 b» O E I G .3? soky ‘ s, i g f ) . D? 99 . ffi¢ Id]! ‘ I 7 9 gIVf Rad52 102 . e _ H .\~ MAN-([41 :3” 3“ ‘j ‘1 I“?! G-b-PDH ...,.-" ‘m.a=-n- ’”25 Figure 2-3 C. Quantitative comparison ofthe Rad52 protein level in rat/52A cells expressing GALl- RAD52F1 or GALl-RAD52F3 Proteins were extracted from cells cultivated in medium containing 2% raffinose and 0.05% galactose. The amount of total protein loaded is indicated above each lane. The Western blot was probed with an anti-Rad52 antibody. after which the membrane was stripped and probed with an anti-G-o-PDl-l antibody. 103 Au 5 823. ESPE .Sch 2 as“: c- v N M.,—NmA—éAs—«Jv IDA—-90 Nmtam l04 D. Examination of RAD52 mRNA transcribed in rut/52.1 cells expressing GALl- RAD52F] Total RNA was prepared from JP166 (RAD52). JPIOOLI (rm/52A), and JP166L1 expressing GALl-RADSZFI cultivated in medium containing 2% raffinose and 0.05% galactose. The radioactive labeled probe hybridizes to nucleotide 243-259 of the RAD52 sequence. The nucleotide positions ofthe first. second and third ATG are indicated. The expected sizes for mRNA transcribed from upstream ofthe first. second or third ATG are also indicated. RAD52 transcripts detected only in JP166L1 GALl-RAD52F1 are indicated by a bracket. The transcript with a 5’ end identical to the endogenous RAD52 transcript is indicated by an arrowhead. l()5 2591n 219 nt 159 nt G: 53 E. 1 AMV—k. 9.3 2: .2. amN LH 2 2%: A052. 1=~v av 6:. a: _ . ‘ ' .. : l .lceN - . Iv.lll Alv 2.289.. I'm]... ‘ _h_mmd<~_-_1_<0 can IV. fl .6 IVfl/Vfl IV 096 99 O. 0.0. I? >9 1 e/ LO $1 0 v9 sl) an“ fan» be» r» l ()6 E. Analysis ofthe MMS sensitivity ohm/52.4 cells expressing GALl-RADSZFIATGI The wild type strain JP166 and the I‘Ut/52/l strain JP166L1 expressing GALl— RADSZFI. GALl-RADSZFIATGI. or the control vector. were cultivated in SC-Ura-Leu medium containing 2% raffinosc to mid-log phase. lO-fold serial dilutions containing 106 to 10 cells of each strain were spotted on SC-Ura-Leu plates containing 2% raffmose and 0.05% galactose with or without 0.005% MMS. F. Examination of Rad52 protein expression in rat/52.1 cells expressing GALl- RADSZF l ATGl Proteins were extracted from cells cultivated in medium containing 200 raffinose and 0.05% galactose for induction of Rad52 expression. The Western blot was probed with an anti-Rad52 antibody, after which the membrane was stripped and probed with an anti-G-6-PDH antibody. l()7 Tn 0.59,.— _ =:._-e-u ill- REE e «U .43 .2.» \l \/ .W .V o.» .92 r.— no 6» 9e 5v V. /O €335 58:: .889» m u M .u:_,._~mnm < ll() B. Complementation of the MMS sensitive phenotype of rad59A cells with GALl- RAD59 JP166W2 (rm/59x1) strain expressing CALI-RAD59 or the control vector was cultivated in SC-Ura-Leu medium containing 2% raffinose to mid-log phase. lO-fold . . . . . (, . . . serial dilutions containing 10 to 10 cells of each strain were spotted on SC-Ura-Leu solid medium containing 2% raffinose and (1.05% galactose with or without 0.01% MMS. 111 925. 02 YN 2%: 222 $36 .539.» amfi<¢-—A M m QNmn20 kb terminal restriction fragments. Those telomeres are maintained by a Alternative Lengthening of Telomere (ALT) pathway. ALT exists in about 30% of human cell lines. Rapid elongation of telomeres following gradual shortening has been observed in human ALT cells [14]. The long and heterogeneous telomeres observed in ALT cells are similar to those in type II yeast survivors, suggesting that human ALT is mediated by a recombination process similar to that which occurs in the type II survivor pathway in S. 130 cerevisiae [l l, 15]. Indeed, DNA sequences can be copied from telomere to telomere in human ALT cells [16]. Human ALT cells contain ALT-associated PML bodies (APB), novel promyelocytic leukemia (PML) bodies [17]. Telomeric DNA, Rad51 and Rad52 have been detected in APB. Noticeably, the appearance of APB coincides with the activation of ALT. The existence of Rad52 in APB suggests a possible role for Rad52 in human ALT. Rad52 is involved in virtually all homologous recombination in yeast. It is conserved throughout eukaryotes, especially at its N-terminus which contains the core activities, including DNA-binding and self-association [18-20]. This region alone is able to catalyze homologous pairing [18, 21]. Rad52 interacts with Rad59 through the N- terminus [22, 23]. This interaction stimulates Rad52’s function in single strand annealing in vivo [24].The C-terminus of Rad52 interacts with replication protein A (RPA) and Rad51. Rad52 can efficiently anneal short complementary oligonucleotides [19]. However, it needs RPA to eliminate secondary structures in long DNA molecules for efficient annealing [25]. Rad52 acts as a mediator between RPA and Rad51 to facilitate Rad51 nucleation onto ssDNA substrates [26-29]. Rad52 also stimulates strand exchange by stabilizing Rad51 presynaptic filaments [30]. Although the functions of Rad52 in homologous recombination have been studied in detail, it remains unclear as to how Rad52 differentially participates in the two telomere maintenance pathways in survivor cells. Similarly, while an increasing number of rad52 mutants have been identified and characterized, the effects of Rad52 mutants on survivor pathways have not yet been studied. In this work, the well-conserved N-terminus of Rad52 was subjected to random mutagenesis to identify residues that are critical for its 131 functions in survivor pathways. 1 identified mutations differentially affecting the two survivor pathways. R70G, K159E and R171S mutations cause defects in the type II pathway, and this phenotype correlates well with a reduced efficiency in carrying out direct repeat recombination. D164G results in defects in both type I and type II pathways. It also leads to a severe defect in interchromosomal recombination. Interestingly, while the overall recombination efficiency between two direct repeats remains normal, nearly all the recombinants are generated via intrachromatid pop-out events. These mutants may provide tools for detailed studies of telomere maintenance in the absence of telomerase. MATERIALS AND METHODS Yeast Strains, media, aflgenetic methogg Yeast strains used in this study are listed in Table 3-1. JP166 was generously provided by Dr. John Prescott (University of California, San Francisco). A type I survivor strain. JP166SIO, was created by selecting for 5-fiuoroorotic acid (5-FOA) resistant clones of JP166 followed by five re-streakings on YPD. JP166L1, a rad52A strain derived from JP166, was constructed by replacing the RAD52 with the HIS3 gene. JP166L041 was constructed by replacing the RAD5] gene with KanMX6 [31] first, then replacing the RAD52 with the HIS3 gene. JP166L042 was constructed by replacing the RAD59 gene with KanMX6 first, then replacing RAD52 with HIS3. All gene disruptions were complete deletions of the open reading frames and were constructed by transfomiing the cells with a PCR-generated gene disruption cassette [32]. W2078 and W2014-5C were kindly provided by Dr. Rodney Rothstein (Columbia University). 137 ‘- S. cerevisiae strains were propagated at 30°C in dropout media lacking the amino acids required for plasmid selection. Yeast transformation was performed according to Agatep. R. et al. [33]. Plasmids and mutations Plasmids used in this study are listed in table 3-2. Vector pRSG415 (generously provided by Dr. John Prescott, UCSF) used for galactose-inducible expression, contains the GAL] promoter and the C YC l terminator [34] in pRS415 backbone, a CEN vector with a LE U2 selectable marker (Stratagene). An Smal site was previously engineered between the GAL] promoter and the C YC I terminator. The RAD52 coding sequence starting from 1 17 bp downstream of the first ATG to the stop codon was subcloned into the Smal site of pRSG4] 5. The single BamHI site in the coding sequence was deleted by a silent mutation. A BamHI site was introduced into the 5’ end of the coding sequence. An SpeI site was introduced at nucleotide 71 l for cloning purposes, resulting in a lysine to serine mutation. This mutation does not affect Rad52 functions (data not shown). The resulting construct was designated as pLLl. pLL2 was constructed by ligating the Sall fragment containing RAD52 and its genomic promoter [35], amplified by PCR from genomic DNA of JP166, into the Sall site of pRS415. R70G, K159E, D164G and R171S mutations were introduced into pLL2 individually by site-directed mutagenesis (Stratagene) to create pLL2a, pLL2b, pLL2c and pLL2d, respectively. Mutations were confirmed by DNA sequencing. The XhoI/Spel fragment from these plasmids was inserted into pRS4l4 (Stratagene) to generate pLL3 (for RAD52), pLL3a (for R70G), pLL3b (for K159E), pLL3c (for D164G) and pLL3d (for R1718). 133 Random mutagenesis and yeast co-transformation Random mutagenesis of the RAD52 gene was carried out using Taq polymerase (Promega) as previously described [36]. Reaction mixtures contained lx mutagenic PCR buffer (7 mM MgC12, 50 mM KCl, 10 mM Tris-HCI pH 8.3, 0.01% (w/v) gelatin), 1x dNTP mix (200 uM dGTP, 200 [1M dATP, 1 mM dCTP and 1 mM dTTP), 20 fmoles of pLLl as input DNA, 30 pmoles of each primer (5”-CATTTTCGGTTTGTATTACTTC - 3‘ (anneals to a sequence within the GAL] promoter); 5’-TTTTCACCAGGTTCTTCGT CG -3’), with addition of Man at 0.5 mM. The reaction was carried out for 30 cycles for l min at 94°C, 1 min at 45°C, and 3 min at 72°C. To create a rad52 library, the purified mutagenized PCR fragments and linear pLLl lacking the BamHI/Spel fragment were co-introduced into JP166SIO.Yeast co- transformation was performed using the LiAc/ss-DNA/PEG method [33]. Genetic screen for rad52 alleles Transformants of JP166SIO carrying randomly mutagenized rat/52 on pRSG415 were replica plated first onto SC-Leu plates containing 2% raffinose and 0.8% galactose with 0.001% methylmethanesulfonate (MMS), then onto SC-Leu plates containing 2% glucose with 0.001% MMS as a growth control. Clones displaying increased MMS sensitivity were recovered from glucose-containing plates and patched on test plates (SC- Leu plates containing 2% raffinose and 0.8% galactose) and control plates (SC-Leu plates containing 2% glucose). Plates were incubated for 3 days. Cells from the test plates were then patched onto test plates and control plates one additional time. Clones showing a deficient growth phenotype on test plates, judged by small colonies or the lack of overall growth, were then streaked on test plates once. The candidates were recovered from the 134 control plates. Plasmids were recovered and re-introduced into JP166S10 to confirm the phenotype. Plasmids were subjected to DNA sequencing to identify mutations. Protein preparation and Western blotting analysis Yeast proteins were prepared following a procedure from the laboratory of Steven Hahn (gwwfligcflgfizyhhgflin). Briefly, yeast cultures were harvested at ODboo ~1.0, and washed in cold extraction buffer (200 mM Tris pH 8.0, 150 mM (NH4)2SO4, 10% glycerol, 1 mM EDTA) containing 2 mM DTT and protease inhibitor (Roche). Cells were then resuspended in cold extraction buffer with 2 mM DTT and protease inhibitor. 60% volume of acid-washed glass beads (425 - 600 microns, Sigma) were added to each sample. Samples were then vertexed at top speed for 1 min at 4°C for 5 times. Between each vortex, samples were kept on ice for at least 1min. Cell debris and glass beads were removed by centrifugation. Equal amounts of protein, detemiined by the Bradford method (Pirece), were separated on 10% SDS-PAGE and blotted onto polyvinylidene difiuoride membrane (Amersham Pharmacia Biotech). Rad52 was detected with a goat anti-Rad52 antibody RAD52 yC-l7 (Santa Cruz Biotech) followed by a rabbit anti—goat IgG HRP (Sigma). Membranes were stripped and probed with a rabbit anti-G-6-PDH antibody (Sigma) followed by a goat anti-rabbit IgG HRP (Sigma) for a loading control. Determination of MMS sensitivity Yeast cells were grown in appropriate dropout media to mid-log phase. Cells were collected and washed twice with ddeO. lO-fold serial dilutions containing 10 to 10° cells were spotted on plates with or without methlmethane sulfonate at concentration 135 specified in figure legends. Plates were incubated at 30°C for three to four days and photographed. LiquLid growth potential assav JP166L1, JP166041, or JP166042 carrying either wild type RAD52 or mutant rad52 were selected using 5-FOA for loss of pRS3 16TLC1. Individual tchA colonies for each strain were inoculated into appropriate dropout media. Growth potential was monitored as described previously [8, 10]. Briefly, every 21 hours, cells were counted using a hemocytometer and inoculated into fresh media at 3x105 cells/ml. The growth potential of each culture was presented as an average cell density at the end of each day. Formation of survivors Survivors were obtained as previously described with slight modifications [9, l 1]. For liquid assays, 3 independent transformants of JP166L1 carrying either wild type RAD52 or mutant rad52 were selected by 5-FOA to lose the pR316TLC1 plasmid. 10 individual tlclA colonies from each transformant were inoculated into appropriate dropout media and cultivated for 3 days. Cells were then inoculated into fresh media with a 1:10.000 dilution. This process was repeated five times to allow cellular senescence to occur and survivors to appear. Alternatively, survivors were obtained by single colony assay. 10 to 20 individual tchA colonies from each transformant were streaked on solid plates until colony sizes reached ~0.5 mm. For strains displaying high mortality during senescence, more colonies were picked from the first streaks for subsequent re-streaking in order to obtain a certain number of survivors. All plates were incubated for three to four days before colonies were picked. 136 DNA preparation and Southern blotting analysis Cell pellets collected from each overnight culture were suspended in SEB (1 M D- Sorbital, 100 mM EDTA (pH 8.0), 14.4 mM 2-mercaptoethanol) containing 1 mg/ml of zymolyase (ICN Biomedicals. Inc) and incubated at 37° with shaking for 30 min. After brief centrifugation, the pellets were suspended in EDS (50 mM EDTA pH8.0, 2% SDS, 0.025 N NaOH) and incubated at 65°C for 15 min. Following addition of NH4OAC, samples were precipitated at 4°C for 30 min. The supernatant was precipitated with isopropanol. The resulting pellets were resuspended in TE (10 mM Tris-HCl pH 8.0, lmM EDTA pH 8.0) containing 20 ug/ml RNase. Equal amounts of genomic DNA judged by ethidium bromide staining were digested with XhoI to completion and resolved in 1% agarose gel and then blotted onto positively charged nylon membrane (Amersham Pharmacia Biotech) in 1.5M NaCl, 0.5M NaOH following manufacture’s manual. After UV-crosslinking, the membrane was hybridized to a yeast telomere probe (5’- TGTGGTGTGTGGGTGTGGTGT-3’) labeled with [7-33P] dATP using T4 polynucleotide kinase (New England Biolabs). Hybridization was carried out in 0.5M N3H3PO4/N33HPO4 pH 7.2, 7% SDS, 1 mM EDTA pH 8.0 at 55°C for overnight (modified from Church and Gilbert [37]). The membranes were exposed to phosphorescent screens and the images were scanned with Phosphorimager (Molecular Dynamic). Determination of mitotic recombination rates Mitotic recombination between leuZ—AEcoRI and leu2-ABstEII heteroalleles was examined in diploid strains (W2078; for interchromosomal recombination) or in haploid strains (W2014-5C; direct-repeat recombination) as described previously [38]. Briefly, 137 wild type RAD52 or mutant rad52 on pRS4l4 backbone (pLL3, pLL3a, pLL3b, pLL3e and pLL3d) were introduced into W2078 or W2014-5C. Fresh transformants were inoculated into SC-Trp medium at 2x104 cells/ml and grown to 2x107 cells/ml. Plating efficiency and the number of recombinants were determined by plating an appropriate number of cells on SC-Trp and SC-Trp-Leu plates, respectively. For each mutant rad52, five to seven independent transformants were analyzed. Recombination rate and standard deviation were determined as previously described [39]. A two-tailed t-test was used to determine the significance of differences between rad52 and RAD52. To detect the deletion of the URA3 marker in direct-repeat recombination events, Leif recombinants from SC -Trp-Leu plates were patched onto SC-Ura plates to determine the percentage of Ura’ recombinants. RESULTS Experimental strategv There are five in-frame ATG triplets at the 5’ end of RAD52. In Chapter II, I have shown that the third, forth and the fifth ATG can serve as the translation initiation site in vivo, and that Rad52 translated from the fifth ATG has the same function as that translated from the third ATG. Therefore I cloned RAD52 starting from the fifth ATG into the pRSG415 vector to create pLLl, in which RAD52 expression is under the control of the galatose-inducible GAL] promoter. The well—conserved N-terminal region, about 230 amino acids, of Rad52 was targeted for random mutagenesis using pLLl as the template by error-prone PC R-based mutagenesis [36]. The mutagenized PCR products were introduced into a survivor strain JP166S10, along with the linearized pLLl lacking 138 the targeted region of RAD52. Gap repair of the linear vector with the PCR fragment gives rise to a circular plasmid (Figure 3-1). Approximately 24,000 transformants were screened by a two-step screening scheme illustrated in Figure 3-2. The first step was to screen for mutants conferring higher sensitivity to the DNA damaging agent MMS. About 1,700 clones displayed increased MMS sensitivity. These clones were further screened for defects in survivor growth (Figure 3-3). Seventy-three candidate clones were isolated. DNA sequencing revealed 57 different rad52 mutants (Table 3-3), among which 7 mutants had single amino acid changes, 19 had two, 17 had three, and 14 had four to six mutations. 37 of the 94 mutated residues are conserved between HsRad52 and ScRad52 (Figure 3-4). Some residues were mutated in 6-8 different mutants (Figure 3-4). This high frequency of mutation at certain positions suggests that these residues are probably critical for Rad52 function. It is also possible that these mutations were generated at certain hot spots during PCR mutagenesis. However, at least some residues, for example, Asp164 and Arg171, are indeed critical for Rad52 functions (see below). The growth defects displayed by these rad52 alleles are specific for survivor cells, since a wild type strain, JP166, harboring these alleles showed normal growth phenotype (data not shown). To study whether the growth defects of survivor cells were caused by defects in telomere maintenance, four single mutations, R70G, K159E, D164G and R1718, were examined in detail. These four residues are evolutionarily conserved and some were studied by others [21, 38, 40]. These mutations were introduced into RADSZ carried on pRS415 or pRS414, low copy vectors in which RAD52 is controlled by its own 139 promoter. The constructs were introduced into rad52A strains to study the effects of these mutations on telomerase-independent telomere maintenance. Mutant rad52 displays increased MMS sensitivity The four rad52 mutants were first examined for their ability to complement the MMS sensitivity of a rad52A strain JP166L1 by spot assays. Cells expressing R70G, K159E, or R1718 showed similar sensitivity to MMS as cells carrying the control vector (Figure 3-5). Cells expressing D164G were more resistant to MMS treatment than the other three mutants. However, they were significantly more sensitive than cells expressing RAD52 (Figure 3-5). Therefore, R70G, K159E, and R1718 cannot complement the MMS sensitivity of rad52A, and D164G partially complemented rad52A. Next I examined whether these mutations caused defects in Rad52 protein expression. While R70G and R1718 did not affect Rad52 level, K159E and D164G reduced Rad52 expression (Figure 3-6). However, a low level of protein expression is not likely the cause for the phenotypes of K159E and D164G. D164G, while expressed at a lower level, is the only allele that could partially complement rad52A. In addition, though present at similar levels, K159E and D614G alleles display distinctive phenotypes in survivor pathways and perform homologous recombination with different efficiencies (See below). Rad52 mutations lead to early senescence To study the effects of Rad52 mutations on survivor pathways, we examined the growth potential of tchA cells expressing rad52. Growth potential was measured by diluting liquid cultures to 3x10°cells/m1 and examining cell densities every 21 hours. Cell 140 densities for each day were plotted to generate “grow-out-of senescence” curves. A yeast culture gradually loses the capacity to repopulate as cells enter senescence due to telomere shortening. This process is reflected by decreased cell densities on grow-out curves. When cells recover from senescence, cell density will return to a normal level, indicating survivors have been generated [9, 10]. The abilities of rad52 mutants to generate survivors were compared to that of RAD52. R70G, K159E, D164G and R1718 entered senescence on the third or forth day after telomerase activity was eliminated by shuffling out the URA3 plasmid bearing the sole copy of TLC], whereas RAD52 cells entered senescence on the fifth and sixth day, suggesting that these mutations cause early senescence (Figure 3-7). All four mutants recovered from senescence, indicating that these mutations do not have an apparent effect on survivor generation. Rad52 mutations differentially affect the two sgrvivor pathways Yeast cells employ either a type I or type II pathway to generate survivors [9]. These two survivor pathways have different genetic requirements. Loss of certain genes causes cells to preferentially utilize one survivor pathway over the other, or even completely blocks the second pathway. To examine whether Rad52 mutations deferentially affect the two survivor pathways. I analyzed survivor types by Southem blot analysis. The two types of survivors arise at different frequencies depending on growth conditions [11]. In general, most of the survivors are type I when there is no growth competition, such as growing cells on solid medium where individual colonies form independently (single colony assay) [11]. However, when growth competition exists, such as growing cells in liquid culture (liquid assay), most ofthe survivors are type 11 due to their higher growth rate [11]. The exact ratio between the two survivor types under 141 these conditions is strain-dependent [9, 11]. When cultivated in liquid medium, RAD52 generated roughly an equal number of type I and type 11 survivors. Like RAD52, D164G generated both type I and type II survivors at a similar frequency. In contrast, R70G, K159E and R1718 generated only type I survivors. (Figure 3-8 and Table 3-4), suggesting that these three mutations have negative effects on the type II pathway. When streaked on solid medium, R70G, K159E and R1718 showed similar phenotypes as RAD52 in that type I survivors occurred predominantly. In contrast, 40% of the survivors generated by D164G were type II. The increase of type 11 survivors in single colony assays suggests that D164G mutation has a negative effect on the type I pathway, or results in a more active type II pathway, or both (Figure 3-9 and Table 3-4). Taken together, the above results indicate that R70G, K159E and R1718 survive senescence preferentially through type I pathway. D164G has a preference for type II pathway as shown by single colony assays. However, the frequency of type 11 survivors was not increased in liquid assay, which favors type II pathway. It is possible that D164G is also defective in generating type 11 survivors, though to a lesser degree than generating type I survivors. Alternatively, the D164G mutation may have negative effects on the proliferation of type 11 survivors. Therefore, the growth advantage of type 11 survivors over type 1 survivors in liquid Cultures is diminished, and results in changes in the ratio of the two survivor types. To further provide genetic evidence for the differential utilization of type I or II survivor pathways by these rad52 mutants, we introduced each allele into rad52A rad5 IA (JP166041) and rad52A rad59A (JP166042) strains. Deleting RAD5] or RAD59 forces cells to use, if available, only the type II or type I pathway. respectively. Liquid growth potential assays were performed to examine cell growth and survivor generation. When introduced into the rad5lA strain, R70G, 159E and R1718 showed declined growth potential similar to that of RAD52 after telomerase activity was eliminated. However, they did not generate survivors, whereas RAD52 recovered from senescence (Figure 3- 10), indicating that these three mutants rely mainly on type I pathway to generate survivors. In the absence of Rad51, none of these rad52 mutants displayed discernible type 11 survival, indicating that the mutations cause severe defects in the type II pathway. In contrast, two out of five D164G were able to generate survivors, at much later time points compared to RAD52 (Figure 3-10). Thus, D164G is also defective in the type II pathway, though to a lesser degree than the other three mutants. The same rad5/A rad52 mutant strains were also tested on solid plates for survivor generation. 24 samples were tested for each strain. Consistent with the liquid assay, cells expressing R70G, K159E or R1718 did not generate any survivors, whereas D164G generated survivors at a reduced frequency (data not shown). Together, these results suggest that all four mutants are defective in eliciting type 11 survival. When introduced into the rad59A strain, all four mutants were able to generate survivors in liquid media (Figure 3-11). However, cells expressing K159E or D164G displayed a more profound senescence phenotype (Figure 3-1 1). The accelerated decline in growth potential is similar to that of rad52A described previously [9]. Deletion of Rad59 does not change the ability of the mutants to generate survivors as all the mutants were still able to survive senescence. Since rad59A mutation selectively eliminates type II pathway, these results indicate that all four mutants maintain an appreciable portion of the type I survival function. 143 rad52 mutants perform homologous recombination with different efficiencies Based on the telomere DNA sequences in survivor strains and the genetic requirement of the two survivor pathways, it has been proposed that the type I pathway arises from recombination between the Y’ elements on different chromosomes and the type II pathway may be caused by recombination between the telomere repeats on the same (looping back) or different chromosomes [9, 11, 41]. Since the four Rad52 mutations appear to display differential effects on survivor pathways, it is possible that corresponding effects can be seen in selective homologous recombination events. To test this, I examined the efficiency of recombination between two homologous alleles present on different chromosomes or as tandem repeats on the same chromosome. The interchromosomal heteroallelic recombination efficiency of each rad52 mutant was determined by introducing the mutant into a homozygous rad52A strain carrying two nonfunctional leuZ alleles, leuZABstEII and leu2AEcoRI (Figure 3-12A) [38, 42]. In the absence of RAD52, the rate of LE U2 recombinant formation was 8-fold lower than in wild type. R70G and R1718 displayed a 2 to 3-fold increase in recombination rate compared to wild type (Figure 3-12B). K159E had a recombination rate similar to RAD52. In contrast, D164G reduced the recombination rate by 14-fold. These results indicate that while R70G, K159E and R1718 mutations do not have significant effects on interchromosomal recombination, the D164G mutation causes a severe defect similar to rad52A. The direct-repeat recombination rate was examined by using a haploid rad52A strain carrying the leuZABs/EII and the leuZAEcoRI alleles as tandem repeats on chromosome V (Figure 3-12C) [38, 43]. A URA3 marker is flanked by the two leuJ 144 alleles. LEU2 recombinants can be generated via different mechanisms, including pop- out events that result in the loss ofthe intervening URA3 marker, gene conversion events that replace one of the two repeats, unequal sister chromatid exchange that produces a triplication product and events that create disomes (Figure 3-12C) [38]. The preference for certain mechanisms reflects changes in the functions of homologous recombination proteins, and will inevitably affect the outcome of biological processes that are mediated by such events. The percentage of pop-out is of particular interest since such events could have devastating effects on telomere maintenance if it occurred between telomere repeats. The recombination rate was 30-fold lower in rat/52A cells than the IMD52 cells. R70G and R1718 caused a moderate decrease of 1.5 to 2-fold in recombination rates. K159E displayed a 15-fold reduction in the recombination rate. The tendency of popping out the intervening URA3 gene of these three mutants as evidenced by assaying the percentage of Leui/Ura' recombinants was similar to that of RAD52, which was ~60%. D164G displayed overall recombination efficiency similar to that of RAD52. Interestingly, nearly all of the Leu+ recombinants generated in D164G strain were produced by pop-out events ((Figure 3-12D). Thus, R70G, K159E and R1718 negatively affect the recombination between two direct repeats. These mutations do not change the tendency of deletion events. As much as 40% of the recombinants retain the intervening sequence. In contrast, D164G does not affect overall efficiency of direct-repeat recombination. However, it results in elevated pop-out activity. Most of the recombinants produced by D164G lose the intervening sequence. DISCUSSION Rad52 plays a central role in yeast survivor pathways [5, 9, 10]. tchA rad52A mutants cannot survive senescence. The highly conserved N-terminal region of Rad52 can catalyze homologous pairing [21]. Thus, the core activity of Rad52 appears to reside in this region [18, 21]. In this study, we identified amino acid residues critical for Rad52 functions in survivor pathways by screening for rad52 mutants that were randomly mutated at the N-terminal region. The functions of 4 residues, Arg70, Ly5159, Asp164 and Argl71, were examined in detail. We demonstrated that mutations of these residues differentially affect the two survivor pathways, as well as different homologous recombination events (summarized in Figure 3-13). R70G, R1718 and K159E cause defects in the type II pathway specifically. These three mutants perform interchromosomal recombination at near wild type efficiencies, but are moderately defective in direct-repeat recombination. D164G is defective in both type I and type II pathways. It also has a severe defect in interchromosomal recombination. While D164G does not result in an apparent reduction in direct-repeat recombination, most of the recombinants are produced through the intrachromatid pop-out mechanism. The correlation between the two survivor pathways and different recombination events provides further support that the type I pathway is mediated by recombination between telomeres on different chromosomes and the type II pathway is mediated most likely by telomere “looping back” [9]. Effects of Rad52 mflations on sgrvivor pathways The ratio between the two types of survivor is strain-dependent [5, 9, 11]. The changes in the activity of the two pathways can be assessed by comparing this ratio 146 between tlclA cells expressing wild type RAD52 and those expressing mutant rad52. While the tch RAD52 strain generates a similar number of type I and 11 survivors when grown in liquid medium, it generates only type I survivors when streaked for individual colonies. R70G, K159E, and R]7]S are defective in the type II pathway since tic/A cells expressing these mutants do not generate any type 11 survivors in either liquid or single colony assays (Figure 3-8 & Table 3-4). Consistent with this defect, these cells can not recover from senescence in the absence of Rad51 where cells rely solely on the type II pathway for survival (Figure 3-10). These alleles appear to have normal type I activity since they can recover from senescence and proliferate normally even in the absence of Rad59 (Figure 3-1 1). Several pieces of evidence suggest that DI64G is defective in both type I and type II pathways. There is a significant increase in type 11 survivors in single colony assays that favors the type I pathway, suggesting that D164G mutation results in a defective type I pathway, and/or an elevated type II activity. However, in the absence of Rad51 where cells rely on the type II pathway to survive through senescence, D164G generates survivor at a later time point and with a reduced frequency, suggesting that the D614G mutation has a negative effect on the type II pathway as well. Thus, the change in the ratio of the survivor types is possibly due to different degrees of defects in the two pathways. Contrithtion of inter- and intra- chromosomal recombination to the two sgrvivor pathways Based on the genetic requirement of the survivor pathways and the nature of the survivor telomeres, it has been widely accepted that the type I pathway is mediated by recombination between Y’ elements on different chromosomes, and the type II pathway 147 is mediated by recombination between the telomere repeats [9, l 1, 41]. However, for the type II pathway, it remains unclear as to whether the recombination between telomere repeats is via interchromosomal or intrachromosomal mechanisms. The defects of R 70G, K159E, and R] 7 ] S in the type II pathway correlate well with their defects in direct-repeat recombination, and their ability to carry out the type I pathway is consistent with their near wild type efficiency in perfomting interchromosomal recombination. The defect of D164G in the type I pathway is in accord with its marked deficiency in interchromosomal recombination. These results support the role of interchromosomal recombination in the type I pathway. Our results also suggest that type II survivor telomeres are maintained mainly by telomere looping back to copy the T(G)3-3(TG)I-6 repeats intrachromosomally. Consistent with this idea, the high incidence of excising the intervening sequence between two direct repeats may account for the defects of D164G in the type II pathway. There is evidence that telomeres form t-loops, in which telomeres loop back and the 3’ single-stranded tails of G-strands pair with the duplex telomeric DNA [44, 45]. t-loops have been observed in evolutionarily unrelated organisms [44, 45], suggesting they are a conserved feature of eukaryotic telomeres. Similar structures have been proposed to mediate telomere length regulation and the transcriptional regulation of genes placed in subtelomeric region in yeast [46-49]. This structure is likely disintegrated by illegitimate excision of the intervening sequence between the paired regions. Indeed, a mutation in TRF2, a human telomere binding protein. induces t-loop deletion and results in rapid telomere shortening [50]. Possible structural basis of the observed phenotypes of rad52 muLtants 148 The R70G, RI7IS and [(1595 alleles display similar phenotypes in homologous recombination. They are able to perform interchromosomal recombination normally, but are moderately defective in recombination between two direct repeats located on the same chromosome. Genetic studies show that intrachromosomal recombination requires RAD59 [41, 51]. Consistent with their defects in direct repeats recombination, all three mutant alleles are defective in the Rad59-depedent type II pathway. Rad59 physically interacts with Rad52 at the N-temiinal region of Rad52 that coincides with its self- association region [24] [52]. Thus, the defects of these mutants in the type II pathway and direct-repeat recombination could result from the disruption of Rad52-Rad59 interaction. It would be important to examine the interaction between Rad59 and Rad52 mutants by co-immunoprecipitation experiments. However, no suitable antibodies are available at present. Arg55 and Arg156 of the human Rad52, corresponding to Arg70 and Arg171 of the yeast Rad52, are important for DNA binding [21]. Substituting these residues with alanine results in severe defects in ssDNA and/or dsDNA binding by HsRad52. Arg70 and Arg17l of ScRad52 do not appear to be essential for DNA binding, since R70G and R] 7 IS are normal in interchromosomal recombination. This is further supported by the previously identified R70A and R]7]A alleles that show no defects in interchromosomal or direct-repeat recombination [38]. D164G allele is defective in interchromosomal recombination. It has no obvious defect in direct-repeat recombination. However, it preferentially generates pop-out recombinants. D164G is also defective in both type I and type II pathways. These phenotypes suggest that D164G mutation might compromise the DNA binding and/or 149 self-association functions, which will lead to defects in multiple genetic processes. It is possible that this allele is also defective in interaction with Rad59 for the same reason discussed for the other three mutant alleles. Secondary structure prediction places Lys159, Asp164 and Arg171 on the same or-helix. Their spacing is such that Asp164 and Argl 71 are located on the same side of the or-helix, and Ly3159 is located on the opposite side. The crystal structure of the N- terrninal domain of human Rad52 reveals that the corresponding residues, Lysl44, Asp149 and Arg156, are also located on a oz-helix. Arg156 is part ofa DNA binding site. In the ring structure formed by Rad52 monomers, Lysl44 forms a hydrogen bond with Asp149 of the neighboring monomer. This structure predicts similar phenotypes for mutations at Lysl44 and Aspl49 which disrupt the hydrogen bond. However, K159E and D164G mutations of the yeast Rad52 display different phenotypes. The difference could be due to, at least in part, different degrees of disruption of the interactions between Rad52 monomers. It is possible that there are subtle structural differences between ScRad52 and HsRad52. The fact that Argl 71 of ScRad52 appears not to be important for DNA binding supports such notion. SUMMARY In summary, I identified 57 rad52 alleles with defects in responding to MMS toxicity. Of these, I characterized in greater details 4 alleles for their phenotypes in telomerase-independent telomere maintenance and homologous recombination. rad52R70G, rad52K159E and rad52RI7IS have defects specifically in the type II survivor pathway. rad5ZD]64G is defective in both type I and type II pathways. The defects in telomere maintenance correlate well with mutant phenotypes in homologous 150 recombination. A mutant with a defect in the type II pathway is also defective in intrachromosomal direct-repeat recombination. A mutant with a defect in the type I pathway is also defective in interchromosomal heteroallelic recombination. These results provide further support for the proposed mechanisms of the two telomere maintenance pathways in the absence of telomerase [9, 1 1, 41]. Since the amino acid residues mutated in these alleles are highly conserved, analogous mutations in Rad52 homologues may have similar effects on their functions. 151 APPENDIX 3: FIGURES AND TABLES FOR CHAPTER 3 Table 3-1 S. cerevisiae strains used in this study strain genotype JP166" MATa his3A leu2A ura3A tlclA pRS3/6TLCI I MA Ta his3A leuZA ura3A tlclA rad52A::HlS3 J P166L1 ’ pRS3/6TLCY JP166L041h MA Ta his3A leuZA ura3A tlc] A rad5 ] A .° 5K anMX 6 rad52A .' .‘H1S3 pRS3 1 6 TLC] MA Ta his3A leuZA ura3A tlclA rad59AssKanMX6 JP166L042h rad52A.':I-IIS3 pRS3 ] 6 TLC 1 , MA Ta rad5 2: .'H]S5 S UP4 -o: .' CA N ] -H]S3 .° :sup4 + W2014-5C ‘ leu2-AEcoRI: .' URA 3 .'.'leu2-A BstEII W2078 " MA T a/oz rad52::HIS5/ rad52::H]S5 leu2-AEcoRI/ leuZ-ABstEll " From Dr. John Prescott (University of California, San Francisco) ” Derivative ofJPl66, This study. " From R. Rothstein’s laboratory {Mortensen, 2002 #4}. 153 Table 3-2 Plasmids used in this study Plasmid description pRS415 " a CEN vector with a LE U2 marker pRS414 " a CEN vector with a T RP] marker pLL2 RAD52 and its own promoter is cloned into pRS415 pLL2a R70G mutation is introduced into pLL2 pLL2b K159E mutation is introduced into pLL2 pLL2c D164G mutation is introduced into pLL2 pLL2d R1718 mutation is introduced into pLL2 pLL3 The XhoI/Spel fragment from LL2 containing RAD52 and its promoter is cloned into pRS4l4 pLL3a The XhoI/Spel fragment from LL23 containing rad52R 70G and its promoter is cloned into pRS414 pLL3b The XhoI/Spel fragment from LL2 containing racl52Kl59E and its promoter is cloned into pRS414 pLL3c The XhoI/Spel fragment from LL2 containing rad52D/64G and its promoter is cloned into pRS414 pLL3d The XhoI/Spel fragment from LL2 containing rad52R] 71S and its promoter is cloned into pRS4l4 pRSG415 b a GALl promoter and a C YC l terminator is cloned into pRS415 pLLl RAD52 coding sequence lacking the first 1 l7 nucleotide is cloned into pRSG415 between the GAL] promoter and the C YC 1 terminator. The single BamHl in the RAD52 sequence is removed by a silent mutation. An SpeI is inserter at position 71 1. " From Stratagene ” From Dr. John Prescott (University of California, San Francisco) 154 Figure 3-1 Construction ofrad52 library pLLl harbors RAD52 starting from the fifth ATG under the control of the GAL] promoter. RAD52 expression is induced by galactose, but repressed by glucose. The 5‘~70() bp of RAD52 was replaced with randomly mutagenized PCR fragments of the same region by gap repair. pLL l mut represents the resulting plasmids. Tm 23.x..— 6 Figure 3-2 Screening scheme for rad52 mutants JP166810 (a survivor strain) carrying wild type RAD52, mutant rad52 or control vector were plated on SC-Leu/glucose (noninducing) plates, after which colonies were replica plated onto SC-Leu/galactose (inducing) plates containing 0.001% MMS, then onto SC-Leu/glucose plates containing 0.001% MMS. Clones showed higher sensitivity to MMS on galactose—containing plates were recovered from glucose-containing plates, and patched onto SC-Leu/galactose plates twice to identify the ones that grew poorly. At the same time cells were also patched onto SC-Leulglucose plates serving as control for cell growth, as well as for recovering candidates. Candidate clones were further streaked on SC -Leu plates to examine the growth phenotype. Plasmid DNA recovered from the candidates was used to transform JP166810 and other strains mentioned in the text, and for DNA sequencing. 3 as”: com BOBoq- ROBE- .m0\:oq-\m22 cox...) j .4\ . \\ xmozm \ & //.K\fl E923- [R /O O O O I O O O O O o o o o o o 83.? O O O \ + 20.53- :23 \VWEM/ /_/C\ a... Q _ My m £8 8.602 All ax mEEmEQ 5.503. g 2033- h \_\ _ M36538 29.3%st a. 2025:5200 20:64- Figure 3-3 Growth phenotype ofJP166810 expressing rad52 alleles This figure shows a plate from the last step of the screening procedure illustrated in Figure 3-2. JPl66810 harboring the control vector expresses the genomic copy of R.-ID52. It served as the control for comparing growth phenotype. JP166810 expressing RAD52 showed normal growth. JP166810 expressing rad52-215 or rad52-507 displayed deficient growth phenotype. These two clones were identified as candidates. The other four clones shown in the figure displayed less severe or no apparent growth phenotype. These clones were not investigated further. 159 m3. -Nwwau m m 95$... m~Nnmth .869» 160 Table 3-3 rad52 alleles identified in this study Mutant I.D. Mutations rad52- 101 K60E rad52-102 K61N rad52-103 P64H Single mutants rad52-104 R7OG rad52-105 K159E rad52- 106 D 164G rad52—107 R1718 rad52-201 E113G; V193E rad52—202 R70G; N203I rad52-203 K60E; F2068 rad52-204 D164V; N204D rad52-205 K61E; R217G rad52-206 K167R; R1 71G rad52-207 F738; D164G rad52-208 E65D; N232D rad52-209 K69E; N 179D Double mutants rad52-210 R171G; 11908 rad52-21 1 K61N; N204Y rad52-212 R207G; Q229P rad52-213 K61E; T75P rad52-214 K69E; N242Y rad52-215 R1718; E202K rad52-216 K57N; Y80D rad52-217 D41V; K61E rad52-218 1120L; F1951 rad52-219 R1718;N219S 161 Table 3-3 rad52 alleles identified in this study (continued) Triple mutants Mutant I.D. Mutations rad52-301 V1261; K167R; T218A rad52-302 T163M; N2048: E223A rad52-303 R207G; N2428; P245L rad52-304 R171G; I212N; N242D rad52-305 T163K; T2208; N2328 rad52-306 R207G; E21 1V; Q229L rad52-307 1120M; V193M; H222L rad52-308 D41H; D164G; Q229R rad52-309 T75A; D164G; N204D rad52-310 N146D; L237M; 8247P rad52-31 l Y80D; R858; F1958 rad52-312 N97D; K117M; Y141F rad52-313 K61E; 178T; K192R rad52-314 K184N; D210E; P2318 rad52-315 E52N; T1348; D164G rad52-316 Y66H; L2058; E223V rad52-317 K184E; R207G; N244Y Quadruple rad52-401 F738; T75A; D164G; D199G mumms rad52-402 R77G; 181T; L187P; 8215N rad52-403 F1 101; 1120V; N175D; N203D Quintuple rad52-501 D53G; D201E;R217K;N232D;D246Y mutants rad52-502 F47Y; D112G;K167R; E214V; Q227R rad52-503 K159N; E211G; 8213R; L2218; K233E rad52-504 L621; G63R; 1120V; E155G; D199G rad52-505 E42K; V86A; D164N; D201V; D210E rad52-506 K167E; N179S; D199G; E223K; Q239H rad52-507 E147V; P231T; R234G; 8239N; N244Y rad52-508 P64L; G74E; W84R; Q1 15R; L237S rad52-509 V46A; T101A; 1190N; E223G; V2401 Sextuple mutants rad52-601 Y80N; SIOSR; F1951; P1978; 8215G; Y230H rad52-602 N97Y; L111M;T163M; D201E;T218A; L221F 162 Figure 3-4 Distribution of mutations Ninety-four residues were mutated in fifty-seven mutants. Only residues shared by at least two mutants are shown. Seven mutants have single amino acid substitution. Those mutated residues are indicated by open boxes. Residues indicated by open ovals are conserved between thad52 and 8crad52 (Only residues up to amino acid 171 are shown.) The numbers next to certain residues indicate the number of mutants that have mutations at that position. 163 Tm «Eur— :S 28 SS. 38 28. \éN 82 :2 :2/ SS 3:. ~32 3? \ mi..— 23 8; $2 :3 .5: 3:: 3.2”.— MNS i 2.2 a q ,. , . . . u . 8N :— ezEmE BEBE 08mm 2: mcthm mwcmazEmouopEzz Q 2668 cotomcou O as. 2: HE E 0 ME IE m? an. e a... .3 cm— 164 Figure 3-5 MMS sensitivity assay ofrad52 mutants JP166L1 (a rat/52A strain) expressing wild type RAD52, rad52R70G, rad52/<1 5 9E, rat/520164G, rad52/81718, or the control vector were cultivated ovemight to mid-log phase. 10-fold serial dilutions containing 10 to 10° cells of each strain were spotted on plates with or without 0.005% MMS. MMS sensitivity was evaluated after3 or 4 days of incubation. 165 m2: cZ an 2.3.: m2 2 =\emce.¢ .589, J . mam—v— DE»— mg;— Owe—Q $9.2 \ (V3 51"“ ) I'l99ldl‘ 166 Figure 3-6 Steady state protein levels of rad52 alleles Yeast proteins (60 pg) were prepared from strains with the indicated genotypes. The proteins were separated in 10% SDS-PAGE. The membrane was probed with an antibody against Rad52, after which the membrane was stripped and probed with an antibody against G-6-PDH. 167 3. 2%: Q/V/ 0.1.0. . \, Pu/J/J / x ‘ é/J r \ PJ/Jv . Abv . OAIV )9 « :Vc . 9c 1.»? . JV». \( r94 0. 04 .er Ox Afimwhaisavoas IDLéAy chmm 168 Figure 3-7 Growth potential of rad52 mutants through senescence and recovery process JP166L1 (rat/52A) expressing wild type RAD52, rad52R7UG, rad52/(I595, rarl52D/64G or rat/52RI7IS was plated on solid SC-Leu medium containing 5-FOA to lose the sole copy of TLC] bearing on a URA3 plasmid. Individual tchA colonies were cultivated in liquid SC-Leu medium to examine growth potential through senescence and recovery process. Cells were counted every 21 hours and inoculated into fresh media at 3x105 cells/ml. This process was repeated for 11 days to generate “grow-out-of- senescence” curves. The curves shown are the average of 5 samples for each genetic background. Error bars represent standard deviation. 169 23.30 E gun :Sawnmmvnue 8&8; ho+moo€ 8&2... mrhwm I Nmn—(m O 23.30 5 £30 :omehmmvmmv 1.1.1 1.1 liliill.-l-l l- l 1 868.. ho+wccé Smog man .9. I Nan—<1 O (rm/sues) Kitsuea Iiao (rm/sum) Ausuaa "as E e. a: 23.30 5 £30 :cwmmhcmvmwv 9.05 - ~35. 9 23.30 5 gen :cwmmhomemww 02m - ~35. o mo+mcc€ hc+mco€ mo+moc€ wo+moo.—. hc+mccé mo+moc.—. (rm/Silas) Misuea Mao (Iw/suao) Mtsuea "ea 1 70 Figure 3-8 Southem blot analysis of telomeric DNA in survivors obtained by liquid assay JP166L1 expressing wild type RAD52, rad52R70G, rad52/(I595, rad52D164G or rad52/{INS was selected on solid medium containing 5-FOA to lose the pR8316TLC1 plasmid. Individual tlclA colonies were inoculated into liquid SC-Leu medium and cultivated for 3 days. Cultures were then diluted 1210,000 with fresh medium. This process was repeated five times until survivors appeared. Genomic DNA was digested with Xhol, which cuts within the Y’ elements once. The filters were hybridized to a poly(dG-dT) probe. Type II telomeres, indicated by triangles above lanes, are characterized by multiple bands with various sizes. All others samples are type I telomeres. A: RADS2. B: rad52R70G. C: rad52Kl59E. D: rad52D164G. E: rad52RI7IS. 2 use... a: £23: .m uuuuununuuuuuuu nu MQWNENMEE .0 DE RN35. .m— ztntn uuuuuxqunm .. .u rt. r.» b b» w it "mmumru-u-¢.nnam DDDD DD .5) DVEQNWEE d ”.2 2 I. p» > > >>> .. a Z : t z x N» > >>> NwQ—xk .< m.-.a........ 172 -. «g 3 ‘. .. .. I . nut-unou... O- P Figure 3-9 Southern blot analysis of telomeric DNA in survivors obtained by single colony assay JP166L1 expressing wild type R.AID52, rat/53R7IIG, rad52/(I595, rad52D/64G or rad52/{I718 was selected on solid medium containing 5-FOA to lose the pRS3 16TLC1 plasmid. Individual tlc].~l colonies were streaked on solid SC-Leu medium. Single colonies were then picked for the next streaking. This process was repeated five to seven times until survivors appeared. Survivors were cultivated in liquid medium for overnight and collected for analyzing telomeres. Genomic DNA was digested with (17201, which cuts within the Y” elements once. The filters were hybridized to a poly(dG-dT) probe. Type 11 telomeres are indicated by triangles above lanes. All other samples are type 1 telomeres. A: RAD52. B: rad52R70G. C: rad52/(I595. D: rad520164G. E: rat/52R] 713. 173 Otl...‘l.t I. In. I {11“vauulb'.‘ 0|. h? 2N5»? .u '0' .q. '! "ll-lb-,-l'uul‘l Mowcamiau .U .3. 2:“: 3 wrtmuummwuumun run..- . .5551. to I. >>. z .. I m - D. D DD 5’ O D D ,DXCQNhfiE .A— ..:JI.. .- S I... 1--....“ .............. 333.338.; II. ‘OO'““.O.‘.I|IIOII I-O.0-O... O.. . o u; .. . a. 1 IlI”---Il."olu rt“...3im I l I III-sol” L-’ I- .03 23.3: .m NMQTQ .< 174 Table 3-4 Differential utilization of the two survivor pathways by rad52 mutants Single colony assay Liquid assay Type 1 Type II TypeI/typeII Type 1 Type II TypeI/typeII tlclA RAD52 30 0 N/A 16 14 1.1 f [(15312 R 70 G 30 0 N/A 30 0 N/A gig/(1595 30 0 N/A 30 0 N/A iifiifzo/mo 25 17 1.5 17 12 1.4 xiii/21 715 30 0 N/A 30 0 N/A N/A: not applicable Figure 3-10 Growth potential of rad52 mutants in a rad5/A strain through senescence and recovery process RA D52 or rad52 mutants were introduced into J P166041 (rad5/A rat/5221) strain which relies solely on the type 11 pathway to survive through senescence. After telomerase activity was eliminated by shuffling out the URA3 plasmid bearing the sole copy of TLC], individual tlclA colonies were inoculated into SC-Leu medium. Cell density was examined every 21 hours and cells were inoculated into fresh media at 3x10S cells/ml. This process was repeated for 10-12 days to generate grow-out-of senescence curves. The curves shown are the average of 5 samples for each genetic background except for D614G. D164Ga represents three samples that did not recover from senescence. D164Gb and D164Gc represent the two samples that recovered from senescence at day 7 and day 12 of the experiment. respectively. The error bars represent standard deviation. S. or m m w m 23.30 5 9:3 h m m mgr”...— I v 23.30 5 £30 5 o m man 5. I v wauug <34»? $93.. =_ 3:3:E 3.3: .8 NWQVQ n N «35. o m N ~35. o mo+wocé oc+mooé hc+mccé wc+wco€ mo+moo€ cc+moo. P 22.25. P wc+wooé (Ina/$1160) Mrsuaa 116:) (rm/911601511suaa 1160 Nrwwcpmwnomvnmw 09.9.0 x 90334 or m m a.-.” 2%: 33.30 E 230 9.3.30 E 930 h m m 03m I v 835 o ~35. o m N ~35. o mo+wcc._. mo+wcc._. hc+mcoé wc+moo€ mc+woo._. oo+moo€ sc+wocé wo+moo€ (Iw/suao) Misuao 1163 (rm/sues) Mtsuea 1163 177 Figure 3-1 1 Growth potential of rad52 mutants in a rad59A strain through senescence and recovery pI'OCCSS RA D52 or rad52 mutants were introduced into J Pl66042 (rud59A rad52A) strain which relies solely on the type I pathway to survive through senescence. After telomerase activity was eliminated by shuffling out the URA3 plasmid bearing the sole copy of TLC], individual t/clA colonies were inoculated into SC-Leu medium. Cell density was examined every 21 hours and cells were inoculated into fresh media at 3x105 cells/ml. This process was repeated for 12 days to generate grow-out-of-senescence curves. The curves shown are the average of 5 samples for each genetic background. The error bars represent standard deviation. 178 ervcwmwhomvmww 85:: <33: 8::- =_ 352...: 3:: s 23$ :A 2:“: 93.30 5 £30 8&8; 3.58... 2102: mEE I 393. O 23.30 5 £30 vawovmwhomvnww : - I-. . - I- - I .wo+m8.w \t 3&8; \ “ mo+Mocé mans. - ~35. o (Iw/suao) Misuse Ileo (Iw/suaa) Kasuea "80 23.30 5 £60 erwmemhmmvan an 035 - ~35. 0 9.3.30 5 £25 Newwcwmmhomvmmw 02m - ~35. o mc+mcc._. salmon. v wc+moo€ wo+mocé ho+moo€ war-moo... (um/sues) Misuse Ilao (Ins/sues) Misuse "63 179 Figure 3-12 Effect of Rad52 mutations on interchromosomal recombination and direct- repeat recombination A. Possible recombination events in interehromosomal recombination. ln interchromosomal recombination. LEU2 recombinants can arise through (a) reciprocal exchange, (b) gene conversion ofAB-s'IElI allele, and (c) gene conversion of 411?le allele. lb’l) 38$ mom--3;- i: 3:253 ._n S :3 £3.29th :5: BEGU< £3. 23:... T i $72.2 E1 T T l.r 3.": 3. Jill i ~=2 Es T 1 m3 ~=2 Ii ~23 T i 33 E l\\lT ~33 $3 {I is ~22 IT Tl E 18] B. Effect of Rad52 mutations on interchromosomal recombination. Recombination rates are calculated as events per cell per generation. Relative recombination efficiency is also indicated. A two-tailed t-test is used to determine the significance of differences between rad52 and RAD52. P value: D164G 0.032; K159E: 0.599; R700 0.064; Rl7lS 0.064. E 3. as“: v—vmma 2:: Dana mam—v— Uve—D nmn? 3.3352. _ 39:. _ .53583353 33353332353 M 3. 232... 3535833333 23353332332 _ _ 3 _ :2353E33 3333:3839. 2.353.333.53— 2353.23.53:— \/ \. mg;— mOmU— OAK-— /\ / .3352— : 2.»? 33.5533 _ 33.; 5353.33-33.33 332.533-3333 lX‘) REFERENCES IN) Shampay, J ., J.W. Szostak, and EH. Blackburn, DNA sequences oftelomeres maintained in yeast. Nature, 1984. 310(5973): p. 154-7. Louis, E]. and J .E. Haber, The structure and evolution o/‘subtelomeric Y ' repeats in Saccharomyces cerevisiae. Genetics, 1992. 131(3): p. 559-74. Louis, B]. and J .E. Haber, The subtelomeric Y ' repeatfitmilv in Saccharomt‘ces cerevisiae: an experimental systemfor repeated sequence evolution. Genetics, 1990. 124(3): p. 533-45. Zakian, V.A., Structure.junction. and replication ofSacc/zaromt'ces cerevisiae telomeres. Annu Rev Genet, 1996. 30: p. 141-72. Lundblad, V. and EH. Blackburn, An alternative pathway/oryeast telomere maintenance rescues estl- senescence. Cell, 1993. 73(2): p. 347-60. Lundblad, V. and J .W. Szostak, A mutant with a defect in telomere elongation leads to senescence in yeast. Cell, 1989. 57(4): p. 633-43. Lendvay, T.S., et al., Senescence mutants of Sacclmronuces cerevisiae with a defect in telomere replication identifi' three additional EST genes. Genetics, 1996. 144(4): p. 1399-412. Singer, MS. and DE. Gottschling, T LC 1: template RNA component of Saccharomyces cerevisiae telomerase. Science, 1994. 266(5184): p. 404-9. Chen, Q., A. ijma, and CW. Greider, Two survivor pathways that allow growth in the absence oftelomerase are generated by distinct telomere recombination events. Mol Cell Biol, 2001. 21(5): p. 1819-27. Le, S., et al., RAD50 and RAD5] de me two pat/mars that collaborate to maintain telomeres in the absence oftelomerase. Genetics, 1999. 152(1): p. 143- 52. l 90 1]. 13. 15. l6. l7. 19. 21. Teng, SC. and V.A. Zakian, T elomere-telomere recombination is an efficient bypass pathway for telomere maintenance in Saccharomyces cerevisiae. Mol Cell Biol, 1999. 19(12): p. 8083-93. Bryan, T.M., et al., E videncefor an alternative mechanism for maintaining telomere length in human tumors and tumor-derived cell lines. Nat Med, 1997. 3(11): p. 1271-4. Henson, J.D., et al., Alternative lengthening oftelomeres in mammalian cells. Oncogene, 2002. 21(4): p. 598-610. Reddel, R.R., T.M. Bryan, and J .P. Mumane, lmmortalized cells with no detectable telomerase activity. A review. Biochemistry (Mose), 1997. 62(1 1): p. 1254-62. Lundblad, V., Telomere maintenance without telomerase. Oncogene, 2002. 21(4): p. 522-31. Dunham, M.A., et al., T elomere nu’tintenance by recombination in human cells. Nat Genet, 2000. 26(4): p. 447-50. Yeager, T.R., et al., Telomerase-negative immortalized human cells contain a novel type ofpromyelocytic leukemia (PML) body. Cancer Res, 1999. 59(17): p. 4175-9. Asleson, E.N., R-J. Okagaki, and D.M. Livingston, A core activity associated with the N terminus of the yeast RAD52 protein is revealed by RAD5] overexpression suppression ofC—terminal rad52 truncation alleles. Genetics, 1999. 153(2): p. 681-92. Mortensen, U.H., et al., DNA strand annealing is promoted by the yeast Rad52 protein. Proc Natl Acad Sci U S A, 1996. 93(20): p. 10729-34. Shinohara, A., et al., Rad52forms ring structures and co-operates with RPA in single-strand DNA annealing. Genes Cells, 1998. 3(3): p. 145-56. Kagawa, W., et al., C rystal structure ofthe homologous-pairing domain from the human Rad52 recombinase in the undecamericfbrm. Mol Cell, 2002. 10(2): p. 359-71. 191 24. 25. 27. 29. 30. Cortes-Ledesma, F., F. Malagon, and A. Aguilera, A Novel Yeast Mutation, rad52-L89F, Causes 0 Specific Defect in Rad51-Independent Recombination That Correlates With a Reduced Ability of Rad52-L89F to Interact With Rad5 9. Genetics, 2004. 168(1): p. 553-7. Symington, L.S., Role of RAD52 epistasis group genes in homologous recombination and double—strand break repair. Microbiol Mol Biol Rev, 2002. 66(4): p. 630-70, table of contents. Davis, A.P. and LS. Symington, T heyeast recombinational repair protein Rad5 9 interacts with Rad52 and stimulates single-strand annealing. Genetics, 2001. 159(2): p. 515-25. Sugiyama, T., J .11. New, and SC. Kowalczykowski, DNA annealing by RAD52 protein is stimulated by specific interaction with the complex of replication protein A and single-stranded DNA. Proc Natl Acad Sci U S A, 1998. 95(1 1): p. 6049-54. ' Sugiyama, T. and SC. Kowalczykowski, Rad52 protein associates with replication protein A (RPA)-single-stranded DNA to accelerate Rad51-mediated displacement of RPA and presynaptic complex formation. J Biol Chem, 2002. 277(35): p. 31663-72. Shinohara, A. and T. Ogawa, Stimulation by Rad52 ofyeast Rad5] -n-1ediated recombination. Nature, 1998. 391(6665): p. 404-7. New, J.H., et al., Rad52 protein stimulates DNA strand exchange bt Rad5] and replication protein A. Nature, 1998. 391(6665): p. 407-10. Sung, P., Function of yeast Rad52 protein as a mediator between replication protein A and the Rad51 recombinase. J Biol Chem, 1997. 272(45): p. 28194-7. New, J .H. and SC. Kowalczykowski, Rad52 protein has a second stimulatmy role in DNA strand exchange that complements replication protein-A function. .1 Biol Chem, 2002. 277(29): p. 26171-6. Longtine, M.S., et al., Additional modules/or wrsatile and economical PC R- based gene deletion and modification in .S‘accharomt'ces cerevisiae. Yeast, 1998. 14(10): p. 953-61. 192 33. 35. 37. 38. 39. 40. 41. 42. Brachmann, C.B., et al., Designer deletion strains derived from Saccharon‘zyces cerevisiae SZ88C.‘ a useful set of strains and plasmids for PCR—mediated gene disruption and other applications. Yeast, 1998. 14(2): p. 1 15-32. Agatep, R., R.D. Kirkpatrick, D.L. Parchaliuk, R.A. Woods, and RD. Gietz, Transformation of Sacchromyces cerevisiae by the lithium acetate/single-stranded carrier DNA/polyethylene glycol (LiAc/ss-DNA/PE G ptotocol. Technical Tips Online (http://tto.trends.com), 1998. Mumberg, D., R. Muller, and M. Funk, Yeast vectors/or the controlled expression of heterologous proteins in different genetic backgrounds. Gene, 1995. 156(1): p. 119-22. Adzuma, K., T. Ogawa, and H. Ogawa, Primary structure of the RA D52 gene in Saccharomyces cerevisiae. Mol Cell Biol, 1984. 4(12): p. 2735-44. Cadwell, R.C.J., G.F., PCR primer- (1 labratoly manual, C.W.D. Dieffenbach, G. S., Editor. 1995. p. 583-589. Church, GM. and W. Gilbert, Genomic sequencing. Proc Natl Acad Sci U S A, 1984. 81(7): p. 1991-5. Mortensen, U.H., et al., A molecular genetic dissection of the evolutionarily conserved N terminus of yeast Rad52. Genetics, 2002. 161(2): p. 549-62. Lea, D.E., Coulson, C. A., The distribution in the numbers of'mutants in bacterial populations. Journal of genetics, 1949. 49: p. 264-285. Bai, Y., A.P. Davis, and LS. Symington, A novel allele ofRAD52 that causes severe DNA repair and recombination deficiencies only in the absence o_f'RAD5 l or RAD59. Genetics, 1999. 153(3): p. 1117-30. lra, G. and J .E. Haber, Characterization ofRAD5l-independent break-imluced replication that acts preferentiallv with short homologous sequences. Mol Cell Biol, 2002. 22(18): p. 6384-92. Smith, J. and R. Rothstein, A mutation in the gene encoding the Saccharomiices cerevisiae single-stranded DNA-bindin g protein Rfal stimulates a RAD52- 193 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. independent pathwayfor direct-repeat recombination. Mol Cell Biol, 1995. 15(3): p. 1632-4]. Smith, J. and R. Rothstein, An allele ofRFAl suppresses RAD52-dependent double-strand break repair in Saccharmnvces cerevisiae. Genetics, 1999. 151(2): p. 447-58. Griffith, J .D., et al., Mammalian telomeres end in a large duplex loop. Cell, 1999. 97(4): p. 503-14. Munoz-Jordan, J .1... et al., t-loops at tripanosome telomeres. Embo J, 2001. 20(3): p. 579-88. Li, B. and A.J. Lustig, A novel mechanismfor telomere size control in Saccharomyces cerevisiae. Genes Dev, 1996. 10(1 1): p. 1310-26. de Bruin, D., et al., T elomere looping permits gene activation by a downstream UAS in yeast. Nature, 2001. 409(6816): p. 109-13. de Bruin, D., et al., Telomerefolding is required for the stable maintenance of' telomere position ef‘ects in yeast. Mol Cell Biol, 2000. 20(21): p. 7991-8000. Grunstein, M., Molecular model/or telomeric heterochromatin in yeast. Curr Opin Cell Biol, 1997. 9(3): p. 383-7. Wang, R.C., A. Smogorzewska, and T. de Lange, Homologous recombination generates T-loop-sized deletions at human telomeres. Cell, 2004. 119(3): p. 355- 68. Bai, Y. and LS. Symington, A Rad52 homolog is required for RAD51- independent mitotic recombination in Saccharom)'ces cerevisiae. Genes Dev, 1996. 10(16): p. 2025-37. Davis, A.P. and LS. Symington, The Rad52-Rad5 9 complex interacts with Rad5] and replicationprotein A. DNA Repair (Amst), 2003. 2(10): p. 1 127-34. 1 94 MICHOGAN SYAYE UNIVERSITY LIBH AAAAA llifllllllllllllllljlllllllIllllzllllll