RecQ-like proteins are a family of DNA helicases that are evolutionary conserved from prokaryotes to eukaryotes. A large amount of experimental evidence suggests these proteins have a major role in the maintenance of genome stability. In humans five RecQ like helicase have been identified (RecQL1, BLM, WRN, RecQL4, and RecQL5), three of which are associated with rare genetic disorders with sever chromosomal and developmental abnormalities, and an elevated predisposition to cancer. Among the disease associated RecQ-like helicases, BLM and WRN have been subject to extensive research, while our collective knowledge about the function of RecQL4 is still very limited. Similarly, little is known about the role of RecQL1 and RecQL5 in maintenance of genome integrity. In the past studies of Sgs1, the Saccharomyces cerevisiae homolog of RecQ, have been very informative regarding BLM function. Thus, here we sought to further investigate BLM, RecQL1, and RecQL4 by using yeast as a model. By constructing humanized yeast strains we evaluated the ability of these genes to complement defects observed in sgs1∆. In Chapter 2, our investigation led to the development of a novel chimeric system, which was able to complement some defects of the sgs1∆ strain. In Chapter 3, by taking advantage of this chimeric system, we evaluated 41 known BLM variants identified in the general human population. This study resulted in identification of six novel variants that completely impaired BLM function and three novel variants that partially impaired BLM function. In Chapter 4 we conducted multiple yeast 2-hybrid screens in search for novel protein-protein interaction for RecQL1 and RecQL4. We have identified two new putatively interacting partners for RecQL1 and three putatively interacting partners for RecQL4. In Chapter 5, functional characterization of RecQL1, BLM, RecQL4 and RecQL5 in yeast suggested genetic interaction between BLM and RecQL4 and RecQL5. Finally, in Chapter 6 a random mutagenesis screen of BLM has led to identification a mutation that impairs BLM function by disrupting the HRDC domain. This mutant suggests that the HRDC domain of BLM has an important role in proper functionality of this helicase.

The RecQ family of helicases has been termed the “Caretakers of the Genome,” and rightfully so. These proteins are highly conserved from bacteria to humans and have been implicated in functions from homologous recombinatorial repair to damage checkpoint response to telomere maintenance and more. Mutant genes of three of the human RecQ helicases lead to syndromes characterized by a high incidence of cancer, premature aging and early death. Despite their implications in several biological functions and importance to the integrity of the human genome and suppression of cancer, many aspects of the RecQ family structure and function remain unknown. To date, much is known about the catalytic function of the helicase domain and accompanying domains, but considerably less is known about the non-catalytic N-terminus in these proteins, which, in many cases, including those human orthologs involved in disease, can make up about half of the total protein length. While experiments have been able to identify protein partners that interact with the N-terminal region, few are able to narrow the binding sites to minimally functional parts and fewer still describe any detail regarding the structural features of these binding areas. In fact, some reviews have generally described the N-terminus as “featureless,” a concept we challenge in our studies. Many of the N-termini of these RecQs have long been known to contain large stretches of acidic residues, a feature of intrinsically disordered regions. These regions/proteins are rich in charged and polar residues, lack compactness that makes crystallography possible, and have flexible and dynamic conformations that are prevalent in “high specificity, low affinity” interactions. Disordered proteins are well-known to be hot spots for protein/protein interactions and post-translational modifications, amongst other functions. Considering these facts, and recognizing the ties between these and what we know about the N-termini of the RecQs, we hypothesized that these proteins likely have long disordered termini. In Chapter 3, we confirm the presence of disorder at the Top3/Rmi1 binding site on Sgs1, the Saccharomyces cerevisiae RecQ helicase. We show that even in a disordered state, this binding region is not “featureless,” but in fact contains a transient alpha-helical molecular recognition element that is necessary to facilitate complex formation between Sgs1, Top3 and Rmi1. Loss of helical structure at this site leads to increased genomic instability and sensitivity to DNA damaging agents. Based on these results, we suggest that there are likely many more such elements in the N-terminus that that are important for other Sgs1 protein/protein interactions and provide an estimate for the number of interactions in this region. In Chapter 4, we evaluate the prevalence of disorder in a set of Chromatin Processes proteins in an effort to establish a role for disorder with regards to maintaining chromatin integrity. In our bioinformatics study, we found that disorder is overrepresented in the Chromatin Processes proteins, and that a major driving force for disorder in these proteins is protein/protein interaction and post-translational modification. We also show a biological connection to disorder and increased protein/protein interaction by investigating these parameters in the context of the DNA damage checkpoint response and in complex formations. Mediators between highly structured kinases in the checkpoint were the most interactive proteins and over half of all predicted interaction sites occurred in disordered areas. Complexed proteins often contained one protein with a high number of disordered sites and a high number of predicted interactions, while the rest were considerably more ordered. Chapter 5 explores a Sgs1 interaction partner, Rmi1 and uses bioinformatics to design structurally-based point mutations in an effort to further elucidate Rmi1 function in yeast, which remains largely unknown outside of its enhancement of Top3/Sgs1 catalytic function. Using AGADIR, which predicts alpha-helical structure and is particularly useful in our hands for guided-mutagenesis in disordered regions, we identified several point mutations that lead to Δrmi1 phenotypes or intermediate growth on hydroxyurea. We hypothesize that these mutants are important in maintaining Rmi1 stability. Together, these studies suggest an important change in how the field approaches further studies into the RecQ helicases; traditional methods of primary sequence comparisons and crystal structures limit the study of disordered regions that are still functionally important. Future care should be given to consider the conservation of structure or structural elements in the RecQs over strict alignments when comparing functional regions between orthologs. Our studies also suggest that it is highly likely that structural motifs for important protein interactions in RecQs are being overlooked because they are not readily obvious using traditional methods. By understanding these motifs and the interactions they facilitate, we may be able to more easily identify polymorphisms in patients with genomically unstable conditions like cancer and, having better understood the biological process these structures facilitate, design drugs to counteract detrimental effects.

<p>The genetic information necessary for the survival and propagation of a species is contained within a physical structure, DNA. However, this molecule is sensitive to damage arising from both exogenous and endogenous sources. DNA damage can prevent metabolic processes such as replication and transcription; thus, systems to bypass or repair DNA lesions are essential. One type of lesion in particular - the double strand break (DSB) - is extremely dangerous as inappropriate repair of DSBs can lead to deletions, mutations and rearrangements. Homologous recombination (HR) uses a template with sequence homology to the region near the DSB to restore the damaged molecule. However, this high-fidelity pathway can contribute to genome instability when recombination occurs between diverged substrates. To further our understanding of the regulation of HR during vegetative growth, we have used the budding yeast Saccharomyces cerevisiae as a model system and a plasmid-based assay to model repair of a DSB. In the first part of this work, the molecular structures of noncrossover (NCO) and crossover (CO) products of recombination were examined. While the majority of NCOs had regions of heteroduplex DNA (hDNA) on one side of the gap in the repaired allele and no change to the donor allele, most COs had two tracts of hDNA. They were present on opposite sides of the gap, one in each allele. Our results suggest that the majority of NCOs are generated through synthesis-dependent strand annealing (SDSA), and COs are the result of constrained cleavage of a Holliday junction (HJ) intermediate. To clarify the mechanisms regulating NCO production, the effects of three DNA helicases - Mph1, Sgs1 and Srs2 - on the structures of NCO events were examined. All three helicases promote NCO formation by SDSA, but Sgs1 and Srs2 also assist in NCO formation arising from an HJ-containing intermediate, consistent with HJ-dissolution. To study how CO products are generated, we have investigated the contribution of the following candidate HJ resolvases to the structures of CO events: Mus81, Yen1 and Rad1. The results suggest that Rad1 is important to normal CO formation in this assay, but Mus81 and Yen1 are largely dispensable. Together, this work advances our knowledge of how the NCO versus CO outcome is determined during HR, expanding our understanding of how mitotic recombination is regulated.</p> / Dissertation

Genome instability is a hallmark of human cancers. Patients with Bloom’s syndrome, a rare chromosome breakage syndrome caused by inactivation of the RecQ helicase BLM, result in phenotypes associated with accelerated aging and develop cancer at a very young age. Patients with Bloom’s syndrome exhibit hyper-recombination, but the role of BLM and increased genomic instability is not fully characterized. Sgs1, the only member of the RecQ family of DNA helicases in Saccharomyces cerevisiae, is known to act both in early and late stages of homology-dependent repair of DNA damage. Exo1, a 5′–3′ exonuclease, first discovered to play a role in mismatch repair has been shown to participate in parallel to Sgs1 in processing the ends of DNA double-strand breaks, an early step of homology-mediated repair. Here we have characterized the genetic interaction of SGS1 and EXO1 with other repair factors in homology-mediated repair as well as DNA damage checkpoints, and characterize the role of post-translational modifications, and protein-protein interactions in regulating their function in response to DNA damage. In S. cerevisiae cells lacking Sgs1, spontaneous translocations arise by homologous recombination in small regions of homology between three non-allelic, but related sequences in the genes CAN1, LYP1, and ALP1. We have found that these translocation events are inhibited if cells lack Mec1/ATR kinase while Tel1/ATM acts as a suppressor, and that they are dependent on Rad59, a protein known to function as one of two sub-pathways of Rad52 homology-directed repair. Through a candidate screen of other DNA metabolic factors, we identified Exo1 as a strong suppressor of chromosomal rearrangements in the sgs1∆ mutant. The Exo1 enzymatic domain is located in the N-terminus while the C-terminus harbors mismatch repair protein binding sites as well as phosphorylation sites known to modulate its enzymatic function at uncapped telomeres. We have determined that the C-terminus is dispensable for Exo1’s roles in resistance to DNA-damaging agents and suppressing mutations and chromosomal rearrangements. Exo1 has been identified as a component of the error-free DNA damage tolerance pathway of template switching. Exo1 promotes template switching by extending the single strand gap behind stalled replication forks. Here, we show that the dysregulation of the phosphorylation of the C-terminus of Exo1 is detrimental in cells under replication stress whereas loss of Exo1 suppresses under the same conditions, suggesting that Exo1 function is tightly regulated by both phosphorylation and dephosphorylation and is important in properly modulating the DNA damage response at stalled forks. It has previously been shown that the strand exchange factor Rad51 binds to the C-terminus of Sgs1 although the significance of this physical interaction has yet to be determined. To elucidate the function of the physical interaction of Sgs1 and Rad51, we have generated a separation of function allele of SGS1 with a single amino acid change (sgs1-FD) that ablates the physical interaction with Rad51. Alone, the loss of the interaction of Sgs1 and Rad51 in our sgs1-FD mutant did not cause any of the defects in response to DNA damaging agents or genome rearrangements that are observed in the sgs1 deletion mutant. However, when we assessed the sgs1-FD mutant in combination with the loss of Sae2, Mre11, Exo1, Srs2, Rrm3, and Pol32 we observed genetic interactions that distinguish the sgs1-FD mutant from the sgs1∆mutant. Negative and positive genetic interactions with SAE2, MRE11, EXO1, SRS2, RRM3, and POL32 suggest the role of the physical interaction of Sgs1 and Rad51 is in promoting homology-mediated repair possibly by competing with single-strand binding protein RPA for single-stranded DNA to promote Rad51 filament formation. Together, these studies characterize additional roles for domains of Sgs1 and Exo1 that are not entirely understood as well as their roles in combination with DNA damage checkpoints, and repair pathways that are necessary for maintaining genome stability.

Double-strand breaks and stalled forks arise when the replication machinery encounters damage from exogenous sources like DNA damaging agents or ionizing radiation, and require specific DNA helicases to resolve these structures. Sgs1 of Saccharomyces cerevisiae is a member of the RecQ family of DNA helicases and has a role in DNA repair and recombination. The RecQ family includes human genes BLM, WRN, RECQL4, RECQL1, and RECQL5. Mutations in BLM, WRN, and RECQL4 result in genetic disorders characterized by developmental abnormalities and a predisposition to cancer. All RecQ helicases have common features including a helicase domain, an RQC domain, and a HRDC domain. In order to elucidate the role of these domains and to identify additional regions in Sgs1 that are required for the maintenance of genome integrity, a series of systematic truncations to the C terminus of Sgs1 were created. We found that ablating the HRDC domain does not cause an increase in accumulating gross chromosomal rearrangements (GCRs). But deleting the RQC domain and leaving the helicase domain intact resulted in a rate similar to that of a helicase-defective mutant. Additionally, we exposed these truncation mutants to HU and MMS and demonstrated that losing up to 200 amino acids from the C terminus did not increase sensitivity to HU or MMS, whereas losing 300 amino acids or more led to sensitivity similar to that of an sgs1∆ cell. These results suggest that the RQC domain, believed to mediate protein-protein interactions and required for DNA recognition, is important for Sgs1’s role in suppressing GCRs and sensitivity to HU and MMS, whereas the HRDC domain that is important for DNA binding is not necessary. RecQL5 is a RecQ-like helicase that is distinct from the other members through its three different isoforms, RecQL5α, RecQL5β, and RecQL5ɣ. It has a helicase domain and an RQC domain, but lacks the HRDC domain that other RecQ-like helicases possess. In contrast to Blm, Wrn, and RecQL4, no human disorder has been associated with defects in RecQL5. For this reason the role of RecQL5 in the cell has remained largely unknown. To try to elucidate the pathways RecQL5 may be involved in we performed a yeast two hybrid to identify RecQL5-interacting proteins. We found that RecQL5 interacts with Hlp2, an ATP-dependent RNA helicase, and Ube2I, a SUMO-conjugating enzyme. These novel interactions shed light on a potential role of RecQL5 in the cell as a transcriptional regulator. Saccharomyces cerevisiae, Rrm3, is a 5’-3’ DNA helicase that is part of the Pif1 family of DNA helicases and is conserved from yeast to humans. It was initially discovered as a suppressor of recombination between tandem arrays and ribosomal DNA (rDNA) repeats. In its absence there are increased rates of extra-chromosomal rDNA circles, and cells accumulate X-shaped intermediates at stalled forks. Rrm3 may be involved in displacing DNA-protein blocks and unwinding DNA to facilitate fork progression. We used stable isotope labeling by amino acids in cell culture (SILAC)- based quantitative mass spectrometry in order to determine proteins that deal with the stalled fork in the absence of Rrm3. We found that in the absence of Rrm3 and increased replication fork pausing, there is a requirement for the error-free DNA damage bypass factor Rad5 and the homologous recombination factor Rdh54 for fork recovery. We also report a novel role for Rrm3 in controlling DNA synthesis upon exposure to replication stress and that this requirement is due to interaction with Orc5, a subunit of the origin recognition complex. Interaction of Orc5 was found to be located within a 26-residue region in the unstructured N-terminal tail of Rrm3 and loss of this interaction resulted in lethality with cells devoid of the replication checkpoint mediator Mrc1, and DNA damage sensitivity with cells lacking Tof1. In this study we describe two independent roles of Rrm3, a helicase-dependent role that requires Rad5 and Rdh54 for fork recovery, and a helicase-independent role that requires Orc5 interaction to control DNA synthesis. Our data provides novel insight into the role of DNA helicases and their role in protecting the genome. Through yeast genetics it was possible to determine the importance of the C terminus of Sgs1 and elucidate new RecQL5 interacting partners that shed light onto roles for RecQL5 distinct from other RecQ like helicases. Quantitative mass spectrometry allowed us to take on a more global view of the cell and determine how it responds to replication fork pausing in the absence of Rrm3. Using both proteomics and yeast genetics we were able to better understand how these DNA helicases contribute to maintaining genome stability.

In Saccharomyces cerevisiae DNA postreplication repair (PRR) functions to bypass replication-blocking lesions to prevent damage-induced cell death. PRR employs two different mechanisms to bypass damaged DNA. While translesion synthesis (TLS) has been well characterized, little is known about the molecular events involved in error-free bypass although it has been assumed that homologous recombination (HR) is required for such a mode of lesion bypass. We undertook a genome-wide, synthetic genetic array (SGA) screen for novel genes involved in PRR and observed evidence of genetic interactions between error-free PRR and HR. We were screening for synthetic lethality which occurs when the combination of two mutations leads to an inviable organism, however, either single mutation allows for cell viability. In addition, we screened for conditionally synthetic lethal interaction which occurs when the combination of two mutations is inviable only in the presence of a DNA-damaging agent. This screen identified and assigned four genes, CSM2, PSY3, SHU1 and SHU2, whose products form a stable Shu complex, to the error-free PRR pathway. Previous studies have indicated that the Shu complex is required for efficient HR and that inactivation of any one of these genes is able to suppress the severe phenotypes of top3 and sgs1. We confirmed and further extended some of the reported observations and demonstrated that error-free PRR mutations are also epistatic to sgs1. Based on the above analyses, we propose a model in which error-free PRR utilizes the Shu complex to recruit HR to facilitate template switching, followed by double-Holliday junction resolution by Sgs1-Top3. Null mutations of HR genes including rad51, 52, 54, 55 and 57 are known to confer characteristic synergistic interactions with TLS mutations. To our surprise, null mutations of genes encoding the Mre11-Rad50-Xrs2 (MRX) complex, which is also required for HR, are epistatic to TLS mutations. The MRX complex confers an endo/exonuclease activity required for the detection and processing of DNA double-strand breaks (DSBs). Our results suggest that the MRX complex functions in both TLS and error-free PRR and that this function requires the nuclease activity of Mre11. This is in sharp contrast to other known HR genes that only function downstream of error-free PRR. Furthermore, we found that inactivation of SGS1 significantly inhibits proliferating cell nuclear antigen (PCNA) monoubiquitination and is epistatic to mutations in TLS, suggesting that Sgs1 also functions at earlier steps in DNA lesion bypass. We also examined the roles of Sae2 and Exo1, two accessory nucleases involved in DSB resection, in PRR. We found that while Sae2 is primarily required for TLS, Exo1 is exclusively involved in error-free PRR. In light of the distinct and overlapping activities of the above nucleases in the resection of DSBs, we propose that the distinct single-strand nuclease activities of MRX, Sae2 and Exo1 dictate the preference between TLS and error-free PRR for lesion bypass. While both PRR pathways are dependent on the ubiquitination of PCNA, error-free PRR utilizes non-canonical Lys63-linked polyubiquitinated PCNA to signal lesion bypass. This mechanism is dependent on the Mms2-Ubc13 complex being in close proximity to PCNA, a process thought to be dependent on Rad5. Rad5 is a member of the SWI/SNF family of ATPases that contains a RING finger motif characteristic of an E3 Ub ligase. Previous in vitro experiments demonstrated the ability of Rad5 to promote replication fork regression, a function dependent on its helicase/ATPase activity. We therefore created site-specific mutants defective in either Rad5 RING finger or helicase/ATPase activity, or both, in order to examine their genetic interactions with known TLS and error-free PRR genes. Our results indicate that both the Rad5 RING finger motif and the helicase/ATPase activity are exclusively involved in error-free PRR. To our surprise, like the Rad5 RING finger, lack of the helicase/ATPase activity also abolishes the Lys63-linked PCNA polyubiquitin chain formation, suggesting that either the Rad5 helicase/ATPase-promoted replication fork regression signals PCNA polyubiquitination or this domain has a yet unidentified activity. In summary, results obtained from this thesis dissertation have revealed novel mechanisms of yeast PRR in S. cerevisiae, a mechanism that appears to be evolutionarily conserved throughout eukaryotes, from yeast to humans.

In Saccharomyces cerevisiae DNA postreplication repair (PRR) functions to bypass replication-blocking lesions to prevent damage-induced cell death. PRR employs two different mechanisms to bypass damaged DNA. While translesion synthesis (TLS) has been well characterized, little is known about the molecular events involved in error-free bypass although it has been assumed that homologous recombination (HR) is required for such a mode of lesion bypass. We undertook a genome-wide, synthetic genetic array (SGA) screen for novel genes involved in PRR and observed evidence of genetic interactions between error-free PRR and HR. We were screening for synthetic lethality which occurs when the combination of two mutations leads to an inviable organism, however, either single mutation allows for cell viability. In addition, we screened for conditionally synthetic lethal interaction which occurs when the combination of two mutations is inviable only in the presence of a DNA-damaging agent. This screen identified and assigned four genes, CSM2, PSY3, SHU1 and SHU2, whose products form a stable Shu complex, to the error-free PRR pathway. Previous studies have indicated that the Shu complex is required for efficient HR and that inactivation of any one of these genes is able to suppress the severe phenotypes of top3 and sgs1. We confirmed and further extended some of the reported observations and demonstrated that error-free PRR mutations are also epistatic to sgs1. Based on the above analyses, we propose a model in which error-free PRR utilizes the Shu complex to recruit HR to facilitate template switching, followed by double-Holliday junction resolution by Sgs1-Top3. Null mutations of HR genes including rad51, 52, 54, 55 and 57 are known to confer characteristic synergistic interactions with TLS mutations. To our surprise, null mutations of genes encoding the Mre11-Rad50-Xrs2 (MRX) complex, which is also required for HR, are epistatic to TLS mutations. The MRX complex confers an endo/exonuclease activity required for the detection and processing of DNA double-strand breaks (DSBs). Our results suggest that the MRX complex functions in both TLS and error-free PRR and that this function requires the nuclease activity of Mre11. This is in sharp contrast to other known HR genes that only function downstream of error-free PRR. Furthermore, we found that inactivation of SGS1 significantly inhibits proliferating cell nuclear antigen (PCNA) monoubiquitination and is epistatic to mutations in TLS, suggesting that Sgs1 also functions at earlier steps in DNA lesion bypass. We also examined the roles of Sae2 and Exo1, two accessory nucleases involved in DSB resection, in PRR. We found that while Sae2 is primarily required for TLS, Exo1 is exclusively involved in error-free PRR. In light of the distinct and overlapping activities of the above nucleases in the resection of DSBs, we propose that the distinct single-strand nuclease activities of MRX, Sae2 and Exo1 dictate the preference between TLS and error-free PRR for lesion bypass. While both PRR pathways are dependent on the ubiquitination of PCNA, error-free PRR utilizes non-canonical Lys63-linked polyubiquitinated PCNA to signal lesion bypass. This mechanism is dependent on the Mms2-Ubc13 complex being in close proximity to PCNA, a process thought to be dependent on Rad5. Rad5 is a member of the SWI/SNF family of ATPases that contains a RING finger motif characteristic of an E3 Ub ligase. Previous in vitro experiments demonstrated the ability of Rad5 to promote replication fork regression, a function dependent on its helicase/ATPase activity. We therefore created site-specific mutants defective in either Rad5 RING finger or helicase/ATPase activity, or both, in order to examine their genetic interactions with known TLS and error-free PRR genes. Our results indicate that both the Rad5 RING finger motif and the helicase/ATPase activity are exclusively involved in error-free PRR. To our surprise, like the Rad5 RING finger, lack of the helicase/ATPase activity also abolishes the Lys63-linked PCNA polyubiquitin chain formation, suggesting that either the Rad5 helicase/ATPase-promoted replication fork regression signals PCNA polyubiquitination or this domain has a yet unidentified activity. In summary, results obtained from this thesis dissertation have revealed novel mechanisms of yeast PRR in S. cerevisiae, a mechanism that appears to be evolutionarily conserved throughout eukaryotes, from yeast to humans.