The cell division of Saccharomyces cerevisiae is controlled by the action of pheromones at the G1 phase of the cell cycle. A general method was developed for the isolation of constitutive mutants in the pheromone response pathway. Recessive alleles of the SCG1 gene (encoding the α subunit of a G protein) were isolated as well as a dominant mutation in the STE4 gene (encoding the β subunit of a G protein). Analysis of double mutants suggested that the STE4 gene product functions after the SCG1 product but before the STE5 gene product. Double mutants carrying either scg1 or STE4Hp1 constitutive alleles together with the temperature-sensitive unresponsive mutation, ste5-3ts, showed arrest and recovery when shifted from 34° C to 22° C. Recovery from the constitutive signal was independent of the receptor. The STE4Hp1 sst2 ste5ts triple mutant was not able to recover from arrest, suggesting that an SST2-dependent mechanism is involved in recovery of the STE4Hp1 mutant from constitutive arrest. In contrast, the scg1-7 sst2 ste5ts triple mutant recovered only partially suggesting that even though SST2 gene product is probably involved in recovery of the scg1-7 mutant, this mutant can recover by an SST2-independent mechanism. This implies existence of another, SST2-independent postreceptor recovery mechanism. The scg1-null mutant do not recover from constitutive arrest (J. Hirschman, personal communication). Both recovery mechanisms probably operate at the G protein step. Isolation of a constitutive allele of STE5 allowed the definition of its site of action as being after the STE4-controlled step. In addition, constitutive activation of the pheromone pathway by STE5Hp1 mutation was found to be partially dependent on the STE4 and STE18 gene products, the β and γ subunits of a G protein. A comprehensive genetic model is presented to explain the mechanisms of signal transduction and recovery.

Chromatin structure plays an essential role in the regulation of many nuclear processes such as transcription, replication, recombination, and repair. It is generally accepted that chromatin remodeling is a prerequisite step in gene activation. Over recent years, large multisubunit enzymes that regulate the accessibility of nucleosomal DNA have emerged as key regulators of eukaryotic transcription. It seems likely that similar enzymes contribute to the efficiency of DNA replication, recombination, and repair. These chromatin remodeling complexes can be classified into two broad groups: (1) the ATP-dependent enzymes, which utilize the energy of ATP hydrolysis to increase the accessibility of nucleosomal DNA; and (2) histone modifying enzymes that phosphorylate, acetylate, methylate, ubiquitinate, or ADP-ribosylate the nucleosomal histones (for review see Kingston and Narlikar, 1999; Muchardt and Yaniv, 1999; Brown et al., 2000; Vignali et al., 2000; Strahl and Allis, 2000). The mechanism by which these two groups of large, multi-subunit enzymes function to alter chromatin structure is enigmatic. Studies suggest that ATP-dependent and histone acetyltransferase chromatin remodeling enzymes have widespread roles in gene expression and perform both independent and overlapping functions. Interestingly, although both groups of enzymes appear to be distinct, several features of these enzymes have been conserved from yeast to man. Thus, understanding the role of these similar features will be essential in order to elucidate the function of remodeling enzymes, their functional interrelationships, and may uncover the fundamental principals of chromatin remodeling. In this study, we use a combination of yeast molecular genetics and biochemistry to dissect out the function of individual parts of these chromatin remodeling machines and to understand how these large macromolecular assemblies are put together. In addition, we also investigate the mechanism by which the ATP-dependent enzymes exert their regulatory effects on chromatin structure. Structure/function analysis of Saccharomyces cerevisiaeSwi3p (conserved in SWI/SNF complexes across all eukaryotic phyla) reveals a unique scaffolding role for this protein as it is essential for assembly of SWI/SNF subunits. We have also characterized a novel motif that has homology to the Myb DNA binding domain, the SANT domain, and that is shared among transcriptional regulatory proteins implicated in chromatin remodeling. Mutational analysis of this domain in yeast Swi3p (SWI/SNF), Rsc8/Swh3p (RSC), and Ada2p (GCN5 HATs) reveals an essential function for the SANT domain in chromatin remodeling. Moreover, our studies suggest that this novel motif may be directly involved in mediating a functional interaction with chromatin components (i.e. histone amino terminal domains). We have also directly compared the activities of several members of the ATP-dependent chromatin remodeling enzymes. Surprisingly, we find that these enzymes utilize similar amounts of ATP to increase nucleosomal DNA accessibility. In as much, we show that changes in histone octamer comformation or composition is not a requirement or consequence of chromatin remodeling by SWI/SNF. Taken together, these data suggest a similar mechanism for ATP-utilizing chromatin remodeling enzymes in which disruption of histone-DNA contacts occur without consequence to the structure of the histone octamer. These data have striking implications for how we view the mechanism of chromatin remodeling.

CP1 is a sequence specific DNA-binding protein of the yeast Saccharomyces cerevisiae which recognizes the highly conserved centromere DNA element I (CDEI) of yeast centromeres. The gene encoding CP1, which was designated CEP1 for centromere protein 1, was cloned and sequenced. CEP1 encodes a highly acidic protein of molecular weight 39,400. CEP1 was mapped to a position 4.6 centiMorgans centromere distal to SUP4 on the right arm of chromosome X. Phenotypic analysis of cep1 mutants demonstrated that yeast strains lacking CP1 are viable but have a 35% increase in cell doubling time, a ninefold increase in the rate of mitotic chromosome loss, and are methionine auxotrophs. Detailed analysis of the mitotic chromosome-loss phenotype showed that the loss is primarily due to chromosome nondisjunction (2:0 segregation). During meiosis cep1 null mutants exhibited aberrant segregation of centromere containing plasmids, chromosome fragments, and chromosomes. The predominant missegregation event observed was precocious sister segregation. The mutants also displayed a nonrandom 20% decrease in spore viability. Missegregation of chromosomes accounted for some but not all of this decreased spore viability, the remainder of which is presumed to be related to the pleiotropic consequences of the cep1 mutation. Together with the observed mitotic missegregation phenotype the results are interpreted as suggesting that CP1 promotes sister chromatid-kinetochore adhesion. The following conclusions are based on my mutational analysis of CP1: (1) CP1 is normally present in functional excess, (2) the C-terminal 143 amino acids are sufficient for full CP1 function in chromosome segregation and methionine metabolism, and (3) while DNA binding is apparently necessary for function, DNA binding per se is not sufficient. All of the mutations which caused an observable phenotype affected both centromere function and methionine metabolism. In addition, a direct correlation was observed in the degree to which both phenotypes were affected by different mutations. None of the mutant proteins displayed trans-dominant effects in a wild type background; however, two nonfunctional DNA binding-competent mutants exerted a dominant negative effect on the ability of PHO4 to suppress cep1 methionine auxotrophy. The data are consistent with a model in which CP1 performs a similar function at centromeres and promoters.

The work presented here describes novel functions for centrosome proteins, specifically for pericentrin and centriolin. The first chapter describes the involvement of pericentrin in ciliogenesis. Cells with reduced pericentrin levels were unable to form primary cilia in response to serum starvation. In addition we showed novel interactions between pericentrin, intraflagellar transport (IFT) proteins and polycystin 2 (PC2). Pericentrin was co-localized with IFT proteins and PC2 to the base of primary cilia and motile cilia. Ciliary function defects have been shown to be involved in many human diseases and IFT proteins and PC2 have been implicated in these diseases. We conclude that pericentrin is required for assembly of primary cilia possibly as an anchor for other proteins involved in primary cilia assembly. The second chapter describes identification of centriolin, a novel centriolar protein that localizes to subdistal appendages and is involved in cytokinesis and cell cycle progression. Depletion of centriolin leads to defects in the final stages of cytokinesis, where cells remain connected by thin intercellular bridges and are unable to complete abscission. The cytokinesis defects seemed to precede the G0/G1 p53 dependant cell cycle arrest. Finally, the third chapter is a continuation of the cytokinesis study and it identifies pericentrin as an interacting partner for centriolin. Like centriolin, pericentrin knockdown induces defects in the final stages of cytokinesis and leads to G0/G1 arrest. Moreover, pericentrin and centriolin interact biochemically and show codependency in their centrosome localization. We conclude that pericentrin and centriolin are members of the same pathway and are necessary for the final stages of cytokinesis.

Nonsense-mediated mRNA decay (NMD) is an important mRNA surveillance mechanism conserved in eukaryotes. This thesis explores several interesting aspects of the NMD pathway. One important aspect of NMD which is presently the subject of intense controversy is the subcellular localization of NMD. In one set of experiments, the decay kinetics of the ade2-1 and pgk1 nonsense mRNAs (substrates for NMD) were investigated in response to activating the NMD pathway to determine if cytoplasmic nonsense mRNAs are immune to NMD in the yeast system. The results of these studies demonstrated that activation of NMD caused rapid and immediate degradation of both the ade2-1 and the early nonsense pgk1 steady state mRNA populations. The half lives of the steady state mRNA populations for both ade2-1 and pgk1 (early nonsense) were shortened from >30 minutes to approximately 7 minutes. This was not observed for pgk1mRNAs that contained a late nonsense codon demonstrating that activation of NMD specifically targeted the proper substrates in these experiments. Therefore, in yeast, nonsense mRNAs residing in the cytoplasm are susceptible to NMD. While these findings are consistent with NMD occurring in the cytoplasm, they do not completely rule out the possibility of a nuclear-associated decay mechanism. To investigate the involvement of the nucleus in NMD, the putative nuclear targeting sequence identified in Nmd2p (one of the trans-acting factors essential for NMD) was characterized. Subcellular fractionation experiments demonstrated that the majority of Nmd2p localized to the cytoplasm with a small proportion detected in the nucleus. Specific mutations in the putative nuclear localization signal (NLS) of Nmd2p were found to have adverse effects on the protein's decay function. These effects on decay function, however, could not be attributed to a failure in nuclear localization. Therefore, the residues that comprise the putative NLS of Nmd2p are important for decay function but do not appear to be required for targeting the protein to the nucleus. These results are in accordance with the findings above which implicate the cytoplasm as an important cellular compartment for NMD. This thesis then investigates the regulatory roles of the trans-acting factors involved in NMD (Upf1p, Nmd2p, and Upf3p) using a novel quantitative assay for translational suppression, based on a nonsense allele of the CAN1 gene (can1-100). Deletion of UPF1, NMD2, or UPF3 stabilized the can1-100 transcript and promoted can1-100 nonsense suppression. Changes in mRNA levels were not the basis of suppression, however, since deletion of DCP1 or XRN1 or high-copy can1-100 expression in wild-type cells caused mRNA stabilization similar to that obtained in upf/nmd cells but did not result in comparable suppression. can1-100 suppression was highest in cells harboring a deletion of UPF1, and overexpression of UPF1 in cells with individual or multiple upf/nmd mutations lowered the level of nonsense suppression without affecting the abundance of the can1-100 mRNA. These findings indicate that Nmd2p and Upf3p regulate Upf1p activity and that Upf1p plays a critical role in promoting termination fidelity that is independent of its role in regulating mRNA decay.

The overall emphasis in this thesis is the study of the regulation of virus infections by natural killer (NK) cells. In initial analyses, vaccinia virus (VV)-infected cells were found to be more sensitive to NK cell-mediated lysis during a discrete period of time post-infection. This enhanced susceptibility to lysis correlated with enhanced triggering (but not binding) of the effector cells and a concomitant decrease in target cell H-2 class I antigen expression. Furthermore, VV-infected cells became resistant to lysis by allospecific cytotoxic T lymphocytes (CTL) at a time when they were very sensitive to killing by NK cells or VV-specific CTL. This suggested that alterations in class I MHC antigens may affect target cell sensitivity to lysis by NK cells. The hypothesis that viral peptide charging of H-2 class I molecules can modulate target cell sensitivity to NK cell-mediated lysis was tested by treating target cells with synthetic viral peptides corresponding to the natural or minimal immunodominant epitopes defined for virus-specific CTL, and then target cell susceptibility to NK cell-mediated lysis was assessed. None of the 12 synthetic viral peptides used were able to significantly alter target cell lysis by NK cells under any of the conditions tested. In order to determine if H-2 class I molecules were required in the regulation of a virus infection by NK cells in vivo, intact or NK depleted (treated with anti-asialo GM1 antiserum) β2-microglobulin-deficient [β2m (-/-)] mice, which possess a defect in H-2 class I antigen expression, were infected with the prototypic NK-sensitive virus, murine cytomegalovirus (MCMV). In anti-asialo GM1-treated β2m (-/-) mice, as well as in β2m + (H-2 class I normal) control mice also treated with anti-asialo GM1 a significant enhancement in splenic MCMV titers as compared to NK-intact animals, was observed. When thymocyte expression of H-2 class I molecules (H-2Db) in normal mice was analyzed, it was found that following MCMV infection, H-2Db expression was significantly greater than the low level of expression found in uninfected thymocytes. In marked contrast, thymocytes from β2m (-/-) mice did not display any detectable H-2Db before or after infection. These in vivoresults demonstrate that NK cells can regulate a virus infection, at least in the case of MCMV, independent of H-2 class I molecule expression. Thymocytes from uninfected normal mice were found to be very sensitive to NK cell-mediated lysis, whereas those from MCMV-infected animals were completely resistant, presumably due to the protective effects of MCMV-induced interferon (IFN). However, thymocytes from MCMV-infected β2m (-/-) mice were only slightly protected from lysis by NK cells, consistent with the inverse correlation between MHC class I antigen expression and sensitivity to NK cell-mediated lysis. These results provide in vivoevidence suggesting a requirement for MHC class I molecules in IFN-mediated protection from lysis by NK cells. In addition to the analysis of H-2 class I molecules on target cells, the identity of a molecule present on the surface of all NK cells and other cytotoxic effector cells, which is recognized by a monoclonal antibody (mAb) generated in this laboratory designated CZ-1, and can also modulate NK cell triggering, was also of interest. This laboratory has previously reported that this antigen is upregulated on cytotoxic (and other) lymphocytes following a virus infection in vivo, or upon activation in vitro. Using competitive FACS analysis and fibroblasts transfected with various isoforms of CD45, it was found that mAb CZ-1 recognizes a sialic acid-dependent epitope associated with a subpopulation of CD45RB molecules.

P22 lysozyme, encoded by gene 19, is an essential phage protein responsible for hydrolyzing the bacterial cell wall during lytic infection. P22 lysozyme is related to T4 lysozymein its mode of action, substrate specificities, and in its structure. Gene 19 was located on the phage genome, subcloned, and then sequenced. lysozyme was produced in large quantities and purified for biochemical characterization and for crystallograpic studies. Gene 19consists of 146 codons, and encodes a protein with a molecular weight of 16,117. Amber mutations were created in gene 19 by in vitro primer-directed mutagenesis. The mutations were crossed by homologous recombination onto the phage genome. The phages bearing the amber mutations in gene 19 were screened for the ability to grow on six different amber suppressor strains. Amino acid substitutions that resulted in nonfunctional or less functional lysozyme were determined. Of 60 possible amino acid substitutions at 11 different sites in P22 lysozyme, 20 are deleterious. The phage bearing amber mutations in gene 19that failed to grow on given suppressor strains were reverted and second site intragenic revertants were obtained. The mutations were sequenced. A substitution of serine for glutamine at residue 82 is compensated for by changing residue 46 from serine to leucine. This single change enables the phage to form a plaque at 300C but not at 400C. When the triple change asn42->lys; ser46->leu; and ser43->pro is present the lysozyme produced is no longer temperature sensitive. The crystal structure of P22 lysozyme is not yet solved. Assuming that the structures of T4 lysozyme and P22 lysozyme are similar, one can examine the positions of equivalent residues in the T4 lysozyme structure. The spatial arrangement of the residues changed by the secondary site mutations and the original substitution can then be visualized. The mutations discussed above all map far from the original mutation on the T4 three dimensional model. A substitution of leucine for tyrosine at position 22 is compensated for by the double mutation of arg18->ser and ser23->lys. When the equivalent residues are mapped on the T4 three dimensional model the changes map in close proximity to the original mutation.

The human erythrocyte glucose transport protein (GLUT1) interacts with, and is regulated by, cytosolic ATP. This study asks the following questions concerning ATP modulation of GLUT1 mediated sugar transport. 1) Which region(s) of GLUT1 form the adenine nucleotide-binding domain? 2) What factors influence ATP modulation of sugar transport? 3) Is ATP interaction with GLUT1 sufficient for sugar transport regulation? The first question was addressed through peptide mapping, n-terminal sequencing, and alanine scanning mutagenesis of GLUT1 using [32P]-azidoATP, a photoactivatable ATP analog. We then used a combination of transport measurements and photolabeling strategies to examine how glycolytic intermediates, pH, and transporter oligomeric structure affect ATP regulation of sugar transport. Finally, GLUT1 was reconstituted into proteoliposomes to determine whether ATP is sufficient for the modulation of GLUT1 function in-vitro. This thesis presents data supporting the hypothesis that residues 332-335 contribute to the efficiency of adenine nucleotide binding to GLUT1. In addition, we show that AMP, acidification, and conversion of the transporter to its dimeric form antagonize ATP regulation of sugar transport. Finally, we present results that support the proposal that ATP interaction with GLUT1 is sufficient for transport modulation.

The yeast SWI/SNF complex is the prototype of a subfamily of ATP-dependent chromatin remodeling complexes. It consists of eleven stoichiometric subunits including Swi2p/Snf2p, Swi1p, Snf5p, Swi3p, Swp82p, Swp73p, Arp7p, Arp9p, Snf6p, Snf11p, and Swp29p, with a molecular weight of 1.14 mega Daltons. Swi2p/Snf2p, the catalytic subunit of SWI/SNF, is evolutionally conserved from yeast to human cells. Genetic evidence suggests that SWI/SNF is required for the transcriptional regulation of a subset of genes, especially inducible genes. SWI/SNF can be recruited to target promotors by gene specific activators, and in some cases, SWI/SNF facilitates activator binding. Biochemical studies have demonstrated that purified SWI/SNF complex can hydrolyze ATP, and it can use the energy from ATP hydrolysis to generate superhelical torsion, mobilize mononucleosomes, enhance the accessibility of endonucleases to nucleosomal DNA, displace H2A/H2B dimers, induce dinucleosome and altosome formation, or evict nucleosomes. A human homolog of Swi2p/Snf2p, BRG1, is the catalytic subunit of the human SWI/SNF complex. Interestingly, isolated BRG1 alone is able to remodel a mononucleosome substrate. Importantly, mutations in mammalian SWI/SNF core subunits are implicated in tumorigenesis. Therefore, it remains interesting to characterize the role(s) of each subunit for SWI/SNF function. In this thesis project, I dissected SWI/SNF chromatin remodeling function by investigating the role of the SANT domain of the Swi3p subunit. Swi3p is one of the core components of SWI/SNF complex, and it contains an uncharacterized SANT domain that has been found in many chromatin regulatory proteins. Earlier studies suggested that the SANT domain of Ada2p may serve as the histone tail recognition module. For Swi3p, a small deletion of eleven amino acids from the SANT domain caused a growth phenotype similar to that of other swi/snf mutants. In chapter I, I have reviewed recent findings in the function of chromatin remodeling complexes and discuss the molecular mechanism of their action. In chapter II, I characterized the role of the SANT domain of Swi3p. I found that deletion of the SANT domain caused a defect in a genome-wide transcriptional profile, SWI/SNF recruitment, and more interestingly impairment of the SANT domain caused the dissociation of SWI/SNF into several subcomplexes: 1) Swi2p/Arp7p/Arp9p, 2) Swi3p/Swp73p/Snf6p, 3) Snf5p, and 4) Swi1p. Artificial tethering of SWI/SNF onto a LacZ reporter promoter failed to activate the reporter gene in the absence of the SANT domain, although Swi2p can be recruited to the LacZ promoter. We thus demonstrated that the Swi3p SANT domain is critical for Swi3p function and serves as a protein scaffold to integrate these subcomplexes into an intact SWI/SNF complex. In Chapter III, I first characterized the enzymatic activity of the subcomplexes, especially the minimal complex of Swi2p/Arp7p/Arp9p. We found that this minimal subcomplex is fully functional for chromatin remodeling in assays including cruciform formation, restriction enzyme accessibility in mononucleosomal and nucleosomal array substrates, and mononucleosome mobility shift. However, it is defective in ATP-dependent removal of H2A/H2B dimers. Moreover, we found that Swi3p and the N-terminal acidic domain of Swi3p strongly interact with GST-H2A and H2B but not GST-H3 or H4 tails. We purified a SWI/SNF mutant (SWI/SNF-Δ2N) that lacks 200 amino acids within the N-terminal acidic domain of Swi3p. Intriguingly, SWI/SNF-Δ2N failed to catalyze ATP-dependent dimer loss, although this mutant SWI/SNF contains all the subunits and has intact ATP-dependent activity in enhancing restriction enzyme accessibility. These data help to further understand the molecular mechanism of SWI/SNF, and show that H2A/H2B dimer loss is not an obligatory consequence of ATP-dependent DNA translocation, but requires the histone chaperone function of the Swi3p subunit. Based on these findings, we proposed a new model of the structural and functional organization of the SWI/SNF chromatin remodeling machinery: SWI/SNF contains at least four distinct modules that function at distinct stages of the chromatin remodeling process. 1) Swi1p and Snf5p modules directly interact with gene specific activators and function as the recruiter; 2) Swi2p/Arp7p/Arp9p generates energy from ATP hydrolysis and disrupts histone/DNA interactions; and 3) Swi3p/Swp73p/Snf6p may play dual roles by integrating each module into a large remodeling complex, as well as functioning as a histone H2A/H2B chaperone to remove dimers from remodeled nucleosomes. Chapter IV is a perspective from current work in this project. I first discuss the interest in further characterizing the essential role of Snf6p, based on its activation of LacZ reporter on its own. Using in vitro translated protein and co-IP studies, I tried to pinpoint the requirement of the SANT domain for SWI/SNF assembly. I found that Swi3p directly interacts with Swp73p, but not with other subunits. When Swi3p is first incubated with Swp73p, Swi3p also interacts with Snf6p, indicating that Swi3p indirectly interacts with Snf6p, therefore forming a subcomplex of Swi3p/Swp73p/Snf6p. This subcomplex can also be reconstituted using in vitro co-translation. Consistent with the TAP preparation of this subcomplex, partial deletion of the SANT domain of Swi3p does not affect the assembly of Swi3p/Swp73p/Snf6p in vitro. However, the assembly of SWI/SNF complex was not detected in the presence of eight essential in vitro translated subunits or from co-translation of all the subunits. I have discussed the interest in further characterizing the histone chaperone role of the Swi3p N-terminal acidic domain and the role of other core subunits of SWI/SNF such as Snf6p for transcriptional regulation.

The regulation of protein function through oligomerization is a common theme in biological systems. In this work, I have focused on the effects of the oligomeric states of the human Rad52 protein on activities related to DNA binding. HsRad52, a member of the RAD52 epistasis group, is thought to play an important and as yet undefined role in homologous recombination. HsRad52 preferentially binds to ssDNA over dsDNA and stimulates HsRad51-mediated strand exchange (Benson et al., 1998). In either the presence or absence of DNA, HsRad52 has been observed to form both 10 nm ring-like structures as well as higher order oligomers consisting of multiple 10 nm rings (Van Dyck et al., 1998; Van Dyck et al., 1999). Earlier protein-protein interaction studies mapped the domain responsible for HsRad52 self-association in the N-terminus (residues 85-159) (Shen et al., 1996). Data presented here identifies a novel self-association domain in the C-terminus of HsRad52 that is responsible for the formation of higher order oligomers. VanDyck et al. observed DNA ending binding complexes consisting of multiple rings (Van Dyck et al., 1999). They proposed that these higher order oligomers may be functionally relevant. In this work, we demonstrate that DNA binding depends on neither ring shaped oligomers nor higher order oligomers but that activities of HsRad52 that require simultaneous interaction with more than one DNA molecule depend on the formation of higher order oligomers consisting of multiple HsRad52 rings. Early studies of HsRad52 proposed that the DNA binding domain resides in the highly conserved N-terminus of the protein (Park et al., 1996). A series of studies using truncation mutants of HsRad52 have provided evidence that supports this hypothesis. For example, we demonstrated that a truncation mutant containing only the first 85 residues of the protein is still able to bind DNA (Lloyd, submitted 2002). In this study, we demonstrate that aromatic (Y65, F79 and Y81) and hydrophobic (L43, I52 and I66) residues within the N-terminus contribute to DNA binding by either directly contacting the DNA or by stabilizing the structure of the protein. In summary, through the work presented in this dissertation, we have determined that the formation of 10 nm rings is mediated by a self-association domain in the N-terminus and that the formation of higher order oligomers consisting of multiple HsRad52 rings is mediated by an additional self-association domain in the C-terminus. We have correlated the oligomeric properties of HsRad52 with its biochemical functions related to DNA binding. Additionally, we have demonstrated that aromatic and hydrophobic residues contribute to DNA binding. Further studies will differentiate between the contribution of these residues to the DNA binding by stabilizing the overall structure of the protein versus making specific DNA contacts. Additional studies will also address how the oligomeric state of HsRad52 contributes to its role in HsRad51-mediated strand exchange.