β Clamp Directs Localization of Mismatch Repair in Bacillus subtilis

Molecular Cell

Article b Clamp Directs Localization of Mismatch Repair in Bacillus subtilis Lyle A. Simmons,1 Bryan W. Davies,1 Alan D. Grossman,1 and Graham C. Walker1,* 1Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02139, USA *Correspondence: [email protected] DOI 10.1016/j.molcel.2007.10.036

SUMMARY

MutS homologs function in several cellular pathways including mismatch repair (MMR), the process by which mismatches introduced during DNA replication are corrected. We demonstrate that the C terminus of Bacillus subtilis MutS is necessary for an interaction with b clamp. This interaction is required for MutS-GFP focus formation in response to mismatches. Reciprocally, we show that a mutant of the b clamp causes elevated mutation frequencies and is reduced for MutS-GFP focus formation. MutS mutants defective for interaction with b clamp failed to support the next step of MMR, MutL-GFP focus formation. We conclude that the interaction between MutS and b is the major molecular interaction facilitating focus formation and that b clamp aids in the stabilization of MutS at a mismatch in vivo. The striking ability of the MutS C terminus to direct focus formation at replisomes by itself, suggests that it is mismatch recognition that licenses MutS’s interaction with b clamp. INTRODUCTION Replicative DNA polymerases are responsible for duplicating entire genomes with high fidelity (for review see Johnson and O’Donnell, 2005). Despite the high fidelity of this process (e.g., Kool, 2002), replication errors still occur. Correction of these errors requires the enlistment of mismatch repair (MMR) proteins to restore the proper coding sequence (for review see Kunkel and Erie, 2005). MMR is an important, highly conserved repair pathway that is found in both prokaryotes and eukaryotes (Culligan et al., 2000). In bacteria, deletion of MMR genes results in a several hundred-fold increase in mutation frequency (Cox et al., 1972). In humans, several primary tumors from cancers including hereditary nonpolyposis colorectal carcinoma (HNPCC) and Turcot syndrome (Fishel et al., 1993; Hamilton et al., 1995) have MMR defects, suggesting that increased mutation frequency could contribute to the development of these cancers. The function of MMR proteins is not limited to their clearly defined roles in MMR. For example, in bacteria MMR proteins can also participate in an antirecombination mechanism preventing recombination between nonidentical DNA sequences and thereby serving

to preserve species barriers at a molecular level (Rayssiguier et al., 1989). In eukaryotes, MMR proteins have been shown to function in meiosis (Hunter and Borts, 1997; Williamson et al., 1985) and are important for maintaining repetitive DNA sequences (microsatellites) (Orth et al., 1994). Moreover, MMR proteins contribute to several other cellular processes including apoptosis (Hickman and Samson, 2004) and DNA damage checkpoint activation (Yoshioka et al., 2006). The central and most intensively studied role for MMR is in the repair of mismatched nucleotides or insertion-deletion loops introduced principally as polymerase errors. In both prokaryotes and eukaryotes, MMR is initiated through the binding of a MutS homolog to mismatched DNA (for review see Schofield and Hsieh, 2003). In the paradigmatic E. coli system, MutS recognizes a mismatch followed by the recruitment of MutL (Schofield et al., 2001). MutL coordinates the actions of the nicking endonuclease MutH (e.g., Hall and Matson, 1999) and the loading of UvrD helicase (Hall et al., 1998) at the incised nick to result in the ATP-dependent separation of the DNA strand encoding the mismatch (Oeda et al., 1982). The mismatch-containing strand can then be degraded by one of several exonucleases (Viswanathan et al., 2001). The single-strand region is filled in by DNA Pol III and sealed by ligase to complete the repair process (Lahue et al., 1989). In eukaryotes, a heterodimer of MutS homologs (Saccharomyces cerevisiae MSH2-MSH6 [MutSa] [Prolla et al., 1994b] and hMSH2-hMSH6 in humans [Acharya et al., 1996]) is responsible for the recognition of most mismatches and small insertion or deletion loops (Drummond et al., 1995). These proteins recruit heterodimeric MutL homologs, (S. cerevisiae MLH1-PMS1 [MutLa], and human hMLH1-hPMS2 [Prolla et al., 1994a]), which function analogously to E. coli MutL and appear important for coordinating the actions of the remaining proteins in the pathway. The purified human proteins have been used to reconstitute MMR in vitro (Dzantiev et al., 2004). These studies have revealed that MMR excision from a 50 strand break requires MutSa, an exonuclease (EXO1), and single-stranded binding protein (RPA). MMR excision from a strand break 30 to the mismatch requires the proteins mentioned above in addition to MutLa, the clamp loader (RFC), and the replication processivity clamp (PCNA). These experiments indicate that the protein assemblies and mechanisms required for excision and repair are different depending on the directionality of the strand break relative to the mismatch. Theses studies also demonstrate the dependence on a processivity clamp for 30 directed repair. In B. subtilis, both MutS and MutL fused to green fluorescent protein (GFP) localize as discrete foci in cells exposed to the

Molecular Cell 29, 291–301, February 15, 2008 ª2008 Elsevier Inc. 291

Molecular Cell b Clamp Recruits MutS

Figure 1. MutS Interacts with the b Clamp through the C-Terminal 58 Amino Acids The peptide array and far-western blot were probed with b clamp bearing a single Myc tag (b-Myc). (A) Shown are 10-mer peptides of interest from the MutS N terminus that failed to bind b or the Cterminal 58 amino acids that bound b-Myc. The amino acid sequence is indicated, and the b clamp binding motif or N-terminal motif is highlighted in gray. Ponceau staining of the peptides is shown in the left-most panel. (B) b-Myc, MutS, MutS800, and His-DnaE probed with b-Myc. (C) Interaction between MutS or MutS800 with b clamp covalently linked to a sensor chip was completed using surface plasmon resonance as measured with a BIAcore biosensor (Experimental Procedures). Representative SPR traces for MutS 1.0 mM (blue) and 4.0 mM (red), and MutS800 2.0 mM (green) and 10.0 mM (purple) are shown.

mismatch-inducing agent 2-aminopurine (2-AP) (Smith et al., 2001). A new challenge in understanding MMR is to determine the protein interactions that allow for the coordinated assembly of MMR foci in vivo. B. subtilis serves as a particularly useful system for addressing this issue because B. subtilis is the only bacterium in which MMR proteins have been visualized in vivo. In addition, like eukaryotes and most bacteria, B. subtilis uses a DNA methylase-independent mechanism for strand discrimination indicating that B. subtilis serves as a valuable model system for understanding this process. Experiments in several systems have shown that MutS homologs interact with replication processivity clamps (b clamp in bacteria and PCNA in eukaryotes) (Gu et al., 1998; Kleczkowska et al., 2001; Lau and Kolodner, 2003; Lee and Alani, 2006; Lopez de Saro et al., 2006; Umar et al., 1996). Disruption of the interaction between MSH6 and PCNA in S. cerevisiae results in mild mutator phenotypes in vivo (for review see Schofield and Hsieh, 2003). The only modifications shown to result in strong mutator phenotypes encompass large deletions of the N-terminal domain that forms a flexible linker to PCNA (Shell et al., 2007). In E. coli, internal deletion of a b clamp binding motif renders MutS refractory to b clamp binding in vitro, yet in vivo the corresponding mutS allele confers a wild-type MMR phenotype (Lopez de Saro et al., 2006). Collectively, MutS homologs from several organisms have been shown to bind their resident processivity clamps; however, the in vivo significance for this interaction is unclear. We examined the mechanism that governs the focus formation response of MutS to mismatches in B. subtilis. We studied an as-

sortment of MutS variants altered in their C terminus and found that interaction between MutS and b clamp is required for MutS-GFP focus formation. Furthermore, we found that the C-terminal 58 amino acids are not only necessary but also sufficient for focus formation and that these foci exhibit the same subcellular distribution as the replisome. We also found that MutS deleted for this C-terminal region is reduced for binding to b clamp, although this purified mutant protein retains the ability to bind a mismatch in vitro. Taken together, we conclude that interaction between MutS and b is the major molecular interaction that facilitates focus formation and that b clamp aids in the stabilization of MutS at a mismatch in vivo. RESULTS The C Terminus of MutS Is Required for Interaction with b Clamp Several lines of evidence indicate that processivity clamps play an important role in MMR (Kleczkowska et al., 2001; Lopez de Saro et al., 2006; Umar et al., 1996). B. subtilis MutS contains a putative b clamp binding motif (806QLSFF810). It has been hypothesized that b clamp binding motifs modulate the interaction between a variety of proteins and the b clamp (Dalrymple et al., 2001). In vitro studies of E. coli MutS have shown that both its N- and C-terminal regions interact with b clamp (Lopez de Saro et al., 2006). The N-terminal MutS motif is necessary for MMR in E. coli, while the C-terminal MutS motif appears to be dispensable (Lopez de Saro et al., 2006) in vivo. As a first step toward determining whether b clamp contributes to mismatch-dependent MutS-GFP focus formation in B. subtilis, we used a peptide array approach to identify MutS sequences capable of interacting with the b clamp. Unlike E. coli, we found that peptides containing the conserved N-terminal motif (9QQYL12) failed to bind b (Figure 1A). Instead, we found that

292 Molecular Cell 29, 291–301, February 15, 2008 ª2008 Elsevier Inc.

Molecular Cell b Clamp Recruits MutS

the predicted C-terminal b clamp binding motif of MutS (806QLSFF810) does indeed bind b clamp (Figure 1A) and that one other 10-mer peptide corresponding to a region within the last 58 amino acids of MutS also bound b (Figure 1A). We did find that several N-terminal MutS 10-mer peptides did bind b (see Figure S1A available online). To determine whether any of these regions were important for interaction with b in the folded MutS protein, we performed a far-western blot with purified intact MutS (amino acids 1–858) and a protein deleted for the C-terminal 58 amino acids (referred to as MutS800). We found that MutS bound a C-terminal Myc (single) tagged derivative of b clamp (b-Myc). The interaction was comparable to that for DnaE (an essential polymerase) and b-Myc (Figure 1B). In contrast, no interaction between MutS800 and b-Myc was detected (Figure 1B). By covalently coupling b-Myc to a flow cell surface, we were able to use the more sensitive technique of surface plasmon resonance (SPR) to show that a very minimal interaction with b can be detected when MutS800 is present at high concentrations (Figure 1C). In these experiments, binding is indicated by a trace with resonance units above zero. For each trace shown, we subtracted the injection signal for each protein over a blank surface. We found that when we injected MutS with increasing concentrations (Figure 1C) we observed an increase in binding and an increase in retention of MutS on the b-Myc-coupled surface (Koff) indicative of protein-protein interaction over a range of 1.0–4.0 mM. In contrast, when MutS800 was injected we observed no interaction with b-Myc above the blank control surface at 2.0 mM (Figure 1C, green trace). When we increased the concentration of MutS800 to 10.0 mM we observed a binding signal less then MutS at 4.0 mM (Figure 1C). We attempted this analysis with MutS850; however, this protein showed significant binding to the blank surface, precluding further use of this technique to evaluate MutS850 interaction with b clamp (data not shown). With these data, we conclude that MutS binds b clamp and MutS800 is significantly reduced for binding b clamp. We also conclude that the weak binding observed by SPR analysis between MutS800 and b clamp is likely mediated by the multiple N-terminal peptides that we identified in our peptide array analysis (Figure S1A). We further conclude that the C-terminal 58 amino acids represent the crucial site for interaction with b clamp in vitro. It seems likely that after MutS interacts with a mismatch it could change conformation in such a way that the other interaction sites that we identified in the peptide array could become important for forming a productive MMR complex between MutS and b clamp. Intriguingly, our observations reveal that the functional interaction between MutS and b clamp is different for E. coli and B. subtilis. In E. coli, the N-terminal site of interaction between MutS and b is critical for MMR activity in vivo (Lopez de Saro et al., 2006). In B. subtilis, the MutS C terminus is critical for interaction between MutS and b clamp and for MMR activity in vivo (see below). It should be noted that we considered a role for MutL in contributing to MutS-GFP localization, but we did not find support for this possibility. We describe in the Supplemental Data that in a DmutL strain background MutS-GFP foci are unaffected (Supplemental Results and Table S1).

The C Terminus of MutS Is Required for MMR and Focus Formation In Vivo, but Not Mismatch Binding In Vitro We show that at least two nonoverlapping 10-mer peptides in the C-terminal 58 amino acids have the capability to bind b clamp. We also found that MutS800 is significantly reduced for binding to b, suggesting that the entire C-terminal region might be involved in interaction with b and thus might influence MMR and focus formation. Therefore, we constructed a set of mutS-gfp alleles that successively removed segments of this 58 amino acid C-terminal region from MutS within the MutS-GFP fusion context (Figure 2A). All of these truncated mutS alleles were present at the normal mutS gene location on the chromosome and provide the only source of MutS protein in the cell. Because integration of these alleles disrupts mutL expression, mutL was ectopically expressed from the amyE locus (Experimental Procedures). All of the C-terminal mutS deletion mutants exhibited profound deficiencies for MMR as assayed by mutation frequency (Figure 2B). Strikingly, truncation of only the last 8 amino acids of MutS (MutS850-GFP) resulted in nearly a complete loss of MMR activity in vivo (Figure 2B). Furthermore, all of the MutS deletion proteins accumulated to levels similar to those observed for wild-type protein in vivo (Figure 2B). To ensure this result was not influenced by the GFP component of MutSGFP, we also replaced the native mutS gene with each corresponding mutS truncation allele lacking the GFP fusion. We found that all of these mutS C-terminal deletion alleles resulted in profound MMR deficiencies (Figure S2). All of the MutS-GFP derivatives with truncation of the MutS C terminus exhibited major defects in focus formation when cells were treated with 2-AP to introduce replication errors (Figures 3C and 3F). These results suggest that interaction between MutS and b clamp through the MutS C terminus might contribute to stabilizing MutS at mismatches in vivo. Another explanation could be that the MutS C-terminal truncation mutants are unable to bind mismatches. To discriminate between these possibilities, we tested MutS800 and MutS850 for mismatch binding in vitro. We found that mismatch binding by purified MutS800 and MutS was nearly indistinguishable (Figure 3D). Because MutS850 only truncated the last eight amino acids, we examined this protein for binding to b clamp by far-western blotting. We found that MutS850 was reduced for binding b but was not completely defective, and this protein also bound a mismatch in vitro (Figures S1B and S1C). Taken together, the only biochemical difference that we have identified between MutS and MutS800 (or MutS850) is interaction with b clamp. In E. coli MutS, removal of the C-terminal region (53 amino acids) causes defects in tetramer formation and mismatch binding (Bjornson et al., 2003), and the truncated protein is defective in vivo when expressed from the native locus (Calmann et al., 2005). Whether or not tetramer formation by E. coli MutS provides an in vivo function is unclear (Mendillo et al., 2007; Miguel et al., 2007). Our finding that MutS800 and MutS850 bind a mismatch in vitro suggests that our deletion mutants do not suffer from the same biochemical defects as the E. coli protein. These data support the model that removal of C-terminal residues from MutS disrupts interaction between MutS and b clamp, which uncovers a defect in the cellular response to mismatches in vivo, but not mismatch binding in vitro. We conclude that, in vivo,

Molecular Cell 29, 291–301, February 15, 2008 ª2008 Elsevier Inc. 293

Molecular Cell b Clamp Recruits MutS

Figure 2. The MutS C Terminus Is Required for MMR Activity In Vivo (A) Diagram of B. subtilis MutS domain structure and segments of the protein absent from the truncated mutants. The domains were determined by alignment with Thermus aquaticus MutS, and the domain nomenclature follows that of the crystal structure (Obmolova et al., 2000). Briefly, domains I and IV are involved in mismatch binding, domain III is the central core of the protein, domain II resembles an RNase H-like fold, and domain V is an ABC ATPase. (B) Corresponding mutation frequencies associated with strains lacking the indicated C-terminal amino acids of MutS. Relative stability of the resultant proteins was determined by immunoblot analysis and is presented above the histogram. (C) Mutation frequency resulting from alanine scanning mutagenesis of the MutS b clamp binding motif. An immunoblot of the in vivo stability of the resulting proteins is also shown. The alanine substitutions in the MutS b clamp binding motif are indicated in bold type. Error bars represent the standard deviation from at least triplicate experiments.

interaction between MutS and b clamp helps stabilize MutS at mismatches and stimulate focus formation. The Putative b Clamp Binding Motif Is Important for MMR and MutS Focus Formation In Vivo To examine the importance of the putative b clamp binding motif to MMR, we performed comprehensive alanine scanning mutagenesis of this penta-amino acid motif (Figure 2C). We found that substitution of the second to last amino acid position (806QLSFF810) in this motif with an alanine indicated that it was important for MMR activity in vivo. We also found that substituting the first two positions (806QLSFF810) or the last two positions (806QLSFF810) with alanines conferred an increase in mutation frequency similar to replacing the entire motif with alanines. Even so, replacement of the entire motif with alanines (MutS5A) resulted in only an 8-fold increase in mutation frequency (Figure 2C) compared with 50-fold for the DmutS strain when both are normalized to the mutS-gfp strain. These results suggest that, although this motif contributes to MutS function in vivo, considerable MMR can occur in its absence. We asked whether a correlation exists between the mild mutator phenotype observed for MutS5A and a decrease in focus formation. We challenged our MutS5A-GFP strain with 2-AP and observed an 2-fold (MutS5A-GFP 21%, n = 1102; MutS-GFP 42%, n = 1309) reduction in focus formation (Figures 3B and 3F) indicating that MutS5A is less proficient for focus

formation than MutS-GFP in response to replication errors. We also determined the subcellular position of MutS5A-GFP relative to MutS-GFP. We found that the position of MutS5A-GFP and MutS-GFP were nearly identical, indicating that the frequency of MutS5A-GFP focus formation is altered, but not the subcellular position at which foci form (Figure 3G). Immunoblot analysis of these MutS-GFP b clamp binding motif mutants revealed that the levels of these proteins were indistinguishable from wild-type MutS fused to GFP in vivo (Figure 2C). Finally, as an additional control we constructed fusions of MutS, MutS5A, and MutS800 to a monomeric version of GFP and found that the localization properties of these proteins were almost identical to the more commonly used GFPMUT2 employed in the above analyses (Table S2). We conclude that the b clamp binding motif of B. subtilis MutS contributes to both MutS focus formation and MMR; however, the system can function moderately well in the absence of the motif. MutS800 and MutS5A Are Reduced for MutL Recruitment Previous experiments examining MutS-GFP and MutL-GFP localization in B. subtilis demonstrated that 2-AP treatment failed to stimulate MutL-GFP focus formation in strains lacking the mutS gene (Smith et al., 2001). Taken together with biochemical studies from E. coli demonstrating that MutL directly binds to MutS (Wu and Marinus, 1999), these observations indicate that MutS-GFP foci must form prior to MutL-GFP foci in vivo. We therefore asked whether our C-terminal MutS variants impaired for their own focus formation and MMR were also impaired for supporting MutL-GFP focus formation.

294 Molecular Cell 29, 291–301, February 15, 2008 ª2008 Elsevier Inc.

Molecular Cell b Clamp Recruits MutS

Figure 3. The C Terminus of MutS Is Required for Focus Formation All samples shown are representative. Cells were treated with 600 mg/ml 2-AP to induce mismatches. Shown are (A) MutS-GFP foci, (B) MutS5A-GFP, and (C) MutS800-GFP. (D) MutS, MutS800, and () control (no protein) were incubated with homoduplex DNA (T/A) or mismatched heteroduplex DNA (T/G) as described (Biswas and Hsieh, 1996). MutS bound to DNA was retained on the nitrocellulose membrane and exposed to film. (E) Percent of cells with foci untreated. (F) Percent of cells with foci 45 min following addition of 2-AP. (G) Comparison of the cellular position of MutSGFP single foci (black bar) and MutS5A-GFP foci (white bar) following 2-AP addition. The number of cells scored for untreated and 2-AP treated in (D) and (E) are as follows: MutS-GFP (n = 837, n = 1309); MutS5A-GFP (n = 1044, n = 1102); MutS800-GFP (n = 396, n = 958); MutS810-GFP (n = 567, n = 410); MutS830-GFP (n = 382, n = 944); MutS840-GFP (n = 595, n = 1022), MutS850-GFP (n = 771, n = 801). The cell membrane is stained with FM4-64 and is colored in red, while MutS-GFP foci are in green. The white bar indicates 2 mm. White dots are adjacent to MutS-GFP foci.

To this end, we utilized strains in which the mutS800 and mutS5A alleles were located at their native chromosomal locus. The mutL-gfp allele was integrated at an ectopic locus (amyE) and placed under control of an IPTG-inducible promoter. We used this strategy because the integration of the mutS alleles disrupts expression of the mutL gene directly downstream (Smith et al., 2001). MutL-GFP forms foci in response to 2-AP treatment from an ectopic locus at the same frequency as MutL-GFP expressed from the native locus (Smith et al., 2001). The mutS800 allele supported only a small frequency of MutL-GFP focus formation (3%, n = 2866), a level nearly identical to the DmutS allele (2%, n = 1042), indicating that mutS800 is completely defective for recruiting MutL-GFP to mismatches (Figures 4C, 4D, and 4F). The mutS5A allele displays an intermediate ability in forming its own focus (2-fold [n = 1102] reduction) and a similar, although more severe, reduction for supporting MutL-GFP foci (a 4-fold [n = 690] reduction) (Figures 4B and 4F). The MutL-GFP foci that formed in a mutS5A genetic background were the same, qualitatively, as those that formed in a mutS+ background. We conclude that localization of MutL-GFP to mismatches requires an MMR-proficient mutS allele.

The MutS C Terminus Is Necessary and Sufficient to Localize YFP to b Clamp In Vivo Biochemical studies have revealed that E. coli MutS homodimers form clamplike structures that can slide away from a mismatch along the DNA (Acharya et al., 2003). This led to the demonstration that a single mismatch could stimulate the loading of multiple MutS homodimers (Acharya et al., 2003). This mechanism provided a facile explanation for the MutS focus formation we had observed (Smith et al., 2001). Nevertheless, we asked whether the C terminus of MutS is not only necessary but also sufficient for MutS focus formation. We used a strain expressing yellow fluorescent protein (YFP) from an ectopic locus using an IPTG-inducible promoter. As expected, YFP alone was diffuse in the cell and failed to show any specific localization pattern (Figure 5A). However, when we linked YFP to the C-terminal 58 amino acids of MutS (MutSCYFP), we found that this protein localized as foci in virtually every cell in a wild-type background in the absence of 2-AP addition. A similar result was obtained in a strain deleted for both mutS and mutL (Figure 5B). Thus focus formation by the MutSC-YFP protein does not require interaction with intact MutS or MutL. The foci formed in the mutSL+ strain are qualitatively more intense, suggesting that intact MutS may increase the number of MutSC-YFP in each focus even though intact MutS is not absolutely required (Figure 5). MutSC-YFP formed foci in 82% of cells (n = 326) in a DmutSL strain, indicating that in vivo the C terminus of MutS is sufficient to localize as a focus. This result was very intriguing because it suggests that DNA binding by MutS is not required for focus

Molecular Cell 29, 291–301, February 15, 2008 ª2008 Elsevier Inc. 295

Molecular Cell b Clamp Recruits MutS

Figure 4. MutS Mutant Proteins Are Impaired for Supporting MutL-GFP Focus Formation Cells were grown in S750 minimal medium and treated with 600 mg/ml 2-AP to induce mismatches 45 min prior to visualization. Cell images shown are representative of several experiments. Shown are (A) MutL-GFP in a wild-type mutS background, (B) MutL-GFP in a mutS5A (b clamp binding motif replaced with alanines) background, (C) MutLGFP in a mutS800 background, and (D) MutLGFP in a DmutS background. (E) Percent of cells with MutL-GFP foci untreated. (F) Percent of cells with MutL-GFP foci 45 min following the addition of 2-AP. (G) Comparison of the subcellular position of MutL-GFP foci in a mutS (black bars) and mutS5A (white bars) genetic background (n = 150). Number of cells scored for the graphs in (E) and (F) with the untreated number followed by the 2-AP-treated number are as follows: MutLGFP (n = 1396, n = 891) in a wild-type mutS background; (n = 1028, n = 1042) in a DmutS background; (n = 2543, n = 2866) in a mutS800 background; and (n = 1142, n = 884) in a mutS5A background.

formation, but instead interaction with b clamp is sufficient for focus formation. It has been shown in both B. subtilis and E. coli that b clamp fused to GFP formed discrete foci, suggesting that b clamp is concentrated to sites of replication in vivo (Kongsuwan et al., 2002; Meile et al., 2006). We performed a similar analysis, except that b-GFP was expressed from its native promoter at the native chromosomal locus, and this fusion protein also formed discrete foci in vivo (Figure 5E). The simplest explanation is that MutSC-YFP binds to the b clamp, and it is this interaction that allows for visualization of foci in vivo. The Subcellular Position of MutSC and b Clamp Is Consistent with the Replisome We asked whether both b-GFP and MutSC-YFP formed foci at cellular positions consistent with the replisome (defined as replicative DNA polymerase and associated proteins at the replication fork). Indeed, both b-GFP and MutSC-YFP localization for single-focus cells was indistinguishable from the subcellular position of DnaX-GFP (replisome) (Figure 5F). An alternative explanation for the MutSC-YFP focus is that the C terminus is highly

multimerizing, which allows for us to visualize this protein fragment as a focus. Because MutSC-YFP shows the same localization pattern as the replisome (DnaX-GFP), we find this explanation unlikely. Our observation that the frequency of MutSC-YFP foci in cells untreated with a mismatch-inducing agent was so high (82%) relative to the frequency of focus formation by MutS-GFP (R5%) was of interest. These data suggest that, in the absence of a mismatch, MutS conformation is such that interaction with b clamp is weak or nonexistent. In contrast, in the presence of a mismatch, MutS undergoes a conformational change licensing the MutS C terminus to interact with b clamp. A Mutant b Clamp Is Reduced for Supporting MutS-GFP Focus Formation and MMR In Vivo The data presented above are consistent with a MutSb clamp interaction providing an important step in MMR. In support of this hypothesis, we have determined that a previously isolated allele (dnaN5) encoding a mutant form of b confers a mutator phenotype (Figure 6), and this allele confers a reduction in MutS-GFP focus formation (see below). The dnaN5 (referred to as bG73R) allele was isolated because it confers a temperature-sensitive replication phenotype at 49 C (Karamata and Gross, 1970). Sequencing of the dnaN5 allele revealed that it encodes a G73R missense mutation (Ogasawara et al., 1986). We found that strains bearing bG73R had elevated mutation frequencies (Figures 6A and 6B). The bG73R conferred an 10-fold mutator at 30 C and 25-fold at 37 C (Figures 6A and 6B). An epistasis analysis demonstrated that the mutation

296 Molecular Cell 29, 291–301, February 15, 2008 ª2008 Elsevier Inc.

Molecular Cell b Clamp Recruits MutS

sult. MutS-GFP foci were reduced 3-fold (13%, n = 679) at 37 C. In contrast, the percent of cells with MutS-GFP foci in a wild-type background was unchanged at 37 C (36%, n = 699) (Figure 6F). We performed an immunoblot analysis of wild-type b and bG73R using anti-serum against b and found that these proteins accumulated to indistinguishable levels in vivo at 30 C and 37 C (Figure 6E). The amino acid substitution in B. subtilis bG73R is located in a surface-exposed region on the outer rim of the protein (Figures 6C and 6D). This position in b might be important for MutS binding to b. In support of this, the G73R mutation in B. subtilis b is located in a position of b corresponding to a position in E. coli b that binds d0 , a component of clamp loader, and the little finger domain of DinB (Pol IV) and UmuC (Pol V) (Bunting et al., 2003; Jeruzalmi et al., 2001). We conclude that the mutator phenotype of bG73R results from decreased focus formation of MutS-GFP in vivo. DISCUSSION

Figure 5. The C Terminus of MutS Is Both Necessary and Sufficient for Focus Formation YFP was expressed from the amyE locus and visualized by fluorescence microscopy. Shown are (A) YFP and membrane; (B) MutSC-YFP in a mutSL strain and membrane; (C) MutSC-YFP and membrane in a wild-type strain; (D) MutSC-YFP and membrane in a mutL strain. The white bar indicates 2 mm. The percent of cells with foci 45 min after addition of IPTG are shown in each panel, and the numbers of cells scored are indicated in each panel. (E) Representative b clamp-GFP foci are shown during normal growth in S750 minimal medium. (F) Subcellular position of MutSC-YFP (white bars), b clamp-GFP (gray bars), and DnaX-GFP (black bars) are presented as the percent of cells versus the distance from the nearest pole.

frequency conferred by bG73R was not additive or synergistic with a DmutSL allele (Figures 6A and 6B), indicating that the mutator phenotype of bG73R results from loss of MutSL-dependent MMR. To further confirm that the bG73R mutator was MMR dependent and DNA damage independent, we examined phage induction in single cells by visualizing a YFP fusion to a PBSX-regulated gene (Britton et al., 2007). We found that bG73R did not display any detectable PBSX induction (Figure S3) above our wild-type control strain. The MutSL-dependent mutator phenotype caused by the bG73R mutation stimulated us to examine whether MutS-GFP focus formation was decreased in this background. Indeed, we found that MutS-GFP was reduced for focus formation in cells treated with the mismatch-inducing agent at 30 C (24%, n = 1248) relative to MutS-GFP foci in a wild-type background (36%, n = 799). We found that increasing the temperature to 37 C exacerbated this re-

In this study, we found that the C terminus of MutS is required for MMR and localization as a focus in response to replication errors. We have shown that this region of MutS is also required for interaction with b clamp. Furthermore, the C terminus of MutS is also sufficient for focus formation, even in the absence of mismatches. We interpret these results to mean that interaction between MutS and b is the major molecular interaction that facilitates focus formation and that b clamp aids in the stabilization of MutS at mismatches in vivo. This conclusion is supported by experiments demonstrating that a mutant form of MutS impaired for focus formation is a mutator in vivo, and this protein still binds a mismatch in vitro. These data strongly argue that, in a living cell, mismatch binding and focus formation are more complicated than MutS simply recognizing and binding to the error and recruiting several more MutS dimers to the site of the mismatch. These data also reinforce the idea that focus formation is physiologically important for MMR activity in vivo as we find a distinct correlation between mutS alleles that are defective for MMR activity and focus formation. In a reciprocal approach, we show that a mutant form of the b clamp is impaired for MMR and MutS-GFP focus formation. This observation strengthens our argument that b clamp is critical for MMR and MutS-GFP focus formation. We also show that a DmutSL allele is epistatic to our b clamp mutant, indicating these proteins function in the same pathway. Our findings that the C terminus of B. subtilis MutS is critical for interaction with b and MMR activity in vivo are in contrast to observations in E. coli. In E. coli, the C-terminal b clamp binding motif of MutS is important for interaction with b clamp in vitro, although the N-terminal interaction site is required for MMR activity in vivo (Lopez de Saro et al., 2006). The differences in b clamp binding observed between these two systems could reflect the use of methylation-directed repair by E. coli. Taken together, we conclude that b clamp helps recruit MutS into a focus at mismatches in B. subtilis. b Clamp Recruits MutS-GFP into a Focus in Response to Mismatches Because we deleted the entire domain structure of MutS responsible for DNA binding and still observed foci, we propose that the

Molecular Cell 29, 291–301, February 15, 2008 ª2008 Elsevier Inc. 297

Molecular Cell b Clamp Recruits MutS

Figure 6. A Mutant Form of the b Clamp Is Reduced for Supporting MutS-GFP Formation (A) Relative mutation frequency of the bG73R strain compared with MMR-deficient strains at 30 C, and (B) 37 C. The error bars represent the standard deviation from at least triplicate experiments. (C) Homology model of B. subtilis b clamp with the G73R substitution shown (green). Residues colored in red and yellow are conserved with residues in E. coli b that comprise the hydrophobic pocket (red) and conserved residues important for interaction between E. coli clamp loader component d0 and b (yellow). (D) Enlargement of the region containing the G73R mutation (green). (E) Immunoblot of b and bG73R protein levels in whole-cell lysates at 30 C and 37 C. Protein load was normalized to cell number. b is in lanes 1 and 3, and bG73R is in lanes 2 and 4. (F) Quantitation of MutS-GFP focus formation in wild-type and bG73R backgrounds at 30 C and 37 C.

MutS C terminus has an intrinsic ability to bind b and localize as a focus. We propose that, in the absence of a mismatch, intact MutS assumes a conformation that negatively controls the interaction between MutS and b clamp. Once MutS binds a mismatch, MutS assumes a conformation that licenses interaction with b clamp through the MutS C terminus. We propose that this conformational change takes place in MutS and not b clamp because our MutS C-terminal fragment appears to bind b, indicating that b clamp is constitutively able to interact with the MutS C terminus. Several models have been proposed suggesting that MutS recognizes and then translocates or diffuses away from the mismatch (Allen et al., 1997; Gradia et al., 1999). Placing these observations into context with ours, we propose that formation of the MutSb clamp complex is a required step to stimulate the repetitive loading of MutS dimers at a mismatch in a living cell. Another interpretation could be that MutS is bound to b clamp during replication and it dissociates from b once a mismatch is detected, resulting in the recruitment of more MutS dimers and focus formation. We find this explanation unlikely. We show that b clamp forms discrete foci during normal growth. If MutS constitutively associates with b during replication, we should be able to visualize MutS-GFP foci in the absence of mismatches by virtue of the interaction with b clamp. To the contrary, we only observe MutS foci after inducing mismatches or by expression of the MutS C terminus and thus bypassing the requirement for a mismatch to visualize foci.

Processivity Clamps May Contribute to Strand Discrimination in Organisms Lacking dam-Directed Methylation All MMR systems must recognize the mismatch and identify the strand bearing the mismatch. Early studies of the methylation-independent MMR system (hex) of S. pneumoniae led to the suggestion that single-strand breaks in the nascent strand were used to direct the repair system (Lacks et al., 1982). Examination of MMR using purified components from the human system has demonstrated both 30 and 50 repair from a single-strand break (Dzantiev et al., 2004; Zhang et al., 2005). Considered together with ours, these results indicate that interaction between MutS and b clamp might help localize MutS to the site of replication, bringing MutS into close proximity with the 30 terminus or the 50 end of an Okazaki fragment at a replication fork, allowing for the system to direct mismatch removal from these sites. Recently, it was shown that human MutL homolog hPMS2 contains an endonuclease activity that is required for MMR in vitro from a 30 strand break (Kadyrov et al., 2006). The endonuclease active site is conspicuously conserved among many organisms that lack methyl-directed MMR, including B. subtilis MutL (Kadyrov et al., 2006). Moreover, in human, S. cerevisiae, E. coli, and now B. subtilis, MutS homologs and some MutL homologs have been shown to interact with their cognate processivity clamps (PCNA or b) (Kleczkowska et al., 2001; Lee and Alani, 2006; Lopez de Saro et al., 2006; Shell et al., 2007). Together with our localization results, these data collectively support the model that MMR proteins assemble at mismatches located at sites of replication, stimulating the formation of MutS and MutL repair complexes through the direction of b clamp. We propose a model where MutS is stabilized by b clamp at mismatches in vivo located at or near the site of

298 Molecular Cell 29, 291–301, February 15, 2008 ª2008 Elsevier Inc.

Molecular Cell b Clamp Recruits MutS

DNA replication. MutS then recruits MutL where MutL could incise a nick to help direct removal and repair of the mismatch. EXPERIMENTAL PROCEDURES

tibodies used to detect b clamp were raised against b and are from Dr. C. Lee (Massachusetts Institute of Technology). The primary was used in a 1:5000 dilution, and the goat anti-rabbit HRP-conjugated secondary antibody was used in a 1:200 dilution.

Growth Conditions, Medium, and Mutagenesis Experiments For microscopy all strains were grown at 30 C in S750 minimal medium. The S750 medium was supplemented with 1% glucose, 0.1% glutamate, 40 mg/ml tryptophan, and 40 mg/ml phenylalanine or 0.5% casamino acids for Figure 5. Antibiotic concentrations were the same as previously described (Smith et al., 2001). Mutagenesis experiments were accomplished as described (Smith et al., 2001) except that for the data in Figure 2, 600 mg/ml 2-AP was added for 1 hr prior to harvesting and plating of the cells.

Surface Plasmon Resonance A BIAcore 3000 biosensor was used for our analysis of MutS or MutS800 binding to b clamp-Myc. Standard research-grade CM5 sensor chips were used for the immobilization of 10,200 resonance units of b clamp. Immobilization of b was accomplished using the amine coupling kit (BIAcore) following the manufacturer’s recommendations. b clamp was immobilized by using 10 mM sodium acetate buffer (pH 4.0). MutS and MutS800 were diluted in binding buffer prior to injection of the blank or b clamp-coupled surfaces.

Construction of B. subtilis Strains Expressing Mutant Forms of mutS All of the mutS alleles were constructed using plasmid pBTS240 that has been previously described (Smith et al., 2001). To verify that the intended mutS missense mutations were properly integrated, we PCR amplified the 30 end of the full-length mutS gene fused to gfp from the indicated strains and sequenced the PCR product (data not shown). The mutS alleles encoding the C-terminal deletion mutants and the mutS5A allele were cloned into plasmid pJL74 (LeDeaux and Grossman, 1995) using restriction sites EcoRI and XhoI. These alleles were integrated into the chromosome by transformation at the normal mutS gene location placing expression under the native mutS promoters.

Supplemental Data Supplemental Data include Supplemental Results, Supplemental Experimental Procedures, Supplemental References, three tables, and three figures and can be found with this article online at http://www.molecule.org/cgi/content/ full/29/3/291/DC1/.

Cell Treatment and Live Cell Microscopy Microscopy and cell treatment for microscopy were done essentially as described (Smith et al., 2001). Briefly, cells were grown in S750 minimal medium at 30 C. Where indicated, cultures were split and one was treated with 2-AP for 1 hr in mid-log phase (OD600 0.2–0.4), while the control culture was untreated. Aliquots (300 ml) of cells were extracted from each culture and treated with the vital membrane stain FM4-64 (Molecular Probes) and DAPI to visualize the DNA. Images were captured with a Nikon E800 microscope and OpenLab software (Improvision). Images were adjusted, colored, and merged using OpenLab software. Protein Purification, Far-Western Blotting, Peptide Array Analysis, and Immunoblotting MutS, MutS800, MutS850, and b-Myc were cloned into pET11T and overexpressed (primer sequences available upon request) in E. coli BL21DE3 cells using standard procedures. Cell pellets containing b-Myc, MutS, MutS800, and MutS850 were resuspended in buffer A (0.02 M K2HP04 [pH 7.4], 50 mM KCl, 10% glycerol, 0.1 M EDTA, and 0.1 M DTT) + 150 mM KCl, 20 mM SpCl3, 1 mM AEBSF, and 0.3 mg/ml lysozyme. For each protein, cells were ruptured by freeze/thaw lysis after incubation on ice for 20 min. The lysate was clarified by centrifugation prior to dilution with buffer A and application to a monoQ column for purification. An elution gradient of 0.1–1 M KCl was used with each protein eluting near 700 mM KCl. Purity of proteins was determined to be greater than 90% as judged by SDS-PAGE. Purified His6-DnaE was from Dr. Neil Brown (University of Massachusetts, Worcester). Far-western blotting was done with slight modification from a procedure already described (Ludlam et al., 2001). Briefly, equal amounts of the indicated proteins were immobilized onto a nitrocellulose membrane (Protran, Whatman) by spotting. Following protein immobilization, each membrane was blocked in phosphate-buffered saline (PBS) + 2% nonfat milk for 1 hr. After three washes for 10 min with TS/Tween, 2.6 mg/ml b-Myc was added to binding buffer (40 mM HEPES-KOH [pH 7.6], 10 mM MgCl2, 4 mM DTT, 1 mM ATP, 0.1 mg/ml BSA, and 0.1% glutaraldehyde) for 1 hr. Each membrane was washed 3X in TS/Tween for 20 min. The membranes were then incubated overnight at 4 C with anti-Myc monoclonal antibodies (1:1000) in TS/Tween with 2% nonfat milk. Membranes were washed 3X for 30 min with TS/Tween followed by incubation with goat anti-mouse HRP-conjugated (Pierce) antibodies (1:500) for 1 hr. Following three washes in TS/Tween for 30 min, membranes were developed with chemiluminescence (Super-Signal Altra, Pierce). Peptide arrays were performed as described (Beuning et al., 2006), and immunoblots were done as described (Simmons et al., 2007). Primary an-

ACKNOWLEDGMENTS We thank Dr. Brad Smith, Dr. Adam Breier, Mary Ellen Wiltrout, and members of the Walker and Grossman labs for careful reading of this manuscript. We thank Dr. Alexi Goranov for help with construction of the dnaN-GFP allele. We would also like to thank Dr. Caroline Koehrer for help with the SPR analysis and Drs. Uttam RajBhandary and H. Gobind Khorana for use of their BIAcore 3000 instrument. G.C.W. is an American Cancer Society Research Professor and is supported by NCI grant CA21615-27 and the Massachusetts Institute of Technology Center for Environmental Health Sciences. This work was also supported by NIH grant GM41934 to A.D.G. and a postdoctoral fellowship from the NCI to L.A.S. A graduate scholarship from the National Sciences and Engineering Research Council of Canada supported B.W.D. Received: August 7, 2007 Revised: September 28, 2007 Accepted: October 31, 2007 Published: February 14, 2008 REFERENCES Acharya, S., Wilson, T., Gradia, S., Kane, M.F., Guerrette, S., Marsischky, G.T., Kolodner, R., and Fishel, R. (1996). hMSH2 forms specific mispair-binding complexes with hMSH3 and hMSH6. Proc. Natl. Acad. Sci. USA 93, 13629–13634. Acharya, S., Foster, P.L., Brooks, P., and Fishel, R. (2003). The coordinated functions of the E. coli MutS and MutL proteins in mismatch repair. Mol. Cell 12, 233–246. Allen, D.J., Makhov, A., Grilley, M., Taylor, J., Thresher, R., Modrich, P., and Griffith, J.D. (1997). MutS mediates heteroduplex loop formation by a translocation mechanism. EMBO J. 16, 4467–4476. Beuning, P.J., Sawicka, D., Barsky, D., and Walker, G.C. (2006). Two processivity clamp interactions differentially alter the dual activities of UmuC. Mol. Microbiol. 59, 460–474. Biswas, I., and Hsieh, P. (1996). Identification and characterization of a thermostable MutS homolog from Thermus aquaticus. J. Biol. Chem. 271, 5040–5048. Bjornson, K.P., Blackwell, L.J., Sage, H., Baitinger, C., Allen, D., and Modrich, P. (2003). Assembly and molecular activities of the MutS tetramer. J. Biol. Chem. 278, 34667–34673. Britton, R.A., Kuster-Schock, E., Auchtung, T.A., and Grossman, A.D. (2007). SOS induction in a subpopulation of structural maintenance of chromosome (Smc) mutant cells in Bacillus subtilis. J. Bacteriol. 189, 4359–4366.

Molecular Cell 29, 291–301, February 15, 2008 ª2008 Elsevier Inc. 299

Molecular Cell b Clamp Recruits MutS

Bunting, K.A., Roe, S.M., and Pearl, L.H. (2003). Structural basis for recruitment of translesion DNA polymerase Pol IV/DinB to the beta-clamp. EMBO J. 22, 5883–5892. Calmann, M.A., Nowosielska, A., and Marinus, M.G. (2005). The MutS C terminus is essential for mismatch repair activity in vivo. J. Bacteriol. 187, 6577–6579. Cox, E.C., Degnen, G.E., and Scheppe, M.L. (1972). Mutator gene studies in Escherichia coli: the mutS gene. Genetics 72, 551–567. Culligan, K.M., Meyer-Gauen, G., Lyons-Weiler, J., and Hays, J.B. (2000). Evolutionary origin, diversification and specialization of eukaryotic MutS homolog mismatch repair proteins. Nucleic Acids Res. 28, 463–471. Dalrymple, B.P., Kongsuwan, K., Wijffels, G., Dixon, N.E., and Jennings, P.A. (2001). A universal protein-protein interaction motif in the eubacterial DNA replication and repair systems. Proc. Natl. Acad. Sci. USA 98, 11627–11632. Drummond, J.T., Li, G.M., Longley, M.J., and Modrich, P. (1995). Isolation of an hMSH2-p160 heterodimer that restores DNA mismatch repair to tumor cells. Science 268, 1909–1912. Dzantiev, L., Constantin, N., Genschel, J., Iyer, R.R., Burgers, P.M., and Modrich, P. (2004). A defined human system that supports bidirectional mismatch-provoked excision. Mol. Cell 15, 31–41. Fishel, R., Lescoe, M.K., Rao, M.R., Copeland, N.G., Jenkins, N.A., Garber, J., Kane, M., and Kolodner, R. (1993). The human mutator gene homolog MSH2 and its association with hereditary nonpolyposis colon cancer. Cell 75, 1027–1038. Gradia, S., Subramanian, D., Wilson, T., Acharya, S., Makhov, A., Griffith, J., and Fishel, R. (1999). hMSH2-hMSH6 forms a hydrolysis-independent sliding clamp on mismatched DNA. Mol. Cell 3, 255–261. Gu, L., Hong, Y., McCulloch, S., Watanabe, H., and Li, G.M. (1998). ATPdependent interaction of human mismatch repair proteins and dual role of PCNA in mismatch repair. Nucleic Acids Res. 26, 1173–1178. Hall, M.C., and Matson, S.W. (1999). The Escherichia coli MutL protein physically interacts with MutH and stimulates the MutH-associated endonuclease activity. J. Biol. Chem. 274, 1306–1312. Hall, M.C., Jordan, J.R., and Matson, S.W. (1998). Evidence for a physical interaction between the Escherichia coli methyl-directed mismatch repair proteins MutL and UvrD. EMBO J. 17, 1535–1541. Hamilton, S.R., Liu, B., Parsons, R.E., Papadopoulos, N., Jen, J., Powell, S.M., Krush, A.J., Berk, T., Cohen, Z., Tetu, B., et al. (1995). The molecular basis of Turcot’s syndrome. N. Engl. J. Med. 332, 839–847.

Kool, E.T. (2002). Active site tightness and substrate fit in DNA replication. Annu. Rev. Biochem. 71, 191–219. Kunkel, T.A., and Erie, D.A. (2005). DNA mismatch repair. Annu. Rev. Biochem. 74, 681–710. Lacks, S.A., Dunn, J.J., and Greenberg, B. (1982). Identification of base mismatches recognized by the heteroduplex-DNA-repair system of Streptococcus pneumoniae. Cell 31, 327–336. Lahue, R.S., Au, K.G., and Modrich, P. (1989). DNA mismatch correction in a defined system. Science 245, 160–164. Lau, P.J., and Kolodner, R.D. (2003). Transfer of the MSH2.MSH6 complex from proliferating cell nuclear antigen to mispaired bases in DNA. J. Biol. Chem. 278, 14–17. LeDeaux, J.R., and Grossman, A.D. (1995). Isolation and characterization of kinC, a gene that encodes a sensor kinase homologous to the sporulation sensor kinases KinA and KinB in Bacillus subtilis. J. Bacteriol. 177, 166–175. Lee, S.D., and Alani, E. (2006). Analysis of interactions between mismatch repair initiation factors and the replication processivity factor PCNA. J. Mol. Biol. 355, 175–184. Lopez de Saro, F.J., Marinus, M.G., Modrich, P., and O’Donnell, M. (2006). The beta sliding clamp binds to multiple sites within MutL and MutS. J. Biol. Chem. 281, 14340–14349. Ludlam, A.V., McNatt, M.W., Carr, K.M., and Kaguni, J.M. (2001). Essential amino acids of Escherichia coli DnaC protein in an N-terminal domain interact with DnaB helicase. J. Biol. Chem. 276, 27345–27353. Meile, J.C., Wu, L.J., Ehrlich, S.D., Errington, J., and Noirot, P. (2006). Systematic localisation of proteins fused to the green fluorescent protein in Bacillus subtilis: identification of new proteins at the DNA replication factory. Proteomics 6, 2135–2146. Mendillo, M.L., Putnam, C.D., and Kolodner, R.D. (2007). Escherichia coli MutS tetramerization domain structure reveals that stable dimers but not tetramers are essential for DNA mismatch repair in vivo. J. Biol. Chem. 282, 16345–16354. Miguel, V., Pezza, R.J., and Argarana, C.E. (2007). The C-terminal region of Escherichia coli MutS and protein oligomerization. Biochem. Biophys. Res. Commun. 360, 412–417. Obmolova, G., Ban, C., Hsieh, P., and Yang, W. (2000). Crystal structures of mismatch repair protein MutS and its complex with a substrate DNA. Nature 407, 703–710.

Hickman, M.J., and Samson, L.D. (2004). Apoptotic signaling in response to a single type of DNA lesion, O(6)-methylguanine. Mol. Cell 14, 105–116.

Oeda, K., Horiuchi, T., and Sekiguchi, M. (1982). The uvrD gene of E. coli encodes a DNA-dependent ATPase. Nature 298, 98–100.

Hunter, N., and Borts, R.H. (1997). Mlh1 is unique among mismatch repair proteins in its ability to promote crossing-over during meiosis. Genes Dev. 11, 1573–1582.

Ogasawara, N., Moriya, S., Mazza, G., and Yoshikawa, H. (1986). A Bacillus subtilis dnaG mutant harbours a mutation in a gene homologous to the dnaN gene of Escherichia coli. Gene 45, 227–231.

Jeruzalmi, D., Yurieva, O., Zhao, Y., Young, M., Stewart, J., Hingorani, M., O’Donnell, M., and Kuriyan, J. (2001). Mechanism of processivity clamp opening by the delta subunit wrench of the clamp loader complex of E. coli DNA polymerase III. Cell 106, 417–428.

Orth, K., Hung, J., Gazdar, A., Bowcock, A., Mathis, J.M., and Sambrook, J. (1994). Genetic instability in human ovarian cancer cell lines. Proc. Natl. Acad. Sci. USA 91, 9495–9499.

Johnson, A., and O’Donnell, M. (2005). Cellular DNA replicases: components and dynamics at the replication fork. Annu. Rev. Biochem. 74, 283–315. Kadyrov, F.A., Dzantiev, L., Constantin, N., and Modrich, P. (2006). Endonucleolytic function of MutLalpha in human mismatch repair. Cell 126, 297–308. Karamata, D., and Gross, J.D. (1970). Isolation and genetic analysis of temperature-sensitive mutants of B. subtilis defective in DNA synthesis. Mol. Gen. Genet. 108, 277–287. Kleczkowska, H.E., Marra, G., Lettieri, T., and Jiricny, J. (2001). hMSH3 and hMSH6 interact with PCNA and colocalize with it to replication foci. Genes Dev. 15, 724–736. Kongsuwan, K., Dalrymple, B.P., Wijffels, G., and Jennings, P.A. (2002). Cellular localisation of the clamp protein during DNA replication. FEMS Microbiol. Lett. 216, 255–262.

Prolla, T.A., Christie, D.M., and Liskay, R.M. (1994a). Dual requirement in yeast DNA mismatch repair for MLH1 and PMS1, two homologs of the bacterial mutL gene. Mol. Cell. Biol. 14, 407–415. Prolla, T.A., Pang, Q., Alani, E., Kolodner, R.D., and Liskay, R.M. (1994b). MLH1, PMS1, and MSH2 interactions during the initiation of DNA mismatch repair in yeast. Science 265, 1091–1093. Rayssiguier, C., Thaler, D.S., and Radman, M. (1989). The barrier to recombination between Escherichia coli and Salmonella typhimurium is disrupted in mismatch-repair mutants. Nature 342, 396–401. Schofield, M.J., and Hsieh, P. (2003). DNA mismatch repair: molecular mechanisms and biological function. Annu. Rev. Microbiol. 57, 579–608. Schofield, M.J., Nayak, S., Scott, T.H., Du, C., and Hsieh, P. (2001). Interaction of Escherichia coli MutS and MutL at a DNA mismatch. J. Biol. Chem. 276, 28291–28299.

300 Molecular Cell 29, 291–301, February 15, 2008 ª2008 Elsevier Inc.

Molecular Cell b Clamp Recruits MutS

Shell, S.S., Putnam, C.D., and Kolodner, R.D. (2007). The N terminus of Saccharomyces cerevisiae Msh6 is an unstructured tether to PCNA. Mol. Cell 26, 565–578.

Williamson, M.S., Game, J.C., and Fogel, S. (1985). Meiotic gene conversion mutants in Saccharomyces cerevisiae. I. Isolation and characterization of pms1-1 and pms1-2. Genetics 110, 609–646.

Simmons, L.A., Grossman, A.D., and Walker, G.C. (2007). Replication is required for the RecA localization response to DNA damage in Bacillus subtilis. Proc. Natl. Acad. Sci. USA 104, 1360–1365.

Wu, T.H., and Marinus, M.G. (1999). Deletion mutation analysis of the mutS gene in Escherichia coli. J. Biol. Chem. 274, 5948–5952.

Smith, B.T., Grossman, A.D., and Walker, G.C. (2001). Visualization of mismatch repair in bacterial cells. Mol. Cell 8, 1197–1206. Umar, A., Buermeyer, A.B., Simon, J.A., Thomas, D.C., Clark, A.B., Liskay, R.M., and Kunkel, T.A. (1996). Requirement for PCNA in DNA mismatch repair at a step preceding DNA resynthesis. Cell 87, 65–73. Viswanathan, M., Burdett, V., Baitinger, C., Modrich, P., and Lovett, S.T. (2001). Redundant exonuclease involvement in Escherichia coli methyldirected mismatch repair. J. Biol. Chem. 276, 31053–31058.

Yoshioka, K., Yoshioka, Y., and Hsieh, P. (2006). ATR kinase activation mediated by MutSalpha and MutLalpha in response to cytotoxic O6-methylguanine adducts. Mol. Cell 22, 501–510. Zhang, Y., Yuan, F., Presnell, S.R., Tian, K., Gao, Y., Tomkinson, A.E., Gu, L., and Li, G.M. (2005). Reconstitution of 50 -directed human mismatch repair in a purified system. Cell 122, 693–705.

Molecular Cell 29, 291–301, February 15, 2008 ª2008 Elsevier Inc. 301