INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The Saccharomyces cerevisiae Rad27p belongs to a family of evolutionarily conserved exo- and endonucleases that are involved in DNA replication and repair (12, 13, 36, 41, 46). Biochemical studies, which have included analyses of cell extracts that are competent for simian virus 40 DNA replication, indicated that Fen1p, the mammalian homolog of Rad27p, is required in lagging-strand DNA replication (11, 22, 26, 50, 51, 53). In vitro, Fen1p and Rad27p display a 5'-to-3' exonuclease activity and an endonuclease activity that is specific to nucleic acid branch structures (1, 14, 20). The 5'-to-3' exonuclease activity is thought to be required to remove single ribonucleotides that remain at the 5' ends of Okazaki fragments following cleavage of the RNA primer by RNase H (51). The endonuclease activity is hypothesized to be required for the cleavage of flap structures that are generated as the result of DNA synthesis from an upstream Okazaki fragment that displaces a downstream RNA primer (12, 13). Both activities are thought to be directed by the same catalytic site in Fen1p and Rad27p (19, 21, 29). Subsequent studies have shown that Fen1p and Rad27p track along the length of the single-stranded DNA tail before cleaving at the branch junction (16, 33). This cleavage reaction also is stimulated by polymerase processivity factor PCNA through the enhancement of Fen1p binding at the point where the flap anneals to the template, i.e., the cleavage site (5, 9, 25, 28, 49, 56).

Genetic analyses of S. cerevisiae also have supported a role for RAD27 in DNA replication. rad27 mutants display a conditional growth phenotype. At 37°C, rad27 cells arrest in S phase and display a single nuclear body (37, 42, 52). Consistent with a role in DNA replication, strains bearing rad27 mutations in conjunction with a pol3 (Pol) mutation are inviable (7, 27). Some of the defects observed in rad27 strains, including conditional viability, are suppressed by overexpression of 5'-to-3' exonuclease Exo1p, suggesting that some of the Rad27p functions are redundant (47). This idea is also supported by the observation that rad27 exo1 strains are inviable (35, 47).

In addition to exhibiting DNA replication defects, rad27 strains display DNA repair defects and a chromosome instability phenotype. The rad27 strains are highly sensitive to the DNA-damaging agent methyl methanesulfonate (MMS) but are only moderately sensitive to UV light and are insensitive to gamma irradiation (37, 52). The rad27 mutants also display synthetic lethality with mutants defective in the RAD52 double-strand break repair (DSBR) pathway, suggesting that replication lesions generated in the absence of Rad27p are processed by homologous recombination (45, 48). In assays that measure chromosome instability, rad27 strains display a complex mutator phenotype. The rad27 mutants have a high rate of forward mutations to canavanine resistance; the mutations primarily consist of duplications within the CAN1 gene of 5- to 100-bp DNA sequences that are flanked by direct repeats (48). Duplication mutations similar to those found in rad27 strains also are observed in human tumors and inherited diseases (reviewed in reference 48). In an assay that measures dinucleotide repeat tract instability, rad27 strains display instability rates similar to those observed for mismatch repair mutants (23, 42). Double-mutant and mutation spectrum analyses, however, showed that Rad27p plays a role in mutation avoidance that is distinct from mismatch repair (48). Based on these observations and the finding that rad27 rad52 double mutants are not viable, Tishkoff et al. (48) hypothesized that Rad27p endonuclease activity prevents mutations by efficiently cleaving 5' flaps generated through the displacement of downstream Okazaki fragments by extension of the upstream fragment. In their model, the absence of Rad27p cleavage results in DNA breakage at the flap junction that is repaired by a nonmutagenic DSBR mechanism or by a slip-pairing DSBR mechanism that results in duplication mutations.

The wide range of defects observed in rad27 strains in DNA replication, DNA repair, and chromosome instability assays suggests that Rad27p may perform partly or completely different functions in each of these processes. Alternatively, these defects could be due to the disruption of a single function. To explore this issue, we characterized two rad27 alleles, rad27-G67S and rad27-G240D, that were identified in a screen for mutants that displayed both a repeat tract instability and mutator phenotype. Strains containing either rad27 mutation were viable at 37°C and highly resistant to MMS and did not display synthetic lethality with an exo1 mutation. However both rad27 mutations conferred synthetic lethality with the rad52 mutation. The two rad27 mutations produced distinct properties in canavanine resistance and repeat tract instability assays. In addition, the Rad27 mutant proteins displayed differences in their biochemical properties. This suggests that distinct structures in Rad27p are responsible for different functions in replication and repair.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Media and chemicals. Yeast strains were grown in either yeast extract-peptone-dextrose (YPD) or minimal selective media (39). Sporulation plates were prepared as described previously (6). When required, canavanine was included in minimal selective media at 60 mg/liter (39) and MMS (Aldrich) was included in YPD media at 0.004 to 0.020% (vol/vol). 5-Fluoro-orotic acid (5-FOA) was purchased from U.S. Biologicals and used as described previously (2).

S. cerevisiae strains. The genotypes of all strains used in these studies are shown in Table 1. All strains were derived from the isogenic FY strain background (55). The msh2::hisG, rad52::URA3, exo1::HIS3, and rad27::HIS3 alleles have complete or nearly complete coding region deletions of their respective genes and were introduced into FY23 and FY86 by single-step transplacement. Double-mutant combinations of the alleles described in Table 1 were made by standard crosses (39). The genotypes of all of the alleles was confirmed by PCR analysis and DNA sequencing of chromosomal DNA isolated from the transformed strains.

Plasmids. pEAA67 (MLH1 PMS1 MSH2 ARSH4 CEN6 URA3) and pK5 ([GT]₁₄G-LACZ 2µm LEU2; kindly provided by Richard Kolodner) were introduced into wild-type strain FY86 prior to mutagenesis (see below) to avoid identifying mutators resulting from defects in the previously characterized MLH1, MSH2, and PMS1 genes. The MSH2, MLH1, and PMS1 genes in pEAA67 are each expressed under their native promoters. pK5 contains an out-of-frame GT repeat sequence inserted into the LACZ open reading frame. Previous studies indicated that nearly 100% of mismatch repair-defective colonies containing pK5 display a blue colony phenotype due to DNA slippage events in the GT repeat tract; in contrast, less than 0.5% of wild-type colonies show a blue colony phenotype (44).

Rad27p substrates. Oligonucleotides were designed to form nicked or flap substrates. For the nicked substrate, two primers were annealed to a template such that the upstream and downstream primers formed a nick. Flap substrates were generated by including sequences not complementary to the template at the 5' end of the downstream primer. These sequences form the unannealed 5' tail or flap. Upstream primers were annealed to the template such that they formed a nick at the base of the flap. Oligomer sequences are listed in Table 2.

Genetic techniques. Yeast was transformed with DNA using the lithium acetate method as described by Geitz and Schiestl (8). Tetrads were dissected on YPD plates immediately after Zymolyase treatment using previously established methods (39). All tetrads that yielded four, three, and sometimes two and one viable spores were examined for relevant genetic markers by PCR or by segregation of a linked marker (e.g., rad27::HIS3 rad52::URA3 exo1::HIS3).

View this table: [in this window] [in a new window]   TABLE 3.   Median frequencies of forward mutations in wild type, rad27, and msh2 strainsa

Nucleic acid techniques. All restriction endonucleases were purchased from New England Biolabs (Beverly, Mass.) and used according to the manufacturer's specifications. Taq DNA polymerase was purchased from Perkin-Elmer Cetus. Yeast chromosomal DNA was prepared as described by Holm et al. (18). PCR was performed as described previously (40), and amplification conditions and primer sequences for the different reactions are available upon request. The DNA primer synthesis and DNA sequencing was performed at the Cornell Biotechnology Analytical/Synthesis facility.

Expression and purification of wild-type and mutant Rad27 proteins. The expression and purification of wild-type and mutant Rad27p will be described in detail elsewhere. In brief, Rad27, rad27-G67S, and rad27-G240D proteins were expressed in Escherichia coli using T7 expression vector pET-24b. After transformation into E. coli strain BL21(DE3) codon plus (Stratagene, Inc.), cultures were grown until the optical density at 600 nm reached ~0.5. The addition of IPTG (isopropyl--D-thiogalactopyranoside) results in the expression of wild-type or mutant Rad27p with the addition of a six-His tag at the C terminus. After lysis by a French press, proteins were separated in successive chromatographic columns containing Ni⁺-agarose (Qiagen, Inc.), carboxymethyl-Sepharose, Mono-S, hydroxyapatite, and phenyl-Sepharose resins to obtain highly purified (>95%) enzyme fractions.

Nuclease assays. Amounts of substrate indicated in the figure legends and wild-type or mutant Rad27p were incubated in reaction buffer (30 mM HEPES [pH 7.6] diluted from a 1 M stock, 40 mM KCl, 8 mM MgCl_2, 0.01% NP-40, and 0.1 mg of bovine serum albumin/ml) in a final volume of 20 µl. Assay mixtures were incubated at 30°C for 15 min, and reactions were stopped by the addition of 10 µl of termination dye (95% formamide [vol/vol] with bromophenol blue and xylene cyanole). After a 95°C incubation for 5 min, samples were separated on a 12 or 18% polyacrylamide-7 M urea denaturing gel. The gels were quantified using a PhosphorImager (Molecular Dynamics) and analyzed using Imagequant, version 1.2, software from Molecular Dynamics. In all studies, the amounts of substrate and product(s) were quantitated and the percentages of product formed were determined by the product/(substrate plus product) ratio. This method allowed for the correction of any loading errors between lanes. All assays were performed at least in triplicate, and representative assays are shown.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of rad27 alleles. The rad27-G67S and rad27-G240D mutations were acquired in a two-part screen to identify mutants that displayed both mutator and DNA slippage phenotypes (see Materials and Methods) (57). The rad27 mutations map to the N (G67) and I (G240) nuclease domains. These domains are highly conserved in eukaryotic, prokaryotic, and viral 5' nucleases, which include the Fen1p family of DNA flap endonucleases, XPG endonuclease, T4 phage RNase H, and E. coli DNA polymerase I (7, 41). The glycine 67 residue is conserved only in the DNA flap endonuclease family, while the glycine 240 residue is conserved among all family members (41). Other mutations in these nuclease domains, none of which correspond to the G67S and G240D mutations described here, affect Rad27p nuclease activity on DNA flap substrates but do not necessarily disrupt DNA binding (7, 41).

View larger version (28K): [in this window] [in a new window]   FIG. 1.   Resistance of rad27 strains to MMS. Saturated cultures of wild-type (FY86), rad27 (EAY545), rad27-G67S (EAY595), and rad27-G240D (EAY597) strains were diluted in water and spotted in 10-fold serial dilutions (101 to 105) onto YPD media containing 0 to 0.020% MMS. The plates were photographed after a 3-day incubation at 30°C.

rad27-G67S and -G240D mutations confer distinct phenotypes in forward mutation and DNA slippage assays. rad27 strains display a strong mutator phenotype that was assessed by measuring resistance to toxic arginine analog canavanine. In strains bearing mismatch repair mutations such as the msh2 strain, canavanine resistance primarily results from forward mutations in CAN1 which include nucleotide misincorporation and single-nucleotide-deletion events (30). In rad27 strains, these forward mutations primarily result from duplication events that occur between short repeated sequences in the CAN1 gene (48). Compared to the wild type, rad27 strains showed a 150-fold increase in canavanine resistance frequency, a value that is twofold higher than those observed in msh2 strains. The rad27-G67S and rad27-G240D strains displayed canavanine resistance frequencies that were 20- and 61-fold higher, respectively, than that for the wild type. The difference in frequency between the rad27-G67S and rad27-G240D strains was statistically significant (P = 0.0008). A previous double-mutant analysis indicated a nonepistatic relationship between rad27 and mismatch repair mutations (48). As shown in Table 3, the frequency of canavanine resistance in rad27 msh2 strains (535-fold increase) was greater than the frequencies of the single mutant strains (msh2, 74-fold increase; rad27, 155-fold increase). A similar observation was made in the analysis of msh2 rad27-G67S and -G240D double mutants.

View larger version (33K): [in this window] [in a new window]   FIG. 2.   rad27-G240D and rad27-G67S strains display different insertion-deletion phenotypes at the CAN1 locus. Independent Canr colonies were obtained from EAY595 (rad27-G67S), EAY597 (rad27-G240D), EAY545 (rad27), and FY86 (wild-type) strains, and a DNA fragment containing the CAN1 open reading frame was amplified by PCR from each of these resistant colonies and digested with HphI, resulting in 490-, 411-, 314-, 252-, 249-, and 207-bp fragments, which could be detected by gel electrophoresis (57). Two smaller bands of 87 and 46 bp could not be detected. Lanes A and B, DNA fragments from the wild-type CAN1 locus in FY86; lanes 1 to 12 and 13 to 25, HphI-digested CAN1 DNA from Canr rad27-G67S and rad27-G240D, respectively. Asterisks, lanes containing bands that differ from the wild-type CAN1 pattern.

Biochemical analysis of Rad27 mutant proteins. To examine the biochemical properties of rad27-G67S and rad27-G240D proteins, we expressed Rad27p and the two mutant proteins in E. coli. Rad27p and rad27-G67Sp were expressed and purified similarly, with final preparations at more than 95% homogeneity (data not shown). Expression of rad27-G240Dp resulted in a number of truncation products whose presence was not reduced by the addition of protease inhibitors. Through additional chromatographic steps it was found that the final protein fraction contained full-length rad27-G240Dp, active for catalysis, and one truncation product representing ~20% of the preparation that did not contain any detectable activity (data not shown). Using these recombinant enzymes, we examined the activity of each mutant on a set of standard nicked and flap substrates to determine the effect of the point mutations (16).

View larger version (71K): [in this window] [in a new window]   FIG. 3.   Ability of rad27-G67Sp and rad27-G240Dp to bind to a flap substrate. Wild-type and mutant Rad27p was incubated with a substrate containing a six-nucleotide flap radiolabeled at the 3' end of the downstream primer. Substrate (5 fmol) containing primers D6nt,T44, and U25 was incubated with increasing amounts of enzyme, as indicated, in reaction buffer (see Materials and Methods) without MgCl2 at 25°C for 8 min in a total volume of 20 µl. After the addition of 2 µl of 50% glycerol, DNA and DNA-enzyme complexes were separated by electrophoresis on a 1% agarose-0.5% polyacrylamide gel in 0.25× Tris-borate-EDTA at 4°C. The gel was dried onto DE81 Whatman filter paper, and products were detected using a PhosphorImager. Reactions mixtures contained either Rad27p, rad27-G67Sp (G67S), or rad27-G240Dp (G240D), as noted. Lanes "boiled," substrate incubated at 100°C prior to loading to indicate the location of unannealed substrate. A schematic diagram of the substrate is at the top. The lengths of substrates and products are in nucleotides. Asterisk, position of the radiolabeled nucleotide.

View larger version (38K): [in this window] [in a new window]   FIG. 4.   Endonucleolytic cleavage of flap substrates by rad27-G67Sp and rad27-G240Dp. (A) Wild-type or mutant Rad27p was incubated with a substrate containing a six-nucleotide flap. Substrate (5 fmol) containing primers D6nt, T44, and U25 was incubated with increasing amounts of enzyme (10, 25, 50, 100, 500, and 1,000 fmol) at 30°C as described in Materials and Methods and separated by electrophoresis on a denaturing 12% polyacrylamide gel. Reaction mixtures contained either Rad27p (lanes 2 to 7), rad27-G67Sp (G67S; lanes 9 to 14), or rad27-G240Dp (G240D; lanes 16 to 21). Reaction mixtures in lanes 1, 8, and 15 contained only substrate. A schematic diagram of the substrate is at the top. The lengths of substrates and products are in nucleotides. Asterisk, position of radiolabeled nucleotide. (B) Increasing amounts of Rad27p (solid lines), rad27-G67Sp (G67S; dotted lines), and rad27-G240Dp (G240D; dashed lines) were incubated with substrates containing a nick (no flap) (circle) or a 6-nucleotide (square) or 15-nucleotide (triangle) flap as described for panel A. Results of the experiments with the six-nucleotide flap substrate are graphical representations of the gels presented in panel A. The products were detected using a PhosphorImager, quantitated, and presented as the percentages of substrate converted to product versus amounts of enzyme. The nicked substrate contains primers Dnick, T44, and U25, and the 15- nucleotide flap contains primers D15nt, T44, and U25.

View larger version (27K): [in this window] [in a new window]   FIG. 5.   Endo- and exonucleolytic cleavage of 3'-end labeled substrates by rad27-G67Sp and rad27-G240Dp. (A) Wild-type or mutant Rad27p was incubated with a substrate containing a six-nucleotide flap radiolabeled at the 3' end of the downstream primer. Substrate (5 fmol) containing primers D6nt, T44, and U25 was incubated with increasing amounts of enzyme (10, 25, 50, 100, 500, and 1,000 fmol) at 30°C as described in Materials and Methods and separated by electrophoresis on a denaturing 18% polyacrylamide gel. Reaction mixtures contained either Rad27p (lanes 2 to 7), rad27-G67Sp (G67S; lanes 9 to 14), or rad27-G240Dp (G240D; lanes 16 to 21). Reaction mixtures in lanes 1, 8, and 15 contained only substrate. A schematic diagram of the substrate is at the top. The lengths of substrates and products are in nucleotides. Asterisk, position of radiolabeled nucleotide. (B) Rad27p or rad27-G67Sp was incubated with a substrate containing a nick (lanes 1 to 14) or one-nucleotide gap (lanes 15 to 28) radiolabeled at the 3' end of the downstream primer. Substrate (5 fmol) was incubated with increasing amounts of enzyme (10, 20, 40, 60, 80, and 100 fmol) at 30°C as described in Materials and Methods and separated by electrophoresis on a denaturing 18% polyacrylamide gel. Reaction mixtures contained either Rad27p (lanes 2 to 7 and 16 to 21) or rad27-G67Sp (G67S; lanes 9 to 14 and 22 to 28). Reaction mixtures in lanes 1, 8, 15, and 22 contained only substrate. Schematic diagrams of the substrates are shown. The nicked substrate contains primers Dnick, T44, and U25, and the one-nucleotide gap substrate contains primers Dnick, T44, and U24.

Activity of wild-type and mutant Rad27 proteins on double-flap and bubble substrates. Double-flap structures have been hypothesized to reflect an intermediate formed in vivo in end-joining and homologous-recombination pathways (14). These structures might also be an intermediate in nick translation reactions widely used in DNA replication and repair. Several homologs of Rad27p from Eucarya and Archaea were shown to display an enhanced endonuclease activity on substrates containing a double-flap structure. This enhanced activity was identified by altering the length of the upstream primer (14, 31, 49). For the experiments presented in Fig. 6, the upstream primer for the six-nucleotide flap substrate (Fig. 4A and 5A) was extended one nucleotide to form a 3' flap adjacent to the 5' flap formed by the downstream primer. We tested the activity of each mutant on such a substrate (Fig. 6). Both the wild type (lanes 2 to 6) and rad27-G67Sp (lanes 7 to 11) easily cleaved this substrate. In fact, rad27-G67Sp cleaved the double flap (87%) more efficiently than the conventional flap, almost reaching the cleavage level of Rad27p (95%). Although rad27-G240Dp effectively failed to cleave standard flap and nick substrates, the double-flap structure was cleaved (60%) (lanes 12 to 15). However, after endonucleolytic cleavage, neither rad27-G67Sp nor rad27-G240Dp was able to carry out subsequent exonucleolytic cleavage (data not shown).

View larger version (65K): [in this window] [in a new window]   FIG. 6.   Cleavage of a double-flap substrate by rad27-G67Sp and rad27-G240Dp. Wild-type or mutant Rad27p was incubated with a substrate containing a six-nucleotide 5' flap and a one-nucleotide 3' flap radiolabeled at the 5' end of the downstream primer. Substrate (5 fmol) containing primers D6nt, T44, and U26 was incubated with increasing amounts of enzyme (10, 50, 100, 500, and 1,000 fmol) at 30°C as described in Materials and Methods and separated by electrophoresis on a denaturing 12% polyacrylamide gel. Reaction mixtures contained either Rad27p (lanes 2 to 6), rad27-G67Sp (G67S; lanes 7 to 11), or rad27-G240Dp (G240D; lanes 12 to 16). Lane 1 contained only substrate. A schematic diagram of the substrate is shown. The lengths of substrates and products are in nucleotides. Asterisk, position of radiolabeled nucleotide.

View larger version (82K): [in this window] [in a new window]   FIG. 7.   Cleavage of intermediates of repeat expansion by rad27-G67Sp and rad27-G240Dp. Increasing amounts of either wild-type or mutant Rad27p (10, 20, 40, 60, 80, and 100 fmol) were incubated with a 3' radiolabeled bubble substrate containing primers Dbubble and Tbubble. Each primer contained both 5' and 3' regions of complementarity resulting in the formation of a bubble. Reaction mixtures contained Rad27p (lanes 2 to 7), rad27-G67Sp (G67S; lanes 9 to 14), or rad27-G240Dp (G240D; lanes 16 to 21). Reaction mixtures in lanes 1, 8, and 15 contained only substrate. Schematic diagrams of the substrates are shown. The lengths of substrates and products are in nucleotides. Asterisk, position of radiolabeled nucleotide.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We isolated rad27 alleles that displayed specific properties in chromosome stability assays and in their ability to cleave substrates that mimic structures formed during lagging-strand DNA synthesis. These alleles contain mutations that map to distinct nuclease domains (N for rad27-G67Sp, I for rad27-G240Dp) that were previously identified in RAD27 (7, 41). Like the rad27 allele, the rad27-G67S and -G240D alleles were inviable in the recombination-deficient rad52 strain background. Unlike rad27 strains, rad27-G67S and -G240D strains were resistant to MMS, were viable in an exo1 strain background, and were also viable at 37°C. The rad27-G67S and -G240D strains, however, could be distinguished from each other with respect to their chromosome instability phenotype and flap cleavage activities. These properties have made the mutant strains particularly useful in probing the role of Rad27p in maintaining chromosomal integrity. In chromosome stability assays, rad27-G67S strains displayed a higher frequency of repeat tract instabilities than of CAN1 insertion-deletion events; in contrast, the rad27-G240D strains displayed the opposite phenotype. In biochemical assays, rad27-G67Sp and rad27-G240Dp were found to have distinct defects in nuclease function, suggesting that catalytic deficiency is the basic cause of the biological defects in the mutant strains. rad27-G67Sp displayed a weak exonuclease activity but showed a double-flap endonuclease activity that was similar to that for the wild type and a single-flap endonuclease activity that was reduced only twofold. In contrast, rad27-G240Dp was devoid of exonuclease activity, displayed a double-flap endonuclease activity that was 60% of that for the wild type, and showed an extremely weak single-flap endonuclease activity.

Models to explain the mutator and DNA slippage phenotypes observed in rad27 strains. Models have been developed to correlate the biochemical and genetic phenotypes exhibited by rad27 mutants. Tishkoff et al. (48) suggested that Rad27p endonuclease activity plays an important role in preventing insertion-deletion mutations that are observed in rad27 strains by cleaving 5' flaps generated through the displacement of downstream Okazaki fragments by extension of the upstream fragment. Any delays in this cleavage increase the probability of DNA breakage at the flap junction. DSBR mechanisms are then thought to act on these break sites. Support of this model was provided by genetic studies which showed that rad27 mutants are inviable in recombination-defective strain backgrounds and also display a duplication mutator phenotype that is thought to result from DSBR (45, 48).

View larger version (15K): [in this window] [in a new window]   FIG. 8.   Proposed model for duplications and expansions. Displacement from DNA synthesis causes the formation of a flap structure (steps 1 and 2). Some sequences are able to adopt a secondary structure that forms either foldbacks or bubble intermediates (arrows). Rad27p binds to the 5' end of the foldback or bubble and exonucleolytically degrade the annealed portions, resulting in the formation a more preferred flap structure (steps 2a and 3a). Step 4a, cleavage of the flap followed by synthesis and ligation complete synthesis. In this model, absence of or reduction in Fen1p activity (cross) permits the ligation of the bubble intermediate to an upstream primer (step 3b). Resolution of this structure by repair DNA synthesis or a subsequent round of DNA replication leads to expansion (step 4b).

Correlations between the genetic and biochemical defects in rad27-G67S strains. We hypothesize that the high ratio of DNA slippage to Can^r mutator events observed in rad27-G67S strains resulted largely from defects in nuclease function that are important for degrading bubble intermediates, which can lead to DNA slippage events. As shown in Fig. 5 and 7, rad27-G67Sp displayed an extremely weak exonuclease activity on bubble, nicked, and gapped substrates. This defect correlates well with the relatively high level of DNA slippage events observed in rad27-G67S strains (Table 5). Genetic studies involving the Exo1 double-stranded 5'-to-3' exonuclease were also consistent with the conclusion that the exonuclease defect is a major cause of the mutant phenotype in rad27-G67S strains. As shown in Table 3, the CAN1 mutator defect observed in rad27-G67S strains was almost completely suppressed by Exo1p overexpression, whereas the rad27 and rad27-G240D alleles were only partially suppressed. In contrast to its weak exonuclease activity, rad27-G67Sp displayed a flap endonuclease activity that was reduced only about twofold compared to that for the wild type for single-flap structures (Fig. 4B) and was similar to that for the wild type for double-flap structures (Fig. 6). This relatively active endonuclease activity correlates well with the low frequency of Can^r mutations and the weak insertion-deletion phenotype that was observed in rad27-G67S strains (Table 3).

Analysis of rad27-G240Dp. Our results demonstrate that the catalytic defect in rad27-G240Dp is distinctly different. As shown in Fig. 4, rad27-G240Dp displayed an extremely weak flap endonuclease activity on a six-nucleotide flap substrate and a moderate activity on a double-flap substrate. In contrast to what was found for rad27-G67S, Exo1p overexpression only partially suppressed the CAN1 mutator phenotype of rad27-G240D. The level of suppression, to ~15-fold higher than the wild-type level, was similar to that observed for rad27 strains overexpressing Exo1p (Table 3). This genetic observation, in combination with the relatively high frequency of the Can^r and strong-insertion-deletion phenotype observed in rad27-G240D strains, is consistent with a substantial defect in rad27-G240Dp endonuclease activity.

Genetic and biochemical studies are consistent with a structural role for Rad27p. The biochemical characteristics of the two mutant forms of Rad27p are not able to completely explain the relative differences of the mutant strains in canavanine resistance and repeat tract instability assays as well as their resistance to MMS. A likely explanation is that Rad27p serves a structural role that is also perturbed by the mutations and that Rad27p has distinct structural and catalytic functions. In this scenario, a combination of Rad27p structural and catalytic defects results in the observed phenotypes. In support of this idea, Rad27p has been shown to interact with other replication factors (see above) and recent reports have suggested that large complexes of proteins exist in eukaryotic cells that contain both DNA repair and replication factors. The absence of key components could therefore disrupt these complexes (54).

Current and future analyses of Rad27p mutants. Although the crystal structure of S. cerevisiae Rad27p has not been solved, the structure of several homologs including T5 exonuclease (4), T4 RNase H (32), Methanococcus jannaschii Fen1p (21), and Pyrococcus furiosus Fen1p (PfFen1p) (19) are available. Based on the crystal structure of PfFen1p, the glycines at positions 67 and 240 are located within different regions of the enzyme. Glycine 67 is part of an  helix proposed to form one side of the catalytic groove of the enzyme. Mutation of this residue is consistent with the observed reduction in catalytic activity. Glycine 240 is located within a helix-turn-helix motif. This region is proposed to participate in the binding of Fen1p to the template strand of the substrate. In our studies, we do not detect major changes in the binding of the rad27-G240Dp to a flap substrate. However, the mutation may have more-subtle effects that alter the substrate sequence or structure requirements for efficient cleavage. We are currently performing extensive measurements of substrate specificity with this mutant.