Pol32, a Subunit of Saccharomyces cerevisiae DNA Polymerase , Suppresses Genomic Deletions and Is Involved in the Mutagenic Bypass Pathway

Corresponding author: Meng-Er Huang, “Génétique et Développement,” Faculté de Médecine, 2 ave. du Professeur Léon Bernard, 35043 Rennes, France., huang{at}univ-rennes1.fr (E-mail)

Communicating editor: A. NICOLAS

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The Pol32 subunit of S. cerevisiae DNA polymerase (Pol) plays an important role in replication and mutagenesis. Here, by measuring the CAN1 forward mutation rate, we found that either POL32 or REV3 (which encodes the Pol catalytic subunit) inactivation produces overlapping antimutator effects against rad mutators belonging to three epistasis groups. In contrast, the msh2 pol32 double mutant exhibits a synergistic mutator phenotype. Can^r mutation spectrum analysis of pol32 strains revealed a substantial increase in the frequency of deletions and duplications (primarily deletions) of sequences flanked by short direct repeats, which appears to be RAD52 and RAD10 independent. To better understand the pol32 and rev3 antimutator effects in rad backgrounds and the pol32 mutator effect in a msh2 background, we determined Can^r mutation spectra for rad5, rad5 pol32, rad5 rev3, msh2, msh2 pol32, and msh2 rev3 strains. Both rad5 pol32 and rad5 rev3 mutants exhibit a reduction in frameshifts and base substitutions, attributable to antimutator effects conferred by the pol32 and rev3 mutations. In contrast, an increase in these two types of alterations is attributable to a synergistic mutator effect between the pol32 and msh2 mutations. Taken together, these observations indicate that Pol32 is important in ensuring genome stability and in mutagenesis.

PROGRESS in understanding the relationship between genomic instability and cancer susceptibility has depended heavily on research in model organisms. The yeast Saccharomyces cerevisiae provides an ideal system for use in mutator, antimutator, and mutation spectrum analyses because of the facility of both genetic and molecular manipulations. Early studies of mutagenesis relied basically on genetic analysis of reversion, suppression, or forward mutation (reviewed in SARGENTINI and SMITH 1985 ; FRIEDBERG et al. 1995 ). However, these systems did not allow the precise DNA sequence alterations involved to be identified, nor were they able to detect all types of mutation that could occur. Mutation spectrum analysis by DNA sequencing has overcome these limitations. The data obtained on specific DNA sequence changes (base substitution, frameshift, larger deletion and insertion, inversion, and gross chromosomal rearrangement), the rates at which they arise, their site specificity, and the influence of mutator and antimutator alleles on these changes have all yielded insights into the processes that ensure genomic stability (reviewed in KUNZ et al. 1998 ). The best known example in this regard concerns the relationship between postreplicative mismatch repair (MMR) and microsatellite sequence instability (reviewed in HARFE and JINKS-ROBERTSON 2000A ). Mutations in different components of the MMR system can differentially affect the microsatellite sequence stability, and individual mutants display distinct mutation spectra (JOHNSON et al. 1996 ; MARSISCHKY et al. 1996 ; GREENE and JINKS-ROBERTSON 1997 ). These results have provided the basis of a widely accepted model in which Msh2 interacts with either Msh3 or Msh6 to form heterodimers with distinct but overlapping recognition specificities. The Msh2-Msh6 complex recognizes base:base mismatches as well as single-base deletion or addition mispairs. The Msh2-Msh3 complex also recognizes single-base deletion or addition mispairs but is primarily responsible for the recognition of larger deletion or addition mispairs. MutL homolog heterodimers bind to the Msh2-Msh6 or Msh2-Msh3 complex to effect repair. The importance of human MMR in mutation avoidance is highlighted by the demonstration that mutations in MMR genes segregate with a cancer predisposition syndrome, hereditary nonpolyposis colorectal cancer (reviewed in HARFE and JINKS-ROBERTSON 2000A ; JIRICNY and NYSTROM-LAHTI 2000 ).

Mutation rates and mutational spectra provoked by defects in several replication and repair proteins have also been characterized. For instance, a mutation affecting the proofreading exonuclease domain of the DNA polymerase (Pol) catalytic subunit (pol3-01) causes an increase in the mutation rate. Not surprisingly, the mutation spectrum of pol3-01 strains includes an increase in the accumulation of base substitutions (MORRISON and SUGINO 1994 ). Other alleles of POL3 (pol3-t, pol3-ts1, and pol3-ts11, all mutated in regions outside the exonuclease-encoding domain) and several alleles of POL30 (which encodes the DNA polymerase processivity factor PCNA) greatly increase the rate of deletions involving short direct repeat sequences, suggesting that these alleles increase replication slippage (TRAN et al. 1995 ; KOKOSKA et al. 1998 , KOKOSKA et al. 2000 ; CHEN et al. 1999 ). Null mutations of RAD27 (which encodes an Okazaki fragment-processing enzyme) increase the frequency of duplications (TISHKOFF et al. 1997A ). These duplications involve short direct repeats and are thought to be initiated by extensive strand displacement of a 5' flap on a downstream Okazaki fragment by synthesis of the upstream Okazaki fragment. A single-strand DNA loop subsequently formed by illegitimate (slip) pairing of direct repeats on the displaced 5' flap or a DNA double-strand break (DSB) followed by mutagenic single-strand annealing (SSA) are two possible intermediates in pathways that result in duplication mutations (TISHKOFF et al. 1997A ). Another example is provided by the disruption of REV3 (which encodes the catalytic subunit of Pol ). The resulting antimutator phenotypes of relatively broad specificity suggest a role of Pol in generating multiple mutations (ROCHE et al. 1994 , ROCHE et al. 1995 ; HARFE and JINKS-ROBERTSON 2000B ). Recently, a new mutational spectrum, called gross chromosomal rearrangement (GCR), has been described in S. cerevisiae. Large increases in the rate of accumulation of GCRs can be caused by rfa mutations, which confer repair and recombination defects (CHEN et al. 1998 ; CHEN and KOLODNER 1999 ); by the rad27 mutation, which causes repair and replication defects (CHEN and KOLODNER 1999 ); by mutations in genes like MRE11, RAD50, and XRS2, which are required for the repair of DSBs (CHEN and KOLODNER 1999 ); by mutations in the DNA helicase gene SGS1 or in the TOP3 gene (MYUNG et al. 2001A ); by mutations in genes like RFC5, MEC3, and DPB11, which cause defects in S-phase checkpoint functions (MYUNG et al. 2001B ); and finally, by mutations in genes such as MEC1, DDC2, RAD53, CHK1, PDS1, and DUN1, which are implicated in downstream signal transduction pathways (MYUNG et al. 2001B ). These accumulating data show that the analysis of mutation spectra to determine a particular signature of genomic instability is particularly helpful in understanding the origin of genetic alterations.

Pol , the major replicative DNA polymerase in eukaryotic cells, plays a role in bulk DNA replication, and it is also the primary DNA polymerase in most DNA repair pathways (reviewed in BURGERS 1998 ; WAGA and STILLMAN 1998 ). S. cerevisiae Pol has been defined as comprised of three subunits encoded by the genes POL3, POL31, and POL32 (GERIK et al. 1998 ). The present investigation focuses on the Pol32 subunit. pol32 cells are viable but show replication defects characterized by a higher proportion of large-budded cells with a single but duplicated DNA mass at the mother-bud neck and increased sensitivity to hydroxyurea (HUANG et al. 1997 , HUANG et al. 1999 ; GERIK et al. 1998 ). In vitro, DNA synthesis mediated by Pol lacking Pol32 is inefficient and characterized by frequent pausing (BURGERS and GERIK 1998 ). The pol32 mutant also displays DNA repair defects, with increased sensitivity to ultraviolet (UV) radiation and methylation damage, and it is also defective in UV- and N-methyl-N'-nitro-N-nitrosoguanidine-induced mutagenesis (GERIK et al. 1998 ; HARACSKA et al. 2000 ; HUANG et al. 2000 ). Genetic analyses place POL32 in the same error-prone pathway as REV3 within the RAD6 epistasis group (HUANG et al. 2000 ). Members of the RAD6 epistasis group are involved in error-free and/or error-prone bypass mechanisms (reviewed in FRIEDBERG et al. 1995 ; KUNZ et al. 2000 ; BROOMFIELD et al. 2001 ). RAD6 and RAD18 are required for both pathways. Some members, including RAD5, RAD30, MMS2, UBC13, and PCNA, are primarily involved in the error-free process while others, including REV1, REV3, REV7, and POL32, promote the error-prone process. Rad30, Rev1, and the Rev3/Rev7 complex appear to be low-processivity DNA polymerases (reviewed in KUNZ et al. 2000 ; BROOMFIELD et al. 2001 ). Several lines of evidence also indicate a role of Pol in mutagenic bypass replication. For example, damage-induced mutagenesis is defective in the pol3-13 mutant (GIOT et al. 1997 ). In vitro, Pol replicates through templates bearing O⁶-methylguanine (m6G) damage by inserting a C (30%) or a T (70%) opposite m6G, preferentially in an error-prone manner (HARACSKA et al. 2000 ). A previous genetic epistasis analysis placing POL32 in the same mutagenic bypass pathway as REV3 and the observation that rad mutators depend on a functional Pol32 further highlight the involvement of the Pol replication complex in mutagenesis (HUANG et al. 2000 ).

The present study not only examines the relative contributions of Pol32 and Rev3 to several mutator effects that influence the CAN1 forward mutation rate, but also, and more importantly, determines the canavanine resistance (Can^r) mutation spectrum for wild-type, pol32, rev3, rad5, rad5 pol32, rad5 rev3, msh2, msh2 pol32, and msh2 rev3 strains. The results of these studies indicate that the Pol32 subunit of S. cerevisiae Pol is important in ensuring genome stability and is involved in the mutagenic bypass pathway.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains, media, and general methods:
The S. cerevisiae strains used in this study are listed in Table 1. All strains are derived from FY1679, and some of these have been described previously (HUANG et al. 1999 , HUANG et al. 2000 ). The RAD5 gene was disrupted using the construct rad5::HIS3 from the plasmid pJM112 (MCDONALD et al. 1997 ), kindly provided by Jean McDonald and Roger Woodgate. The MSH2 gene was disrupted using the construct msh2::hisG-URA3-hisG from the plasmid pRDK351 (MARSISCHKY et al. 1996 ), kindly supplied by Richard Kolodner. rad5 pol32 and msh2 pol32 strains were obtained by crossing rad5 and msh2 to pol32 haploids, respectively. rad1 rev3, rad52 rev3, rad6 rev3, rad18 rev3, mms2 rev3, rad5 rev3, and msh2 rev3 strains were obtained by crossing rev3 to the respective isogenic single mutant strains. Haploid disruptant strains were obtained by sporulation and tetrad dissection. All gene disruptions were confirmed by PCR analysis using primer pairs that allowed amplification of both wild-type and mutant alleles.

Yeast strains were grown in standard media (ROSE et al. 1990 ). Cells were grown nonselectively in YPD (1% yeast extract, 2% Bacto peptone, 2% dextrose, and 2% agar for plates). Selective growth was in synthetic complete (SC) medium lacking the appropriate amino acid. SC medium lacking arginine and containing 60 mg of canavanine per liter was used to identify forward mutations in the CAN1 gene.

Measurements of the Can^r spontaneous mutation rate:
Forward mutation to canavanine resistance was determined by fluctuation tests exactly as previously described (HUANG et al. 2000 ). Briefly, cells were diluted to 100 cells/ml in 10 separate cultures for each strain and again grown to 1–2 x 10⁸ cells/ml in YPD. Cells were then harvested by centrifugation, washed once, and resuspended in sterile water. A 100-µl drop of an appropriate dilution was plated onto canavanine-containing medium (1–2 x 10⁷ cells/plate) to identify forward mutations in CAN1 and onto YPD to count viable cell number. Colonies appearing after 3 or 4 days of growth at 30° were counted. The number of Can^r colonies per 10⁷ viable cells among 10 parallel cultures was calculated and the median value from each set of 10 cultures was used to determine the spontaneous mutation rate of a given strain by the method of the median (LEA and COULSON 1948 ). All the rates reported are the average of at least three independent sets of experiments. The 95% confidence intervals for a median rate were calculated as described previously (ZAR 1996 ).

Can^r mutation spectra:
Can^r mutation spectra were determined by PCR amplification of the CAN1 gene followed by DNA sequence analysis of independent Can^r isolates. The majority of mutants analyzed were obtained from plates generated during the fluctuation analysis experiments. In some cases, independent Can^r mutants were isolated by first streaking the strain of interest for single colonies on YPD plates. Then, single colonies were patched onto plates containing canavanine and a single Can^r colony per patch was purified by streaking for single colonies on selective media. Genomic DNA was prepared and the CAN1 gene was amplified by PCR using the primers CANF1 (5'-CTTCAGACTTCTTAACTCC-3') and CANR1 (5'-GAGGGTGAGAATGCGAAAT-3'). The PCR product was then sequenced with the primers CANF2 (5'-GGCATATTCTGTCACGCAG-3'), CANR1 (5'-GAGGGTGAGAATGCGAAAT-3'), CANR2 (5'-GCCTGCAACACCAGTGATA-3'), and CANR3 (5'-GTAACAGGGATGAATGTAG-3'). All DNA sequencing was performed with an Applied Biosystems 373 DNA sequencer using Taq DNA polymerase and dye terminators according to protocols recommended by the manufacturer.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Pol32 and Rev3 are required to generate rad mutators belonging to the three epistasis groups:
Genetic data suggest that the absence of any of the DNA damage repair/bypass pathways in yeast results in the channeling of lesions into alternative pathways (SWANSON et al. 1999 ). The DNA damage bypass pathway contributes to spontaneous mutagenesis. The rev3 and pol32 mutations have been shown to decrease the magnitude of the mutator effect of the rad1, rad52, rad6, rad18, and mms2 mutations (ROCHE et al. 1994 , ROCHE et al. 1995 ; BROOMFIELD et al. 1998 ; HUANG et al. 2000 ). However, these analyses were performed on strains of different background and employed different mutational assays. To investigate and compare the roles of Pol32 and Rev3 in mutagenesis, we introduced the relevant single and double mutations into isogenic backgrounds and measured the spontaneous forward mutation rate of the CAN1 gene by fluctuation tests. Any mutation that inactivates the arginine permease encoded by CAN1 results in canavanine resistance.

No significant differences in Can^r mutation rates were observed between POL32 and pol32 strains (Table 2). The rev3 strain displays a reduced mutation rate. As expected, elimination of the nucleotide excision repair pathway (rad1), the recombinational repair pathway (rad52), and some members of the RAD6-mediated DNA damage bypass pathway (rad6, rad18, mms2, rad5) results in a rad mutator phenotype. Deletion of the POL32 or REV3 gene in various rad mutator backgrounds clearly diminishes the mutator effects, but to different extents in different mutants (Table 2). For example, the rad1 mutator effect is both POL32 and REV3 dependent. For the rad52 mutant, elimination of POL32 lowers the mutation rate, although it remains threefold greater than that in a pol32 strain. Deleting the REV3 gene has a more marked impact on the rad52 mutator effect than does deleting POL32, but the overall mutation rate for the rad52 rev3 strain is still greater than the rate for the rev3 single mutant (Table 2). Therefore, the rad52 mutator effect is partially POL32 and REV3 dependent. For the rad6, rad18, mms2, and rad5 strains, the additional deletion of POL32 reduces their mutation rates to or close to the pol32 level, indicating that their mutator phenotypes are essentially POL32 dependent. By comparison, deletion of REV3 also lowers the mutation rates of these strains but to a level that is greater than that of the rev3 strain. Taken together, we conclude that both Pol32 and Rev3 contribute to the mutator phenotypes of the rad1, rad52, rad6, rad18, mms2, and rad5 mutants and that their effects overlap to a great extent. This is consistent with the fact that Pol32 is in the same mutagenesis pathway as Rev3, as shown by epistasis analysis. By analyzing a plasmid-borne SUP4-o locus, ROCHE et al. 1995 concluded that the rad52 mutator effect is completely REV3 dependent, which is at variance with our observation that the rad52 mutation rate is only partially REV3 dependent. We believe that this difference is due simply to the different assay systems used. In fact, the SUP4-o locus differs from the majority of loci with respect to mutability (DRAKE 1991 ).

Synergistic mutator effect between the pol32 and msh2 mutations:
The mutation rate of the pol32 strain, which is similar to that of wild-type strains, does not provide a true measure of the replication errors made by DNA Pol in the absence of the Pol32 subunit, since many errors are corrected by postreplicative MMR. Furthermore, as described above, the pol32 and rev3 mutations confer antimutator effects when combined with rad mutations belonging to any of the three epistasis groups. Accordingly, we determined the effect of pol32 and rev3 mutations on the Can^r mutation rate in msh2 strains, which lack MMR activity. We found that while the mutation rate of the msh2 strain is 16-fold greater than that of the wild type, a 50-fold increase is observed in the msh2 pol32 strain. This synergy suggests that the pol32 mutation causes defects in a process distinct from MMR. MMR likely plays a role in removing the errors produced during replication by DNA Pol without the Pol32 subunit. In addition to having a highly elevated mutation rate, the msh2 pol32 strain grows relatively slowly in liquid culture, has a low plating efficiency (only 50–60% of cells form colonies), and produces colonies of variable size. In contrast, the rev3 mutation has no significant effect on the mutation rate of the msh2 strain, and the msh2 single mutant and the msh2 rev3 strain have similar growth phenotypes.

The absence of Pol32 results in an increased frequency of deletion:
To understand how Pol32 is involved in replication and mutagenesis, we determined the spectrum of Can^r mutations in wild-type and pol32 strains. The CAN1 assay was selected in view of its sensitivity to a variety of mutational events, including single-base substitutions, single-base frameshifts, larger deletions and insertions, inversions, and translocations. We sequenced the CAN1 locus from 20–30 independent Can^r isolates per strain. However, the spectrum data should be interpreted conservatively because the number of data points is low, although this number of data points is similar to or greater than that reported in most spectrum studies. In the wild-type strain, base substitutions (70% of all mutation events), +1 or -1 frameshifts (19%), and deletions and duplications of several bases (11%) were detected (Table 3). This mutation spectrum is similar to those reported by other investigators (TISHKOFF et al. 1997A ; KOKOSKA et al. 2000 ). We note that neither the two deletions nor the single duplication detected took place in tracts containing direct repeat sequences. In contrast, for 29 independent Can^r isolates from the pol32 strain (corresponding to 31 mutation events), 45% of the mutations were deletions of more than a few base pairs, ranging from 8 to 237 bp in length (Table 3 and Table 4). These deletions took place exclusively in sequences flanked by perfect direct repeats of 3–10 bp in size (Table 4). Some of the repeats are even longer if a single-base difference within a repeat pair is accepted (imperfect repeats). Each deletion involved the specific sequence between the two short direct repeats and one of the repeats (Table 4). Two duplication events (6%), both involving sequences between direct repeats, were also observed. Similarly, the duplicated sequence includes the intervening sequence and one of the two direct repeats flanking the duplicated region. The deletion or duplication of a tract that inactivates the CAN1 gene is not restricted to any particular part of the 1.8-kb open reading frame. However, it is noteworthy that (i) a 16-bp sequence (nucleotides 1324–1339 of CAN1) was deleted in two independent clones, and (ii) a 27-bp sequence (nucleotides 284–310 of CAN1) was deleted in three, and duplicated in one, of four independent clones (Table 4). In the rad5 pol32 strain we subsequently found two 16-bp deletions of positions 1324–1339 of the CAN1 gene (Table 4). These identical changes represent independent events, since each Can^r clone was derived from an independent isolate. However, no particular secondary structure was revealed for the nucleotide sequence of these two regions by using structural prediction programs, which suggests that these deletion hotspots may not be due to such features. Another characteristic of the Can^r mutation spectrum in pol32 strains appears to be the absence of +1 or -1 frameshift mutations. However, this distribution is not significantly different from that seen in wild-type strains, for which this mutation class represents 19% of all events (P > 0.05, two-tailed Fisher exact test).

View this table: In this window In a new window   Table 4. Sequences flanking large deletions and duplications observed in pol32 and rad5 pol32 strains

The mutational spectrum of the rev3 strain was also determined. Of 23 mutagenic events, 87% (20/23) consisted of base substitutions and 9% (2/23) of frameshifts. Taking into account the decrease in overall mutation rate, the specific rates of base substitutions and frameshifts are still below those of the REV3 strain (Table 3). One mutation event resulted in an 8-bp deletion of a sequence flanked by 2-bp direct repeats. The reduction in the rates of base substitutions and frameshifts is likely attributable to the rev3 antimutator effect. This is in keeping with other observations implying that Pol can process a variety of spontaneous DNA lesions (ROCHE et al. 1995 ; HARFE and JINKS-ROBERTSON 2000B ). The striking difference between the mutation spectra of the rev3 and pol32 mutants concerns the relative frequencies of deletions of sequences flanked by short direct repeats.

The frequency of deletion formation does not depend on RAD52 and RAD10:
Deletions of sequences flanked by short stretches of direct repeats may arise from the repair of DSBs by mutagenic SSA (TISHKOFF et al. 1997A ; CHEN et al. 1998 ; KOKOSKA et al. 2000 ). This model predicts that the frequency of deletions should be reduced by elimination of the repair pathway responsible for their production. We tested this by examining the effect of the rad52 and rad10 mutations on deletion formation in the pol32 background. The rad52 mutant is defective in both mutagenic SSA and DSB repair (SUGAWARA and HABER 1992 ; MORTENSEN et al. 1996 ). Rad10 is required for mutagenic SSA, because it removes the nonhomologous tails that arise during SSA (IVANOV and HABER 1995 ). Instead of sequencing all the independent Can^r isolates from the rad52 pol32 and rad10 pol32 double mutants, we performed a more limited analysis: Only those mutants that exhibited apparent deletions or additions in a PCR screen were selected for sequencing. This PCR screen consisted of amplifying several overlapping fragments covering the CAN1 locus and allowed for the detection of deletions or additions as small as 10 bp by gel electrophoresis. The frequency of deletion thus detected in the rad52 pol32 and rad10 pol32 double mutants is 19% (6/32) and 30% (6/20), respectively. Sequence analysis revealed that all of the deleted sequences (ranging from 10 to 376 bp) were flanked by short direct repeats. We conclude that RAD52 and RAD10 are not required for the formation of deletions in the pol32 strain.

Mutation spectra of rad5, rad5 pol32, and rad5 rev3 strains:
To assess in greater detail the POL32 and the REV3 dependence of the rad mutators, the Can^r mutation spectra for rad5, rad5 pol32, and rad5 rev3 strains were determined by DNA sequencing. Rad5 is a component of the error-free repair/bypass pathway (ULRICH and JENTSCH 2000 ; reviewed in KUNZ et al. 2000 ). The rad5 single- and double-mutant strains were chosen for this study because (i) the rad5 strain has a clear-cut mutator phenotype, with a mutation rate up to fourfold greater than that of the wild-type rate; (ii) the rad5 mutator phenotype depends largely on functional POL32 and REV3; and (iii) the rad5 mutation spectrum had not yet been reported.

Sequence analysis of 26 independent Can^r colonies revealed 27 mutation events in rad5 strains (Table 3). There is an increased proportion of frameshifts (37%, all -1 frameshifts), while the fraction of base substitutions is similar to that of wild-type strains. Taking into account the overall increase in the mutation rate (Table 2), elimination of RAD5 results in an eightfold increase in the rate of frameshifts and a fourfold increase in the rate of base substitutions (Table 3). These two alterations (frameshift and base substitution) account for the total increase in the Can^r mutation rate attributable to the rad5 mutator effect. No multiple-base deletion or insertion mutations were detected among the 26 Can^r colonies analyzed.

In the rad5 pol32 mutant, which has an overall Can^r mutation rate similar to that of wild-type strains or of the pol32 single mutant, the frequencies of deletion (35%) and duplication (4%) resemble those of the pol32 mutant. All these events occurred in sequences flanked by short direct repeats (Table 4). The rates of base substitution and frameshift are fivefold and eightfold, respectively, lower than those for the rad5 mutant (Table 3). Therefore, the decreased rates of these two mutations account for the overall reduction in the mutation rate. The persistent presence of a high proportion of deletion and duplication mutations with a characteristic structure strongly suggests that the pol32 antimutator effect in the rad5 background differs from the pol32 effect in the production of extended deletions—the latter presumably arise by replication slippage between distant short repeats.

Elimination of REV3 also decreases the overall mutation rate of a rad5 strain, but this rate still appears to be somewhat higher than that of the wild-type strain and threefold higher than that of the rev3 mutant (Table 2). Mutational spectrum analysis showed a reduction in the rates of both base substitutions (twofold) and frameshifts (fivefold) compared to their incidence in the rad5 strain. However, these two rates are still twofold and sevenfold greater, respectively, than those for the rev3 strain (Table 3). This result indicates that REV3-dependent processes contribute substantially to both types of change, but that the participation of REV3-independent mechanisms is also important.

Mutation spectra of msh2, msh2 pol32, and msh2 rev3 strains:
To better understand the synergistic mutator effect of the msh2 and pol32 mutations, the Can^r mutation spectrum was determined for msh2, msh2 pol32, and msh2 rev3 strains. As previously reported by other investigators, the Can^r mutations arising in the msh2 mutant are primarily frameshift mutations (75%, usually -1 frameshifts) in short mononucleotide repeat sequences (Table 3). The rest consist of base substitutions (25%), predominantly G to A transitions. Sequencing of 24 Can^r isolates from the msh2 pol32 strain revealed a mutation spectrum similar to that of the msh2 single mutant, consisting of frameshift mutations in microsatellite sequences (71%, primarily -1 frameshifts) and base substitutions (29%). Deletion mutations, frequently identified in the pol32 single mutant, are entirely absent from the msh2 pol32 double mutant (0 of 24). However, this result is not too surprising in view of the much higher rate of frameshifts and base substitutions (see below) in msh2 and msh2 pol32 strains relative to the rate of deletion/duplication (2 x 10^-7) in the pol32 strain. In other words, the frameshifts and base substitutions that result in Can^r mutation outnumber deletions and duplications. The mutation spectrum of the msh2 rev3 mutant resembles that of the msh2 strain, which predominantly accumulates frameshift mutations in microsatellite sequences.

On the basis of the overall mutation rates and the frequency of each type of mutation, we calculated the Can^r mutation rates for specific classes. The base substitution rates are 1.9 x 10^-6 (msh2; a 6-fold increase relative to wild type), 7.2 x 10^-6 (msh2 pol32; a 21-fold increase), and 9.2 x 10^-7 (msh2 rev3, a 3-fold increase) per cell division. The rates at which frameshifts are generated are 5.8 x 10^-6 (msh2; a 64-fold increase relative to wild type), 1.8 x 10^-5 (msh2 pol32; a 193-fold increase), and 5.5 x 10^-6 (msh2 rev3, a 61-fold increase) per cell division. These data suggest that the effect of the pol32 mutation and the loss of Msh2 activity operate synergistically during replication to generate high rates of single-base mispairs and single-base deletion/addition mispairs. In contrast, and not surprisingly, the rev3 mutation does not significantly modify the mutation rate in the msh2 background. The mutator phenotype of an MMR-defective strain likely does not depend on the DNA damage replication bypass pathway.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The results obtained from this study can be summarized as follows: (i) The absence of the Pol32 subunit of Pol causes increases in genomic deletions of sequences flanked by short direct repeats, resembling the deletions associated with some human diseases (CANNING and DRYJA 1989 ; LUZI et al. 1995 ; MAGNANI et al. 1996 ; CHEN et al. 1998 ); (ii) the inactivation of POL32 or REV3 confers distinct but largely overlapping antimutator effects against rad mutators by a reduction in the rates of frameshift and base substitution; and (iii) the pol32 and msh2 mutations show a synergistic mutator effect on frameshift and base substitution mutations. These observations provide in vivo evidence that the Pol subunit encoded by POL32 plays a dual role in DNA replication and DNA damage bypass.

Mechanisms of deletion and duplication formation:
Mutagenic SSA and DNA replication slippage have been proposed as probable mechanisms for the deletion and duplication of sequences between nontandem direct repeats (TRAN et al. 1995 ; TISHKOFF et al. 1997A ; CHEN et al. 1998 , CHEN et al. 1999 ; KOKOSKA et al. 2000 ). In the mutagenic SSA model, prolonged exposure of single-strand template DNA can result in cleavage of this region, forming a DSB. Exonucleolytic resection of the DSB ends uncovers complementary repetitive sequences, and these short tracts can pair out of register (reviewed in PAQUES and HABER 1999 ). In the case of a deletion, the removal of the nonhomologous 3' tails after annealing requires a non-mismatch-related function of Msh2 and Msh3 and the Rad1-Rad10 endonuclease (IVANOV and HABER 1995 ; PAQUES and HABER 1997 ; SUGAWARA et al. 1997 ). After excision of nonhomologous 3' ends and new DNA synthesis, ligation restores two continuous strands but results in a deletion. In the replication slippage model, a decrease in the processivity of polymerase leads to arrest and dissociation from the template (VIGUERA et al. 2001 ), triggering the dissociation of the newly synthesized strand (NSS) from the template strand. The two strands may reassociate in a misaligned configuration. After reannealing, DNA polymerase reloads and synthesis resumes (KOKOSKA et al. 2000 ; Fig 1).

View larger version (21K): In this window In a new window Download PPT slide   Figure 1. Classical model for the generation of deletions (A) and duplications (B) by replication slippage between nontandem short direct repeats. The top and bottom strands of each structure represent NSS and the template strand, respectively. a1, upstream repeat on the NSS; a2, downstream repeat on the NSS; b1, upstream repeat on the template strand; b2, downstream repeat on the template strand; c, unique sequence between repeats. During DNA replication, the absence of Pol32 provokes frequent polymerase arrest, allowing a dissociation of the NSS from the template strand. The two strands may reassociate in a misaligned configuration. In A, pairing involves the downstream repeat on the template strand (b2); in B, pairing involves the upstream repeat on the template strand (b1), requiring a more extensive dissociation of the NSS and template strand duplex. Polymerase reloads on the reannealed 3' end of the NSS and replication resumes. A heteroduplex is thus formed, containing one parental and one recombinant strand, which are separated by a second round of replication. According to this model, illegitimate pairing of the upstream repeat on the NSS with a downstream repeat on the template strand leads to a deletion event (A), while illegitimate pairing of the downstream repeat on the NSS with an upstream repeat on the template strand leads to a duplication event (B).

The present results, combined with the known properties of Pol32, speak in favor of DNA replication slippage as the underlying mechanism of deletion and duplication formation in the pol32 mutant (Fig 1). First, Pol32 is a subunit of Pol . The deletions observed in pol32 resemble mostly those of the Pol mutants pol3-t, pol3-ts1, and pol3-ts11 (TRAN et al. 1995 ; KOKOSKA et al. 1998 , KOKOSKA et al. 2000 ) and some PCNA mutants (CHEN et al. 1999 ), implicating replication in the deletion/duplication process. Second, mutations in RAD52 and RAD10 do not significantly modulate the frequency of deletion formation in the pol32 background, suggesting that deletion events do not involve mutagenic SSA. Finally, in vitro biochemical studies have demonstrated that in the absence of Pol32 the Pol complex has a dramatically decreased processivity characterized by frequent pausing (BURGERS and GERIK 1998 ). Indeed, a decrease in processivity is usually considered as the initial cause of replication slippage. It may be argued that the absence of deletion events in the Can^r mutation spectrum of the msh2 pol32 strain is not in agreement with this model. In reality, this contradiction may be apparent only because the much higher rates of frameshift (1.8 x 10^-5) and base substitution (7.2 x 10^-6) mutations that accumulate in the msh2 pol32 strain obscure the much lower rate of deletion (1.8 x 10^-7), so that the limited number of Can^r colonies analyzed does not allow for the detection of deletions. Alternatively, the absence of CAN1 deletions in the msh2 pol32 strain could be taken as substantiating the mutagenic SSA model, since a non-mismatch-related function of Msh2 and Msh3 is involved in the removal of nonhomologous 3' tails during SSA (SUGAWARA et al. 1997 ). However, the fact that the elimination of RAD52 and RAD10, which both participate in SSA (IVANOV and HABER 1995 ; MORTENSEN et al. 1996 ; PAQUES and HABER 1997 ; SUGAWARA et al. 1997 ), does not affect deletion formation is not in support of mutagenic SSA. Furthermore, it was shown that MSH3 is not required for the formation of deletions flanked by direct repeats in the pol3-ts1 background (KOKOSKA et al. 2000 ). It is noteworthy that the Can^r mutation rate in the msh2 rad27 double mutant is similar to that of the rad27 single mutant (TISHKOFF et al. 1997A ) and that the msh2 or msh6 mutations significantly enhanced the Can^r mutation rate of various pol30 mutations (CHEN et al. 1999 ), but the mutation spectrum of these double mutants has not been analyzed.

Several additional points arising from our observations concerning the replication slippage model should be mentioned (see Fig 1 for details). First, the deleted sequence should be the unique sequence (c in Fig 1) bordered by the upstream repeat (b1) and the downstream repeat (b2) on the template strand (Fig 1). That the downstream rather than the upstream repeat on the template is lost is suggested by those clones with deletions of 8, 13, 16, 16, and 29 bp (Table 4), for which the two repeats are imperfect. The single direct repeat present in the NSS is always the upstream (a1) and never the downstream repeat (a2, Fig 1A; Table 4). Second, a perfect match of three base pairs is sufficient to restart replication following strand slippage. Third, the two large deletions and single duplication identified in the wild-type strain are structurally different from the deletions observed in the pol32 strain, in that they are not associated with nearby repeats, suggesting that the mechanisms responsible for deletion and duplication in wild-type and in pol32 strains are different. Finally, the pathway indicated in Fig 1A appears to be preferred, since deletions are much more frequent than duplications. This difference in frequency can be accounted for in two ways: (i) The absence of Pol32 could delay DNA synthesis on the lagging strand and lead to an increase in single-strand regions on the template strand, thereby favoring the formation of a loop on the template strand (Fig 1A), the net result being a deletion in the NSS and (ii) the formation of a loop on the NSS (Fig 1B) requires the dissociation of a longer NSS/template duplex, involving repeats (a2/b2), sequence (c), and repeats (a1/b1), a process that is likely to be less frequent than a transient dissociation of the shorter NSS/template duplex that involves only one pair of repeats (a1/b2, Fig 1A).

The increase in the frequency of extended deletions in pol32 mutants contrasts with a lack of -1 frameshift mutations. Similarly, the synergism between the msh2 and pol32 mutations with respect to frameshift and base substitution mutations contrasts with the antimutator effect conferred by the pol32 mutation in rad mutator backgrounds. Both these contrasts can be interpreted by making two assumptions. First, extended deletions of sequences flanked by both direct repeats and frameshifts in mononucleotide runs may be produced by replication slippage. Loops of 1 bp (leading to frameshifts) are recognized and repaired much more efficiently by the mismatch repair system than are larger loops formed as a result of slippage between more distant repeats (SIA et al. 1997 ; WIERDL et al. 1997 ). Second, the mechanisms responsible for frameshift mutations in mononucleotide runs and for other in vivo frameshift mutations that do not necessarily involve tandem repeats could be different (reviewed in KUNKEL and BEBENEK 2000 ; GREENE and JINKS-ROBERTSON 2001 ). The first type of frameshift mutation, observed predominately in msh2 and msh2 pol32 strains, should directly reflect primary slippage errors generated by replication polymerases. The second type of frameshift, which mostly occurs in nonrepetitive sequences, might arise in the context of unscheduled DNA replication (e.g., during DNA repair or DNA damage bypass). The antimutator phenotype with respect to frameshifts caused by the pol32 mutation in rad mutator backgrounds may be relevant for the second type of frameshift intermediate. The fact that rev3 has no effect on reducing the frameshift rate in a msh2 background further supports this possibility. These two proposals could explain the respective mutation spectra of the pol32 and msh2 pol32 strains as well as the different rad pol32 double mutants. A similar dual mechanism could be invoked to explain the increase in the rate of base substitutions in the msh2 pol32 double mutant (reflecting the increased replication error) and the decrease in rad pol32 double mutants (reflecting the antimutator effect of pol32 in spontaneous mutagenesis). Interestingly, pol32 was identified recently as a mutator in an exo1 background although no corresponding mutation spectra have been reported (AMIN et al. 2001 ). EXO1 encodes a 5'-to-3' double-strand DNA exonuclease that physically interacts with Msh2 (TISHKOFF et al. 1997B ).

Role of Pol32 in the mutagenic bypass pathway:
S. cerevisiae utilizes several DNA polymerases such as Pol , Pol , and Rev1 to bypass DNA damage (reviewed in KUNZ et al. 2000 ; BROOMFIELD et al. 2001 ). The replicative Pol also appears involved in a bypass pathway (GIOT et al. 1997 ; HARACSKA et al. 2000 ; HUANG et al. 2000 ). In the present study, we show that the deletion of POL32 or REV3 clearly diminishes the rad1, rad52, rad6, rad18, mms2, and rad5 mutator effects, but to different extents in the different mutants. The rad1 mutator effect is entirely POL32 and REV3 dependent. Deletion of POL32 seems to have a larger antimutator effect in the rad6, rad18, mms2, and rad5 backgrounds than does deletion of REV3. The rad52 mutator effect is only partially POL32 and REV3 dependent. In conclusion, a systematic comparison of the roles of Pol32 and Rev3 regarding the antimutator effect reveals distinct but largely overlapping activities. This is confirmed by a detailed CAN1 mutation spectrum analysis of the rad5, rad5 pol32, and rad5 rev3 strains: A reduction in the rate of base substitutions and frameshifts is attributable to the antimutator effects of the pol32 and rev3 mutations, as revealed in the rad5 pol32 and rad5 rev3 double mutants. These data suggest that both Pol32 and Rev3 are necessary to carry out DNA damage bypass in yeast.

There remain the questions of at which step and in which manner Pol32 participates in the process of damage bypass—for example, as a member of the Pol complex or in conjunction with Pol or other proteins. It has been proposed that the fundamental mechanism of replication bypass through a given type of damage can be dissected into the following steps: (i) a switch from a replicative DNA polymerase to a low-processivity translesion DNA polymerase, (ii) insertion of a nucleotide opposite the damaged base, (iii) elongation by incorporation of correct nucleotides, and (iv) a switch from the low-processivity translesion DNA polymerase to the replicative DNA polymerase (reviewed in WOODGATE 1999 ; BAYNTON and FUCHS 2000 ). In this model, a single low processivity translesion DNA polymerase like Pol or Pol carries out both the insertion and the elongation steps. Alternatively, JOHNSON et al. 2000 have proposed a different model, in which the two DNA polymerases act sequentially at the insertion and elongation steps to bypass DNA lesions. They found that in vitro, human Pol functions in damage bypass by inserting deoxynucleotides opposite (6-4) T-T lesions, whereas Pol acts at the subsequent extension step. Pol is inefficient in inserting deoxynucleotides opposite such highly distorting or noninformational lesions as (6-4) T-T and abasic sites (JOHNSON et al. 2000 , JOHNSON et al. 2001 ). More recently, HARACSKA et al. 2001 reported that Pol and Pol together carry out efficient abasic site bypass. Pol inserts a nucleotide opposite the abasic site and Pol then extends from that nucleotide. In the light of this finding, the present observations suggest that Pol and Pol may act sequentially to bypass the spontaneous lesions that arise in different rad mutators. The relative influence of POL32 and REV3 in a given rad mutator may merely reflect (i) the distinct functions of Pol and Pol in mediating the insertion of a mispaired base and/or in extending the NSS from this mispair and (ii) the interactions of Pol32 with other components of the damage bypass machinery, depending on the nature of the damage, the structure of the lesion, and the sequence context. Our data also suggest that Pol32 functionally links the DNA replication and DNA damage bypass pathways. It is logical to assume that the DNA damage bypass apparatus did not evolve entirely independently of the replication apparatus. Since DNA Pol is the first polymerase to encounter a DNA lesion, it will be most efficient if the replication can be converted into a damage bypass apparatus.

In conclusion, the present study underlines different aspects of the roles of Pol32 in DNA replication and in mutagenic damage bypass. These results illustrate the complexity and the highly interconnected nature of the pathways involved in these DNA transactions. The findings that the pol32 mutant exhibits a mutational spectrum that is biased toward deletions and duplications and that the msh2 pol32 double mutant has an enhanced mutator phenotype may have interesting implications for the genetics of human disease. For instance, it has been noted that the mutations associated with some diseases are deletions of sequences flanked by short direct repeats, similar to those observed in the pol32 mutant (CANNING and DRYJA 1989 ; LUZI et al. 1995 ; MAGNANI et al. 1996 ; CHEN et al. 1998 ), suggesting that they may arise by a similar mechanism. Furthermore, the synergistic mutator effect of the msh2 and pol32 mutations suggests that simultaneous mutations of genes involved in DNA mismatch repair and in DNA polymerase accessory subunits could be an important source of genome destabilization. More generally, mutation of a single such gene might be innocuous per se but could confer an increased cancer susceptibility on an individual who harbors mutations in other genes.

	ACKNOWLEDGMENTS

We thank R. D. Kolodner for the plasmid pRDK35 and J. McDonald and R. Woodgate for the plasmid pJM112. We also thank A. Nicolas and R. P. Fuchs for valuable discussion during this work and A. Nicolas, A. Bresson, J. C. Chuat, and K. Smith for the critical reading of the manuscript. This work was supported by the Centre National de la Recherche Scientifique (CNRS, France).

Manuscript received August 27, 2001; Accepted for publication January 25, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AMIN, N. S., M. N. NGUYEN, S. OH, and R. D. KOLODNER, 2001  exo1-dependent mutator mutations: model system for studying functional interactions in mismatch repair. Mol. Cell. Biol. 21:5142-5155[Abstract/Free Full Text].

BAYNTON, K. and R. P. FUCHS, 2000  Lesions in DNA: hurdles for polymerases. Trends Biochem. Sci. 25:74-79[Medline].

BROOMFIELD, S., B. L. CHOW, and W. XIAO, 1998  MMS2, encoding a ubiquitin-conjugating-enzyme-like protein, is a member of the yeast error-free postreplication repair pathway. Proc. Natl. Acad. Sci. USA 95:5678-5683[Abstract/Free Full Text].

BROOMFIELD, S., T. HRYCIW, and W. XIAO, 2001  DNA postreplication repair and mutagenesis in Saccharomyces cerevisiae.. Mutat. Res. 486:167-184[Medline].

BURGERS, P. M., 1998  Eukaryotic DNA polymerases in DNA replication and DNA repair. Chromosoma 107:218-227[Medline].

BURGERS, P. M. and K. J. GERIK, 1998  Structure and processivity of two forms of Saccharomyces cerevisiae DNA polymerase . J. Biol. Chem. 273:19756-19762[Abstract/Free Full Text].

CANNING, S. and T. P. DRYJA, 1989  Short, direct repeats at the breakpoints of deletions of the retinoblastoma gene. Proc. Natl. Acad. Sci. USA 86:5044-5048[Abstract/Free Full Text].

CHEN, C. and R. D. KOLODNER, 1999  Gross chromosomal rearrangements in Saccharomyces cerevisiae replication and recombination defective mutants. Nat. Genet. 23:81-85[Medline].

CHEN, C., K. UMEZU, and R. D. KOLODNER, 1998  Chromosomal rearrangements occur in S. cerevisiae rfa1 mutator mutants due to mutagenic lesions processed by double-strand-break repair. Mol. Cell 2:9-22[Medline].

CHEN, C., B. J. MERRILL, P. J. LAU, C. HOLM, and R. D. KOLODNER, 1999  Saccharomyces cerevisiae pol30 (proliferating cell nuclear antigen) mutations impair replication fidelity and mismatch repair. Mol. Cell. Biol. 19:7801-7815[Abstract/Free Full Text].

DRAKE, J. W., 1991  A constant rate of spontaneous mutation in DNA-based microbes. Proc. Natl. Acad. Sci. USA 88:7160-7164[Abstract/Free Full Text].

FRIEDBERG, E. C., G. C. WALKER and W. SIEDE, 1995 DNA Repair and Mutagenesis. American Society for Microbiology Press, Washington, DC.

GERIK, K. J., X. LI, A. PAUTZ, and P. M. BURGERS, 1998  Characterization of the two small subunits of Saccharomyces cerevisiae DNA polymerase . J. Biol. Chem. 273:19747-19755[Abstract/Free Full Text].

GREENE, C. N. and S. JINKS-ROBERTSON, 1997  Frameshift intermediates in homopolymer runs are removed efficiently by yeast mismatch repair proteins. Mol. Cell. Biol. 17:2844-2850[Abstract].

GREENE, C. N. and S. JINKS-ROBERTSON, 2001  Spontaneous frameshift mutations in Saccharomyces cerevisiae: accumulation during DNA replication and removal by proofreading and mismatch repair activities. Genetics 159:65-75[Abstract/Free Full Text].

GIOT, L., R. CHANET, M. SIMON, C. FACCA, and G. FAYE, 1997  Involvement of the yeast DNA polymerase in DNA repair in vivo.. Genetics 146:1239-1251[Abstract].

HARACSKA, L., S. PRAKASH, and L. PRAKASH, 2000  Replication past O⁶-methylguanine by yeast and human DNA polymerase . Mol. Cell. Biol. 20:8001-8007[Abstract/Free Full Text].

HARACSKA, L., I. UNK, R. E. JOHNSON, E. JOHANSSON, and P. M. BURGERS et al., 2001  Roles of yeast DNA polymerases and and of Rev1 in the bypass of abasic sites. Genes Dev. 15:945-954[Abstract/Free Full Text].

HARFE, B. D. and S. JINKS-ROBERTSON, 2000a  DNA mismatch repair and genetic instability. Annu. Rev. Genet. 34:359-399[Medline].

HARFE, B. D. and S. JINKS-ROBERTSON, 2000b  DNA polymerase introduces multiple mutations when bypassing spontaneous DNA damage in Saccharomyces cerevisiae.. Mol. Cell 6:1491-1499[Medline].

HUANG, M. E., E. CADIEU, J. L. SOUCIET, and F. GALIBERT, 1997  Disruption of six novel yeast genes reveals three genes essential for vegetative growth and one required for growth at low temperature. Yeast 13:1181-1194[Medline].

HUANG, M. E., B. LE DOUARIN, C. HENRY, and F. GALIBERT, 1999  The Saccharomyces cerevisiae protein YJR043C (Pol32) interacts with the catalytic subunit of DNA polymerase and is required for cell cycle progression in G2/M. Mol. Gen. Genet. 260:541-550[Medline].

HUANG, M. E., A. DE CALIGNON, A. NICOLAS, and F. GALIBERT, 2000  POL32, a subunit of the Saccharomyces cerevisiae DNA polymerase defines a link between DNA replication and the mutagenic bypass repair pathway. Curr. Genet. 38:178-187[Medline].

IVANOV, E. L. and J. E. HABER, 1995  RAD1 and RAD10, but not other excision repair genes, are required for double-strand break-induced recombination in Saccharomyces cerevisiae.. Mol. Cell. Biol. 15:2245-2251[Abstract].

JIRICNY, J. and M. NYSTRÖM-LAHTI, 2000  Mismatch repair defects in cancer. Curr. Opin. Genet. Dev. 10:157-161[Medline].

JOHNSON, R. E., G. K. KOVVALI, L. PRAKASH, and S. PRAKASH, 1996  Requirement of the yeast MSH3 and MSH6 genes for MSH2-dependent genomic stability. J. Biol. Chem. 271:7285-7288[Abstract/Free Full Text].

JOHNSON, R. E., M. T. WASHINGTON, L. HARACSKA, S. PRAKASH, and L. PRAKASH, 2000  Eukaryotic polymerases and act sequentially to bypass DNA lesions. Nature 406:1015-1019[Medline].

JOHNSON, R. E., L. HARACSKA, S. PRAKASH, and L. PRAKASH, 2001  Role of DNA polymerase in the bypass of a (6-4) TT photoproduct. Mol. Cell. Biol. 21:3558-3563[Abstract/Free Full Text].

KOKOSKA, R. J., L. STEFANOVIC, H. T. TRAN, M. A. RESNICK, and D. A. GORDENIN et al., 1998  Destabilization of yeast micro- and minisatellite DNA sequences by mutations affecting a nuclease involved in Okazaki fragment processing (rad27) and DNA polymerase (pol3-t). Mol. Cell. Biol. 18:2779-2788[Abstract/Free Full Text].

KOKOSKA, R. J., L. STEFANOVIC, J. DEMAI, and T. D. PETES, 2000  Increased rates of genomic deletions generated by mutations in the yeast gene encoding DNA polymerase or by decreases in the cellular levels of DNA polymerase . Mol. Cell. Biol. 20:7490-7504[Abstract/Free Full Text].

KUNKEL, T. A. and K. BEBENEK, 2000  DNA replication fidelity. Annu. Rev. Biochem. 69:497-529[Medline].

KUNZ, B. A., K. RAMACHANDRAN, and E. J. VONARX, 1998  DNA sequence analysis of spontaneous mutagenesis in Saccharomyces cerevisiae.. Genetics 148:1491-1505[Abstract/Free Full Text].

KUNZ, B. A., A. F. STRAFFON, and E. J. VONARX, 2000  DNA damage-induced mutation: tolerance via translesion synthesis. Mutat. Res. 451:169-185[Medline].

LEA, D. E. and C. A. COULSON, 1948  The distribution of the numbers of mutants in bacterial populations. J. Genet. 49:264-285.

LUZI, P., M. A. RAFI, and D. A. WENGER, 1995  Characterization of the large deletion in the GALC gene found in patients with Krabbe disease. Hum. Mol. Genet. 4:2335-2358[Abstract/Free Full Text].

MAGNANI, C., L. CREMONESI, A. GIUNTA, P. MAGNAGHI, and R. TARAMELLI et al., 1996  Short direct repeats at the breakpoints of a novel large deletion in the CFTR gene suggest a likely slipped mispairing mechanism. Hum. Genet. 98:102-108[Medline].

MARSISCHKY, G. T., N. FILOSI, M. F. KANE, and R. KOLODNER, 1996  Redundancy of Saccharomyces cerevisiae MSH3 and MSH6 in MSH2-dependent mismatch repair. Genes Dev. 10:407-420[Abstract/Free Full Text].

MCDONALD, J. P., A. S. LEVINE, and R. WOODGATE, 1997  The Saccharomyces cerevisiae RAD30 gene, a homologue of Escherichia coli dinB and umuC, is DNA damage inducible and functions in a novel error-free postreplication repair mechanism. Genetics 147:1557-1568[Abstract].

MORRISON, A. and A. SUGINO, 1994  The 3'5' exonucleases of both DNA polymerases and participate in correcting errors of DNA replication in Saccharomyces cerevisiae.. Mol. Gen. Genet. 242:289-296[Medline].

MORTENSEN, U. H., C. BENDIXEN, I. SUNJEVARIC, and R. ROTHSTEIN, 1996  DNA strand annealing is promoted by the yeast Rad52 protein. Proc. Natl. Acad. Sci. USA 93:10729-10734[Abstract/Free Full Text].

MYUNG, K., A. DATTA, C. CHEN, and R. D. KOLODNER, 2001a  SGS1, the Saccharomyces cerevisiae homologue of BLM and WRN, suppresses genome instability and homeologous recombination. Nat. Genet. 27:113-116[Medline].

MYUNG, K., A. DATTA, and R. D. KOLODNER, 2001b  Suppression of spontaneous chromosomal rearrangements by S phase checkpoint functions in Saccharomyces cerevisiae. Cell 104:397-408[Medline].

PÂQUES, F. and J. E. HABER, 1997  Two pathways for removal of nonhomologous DNA ends during double-strand break repair in Saccharomyces cerevisiae.. Mol. Cell. Biol. 17:6765-6771[Abstract].

PÂQUES, F. and J. E. HABER, 1999  Multiple pathways of recombination induced by double-strand breaks in Saccharomyces cerevisiae.. Microbiol. Mol. Biol. Rev. 63:349-404[Abstract/Free Full Text].

ROCHE, H., R. D. GIETZ, and B. A. KUNZ, 1994  Specificity of the yeast rev3 antimutator and REV3 dependency of the mutator resulting from a defect (rad1) in nucleotide excision repair. Genetics 137:637-646[Abstract].

ROCHE, H., R. D. GIETZ, and B. A. KUNZ, 1995  Specificities of the Saccharomyces cerevisiae rad6, rad18, and rad52 mutators exhibit different degrees of dependence on the REV3 gene product, a putative nonessential DNA polymerase. Genetics 140:443-456[Abstract].

ROSE, M. D., F. WINSTON and P. HIETER, 1990 Methods in Yeast Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SARGENTINI, N. J. and K. C. SMITH, 1985  Spontaneous mutagenesis: the role of DNA repair, replication and recombination. Mutat. Res. 154:1-27[Medline].

SIA, E. A., R. J. KOKOSKA, M. DOMINSKA, P. GREENWELL, and T. D. PETES, 1997  Microsatellite instability in yeast: dependence on repeat unit size and DNA mismatch repair genes. Mol. Cell. Biol. 17:2851-2858[Abstract].

SUGAWARA, N. and J. E. HABER, 1992  Characterization of double-strand break-induced recombination: homology requirements and single-stranded DNA formation. Mol. Cell. Biol. 12:563-575[Abstract/Free Full Text].

SUGAWARA, N., F. PÂQUES, M. COLAIACOVO, and J. E. HABER, 1997  Role of Saccharomyces cerevisiae Msh2 and Msh3 repair proteins in double-strand break-induced recombination. Proc. Natl. Acad. Sci. USA 94:9214-9219[Abstract/Free Full Text].

SWANSON, R. L., N. J. MOREY, P. W. DOETSCH, and S. JINKS-ROBERTSON, 1999  Overlapping specificities of base excision repair, nucleotide excision repair, recombination, and translesion synthesis pathways for DNA base damage in Saccharomyces cerevisiae.. Mol. Cell. Biol. 19:2929-2935[Abstract/Free Full Text].

TISHKOFF, D. X., N. FILOSI, G. M. GAIDA, and R. D. KOLODNER, 1997a  A novel mutation avoidance mechanism dependent on S. cerevisiae RAD27 is distinct from DNA mismatch repair. Cell 88:253-263[Medline].

TISHKOFF, D. X., A. L. BOERGER, P. BERTRAND, N. FILOSI, and G. M. GAIDA et al., 1997b  Identification and characterization of Saccharomyces cerevisiae EXO1, a gene encoding an exonuclease that interacts with MSH2. Proc. Natl. Acad. Sci. USA 94:7487-7492[Abstract/Free Full Text].

TRAN, H. T., N. P. DEGTYAREVA, N. N. KOLOTEVA, A. SUGINO, and H. MASUMOTO et al., 1995  Replication slippage between distant short repeats in Saccharomyces cerevisiae depends on the direction of replication and the RAD50 and RAD52 genes. Mol. Cell. Biol. 15:5607-5617[Abstract].

ULRICH, H. D. and S. JENTSCH, 2000  Two RING finger proteins mediate cooperation between ubiquitin-conjugating enzymes in DNA repair. EMBO J. 19:3388-3397[Medline].

VIGUERA, E., D. CANCEILL, and S. D. EHRLICH, 2001  Replication slippage involves DNA polymerase pausing and dissociation. EMBO J. 20:2587-2595[Medline].

WAGA, S. and B. STILLMAN, 1998  The DNA replication fork in eukaryotic cells. Annu. Rev. Biochem. 67:721-751[Medline].

WIERDL, M., M. DOMINSKA, and T. D. PETES, 1997  Microsatellite instability in yeast: dependence on the length of the microsatellite. Genetics 146:769-779[Abstract].

WOODGATE, R., 1999  A plethora of lesion-replicating DNA polymerases. Genes Dev. 13:2191-2195[Free Full Text].

ZAR, J. H., 1996 Biostatistical Analysis. Prentice-Hall International, Englewood Cliffs, NJ.

This article has been cited by other articles:

				P. A. van der Kemp, M. de Padula, G. Burguiere-Slezak, H. D. Ulrich, and S. Boiteux PCNA monoubiquitylation and DNA polymerase {eta} ubiquitin-binding domain are required to prevent 8-oxoguanine-induced mutagenesis in Saccharomyces cerevisiae Nucleic Acids Res., May 1, 2009; 37(8): 2549 - 2559. [Abstract] [Full Text] [PDF]

				L. S. Waters, B. K. Minesinger, M. E. Wiltrout, S. D'Souza, R. V. Woodruff, and G. C. Walker Eukaryotic Translesion Polymerases and Their Roles and Regulation in DNA Damage Tolerance Microbiol. Mol. Biol. Rev., March 1, 2009; 73(1): 134 - 154. [Abstract] [Full Text] [PDF]

				W. Ma, V. Panduri, J. F. Sterling, B. Van Houten, D. A. Gordenin, and M. A. Resnick The Transition of Closely Opposed Lesions to Double-Strand Breaks during Long-Patch Base Excision Repair Is Prevented by the Coordinated Action of DNA Polymerase {delta} and Rad27/Fen1 Mol. Cell. Biol., March 1, 2009; 29(5): 1212 - 1221. [Abstract] [Full Text] [PDF]

				C. M. Stith, J. Sterling, M. A. Resnick, D. A. Gordenin, and P. M. Burgers Flexibility of Eukaryotic Okazaki Fragment Maturation through Regulated Strand Displacement Synthesis J. Biol. Chem., December 5, 2008; 283(49): 34129 - 34140. [Abstract] [Full Text] [PDF]

				A. Deem, K. Barker, K. VanHulle, B. Downing, A. Vayl, and A. Malkova Defective Break-Induced Replication Leads to Half-Crossovers in Saccharomyces cerevisiae Genetics, August 1, 2008; 179(4): 1845 - 1860. [Abstract] [Full Text] [PDF]

				T. van Welsem, F. Frederiks, K. F. Verzijlbergen, A. W. Faber, Z. W. Nelson, D. A. Egan, D. E. Gottschling, and F. van Leeuwen Synthetic Lethal Screens Identify Gene Silencing Processes in Yeast and Implicate the Acetylated Amino Terminus of Sir3 in Recognition of the Nucleosome Core Mol. Cell. Biol., June 1, 2008; 28(11): 3861 - 3872. [Abstract] [Full Text] [PDF]

				I. Iraqui, G. Faye, S. Ragu, A. Masurel-Heneman, R. D. Kolodner, and M.-E. Huang Human Peroxiredoxin PrxI Is an Orthologue of Yeast Tsa1, Capable of Suppressing Genome Instability in Saccharomyces cerevisiae Cancer Res., February 15, 2008; 68(4): 1055 - 1063. [Abstract] [Full Text] [PDF]

				M. Jaszczur, K. Flis, J. Rudzka, J. Kraszewska, M. E. Budd, P. Polaczek, J. L. Campbell, P. Jonczyk, and I. J. Fijalkowska Dpb2p, a Noncatalytic Subunit of DNA Polymerase {varepsilon}, Contributes to the Fidelity of DNA Replication in Saccharomyces cerevisiae Genetics, February 1, 2008; 178(2): 633 - 647. [Abstract] [Full Text] [PDF]

				C. H. Sterling and J. B. Sweasy DNA Polymerase 4 of Saccharomyces cerevisiae Is Important for Accurate Repair of Methyl-Methanesulfonate-Induced DNA Damage Genetics, January 1, 2006; 172(1): 89 - 98. [Abstract] [Full Text] [PDF]

				P. Auerbach, R. A. O. Bennett, E. A. Bailey, H. E. Krokan, and B. Demple Mutagenic specificity of endogenously generated abasic sites in Saccharomyces cerevisiae chromosomal DNA PNAS, December 6, 2005; 102(49): 17711 - 17716. [Abstract] [Full Text] [PDF]

				S. Sabbioneda, B. K. Minesinger, M. Giannattasio, P. Plevani, M. Muzi-Falconi, and S. Jinks-Robertson The 9-1-1 Checkpoint Clamp Physically Interacts with Pol{zeta} and Is Partially Required for Spontaneous Pol{zeta}-dependent Mutagenesis in Saccharomyces cerevisiae J. Biol. Chem., November 18, 2005; 280(46): 38657 - 38665. [Abstract] [Full Text] [PDF]

				Y. Guo, L. L. Breeden, H. Zarbl, B. D. Preston, and D. L. Eaton Expression of a Human Cytochrome P450 in Yeast Permits Analysis of Pathways for Response to and Repair of Aflatoxin-Induced DNA Damage Mol. Cell. Biol., July 15, 2005; 25(14): 5823 - 5833. [Abstract] [Full Text] [PDF]

				P. Garg, C. M. Stith, J. Majka, and P. M. J. Burgers Proliferating Cell Nuclear Antigen Promotes Translesion Synthesis by DNA Polymerase {zeta} J. Biol. Chem., June 24, 2005; 280(25): 23446 - 23450. [Abstract] [Full Text] [PDF]

				B. K. Minesinger and S. Jinks-Robertson Roles of RAD6 Epistasis Group Members in Spontaneous Pol{zeta}-Dependent Translesion Synthesis in Saccharomyces cerevisiae Genetics, April 1, 2005; 169(4): 1939 - 1955. [Abstract] [Full Text] [PDF]

				S. E. Corrette-Bennett, C. Borgeson, D. Sommer, P. M. J. Burgers, and R. S. Lahue DNA polymerase {delta}, RFC and PCNA are required for repair synthesis of large looped heteroduplexes in Saccharomyces cerevisiae Nucleic Acids Res., December 1, 2004; 32(21): 6268 - 6275. [Abstract] [Full Text] [PDF]

				M.-E. Huang, A.-G. Rio, A. Nicolas, and R. D. Kolodner A genomewide screen in Saccharomyces cerevisiae for genes that suppress the accumulation of mutations PNAS, September 30, 2003; 100(20): 11529 - 11534. [Abstract] [Full Text] [PDF]

				S. G. Kozmin, Y. I. Pavlov, T. A. Kunkel, and E. Sage Roles of Saccharomyces cerevisiae DNA polymerases Pol{eta} and Pol{zeta} in response to irradiation by simulated sunlight Nucleic Acids Res., August 1, 2003; 31(15): 4541 - 4552. [Abstract] [Full Text] [PDF]

				S. Bhattacharyya, M. L. Rolfsmeier, M. J. Dixon, K. Wagoner, and R. S. Lahue Identification of RTG2 as a Modifier Gene for CTG{middle dot}CAG Repeat Instability in Saccharomyces cerevisiae Genetics, October 1, 2002; 162(2): 579 - 589. [Abstract] [Full Text] [PDF]

				R. D. Kolodner, C. D. Putnam, and K. Myung Maintenance of Genome Stability in Saccharomyces cerevisiae Science, July 26, 2002; 297(5581): 552 - 557. [Abstract] [Full Text] [PDF]