Genomic Instability Induced by Mutations in Saccharomyces cerevisiae POL1

Corresponding author: Teresa S.-F. Wang, Edwards Bldg., Rm. R270, Stanford University Medical Center, 300 Pasteur Dr., Stanford, CA 94305. E-mail address: twang@pmgm2.stanford.edu

Communicating editor: F. WINSTON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mutations of chromosome replication genes can be one of the early events that promote genomic instability. Among genes that are involved in chromosomal replication, DNA polymerase is essential for initiation of replication and lagging-strand synthesis. Here we examined the effect of two mutations in S. cerevisiae POL1, pol1-1 and pol1-17, on a microsatellite (GT)₁₆ tract. The pol1-17 mutation elevated the mutation rate 13-fold by altering sequences both inside and downstream of the (GT)₁₆ tract, whereas the pol1-1 mutation increased the mutation rate 54-fold by predominantly altering sequences downstream of the (GT)₁₆ tract in a RAD52-dependent manner. In a rad52 null mutant background pol1-1 and pol1-17 also exhibited different plasmid and chromosome loss phenotypes. Deletions of mismatch repair (MMR) genes induce a differential synergistic increase in the mutation rates of pol1-1 and pol1-17. These findings suggest that perturbations of DNA replication in these two pol1 mutants are caused by different mechanisms, resulting in various types of mutations. Thus, mutations of POL1 can induce a variety of mutator phenotypes and can be a source of genomic instability in cells.

MUTATIONS in genes involved in DNA replication, DNA repair, or chromosome segregation can potentially induce a mutator phenotype. Among the genes involved in DNA replication, the eukaryotic replicative DNA polymerases (Pol, Pol, and Pol) play an essential role in determining the faithful transmission of genetic information from one generation to the next. Homologs of these replicative polymerases from yeast to humans share a high degree of conservation in their protein domain organization (DELARUE et al. 1990 ; WANG 1996 ). In addition to replication, these replicative polymerases are also involved in double-strand-break repair (HOLMES and HABER 1999 ), nucleotide excision repair (LINDAHL and WOOD 1999 ), and telomere homeostasis maintenance (DIEDE and GOTTSCHLING 1999 ; ADAMS-MARTIN et al. 2000 ; DAHLEN et al. 2003 ).

In budding yeast, a mutator phenotype has been associated with several mutant alleles of POL2 and POL3, which encode the replicative polymerase and , respectively. Several of these pol2 and pol3 mutants exhibit an increase in frameshifts in homonucleotide runs (TRAN et al. 1997B ; KIRCHNER et al. 2000 ), in microsatellite instability (STRAND et al. 1993 ; KOKOSKA et al. 1998 , KOKOSKA et al. 2000 ), in deletions of sequence flanked by short direct repeats (GORDENIN et al. 1992 ; TRAN et al. 1995 ; KOKOSKA et al. 2000 ), or in base substitutions (MORRISON et al. 1993 ; MORRISON and SUGINO 1994 ). In contrast, little is known about the effect of mutations in POL1 (Pol), the replicative polymerase essential for both initiation at the replication origin and initiation of Okazaki fragments during the lagging-strand synthesis (WAGA and STILLMAN 1998 ). Although Pol1p (Pol) does not seem to be involved in synthesizing the main bulk of cellular DNA as shown in an in vitro reconstituted replication assay containing an SV40-ori-bearing plasmid (WAGA et al. 1994 ; WAGA and STILLMAN 1994 ), its unique role in chromosome initiation has led us to investigate POL1's contribution to mutation avoidance.

Our previous fission yeast studies identified conditional mutants in the catalytic subunit of Pol, which conferred a mutator phenotype at the ura4⁺ locus characterized by an elevated rate of base substitutions and deletion of sequences flanked by short direct repeats (LIU et al. 1999 ; KAI and WANG 2003 ). The mutation rate was exacerbated when these replication mutators were combined with a deletion of Cds1 (LIU et al. 1999 ; KAI and WANG 2003 ), a checkpoint effector kinase essential in fission yeast for tolerating and recovering from replication perturbations (WALWORTH 2000 ; BODDY and RUSSELL 2001 ). These studies suggest that mutations of Pol potentially affect the initiation complex at the replication origin or during Okazaki fragment synthesis, compromising the stability of replication fork and thus generating a mutator phenotype in cells.

In this study, we use the budding yeast Saccharomyces cerevisiae to investigate the contribution of POL1 to mutation avoidance by analyzing two distinct conditional pol1 mutant alleles, pol1-1 and pol1-17. The pol1-1 mutant allele contains a single missense mutation (Gly⁴⁹³ to Arg) within the N-terminal region (PIZZAGALLI et al. 1988 ), while pol1-17 encodes a mutation (Thr¹⁰⁰⁴ to Ile) buried near the most conserved domain of the B-family (-like) polymerases (ITO and BRAITHWAITE 1991 ; HERINGA and ARGOS 1994 ). By using a well-established dinucleotide repeat system (HENDERSON and PETES 1992 ), we investigate the effects of these pol1 alleles on genome stability as well as their genetic interactions with double-strand-break repair and mismatch repair (MMR) genes. Our results indicate that POL1 plays a role in ensuring genomic stability in several ways, since these mutant alleles of POL1 can induce a range of mutator phenotypes.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Media, growth conditions, and general methods:
Standard budding yeast cultivation methods and standard media were utilized (SHERMAN 1991 ). Synthetic complete (SC) medium lacking tryptophan (W), leucine (L), threonine (T), and uracil (U) was used for microsatellite instability assays as described in HENDERSON and PETES 1992 . For sporulation, cells were treated with 0.5% KOAc (pH 7.0) supplemented with half the normal amount of amino acids required for auxotrophic strains. Standard methods were used to switch mating type (HERSKOWITZ and JENSEN 1991 ), utilizing pRB1191 [URA3 CEN4 GAL1/10-HO] and tester strains PT1 and PT2 (plasmid and strains kindly provided by D. Botstein).

Plasmid constructions:
Both pPGI11 and pPGI117 were constructed by restricting the 5.5-kb BamHI-SphI fragment of pPGC11 and pPGC117, which contain the entire pol1-1 or pol1-17 coding region in addition to upstream and downstream untranslated regions, and inserting this fragment into BamHI-SphI-restricted pRB1721 [pUC18-URA3] (kindly supplied by D. Botstein).

Yeast strain constructions:
All strains used in this study are described in Table 1. Transformations were performed using the high-efficiency lithium acetate protocol described in AGATEP et al. 1998 . PGY 2130 and PGY 2140 were generated via a two-step gene replacement, transforming EAS18 (kindly provided by T. Petes) with XbaI-linearized pPGI11 and pPGI117, respectively, and selecting for thermosensitive colonies after treatment with 5-fluoroorotic acid (5-FOA). The pol1-1 and pol1-17 mutations were confirmed by sequencing.

PGY 2170 was generated from a cross between PGY 2005 and MS 73 (kindly supplied by T. Petes). Presence of the pol3-01 allele was determined by sequencing. PGY 2180 was constructed by transforming MS72 (kindly supplied by T. Petes) with pRB1191 and changing the mating type. PGY 2320, PGY 2321, PGY 2330, PGY 2331, PGY 2340, and PGY 2341 were all meiotic segregants from the crosses with RJK 397 (kindly provided by T. Petes), a strain that contains a disruption of the RAD52 locus with the hisG-URA3-hisG cassette. Crosses are described in Table 1. After verifying that these strains were thermosensitive (as appropriate) and methyl methanesulfonate sensitive, 5-FOA was used to induce recombinogenic loss of URA3 and one copy of hisG.

PGY 2221, PGY 2231, and PGY 2241 were meiotic segregants generated from crosses described in Table 1, which were verified to be Leu⁺ and/or thermosensitive (the AMY101 parental strain kindly provided by A. Sugino). The presence of pms1 was confirmed by high papillation on plates containing canavanine sulfate (60 µg/ml) and by PCR with a sense primer in the upstream region and an antisense primer internal to LEU2. A similar strategy was used to generate PGY 2721, PGY 2731, PGY 2741, PGY 2621, and PGY 2631 (parental strains kindly provided by T. Petes). The presence of msh6 or msh3 was verified by PCR with appropriate primers as previously mentioned. Since only MAT segregants were isolated for PGY 2621 and PGY 2631, the mating type was changed with pRB1191.

Diploids from a chromosome-loss assay were constructed by crossing strains specified in Table 1. Strains PGY 2024, PGY 2134, and PGY 2026, used in the MAT conversion assay, were constructed by transforming both PGY 2006 and PGY 2131 with a PCR-amplified fragment containing URA3 flanked by 48 nucleotides of homologous sequence to the 205-kb region on chromosome III and AMY125 with a PCR fragment containing TRP1 flanked by the same sequences.

To measure Pol1p protein levels, wild-type and pol1 mutant strains (PGY 210, DFS4d/5a, and PGY 300) were transformed with a PCR fragment that contained the TAP tag followed by the TRP1 gene of Kluyveromyces lactis (RIGAUT et al. 1999 ), which was amplified from plasmid pBS1479 using the primers 5'-ACTATATAGAATATTCATGAGATCACACAACACATACAAAATACTTACtacgactcactataggg-3' and 5'-GGACGTCGCTACGTTGATATGACTAGCATATTTGATTTCATGCTAAATtccatggaaaagagaag-3' where uppercase sequences correspond to the C terminus of POL1. This generated PGY 1121, PGY 1131, and PGY 1141. Correct integration was verified by PCR and Western analysis.

Microsatellite instability assays:
Microsatellite assays were carried out as in WIERDL et al. 1997 with the following minor modifications. All assays were performed at the semipermissive temperature of 28°, the highest temperature at which viability of the thermosensitive strains was determined to be comparable to the wild type. Yeast strains were grown at the permissive temperature (25°) in liquid SC-WLTU, plated for single colonies onto solid SC-W medium, and incubated at 28°. Colonies resistant to 5-FOA (5-FOA^r) were scored after incubating 5 days at 25°.

Mutation rate analysis for microsatellite instability assays:
Each experiment was calculated from 12 to 24 independent cultures and at least two independent experiments were done per strain. Mutation rate was calculated using the method of the median as described by LEA and COULSON 1949 , which had been used in various other microsatellite instability studies (HENDERSON and PETES 1992 ; STRAND et al. 1993 ; JOHNSON et al. 1996A ; WIERDL et al. 1996 , WIERDL et al. 1997 ; KOKOSKA et al. 1998 ). Confidence intervals were generated as described in WIERDL et al. 1996 . Significance levels between strains were determined by using the Mann-Whitney rank test (Analyze-it, Microsoft Excel), a standard test of significance when populations are nonparametric and the distribution is unknown. Unless specified otherwise, all comparisons are significant with a P level of P < 0.01.

Analysis of poly(GT) tract lengths:
Two methods were used to determine the lengths of the poly(GT) tract. Standard methods were used to directly sequence the poly(GT) tract in Escherichia coli colonies carrying the repeat plasmid rescued from yeast. In addition, "hot" PCR analysis of poly(GT) tracts was done directly from yeast colonies with PCR primers used in previous studies (WIERDL et al. 1997 ). Lengths were compared to standard tracts that had been previously sequenced. The 10-µl PCR reaction contained the following: 5 pmol of each primer, 200 µM of dNTP, 1 µCi of [-³³P]dATP (Amersham, Buckinghamshire, UK), 1x Vent DNA polymerase buffer, and 0.2 units of Taq polymerase. PCR products were analyzed on a 6% denaturing polyacrylamide gel. Radioactive PCR sizing of the poly(GT) tract was done from both bacterial and yeast colonies to confirm that changes present in pSH44 did not result from mutations induced in E. coli during plasmid propagation. No tract length discrepancies were found throughout these studies.

Confidence intervals of 95% were derived using standard statistical methods based on the binomial distribution as described in two introductory statistics texts (FREEDMAN et al. 1991 , p. 348; GLANTZ 1992 , p. 206) with the formula 2. N is the total number of tracts sequenced, and P is the percentage of tracts that are in a particular category (i.e., "deletions," "insertions," and "no change"). These confidence intervals were also used in reporting rates for specific alterations [e.g., rates of deletions, alterations outside of the poly(GT) tract, etc.]. Significant differences among types of alterations were determined using the chi-square test (or Fisher exact test when appropriate).

Analysis of FOA^r colonies that had no change in the poly(GT) tract:
When no changes were detected in the poly(GT) tract after PCR sizing, pSH44 was isolated from the 5-FOA^r cells and digested with HindIII to detect gross sequence changes in the poly(GT)-HIS4-URA3 substrate. These plasmids were subsequently sequenced with various sense and antisense primers running along the length of the HIS4-URA3 reporter.

Spontaneous mutation rate and mutation spectrum in CAN1:
Standard methods were used to determine the forward mutation rate in the CAN1 gene, which confers resistance (Can^r) to the arginine analog canavanine sulfate (XIE et al. 1999 ). Rates were calculated from 11–15 independent cultures and two independent experiments were done per strain. Genomic DNA was isolated from Can^r colonies and the CAN1 open reading frame was PCR amplified with primers 100 bp upstream and downstream. These products were purified (QIAGEN, Chatsworth, CA) and restricted with HphI, which produces six fragments of 480, 411, 303, 252, 249, and 207 bp. These fragments were examined using electrophoresis on 2%-TBE agarose gels.

Plasmid retention assay:
Cells were inoculated into SC-WLTU and grown at room temperature to midlog phase. Approximately 200 cells were plated onto YPA medium and incubated at 28° for 2–3 days until colonies reached 2 mm in diameter. Colonies were then replica plated onto SC and SC-W plates. Sectored colonies that grew on SC-W were counted and compared to the total number of colonies. The percentage of plasmid retention was calculated using the equation P = S/(T - T₀), where P is the proportion of plasmid retention, S is the number of sectored colonies, T is the total number of colonies on SC, and T₀ is the number of colonies that did not grow on SC-W (i.e., did not contain the plasmid when plated). At least 12 cultures per strain were tested and the average percentage reported. Confidence intervals for percentages were generated as described above, replacing the number of tracts, N, with the average number of colonies per culture. Significance among percentages was determined using chi-square analysis.

Chromosome loss assay:
The assay is based on loss of chromosome III. If MATa/ diploids lose chromosome III, which contains the mating-type locus, these cells can mate with haploid tester strains and can be scored using nutritional complementation. To qualitatively assess chromosome loss, diploids were first grown at 21° in YPA medium to exponential phase. For each diploid strain, 1000 cells in 25 µl were dotted onto YPA medium and incubated at 28° for 2 days. Cells were then replica plated onto YPA and 10⁶ MAT tester cells (PT2) were dotted on top of the replica-plated patches. These plates were incubated overnight at 25°, replica plated onto minimal media (SD), and incubated at 30° for 2 days to score for mating events.

MAT locus conversion assay:
To determine whether colonies from the chromosome loss assay were due to MAT locus conversion, a genetic strategy was designed. Leu⁺ diploids PGY 2038 and PGY 2128 were constructed to carry homozygous chromosome III except for the 205-kb region that was heterozygous for URA3. After incubation at 28°, single colonies of 2 x 10⁶ cells were mixed with 1 x 10⁷ cells of haploid PGY 2026 (Leu^-), which has TRP1 in the 205-kb region of chromosome III. Cells were mated for 5 hr at 25° and plated on SC-WL plates that select for mating events. Colonies that grew on these plates were replica plated onto SC-U medium. Triploid cells that have lost chromosome III should be Ura^- and carry the wild type and TRP1 insertion in the 205-kb region. If recombination has resulted in a homozygous MATa locus, triploid cells will be Ura⁺ and carry all three variants (wild type, URA3, and TRP1 insertion) in the 205-kb region. Frequencies of Ura⁺ cells can be used to measure the rate of MAT conversion using the method of the median. The 205-kb region was also amplified in these cells to verify the presence or absence of URA3 and TRP1.

Analysis of Pol1p protein levels:
TAP-tagged (RIGAUT et al. 1999) wild type (PGY 1121), pol1-1 (PGY 1131), and pol1-17 (PGY 1141) were grown at 28° to midlog phase in liquid YPA medium. Protein extracts were prepared by breaking cells using glass beads in cell lysis buffer [150 mM NaCl, 25 mM HEPES (pH 7.6), 2 mM EDTA, 2 mM dithiothreitol, 0.5% NP-40, 2x complete protease inhibitors-EDTA free (Roche Molecular Biochemicals)]. Concentrations of the soluble extract were then quantified, normalized, serially diluted twofold, fractionated on an 8% SDS-polyacrylamide gel, and transferred to a polyvinylidene difluoride membrane using standard procedures. Membranes were probed with peroxidase-anti-peroxidase (PAP) antibody (Sigma, St. Louis), which recognizes the protein A module in the TAP tag.

Viability assays:
Cells were grown to exponential phase and diluted back to 6.3 x 10⁶ cells/ml. Fivefold serial dilutions (5 µl) were spotted on YPA plates and grown at various temperatures.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Thermosensitive mutations in POL1 induce a significant mutator phenotype:
We employed a frequently used plasmid-based assay to measure repeat tract instability (HENDERSON and PETES 1992 ; STRAND et al. 1993 ; SIA et al. 1997 ; KOKOSKA et al. 1998 , KOKOSKA et al. 2000 ). In this assay, a plasmid (pSH44) containing a (GT)₁₆-repetitive tract inserted in frame within the URA3 gene is transformed into polymerase mutant strains rendering these cells Ura⁺ (HENDERSON and PETES 1992 ). Alterations occurring in the poly(GT) tract that result in out-of-frame insertions, deletions, or in the downstream URA3 coding sequence inactivate URA3 expression. The Ura^- cells can then be selected on 5-FOA-containing medium (BOEKE et al. 1987 ).

Mutation rates in two pol1 mutants, pol1-1 and pol1-17, were first compared to two pol3 mutants (pol3-01 and pol3-t) and a pol2 mutant (pol2-4). The pol3-t allele, which is an Asp₆₄₁-to-Asn₆₄₁ change (TRAN et al. 1997A ), resides between two highly conserved regions in the putative polymerase active site (WANG et al. 1997 ; FRANKLIN et al. 2001 ). Both pol3-01 and pol2-4 alleles contain Ala substitutions in the FDIE exonuclease motif that eliminate exonuclease activity and are not thermosensitive (MORRISON and SUGINO 1994 ). All strains were analyzed at the semipermissive temperature of 28°. Rates of alteration in pSH44 are shown in Table 2. Mutation rates in the wild type (PGY 2005), pol3-t (PGY 2160), pol3-01 (PGY 2170), and pol2-4 (PGY 2180) mutants were 10-fold lower compared to previously published values (STRAND et al. 1993 ; KOKOSKA et al. 1998 ). This difference can be attributed to performing the experiments at a lower temperature of 28°. When wild-type cells were assayed at the previously reported temperature of 32°, the rate of 5-FOA^r colonies rose to levels comparable with previously published values.

As previously reported (STRAND et al. 1993 ), the mutation rate in the pol2-4 mutant is similar to wild type. The mutation rate in pol1-17 is comparable with both pol3 mutant alleles; interestingly, the pol1-1 mutant displayed a four- to fivefold higher mutation rate compared to all other mutant polymerase alleles. These results indicate that POL1 plays a role in mutation avoidance, similar to that of POL3.

pol1 mutants induce two classes of mutagenic effects on the (GT)₁₆ tract:
We further analyzed the types of mutations that accumulated in these two pol1 mutants and compared them to the wild type (Table 3). In the wild type, 20% of the plasmids isolated from 5-FOA^r colonies (10 out of 50) did not carry any change in the poly(GT) tract, while 52% of the plasmids harbored one repeat unit additions. Neither proportion was significantly different from previously published data using the Fisher exact test (SIA et al. 1997 ).

Analysis of pSH44 isolated from 5-FOA^r colonies in pol1-1 (PGY2130) mutant cells revealed that only 54 of 145 events occurred in the poly(GT) tract (Table 3), suggesting that microsatellite destabilization was not the principal cause for the elevated mutation rate observed in the pol1-1 mutant (Table 2). In contrast to pol1-1, the pol1-17 mutant carried similar proportions of alterations inside and outside the poly(GT) tract (35/81 outside and 46/81 inside the tract; Table 3), suggesting that the mutation in the pol1-17 strain had a greater impact on the repeat tract than in pol1-1. The extent and nature of microsatellite instability was further evaluated in each strain (Fig 1). The pol1-1 and pol1-17 mutant exhibited a 24- and 10-fold increase in mutation rate, respectively, compared to wild type. Interestingly, the pol1-1 mutation induced a 179-fold higher mutation rate than wild type in sequences outside of the poly(GT) tract, whereas the effect induced by the pol1-17 mutation was 29-fold. When the rates of poly(GT) tract insertions and deletions were analyzed, the pol1-1 mutation was found to preferentially induce deletions, while the pol1-17 mutation induced both deletions and insertions at comparable rates (Fig 1).

View larger version (18K): In this window In a new window Download PPT slide   Figure 1. Normalized rates of poly(GT) tract instability in pol1 mutants. Values from Table 2 and Table 3 were used to determine poly(GT) tract instability by separating the rates of alteration within the poly(GT) tract and outside it. The rate of instability outside the poly(GT) tract was calculated by multiplying the rate of 5-FOA-resistant cells in pol1 mutants by the proportion of alterations that occurred outside the poly(GT) tract, while repeat tract instability was derived by multiplying the fraction of tracts that had alterations by the rate of 5-FOAr cells. The analogous method was used to calculate rates of deletions and insertions in the repeat tract. Rates are calculated with 95% confidence intervals (noted in parentheses) and presented as 10-6/cell division. The rate of instability outside the poly(GT) tract for wild type was 0.073 (0.032–0.11); for pol1-1, 13.0 (11–15); and for pol1-17, 2.1 (1.5–2.6). The rate of instability inside the poly(GT) tract for wild type was 0.29 (0.25–0.33); for pol1-1, 71.0 (55–87); and for pol1-17, 2.8 (2.0–3.6). The rate of deletion for the wild type was 0.081 (0.038–0.12); for pol1-1, 3.8 (2.5–5.1); and for pol1-17, 0.92 (0.54–1.4). The rate of insertion for the wild type was 0.21 (0.16–0.26); for pol1-1, 3.3 (2.1–4.5); and for pol1-17, 1.9 (1.4–2.4). The wild type, pol1-1 mutant, and pol1-17 mutant are denoted by the open bar, shaded bar, and striped bar, respectively. Bar graph shows rates relative to the wild type.

These data indicate that pol1-1 is a stronger mutator than pol1-17, exhibiting a twofold and sixfold higher relative mutation rate inside and outside of the repeat tract, respectively. These results also indicate that mutations in POL1 induce two distinct types of mutations: changes in the microsatellite tract and changes outside the tract, presumably downstream in URA3.

pol1-1 and pol1-17 exhibit differential mutator activities:
To test whether changes outside of the poly(GT) tract in both pol1-1 and pol1-17 occurred in microsatellite sequences endogenous to the URA3 gene, pSH44 was isolated from 5-FOA^r colonies that contain no changes in the poly(GT) tract and was digested with HindIII to produce a 3.6- and a 4.3-kb fragment (Fig 2). The majority of plasmids isolated from pol1-17 did not exhibit any apparent size changes (right, Fig 2), indicating that alterations in the URA3 coding region in pol1-17 were either small changes or point mutations. In contrast, a large portion of 5-FOA^r isolates from pol1-1 colonies had apparent size changes (left, Fig 2). These changes were grouped into two categories for both pol1-17 and pol1-1: gross alterations (including deletions, insertions, and complex changes) and no detectable size change (Table 4). Thus, a greater number of plasmids in the pol1-1 strain (38 out of 54) exhibited gross alterations compared to pol1-17 (5 out of 24; P < 0.01 by the Fisher exact test). When the rates of gross alterations and "no detectable size change" alterations were calculated and compared to the wild type (Table 5), pol1-1 and pol1-17 displayed elevated rates of "no detectable size change" alterations that were 85- and 38-fold, respectively. The pol1-1 mutant had an 300-fold higher rate of gross alterations compared to the wild type (Table 5).

View larger version (31K): In this window In a new window Download PPT slide   Figure 2. Analysis of 5-FOA-resistant clones with no poly(GT) tract changes. For detection of alterations in 5-FOA-resistant clones that showed no frameshift in the poly(GT) tract, the pSH44 plasmid characterized in HENDERSON and PETES 1992 was digested with HindIII, producing a 4.3-kb fragment (backbone) and a 3.6-kb fragment [poly(GT)-URA3 coding sequence]. Alterations were grouped into four classes: (•) no detectable size change (pol1-1: lanes 2, 4, 5, 8, 12, and 13; pol1-17: lanes 1–4 and 6–9); () insertions (pol1-1: lanes 7 and 10); () deletions (pol1-1: lanes 11 and 14; pol1-17: lane 5); and () complex changes (pol1-1: lanes 1, 3, 6, and 9).

To further characterize the types of changes induced in pol1-1 and pol1-17, we sequenced representative samples from each group. As expected, those in the "no detectable size change" category revealed base substitutions in the URA3 gene (data not shown). Those alterations categorized as deletions or complex changes, mostly observed in pol1-1, were found to be either large deletions (2–5 kb) extending into the vector or deletions within the URA3 gene. These deletion mutations were of sequences that had been flanked by short direct repeats and were reminiscent of the changes occurring in mutants of pol⁺ in Schizosaccharomyces pombe (LIU et al. 1999 ) and POL3 (pol) in S. cerevisiae (TRAN et al. 1995 , TRAN et al. 1996 ). Interestingly, all of the insertion mutations in pol1-1 (as in lanes 7 and 10, Fig 2) were identified as Ty1 transposon sequences resulting in URA3 gene disruptions, probably caused by a recombination event involving the ura3-52 locus (the ura3-52 mutation is a Ty1 transposon disruption of the URA3 gene; ROSE and WINSTON 1984 ).

To determine whether the mutator phenotype induced by the pol1-1 and pol1-17 mutations was also observed in a genomic context, the forward mutation rate was measured at the CAN1 locus. Mutation rates (expressed as 10^-7/cell division with 95% confidence intervals in parentheses) for the wild type, pol1-1, and pol1-17 strains were 1.6 (1.4–1.7), 9.7 (8.5–12), and 5.0 (3.6–5.4), respectively. Although pol1-1 and pol1-17 exhibited only a sixfold and threefold increase over the wild type, respectively, the same trend was observed. PCR amplification of 20 Can^r colonies from each strain followed by HphI digestion (see MATERIALS AND METHODS) revealed that all Can^r isolates from the wild type and the pol1-17 mutant showed no visible changes. In the pol1-1 mutant, 4 of 24 isolates showed various deletions and 7 of 24 isolates did not amplify, suggesting that gross deletions occurred in the primer site required for amplification (data not shown). These studies further suggest that the pol1-1 and pol1-17 mutant alleles promote genomic instability through different mechanisms.

Mutations induced in pol1-1 depend on the Rad52p:
To investigate whether Rad52p activity could affect the types of mutations exhibited by pol1-1 and pol1-17, double mutants with rad52 were generated. It was immediately apparent that the pol1-1 rad52 double mutant had a growth defect compared to both the rad52 and the pol1-1 single mutant (Fig 3A). Further analysis revealed that the growth defect in the pol1-1 rad52 double mutant was due to reduced growth rate (data not shown) and decreased viability, evidenced by a reduction in plating efficiency. In contrast, the growth rates in the pol1-17 rad52 double mutant and the pol1-17 single mutant were comparable. However, it was difficult to discern an effect at 30° since the pol1-17 single mutant is inherently compromised at that temperature (Fig 3A).

View larger version (26K): In this window In a new window Download PPT slide   Figure 3. Viability and mutation rates in pol1 rad52 double mutants. (A) Fivefold serial dilutions of each strain, starting from a cell density of 6.3 x 106 cells/ml, were spotted on YPA plates and incubated as described in MATERIALS AND METHODS. (B) Genomic instability outside of the poly(GT) tract. Mutation rates for alterations outside of the poly(GT) tract were derived as described above in Fig 1 using frequency and rate data from Table 6 and Table 7 with 95% confidence intervals noted in parentheses and presented as 10-6/cell division. The rate for rad52 was 2.4 (2.1–2.7); for pol1-1 rad52 6.7 (5.9–7.5); and for pol1-17 rad52 2.2 (1.8–2.6). These rates were divided by the wild-type rate and presented in the bar graph.

As shown in Table 6 and Table 7, the mutation rate in the pol1-1 rad52 double mutant displayed a threefold decrease when compared to the pol1-1 single mutant (compare Table 2, 21 x 10^-6 in the pol1-1 single mutant to 6.4 x 10^-6 in the pol1-1 rad52 double mutant in Table 6). This suggests that Rad52p activity affects the mutation rate in pol1-1. In contrast, the mutation rate in the pol1-17 rad52 double mutant did not differ from that of the pol1-17 single mutant (P = 0.23) or from that of the rad52 mutant (P = 0.16; compare 2.7 x 10^-6 in the pol1-17 rad52 double mutant in Table 6 x 10^-6 in the pol1-17 single mutant in Table 2 and to 1.7 x 10^-6 in rad52 in Table 6), suggesting that the pol1-17 allele was less affected by the absence of Rad52p activity.

View this table: In this window In a new window   Table 6. Rates and types of alterations in pol1 rad52 double mutants within poly(GT) tract

View this table: In this window In a new window   Table 7. Rates and types of alterations in pol1 rad52 double mutants outside of poly(GT) tract

Since the pol1-1 mutation had a 180-fold greater effect compared to wild type on the URA3 gene downstream of the poly(GT) tract, we analyzed the rate of outside-of-the-tract alterations in the pol1 rad52 mutants (Fig 3B). Deletion of RAD52 resulted in a twofold decrease in mutation rate for outside-of-the-(GT)-tract alterations in the pol1-1 strain, whereas in pol1-17, there was no difference (Fig 3B). These results further support the notion that loss of Rad52p activity has a greater effect on the pol1-1 mutant than on the pol1-17 mutant.

Restriction analysis of plasmid pSH44 carrying no changes in the poly(GT) tract from 5-FOA^r isolates of the pol1-1 rad52 double mutant showed that only 13% of plasmids had gross alterations (Table 7), in contrast to the 70% of plasmids that had this mutation type in the pol1-1 single mutant (Table 7; P < 0.01 by Fisher exact test). When the rates of gross alterations and "no detectable size change" alterations were calculated and compared to the wild type, the pol1-1 rad52 double mutant had a 16-fold lower rate of gross alterations compared to the pol1-1 single mutant (from 310-fold relative to wild type in the pol1-1 single mutant to 19-fold in the pol1-1 rad52 double mutant), while their rates for "no detectable size change" were comparable (from 85-fold relative to wild type to 82-fold in the pol1-1 rad52 double mutant; Table 7). Taken together, these results suggest that Rad52p activity contributes to the generation of gross alterations in the pol1-1 strain.

Mutations in pol1 induce plasmid loss and chromosome loss:
Finding that the pol1-1 rad52 double mutant had a noticeable reduction in plating efficiency when plating cells on medium for plasmid selection (data not shown) led us to investigate whether the pol1-17 and pol1-1 mutations could induce plasmid and chromosome loss at the semipermissive temperature (Fig 4).

View larger version (63K): In this window In a new window Download PPT slide   Figure 4. Plasmid and chromosome instability in pol1 rad52 mutants. (A) Each strain was transformed with the pSH44 plasmid to test for plasmid retention as described in MATERIALS AND METHODS. Error bars represent 95% confidence intervals (see MATERIALS AND METHODS). Average percentages of plasmid retention for the wild type, rad52, pol1-1, pol1-17, pol1-1 rad52, and pol1-17 rad52 were 87, 69, 92, 90, 46, and 75%, respectively. (B) Chromosome III loss in pol1 rad52 mutants was analyzed in exponentially growing homozygous diploid strains (genotype shown in Table 1) as described in MATERIALS AND METHODS. Mating events appeared as papillae after 2 days of incubation at 30°. (C) Rate of MAT locus conversion and chromosome III loss. The rate of chromosome III loss was overestimated in the wild type by taking one colony as the median value. Numbers in parentheses denote 95% confidence intervals. No colonies in the wild type exhibited a conversion event; thus the rate of MAT conversion was not determined (ND). URA3/- indicates that one chromosome carries the URA3 insertion while the other contains the intact locus.

Plasmid retention in the pol1-1 and pol1-17 single mutants was 90% and comparable to the wild type. The reduced level (69%) of plasmid retention in the rad52 mutant was epistatic to the pol1-17 rad52 double mutant (75%), suggesting that lack of Rad52p activity did not promote additional plasmid loss in the pol1-17 mutant. In contrast, the pol1-1 rad52 double mutant (46%) displayed a decrease in plasmid retention compared to the pol1-1 (92%) and rad52 (69%) single mutants. This suggests that Rad52p activity helps to maintain plasmid stability in the pol1-1 mutant. Thus, results from the plasmid retention assay revealed further differences between the pol1-1 and pol1-17 mutant alleles.

Chromosome III loss was then analyzed in these mutants as described in MATERIALS AND METHODS. In the single mutants, the level of chromosome stability seems to correlate with their respective mutation rates (Table 2 and Table 3). The weaker mutator (pol1-17) shows a lower level of chromosome loss, whereas the stronger mutator (pol1-1) shows a higher degree of chromosome instability (top row, middle and right, Fig 4B). Interestingly, chromosome loss in pol1-1 did not seem to depend on RAD52 function (right in Fig 4B; compare top and bottom rows), even though RAD52 was required for plasmid stability (Fig 4A). Furthermore, finding that the pol1-17 rad52 double mutant displayed an elevated level of chromosome loss suggests that Rad52p activity is required for preventing chromosome loss, despite the fact that its activity is not required for maintaining plasmid stability (Fig 4A). These data again suggest the intrinsic mechanistic differences between the pol1-1 and pol1-17 alleles.

Mutation at the same amino acid in pol1-1 (Gly⁴⁹³) to Glu has been shown to associate with a hyper-recombination phenotype (AGUILERA and KLEIN 1988 ). Hence, higher papillation observed for the pol1-1 mutant in the chromosome loss assay could possibly result from an increase in recombination activity that generates homozygous diploids for the MAT locus. To test this possibility, a genetic strategy was designed to measure the rate of MAT conversion in the pol1-1 mutant. Wild type (PGY 2038) and pol1-1 (PGY 2128) diploids heterozygous for a URA3 insertion at the 205-kb region of chromosome III were generated. Mating these diploids with PGY2026, which contains a TRP1 insertion in the 205-kb region of chromosome III, can differentiate chromosome loss and MAT locus conversion. Cells that have lost chromosome III will be Ura^-, while cells that have undergone MAT conversion will be Ura⁺. We found that the rate of MAT conversion in the pol1-1 mutant was 7.6 x 10^-7 while the rate of chromosome loss was 4.5 x 10^-5 (Fig 4C). These results support the notion that the high papillation is due primarily to chromosome loss.

Levels of Pol1p in pol1-1 and pol1-17:
It has previously been reported that reduced expression of DNA polymerase in a cell can lead to a mutator phenotype that is manifested by an increase in deletion mutations (KOKOSKA et al. 2000 ). To test whether the deletion mutations observed in pol1-1 and pol1-17 are caused by a decrease of the mutant Pol1p level, the levels of TAP-tagged Pol1p were measured in these mutants and compared to the wild-type TAP-tagged Pol1p (Fig 5). A slight decrease of Pol1p was noted in the pol1-17 mutant. However, there was no apparent difference in Pol1p levels between the pol1-1 mutant and the wild type. These data suggest that the mutator effects exhibited in the pol1-1 mutant are not due to a reduction in Pol1p expression.

View larger version (51K): In this window In a new window Download PPT slide   Figure 5. Levels of Pol1p in wild-type, pol1-1, and pol1-17 strains. Cell extracts were prepared from cultures that had been grown at 28°, the temperature used in all genetics assays. Total protein concentration in lysates was normalized and twofold serial dilutions were performed. TAP-tagged Pol1p was detected with the PAP antibody as described in MATERIALS AND METHODS. Coomassie staining of the total protein on the membrane is shown as a loading control.

Genetic interactions between pol1 mutants and MMR genes:
In budding yeast, there are two heterotetrameric complexes of mismatch repair proteins. One is composed of the MSH2, MSH6, PMS1, and MLH1 gene products and primarily repairs single-base mismatches but also recognizes small insertion/deletion loops (JOHNSON et al. 1996B ; MARSISCHKY et al. 1996 ). The other complex, in which MSH6 is replaced with MSH3, repairs small loops up to eight bases, but is unable to correct loops that are 20 bp (SIA et al. 1997 ). Finding that the majority of the changes in pol1-1 and in pol1-17 were either frameshift or base substitution mutations (as shown by summing together the fraction of frameshifts from Table 3 with the fraction of "no detectable size change" from Table 4) and that a putative active site pol1 mutant (pol1-Y869A) had a strong mutator effect when combined with pms1 (PAVLOV et al. 2001 ) led us to investigate whether various defects in DNA mismatch repair could alter the mutation rate and spectra of these pol1 mutants.

Both pol1-17 and pol1-1 alleles were independently combined with deletions in PMS1, MSH6, or MSH3 to generate double mutants. There was no difference in the growth rate or viability between any of these double mutants and the wild-type or the single-mutant strains, except for a slight but reproducible decrease in viability in the pol1-1 pms1 mutant at 32° (data not shown). The mutation rate (Table 8) and mutation spectra (Table 9) in these mutants were then analyzed.

The mutation rate in the pol1-17 pms1 double mutant was similar to that of the pms1 single mutant, both exhibiting a 620-fold increase over the wild type. In contrast, the 2700-fold mutation rate increase in the pol1-1 pms1 double mutant over the wild type was synergistic with respect to either of the single mutants (Table 8). When the mutation spectrum was determined in pms1 and pol1-1 pms1 (Table 9), almost all changes occurred in the poly(GT) tract (for changes outside of the GT tract, 0/71 in pol1-1 pms1 vs. 5/71 in pms1 is not significant). Moreover, both pms1 and pol1-1 pms1 displayed similar proportions of insertions (33/71 and 32/71) and deletions (33/71 and 39/71), respectively (Table 9). These data suggest that Pol1p carrying the pol1-1 mutation induces alterations in the poly(GT) tract that can ordinarily be corrected by postreplication MMR.

When analyzing the two pol1 mutants in the MSH6 deletion background, both pol1-1 msh6 and pol1-17 msh6 double mutants showed an 300-fold increase in mutation rate compared to wild type (Table 8). When comparing rates to their corresponding single mutants, the pol1-17 msh6 double mutant displays a greater than multiplicative effect, implying that deletion of MSH6 has a greater effect on the pol1-17 allele (Table 9). This is consistent with the pol1-17 mutation's preference for inducing base substitutions (Table 4 and Table 5). The mutation spectra in msh6, pol1-1 msh6, or pol1-17 msh6 did not differ from each other in the fraction of alterations occurring within or outside the poly(GT) tract (see Table 9). More than 60% of the alterations in these double mutants occurred outside of the poly(GT) tract (compare number in "0" column with number of tracts sequenced in Table 9). Since this was the signature spectrum for the pol1-1 mutant as well, 30 pSH44 isolates from FOA^r msh6, pol1-1 msh6, and pol1-17 msh6 colonies were digested with HindIII to detect large changes in the poly(GT)-URA3 sequence. None of the isolates showed any detectable size changes. Hence, these results strongly suggest that MSH6 deletion induces an increase in point mutations in both pol1-1 and pol1-17, with pol1-17 exhibiting a greater effect.

In the MSH3 deletion background, both pol1-1 msh3 and pol1-17 msh3 exhibited an increase in mutation rate compared to their respective single mutants (P < 0.01; Table 8). Combining the pol1-17 and msh3 mutations had an additive effect on the mutation rate, whereas in the pol1-1 msh3 double mutant the increase in mutation rate was four times that of an additive effect. This suggests that deletion of MSH3 has a greater impact on the mutator phenotype of pol1-1 than on the pol1-17 mutant. The mutation spectra in the pol1-1 msh3 and pol1-17 msh3 double mutants were very similar to msh3 alone, carrying no alterations outside of the poly(GT) tract (Table 9).

Taken together, these results indicate that deletion of mismatch repair genes produces a synergistic effect on the mutation rates in pol1 mutants, thus suggesting an interplay between Pol1p and the MMR system.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Analyses of the mutator phenotypes induced in two distinct pol1 conditional mutants have shown that (i) compromising POL1 function can induce genome instability by displaying elevated mutation rates and by promoting plasmid and chromosome loss (Table 2 and Table 3; Fig 4); (ii) specific mutations of POL1 can generate distinct types of mutations resulting in gross alterations, frameshifts, and base substitutions (Table 3, Table 4, and Table 5; Fig 2 and Fig 3); and (iii) the distinct types of genomic instability induced by mutations in POL1 differentially require MSH3 or MSH6 in postreplicative MMR (Table 8 and Table 9).

Induction of repeat tract alterations in pol1 mutants by polymerase slippage:
Studies analyzing several mutator alleles of POL3 have reported dinucleotide repeat instability using the assay system employed in this study (STRAND et al. 1993 ; KOKOSKA et al. 1998 , KOKOSKA et al. 2000 ). The poly(GT) tract instability in pol3 mutants is thought to reflect catalytic defects that induce polymerase slippage, since these pol3 mutant alleles mapped to regions in either the polymerase active site or the exonuclease domain. In contrast to pol3 mutations, a mutation in POL2 has shown nominal effects on dinucleotide frameshift (STRAND et al. 1993 ). Both pol1 mutations described in this study destabilize the microsatellite tract (Fig 1), suggesting that these mutations in POL1 induce an increased polymerase slippage in the repeat tract similar to that proposed in the pol3 studies (KOKOSKA et al. 1998 , KOKOSKA et al. 2000 ).

In vitro reconstituted SV40 replication experiments have indicated that Pol1p's synthetic contribution is limited to the initiator DNA (iDNA; 25 nucleotides) synthesis. Moreover, these biochemical studies have suggested that during lagging-strand synthesis, the RNA-iDNA is completely removed by RNase H and Fen-1. These in vitro results suggest that Pol1p contibutes little to the main bulk of genomic synthesis (WAGA and STILLMAN 1998 ). Thus, errors caused by the catalytic function of Pol1p are thought to have a negligible contribution to maintaining the integrity of the genome. Here, we showed that mutations in POL1 could have an in vivo effect comparable to POL3 (see Table 2 and Table 3; Fig 1).

Moreover, the synergistic increase in microsatellite destabilization in the absence of MMR (Table 8 and Table 9), particularly in pol1-1, suggests that a great number of the frameshift alterations induced in these two pol1 mutants are corrected by postreplication MMR. This suggests that Pol1p has a synthetic contribution in genomic replication. Alternatively, mutations in pol1 may lead to lower processivity of Pol1p, hence facilitating primer-template misalignments, which could have an indirect effect on the synthetic activity of Pol3p and/or Pol2p and result in alterations in the poly(GT) tract.

Possible mechanisms that induce the pol1-1 mutator phenotype:
The mutator phenotype exhibited in pol1-1 suggests that the pol1-1 mutation causes both polymerase slippage and double-strand breaks (DSBs). Previous studies have also suggested that the pol1-1 mutant is prone to polymerase slippage since it displays an elevated level of CAG tract instability and an increase in excision of 80-bp hairpins (RUSKIN and FINK 1993 ; SCHWEITZER and LIVINGSTON 1998 ). A large fraction of alterations in the pSH44 plasmids with no (GT)-tract changes isolated from pol1-1 exhibited gross alterations (Table 4). Of these, many were deletions of sequence flanked by short direct repeats that were dependent on Rad52p for their generation (Table 7). As previously proposed in studies with pol3, these deletions are likely the result of Rad52p-dependent polymerase slippage and/or the decrease in cellular Pol3p levels (TRAN et al. 1995 ; KOKOSKA et al. 2000 ). As shown in this study (Fig 5), the mutation in the pol1-1 strain does not affect the cellular Pol1p level. Hence, the deletion mutator phenotype observed in the pol1-1 mutant is a result of an intrinsic defect and not due to a decrease in Pol1p levels as seen in Pol3p (KOKOSKA et al. 2000 ). It is possible that the mutation in pol1-1 causes a conformational change in Pol1p that compromises protein-protein interactions, thus resulting in a mutator phenotype.

Another portion of the mutations seen in pol1-1 were insertions resulting from recombination with the ura3-52 locus, which normally is the consequence of DSBs. Since induced DSBs are not repaired in a rad52 background, they can be indirectly reflected in plasmid retention and chromosome loss assays. Consistent with the notion that DSBs are occurring in pol1-1, the pol1-1 rad52 double mutant shows a decreased rate of plasmid retention compared to the pol1-1 single mutant.

pol1-1, however, displays a comparable extent of chromosome loss independent of RAD52 (compare pol1-1 and pol1-1 rad52 in Fig 4B). It is unlikely this is due to the hyper-recombination properties of pol1-1 (AGUILERA and KLEIN 1988 ; LUCCHINI et al. 1990 ) since the rate of MAT locus conversion was 100-fold lower than the rate of chromosome loss (Table 8). The similar extent of chromosome loss observed in pol1-1 and pol1-1 rad52 mutants may in turn reflect a saturation of chromosome loss events that are unable to be differentiated in the assay.

Although the pol1-1 mutation is remotely located from the polymerase active site, there is an 85-fold induction of base substitutions in pol1-1 (Table 5). Several scenarios could explain this result. It is possible that the Gly-to-Arg substitution in pol1-1 may induce a conformational change that affects its catalytic function, thus resulting in base substitutions. Alternatively, the strong positive charge of this Arg residue could alter the accessibility or affinity of polymerase to the DNA backbone, thus indirectly affecting the polymerase DNA synthetic function.

These base substitutions could also be generated by activation of translesion synthesis. A recent fission yeast study has shown that mutagenic synthesis by DinB, a translesion polymerase, accounts for the elevated mutation rate and accumulation of point mutations in a pol mutant when the cell activates checkpoint function in response to replication stress (KAI and WANG 2003 ). Thus, mutation in pol1-1 could cause a replication perturbation, thereby inducing the checkpoint response that activates mutagenic translesion synthesis, resulting in base substitution mutations.

The possible mechanisms that induce the pol1-17 mutator phenotype:
The pol1-17 mutant exhibits a higher base substitution mutator phenotype than the pol1-1 mutant does. Consistent with this mutator phenotype in pol1-17, the pol1-17 msh6 double mutant exhibited a synergistic increase in mutation rate and the expected mutation spectrum (Table 8 and Table 9). This suggests that the pol1-17 mutation induces base substitutions that are often corrected by MSH6. This effect may reflect the pol1-17 mutation's location in the polymerase structure, since the pol1-17 mutation maps to the active site of Pol between the metal-activator-binding region I and nucleotide-binding region III (COPELAND et al. 1993 ; COPELAND and WANG 1993 ; DONG and WANG 1995 ). A mutation in the Pol1p catalytic domain may compromise the primer-template interaction, resulting in a decrease of fidelity and thus generating base substitutions. The ability of the pol1-17 mutation to generate base substitutions supports the notion that Pol1p contributes to the synthesis of genomic DNA sequence. As discussed above for the mechanisms that induce pol1-1 mutator phenotype, the pol1-17 mutation could also cause replication stress and induce mutagenic synthesis via the checkpoint response, generating base substitutions (KAI and WANG 2003 ).

The pol1-17 mutant also displays a small fraction of deletion mutations (Table 4) that suggest a polymerase slippage mechanism. The slightly reduced level of Pol1p in the pol1-17 strain (Fig 5) may promote these deletion mutations as seen in Pol3p (KOKOSKA et al. 2000 ). The plasmid retention assay supports the notion that the pol1-17 mutation does not seem to induce a large level of DSBs (compare rad52, pol1-17, and pol1-17 rad52 in Fig 4A), which correlates with the reduction of gross alterations in this mutant. However, the incidence of DSBs reflected in the chromosome loss assay is surprisingly high (pol1-17 rad52, Fig 4B). This difference may reflect that DSBs occur more frequently in a genomic context than in a plasmid, since there are many more origin and Okazaki fragment initiations in chromosome III than in the pSH44 plasmid, which contains only one replication origin. Hence, a subtle initiation defect in pol1-17 may be amplified as the chance for generating DSB increases in the context of the chromosome.

POL1's role in mutation avoidance:
Results of this study suggest that POL1 participates in mutation avoidance in several ways: by suppressing gross alterations, frameshift mutations, and base substitutions. Because Pol1p has an essential role in initiation and lagging-strand synthesis (WAGA and STILLMAN 1998 ; BELL and DUTTA 2002 ), it is required to interact with a wide variety of proteins involved in these processes. Dysfunction of the replication complex induced by a mutation in POL1 can stall the replication fork, resulting in mutagenic single-strand DNA. Thus the induction of DSBs or polymerase slippage in the pol1-1 mutant may reflect perturbations in the replication complex.

Studies in fission yeast have shown that a mutation in pol results in upregulation of DinB for mutagenic synthesis (KAI and WANG 2003 ). Furthermore, a mutant allele of Pol, polts13, is able to activate checkpoint effector Cds1 kinase activity and displays synthetic lethality in a cds1 deletion background (BHAUMIK and WANG 1998 ). These fission yeast studies suggest that mutations in Pol perturb the integrity of either the replication complex or the replication fork movement (BHAUMIK and WANG 1998 ; LIU et al. 1999 ; KAI and WANG 2003 ). Interestingly, fission yeast polts13 carries a three-amino-acid deletion that is localized three residues upstream of the pol1-1 allele in a highly conserved N-terminal region of Pol.

The fission yeast pol⁺ mutational studies, together with the results from this study, suggest that mutations in this highly conserved N-terminal region of Pol1p, thought to map to the surface of the polymerase, will compromise interactions with other cellular factors during initiation and lagging-strand synthesis, thus affecting genome stability. Thus mutations in the protein-protein interaction domains of replication proteins not only may provide a larger mutational target for mutator phenotype induction, but also may affect a variety of cellular processes essential for mutation avoidance.

	ACKNOWLEDGMENTS

We thank Tom Petes and David Botstein for providing us various plasmids and yeast strains for our studies. We particularly thank members of our lab for helpful discussions, Carlos Perez and Rose Borbely for their excellent technical help, and Ekaterina Schwartz for helpful advice in budding yeast work. This work is supported by a grant (no. CA14835) from the National Cancer Institute of the National Institutes of Health. P.J.A.G. is a recipient of a Predoctoral Fellowship from the Howard Hughes Medical Institute and Cancer Biology Predoctoral Training Program (grant no. CA09302) from Stanford University.

Manuscript received December 5, 2002; Accepted for publication May 6, 2003.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ADAMS-MARTIN, A., I. DIONNE, R. J. WELLINGER, and C. HOLM, 2000  The function of DNA polymerase at telomeric G tails is important for telomere homeostasis. Mol. Cell. Biol. 20:786-796.[Abstract/Free Full Text]

AGATEP, R., R. D. KIRKPATRICK, D. L. PARCHALIUK, R. A. WOODS and R. D. GIETZ, 1998 Transformation of Saccharomyces cerevisiae by the lithium acetate/single-stranded carrier DNA/polyethylene glycol (LiAC/ss-DNA/PEG) protocol. Technical Tips Online (www.umanitoba.ca/faculties/medicine/units/gietz/Trafo.html).

AGUILERA, A. and H. L. KLEIN, 1988  Genetic control of intrachromosomal recombination in Saccharomyces cerevisiae. I. Isolation and genetic characterization of hyper-recombination mutations. Genetics 119:779-790.[Abstract/Free Full Text]

BELL, S. P. and A. DUTTA, 2002  DNA replication in eukaryotic cells. Annu. Rev. Biochem. 71:333-374.[Medline]

BHAUMIK, D. and T. S.-F. WANG, 1998  Mutational effect of fission yeast Pol on cell cycle events. Mol. Biol. Cell 9:2107-2123.[Abstract/Free Full Text]

BODDY, M. N. and P. RUSSELL, 2001  DNA replication checkpoint. Curr. Biol. 11:R953-R956.[Medline]

BOEKE, J. D., J. TRUEHEART, G. NATSOULIS, and G. R. FINK, 1987  5-Fluoroorotic acid a selective agent in yeast molecular genetics. Methods Enzymol. 154:164-175.[Medline]

BUDD, M. and J. L. CAMPBELL, 1987  Temperature-sensitive mutation in the yeast DNA polymerase 1 gene. Proc. Natl. Acad. Sci. USA 84:2838-2842.[Abstract/Free Full Text]

COPELAND, W. C. and T. S.-F. WANG, 1993  Mutational analysis of the human DNA polymerase . The most conserved region in -like DNA polymerases is involved in metal-specific catalysis. J. Biol. Chem. 268:11028-11040.[Abstract/Free Full Text]

COPELAND, W. C., N. K. LAM, and T. S.-F. WANG, 1993  Fidelity studies of the human DNA polymerase . The most conserved region among -like DNA polymerases is responsible for metal-induced infidelity in DNA synthesis. J. Biol. Chem. 268:11041-11049.[Abstract/Free Full Text]

DAHLEN, M., P. SUNNERHAGEN, and T. S. WANG, 2003  Replication proteins influence the maintenance of telomere length and telomerase protein stability. Mol. Cell. Biol.