The Saccharomyces cerevisiae RAD9, RAD17, RAD24 and MEC3 Genes Are Required for Tolerating Irreparable, Ultraviolet-Induced DNA Damage (original) (raw)

Abstract

In wild-type Saccharomyces cerevisiae, a checkpoint slows the rate of progression of an ongoing S phase in response to exposure to a DNA-alkylating agent. Mutations that eliminate S phase regulation also confer sensitivity to alkylating agents, leading us to suggest that, by regulating the S phase rate, cells are either better able to repair or better able to replicate damaged DNA. In this study, we determine the effects of mutations that impair S phase regulation on the ability of excision repair-defective cells to replicate irreparably UV-damaged DNA. We assay survival after UV irradiation, as well as the genetic consequences of replicating a damaged template, namely mutation and sister chromatid exchange induction. We find that RAD9, RAD17, RAD24, and MEC3 are required for UV-induced (although not spontaneous) mutagenesis, and that RAD9 and RAD17 (but not REV3, RAD24, and MEC3) are required for maximal induction of replication-dependent sister chromatid exchange. Therefore, checkpoint genes not only control cell cycle progression in response to damage, but also play a role in accommodating DNA damage during replication.

WE previously showed that in wild-type Saccharomyces cerevisiae, the rate of progression of an ongoing S phase is slowed in the presence of a DNA-alkylating agent (Paulovich and Hartwell 1995). In contrast, checkpoint mutants replicate damaged and undamaged DNA at comparable rates. This regulation is dependent on the RAD9, RAD17, RAD24, RAD53, MEC1, MEC3, PRI1, RFA1, and RFC5 genes (Paulovich and Hartwell 1995; Longhese et al. 1996a,b; Marini et al. 1997; Paulovich et al. 1997; Sugimoto et al. 1997), but is not dependent on genes from a variety of DNA repair pathways, including base and nucleotide excision repair, recombinational repair, mismatch repair, and post-replication repair (Paulovich et al. 1997). Mutations that eliminate S phase regulation also confer sensitivity to alkylating agents, leading us to suggest that, by regulating the S phase rate, cells are either better able to repair or better able to tolerate DNA damage (Paulovich et al. 1997). Mutating known repair genes, singly or in combination, causes further slowing of the S phase rate rather than eliminating S phase regulation (Paulovich et al. 1997), suggesting that S phase slowing may be a result of accommodating unrepaired lesions during replication rather than the result of repairing lesions.

Others have shown that rad1Δ mutants, despite a global defect in nucleotide excision repair (Reynolds and Friedberg 1981; Wilcox and Prakash 1981), are able to replicate damaged DNA and survive after low levels of UV irradiation (James et al. 1978). Therefore, rad1Δ cells must have some mechanism for tolerating irreparable DNA damage. However, there are genetic consequences to replicating the damaged DNA template. Daughter strand gaps are formed (Prakash 1981), presumably straddling polymerase-blocking lesions in the template strand, and both mutation (James et al. 1978) and sister chromatid exchange (SCE; Kadyk and Hartwell 1993) are induced (Figure 1).

The reconstitution of high-molecular-weight DNA after daughter strand gap formation in rad1Δ cells is called postreplication repair, and it has been shown to be dependent on RAD6, RAD18, and RAD52 (Prakash 1981). The molecular mechanism responsible for reconstituting intact daughter strands in rad1Δ yeast has not been elucidated, and two models (Figure 1, box) have been proposed (reviewed in Naegeli 1994; Friedberg et al. 1995). In the first model, the free 3″ end of the daughter strand gap invades the sister chromatid and a Holliday junction is formed. Branch migration could then occur across the damaged template, resulting in recombinational bypass of the lesion without the need to replicate across it. Resolution of the junction would result in SCE (a combination of gene conversion and reciprocal exchange). In the second model, the free 3″ end of the daughter strand gap invades the sister chromatid duplex, ultimately using the nascent strand of the sister chromatid as a template on which to replicate around the lesion. A second template switch would then occur downstream of the lesion, allowing the polymerase to replicationally bypass the lesion and resulting in SCE (exclusively gene conversion).

Figure 1.

Model proposed to explain how cells replicate despite the presence of irreparable, UV-induced, covalent modification of the DNA template. Thin lines represent nascent strands, and arrows represent 3″ ends. Mechanisms proposed for RAD52-dependent SCE (inside box) are hypotheses and are not based on data from yeast. The reader is referred to the Introduction for a discussion of the model. Note that the existence of a RAD52-dependent recombination pathway and a REV3-dependent mutagenic pathway does not preclude the existence of other nonrecombinational, nonmutagenic pathways of DNA damage tolerance.

More is known about the UV induction of mutations. REV3 encodes an inessential DNA polymerase necessary for UV-induced mutagenesis in vivo (Lemontt 1971; Morrison et al. 1989). Rev3p and Rev7p form a complex called Polζ, which in vitro is able to replicate over a damaged template much more efficiently than other major replicases (Nelson et al. 1996). Replication across a damaged template results in the insertion of noncognate nucleotides. This is the so-called error-prone pathway that results in mutation (Figure 1).

We hypothesized that the mutagenic and/or recombinogenic mechanisms of accommodating lesions in the template may require special modes of replication that are slower than normal replicative DNA synthesis and that may depend on checkpoint gene function. This could explain why checkpoint mutants fail to slow the rate of ongoing replication in response to DNA damage and would predict that checkpoint mutants would be defective in replication-dependent SCE and/or induced mutagenesis. To test this hypothesis, we assayed the induction of SCE and mutation in checkpoint mutants in a rad1Δ excision repair-defective background after UV irradiation. We chose to work in a rad1Δ background because in wild-type cells, most lesions can be repaired before replication. We find that RAD9, RAD17, RAD24, and MEC3 are all required for UV-induced mutagenesis, and that RAD9 and RAD17 (but not REV3, RAD24, and MEC3) are required for maximal UV induction of SCE in rad1Δ cells.

MATERIALS AND METHODS

Media and growth conditions: YEPD and dropout media have been described (Sherman et al. 1981). YM-1 is described in Hartwell (1967).

Yeast strains: All strains used in this study were constructed in the A364a background. The rad9Δ::URA3 (yMP11498), rad-17Δ::URA3 (yMP11474), and rad24Δ::URA3 (yMP11477) strains were constructed by PCR-based gene replacement (Baudin et al. 1993) using oligonucleotides designed to delete >90% of the coding sequences. The oligonucleotides used for each deletion were as follows: rad9Δ::URA3 (yMP11498), OLIGO 1: 5″-CAT AGT GAG AAA ATC TTC AAC ATC AGG GCT ATG TCA GGC CGG AAA CAG CTA TGA TG-3″ and OLIGO 2: 5″-CAT CTA ACC TCA GAA ATA GTG TTG TAT ATA TCA TTG TCC GTG TAA AAC GAC GGC CAG T-3″; rad17Δ::URA3 (yMP11474), OLIGO 1: 5″-CGG TGT GGA AAC AAA GTA GTT GAA GGA TTT CAA CTA TGC GGG AAA CAG CTA TGA CCA TG-3″ and OLIGO 2: 5″-GAA TGA AGT TCT GCG TTT TCT GCG ATG CTG GAT ATT GAC TTG TAA AAC GAC GGC CAG T-3″; and rad24Δ::URA3 (yMP11477), OLIGO 1: 5″-CAC CAC TAA TTA TCA AGT TTG TTC CTG TCT GAA TGA TAT GGG GAA ACA GCT ATG ACC ATG-3″ and OLIGO 2: 5″-CCT GGG GTT TTC TCG TCA AAT TTA AAG AGT AAA AAG TTA GAG TGT AAA ACG ACG GCC AGT-3″. PCR using pJJ242 (Jones and Prakash 1990) as a template generated an intact URA3 gene flanked by sequence homology to either RAD9, RAD17, or RAD24. Each PCR product was used to transform the parent strain yMP10502, resulting in the construction of rad9Δ::URA3 (yMP11498), rad-17Δ::URA3 (yMP11474), and rad24Δ::URA3 (yMP11477) derivatives by one-step gene replacement (Rothstein 1983). Each deletion strain was confirmed by PCR analysis of genomic DNA followed by diagnostic restriction mapping (data not shown).

The rad9Δ (7835-25b), rad17Δ (DLY 279), and rad24Δ (DLY 270) strains (A364a) were kindly provided by Ted Weinert and David Lydall (Weinert and Hartwell 1990; Weinert et al. 1994; Lydall and Weinert 1995) and subsequently crossed to other strains in the A364a background to generate the checkpoint deletion strains used in this study. The identities of all strains used in the experiments in this study were confirmed by complementation testing with our newly constructed rad9Δ::URA3 (yMP11498), rad17Δ::URA3 (yMP11474), and rad24Δ::URA3 (yMP11477) strains (described above), using methyl methanesulfonate (MMS) sensitivity as an assay.

The rad6Δ disruption (A364a background) was kindly provided by Barbara Garvik. It was derived from yMP10381 by one-step gene replacement (Rothstein 1983) using a 2-kb BamHI-HindIII fragment of plasmid pJJ211 (kindly provided by Louise Prakash), which contains an allele of RAD6 in which an internal 0.6-kb EcoRI fragment is replaced with LEU2. The deletion was confirmed by genomic Southern blotting (data not shown).

The rad52Δ disruption was constructed by PCR-based gene replacement (Baudin et al. 1993) using oligonucleotides designed to delete >90% of the coding sequences. The oligonucleotides used were as follows: OLIGO 1: 5″-GAA AAA TAT AGC GGC GGG CGG GTT ACG CGA CCG GTA TCG AAG GAA ACA GCT ATG ACC ATG-3″ and OLIGO 2: 5″-GAT GCA AAT TTT TTA TTT GTT TCG GCC AGG AAG CGT TTC AGT TGT AAA ACG ACG GCC AGT-3″. PCR using pJJ250 (Jones and Prakash 1990) as a template generated an intact LEU2 gene flanked by sequence homology to RAD52. The PCR product was used to transform parent strain yMP10381, resulting in the construction of rad52Δ::LEU2 (yMP11450) by one-step gene replacement (Rothstein 1983). The deletion strain was confirmed by PCR analysis of genomic DNA followed by diagnostic restriction mapping (data not shown).

The rad27Δ (yMP11400) strain was kindly provided by Elizabeth Jensen of the Seattle Project. It was constructed by PCR-based gene replacement (Baudin et al. 1993) using oligonucleotides designed to delete >90% of the coding sequences. The oligonucleotides used were as follows: OLIGO 1: 5″-TGG AAA GAA ATA GGA AAC GGA CAC CGG AAG AAA AAA TAT GAG GAA ACA GCT ATG ACC ATG-3″ and OLIGO 2: 5″-CCC TCA TCT TCT TCC CTT TGT GAC TTT ATT CTT ATT TTT GGT TGT AAA ACG ACG GCC AGT-3″. PCR using pJJ250 (Jones and Prakash 1990) as a template generated an intact LEU2 gene flanked by sequence homology to RAD27. The PCR product was used to transform parent strain yMP10381, resulting in the construction of rad27Δ::LEU2 (yMP11450) by one-step gene replacement (Rothstein 1983). The deletion strain was confirmed by PCR analysis of genomic DNA followed by diagnostic restriction mapping (data not shown). As this strain was temperature sensitive for mitotic growth at 30°, all manipulations with this mutant were done at 23°.

YEF615 was kindly provided by Eric Foss. It was derived from yMP11450 by selecting for a His+ sister chromatid recombinant and then subsequently for a Ura− derivative of the recombinant.

The msh2Δ, rad18Δ, and rev3Δ constructions were described in Paulovich et al. (1997). The SCR::URA3 construct and the rad1Δ allele were introduced in all strains in this study by genetic crosses originating with either 8202SCR or 8271-1a (A364a derivatives), which are listed in Table 1 and described in Kadyk and Hartwell (1992, 1993).

Detection of SCEs: The SCR::URA3 sister chromatid recombination substrate has been described and diagrammed previously in detail (Kadyk and Hartwell 1992, 1993). Briefly, tandem, nonidentical ade3Δ deletions, oriented head-to-tail and containing a 305-bp region of overlap, were integrated near the centromere of chromosome III between LEU2 and HIS4. All strains used in this study also contain the ade3-130 deletion mutation, which prevents recombination between the endogenous ade3-130 locus and the SCE substrate (Kadyk and Hartwell 1992). Both the endogenous ade3-130 mutation and the tandem ade3Δ deletions in the recombination substrate confer auxotrophy for both adenine and histidine (Kadyk and Hartwell 1992). Strains bearing these ade3Δ alleles arrest within one cell cycle of being plated on C-His medium, and, therefore, all recombinants studied occur in the same cell cycle that receives irradiation (Kadyk and Hartwell 1992). Either an unequal reciprocal recombination event or a gene conversion event between tandem deletions on sister chromatids can reconstitute a functional ADE3 gene, and, hence, these events can be selected on C-His medium (Kadyk and Hartwell 1992). Intrachromosomal recombination cannot be detected in our assay because it results in the formation of an episome containing ADE3 and URA3 but lacking a centromere and an origin of replication. These recombinants are not recovered as viable His+ colonies (see Kadyk and Hartwell 1992).

Kill curves, UV-induced mutagenesis, and UV-induced SCE rates: Cells (4 × 108) were harvested by centrifugation from log phase cultures grown overnight at 30° in YM-1 + 2% glucose medium. Cells were resuspended in 133 ml YM-1 + 2% glucose medium and synchronized in the G1 phase by the addition of α factor to a final concentration of 5 μm. After a 2.5-hr incubation at 30°, cells were harvested by centrifugation and resuspended in 20 ml H2O that had been precooled to 4° to inhibit cell cycle progression during subsequent UV irradiation. After sonication to separate cells, an additional 190 ml of precooled H2O (4°) was added to the cells. We determined in pilot experiments (data not shown) that cell densities >107 cells/ml resulted in a decrease in the effective dose of UV irradiation that each cell was exposed to (based on cell killing), presumably as a result of the cells shielding each other from the UV source. Therefore, 10-ml aliquots (~4 × 106 cells/ml) were transferred to standard 100 × 15-mm plastic petri dishes and exposed, individually and while constantly swirling, to varying doses of UV irradiation. UV irradiation was delivered by a UVP source (model UVS-28; Upland, CA) at a dose rate of 0.04 J/m2/sec (0–5 J/m2 total dose range), 0.083 J/m2/sec (5–10 J/m2 total dose range), or

0.84 J/m2/sec (10–75 J/m2 total dose range). Typically, to obtain enough irradiated cells, three aliquots were irradiated for each UV dose examined in a given experiment. The three aliquots were pooled on ice immediately after UV exposure. Pooled, irradiated samples for each UV dose were harvested by centrifugation and resuspended in 1 ml H2O (4°) + 0.1 mg/ml pronase to degrade residual α factor. (Cells were maintained at 4° to inhibit their dividing before plating.) After resonication and dilution into sterile normal saline (4°), total cell concentration was determined using a Coulter channelizer (Coulter Electronics, Hialeah, FL). Viable cells per milliliter were determined by plating serial dilutions onto C + 2% glucose medium and scoring the number of colony-forming units (CFU) after 3 days of growth at 30°. Viability was calculated as CFU per milliliter per total cells per milliliter. SCE rates were determined by plating serial dilutions onto C-His + 2% glucose medium, selecting for sister chromatid recombinants. SCE rates were expressed as His+ cells per 106 viable cells for each UV dose tested. Mutation rates were determined by plating serial dilutions onto C-Arg-Ser + canavanine + 2% glucose medium, selecting for forward mutation to canavanine resistance. Mutation rates were expressed as canavanine-resistant cells per 106 viable cells for each UV dose tested. For both mutation and SCE, the rates after UV induction were determined by subtracting the observed rate of events (SCE or mutation) in the starting culture (0 J/m2) from the observed rate after exposure to a given dose of UV irradiation. In a few cases, where the numbers of recombinants or mutants recovered were small (in mutants defective in these processes), the difference between the observed rate after UV exposure and the observed rate of events (SCE or mutation) in the starting culture (0 J/m2) gave a low negative number; in these cases, the point was plotted as zero events induced. Error bars on all kill curve data were calculated as described previously (Paulovich et al. 1997). All incubations after UV irradiation were done in the dark to minimize the effects of photoreactivation.

Measurement of spontaneous SCE rates: Spontaneous SCE rates were determined by the method of the median (Lea and Coulson 1948). Liquid cultures were grown to saturation in YM-1 + 2% glucose at 30°, sonicated, diluted into sterile normal saline, and plated onto C + 2% glucose medium at a density of ~100 CFU per standard 100 × 15-mm plastic petri dish. After 3 days of growth at 30°, 20 individual colonies per strain were dissected on agar slabs using a sterile scalpel. Each agar slab (containing one individual colony) was transferred into 3 ml sterile normal saline, vortexed vigorously to resuspend the cells, and subjected to sonication to individualize the cells. Exactly 100 μl was removed from each of the 20 samples, and all 20 aliquots (100 μl each) were pooled in a separate tube to be used to assess the average CFU/colony. The pooled sample (2.0 ml) was diluted serially into sterile normal saline, plated onto C + 2% glucose medium, and incubated at 30° for 3 days to determine the average number of viable cells(CFU) in the 20 colonies examined. The remainder (2.9 ml) of each of the 20 samples was processed individually to determine the number of sister chromatid recombinants per colony. Each 2.9-ml sample was harvested by centrifugation and resuspended in 100 μl sterile normal saline, and the entire 100-μl sample was plated onto C-His medium to select for sister chromatid recombinants. After incubation at 30° for 3 days, the number of His+ colonies was scored for each of the 20 plates, and the median number of His+ colonies for all 20 plates was determined. The rate of production of recombinants was then calculated by the method of the median (Lea and Coulson 1948). Standard deviations were obtained by determination of SCE rates as described above for multiple isolates of any given mutant examined, as indicated in the Table 2 legend.

Measurement of spontaneous mutation rates: Spontaneous mutation rates were determined by the method of the median (Lea and Coulson 1948). Liquid cultures were grown to saturation in YM-1 + 2% glucose at 30°, sonicated, diluted into sterile normal saline, and plated onto C + 2% glucose medium at a density of ~100 CFU per standard 100 × 15-mm plastic petri dish. After 3 days of growth at 30°, 20 individual colonies per strain were dissected on agar slabs using a sterile scalpel. Each agar slab (containing one individual colony) was transferred into 5 ml C + 2% glucose medium, vortexed vigorously to resuspend the cells, and subjected to sonication to individualize the cells. Each of the 20 samples was then incubated at 30° for ~24 hr to allow further expansion of the population so that a significant number of Canr mutants could be obtained. (This extra step was not necessary during measurement of spontaneous SCE rates, which are ~10-fold higher than spontaneous mutation rates.) After this additional 24-hr incubation, exactly 100 μl was removed from each of the 20 samples, and all 20aliquots (100 μl each) were pooled in a separate tube to be used to assess the average CFU per colony. The pooled sample (2.0 ml) was diluted serially into sterile normal saline, and aliquots were plated onto C + 2% glucose medium and incubated at 30° for 5 days to determine the average number of viable cells (CFU) in the 20 colonies examined.

The remainder (4.9 ml) of each of the 20 samples was processed individually to determine the number of Canr mutants per culture. Exactly 0.3 ml was removed from each sample, sonicated, and plated onto C-Arg-Ser + canavanine + 2% glucose medium to select for Canr mutants. After incubation at 30° for 5 days, the number of Canr colonies was scored for each of the 20 plates, and the median number of Canr colonies for all 20 plates was determined. The rate of production of mutants was then calculated by the method of the median (Lea and Coulson 1948). Standard deviations were obtained by determination of mutation rates as described above for multiple isolates of any given strain examined, as indicated in the Table 2 legend.

RESULTS

After low-dose UV irradiation, rad1Δ mutants survive but undergo induction of SCE and mutation: Because rad1Δ mutants are completely defective in the incision step of nucleotide excision repair (Reynolds and Friedberg 1981; Wilcox and Prakash 1981), UV irradiation of rad1Δ mutants results in the introduction of irreparable covalent changes in the DNA molecule. Therefore, to proliferate after UV irradiation, rad1Δ cells are forced to replicate a damaged DNA template, resulting in stimulation of both SCE (Kadyk and Hartwell 1993) and mutation (James et al. 1978; Figure 1).

For the series of experiments described in this study, we designed an excision repair-defective rad1Δ haploid yeast strain in which we were able to measure the induction of SCE and the induction of mutation. To ensure that all strains studied were distributed uniformly in the cell cycle, cells were synchronized in the G1 phase with α factor before UV irradiation. Hence, in our experiments, SCE is induced during or after the ensuing S phase, when sister chromatids are present.

As reported previously (Cox and Parry 1968), the

TABLE 2

Spontaneous mutation and recombination rates

a

All rates were determined by the method of the median (Lee and Coulson 1948). Twenty independent colonies were analyzed for each strain. Rates are the average or mean of two to three independent experiments (except for the msh2Δ-containing strains) performed on independent segregants (as indicated below). The range or standard deviation is shown for each data point. Strains used are as follows: wild type (yMP10381, yMP10625, yMP10628);rad9Δ (yMP10899, yMP11030, yMP11038);rad17Δ (yMP11089, yMP11094, yMP11097); rad24Δ (yMP10533, yMP11006, yMP11011, yMP11020);mec3Δ (yMP11530, yMP11446);rad1Δ (yMP10619, yMP10621, yMP10633); rad1Δ rad9Δ (yMP11048, yMP11049); rad1Δ rad17Δ (yMP11095, yMP11096);rad1Δ rad24Δ (yMP11005, yMP11007); rad1Δ mec3Δ (yMP11430, yMP11434, yMP11444); rad1Δ rev3Δ (yMP10618, yMP11463); rad1Δ rad52Δ (yMP11458, yMP11459);msh2Δ (yMP11449);msh2Δ rad17Δ (yMP11385);msh2Δ rad24Δ (yMP11354).

TABLE 2

Spontaneous mutation and recombination rates

a

All rates were determined by the method of the median (Lee and Coulson 1948). Twenty independent colonies were analyzed for each strain. Rates are the average or mean of two to three independent experiments (except for the msh2Δ-containing strains) performed on independent segregants (as indicated below). The range or standard deviation is shown for each data point. Strains used are as follows: wild type (yMP10381, yMP10625, yMP10628);rad9Δ (yMP10899, yMP11030, yMP11038);rad17Δ (yMP11089, yMP11094, yMP11097); rad24Δ (yMP10533, yMP11006, yMP11011, yMP11020);mec3Δ (yMP11530, yMP11446);rad1Δ (yMP10619, yMP10621, yMP10633); rad1Δ rad9Δ (yMP11048, yMP11049); rad1Δ rad17Δ (yMP11095, yMP11096);rad1Δ rad24Δ (yMP11005, yMP11007); rad1Δ mec3Δ (yMP11430, yMP11434, yMP11444); rad1Δ rev3Δ (yMP10618, yMP11463); rad1Δ rad52Δ (yMP11458, yMP11459);msh2Δ (yMP11449);msh2Δ rad17Δ (yMP11385);msh2Δ rad24Δ (yMP11354).

rad1Δ mutant is extremely sensitive to the killing effects of UV irradiation (Figure 2A). Doses of irradiation that do not lower the viability of the wild-type, excision repair-proficient strain are highly lethal to the rad1Δ excision repair-defective strain. Furthermore, there is a dramatic stimulation of mutation and SCE at this low dose range (Kadyk and Hartwell 1993) in the excision-defective strain, but not in the wild-type strain (Figure 2, B and C),

Figure 2.

Excision repair-defective rad1Δ cells that survive exposure to UV irradiation undergo induction of SCE and mutation. RAD1 or rad1Δ cells were synchronized in the G1 phase of the cell cycle and exposed to varying doses of UV irradiation. After UV irradiation, cells were processed as described in materials and methods to determine viability (A), induction of forward mutation to Canr (B), and induction of SCE (C). All curves are the mean of at least three independent experiments that were performed on independent segregants (as indicated below). The standard deviation is shown for each data point, although in some cases, it is concealed under the symbol used for the data point. Strains used are as follows: wild type (yMP10381, yMP10625, yMP10628); rad1Δ (yMP10619, yMP10621, yMP10633).

Figure 3.

Wild-type cells that survive exposure to UV irradiation undergo induction of SCE and mutation. Rad+ cells were synchronized in the G1 phase of the cell cycle and exposed to varying doses of UV irradiation. After UV irradiation, cells were processed as described in materials and methods to determine viability (A), induction of forward mutation to Canr (B), and induction of SCE (C). All curves are the mean of at least three independent experiments performed on independent segregants (as indicated below). The standard deviation is shown for each data point, although in some cases, it is concealed under the symbol used for the data point. Strains used are yMP10381, yMP10625, and yMP10628.

and we conclude that the rad1Δ mutant is hypermutable and hyperrecombinogenic in response to UV irradiation. Interestingly, excision repair-defective human XP-A cells are also hypermutable compared with wild-type mammalian cells in response to UV irradiation (Konze-Thomas et al. 1982). These data are consistent with the model in which irreparable lesions can be accommodated during replication by processes that result in SCE and mutagenesis (Figure 1).

SCE and mutation are also induced in the wild-type strain, although at much higher doses of UV irradiation: Wild-type cells arrest in the G1 phase in response to UV irradiation, maximizing the probability that nucleotide excision repair of UV photoproducts will be completed before replication begins (Siede et al. 1993, 1994). Therefore, if SCE and mutagenesis result from replication over unrepaired lesions, one might predict that wild-type cells, benefitting from DNA excision repair and the G1 checkpoint, may not undergo induction of mutation and SCE. Alternatively, at high doses of UV irradiation, some lesions in wild-type cells might escape repair and therefore need to be accommodated during replication, just as irreparable lesions must be accommodated in rad1Δ cells. This model predicts that SCE and mutation should be induced in the wild type at high doses of UV.

We determined the response of wild-type cells to higher doses of UV irradiation than rad1Δ cells can tolerate. To achieve comparable killing in wild-type and rad1Δ backgrounds, wild-type cells must be exposed to approximately sevenfold higher doses of UV irradiation (Figure 3A), as compared to rad1Δ cells (Figure 2A). Although there was only a small induction of mutation in wild-type cells at low doses of UV irradiation (see Figure 2B), at higher doses, wild-type cells undergo mutation induction at least as effectively as rad1Δ cells (Figure 3B). Similarly, although there was very little induction of SCE in wild-type cells at low doses of UV irradiation (see Figure 2C), at higher doses (Figure 3C), wild-type cells undergo SCE induction as effectively as rad1Δ cells (Figure 2C). In fact, both rad1Δ and wild-type cells reach comparable mutation induction (~75–100 Canr mutants per 106 viable cells) and SCE induction (~100 recombinants per 106 viable cells) at roughly the same level (50% viability) of survival (compare Figures 2 and 3). These data are consistent with a model in which damage that cannot be repaired in a rad1Δ cell or that is replicated before completion of repair in a RAD1 cell may be accommodated using SCE and induced mutagenesis.

REV3 is required for the induction of mutation, but not for the induction of SCE in a rad1Δ background after UV irradiation: It had previously been shown that the rev3 mutant is defective in UV-induced mutagenesis (Lemontt 1971; Morrison et al. 1989), and that in vitro, Rev3p and Rev7p form a complex called Polζ that is able to replicate over damaged template much more efficiently than other major replicases (Nelson et al. 1996). To confirm that mutation induction in our rad1Δ strains is dependent on REV3, we constructed the rad1Δ rev3Δ double mutant and determined its response to UV irradiation. The rad1Δ rev3Δ double mutant is more sensitive to UV damage than either the rad1Δ or the rev3Δ single mutants (Figure 4A), consistent with the hypothesis that Rev3p is required for translesion synthesis across sites of unrepaired damage. Also consistent with this hypothesis, mutation induction in the rad1Δ background is dependent on REV3 (Figure 4B).

Because it was possible that the Rev3 polymerase replaced the replicative polymerase at sites of damage, and because all UV-induced SCE in a rad1Δ background were previously shown to be replication dependent (Kadyk and Hartwell 1993), it was possible that Rev3p also carried out bypass synthesis at sites of damage, resulting in SCE (see Figure 1). This model predicts that the induction of SCE should be dependent on REV3.

Figure 4.

REV3 is required for the induction of mutation, but not for the induction of SCE in a rad1Δ background after UV irradiation. rad1Δ or rad1Δ rev3Δ cells were synchronized in the G1 phase of the cell cycle and exposed to varying doses of UV irradiation. After UV irradiation, cells were processed as described in materials and methods to determine viability (A), induction of forward mutation to Canr (B), and induction of SCE (C). All curves are the mean of at least three independent experiments performed on independent segregants (as indicated below). The standard deviation is shown for each data point, although in some cases, it is concealed under the symbol used for the data point. Strains used are as follows: rad1Δ (yMP10619, yMP10621, yMP10633); rad1Δ rev3Δ (yMP10618, yMP10622, yMP11463).

However, we found that the rad1Δ rev3Δ double mutant induced the same level of SCE as the rad1Δ single mutant after UV irradiation (Figure 4C). Therefore, we conclude that REV3 is not required for replication-dependent SCE in a rad1Δ background in response to UV irradiation.

RAD52 is required for the induction of SCE, but not for the induction of mutation in a rad1Δ background after UV irradiation: It had previously been shown that the rad1Δ rad52 mutant is defective in SCE (Kadyk and Hartwell 1993).

To confirm that SCE induction in our rad1Δ strains is dependent on RAD52, we constructed the rad1Δ rad52Δ double mutant and determined its response to UV irradiation. The rad1Δ rad52Δ double mutant is more sensitive to UV damage than either the rad1Δ or the rad52Δ single mutants (Figure 5A; Kadyk and Hartwell 1993), consistent with the hypothesis that Rad52p is required for recombinational bypass of irreparable damage (see Figure 1). Also consistent with this hypothesis, SCE induction in the rad1Δ background is completely dependent on RAD52 (Figure 5C; Kadyk and Hartwell 1993). In contrast, the rad1Δ rad52Δ mutant was proficient at induced mutagenesis (Figure 5B). In fact, the double mutant is actually hypermutable because it reaches comparable levels of mutagenesis at lower doses of UV irradiation than the rad1Δ single mutant. The high induction of mutation in the rad1Δ rad52Δ mutant is consistent with a model in which lesions that cannot be tolerated recombinationally (because of deletion of RAD52) are channeled into a mutagenic repair pathway. Interestingly, both rad1Δ and rad1Δ rad52Δ mutants reach comparable mutation induction (~65 Canr mutants per 106 viable cells) at roughly the same level (40% viability) of

Figure 5.

RAD52 is required for the induction of SCE, but not for the induction of mutation in a rad1Δ background after UV irradiation. rad1Δ or rad1Δ rad52Δ cells were synchronized in the G1 phase of the cell cycle and exposed to varying doses of UV irradiation. After UV irradiation, cells were processed as described in materials and methods to determine viability (A), induction of forward mutation to Canr (B), and induction of SCE (C). All curves are the mean of at least three independent experiments performed on independent segregants (as indicated below). The standard deviation is shown for each data point, although in some cases, it is concealed under the symbol used for the data point. Strains used are as follows: rad1Δ (yMP10619, yMP10621, yMP10633); rad1Δ rad52Δ (yMP11458, yMP11459, yMP11461).

survival (compare Figures 2, A and B, and 5, A and B), raising the possibility that accumulation of lethal mutations is what kills the cells.

Deletion of RAD50 does not suppress the UV sensitivity of rad1Δ rad52Δ**:** One possible explanation for the extreme sensitivity of rad1Δ rad52Δ is that cells are attempting to execute SCE at sites of lesions, but that in the absence of Rad52p, lethal recombination intermediates are formed. This phenomenon has been observed in spo13 rad52Δ yeast mutants attempting homologue recombination in meiosis (Malone et al. 1991). Spore viability of a spo13 rad52Δ homozygous diploid can be rescued by deletion of RAD50, which blocks an earlier step in recombination and prevents formation of lethal recombination intermediates (Malone et al. 1991). We tested the possibility that this was occurring in our experiments by assaying sensitivity to UV irradiation in a rad1Δ rad50Δ rad52Δ triple mutant. If the UV sensitivity of the rad1Δ rad52Δ mutant resulted from the accumulation of RAD50-dependent lethal recombination intermediates, then we would expect deletion of RAD50 to partially suppress the UV sensitivity of this strain. However, the rad1Δ rad50Δ and rad1Δ rad50Δ rad52Δ mutants had identical UV kill curves (data not shown), making this model unlikely.

RAD9, RAD17, RAD24, and MEC3 function to help rad1Δ cells tolerate irreparable UV-induced damage: To investigate whether RAD9, RAD17, RAD24, and MEC3 [all of which play a role in the S phase DNA damage response (Longhese et al. 1996a; Paulovich et al. 1997)] are required for the survival of rad1Δ cells after UV irradiation, we determined the sensitivity of double mutants to UV damage. rad1Δ rad9Δ, rad1Δ rad17Δ, rad1Δ rad24Δ, and rad1Δ mec3Δ double mutants are all 5 to 10 times more sensitive to UV irradiation than the rad1Δ single mutant (see Figure 6A, 10 J/m2/sec dose; note that the rad1Δ rad9Δ mutant is slightly less sensitive than the others). We conclude that in the absence of nucleotide excision repair, checkpoint genes must play some essential role in tolerating UV-induced damage. A similar result has been seen in a rad16Δ excision repair-defective background (Kiser and Weinert 1996).

RAD9, RAD17, RAD24, and MEC3 are required for UV-induced mutagenesis in a rad1Δ background: Because RAD9, RAD17, RAD24, and MEC3 are required for slowing S phase progression in response to DNA damage (Paulovich et al. 1997), and because replicating irreparably damaged DNA results in mutation (James et al. 1978), we examined whether these genes are necessary for induced mutagenesis in rad1Δ cells. We measured the rates of forward mutation at the CAN1 locus after UV irradiation in rad1Δ rad9Δ, rad1Δ rad17Δ, rad1Δ rad24Δ, and rad1Δ mec3Δ double mutants. CAN1 encodes an arginine permease (Ahmad and Bussey 1986), and any loss-of-function mutation at this locus confers resistance to canavanine. Hence, there should be no bias as to the spectrum of mutation we can detect. We found that deletion of any one of these four checkpoint genes drastically reduces induced mutagenesis (Figure 6B), comparable to deletion of the REV3 locus (Figure 4B). Furthermore, we found that all possible combinations of rad9Δ, rad17Δ, and rad24Δ in a rad1Δ background also confer complete defects in induced mutagenesis (see Figure 7B). We conclude that RAD9, RAD17, RAD24, and MEC3 are required for REV3-dependent, UV-induced mutagenesis in a rad1Δ background.

RAD9 and RAD17, but neither RAD24 nor MEC3, are required for maximal UV induction of replication-dependent SCE in a rad1Δ background: Because RAD9, RAD17, RAD24, and MEC3 are required for slowing S phase progression in response to DNA damage (Paulovich et al. 1997), and because replicating damaged DNA results in SCE (Kadyk and Hartwell 1993), we were interested in asking whether these genes are necessary for UV-induced SCE. We measured the induction of SCE after UV irradiation in rad1Δ rad9Δ, rad1Δ rad17Δ, rad1Δ rad24Δ, and rad1Δ mec3Δ double mutants. We found that rad1Δ rad24Δ and rad1Δ mec3Δ double mutants induce levels of SCE comparable to the rad1Δ single mutant after UV irradiation (Figure 6C). However, rad1Δ rad9Δ and rad1Δ rad17Δ mutants induce only about half of the level of SCE induced in the rad1Δ single mutant (Figure 6C). We conclude that UV-induced SCE in a rad1Δ background is not dependent on RAD24 and MEC3, but is partially dependent on RAD9 and RAD17.

The partial SCE defects of rad9Δ and rad17Δ are not additive: We determined whether Rad9p and Rad17p were acting in the same pathway for SCE by constructing the rad1Δ rad9Δ rad17Δ triple mutant and measuring its SCE profile. As can be seen in Figure 7C, the rad1Δ rad9Δ rad17Δ mutant has a partial defect in the induction of SCE that is comparable to that of either the rad1Δ rad9Δ or rad1Δ rad17Δ mutants (see Figure 6C). We conclude that the SCE induction defects in rad1Δ rad9Δ and rad1Δ rad17Δ are not additive, and, hence, that Rad9p and Rad17p act in the same pathway of SCE induction.

UV induction of SCE in rad1Δ rad24Δ is partially dependent on RAD9 and suppressed by RAD17: As shown in Figure 6C, rad1Δ rad24Δ has the same SCE profile as the rad1Δ single mutant, in contrast to the rad1Δ rad9Δ and rad1Δ rad17Δ mutants, which show partial defects in induction of SCE (Figure 6C). To determine if the SCE observed in the rad1Δ rad24Δ mutant was induced via the RAD9- and RAD17-dependent pathway, we measured induction in rad1Δ rad9Δ rad24Δ and rad1Δ rad17Δ rad24Δ triple mutants. The rad1Δ rad9Δ rad24Δ triple mutant showed the same induction kinetics as the rad1Δ rad9Δ mutant (compare Figures 6C and 7C), and, therefore, we conclude that SCE that is induced in rad1Δ rad24Δ, just like that which

Figure 6.

Role of RAD9, RAD17, RAD24, and MEC3 in SCE and induced mutagenesis in UV-treated rad1Δ cells. rad1Δ rad9Δ, rad1Δ rad17Δ, rad1Δ rad24Δ, or rad1Δ mec3Δ double-mutant cells were synchronized in the G1 phase of the cell cycle and exposed to varying doses of UV irradiation. After UV irradiation, cells were processed as described in materials and methods to determine viability (A), induction of forward mutation to Canr (B), and induction of SCE (C). All curves are the mean of at least three independent experiments performed on independent segregants (as indicated below). The standard deviation is shown for each data point, although in some cases, it is concealed under the symbol used for the data point. Strains used are as follows: wild type (yMP10381, yMP10625, yMP10628); rad1Δ (yMP10619, yMP10621, yMP-10633); rad1Δ rad9Δ (yMP11048, yMP11049, yMP11050); rad1Δ rad17Δ (yMP11095, yMP11096, yMP11100); rad1Δ rad24Δ (yMP11005, yMP11007, yMP11013); rad1Δ mec3Δ (yMP11430, yMP11434, yMP11444).

is induced in the rad1Δ single mutant, is partially dependent on RAD9.

Surprisingly, the rad1Δ rad17Δ rad24Δ triple mutant showed a hyper-recombination phenotype (Figure 7C), stimulating an SCE level several times higher than any strain examined so far. We conclude that RAD17 suppresses SCE in the rad1Δ rad24Δ mutant. [Interestingly, the rad17Δ single mutant has a similar hyper-recombinogenic phenotype in response to UV (see below and Figure 8C).]

The hyper-recombinogenic phenotype of rad1Δ rad17Δ rad24Δ is partially dependent on RAD9: To test whether the SCE in rad1Δ rad17Δ rad24Δ was induced through the RAD9-dependent SCE pathway, we constructed the rad1Δ rad9Δ rad17Δ rad24Δ mutant and measured its response to UV irradiation. The quadruple mutant has an approximately fourfold lower level of induction of SCE than the rad1Δ rad17Δ rad24Δ triple mutant (Figure 7C). Hence, the hyper-recombination phenotype of the rad1Δ rad17Δ rad24Δ mutant is partially dependent on RAD9, and we conclude that SCE induced in rad1Δ rad17Δ rad24Δ is partially occurring through the RAD9-dependent pathway. Interestingly, however, the maximal induction in the quadruple mutant is still two to four times greater than the maximal level of induction in either a rad1Δ (Figure 2C), rad1Δ rad9Δ (Figure 6C), or a rad1Δ rad9Δ rad24Δ mutant (Figure 7C), suggesting that the hyper-recombination in the triple mutant occurs through a non-RAD9-dependent mechanism.

Checkpoint mutants in a rad1Δ background do not have elevated levels of spontaneous SCE or mutagenesis: Induction of mutation and SCE may be saturable (see Figures 2, 3, 4, 5 and 6), and the biological basis for the possible plateau in these induction curves has not been determined. Nonetheless, the possibility that these processes were saturable raised the issue that the reason checkpoint mutants might appear to be defective in UV-induced mutagenesis or SCE was simply that the spontaneous rates of these events were elevated in the checkpoint mutants. If the spontaneous levels were already at saturation, then the mutants would not be able to further induce SCE and mutation in response to UV irradiation and would, therefore, appear in our assay to be defective in these processes. To address this possibility, we determined the spontaneous rates of mutation and of SCE in every single and double mutant examined in this study (Table 2). None of the double mutants examined in this study has a spontaneous rate of SCE that is higher than the rad1Δ single mutant. We conclude that in no case does an elevated rate of spontaneous SCE explain the UV induction defects observed in the double mutants.

With the exception of rad1Δ mec3Δ, rad1Δ rad17Δ, and rad1Δ rad52Δ, none of the double mutants examined in this study has a spontaneous rate of mutation that is significantly higher than the rad1Δ single mutant. The rad1Δ rad17Δ and rad1Δ mec3Δ mutants have slightly higher spontaneous mutation rates (23.66 × 10−7 and 25.30 × 10−7, respectively) than the rad1Δ mutant (15.05 × 10−7); however, mutagenesis can be induced to at least a rate of >1000 × 10−7 (Figures 2, 3, 4, 5, 6 and 7), which provides an ample margin to detect induction if the double mutant is proficient. For example, the rad1Δ rad52Δ mutant has a higher spontaneous rate (122.58 × 10−7) than the rad1Δ mutant (15.05 × 10−7). Nonetheless, the elevated spontaneous rate in this mutant did not prevent further induced mutagenesis in response to UV, to an even greater extent than in the rad1Δ single mutant (see Figure 5B). We conclude that in no case does an elevated rate of spontaneous mutagenesis explain the defects in UV induction observed in the double mutants.

RAD9, RAD17, RAD24, and MEC3 are required for UV-induced mutagenesis in an excision repair-proficient background: UV-induced mutations are fixed at different points in the cell cycle in rad1Δ and wild-type cells (reviewed in Friedberg et al. 1995). Mutations arise with high probability prereplicatively in excision repair-proficient cells and predominantly postreplicatively in excision repair-defective cells (Eckardt and Haynes 1977; James and Kilbey 1977; Kilbey et al. 1978; Eckardt et al. 1980; Siede et al. 1983), suggesting that there may be more than one mechanism of induced mutagenesis (both REV3 dependent).

If there are mechanistic differences between induced mutagenesis in RAD1 and rad1Δ cells, mutagenesis in these two backgrounds may show different dependencies on the checkpoint genes. Therefore, having determined that RAD9, RAD17, RAD24, and MEC3 are required for mutation induction in a rad1Δ background (Figure 6B), we determined whether they are required for mutagenesis in an excision repair-proficient background. Previous studies (Prakash 1974; Lawrence and Christensen 1976) have shown that the rad9 mutant may be defective in UV- and chemically induced mutagenesis in RAD1 cells. rad9Δ, rad17Δ, rad24Δ, and mec3Δ single mutants display a mild degree (5–10-fold) of sensitivity to UV irradiation, as compared to the wild type (Figure 8A; Lydall and Weinert 1995). [Interestingly, in this RAD1 background, the rad9Δ mutant is more UV sensitive than the other single mutants (Figure 8A),

Figure 7.

Effects of combining checkpoint deletions on UV induction of SCE and mutagenesis in a rad1Δ background. rad1Δ rad9Δ rad17Δ, rad1Δ rad9Δ rad24Δ, rad1Δ rad17Δ rad24Δ, or rad1Δ rad9Δ rad17Δ rad24Δ cells were synchronized in the G1 phase of the cell cycle and exposed to varying doses of UV irradiation. After UV irradiation, cells were processed as described in materials and methods to determine viability (A), induction of forward mutation to Canr (B), and induction of SCE (C). All curves are the average or mean of two to three independent experiments performed on independent segregants (as indicated below). The range or standard deviation is shown for each data point, although in some cases, it is concealed under the symbol used for the data point. Strains used are as follows: wild type (yMP10381, yMP10625, yMP10628); rad1Δ (yMP10619, yMP10621, yMP10633); rad1Δ rad9Δ rad17Δ (yMP11550, yMP11557); rad1Δ rad9Δ rad24Δ (yMP11436, yMP11438); rad1Δ rad17Δ rad24Δ (yMP11452, yMP11453, yMP11454); rad1Δ rad9Δ rad17Δ rad24Δ (yMP11554, yMP11556).

whereas in the rad1Δ background, rad9Δ conferred a slightly lesser sensitivity to UV irradiation than did the other mutants (Figure 6A).] We find that rad9Δ, rad24Δ, and mec3Δ mutants show decreased induction of mutation compared to wild type after UV irradiation (Figure 8B). Whereas the rad17Δ mutant may be slightly hypomutable compared to the wild type, it is not as defective as the other mutants (Figure 8B). Hence, UV-induced mutagenesis in an excision repair-proficient background is largely dependent on RAD9, RAD24, and MEC3, and to a lesser degree dependent on RAD17. Therefore, mutating these checkpoint genes does not allow us to genetically separate the mutagenesis pathway in RAD1 cells from that in rad1Δ cells.

RAD9, RAD17, RAD24, and MEC3 are not required for UV-induced SCE in an excision repair-proficient background: In the rad1Δ mutant, without excision repair and without replication, recombinogenic, single-stranded DNA gaps cannot be generated. Hence, the SCE that occurs in a rad1Δ background is completely dependent on DNA replication (Kadyk and Hartwell 1993). UV-induced SCE in a rad1Δ mutant is presumably the result of postreplication repair of daughter strand gaps formed during replication across UV lesions in the template strand (see Figure 1 and Introduction).

SCE occurring in wild-type cells synchronized in the G1 phase before irradiation (Figure 3C) may result from the same mechanism (postreplication repair) if some lesions escape surveillance and are replicated before excision repair. However, unlike rad1Δ cells, wild-type cells have a second possible mechanism for SCE induction. If, in wild-type cells, replication forks were to encounter nucleotide excision repair-induced single-stranded gaps, this could result in the formation of one intact and one broken sister chromatid. After replication, the broken chromatid could then be repaired by homologous recombination using the intact sister as a template.

We have shown that SCE in rad1Δ cells is partially dependent on RAD9 and RAD17, but not dependent on RAD24 and MEC3. Given the possibility that SCE is induced via different mechanisms in rad1Δ cells compared to wild type, we determined whether any of these genes are required for SCE in an excision repair-proficient background. We find that none of the checkpoint genes is necessary for SCE induction in a RAD1 background (Figure 8C). Furthermore, the rad17Δ and mec3Δ mutants are actually hyper-recombinogenic (Figure 8C). We conclude that none of these checkpoint genes is required for UV-induced SCE in a RAD1 background, and that Rad17p and Mec3p actually suppress SCE in excision repair-proficient cells. These differences in genetic requirements for SCE in wild-type vs. rad1Δ cells are consistent with the model that SCE occurs at least in part through different mechanisms in nucleotide excision repair-competent and -deficient cells.

RAD1, REV3, RAD9, RAD17, RAD24, and MEC3 are not required for spontaneous SCE or mutation during mitotic growth: As previously discussed, the apparent defect in induced mutagenesis in these checkpoint mutants could result if the mutants had levels of spontaneous mutagenesis so high that further induction would be impossible, either because of saturation of the mutagenesis system or because of accumulation of lethal mutations. However, as can be seen from the data in Table 2, rad1Δ, rev3Δ, rad9Δ, rad17Δ, rad24Δ, and mec3Δ single mutants have rates of spontaneous SCE and mutation comparable to that of the wild type. We conclude that in no case can the observed defects in induced mutagenesis in checkpoint mutants be explained by elevated rates of spontaneous mutagenesis. Note that despite the fact that REV3, RAD9, RAD17, RAD24, and MEC3 are required for UV-induced mutagenesis (Figure 6A), none of these genes is necessary for a wild-type rate of spontaneous mutagenesis (Table 2).

Dramatic elevations in the spontaneous mutation rate are seen in the rad1Δ mutant (15.05 × 10−7) and especially in the rad52Δ mutant (45.39 × 10−7; Von Borstel et al. 1971) compared to the wild type (4.11 × 10−7; Table 2). The elevated level observed in the rad52Δ mutant is consistent with a model in which lesions are channeled into the mutagenic repair pathway when recombinational repair is not available (see Figure 1). Furthermore, the elevation in spontaneous mutation rates in the rad1Δ and the rad52Δ single mutants is synergistic because the increase in the rad1Δ rad52Δ double mutant is greater than the sum of the increases in the single mutants, consistent with a model in which excision repair and recombination repair may compete for common lesions.

The mutator phenotype of the mismatch repair mutant msh2Δ is not dependent on RAD17 or RAD24: msh2Δ cells, defective in mismatch repair, have an elevated spontaneous mutation rate compared to that of the wild type (Table 2; Reenan and Kolodner 1992). Mutations presumably result when errors made by cellular replicases go unrepaired. Because checkpoint genes are necessary for the S phase DNA damage response (Paulovich and Hartwell 1995; Longhese et al. 1996a; Paulovich et al. 1997), and because they are required for UV-induced mutagenesis, we asked whether the high

Figure 8.

Effects of checkpoint deletions on UV induction of SCE and mutagenesis in an excision repair-proficient background. rad9Δ, rad17Δ, rad24Δ, and mec3Δ single mutants were synchronized in the G1 phase of the cell cycle and exposed to varying doses of UV irradiation. After UV irradiation, cells were processed as described in materials and methods to determine viability (A), induction of forward mutation to Canr (B), and induction of SCE (C). All curves are the average or mean of two to three independent experiments performed on independent segregants (as indicated below). The range or standard deviation is shown for each data point, although in some cases, it is concealed under the symbol used for the data point. Strains used are as follows: wild type (yMP10381, yMP10625, yMP10628); rad9Δ (yMP10899, yMP11030, yMP11038); rad17Δ (yMP11089, yMP11094, yMP11097); rad24Δ (yMP10533, yMP11006, yMP11011, yMP11020); mec3Δ (yMP11446, yMP11447).

level of spontaneous mutation in the msh2Δ mutant was dependent on two representative checkpoint genes, RAD17 or RAD24. We constructed rad17Δ msh2Δ and rad24Δ msh2Δ double mutants and measured their spontaneous mutation rates. Both double mutants have spontaneous mutation rates similar to the msh2Δ single mutant (Table 2). We conclude that neither RAD17 nor RAD24 is necessary for spontaneous mutagenesis in the msh2Δ mutant, and, therefore, that spontaneous damage normally repaired by the MSH2 pathway is not channeled into the RAD17- and RAD24-dependent mutagenic pathway in the msh2Δ mutant.

Mutation of RAD9, RAD17, or RAD24 is synthetic lethal with mutation of RAD27: A mutator phenotype has recently been described for the rad27Δ mutant (Tishkoff et al. 1997). The mutator phenotype of rad27Δ is not dependent on mismatch repair (Tishkoff et al. 1997). Moreover, the spectrum of mutations that arise in the rad27Δ mutant is unique (Tishkoff et al. 1997). Because checkpoint genes are necessary for UV-induced mutagenesis, we wanted to determine if the high rate of spontaneous mutation in the rad27Δ mutant was occurring through the checkpoint-dependent mutational system. When we attempted to construct rad9Δ rad27Δ, rad17Δ rad27Δ, and rad24Δ rad27Δ double mutants by crossing haploid parent strains carrying either rad27Δ or a checkpoint gene deletion, viability predominantly segregated 3:1 in these tetrads, and no rad9Δ rad27Δ, rad17Δ rad27Δ, or rad24Δ rad27Δ double mutants were recovered. (Furthermore, the majority of dead segregants were inferred to be double mutants.) We conclude that mutation of RAD9, RAD17, or RAD24 is synthetic lethal with mutation of RAD27, raising the interesting possibility that spontaneous lesions occurring in a rad27Δ background might be handled through the checkpoint-dependent mutational system. [Of course, there may also be other functions of checkpoint genes that are essential in rad27Δ cells. Synthetic lethality between rad9Δ and rad27Δ has been shown by others (Vallen and Cross 1995).]

rad1Δ rad18Δ and rad1Δ rad6Δ double mutants have low plating efficiency: Both RAD6 and RAD18 confer severe sensitivity to UV irradiation (Cox and Parry 1968; Friedberg et al. 1995), and both are necessary for postreplication repair (Prakash 1981) and for UV-induced mutagenesis (Lawrence and Christensen 1976). Therefore, we wanted to determine whether either gene was necessary for sister chromatid recombination. We crossed a rad1Δ parent strain to a strain carrying either rad6Δ or rad18Δ. We were able to recover the double mutants at the expected frequency. However, both the rad1Δ rad18Δ and rad1Δ rad6Δ double mutants had a severe growth defect in our A364a background that was associated with plating efficiencies of 1.5 and 22%, respectively. The low plating efficiencies of these double mutants made it technically unfeasible to determine their UV induction curves for mutation and SCE.

DISCUSSION

Checkpoint genes are involved in replicating irreparable DNA damage: The role of RAD9, RAD17, RAD24, and MEC3 in the regulation of cell cycle progression in response to DNA damage is well established (Weinert and Hartwell 1988, 1990, 1993; Weinert et al. 1994). Some of these genes have also been implicated in the processing of DNA damage (Lydall and Weinert 1995). We find that rad1Δ rad9Δ, rad1Δ rad17Δ, rad1Δ rad24Δ, and rad1Δ mec3Δ mutants are all more sensitive to UV-induced killing than the rad1Δ single mutant, suggesting a role for these genes in helping cells survive DNA damage, even in the absence of nucleotide excision repair (Figure 6A; see Kiser and Weinert 1996 for similar results in rad16Δ background). Furthermore, we show that mutation induction in rad1Δ cells is dependent on RAD9, RAD17, RAD24, and MEC3, and that SCE induction is partially dependent on RAD9 and RAD17 (Figure 6, B and C). Therefore, checkpoint genes not only control cell cycle progression in response to damage, but they also play a role in accommodating DNA damage during replication.

Checkpoint genes may play either a direct or an indirect role in mutagenesis and replication-dependent SCE: One possible explanation for the dependence of UV-induced mutagenesis on Rad9p, Rad17p, Rad24p, and Mec3p is that these proteins may be directly involved in translesion synthesis. For example, they may be physically associated with the Rev3p-containing Polζ complex. Specifically, because RAD24 contains areas of similarity to RF-C subunits and shows genetic interactions with RFC1 (Lydall and Weinert 1997), perhaps in response to DNA damage, Rad24p becomes associated with RF-C, and this association results in loading of the Rev3 polymerase rather than the normal replicase. Another possibility is that a putative Rad17p-dependent exonuclease [Rad17p is homologous to a 3″-5″ exonuclease from Ustilago; see Lydall and Weinert (1995)] may degrade the 3″ end of nascent strands terminated at UV lesions, allowing the Rev3p polymerase the opportunity to load onto the primer and subsequently replicate over the UV lesion. Alternatively, the role of checkpoint genes in induced mutagenesis may be indirect. For example, transcriptional induction of a variety of genes in response to UV irradiation has been shown to be dependent on the checkpoint genes in both excision repair-proficient and excision repair-deficient cells (Aboussekhra et al. 1996; Kiser and Weinert 1996). Therefore, it is possible that the dependence of induced mutagenesis on checkpoint genes reflects defects in checkpoint mutants in the transcriptional induction of genes that are directly involved in translesion synthesis.

In contrast to induced mutagenesis, replication-dependent SCE in rad1Δ cells does not require Rad24p and Mec3p, but does in part depend on Rad9p and Rad17p (Figure 6C). As with mutation induction, Rad9p and Rad17p may play either a direct or an indirect role in SCE. For example, they may associate with the replication complex at sites of damage to facilitate template switching and bypass of lesions (Figure 1, box), or they may be an integral component of the recombinational apparatus involved in repairing daughter strand gaps. The observation that some but not all the checkpoint proteins are required for SCE may favor models of direct involvement in the process over more indirect models, such as a general requirement for transcriptional induction, as discussed above. However, there certainly are documented differences in phenotypes among checkpoint mutants with respect to DNA damage-induced transcription of repair genes (Kiser and Weinert 1996). Nonetheless, the possibility that the observed partial defects are caused by transcriptional induction defects in the checkpoint mutants seems less likely because rad1Δ rad17Δ rad24Δ has a hyper-recombinogenic phenotype (Figure 7C), yet it would presumably have at least as much of a transcriptional induction defect as the rad1Δ rad17Δ mutant, which is partially defective in SCE (Figure 6C).

What is Rad17p's role in SCE? RAD17 is unique among all genes examined in that it is required for SCE in some backgrounds, whereas it acts as a suppressor of SCE in other backgrounds. For example, maximal induction of SCE in the rad1Δ rad17Δ (Figure 6C) mutant is about half of the maximal induction in the rad1Δ single mutant, demonstrating partial dependence of replication-dependent SCE on RAD17. Paradoxically, deletion of RAD17 in the rad1Δ rad24Δ strain results in a hyper-recombination phenotype, with maximal level of induction of SCE in the rad1Δ rad17Δ rad24Δ triple mutant being four to five times higher than that in the rad1Δ rad24Δ double mutant (Figure 7C). Hence, it appears as though, at least in an excision repair-defective background, the ability of the deletion of RAD17 to cause a hyper-recombination phenotype is limited to backgrounds in which RAD24 has also been deleted. These epistatic interactions are depicted in a genetic model (suggested by Ted Weinert) in Figure 9. (Interestingly, Rad9p opposes the activity of a Rad17p-dependent accumulation of single-strand DNA at telomeres in cdc13 cells at nonpermissive temperature, whereas in UV-irradiated rad1Δ cells, Rad9p and Rad17p act together to facilitate SCE. While we do not understand the biochemical basis of this difference, it may be caused by different substrates in the two systems.)

Relationship between defects in replication-dependent SCE and mutagenesis and defects in S phase regulation: In response to alkylation damage, wild-type yeast cells dramatically slow the rate at which chromosomal DNA is replicated (Paulovich et al. 1995). The mec1-1 and rad53 mutants, in contrast, replicate damaged and undamaged DNA at roughly the same rate. rad9Δ, rad17Δ, and rad24Δ have an intermediate phenotype; they replicate damaged DNA slower than mec1-1 and rad53, but faster than wild type in response to damage (Paulovich et al. 1995, 1997). We have observed similar effects of these mutations during S phase in a rad1Δ background in response to UV irradiation (data not shown). One model we proposed (Paulovich et al. 1997) to explain these two different phenotypic classes is that slowing of S phase is the result of cells' replicating over DNA lesions using either translesion or bypass synthesis. If mec1-1 and rad53 were defective in both translesion and bypass synthesis, whereas rad9Δ, rad17Δ, and rad24Δ were defective in only one of these processes, then these differences could explain the two phenotypic classes of defects in S phase regulation. So far, our data are consistent with this model because, despite the fact that rad9Δ, rad17Δ, and rad24Δ are defective in UV-induced mutagenesis (Figure 6B), all mutants are able to carry out at least some sister chromatid recombination (Figure 6C). The sister chromatid recombination occurring during S phase in these mutants could account for a partial slowing of S phase in response to damage. It will be interesting to measure UV induction of mutation and SCE in rad1Δ mec1 and rad1Δ rad53 strains, as this model predicts that these mutants will be completely defective in both mutation and SCE induction.

Relationship between defects in replication-dependent SCE and mutagenesis and defects in postreplication repair: UV photoproducts block the progress of DNA polymerases and consequently cause the formation of daughter strand gaps across lesions during replication (Prakash 1981; reviewed in Friedberg et al. 1995). The reconstitution of high molecular weight DNA after daughter strand gap formation is called postreplication repair, and it has been shown to be dependent on RAD6, RAD18, and RAD52 in a rad1Δ background (Prakash 1981). Because rad1Δ rev3 mutants are not defective in postreplication repair (Prakash 1981) but are defective in UV-induced mutagenesis (Morrison et al. 1989), it appears as though translesion synthesis is not required for the repair of daughter strand gaps. In contrast, the lack of SCE induction in the rad1Δ rad52Δ mutant (Figure 5; Kadyk and Hartwell 1993) is correlated with its defect in postreplication repair (Prakash 1981), suggesting that daughter strand gap repair largely reflects recombinational bypass of

Figure 9.

Genetic pathway showing epistatic interactions between RAD9, RAD17, RAD24, MEC3, RAD52, and REV3. (A) In the rad1Δ mutant, Rad9p, Rad17p, Rad24p, and Rad52p all regulate SCE. Our data can be explained by a model in which Rad9p and Rad52p positively regulate SCE, Rad24p negatively regulates SCE, and Rad17p negatively regulates both Rad24p and SCE. RAD9, RAD17, RAD24, MEC3, and REV3 are all required for UV-induced mutagenesis in the rad1Δ mutant. (B) In excision repair-proficient (RAD1) cells, Rad17p and Mec3p are suppressors of SCE, and RAD9, RAD17, RAD24, MEC3, and REV3 are all required for UV-induced mutagenesis.

lesions. Consistent with this conclusion, the rad1Δ rad52Δ mutant is hypermutable (and therefore presumably carrying out lots of translesion synthesis), yet has a defect in daughter strand gap repair (Prakash 1981). Given our finding that rad1Δ rad9Δ and rad1Δ rad17Δ are partially defective in replication-dependent SCE, it will be of interest to determine if these mutants have a defect in postreplication repair.

Are lesions channeled between mutagenic and recombinogenic pathways? Others have suggested that DNA lesions may be tolerated by both recombination and mutation, but that cells somehow channel lesions preferentially into one pathway over the other (Schiestl et al. 1990; Rong et al. 1991; Aboussekhra et al. 1992; Heude and Fabre 1993). Our data are consistent with the hypothesis that the recombination pathway is preferred over the mutation pathway. In wild-type cells, SCE is induced at lower doses of UV than induction of mutation (Figure 2, B and C). One might argue that this results from there being more than one type of lesion induced by UV, and that a more prevalent lesion is replicated strictly by the recombinational pathway, whereas a less prevalent lesion is replicated strictly by the mutagenic pathway. However, at least in a rad1Δ background, when the option of recombination is taken away by deletion of RAD52 (Figure 3B), cells are hypermutable in response to UV irradiation, suggesting that lesions that would have been handled recombinationally are now secondarily channeled to the mutagenic pathway.

Interestingly, the rad1Δ rev3Δ mutant is not hyper-recombinogenic in response to UV (Figure 4C), suggesting that lesions normally handled by the REV3 pathway may not be channeled into the SCE pathway. Alternatively, the number of lesions normally handled by the REV3 pathway may be negligible relative to the number of lesions normally handled by SCE. (The SCE rates we measure are likely to underestimate the actual rates because we are only able to detect unequal recombination events.) Therefore, channeling these lesions into SCE (by deletion of REV3) would not result in a detectable increase in the SCE rate.

SCE: Measurement of SCE induction is widely used to determine the genotoxicity of chemicals in hopes of predicting their carcinogenicity (Perry and Evans 1975). Additionally, measurement of SCE induction in peripheral blood lymphocytes in people is used to assess the exposure of individuals to DNA-damaging agents (reviewed in Tucker et al. 1993) or to assess the inherent genetic instability of their cells in hopes of correlating cellular phenotypes with disease (e.g., Bloom's syndrome; Chaganti et al. 1974). Amazingly, despite its widespread use as a marker of genome damage, we know virtually nothing about the molecular mechanisms underlying the process of SCE or its regulation.

Work in yeast has suggested that SCE may occur via more than one mechanism. For example, there are qualitative differences between SCEs induced in excision repair-defective rad1Δ cells and SCEs induced in wild-type cells. UV-induced SCE that occurs in a rad1Δ background is completely dependent on lesions being present during DNA replication, whereas SCE in a wild-type background can be induced by UV damage occurring in the G2 phase, without an intervening S phase (Kadyk and Hartwell 1993). Furthermore, SCE in the rad1Δ background is completely caused by gene conversion, whereas SCE in a wild-type background is a mixture of gene conversion and reciprocal recombination (Kadyk and Hartwell 1993). We show that SCE in a rad1Δ background is partially dependent on RAD9 and RAD17, whereas SCE in a wild-type background is independent of these genes. All these observations support the hypothesis that G2 SCE in wild-type cells is mechanistically different from the replication-dependent SCE observed in rad1Δ strains.

One possible basis of these differences in SCE in wild-type and rad1Δ cells is that replication-dependent SCE in a rad1Δ cell and replication-independent (G2) SCE in a wild-type cell are stimulated by different substrates. For example, in the rad1Δ mutant, in the absence of excision repair and without replication, recombinogenic, single-stranded DNA gaps cannot be generated. Hence, in the rad1Δ mutant, recombination is dependent on replication over UV lesions, which results in the formation of recombinogenic daughter strand gaps (Figure 1). In this case, the SCE-inducing substrate is presumably a single-stranded gap across a template that contains an irreparable UV photoproduct (Figure 1). In contrast, wild-type cells have at least two mechanisms to generate recombinogenic lesions. Lesions that happened to be replicated rather than repaired could presumably stimulate daughter strand gap formation and SCE in the wild type, just as in the rad1Δ cells. A second possible source of recombinogenic lesions in wild-type cells is nucleotide excision repair. For example, perhaps in nucleotide excision repair-proficient cells, replication forks occasionally encounter single-stranded excision gaps. This could result in replication fork breakage and the formation of one broken and one intact sister chromatid (reviewed in Kuzminov 1995), followed by repair of the broken sister using SCE in a process that is not dependent on RAD9 and RAD17.

Acknowledgement

We thank Russell Dorer and Dave Toczyski for formative discussions. We thank members of the Hartwell lab for comments on this manuscript, Rebecca Margulies for help with pilot experiments during the early stages of this work, Nick Terzopoulos for media preparation, and Laura Reiter for tetrad dissection. We especially thank Christopher Lawrence for helpful comments on this manuscript, and Ted Weinert for helpful comments on the manuscript and for suggesting the epistasis model depicted in Figure 9. This work was supported by General Medical Sciences grant GM-17709 from the National Institutes of Health. A.G.P. was supported by a Merck Distinguished Fellow Award and a Medical Scientist Training Program (MSTP) training grant from the National Institutes of Health. L.H.H. is a Research Professor of the American Cancer Society.

Footnotes

Communicating editor: M. D. Rose

LITERATURE CITED

Aboussekhra

A

,

Chanet

R

,

Adjiri

A

,

Fabre

F

,

1992

Semidominant suppressors of Srs2 helicase mutation of Saccharomyces cerevisiae map in the RAD51 gene, whose sequence predicts a protein with similarities to prokaryotic RecA proteins

.

Mol. Cell. Biol.

12

:

3224

–

3234

.

Aboussekhra

A

,

Vialard

J E

,

Morrison

D E

,

de la Torre-Ruiz

M A

,

Cernakova

L

et al. ,

1996

A novel role for the budding yeast RAD9 checkpoint gene in DNA damage-dependent transcription

.

EMBO J.

15

:

3912

–

3922

.

Ahmad

M

,

Bussey

H

,

1986

Yeast arginine permease: nucleotide sequence of the CAN1 gene

.

Curr. Genet.

10

:

587

–

592

.

Baudin

A

,

Ozier-Kalogeropoulis

O

,

Denouel

A

,

Lacroute

F

,

Cullin

C

,

1993

A simple and efficient method for direct gene deletion in Saccharomyces cerevisiae

.

Nucleic Acids Res.

21

:

3329

–

3330

.

Chaganti

R S

,

Schonberg

S

,

German

J

,

1974

A manyfold increase in sister chromatid exchanges in Bloom's syndrome lymphocytes

.

Proc. Natl. Acad. Sci. USA

71

:

4508

–

4512

.

Cox

B S

,

Parry

J M

,

1968

The isolation, genetics and survival characteristics of ultraviolet light-sensitive mutants in yeast

.

Mutat. Res.

6

:

37

–

55

.

Eckardt

F E

,

Haynes

R H

,

1977

Induction of pure and sectored mutant clones in excision-proficient and deficient strains of yeast

.

Mutat. Res.

43

:

327

–

338

.

Eckardt

F E

,

Teh

S J

,

Haynes

R H

,

1980

Heteroduplex repair as an intermediate step of UV-mutagenesis in yeast

.

Genetics

95

:

63

–

80

.

Friedberg

E C

,

Walker

G C

,

Siede

W

,

1995

DNA Repair and Mutagenesis

.

ASM Press

,

Washington, DC

.

Hartwell

L H

,

1967

Macromolecular synthesis in temperature sensitive mutants of yeast

.

J. Bacteriol.

93

:

1662

–

1670

.

Heude

M

,

Fabre

F

,

1993

a/alpha-control of DNA repair in the yeast Saccharomyces cerevisiae: genetic and physiological aspects

.

Genetics

133

:

489

–

498

.

James

A P

,

Kilbey

B J

,

1977

The timing of UV mutagenesis in yeast: a pedigree analysis of induced recessive mutation

.

Genetics

87

:

237

–

248

.

James

A P

,

Kilbey

B J

,

Prefontaine

G J

,

1978

The timing of UV mutagenesis in yeast: continuing mutation in an excision defective (rad1-1) strain

.

Mol. Gen. Genet.

165

:

207

–

212

.

Jones

J S

,

Prakash

L

,

1990

Yeast Saccharomyces cerevisiae selectable markers in pUC18 polylinkers

.

Yeast

6

:

363

–

366

.

Kadyk

L C

,

Hartwell

L H

,

1992

Sister chromatids are preferred over homologues as substrates for recombinational repair in Saccharomyces cerevisiae

.

Genetics

132

:

387

–

402

.

Kadyk

L C

,

Hartwell

L H

,

1993

Replication-dependent sister chromatid recombination in rad1 mutants of Saccharomyces cerevisiae

.

Genetics

133

:

469

–

487

.

Kilbey

B J

,

Brychcy

T

,

Nasim

A

,

1978

Initiation of UV mutagenesis in Saccharomyces cerevisiae

.

Nature

274

:

889

–

891

.

Kiser

G L

,

Weinert

T A

,

1996

Distinct roles of yeast MEC and RAD checkpoint genes intranscriptional induction after DNA damage and implications for function

.

Mol. Biol. Cell

7

:

703

–

718

.

Konze-Thomas

B

,

Hazard

R M

,

Maher

V M

,

McCormick

J J

,

1982

Extent of excision repair before DNA synthesis determines the mutagenic but not the lethal effect of UV radiation

.

Mutat. Res.

94

:

421

–

434

.

Kuzminov

A

,

1995

Collapse and repair of replication forks in Escherichia coli

.

Mol. Microbiol.

16

:

373

–

384

.

Lea

D E

,

Coulson

C A

,

1948

The distribution of numbers of mutants in bacterial populations

.

J. Genet.

49

:

264

–

284

.

Lawrence

C W

,

Christensen

R

,

1976

UV mutagenesis in radiation-sensitive strains of yeast

.

Genetics

82

:

207

–

232

.

Lemontt

J F

,

1971

Mutants of yeast defective in mutation induced by ultraviolet light

.

Genetics

68

:

21

–

33

.

Longhese

M P

,

Fraschini

R

,

Plevani

P

,

Lucchini

G

,

1996a

Yeast pip3/mec3 mutants fail to delay entry into S phase and to slow DNA replication in response to DNA damage, and they define a functional link between Mec3 and DNA primase

.

Mol. Cell. Biol.

16

:

3235

–

3244

.

Longhese

M P

,

Neecke

H

,

Paciotti

V

,

Lucchini

G

,

Plevani

P

,

1996b

The 70 kDa subunit of replication protein A is required for the G1/S and intra-S DNA damage checkpoints in budding yeast

.

Nucleic Acids Res.

24

:

3533

–

3537

.

Lydall

D

,

Weinert

T

,

1995

Yeast checkpoint genes in DNA damage processing: implications for repair and arrest

.

Science

270

:

1488

–

1491

.

Lydall

D

,

Weinert

T

,

1997

G2/M checkpoint genes of Saccharomyces cerevisiae: further evidence for roles in DNA replication and/or repair

.

Mol. Gen. Genet.

256

:

638

–

651

.

Malone

R E

,

Bullard

S

,

Hermiston

M

,

Rieger

R

,

Cool

M

et al. ,

1991

Isolation of mutants defective in early steps of meiotic recombination in the yeast Saccharomyces cerevisiae

.

Genetics

128

:

79

–

88

.

Marini

F

,

Pellicioli

A

,

Paciotti

V

,

Lucchini

G

,

Plevani

P

et al. ,

1997

A role for DNA primase in coupling DNA replication to DNA damage response

.

EMBO J.

16

:

639

–

650

.

Morrison

A

,

Christensen

R B

,

Alley

J

,

Beck

A K

,

Bernstine

E G

et al. ,

1989

REV3, a Saccharomyces cerevisiae gene whose function is required for induced mutagenesis, is predicted to encode a nonessential DNA polymerase

.

J. Bacteriol.

171

:

5659

–

5667

.

Naegeli

H

,

1994

Roadblocks and detours during DNA replication: mechanisms of mutagenesis in mammalian cells

.

Bioessays

16

:

557

–

564

.

Nelson

J R

,

Lawrence

C W

,

Hinkle

D C

,

1996

Thyminethymine dimer bypass by yeast DNA polymerase ζ

.

Science

272

:

1646

–

1649

.

Paulovich

A G

,

Hartwell

L H

,

1995

A checkpoint regulates the rate of progression through S phase in S. cerevisiae in response to DNA damage

.

Cell

82

:

841

–

847

.

Paulovich

A G

,

Margulies

R U

,

Garvik

B M

,

Hartwell

L H

,

1997

RAD9, RAD17, and RAD24 are required for S phase regulation in Saccharomyces cerevisiae in response to DNA damage

.

Genetics

145

:

45

–

62

.

Perry

P

,

Evans

H J

,

1975

Cytological detection of mutagencarcinogen exposure by sister chromatid exchange

.

Nature

258

:

121

–

125

.

Prakash

L

,

1974

Lack of chemically induced mutation in repair-deficient mutants of yeast

.

Genetics

78

:

1101

–

1118

.

Prakash

L

,

1981

Characterization of postreplication repair in Saccharomyces cerevisiae and effects of rad6, rad18, rev3 and rad52 mutations

.

Mol. Gen. Genet.

184

:

471

–

478

.

Reenan

R A

,

Kolodner

R D

,

1992

Characterization of insertion mutations in the Saccharomyces cerevisiae MSH1 and MSH2 genes: evidence for separate mitochondrial and nuclear functions

.

Genetics

132

:

975

–

985

.

Reynolds

R J

,

Friedberg

E C

,

1981

Molecular mechanisms of pyrimidine dimer excision in Saccharomyces cerevisiae: incision of ultraviolet-irradiated deoxyribonucleicacid in vivo

.

J. Bacteriol.

146

:

692

–

704

.

Rong

L

,

Palladino

F

,

Aguilera

A

,

Klein

H

,

1991

The hypergene conversion hpr5-1 mutation of Saccharomyces cerevisiae is an allele of the SRS2/RADH gene

.

Genetics

127

:

75

–

85

.

Rothstein

R J

,

1983

One-step gene disruption in yeast

.

Methods Enzymol.

101

:

202

–

211

.

Schiestl

R H

,

Prakash

S

,

Prakash

L

,

1990

The SRS2 suppressor of rad6 mutations of Saccharomyces cerevisiae acts by channeling DNA lesions in to the RAD52 DNA repair pathway

.

Genetics

124

:

817

–

831

.

Sherman

F

,

Fink

G R

,

Hicks

J B

,

1981

Methods in Yeast Genetics

.

Cold Spring Harbor Laboratory Press

,

Cold Spring Harbor, NY

.

Siede

W

,

Eckardt

F

,

Brendel

M

,

1983

Analysis of mutagenic DNA repair in a thermoconditional mutant of Saccharomyces cerevisiae. I. Influence of cycloheximide on UV-irradiated rev2 ts cells

.

Mol. Gen. Genet.

190

:

406

–

416

.

Siede

W

,

Friedberg

A S

,

Friedberg

E C

,

1993

RAD9-dependent G1 arrest defines a second checkpoint for damaged DNA in the cell cycle of Saccharomyces cerevisiae

.

Proc. Natl. Acad. Sci. USA

90

:

7985

–

7989

.

Siede

W

,

Friedberg

A S

,

Dianova

I

,

Friedberg

E C

,

1994

Characterization of G1 checkpoint control in the yeast following exposure to DNA-damaging agents

.

Genetics

138

:

271

–

281

.

Sugimoto

K

,

Ando

S

,

Shimomura

T

,

Matsumoto

K

,

1997

Rfc5, a replication factor C component, is required for regulation of Rad53 protein kinase in the yeast checkpoint pathway

.

Mol. Cell. Biol.

17

:

5905

–

5914

.

Tishkoff

D X

,

Filosi

N

,

Gaida

G M

,

Kolodner

R D

,

1997

A novel mutation avoidance mechanism dependent on S. cerevisiae RAD27 is distinct from DNA mismatch repair

.

Cell

88

:

253

–

263

.

Tucker

J D

,

Auletta

A

,

Cimino

M C

,

Dearfield

K L

,

Jacobson-Kram

D

et al. ,

1993

Sister-chromatid exchange: second report of the Gene-Tox Program

.

Mutat. Res.

297

:

101

–

180

.

Vallen

E A

,

Cross

F R

,

1995

Mutations in RAD27 define a potential link between G1 cyclins and DNA replication

.

Mol. Cell. Biol.

15

:

4291

–

4302

.

Von Borstel

R C

,

Cain

K T

,

Steinberg

C M

,

1971

Inheritance of spontaneous mutability in yeast

.

Genetics

69

:

17

–

27

.

Weinert

T A

,

Hartwell

L H

,

1988

The RAD9 gene controls the cell cycle response to DNA damage in Saccharomyces cerevisiae

.

Science

241

:

317

–

322

.

Weinert

T A

,

Hartwell

L H

,

1990

Characterization of RAD9 of Saccharomyces cerevisiae and evidence that its function acts posttranslationally in cell cycle arrest after DNA damage

.

Mol. Cell. Biol.

10

:

6554

–

6564

.

Weinert

T A

,

Hartwell

L H

,

1993

Cell cycle arrest of cdc mutants and specificity of the RAD9 checkpoint

.

Genetics

134

:

63

–

80

.

Weinert

T A

,

Kiser

G L

,

Hartwell

L H

,

1994

Mitotic checkpoint genes in budding yeast and the dependence of mitosis on DNA replication and repair

.

Genes Dev.

8

:

652

–

665

.

Wilcox

D R

,

Prakash

L

,

1981

Incision and postincision steps of pyrimidine dimer removal in excision-defective mutants of Saccharomyces cerevisiae

.

J. Bacteriol.

148

:

618

–

623

.

Author notes

1

Present address: Massachusetts General Hospital, Medical Services, Boston, MA 02114.

© Genetics 1998