Abstract
The DNA polymerase δ (Pol3p/Cdc2p) allele pol3-t of Saccharomyces cerevisiae has previously been shown to increase the frequency of deletions between short repeats (several base pairs), between homeologous DNA sequences separated by long inverted repeats, and between distant short repeats, increasing the frequency of genomic deletions. We found that the pol3-t mutation increased intrachromosomal recombination events between direct DNA repeats up to 36-fold and interchromosomal recombination 14-fold. The hyperrecombination phenotype of pol3-t was partially dependent on the Rad52p function but much more so on Rad1p. However, in the double-mutant rad1Δ rad52Δ, the pol3-t mutation still increased spontaneous intrachromosomal recombination frequencies, suggesting that a Rad1p Rad52p-independent single-strand annealing pathway is involved. UV and γ-rays were less potent inducers of recombination in the pol3-t mutant, indicating that Pol3p is partly involved in DNA-damage-induced recombination. In contrast, while UV- and γ-ray-induced intrachromosomal recombination was almost completely abolished in the rad52 or the rad1 rad52 mutant, there was still good induction in those mutants in the pol3-t background, indicating channeling of lesions into the above-mentioned Rad1p Rad52p-independent pathway. Finally, a heterozygous pol3-t/POL3 mutant also showed an increased frequency of deletions and MMS sensitivity at the restrictive temperature, indicating that even a heterozygous polymerase δ mutation might increase the frequency of genetic instability.
RECOMBINATION between repeated DNA sequences can occur in meiosis and in mitosis (Petes and Hill 1988; Klein 1995). Mitotic recombination between DNA repeats on the same chromosome, called intrachromosomal recombination, can lead to deletion of sequences located between the repeats, to gene conversion events that retain the duplication, or to triplications (Klein 1995). Genome rearrangements associated with recombination between homologous sequences can cause genetic disease and cancer and they increase in frequency by exposure to cancer-causing chemicals (Bishop and Schiestl 2000, 2001). It is thus important to identify the genetic and environmental factors leading to an increased frequency of such rearrangements, as well as to study the interaction between these factors.
Homologous intrachromosomal recombination events between duplicated sequences resulting in deletions may occur by several different mechanisms, such as intrachromatid exchange, single-strand annealing (SSA), one-sided invasion, unequal sister chromatid exchange, or sister chromatid conversion (Schiestlet al. 1988; Haber 1992; Belmaaza and Chartrand 1994; Galli and Schiestl 1995). Studies were previously carried out on the mechanism of reversion of a duplication of a 400-bp internal fragment of the HIS3 gene separated by the LEU2 gene (Schiestlet al. 1988). Intrachromatid exchange occurs as reciprocal crossing over between the direct repeats, which leaves a single copy of the gene on the chromosome and on the excised DNA fragment bearing the second copy of the gene. Schiestl et al. investigated the contribution of this mechanism to the frequency of such intrachromosomal recombination events by placing an origin of replication onto the integrated plasmid to recover both reciprocal products of an intrachromatid crossing-over event (Schiestlet al. 1988). They found that only a minority of events (∼1%) could be explained by this mechanism. With a different system that forced amplification of the excised circle, Santos-Rosa and Aguilera (1994) found that <10% of the deletion events produced circles. These results indicate that the majority of deletion events do not happen by intrachromatid crossing over, but rather by a nonconservative mechanism. SSA is initiated by a DNA double-strand break (DSB) in the nonhomologous region between the repeats. DNA degradation of single strands from the exposed 5′ ends of the DSB leads to single-strand regions that can anneal once the degradation has proceeded to the repeated sequences. The 3′ tails are processed and nicks are ligated, giving rise to a deletion. Another mechanism yielding deletion events is one-sided invasion, which is initiated by a DSB in one of the duplicated homologous sequences followed by 5′–3′ degradation (Belmaaza and Chartrand 1994). Invasion of the 3′ single strand occurs in the homologous region, leading to D-loop formation and to DNA synthesis. Resolution occurs by continuation of 5′ degradation, single-strand nick formation, and DNA repair synthesis.
Intrachromosomal recombination leading to deletions can also be explained by recombination between sister chromatids as unequal sister chromatid exchange (SCEs) or sister chromatid conversion. Unequal SCEs give rise to a duplication of the disrupting sequence (Schiestlet al. 1988; Galli and Schiestl 1995). The contribution of SCE events was determined by assaying for reciprocal products (Schiestlet al. 1988). Only ∼4% of the recombination events gave such a triplication. This suggests that the majority of events are not due to unequal SCEs.
Intrachromatid exchange, SSA, and one-sided invasion can take place in any phase of the cell cycle, including G1. SCE and sister chromatid conversion events, on the other hand, require the presence of the sister chromatid and thus they can occur in the S-phase or in G2 but not in G1. Intrachromosomal deletion recombination events are induced by a site-specific DSB in G1 and G2 to the same extent. Moreover, DNA single-strand breaks induce intrachromosomal deletion events in dividing but not in cell-cycle-arrested cells (Galli and Schiestl 1998b). This suggests that DNA DSBs are involved and that SSA is the main mechanism by which intrachromosomal deletion events occur (Galli and Schiestl 1998b). Mutations in RAD1, RAD10, and RAD52 are involved in these intrachromosomal deletion events (Schiestl and Prakash 1988, 1990) and Rad1p has been shown on a molecular level to catalyze the excision of the nonhomologous DNA between the recombining duplicated alleles needed for the SSA pathway (Fishman-Lobell and Haber 1992; Ivanov and Haber 1995).
Several mutants with elevated spontaneous intrachromosomal recombination frequencies have been isolated in Saccharomyces cerevisiae (Aguilera and Klein 1988; Klein 1995). Among them, an allele of CDC2/POL3, which encodes the catalytic subunit of the DNA polymerase δ, increases deletion events but not gene conversions (Aguilera and Klein 1988). Polδp, together with Polαp and Polϵp, is an essential function and required for DNA replication. Polαp has a primase activity and is involved in initiation of both the leading- and the lagging-strand syntheses (Brooks and Dumas 1989). Both Polδp and Polϵp can extend the primers formed by Polαp (Burgers 1991; Podust and Hubscher 1993).
The pol3-t mutant allele, initially isolated as tex1 mutant because it increased the rate of excision of a bacterial transposon within the yeast LYS2 gene, also increases intrachromosomal deletion recombination between short repeats of several base pairs separated by long inverted repeats (Gordeninet al. 1992). The molecular analysis of the recombinants of the excision events of the transposon indicates that DNA replication slippage most likely is responsible for these deletion events (Tranet al. 1995; Gordenin and Resnick 1998). In agreement with this model, it has been shown that pol3 mutations increase the frequency of additions and/or deletions of units of microsatellites (defined as repeat units from 1 to 13 bp) as well as minisatellites (>15 bp; Tran et al. 1995, 1996, 1999; Kokoskaet al. 1998). Furthermore, the frequency of deletions between distant short repeats within the LYS2 or the CAN1 genes is also increased many fold (Tranet al. 1995; Kokoskaet al. 2000). Finally, it has been shown that the same mutator phenotype as observed in the pol3 mutations exists after repression of the POL3 gene, indicating that the mutator phenotype may be due to low levels of Pol3p rather than to any other faulty effect of the Pol3p mutant proteins.
Here we report the effect of the temperature-sensitive allele pol3-t on intrachromosomal deletion and interchromosomal recombination, reverse and forward mutation. Moreover, we studied the influence of Rad1p and Rad52p in the pol3-t background to characterize the genetic control of intrachromosomal recombination. Finally, to better understand the role of DNA polymerase δ on DNA-damage-induced recombination, we also studied the effects of Rad1p and Rad52p on UV-, γ-ray-, and methyl methanesulfonate (MMS)-induced intrachromosomal deletion recombination in the pol3-t background.
MATERIALS AND METHODS
Media, genetic, and molecular techniques: Complete media (YPAD), synthetic complete (SC), and drop-out (SD) media were prepared according to standard procedures (Kaiseret al. 1994). Magic Column (Promega, Madison, WI) was used for preparation of small-scale DNA. Other general molecular techniques were carried out according to Maniatis et al. (1989). Yeast transformation was performed using the procedure described in Gietz et al. (1992, 1995).
Yeast strains: The names and genotypes of the strains of S. cerevisiae used are listed in Table 1. Because pol3-t confers a temperature-sensitive phenotype, all pol3-t strains were grown at 25° (Gordeninet al. 1992). Strains TCY1 and TCY2 were constructed by transformation of strains POL-DM and pol3-t-DM with plasmid pRS6, which contains an internal fragment of his3 and a LEU2 marker (Schiestlet al. 1988). This generates duplication within the HIS3 gene, resulting in two incomplete his3 alleles (see below).
Strains TCY3 and TCY4, carrying a deletion from position +40 to +3211 of RAD1, were constructed by two-step gene replacement using the EcoRI-SalI fragment of plasmid pR1.6 (kindly provided by Louise Prakash; Saparbaevet al. 1996) and subsequent 5-fluoroorotic acid (5-FOA) selection (Boekeet al. 1984). Strains AGY30, AGY31, AGY34, and AGY35 were constructed by introducing the pol3-t mutation into strains RSY6, YR1-16, and Y433. This was done by transformation of the cells with plasmid p171 (a gift from Mike Resnick, National Institute of Environmental Health Sciences, Research Triangle Park, NC), which contains a 2.2-kb EcoRV-HindIII fragment containing the pol3-t allele (Kokoskaet al. 1998). The cells were transformed with HpaI-linearized p171. Temperature-sensitive Ura+ colonies that contained the full-length pol3-t allele and a truncated POL3 allele flanking the URA3 gene were isolated. Ura– temperature-sensitive strains carrying just the pol3-t allele were selected after selection on medium containing 5-FOA (Kokoskaet al. 1998). Strains AGY32 and AGY33 carrying the rad52-9 deletion (henceforth called rad52Δ) were constructed by digestion of plasmid pSM22 (from David Schield and R. Mortimer via Louise Prakash) with BamHI and transformation of yeast cells with the BamHI fragment in which the BglII-ClaI fragment in the open reading frame of the RAD52 gene had been replaced by a BamHI-ClaI fragment containing the URA3 gene (Schiestl and Prakash 1990).
Diploid strains AGY36 and AGY37, isogenic to RS112, were constructed by mating AGY30 with AGY35 and RSY6 with AGY35, respectively.
Recombination assays: All strains used carry the same intrachromosomal recombination substrate as strain RSY6 (Schiestlet al. 1988). This substrate consists of two his3 alleles, one with a deletion at the 3′ end and the other with a deletion at the 5′ end, which share 400 bp of homology. These two alleles are separated by the LEU2 marker and by the plasmid DNA sequence. An intrachromosomal recombination event leads to HIS3 reversion and loss of LEU2 (Schiestlet al. 1988). These two copies readily undergo intrachromosomal recombination, resulting in wild-type HIS3 at a frequency of ∼10–4 (Schiestlet al. 1988). Diploid strains RS112 and AGY36 are also heteroallelic for ade2-40 and ade2-101. An interchromosomal gene-conversion event produces ADE2 reversions.
To determine the frequency of spontaneous intrachromosomal recombination, single colonies were inoculated into 5 ml of SC-LEU and incubated at 25° or 30° for 17 hr. Thereafter, cultures were washed twice and counted and appropriate numbers were plated onto SC and SC-HIS plates to determine the surviving fraction and the frequency of intrachromosomal recombination, respectively. Single colonies of the diploid strains RS112 and AGY36 were incubated as above and in addition plated onto SC-ADE plates to determine the frequency of interchromosomal gene conversion. Plates were incubated at 25° for 4 days and colonies were counted thereafter. All HIS3 and ADE2 recombinants were checked for the presence of the pol3-t allele by replica plating and incubation at 37°.
Intrachromosomal recombination was also measured following UV, γ-rays, and MMS exposure. For UV exposure, single colonies were inoculated into SC-LEU at 25° for 17 hr. Thereafter, cells were washed and resuspended in fresh SC-LEU for 4 hr at 30°. Aliquots of 10 ml containing 3 × 107 cells/ml were irradiated in distilled water using a UV source at the dose rate of 3.5 erg/m2/sec. The same number of cells were exposed to γ-rays using a 60Co γ-ray source at 9.1 cGy/sec (Galli and Schiestl 1995, 1998b). Following irradiation, cells were plated as described above. For MMS exposure, single colonies were inoculated into SC-LEU at 25° for 17 hr. Thereafter, cells were washed, resuspended in 5 ml of fresh SC-LEU at the concentration of 3 × 106 cells/ml, and exposed to MMS for 4 hr at 30°. Then cells were washed, counted, and plated as described.
Reverse and forward mutation assay: To measure the spontaneous frequency of reverse mutations at ilv1-92 and arg4-3, single colonies of RSY6 and AGY30 were inoculated into 5 ml YPAD and incubated for 17 hr at 25° or 30°. Then cells were washed and counted and appropriate numbers were plated onto SC, SC-ILV, and SC-ARG to score for the surviving fraction and mutants. Plates were incubated at 25° until colonies were formed.
The spontaneous frequency of forward mutation was determined as follows: single colonies of RSY6 (ARG4) and AGY30 (ARG4) were inoculated in 5 ml YPAD and incubated for 17 hr at 25° or 30°. Then cells were washed and counted and appropriate numbers were plated onto SC and SC-ARG + CAN (60 mg/liter) to score for the surviving fraction and mutants. Plates were incubated at 25° until colonies formed.
Determination of the effect of cell division on the recombination phenotype in the pol3-t mutant: We tested the effect of cell division on the recombination phenotype of pol3-t after growth at 25° and 30°. Single colonies of AGY30 and RSY6 were grown in SC-LEU at 25° for 20 hr. Cells were washed and inoculated for 5 hr in SC-URA to achieve cell-cycle arrest at G0/G1 since they carry the ura3-52 allele (Galli and Schiestl 1995). Cell-cycle arrest was checked by counting unbudded cells under the microscope as previously described (Galli and Schiestl 1995). A total of 250–300 cells were counted per tube and 96.1 ± 0.7% were unbudded. Thereafter, cell cultures were divided into two aliquots; one aliquot was kept in SC-URA medium at 25° and the other one was incubated at 30° for 24 hr. Intrachromosomal recombination was measured at the 0 time point and after 24 hr of incubation.
Data comparison and statistical evaluation: The data were compared either as fold induction compared to the control or as “change in average frequency,” which indicates the number of recombination events after exposure to a certain dose of a genotoxin after subtraction of the spontaneous frequency (Kadyk and Hartwell 1992, 1993; Galli and Schiestl 1998a; Paulovichet al. 1998). Results were statistically analyzed using the Student's t-test.
RESULTS
Effect of pol3-t on spontaneous mitotic recombination: To investigate effects of pol3-t on mitotic recombination we constructed the haploid strains TCY1, TCY2, and AGY30 and the diploid strain AGY36 (Table 1). All these strains contain an intrachromosomal recombination substrate that resulted from integration of plasmid pRS6 at the HIS3 locus (see materials and methods; Schiestlet al. 1988). Intrachromosomal recombination between the two his3 alleles, which share 400 bp of homology, leads to HIS3 reversion and loss of LEU2 (Schiestlet al. 1988). The diploid strains RS112 and AGY36 are heteroallelic for ade2 and can also be used to measure interchromosomal recombination events (Table 1). The pol3-t mutation confers a temperature-sensitive phenotype and growth arrest at 37°; thus, we studied effects of pol3-t mutation on mitotic recombination after growth at 25° and 30° (Gordeninet al. 1992). Single colonies of TCY1, TCY2, RSY6, AGY30, RS112, and AGY36 were incubated in SC-LEU for 17 hr at 25° and 30°. During this incubation period, TCY1, RSY6, and RS112 underwent four to five cell divisions at both temperatures. TCY2, AGY30, and AGY36 underwent three to four cell divisions at 25° and two to four cell divisions at 30°. Appropriate aliquots were plated and incubated. HIS3 leu2 colonies revealed deletion recombination frequencies. At 25°, pol3-t increased intrachromosomal recombination 8-fold in the diploid strain AGY36, 15- and 4-fold, respectively, in the haploid strains AGY30 and TCY2, and, at 30°, 36-fold in AGY36, 22-fold in AGY30, and 18-fold in TCY2 (Table 2). The pol3-t mutation did not significantly increase the frequency of interchromosomal recombination at 25° in the diploid strain AGY36, whereas it caused a significant 14-fold increase at 30°.
S. cerevisiae strains
Effect of the pol3-t mutation on spontaneous mutation frequencies: Another temperature-sensitive mutant of DNA polymerase δ gene CDC2, named hpr6, has been shown to have both hyperrecombination and mutator phenotypes (Aguilera and Klein 1988). In addition, it has been shown that other pol3 alleles increase the mutation frequency at different loci (Morrison and Sugino 1994; Giotet al. 1997; Kokoskaet al. 2000). Therefore, we tested the effects of pol3-t on spontaneous mutation frequencies in our strain backgrounds. As shown in Table 3, the frequency of ILV1 reverse mutation increased in AGY30 6-fold at 25° and 8-fold at 30°. At 25°, pol3-t did not affect the frequency of ARG4 revertants, whereas it caused a small but significant 3-fold increase in reversion frequency at 30° (Table 3). The frequency of forward mutation, determined as frequency of canavanine-resistant (can1R) mutants, increased 18-fold at 25° and 32-fold at 30° in the pol3-t strain (Table 3). Thus, pol3-t, like other pol3 mutants, causes both hyperrecombination and mutator phenotypes.
Since pol3-t had a more pronounced effect on intrachromosomal recombination than on interchromosomal recombination, we decided to further focus our study on intrachromosomal recombination.
Dependence of the pol3-t hyperrecombination phenotype on DNA replication: In yeast, mRNA transcript levels of CDC2/POL3 increase at the boundary of the G1/S phase of the cell cycle (Wang 1991) and return to a low level during or after the S-phase (Campbell and Newlon 1991).
Effect of pol3-t on spontaneous intrachromosomal and interchromosomal recombination frequencies
pol3-t strains showed a pronounced hyperrecombination phenotype following growth at the semipermissive temperature of 30°. We tested the effect of cell division on the hyperrecombination phenotype of pol3-t after growth at 25° and 30° (see materials and methods). Intrachromosomal recombination was measured at the 0 time point and after 24 hr of incubation. At time 0, the frequency of intrachromosomal recombination was 2.34 ± 0.79 × 10–4 in the POL3 strain and 24.5 ± 6.6 × 10–4 in the pol3-t strain. After 24 hr at 25° in G0/G1, the frequency was 2.3 ± 0.85 × 10–4 in the POL3 strain and 16.51 ± 8.27 × 10–4 in the pol3-t strain, and after 24 hr in G0/G1 at 30°, the frequency of intrachromosomal recombination was 1.79 ± 0.77 × 10–4 in the POL3 strain and 21.69 ± 2.8 × 10–4 in the pol3-t strain. Thus, intrachromosomal recombination frequencies did not change after 24 hr of postincubation at 30° as compared to 25° in the absence of cell divisions. In contrast, recombination frequencies of pol3-t cells grown in parallel at 30° for the same amount of time were at least 2.5-fold higher than those at 25°, similar to the data in Table 2. Thus, DNA replication is most likely required for the development of the pol3-t hyperrecombination phenotype.
Effect of pol3-t on reverse and forward mutation
Effect of mutations in rad1 and rad52 on the pol3-t hyperrecombination phenotype: The excision repair gene RAD1 is involved in intrachromosomal recombination (Klein 1988; Schiestl and Prakash 1988) and affects DNA DSB-induced recombination (Ivanov and Haber 1995). To study the effect of RAD1 on the pol3-t hyperrecombination phenotype, we constructed strain AGY31 containing the pol3-t mutation and the rad1 deletion. Single colonies of AGY30 and YR1-16 were incubated at 25° or 30°. Cells were counted and plated as described. In strain YR1-16 POL3 wild type, the rad1 deletion decreased intrachromosomal recombination frequencies 6- to 11-fold (Table 4). In strain AGY31 rad1Δ, the pol3-t mutation did not significantly increase intrachromosomal recombination at 25°, while at 30° intrachromosomal recombination increased 6.5-fold (Table 4). Thus, the rad1 mutation decreased the pol3-t-mediated hyperrecombination phenotype 100-fold at 25° and 36-fold at 30°. Similar results were obtained with TCY1, TCY2, TCY3, and TCY4, another set of wild-type, pol3-t, rad1, and rad1 pol3-t strains (data not shown). Thus, the pol3-t hyperrecombination phenotype in most part is dependent on Rad1p.
Intrachromosomal recombination is also dependent on the RAD52 gene function (Schiestl and Prakash 1988; Klein 1995; Ivanovet al. 1996). The deletion of rad52 in the POL3 strain decreased the intrachromosomal recombination frequency 11- to 15-fold (Table 4). In the rad52Δ pol3-t strain AGY32, intrachromosomal recombination frequencies are 11- and 60-fold higher than those in the rad52 POL3 strain YR52-2 at 25° and 30° (Table 4). Compared to AGY30, the RAD52 wild-type pol3-t strain, the rad52Δ mutation decreased the intrachromosomal recombination frequency 9- and 5-fold at 25° and 30° (Table 4). Thus, the pol3-t hyperrecombination phenotype is partially dependent on Rad52p. In summary, the dependence of the pol3-t hyperrecombination phenotype on Rad1p is much greater than that on Rad52p.
Effect of RAD1 and RAD52 on the hyperrecombination phenotype of pol3-t
The simultaneous deletion of the rad1 and rad52 genes led to a synergistic decrease in intrachromosomal recombination >100-fold in the YR1-18 POL3 strain (Table 4), which has been seen before (Schiestl and Prakash 1988). In the AGY33 strain, which is rad1Δ rad52Δ pol3-t, intrachromosomal recombination frequencies are 6.5- and 57-fold higher than those in YR1-18 at 25° and 30° (Table 4). Thus, the rad1 mutation, together with the rad52 mutation, decreases the pol3-t-mediated hyperrecombination phenotype 246-fold and 45-fold at 25° and 30° (Table 4). However, this is due mostly to the rad1 mutations since the rad1Δ rad52Δ pol3-t mutant showed a frequency similar to that of the rad1Δ pol3-t mutant. In summary, the pol3-t mutation still elevates the recombination frequency in the rad1Δ rad52Δ mutant background.
Effect of pol3-t on UV-induced intrachromosomal recombination: Some alleles of POL3 are deficient in DNA-damage-induced mutagenesis and interchromosomal recombination (Giotet al. 1997). Thus, we determined the effects of pol3-t on intrachromosomal recombination induced by UV, γ-rays, and MMS and the effects of mutations rad1 and rad52 and of the double mutation rad1 rad52 on DNA-damage-induced recombination events. Single colonies of these strains were incubated first at 25° and then for 4 hr at 30° and irradiated with UV and γ-rays or exposed to MMS as described in materials and methods.
UV irradiation induced a significant increase in intrachromosomal recombination at doses of 100 and 500 J/m2 in both the wild-type (RSY6) and the pol3-t (AGY30) strains (Table 5). At 500 J/m2, intrachromosomal recombination increased 4.5-fold (P < 0.01) in the wild type and 1.6-fold (P < 0.05) in the pol3-t strain (Table 5). Since the spontaneous frequency differs greatly between the two strains, it seems justified to also base the comparison on the average number of additional recombinants due to the radiation effect. At 500 J/m2, there was an average of 22 added events in the wild type compared to an average of 8 added events in the pol3-t strain. This indicates that the pol3-t strain shows a mild deficiency in UV-induced intrachromosomal recombination.
The pol3-t mutation did not lower the survival after UV exposure of the RAD wild type and the rad1Δ, rad52Δ, and rad1Δ rad52Δ strains, suggesting that in these mutants UV-induced DNA-damage repair in the pol3-t mutant is as efficient as in the wild type (Table 5). If anything, the pol3-t mutant by itself, as well as in any of the double- and triple-mutant combinations, is slightly more UV resistant compared to the POL3 genotype. The rad1 mutant shows a dose-dependent, significant UV induction of intrachromosomal recombination starting at doses as low as 1 J/m2, regardless of the POL3 genotype. This dose is 100-fold lower than the dose resulting in significant induction in the RAD wild type. The pol3-t mutant still shows some minor defect in induced recombination since the dose of 20 J/m2 resulted in an average of 9 added recombination events in the rad1 mutant vs. an average of 3.7 added events in the rad1 pol3-t mutant (Table 5).
The rad52 mutant as well as the rad1 rad52 mutant were both completely defective in UV-induced intrachromosomal recombination. Both the RAD1 and the RAD52 pathways are involved in spontaneous recombination and the double mutant showed a synergistic decrease in spontaneous recombination, as previously found (Schiestl and Prakash 1988). The present results indicate that only the Rad52p, but not the Rad1p, pathway is involved in UV-induced recombination. In the rad52, as well as the rad1 rad52, background the pol3-t mutation resulted in a UV-induced recombination increase of about twofold.
Effect of pol3-t on UV-induced intrachromosomal recombination in RAD+, rad1Δ, rad52Δ, and rad1Δ rad52Δ strains
Effect of pol3-t on γ-ray-induced intrachromosomal recombination: Irradiation with γ-rays induced significant increases in intrachromosomal recombination at all doses used in both wild-type and pol3-t strains (Table 6). At 1000 Gy, intrachromosomal recombination increased 16-fold (P < 0.001) in the wild type and 3.7-fold (P < 0.05) in the pol3-t mutant (Table 6). At that dose there was an average of 37 added events in the wild type and an average of 23 added events in the pol3-t mutant, which is somewhat less in the mutant but rather similar.
Effect of pol3-t on γ-ray-induced intrachromosomal recombination in RAD+, rad1Δ, rad52Δ, and rad1Δ rad52Δ strains
In the rad1Δ strain, γ-ray exposure elevated recombination frequencies in a dose-dependent manner without any effect of pol3-t, while exposure resulted in a very moderate increase only at the highest dose in the rad52Δ and no significant induction in the rad1Δ rad52Δ strain (Table 6). Interestingly, as for UV-induced recombination in both backgrounds, in the pol3-t mutant there was significant induction at higher doses. In the rad52Δ pol3-t mutant, there was an average of 2.3 added events at a dose of 50 Gy vs. an average of 0.2 added events in the rad52 mutant. Similarly, in the rad1Δ rad52Δ pol3-t mutant at the same dose, there was an average of 0.32 added events vs. an average of 0.022 added events in the rad1Δ rad52Δ strain.
Effect of pol3-t on MMS-induced intrachromosomal recombination: At higher doses of MMS the pol3-t mutation in the RAD wild type, the rad1Δ, and, to a lesser extent, the rad52Δ backgrounds were much more sensitive, which is in agreement with Blank et al. (1994; Table 7). Since the different rad mutant backgrounds caused MMS sensitivity to different degrees, one way to compare the pol3-t effect is to compare sensitivities at a dose giving ∼40–50% viability in the POL3 wild-type strain. In the RAD wild-type strain, exposure to 500 μg/ml MMS resulted in 48% viability with a 60-fold lower viability in the pol3-t strain. In comparison, 200 μg/ml MMS exposure in the rad1Δ strain had a viability of 36% compared to a 6-fold lower viability in the rad1Δ pol3-t strain. At a dose of 50 μg/ml MMS, the rad52Δ strain had a viability of 54% compared to a 7-fold lower viability in the rad52Δ pol3-t strain. Finally, at a dose of 1 μg/ml MMS, the rad1Δ rad52Δ strain had a viability of 39% compared to an ∼3-fold higher viability in the rad1Δ rad52Δ pol3-t strain. Thus, the pol3-t-mediated MMS sensitivity was diminished in both the rad1 and the rad52 background and absent in the rad1 rad52 background.
Effect of pol3-t on MMS-induced intrachromosomal recombination in RAD+, rad1Δ, rad52Δ, and rad1Δ rad52Δ strains
Effect of RAD1, RAD52 deletion on the pol3-t temperature-sensitive phenotype
At the lowest dose of MMS (10 μg/ml), intrachromosomal recombination increased 2-fold (P < 0.01) in the wild type while no increase was seen in the pol3-t mutant (Table 7). At equitoxic doses at the same survival level of 40–50%, at a dose of 500 μg/ml in the wild type and of 200 μg/ml in the pol3-t mutant, MMS increased intrachromosomal recombination 24-fold (P < 0.001) in the wild type and 3.5-fold (nonsignificant) in the pol3-t mutant. At these doses, there was an average of 98 added events in the wild type and an average of 43 added events in the pol3-t mutant. This would allow the conclusion that for all three DNA-damaging agents, the results indicate that the level of DNA-damage-induced recombination is lower in the pol3-t mutant than in the RAD wild type, indicating that Pol3p is partially responsible for DNA-damage-induced intrachromosomal recombination. If, however, the same dose rather than equitoxic doses of MMS is evaluated, there is an equal level of induction of intrachromosomal recombination in the pol3-t mutant.
In the rad1Δ, rad52Δ, and rad1Δ rad52Δ double-mutant background, MMS induced recombination at all doses regardless of the POL3 status (Table 7). However, again as for UV- and γ-ray-induced recombination, in both backgrounds the pol3-t mutant showed higher inducibility. At a dose of 100 μg/ml MMS, there was an average of 27.7 added events in the rad52Δ pol3-t strain compared to an average of 1.1 added events in the rad52Δ strain. At the same dose in the rad1Δ rad52Δ pol3-t strain, there were ∼5 induced events compared to 0.24 induced events for the rad1Δ rad52Δ strain (Table 7).
Effect of rad1 and rad52 mutations on the temperature-sensitive phenotype of pol3-t: Since the rad1 and rad52 mutations partially reduced the hyperrecombination phenotype of the pol3-t mutant, we determined the effect of mutations in these DNA repair pathways on the temperature-sensitive phenotype of pol3-t. The pol3-t mutant and all combinations of double and triple mutants were incubated at the restrictive temperature, and viability was determined after different time points up to 8 hr. After 1 hr, the cells were already arrested and did not grow further. The control POL3 rad mutants were incubated in the same way but kept dividing rapidly at 37°. Thus, the viability must have been high in these strains. After 8 hr, the pol3-t mutant had 0.4% viable cells. At all time points the survival of the double and triple mutants was lower than that of the pol3-t single mutant and at almost all time points starting at the 2-hr point this difference was significant (Table 8). This indicates that both the Rad1p and the Rad52p pathway are partially involved in the repair of lethal DNA lesions in the pol3-t strain at the restrictive temperature.
Effect of pol3-t heterozygosity on recombination and MMS sensitivity: A decrease in the expression of POL3 under the GAL1 promoter is sufficient to cause a mutator phenotype and MMS sensitivity (Kokoskaet al. 2000). Thus, we determined whether a pol3-t/POL3 heterozygous strain showed any effect on recombination efficiency or on MMS sensitivity at the restrictive temperature of 37°. There was a significant difference for both intrachromosomal and interchromosomal recombination at the restrictive temperature in the pol3-t/POL3 heterozygous mutant (Table 9). The heterozygous strain was also more MMS sensitive at every dose and, starting at a dose of 0.2 mg/ml (Table 9), this increase was significant. Thus, the pol3-t/POL3 heterozygous mutant at the restrictive temperature showed an increase in recombination frequency as well as an MMS-sensitive phenotype.
Effect of pol3-t/POL3 heterozygosity on recombination and MMS sensitivity
DISCUSSION
In this study, we found that the pol3-t allele of the POL3/CDC2 gene of S. cerevisiae, which encodes the catalytic subunit of the DNA polymerase δ, increased intra- and interchromosomal recombination in a diploid and intrachromosomal recombination in two haploid strains. Previous studies reported that mutations in POL3/CDC2 genes increased intrachromosomal recombination between homologous and homeologous DNA sequences (Aguilera and Klein 1988; Tranet al. 1997) and increased interchromosomal recombination (Hartwell and Smith 1985; Giotet al. 1997). The elevated frequency of recombination in the pol3-t mutant could be mechanistically similar to the events in the POL3 wild type but just occur more frequently. Thus, we sought to further characterize the pol3-t-mediated hyperrecombination phenotype. The present study adds to the previously published findings by the characterization of effects of rad1 and rad52 mutations on the pol3-t-mediated hyperrecombination phenotype in isogenic backgrounds as well as the effects of exposure to UV, ionizing radiation, and MMS on deletion recombination in the different DNA repair mutants. In addition, we show that the pol3-t-mediated hyperrecombination phenotype is dependent on DNA replication and that even the pol3-t/POL3 heterozygous mutation increased recombination and MMS sensitivity.
Moreover, we found that pol3-t increased both reverse and forward mutation frequencies. Most interestingly, the pol3-01 mutant has abnormal cell-cycle progression due to activation of the S-phase checkpoint, and inactivation of the S-phase checkpoint suppressed the cell-cycle progression defect as well as the mutator phenotype (Dattaet al. 2000). This indicates that activation of the checkpoint might have resulted in the accumulation of the mutations.
The pol3-t-mediated hyperrecombination phenotype requires DNA replication: To determine whether cell division and/or DNA replication affect the hyperrecombination phenotype, we monitored intrachromosomal recombination during a prolonged G1 arrest. The frequency of intrachromosomal recombination did not change during the cell-cycle arrest and did not increase in G1-arrested cells at 30°. Thus, cell division or DNA replication is necessary to increase recombination in pol3-t strains.
Intrachromosomal recombination events leading to deletions are due mainly to DNA DSBs (Galli and Schiestl 1995, 1998b). Single-strand breaks in the region between the HIS3 duplication or exposure to alkylating agents or UV light did not increase deletion recombination unless DNA replication occurred (Galli and Schiestl 1998b, 1999). Thus, single-strand breaks may be converted into DNA DSBs by DNA replication on a single-strand or chemically damaged DNA template (Kaufmann and Paules 1996; Galli and Schiestl 1998b). These DSBs could induce SSA or one-sided invasion.
The hyperrecombination phenotype in pol3-t strains depends partially on Rad52p but much more so on Rad1p: Intrachromosomal recombination events leading to deletions between repeated sequences can occur by several mechanisms: by recombination between the two repeats within one chromatid as intrachromatid exchange; by SSA; by one-sided invasion events; or, alternatively, by recombination between sister chromatids as unequal sister chromatid exchange or sister chromatid conversion (Schiestlet al. 1988; Haber 1992; Belmaaza and Chartrand 1994; Galli and Schiestl 1995, 1998b; Klein 1995). Previous studies on cell-cycle-arrested cells suggested that SSA and/or one-sided invasion are the preferential mechanisms by which such deletions occur (Galli and Schiestl 1998b).
Deletion of RAD1 and RAD52 greatly reduces the frequencies of intrachromosomal deletion events between repeats (Schiestl and Prakash 1988; Thomas and Rothstein 1989; Liefshitzet al. 1995; Saparbaevet al. 1996). Different types of intrachromosomal recombination events are controlled by Rad1p and Rad52p. Deletion events occurring by SSA events are dependent on the Rad1p function (Fishman-Lobell and Haber 1992; Prado and Aguilera 1995; Paques and Haber 1999) whereas conversion events or deletion events occurring by one-sided invasion are probably Rad52p dependent (Klein 1988; Schiestl and Prakash 1988; Prado and Aguilera 1995).
SSA, one of the mechanisms for the deletion recombination events, requires Rad1p and Rad10p if the distance between interacting repeats is >60 bp (Paques and Haber 1999). It has been argued that deletions <60 bp may be due to polymerase slippage (Kokoskaet al. 2000). It seems, however, inconceivable that slippage could occur over a distance of 6 kb in the absence of inverted repeats as would be required for the deletions in our construct. Thus, it is more likely that the events happen by SSA or by one-sided invasion. DNA DSBs could be generated by replication of a single-strand break or single-strand interruption that would be expected to be more prevalent or longer lasting in the pol3-t mutant. Such DSBs could initiate Rad1p-dependent SSA events. Since most pol3-t-induced recombination events are Rad1p mediated, this pathway could account for the majority of the events. On the other hand, long stretches of single-strand DNA on the lagging-strand template in the pol3-t mutant (Gordeninet al. 1992; Kokoskaet al. 1998) could potentially invade the second copy of the HIS3 duplication in our recombination substrate, leading to Rad52p-dependent, one-sided invasion-like events.
The low level of recombination went up significantly in the rad52 mutant in the presence of the pol3-t mutation. In addition, ionizing radiation induced recombination to much higher levels in the pol3-t rad52 double mutant than in the rad52 single mutant. This indicates that pol3-t channels lesions into a Rad52p-independent pathway, like the Rad1p pathway.
Most of the pol3-t-induced recombination events were dependent on Rad1p or Rad52p; however, the pol3-t mutation still increased the frequency in the absence of Rad1p and Rad52p. Other hyperrecombination phenotypes differ in their dependence on the Rad1p- and/or Rad52p-mediated pathways. A mutation in hpr1 increases recombination between DNA repeats up to 2000-fold (Aguilera and Klein 1988; Santos-Rosa and Aguilera 1994). Since hpr1 mutants also show 100-fold elevated frequencies of chromosome loss, it has been suggested that the hyperrecombination phenotype is due to DNA breaks (Santos-Rosa and Aguilera 1994). This increased recombination frequency is completely abolished in a rad1 rad52 double-mutant background (Santos-Rosa and Aguilera 1994), which makes it different from the pol3-t effect. DNA transcription also induces recombination (Voelkel-Meimanet al. 1987; Thomas and Rothstein 1989). Both systems of transcription stimulated recombination. In the GAL10 assay as well as in the HOT1 assay, most events are Rad52p dependent and fewer events Rad1p dependent (Thomas and Rothstein 1989; Zehfuset al. 1990), which makes it different from the pol3-t genetic control. In both assays, a small proportion of the transcription-induced recombination events are Rad1p Rad52p independent (Thomas and Rothstein 1989; Zehfuset al. 1990). Among mutations isolated in a screen for an increase in recombination in a rad1Δ rad52Δ double-deletion mutant strain, the rfa1-D228Y mutant has been found to stimulate intrachromosomal deletion events between repeats up to the wild-type level (Smith and Rothstein 1995). The hyperrecombination phenotype of the rfa mutation is much more dependent on Rad1p than on Rad52p and, out of several possible recombination events between repeats, DNA deletions display the greatest stimulation in the rad1 rad52 double-mutant background (Smith and Rothstein 1999). Thus, the authors proposed that the rfa-mediated hyperrecombination phenotype is most likely caused by a Rad1p Rad52p-independent SSA mechanism (Smith and Rothstein 1999). In a similar way, the weak hyperrecombination phenotype we observed in the rad1Δ rad52Δ pol3-t strain may occur by a Rad1p Rad52p-independent SSA mechanism.
Genetic control of DNA-damage-induced intrachromosomal recombination in strains of different POL3 status: The pol3-t mutant was slightly more UV resistant in all genotypes compared to the POL3 wild type. This might be due to the somewhat longer time available to repair the lesions by excision repair in the pol3-t mutant since the cells grew at 30° for 4 hr prior to UV exposure. There was no such difference in survival for γ-rays or MMS. The pol3-t mutant was MMS sensitive as previously reported for other pol3 mutants (Blanket al. 1994). This MMS sensitivity was diminished in both the rad1 and the rad52 background and absent in the rad1 rad52 double mutant, which may indicate that Pol3p may be important for both the excision repair and the recombination repair pathway of MMS repair.
The pol3-t strain was partially defective in UV- and γ-ray-induced intrachromosomal recombination in agreement with findings by others (Fabreet al. 1991; Giotet al. 1997). DNA polymerase δ is required for both DNA replication and base excision repair (Budd and Campbell 1993, 1995; Blanket al. 1994; Morrison and Sugino 1994). Recently, a new role of the DNA polymerase δ in the DNA DSB repair and DSB-induced mitotic gene conversion has been reported (Holmes and Haber 1999). According to all models of recombination (Orr-Weaver and Szostak 1985), DNA replication is involved in processes such as extension of strand displacement and repair of gaps.
DNA-damage-induced intrachromosomal deletion recombination events are under different genetic control than spontaneous events, suggesting a difference in mechanism. UV and γ-rays induced recombination in the rad1Δ strain, but not at all or very little in the rad52Δ and the rad1Δ rad52Δ strains. This demonstrated that UV- and γ-ray-induced intrachromosomal recombination required Rad52p but not Rad1p in a POL3 background, whereas spontaneous recombination is dependent on both Rad1p and Rad52p functions (Schiestl and Prakash 1988; Thomas and Rothstein 1989; Liefshitzet al. 1995; Saparbaevet al. 1996). Thus, it is possible that different pathways of recombination might be preferred in spontaneous vs. DNA-damage-induced recombination. Interestingly, in the pol3-t background, UV- and γ-ray-induced intrachromosomal recombination was Rad52p independent. Both pathways of hyperrecombination potentially operable in the pol3-t strain, involving DSB formation on a single-strand break or gap template as well as invasion of homologous DNA by long stretches of single-strand DNA on the lagging-strand template, could be more prevalent after additional DNA damage. As much of the pol3-t hyperrecombination pathway was independent of Rad52p so was the UV- or γ-ray-induced recombination in the pol3-t mutant.
Pol3-t heterozygosity results in hyperrecombination and MMS-sensitive phenotypes: Our data indicate that pol3-t/POL3 heterozygosity significantly increased the recombination frequency in both systems as well as the MMS sensitivity at the restrictive temperature. This could be due to the fact that the pol3-t allele may have some dominant effect, such as binding to a multi-enzyme complex as an inactive component, or that just a lower level of Pol3p leads to the recombinagenic effect. Since a lower level of Pol3p obtained by repressing the gene under the GAL1 promoter resulted in a mutator as well as in an MMS-sensitive phenotype (Kokoskaet al. 2000), it is likely that a lower amount of Pol3p in the heterozygous mutant at the restrictive temperature is responsible for the effect in our experiment. In agreement with our finding, it has also been shown that the homozygous as well as the heterozygous pol3-01 mutation is inviable in combination with a mutation in RAD27 (Garyet al. 1999). Our results indicate that a mutation in DNA polymerase δ even in a heterozygous combination might increase the frequency of genetic instability, which might be a risk factor for cancer.
Acknowledgments
This research was supported by Research Career Development award no. ES00299 from the National Institutes of Health to R.H.S.
Footnotes
-
Communicating editor: P. J. Pukkila
- Received June 12, 2002.
- Accepted January 15, 2003.
- Copyright © 2003 by the Genetics Society of America