A Novel Allele of RAD52 That Causes Severe DNA Repair and Recombination Deficiencies Only in the Absence of RAD51 or RAD59

Corresponding author: Lorraine S. Symington, Department of Microbiology and Institute of Cancer Research, Columbia University, 701 W. 168th St., New York, NY 10032., lss5{at}columbia.edu (E-mail)

Communicating editor: M. LICHTEN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

With the use of an intrachromosomal inverted repeat as a recombination reporter, we have shown that mitotic recombination is dependent on the RAD52 gene, but reduced only fivefold by mutation of RAD51. RAD59, a component of the RAD51-independent pathway, was identified previously by screening for mutations that reduced inverted-repeat recombination in a rad51 strain. Here we describe a rad52 mutation, rad52R70K, that also reduced recombination synergistically in a rad51 background. The phenotype of the rad52R70K strain, which includes weak -ray sensitivity, a fourfold reduction in the rate of inverted-repeat recombination, elevated allelic recombination, sporulation proficiency, and a reduction in the efficiency of mating-type switching and single-strand annealing, was similar to that observed for deletion of the RAD59 gene. However, rad52R70K rad59 double mutants showed synergistic defects in ionizing radiation resistance, sporulation, and mating-type switching. These results suggest that Rad52 and Rad59 have partially overlapping functions and that Rad59 can substitute for this function of Rad52 in a RAD51 rad52R70K strain.

HOMOLOGOUS recombination in Saccharomyces cerevisiae requires genes of the RAD52 epistasis group, including RAD50–59, MRE11, and XRS2 (GAME and MORTIMER 1974 ; IVANOV et al. 1992 ; AJIMURA et al. 1993 ; BAI and SYMINGTON 1996 ). In general, these genes are required for mitotic recombination, meiosis, and resistance to DNA-damaging agents that make double-strand breaks, such as ionizing radiation and radiomimetic chemicals. However, the severity of DNA repair and recombination defects is quite variable between different mutants within the rad52 group. The rad52 null mutation results in the most severe phenotype, as measured by various recombination and repair assays (RESNICK 1969 ; HO 1975 ; RESNICK and MARTIN 1976 ; GAME et al. 1980 ; PRAKASH et al. 1980 ; MORTIMER et al. 1981 ). The rate of spontaneous mitotic recombination between inverted repeats of the ade2 gene is reduced by 3000-fold in a rad52 null mutant strain. On the other hand, individual mutations in any other genes of the RAD52 epistasis group reduce recombination by no more than 30-fold (RATTRAY and SYMINGTON 1994 , RATTRAY and SYMINGTON 1995 ). Mating-type (MAT) switching is a specialized mitotic recombination event induced by the HO endonuclease. During MAT switching DNA sequences at the MAT locus are replaced by homologous sequences from a distant, unexpressed donor (HML or HMR) located on the same chromosome. Natural MAT switching is dependent on RAD51–57. However, only RAD52 is required in a modified system where the donor is simultaneously not silenced and located on a plasmid (SUGAWARA et al. 1995 ). When an HO-induced double-strand break (DSB) is made in one homologue in a diploid strain, repair is dependent on RAD52. However, in rad51, rad54, rad55, and rad57 mutants an aberrant repair, termed break-induced replication, occurs, which results in restoration of the broken chromosome arm (MALKOVA et al. 1996 ).

Recombination between direct repeats can occur by a variety of mechanisms (KLEIN 1995 ). When a DSB is made between direct repeats, the ends are processed by a 5'-3' nuclease to reveal complementary single-stranded regions corresponding to the direct repeats. These sequences can then be annealed, resulting in a deletion product. This mechanism for deletion formation, termed single-strand annealing (SSA), is dependent on RAD52, but independent of RAD51, RAD54, RAD55, and RAD57 (SUGAWARA and HABER 1992 ; IVANOV et al. 1996 ). The defect in SSA observed in rad52 mutants can be partially suppressed by the rfa1-D228Y allele, which decreases the affinity of replication factor A (RPA) for single-stranded DNA (SMITH and ROTHSTEIN 1995 , SMITH and ROTHSTEIN 1999 ). The requirement for RAD52 in SSA also depends on the number of repeats, as a DSB in the rDNA locus can be repaired in the absence of RAD52 (OZENBERGER and ROEDER 1991 ). The ubiquitous requirement for RAD52 in DNA recombination and repair suggests a pivotal role of this gene.

Biochemical and two-hybrid studies demonstrate that Rad52 binds to the Rad51 protein, a homologue of bacterial RecA proteins (SHINOHARA et al. 1992 ; MILNE and WEAVER 1993 ; OGAWA et al. 1993 ; DONOVAN et al. 1994 ; HAYS et al. 1995 ; SUNG 1997 ). The observation that high-level expression of RAD51 suppresses certain rad52 mutations, but not the null mutation of RAD52, is consistent with this direct interaction (MILNE and WEAVER 1993 ; SCHILD 1995 ; KAYTOR and LIVINGSTON 1996 ). Genetic and two-hybrid studies have also identified an interaction between Rad52 and RPA (FIRMENICH et al. 1995 ; SMITH and ROTHSTEIN 1995 ; PARK et al. 1996 ; HAYS et al. 1998 ). The Rad51 protein mediates a DNA strand exchange reaction, which is stimulated by the single-stranded DNA binding factor RPA (SUNG 1994 ; SUNG and ROBBERSON 1995 ). In vitro, efficient strand exchange requires the addition of RPA after Rad51 has nucleated onto the single-stranded DNA. If the single-stranded DNA is incubated with Rad51 and RPA simultaneously, the efficiency of strand exchange decreases dramatically, indicating that RPA can be both a stimulator and an inhibitor. The inclusion of the Rad52 protein can relieve the inhibitory effect of RPA and restores the efficiency of strand exchange to that observed by the sequential action of Rad51 and RPA (SUNG 1997 ; BENSON et al. 1998 ; NEW et al. 1998 ; SHINOHARA and OGAWA 1998 ). Thus, Rad52 acts as a cofactor for the Rad51 recombinase in overcoming the competition by RPA for binding to single-stranded DNA. In vitro assays reveal that purified Rad52 is capable of binding to both single-stranded DNA (ssDNA) and double-stranded DNA (dsDNA) and catalyzing annealing reactions between complementary ssDNAs (MORTENSEN et al. 1996 ). Furthermore, the inhibitory effect imposed by RPA on single-strand annealing is overcome by Rad52 (SHINOHARA et al. 1998 ; SUGIYAMA et al. 1998 ).

Homologues of the yeast Rad52 protein have been found in other organisms and in S. cerevisiae itself (BEZZUBOVA et al. 1993 ; MILNE and WEAVER 1993 ; OSTERMANN et al. 1993 ; BENDIXEN et al. 1994 ; MURIS et al. 1994 ; BAI and SYMINGTON 1996 ). The conserved amino acids mostly reside at the N-terminal portion of this family of proteins. The N-terminal region has been suggested to be responsible for binding to DNA, while the C-terminal region is thought to be involved in the interaction with the Rad51 protein (MILNE and WEAVER 1993 ; DONOVAN et al. 1994 ; MORTENSEN et al. 1996 ; SHINOHARA and OGAWA 1998 ). Since the discovery of RAD52, many mutant alleles have been isolated and characterized (RESNICK 1969 ; MALONE et al. 1988 ; BOUNDY-MILLS and LIVINGSTON 1993 ; KAYTOR and LIVINGSTON 1994 , KAYTOR and LIVINGSTON 1996 ; SCHILD 1995 ). Mutations have been identified in both conserved and nonconserved residues of Rad52. The first RAD52 allele isolated, rad52-1, is a missense mutation resulting in an A to V change at position 90, which is conserved among all of the Rad52 family members (RESNICK 1969 ; ADZUMA et al. 1984 ). rad52-1 is indistinguishable from a deletion mutation in almost every aspect examined, including the severe sensitivity to DNA-damaging agents, the deficiency in mating-type switching and other types of mitotic recombination, and the inability to complete meiosis. However, most other rad52 alleles retain some Rad52 function. For example, the rad52-2 mutation is caused by a single amino acid change from P to L at position 64, which is conserved across the whole family of Rad52-like proteins except in the yeast Rad59 protein. Despite the fact that the rad52-2 mutant fails to sporulate successfully and is severely sensitive to methyl methanesulfonate (MMS), the mutant retains significant levels of mitotic recombination and even exhibits a hyper-recombinational phenotype for interchromosomal recombination (MALONE et al. 1988 ; BOUNDY-MILLS and LIVINGSTON 1993 ).

In this study we were interested in identifying factors responsible for RAD51-independent mitotic recombination. We carried out a screen for recombination-defective mutants in a rad51 strain containing an inverted-repeat recombination substrate. From this screen a non-null allele of RAD52 that caused weak -ray sensitivity, reduced mitotic recombination, but was not sporulation defective, was identified. This mutation displayed synergistic effects with the rad51 mutation for inverted-repeat recombination and with the rad59 mutation for -ray sensitivity, mating-type switching, and sporulation.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Yeast strains:
The relevant genotypes of the yeast strains used in this study are given in Table 1. All strains are derivatives of strains W303-1A or W303-1B (THOMAS and ROTHSTEIN 1989 ). Strains containing the ade2-5'-TRP1-ade2-n construct and a deletion of the RAD51 gene were described previously (BAI and SYMINGTON 1996 ). A strain containing a deletion-disruption allele of the RAD59 gene (rad59::LEU2), B368-1A, was constructed by PCR using the following primers: 5' GAGGG A G T C T G T G G C A G T T T AGCACATGCTTTGGACCATTctcgaggagaacttctagta3'; 5' ATATGCGTGCCTTTAGCATCCCTCCAATTTGATAAAAGTCGtcgactacgtcgtaaggccg 3'. Bases in uppercase represent RAD59 sequences and those in lowercase represent LEU2 sequences. Disruption of the RAD59 gene was confirmed by PCR. To construct strain B420, strain W1479-11C was first transformed with plasmid pRS414:MATa to allow mating to B413-8C. Haploid progeny derived from this cross were grown nonselectively and then screened on SC-Trp to identify plasmid-free segregants. All other strains were made by mating the appropriate strains, sporulating and dissecting tetrads from the resulting diploids, and screening haploid progeny for those with the required genotypes.

Plasmids:
pRS416:Erad52 was constructed by inserting a 2.4-kb EcoRI-digested, rad52R70K-containing fragment at the EcoRI site of pRS416. The fragment was made by a PCR reaction performed on the genomic DNA of a rad52R70K strain with the use of the following primers: 5' gataaGAATTCgcctagaatgaaagtaagtgaattagcg 3'; 5' gatgGAATTCaatgaacctaaggattccgctg 3'. pRS416:ERAD52 was made in a similar way except that the inserted fragment was amplified by using wild-type genomic DNA as the PCR template. pRS426:Erad52 was made by cloning the 2.4-kb rad52R70K fragment purified from EcoRI-digested pRS416:Erad52 into the EcoRI site of pRS426.YEp24:RAD51 contains a 3.7-kb RAD51 fragment inserted at the BamHI site of YEp24. YEp24:RAD59 contains a RAD59 fragment inserted at the BamHI site of YEp24. pRS416-SU that carries the leu2 inverted repeats and plasmids pFH800 and pBM272-HO used for induction of the HO endonuclease have been described previously (NICKOLOFF et al. 1986 ; BAI and SYMINGTON 1996 ; MOREAU et al. 1999 ).

Cell growth and genetic methods:
Cells were grown on either YEPD or synthetic complete (SC) media for most procedures (SHERMAN et al. 1986 ). Selection for Ura^- cells was performed on SC medium containing 5-fluoroorotic acid (5-FOA) at 1 mg/ml. Selection for G418-resistant cells was on YEPD medium containing 0.5 mg/ml of G418. To induce expression of the HO endonuclease, cultures were grown in synthetic medium containing 2% raffinose as a carbon source prior to the addition of galactose. Yeast mating, sporulation, and tetrad dissection were performed as described (SHERMAN et al. 1986 ). Cells were grown at 30° unless otherwise indicated. The measurement of recombination rates was as described previously (BAI and SYMINGTON 1996 ). Results for strains not containing the rad52R70K mutation have been presented previously (BAI and SYMINGTON 1996 ) and were included here for comparison. Rates for the ade2 and leu2 inverted repeats and diploid heteroallelic recombination were determined with three independent isolates and the mean values are presented. Rates were determined only once for the Tn903 substrate and the median values are presented.

Mutagenesis:
Mutagenesis was performed as described in BAI and SYMINGTON 1996 .

Determination of -ray sensitivity:
Cells were grown in liquid medium at 30° to mid-log phase. Synthetic medium lacking uracil was used to maintain selection for plasmids. Serial dilutions were made from each culture and aliquots of each dilution were spotted onto solid medium YEPD medium or medium lacking uracil. Cells were radiated in a Gammacell-220 ⁶⁰Co irradiator (Atomic Energy of Canada) with various doses of -rays. Percentage survival was measured after 3–4 days of incubation at 30°. Each experiment was repeated three times on independent transformants and the mean value was presented.

Physical analysis of mating-type switching and single-strand annealing:
Strains to be tested for mating-type switching (W303-1B, B361-4C, B413-13B, and B413-8C) were transformed to Ura⁺ with plasmid pBM272-HO. Ura⁺ plasmid-containing transformants were grown in 5 ml SC-Ura medium for 24 hr. Cells were harvested, washed with water, and used to innoculate 250 ml SC (raffinose)-Ura. Cultures were grown to an OD₆₀₀ of 0.4–0.5. A total of 50 ml of culture was removed for the 0 hr timepoint and then 22.5 ml of 20% galactose was added. One hour after addition of galactose, the cultures were harvested and resuspended in 250 ml of YEPD. Fifty-milliliter samples were removed at 1-hr intervals after induction for DNA analysis. Cells were harvested by centrifugation and washed with water, and the cell pellets were frozen in liquid nitrogen. DNA was extracted from the thawed cell pellets and digested with StyI, and DNA fragments were separated by electrophoresis through 1% agarose gels. DNA fragments were transferred to nylon membranes and hybridized with a 300-bp PCR fragment generated by amplification of MAT sequences distal to the HO cut site.

The strains to be used for physical analysis of HO-induced deletion formation (B420-9D, B420-1B, B420-3A, and B420-6C) were transformed to Trp⁺ using pFH800. Induction of HO was as described above except cells were grown in medium lacking tryptophan. DNA samples were digested with SpeI and DNA fragments separated by electrophoresis through 0.8% agarose gels. DNA fragments were transferred to nylon membranes and hybridized with a 400-bp PCR fragment generated by amplification of sequences from the YCL017 ORF, which is adjacent to LEU2.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Isolation of a novel rad52 allele:
To measure the level of recombination in vivo, a previously described recombination substrate was utilized (RATTRAY and SYMINGTON 1994 ) (Figure 1). The substrate is located on chromosome XV and consists of inverted heteroalleles of the ADE2 gene. Both alleles are inactive, but the substrate can be rearranged by recombination events to form a functional ADE2 gene. Using this substrate, the average rate of recombination was determined to be 1.8 x 10^-4 events/cell/generation in wild-type strains and 3.5 x 10^-5 in rad51 mutants. A screen for recombination mutants was performed in the rad51 background. A rad51 strain (B356-13D) was mutagenized by N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) and mutants displaying reduced sectoring were isolated. Approximately 15,000 colonies were screened. Since a rad52 null mutation was known to dramatically reduce recombination of the ade2 substrate (>3000-fold), isolated mutants were tested for complementation of -ray sensitivity by a rad52 strain (LSY255). Three mutants completely failed to complement the rad52 strain in that diploids from the crosses were as sensitive as a rad52 homozygous diploid strain to -irradiation, indicating that these mutants were likely to have acquired rad52 null mutations. One other mutant, no. 17, partially complemented the rad52 strain in the -ray complementation test, suggesting a partial loss-of-function mutation of RAD52. The reduced-sectoring phenotype of no. 17 was not due to loss or mutation of the inverted-repeat recombination substrate, because diploids from the cross between no. 17 and LSY255 sectored at close to the wild-type level, and no. 17 was the only source for the inverted-repeat substrate. Mutant 17 was then backcrossed to an isogenic rad51 strain (B356-11A). Because the resulting diploids were homozygous for the rad51 mutation and were defective in meiosis, a RAD51-containing plasmid, YEp24:RAD51, was introduced into the diploids to complement the defect. The diploids were sporulated and after tetrad dissection plasmid-free haploid progeny were obtained by counterselecting against the plasmid marker gene URA3 on 5-FOA-containing medium. Haploid progeny were streaked out on YEPD plates and examined for sectoring. The low-sectoring phenotype segregated 2:2 in this backcross, indicating that the unidentified mutation in mutant 17 was a single gene trait. Mutant 17 was derived from a rad51 strain and was thus extremely sensitive to -irradiation. To test whether the unidentified mutation in no. 17 would confer -ray sensitivity by itself, a strain that carried the unidentified mutation but had a wild-type RAD51 gene (B400-1A) was made. B400-1A displayed a mild -ray sensitivity, at a level between that of the rad52 mutant and wild type (see Figure 2). To confirm that the unidentified mutation in no. 17 was allelic to the RAD52 locus, B400-1A was crossed to LSY255 (rad52). Diploids from this cross were capable of sporulation, but at a reduced level. Following tetrad dissection, the haploid progeny were tested for -ray sensitivity. The intermediate-sensitivity phenotype representative of the unidentified mutation always segregated away from the severe-sensitivity phenotype rendered by the rad52 mutation. This result confirmed that the unidentified mutation in mutant 17 was allelic to RAD52, or very closely linked.

View larger version (17K): In this window In a new window Download PPT slide   Figure 1. (A) ade2 inverted-repeat substrate. The substrate contains an intrachromosomal inverted duplication of heteroalleles of the ade2 gene integrated at the HIS3 locus on chromosome XV. The heteroalleles are separated by an active TRP1 gene. One of the ade2 alleles contains a deletion on the 5' side (ade2-5'), and the other one has a frameshift mutation at a NdeI site on the 3' side (ade2-n). Initial strains with nonrecombined substrates (Ade-) form red colonies on nonselective media due to accumulation of a red pigment in the adenine biosynthetic pathway. If recombination between the two alleles produces a wild-type ADE2 gene, a white sector will be formed within the red colony. Thus, the recombination level of a strain can be visually assessed by colony sectoring and quantitatively determined by measuring the frequency of Ade+ prototrophs within a population of cells. Ade+ recombinants can arise by gene conversion and/or events that invert the TRP1 gene. (B) The IS903 inverted-repeat substrate. A fragment of the bacterial transposon Tn903 was integrated into yeast chromosome III between HIS4 and HML loci. The fragment contains the kanamycin resistance gene (kanr) of Tn903, flanked by two inverted copies of the IS903 sequences. The inverted IS903 repeats can undergo an intrachromosomal reciprocal exchange resulting in inversion of the kanr gene. A yeast strain containing a single copy of the kanr gene in its original orientation is sensitive to 0.5 mg/ml of G418. After inversion the kanr gene confers resistance to G418, possibly because the kanr gene is then expressed more efficiently (WILLIS and KLEIN 1987 ). (C) The leu2 inverted-repeat substrate. The substrate is carried on a CEN plasmid and contains two alleles of the yeast LEU2 gene that have been truncated for the 5' and 3' coding regions, respectively. A functional LEU2 gene can be generated by recombination between the inverted leu2 repeats (PRADO and AGUILERA 1995 ). (D) Interchromosomal recombination substrate. One of the ade2 alleles of the diploid contains a fill-in mutation at the 3' NdeI site, and the other one contains a fill-in mutation at the 5' AatII site (HUANG and SYMINGTON 1994 ; BAI and SYMINGTON 1996 ). A functional ADE2 gene can be generated by interchromosomal recombination.

View larger version (19K): In this window In a new window Download PPT slide   Figure 2. The rad52R70K and rad59 mutations synergistically increase sensitivity to -irradiation.

Cloning and identification of the novel rad52 allele:
DNA fragments containing the unidentified rad52 allele and the wild-type RAD52 sequence were amplified by PCR from the genomic DNA of mutant 17 and W303-1A, respectively. On the basis of the sequence of the RAD52 locus, the wild-type fragment expected was 2.4 kb in length, covering a region extending from 462 bp upstream of the first ATG of the RAD52 ORF to 395 bp downstream of the stop codon. The amplified PCR fragment carrying the rad52 allele was of the same size as the corresponding wild-type fragment, indicating that the unidentified mutation in mutant 17 did not involve a large deletion or insertion. The 2.4-kb PCR fragment was cloned into the EcoRI site of pRS416 to create recombinant plasmids containing the mutant allele and the wild-type RAD52, respectively. The mutant plasmid (pRS416:Erad52), upon transformation into LSY255 (rad52 null), partially restored the strain's resistance to -ray radiation to a level similar to that of B400-1A, whereas the wild-type plasmid (pRS416:ERAD52) fully complemented LSY255. By replacing DNA segments from the wild-type sequence with those from the mutant, the location of the unidentified rad52 mutation was narrowed down to the region 5' of the single BglII site in the cloned fragment. DNA sequence analysis revealed a single nucleotide change from G to A at position 209 of the RAD52 ORF, resulting in a R to K missense mutation at position 70 of the Rad52 protein. The observed mutation was identified from independently generated PCR fragments ruling out the trivial possibility of artifact from a PCR-derived nucleotide misincorporation. The rad52 allele in mutant 17 was designated rad52R70K.

rad52R70K and rad51 synergistically reduce recombination:
Mitotic recombination, assayed on the ade2 inverted-repeat substrate, was reduced 5-fold in a rad51 mutant and colonies sectored at a level slightly lower than the wild-type strain (RATTRAY and SYMINGTON 1994 ). Strain B400-1A (rad52R70K), like a rad51 mutant, displayed only slightly reduced colony sectoring compared with wild type. However, mutant 17 (rad51 rad52R70K) was isolated from a rad51 background by the strong reduction in colony sectoring. Thus, the rad52R70K mutation displayed a synergistic effect with the rad51 mutation. The synergistic reduction in recombination was verified by determining the rates of recombination for each strain (Table 2). Compared with wild-type strains, recombination rates in rad51 or rad52R70K single mutants were reduced only 5-fold, whereas the rates in rad51 rad52R70K double mutants were reduced 1900-fold.

RAD59, a recombination gene encoding a Rad52 homologue, was identified in the same genetic screen in which mutant 17 was isolated (BAI and SYMINGTON 1996 ). When tested on the ade2 inverted-repeat substrate, a rad59 mutation resembled the rad52R70K mutation in that rad59 single mutants showed a 4- to 5-fold reduction in recombination rates, whereas rad51 rad59 double mutants were reduced 1100-fold. The similarity in phenotype between rad59 and rad52R70K strains suggested that both might be defective in the same recombination pathway, in which case a double mutant would be expected to show the same recombination rate as the single mutants. The rad59 rad52R70K double mutant showed a 3- to 4-fold reduction in the recombination rate compared with strains containing either mutation alone, indicative of additive effects (Table 2). However, a rad51 rad52R70K rad59 triple mutant showed a rate of recombination similar to those of rad51 rad52R70K and rad51 rad59 strains, suggesting that the additive effect conferred by the rad52R70K and rad59 mutations is dependent on RAD51.

To determine if the rad52R70K mutation has a general effect on recombination, rates were determined using several other recombination substrates (Figure 1). Assayed on a substrate consisting of intrachromosomal inverted IS903 repeats, a rad52R70K or rad51 single mutation reduced recombination slightly, whereas when combined together these two mutations synergistically reduced recombination (Table 3). Similar results were observed for the rad59-1 mutation, a rad59 allele indistinguishable from the null mutation in recombination and repair assays (BAI and SYMINGTON 1996 ). Previous studies showed a reduction >100-fold in the rate of recombination of this reporter in a rad52-1 strain (WILLIS and KLEIN 1987 ). The synergism seen on the IS903 substrate was consistent with that on the ade2 substrate. However, a different result was obtained using a substrate where recombination occurs between inverted leu2 heteroalleles located on a plasmid (PRADO and AGUILERA 1995 ). On this substrate a rad52R70K mutation reduced recombination 3–4-fold, compared with a 100-fold reduction conferred by the rad52-1 allele (PRADO and AGUILERA 1995 ). A rad51 mutation had little or no effect and no synergistic effect was found between rad51 and rad52R70K mutations since a double mutant was not significantly different from a rad52R70K single mutant strain (Table 4). The effect of rad52R70K on interchromosomal recombination was tested by determining the rate of Ade⁺ prototroph formation between ade2 heteroalleles located on homologous chromosomes in diploids. In this assay a rad52R70K or rad59 mutation actually elevated the rate of recombination (Table 5). This contrasts with rad52 null strains, which show a large decrease in the rate of heteroallelic recombination (HOEKSTRA et al. 1986 ; RESNICK et al. 1986 ). The rate in the rad51 strain was below the measurable range of this assay, as were rates in rad51 rad52R70K, rad51 rad59, and rad51 rad52R70K rad59 strains (data not shown). Thus, interchromosomal recombination is dependent on RAD51, but not the functions disabled by rad52R70K or rad59 mutations. The rad52R70K and rad59 mutations exhibited similar defects in mitotic recombination using all of these substrates.

rad52R70K and rad59 synergistically increase -ray sensitivity:
Mutations in genes required for homologous recombination often result in increased sensitivity to DNA damage agents such as ionizing-radiation and radiomimetic chemicals. Among genes of the RAD52 epistasis group in S. cerevisiae, a mutation in RAD59 causes an intermediate sensitivity to -ray radiation, while null mutations in other genes of the group lead to severe sensitivity. Upon exposure to -irradiation, a rad52R70K mutant strain displayed a survival level between that of a rad59 mutant and that of a wild-type strain (Figure 2). The rad52R70K rad59 double mutant was more sensitive to ionizing radiation than either of the single mutants, indicating a synergistic effect between rad52R70K and rad59 mutations in DNA repair. The rad52 rad59 double mutant was as sensitive to -irradiation as the rad52 single mutant.

rad52R70K is suppressed by overexpression of the mutant allele:
In a rad52R70K single mutant strain, the introduction of a high-copy-number plasmid carrying the rad52R70K allele (pRS426:Erad52) partially restored the strain's resistance to -irradiation (Figure 3). The sensitivity of the rad52R70K rad59 double mutant was also partially suppressed by the same plasmid. In neither case was the suppression complete. The rad52R70K mutation could not be suppressed by YEp24:RAD59, a high-copy-number plasmid carrying RAD59, and the rad59 mutation could not be suppressed by pRS426:Erad52.

View larger version (20K): In this window In a new window Download PPT slide   Figure 3. Suppression of the -ray sensitivity of rad52R70K strains by increased copy number of the rad52R70K allele.

Defects caused by the rad52R70K mutation could have arisen from a weakened Rad52 activity and/or a reduced level of Rad52 expression. In either case the mutant phenotype could be suppressed by the overexpression of the mutant allele. However, by Western-blotting analysis using anti-Rad52 antibodies we were unable to detect significant differences in the level of Rad52 between rad59, rad52R70K, rad59 rad52R70K, and wild-type strains (data not shown). Thus, it is reasonable to suggest that the rad52R70K mutant phenotype is due to impaired activity of the Rad52 mutant protein.

The repair of HO-induced DSBs is defective in rad52R70K and rad59 mutants:
-Irradiation creates a variety of DNA lesions in addition to double-strand breaks. To determine whether the rad52R70K, rad59, and rad52R70K rad59 strains are defective in the repair of a single double-strand break, a mating-type switching assay was performed. The repair of an HO endonuclease-induced double-strand break (DSB) was monitored at the DNA level after induction of HO endonuclease for 1 hr. To measure the formation of switched products, the DNA samples were digested with StyI, which cuts within Ya but not Y sequences. The appearance of a 0.9-kb StyI fragment is indicative of repair of the DSB from the HMRa locus (Figure 4). In the wild-type strain, switching was efficient and completed 3 hr after induction of HO. In both rad52R70K and rad59 mutants switching was delayed and most of the cut DNA was not converted to MATa product. This defect was even more severe in the double mutant. Although cut fragment was produced and disappeared with normal kinetics, there was a greater reduction in the formation of switched product compared with the single mutants. The disappearance of the 0.7-kb cut fragment suggests that exonuclease degradation from the DSB occurs normally, but subsequent events are defective in these mutants. To ensure that the mutants were proficient in formation of the single-stranded tail at the break site, DNA samples digested with StyI and BamHI were analyzed by alkaline gel electrophoresis (WHITE and HABER 1990 ; MOREAU et al. 1999 ). Single-stranded tails were formed in all of the mutants with the same kinetics indicating no defect in the nuclease-processing step of the reaction (data not shown). At the 5 hr timepoint, cells from all strains were plated on SC-Ura (to select for those containing the HO plasmid) for phenotypic analysis; 63% of wild type, 6.3% of rad52R70K, 26.7% of rad59, and 1.1 % of the rad52R70K rad59 colonies were MATa maters, consistent with the physical analysis.

View larger version (42K): In this window In a new window Download PPT slide   Figure 4. Kinetics of mating-type switching. A schematic representation of the MATa and MAT loci indicating the locations of the StyI sites and the hybridization probe (top). HO endonuclease produces a 0.7-kb fragment from the 1.8-kb MAT StyI fragment. A 0.9-kb StyI fragment is produced when the mating type switches from MAT to MATa. DNA was isolated from cultures prior to galactose induction (0-hr timepoint) and at 1-hr intervals after HO induction.

The repair of DSBs can occur by a variety of mechanisms. When a DSB is made between direct repeats, repair can occur by single-strand annealing. This nonconservative reaction occurs by 5'-3' degradation from the DSB site to reveal complementary single-stranded regions that can anneal, resulting in deletion of DNA between the repeats (HABER 1995 ). This reaction is dependent on RAD52, but is independent of other tested genes in the RAD52 epistasis group (IVANOV et al. 1996 ). As the DNA binding domain of Rad52 is in the N-terminal half of the protein, it seemed possible that the rad52R70K mutation might affect the DNA binding and annealing properties of the mutant protein. Also, the N-terminal region of Rad52 is conserved with Rad59, suggesting that Rad59 might also participate in strand annealing. Strains containing direct repeats of the leu2 gene, separated by a copy of the URA3 gene and vector sequences containing the HO cut site (HO cs), were used to monitor the efficiency of DSBR by single-strand annealing (Figure 5; SMITH and ROTHSTEIN 1999 ). HO endonuclease was induced for 1 hr and repair by SSA was monitored at the DNA level by the appearance of a 5.7-kb SpeI fragment. In the wild-type strain, cleavage by HO was efficient and all of the cut fragments were converted to deletion products (Figure 5B). For all of the mutants, there was efficient cleavage by HO, but there was a delay in the appearance of deletions and only half of the cut fragments were converted to the deletion products. The rad52R70K rad59 double mutant showed a decrease in the formation of deletions (27.7% deletion product at 5 hr) slightly greater than that of the single mutants (36.3% for rad52R70K and 35.9% for rad59).

View larger version (98K): In this window In a new window Download PPT slide   Figure 5. Kinetics of HO-induced deletion formation. (A) Schematic representation of the leu2 direct repeat substrate showing the locations of the SpeI sites, the HO cleavage site, and hybridization probe. After cleavage by HO, an 8.3-kb fragment is produced from the 13.4-kb SpeI fragment. The single-stranded tails formed after resection from the DSB site can anneal to form a deletion product that is detected as a 5.7-kb SpeI fragment. (B) DNA was extracted at the times shown and digested with SpeI, the position of parental, HO cut fragment, and deletion products are shown to the right of the autoradiogram.

rad52R70K and rad59 cause synergistic meiotic defects:
Recombination gene mutations frequently lead to meiotic defects. A diploid strain homozygous for the rad52R70K mutation showed a sporulation efficiency of 61%, slightly lower than that of a wild-type diploid strain (78%; Table 6). The spore viability of the rad52R70K strain (75%) was also slightly reduced from the wild-type level (98%). A rad59 homozygous diploid strain sporulated at 65% efficiency, with a spore viability of 84%. However, a diploid strain homozygous for both rad52R70K and rad59 mutations displayed only 5% sporulation efficiency and 6% spore viability, which were much lower than either of the single mutant diploid strains. This result strongly suggests that the biological activities abolished by the rad52R70K and the rad59 mutations are required for yeast meiosis.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The isolation of the rad52R70K allele in a genetic screen for recombination deficiency is direct evidence that the arginine residue at position 70 of the Rad52 protein is important for function. The rad52R70K missense mutation causes defects in mitotic recombination and DNA repair. More interestingly, this mutation shows synergistic effects with the rad51 mutation for inverted-repeat recombination and with the rad59 mutation for -ray sensitivity, mating-type switching, and sporulation. However, the rad52R70K allele retains substantial Rad52 function, because in all of the recombination and repair assays the rad52R70K mutation displayed a much less severe defect than a rad52 null mutation.

RAD51-independent recombination:
A single mutation of rad52R70K, like a rad59 mutation, reduces the rate of mitotic recombination only four- to fivefold using inverted-repeat substrates. The synergistic effects of a rad52R70K or rad59 mutation with the rad51 mutation indicate that RAD59 and the rad52R70K disabled activity are involved in a RAD51-independent recombination mechanism. The RAD51-independent mechanism cannot completely substitute for the RAD51-dependent mechanism and vice versa, since the elimination of either one of them decreases recombination rates. Yet the decrease is not substantial, indicating that each type of mechanism is potent by itself. The rad51 rad52R70K rad59 triple mutant showed a similar rate of recombination as did the rad51 rad52R70K and rad51 rad59 double mutants, indicating that RAD59 and the disabled function of rad52R70K are likely to function in the same pathway for mitotic recombination. However, the rad52R70K rad59 double mutant showed a rate of recombination three- to fourfold lower than those of both of the single mutants indicating additive effects. This suggests that RAD59 and the RAD52 activity disabled by the rad52R70K mutation possess biological functions largely overlapping, yet also with slight differences, in recombination events on the ade2 inverted-repeat substrate. The synergism displayed by rad51 with rad52R70K or rad59 for inverted-repeat recombination suggests that there are multiple pathways for recombination of this substrate that are differentially affected by these mutations. Alternatively, Rad51 might have an overlapping function with Rad52 and Rad59. The synergistic defects could also be caused by destabilization of a multiprotein complex. Rad51 and Rad52 are known to interact and we have shown an interaction between Rad52 and Rad59 (A. DAVIS and L. SYMINGTON, unpublished observations).

Most genes of the RAD52 epistasis group were identified by their requirement for the repair of ionizing-radiation-induced DNA damage (GAME and MORTIMER 1974 ), and subsequent studies revealed defects in homologous recombination. However, defects in DNA repair do not always fully correlate with recombination defects. For example, a rad51 mutant is extremely sensitive to ionizing radiation whereas the mutant exhibits heterogeneous phenotypes for recombination. A rad51 mutant is largely proficient in mitotic recombination in assays using certain types of inverted-repeat substrates, including both spontaneous and double-strand break-induced events (RATTRAY and SYMINGTON 1994 ; IVANOV et al. 1996 ). Neither is RAD51 required in a recombination assay using two copies of the yeast MAT sequences as substrates when the donor sequence is simultaneously not silenced and located on a plasmid (SUGAWARA et al. 1995 ). RAD51 is also not required for the formation of Holliday junction intermediates in the rDNA or for sister chromatid joint molecules in meiosis (SCHWACHA and KLECKNER 1997 ; ZOU and ROTHSTEIN 1997 ). However, RAD51 is required for many recombination events. RAD51 is essential for normal mating-type switching and is important for heteroallelic recombination in diploids. rad51 diploids are unable to complete meiosis, but by both physical and genetic assays there is only a 5- to 10-fold reduction in the level of meiotic recombination (SHINOHARA et al. 1992 ). Mutations in several other genes of the RAD52 epistasis group (RAD54, RAD55, and RAD57) are similar to rad51 with respect to heterogeneous phenotypes in recombination. Only the rad52 mutation is nearly homogeneous in recombination and repair phenotypes, in that a rad52 mutant is both sensitive to -irradiation and defective in mitotic recombination of a variety of substrates. These results suggest that cellular controls for DNA repair and the controls for recombination on specific types of substrates, although partially overlapping, involve functions distinct from each other.

RAD59 is required for efficient double-strand break repair:
In this study, RAD59 and the RAD52 function disabled by the rad52R70K mutation were shown to be required for efficient mating-type switching and single-strand annealing, two different DSB-initiated recombination events. By physical analysis, the repair reaction was delayed in both mutants, and there was a 2- to 10-fold reduction in the level of recombination products compared with that of the wild-type strain. There are several possible explanations for the reduction in recombination products. First, as unsynchronized cultures were used for the HO induction experiments, repair could have occurred from a sister chromatid instead of intrachromosomally. The use of the sister chromatid as a donor for repair would not be detected by the physical assay. Although we cannot rule out this possibility, it seems unlikely because both mutants are -ray sensitive and repair of lesions in haploid cells is thought to occur by sister chromatid recombination. Furthermore, rad59 is lethal in combination with rad27, suggesting that the recombinational repair of lesions during S-phase or G2 is defective in rad59 mutants (SYMINGTON 1998 ). Second, the mutants could be defective in the strand invasion or strand-annealing steps of these reactions; this is the model we favor. The displacement of RPA from single-stranded DNA by Rad52 is required for both strand invasion and SSA. A defect in the Rad52-RPA interaction, Rad52 self-interaction, or Rad52-DNA binding could cause the defects observed in the rad52R70K strain. Although the biochemical function of Rad59 is currently unknown, the homology between Rad59 and the N-terminal region of Rad52 and the preponderance of basic residues in Rad59 suggest that it is a DNA binding protein. Rad59 could potentially play a similar role to Rad52 in strand annealing, or the complex of Rad52 and Rad59 may be more efficient in these processes than Rad52 alone. The more severe defect of the rad59 rad52R70K double mutant is consistent with the proteins having overlapping biochemical activities.

RAD52 and RAD59 have partially overlapping functions in meiosis:
As stated earlier, the single rad52R70K and rad59 mutations have little effect on sporulation or spore viability, however, the double mutant is extremely deficient in both spore formation and subsequent viability. This finding suggests that RAD59 and the function of RAD52 disabled by the rad52R70K mutation have partially redundant functions in meiosis. This is the first evidence of a role for RAD59 in meiosis and suggests that RAD59 can carry out a meiotic function of RAD52 normally mediated by the N-terminal region of the Rad52 protein. RAD51 and DMC1, which encodes a meiosis-specific RecA homologue, also have redundant functions in meiosis. Meiotic recombination products are reduced 5- to 10-fold by mutation of either RAD51 or DMC1, but recombination is severely reduced in the rad51 dmc1 double mutant (SHINOHARA et al. 1997A ). Similarly, RAD54 and RDH54, which encodes a Rad54 homologue, have overlapping functions during meiosis (KLEIN 1997 ; SHINOHARA et al. 1997B ).

Comparison of rad52 alleles:
The rad52R70K allele results from a single amino acid change from arginine to lysine at position 70 that resides in the conserved N-terminal region of the Rad52 protein. Arginine 70 is conserved among all of the Rad52 homologues from various organisms except in the yeast Rad59 protein, where the corresponding residue happens to be lysine. The N-terminal part of Rad52 has been suggested to contain a DNA-binding domain (MILNE and WEAVER 1993 ; MORTENSEN et al. 1996 ; SHINOHARA and OGAWA 1998 ). It is possible that the R70K mutation alters the DNA binding capability of Rad52, even though they are similarly charged residues. Alternatively, this position might be important for interaction with other proteins. The N-terminal region of Rad52 is thought to be required for self-association as well as interaction with RPA (PARK et al. 1996 ; HAYS et al. 1998 ). Previously, some other mutations in the N-terminal region of Rad52 have been characterized and found to display diverse phenotypes (Figure 6). The rad52-1 (A90V) allele is indistinguishable from a deletion mutation in nearly every aspect tested, including the -ray sensitivity and meiosis defects (RESNICK 1969 ; ADZUMA et al. 1984 ). The rad52-2 (P64L) allele causes sporulation deficiency and sensitivity to -irradiation and MMS, but the mutant exhibits a hyperrecombinational phenotype for interchromosomal recombination (MALONE et al. 1988 ; BOUNDY-MILLS and LIVINGSTON 1993 ). rad52-76A (N97T), rad52-23A (V128I), and rad52-22B (V162A) alleles give rise to temperature-sensitive phenotypes for cell survival after exposure to MMS or -rays, but unconditionally retain the ability to undergo mitotic and meiotic recombination (KAYTOR and LIVINGSTON 1994 ). rad52-301A (K61N) and rad52-231A (K69D) were characterized as cold sensitive with regard to growth on MMS-containing media, but this phenotype was found only using plates containing a low concentration of MMS (NGUYEN and LIVINGSTON 1997 ). There are several similarities between strains with rad52 mutations in this region of the protein. For example, rad52-2 (P64L) and rad52R70K both increase the rate of interchromosomal recombination, but show slight reductions in the rate of inverted-repeat recombination. Strains containing either the rad52-231A or rad52R70K alleles show similarity in spore viability of homozygous diploids (70% vs. 75% at 25°) and survival to HO-induced breaks. Whether the residual activity of the rad52-2, rad52-231A, and rad52-301A strains depends on RAD59 has yet to be determined. In addition to mutations in the N-terminal portion of Rad52, mutations in the C-terminal portion have also been found to affect Rad52 activity. Among them are a C-terminal truncation allele that exerts dominant negative effects on MMS resistance (MILNE and WEAVER 1993 ), another C-terminal truncation allele that retains partial ability to repair DNA damage and to undergo recombination (BOUNDY-MILLS and LIVINGSTON 1993 ), and a temperature-sensitive missense allele that is proficient in mitotic and meiotic recombination (KAYTOR and LIVINGSTON 1994 ). Attempts to isolate suppressors of the rad52 deletion allele have been unsuccessful, indicating that the Rad52 protein possesses an activity not easily bypassed. Suppressors of nondeletion rad52 alleles have been isolated. The three N-terminal ts alleles are suppressed by any one of a group of suppressors. These suppressors are recessive and involve mutations of different complementation groups (KAYTOR and LIVINGSTON 1996 ). The C-terminal truncation alleles, C-terminal missense ts allele, and an allele of undetermined nature (rad52-20) are suppressible by the overexpression of Rad51, by mating-type heterozygosity, and in some cases by deletion of SRS2 (MILNE and WEAVER 1993 ; KAYTOR and LIVINGSTON 1994 ; KAYTOR et al. 1995 ; SCHILD 1995 ). The suppression of C-terminal mutations by the overexpression of Rad51 is consistent with the finding that the Rad52 C-terminal is responsible for the protein–protein interaction with Rad51. rad52-1 or rad52-2 alleles are not suppressed by any of the above suppressors. Investigation of mutant alleles has improved our understanding of the biological significance of RAD52 and biochemical analysis of these mutants is likely to shed light on the diverse functions of this protein.

View larger version (23K): In this window In a new window Download PPT slide   Figure 6. RAD52 functional domains and mutant alleles. The locations of the DNA-binding domain and Rad51-interaction domain are from published studies. ts, temperature-sensitive alleles; cs, cold-sensitive alleles. The C-terminal truncation alleles and a C-terminal missense ts allele are suppressed by overexpression of RAD51.

In summary, we identified an unusual allele of RAD52 that confers DNA repair and recombination defects similar to those caused by mutation of RAD59. The similarity between the mutant phenotypes, in combination with the synergistic defects of the double mutant strain, suggest that the two proteins function together and/or have overlapping activities. These results provide further support for the idea that vertebrates have other Rad52-like activities that can compensate for the loss of RAD52 function (RIJKERS et al. 1998 ; YAMAGUCHI-IWAI et al. 1998 ).

	ACKNOWLEDGMENTS

We thank members of the Symington lab and C. S. H. Young for helpful discussions and U. Mortensen and W. K. Holloman for critical reading of the manuscript. We thank N. Erdeniz and U. Mortensen for carrying out Western blot analysis of Rad52 in various strains and A. Aguilera, J. Nickoloff, R. Rothstein, and J. Smith for strains and plasmids. This work was supported by grants from the National Institutes of Health (GM41784 and T32 AI07161).

Manuscript received March 19, 1999; Accepted for publication July 12, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ADZUMA, K., T. OGAWA, and H. OGAWA, 1984  Primary structure of the RAD52 gene in Saccharomyces cerevisiae.. Mol. Cell. Biol. 4:2735-2744[Abstract/Free Full Text].

AJIMURA, M., S. H. LEEM, and H. OGAWA, 1993  Identification of new genes required for meiotic recombination in Saccharomyces cerevisiae.. Genetics 133:51-66[Abstract].

BAI, Y. and L. S. SYMINGTON, 1996  A Rad52 homolog is required for RAD51-independent mitotic recombination in Saccharomyces cerevisiae.. Genes Dev. 10:2025-2037[Abstract/Free Full Text].

BENDIXEN, C., I. SUNJEVARIC, R. BAUCHWITZ, and R. ROTHSTEIN, 1994  Identification of a mouse homologue of the Saccharomyces cerevisiae recombination and repair gene, RAD52.. Genomics 23:300-303[Medline].

BENSON, F. E., P. BAUMANN, and S. C. WEST, 1998  Synergistic actions of Rad51 and Rad52 in recombination and DNA repair. Nature 391:401-404[Medline].

BEZZUBOVA, O. Y., H. SCHMIDT, K. OSTERMANN, W. D. HEYER, and J. M. BUERSTEDDE, 1993  Identification of a chicken RAD52 homologue suggests conservation of the RAD52 recombination pathway throughout the evolution of higher eukaryotes. Nucleic Acids Res. 21:5945-5949[Abstract/Free Full Text].

BOUNDY-MILLS, K. L. and D. M. LIVINGSTON, 1993  A Saccharomyces cerevisiae RAD52 allele expressing a C-terminal truncation protein: activities and intragenic complementation of missense mutations. Genetics 133:39-49[Abstract].

DONOVAN, J. W., G. T. MILNE, and D. T. WEAVER, 1994  Homotypic and heterotypic protein associations control Rad51 function in double-strand break repair. Genes Dev. 8:2552-2562[Abstract/Free Full Text].

FIRMENICH, A. A., M. ELIAS-ARNANZ, and P. BERG, 1995  A novel allele of Saccharomyces cerevisiae RFA1 that is deficient in recombination and repair and suppressible by RAD52.. Mol. Cell. Biol. 15:1620-1631[Abstract].

GAME, J. C. and R. K. MORTIMER, 1974  A genetic study of x-ray sensitive mutants in yeast. Mutat. Res. 24:281-292[Medline].

GAME, J. C., T. J. ZAMB, R. J. BRAUN, M. RESNICK, and R. M. ROTH, 1980  The role of the radiation (rad) genes in meiotic recombination in yeast. Genetics 94:51-68[Abstract/Free Full Text].

HABER, J. E., 1995  In vivo biochemistry: physical monitoring of recombination induced by site-specific endonucleases. Bioessays 17:609-620[Medline].

HAYS, S. L., A. A. FIRMENICH, and P. BERG, 1995  Complex formation in yeast double-strand break repair: participation of Rad51, Rad52, Rad55, and Rad57 proteins. Proc. Natl. Acad. Sci. USA 92:6925-6929[Abstract/Free Full Text].

HAYS, S. L., A. A. FIRMENICH, P. MASSEY, R. BANERJEE, and P. BERG, 1998  Studies of the interaction between Rad52 protein and the yeast single-stranded DNA binding protein RPA. Mol. Cell. Biol. 18:4400-4406[Abstract/Free Full Text].

HO, K. S., 1975  The gene dosage effect of the rad52 mutation on X-ray survival curves of tetraploid yeast strains. Mutat. Res. 33:165-172[Medline].

HOEKSTRA, M. F., T. NAUGHTON, and R. E. MALONE, 1986  Properties of spontaneous mitotic recombination occurring in the presence of the rad52-1 mutation of Saccharomyces cerevisiae.. Genet. Res. 48:9-17[Medline].

HUANG, K. N. and L. S. SYMINGTON, 1994  Mutation of the gene encoding protein kinase C 1 stimulates mitotic recombination in Saccharomyces cerevisiae.. Mol. Cell. Biol. 14:6039-6045[Abstract/Free Full Text].

IVANOV, E. L., V. G. KOROLEV, and F. FABRE, 1992  XRS2, a DNA repair gene of Saccharomyces cerevisiae, is needed for meiotic recombination. Genetics 132:651-664[Abstract].

IVANOV, E. L., N. SUGAWARA, J. FISHMAN-LOBELL, and J. E. HABER, 1996  Genetic requirements for the single-strand annealing pathway of double-strand break repair in Saccharomyces cerevisiae.. Genetics 142:693-704[Abstract].

KAYTOR, M. D. and D. M. LIVINGSTON, 1994  Saccharomyces cerevisiae RAD52 alleles temperature-sensitive for the repair of DNA double-strand breaks. Genetics 137:933-944[Abstract].

KAYTOR, M. D. and D. M. LIVINGSTON, 1996  Allele-specific suppression of temperature-sensitive mutations of the Saccharomyces cerevisiae RAD52 gene. Curr. Genet. 29:203-210[Medline].

KAYTOR, M. D., M. NGUYEN, and D. M. LIVINGSTON, 1995  The complexity of the interaction between RAD52 and SRS2.. Genetics 140:1441-1442[Medline].

KLEIN, H. L., 1995  Genetic control of intrachromosomal recombination. Bioessays 17:147-159[Medline].

KLEIN, H. L., 1997  RDH54, a RAD54 homologue in Saccharomyces cerevisiae, is required for mitotic diploid-specific recombination and repair and for meiosis. Genetics 147:1533-1543[Abstract].

MALKOVA, A., E. L. IVANOV, and J. E. HABER, 1996  Double-strand break repair in the absence of RAD51 in yeast: a possible role for break-induced DNA replication. Proc. Natl. Acad. Sci. USA 93:7131-7136[Abstract/Free Full Text].

MALONE, R. E., B. A. MONTELONE, C. EDWARDS, K. CARNEY, and M. F. HOEKSTRA, 1988  A reexamination of the role of the RAD52 gene in spontaneous mitotic recombination. Curr. Genet. 14:211-223[Medline].

MILNE, G. T. and D. T. WEAVER, 1993  Dominant negative alleles of RAD52 reveal a DNA repair/recombination complex including Rad51 and Rad52. Genes Dev. 7:1755-1765[Abstract/Free Full Text].

MOREAU