Genetic Requirements for Spontaneous and Transcription-Stimulated Mitotic Recombination in Saccharomyces cerevisiae

Corresponding author: Sue Jinks-Robertson, 1510 Clifton Rd., Emory University, Atlanta, GA 30322., jinks{at}biology.emory.edu (E-mail)

Communicating editor: L. SYMINGTON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The genetic requirements for spontaneous and transcription-stimulated mitotic recombination were determined using a recombination system that employs heterochromosomal lys2 substrates that can recombine only by crossover or only by gene conversion. The substrates were fused either to a constitutive low-level promoter (pLYS) or to a highly inducible promoter (pGAL). In the case of the "conversion-only" substrates the use of heterologous promoters allowed either the donor or the recipient allele to be highly transcribed. Transcription of the donor allele stimulated gene conversions in rad50, rad51, rad54, and rad59 mutants, but not in rad52, rad55, and rad57 mutants. In contrast, transcription of the recipient allele stimulated gene conversions in rad50, rad51, rad54, rad55, rad57, and rad59 mutants, but not in rad52 mutants. Finally, transcription stimulated crossovers in rad50, rad54, and rad59 mutants, but not in rad51, rad52, rad55, and rad57 mutants. These data are considered in relation to previously proposed molecular mechanisms of transcription-stimulated recombination and in relation to the roles of the recombination proteins.

HOMOLOGOUS recombination repairs single-strand interruptions and double-strand breaks (DSBs) caused by DNA-damaging agents and provides a mechanism to reestablish collapsed replication forks (reviewed by KUZMINOV 1999 ; HOEIJMAKERS 2001 ). Recombination has been extensively studied using the yeast Saccharomyces cerevisiae and can be detected experimentally as a gene conversion and/or a crossover event (reviewed by PAQUES and HABER 1999 ). According to the widely accepted DSB repair model of recombination, broken ends are resected to give 3' single-stranded tails that invade an intact duplex molecule and are used to prime DNA synthesis. The invasion process yields regions of heteroduplex DNA that may contain mismatches, the repair of which results in gene conversion events. The interacting duplexes are ultimately joined by two Holliday junctions, which are resolved to yield either crossover or noncrossover products. In addition to the DSB repair model, recombination also has been proposed to occur by the alternative mechanisms of synthesis-dependent strand annealing (SDSA), break-induced replication (BIR), or single-strand annealing (SSA). Whereas SDSA can give rise to crossover and noncrossover products, BIR and SSA yield only crossovers.

Many proteins that participate in mitotic recombination in yeast have been identified, with most of the corresponding genes falling into the RAD52 epistasis group (reviewed by PAQUES and HABER 1999 ; SUNG et al. 2000 ). Rad51p is the functional homolog of the bacterial RecA strand-exchange protein (SHINOHARA et al. 1992 ; OGAWA et al. 1993 ; SUNG 1994 ); Rad55p and Rad57p form a stable heterodimer that stimulates the in vitro strand-exchange activity of Rad51p when the single-stranded binding protein Rpa is present (SUNG 1997B ); Rad52p likewise stimulates the strand-exchange activity of Rad51p when Rpa is present (SUNG 1997A ; NEW et al. 1998 ; SHINOHARA and OGAWA 1998 ; SONG and SUNG 2000 ), but also exhibits strand-annealing activity (MORTENSEN et al. 1996 ; SUGIYAMA et al. 1998 ); Rad54p has homology to the Swi2p/Mot1p family of chromatin remodeling helicases and stimulates homologous pairing, heteroduplex extension, and DNA remodeling in vitro (EISEN et al. 1995 ; PETUKHOVA et al. 1998 , PETUKHOVA et al. 1999B ; MAZIN et al. 2000 ; VAN KOMEN et al. 2000 ; SOLINGER and HEYER 2001 ); Rad59p has homology to Rad52p, possesses strand-annealing activity, and stimulates the strand-annealing activity of Rad52p (BAI and SYMINGTON 1996 ; PETUKHOVA et al. 1999A ; DAVIS and SYMINGTON 2001 ); and Rad50p, Mre11p, and Xrs2p form a complex that processes the ends of DSBs and promotes sister chromatid interactions (MALONE and ESPOSITO 1981 ; GOTTLIEB et al. 1989 ; MALONE et al. 1990 ; SUGAWARA and HABER 1992 ; IVANOV et al. 1994 ; TRUJILLO et al. 1998 ; TSUBOUCHI and OGAWA 1998 ; USUI et al. 1998 ). Rad52p appears to be required for all types of recombination in yeast, whereas the in vivo roles of the other proteins vary depending on the assay system used and the types of recombination mechanisms that can potentially occur (RATTRAY and SYMINGTON 1994 , RATTRAY and SYMINGTON 1995 ; LIEFSHITZ et al. 1995 ; SUGAWARA et al. 1995 ; JABLONOVICH et al. 1999 ; BARTSCH et al. 2000 ). Both physical and genetic interactions have been detected between members of the RAD52 epistasis group, leading to the idea that the proteins form a multiprotein complex for recombination (reviewed by PAQUES and HABER 1999 ; see also KREJCI et al. 2001 ; ESSERS et al. 2002 ).

Recombination potentially can be influenced by other DNA metabolic processes and, in particular, transcription has been shown to have a stimulatory effect on mitotic recombination in yeast (reviewed by AGUILERA 2001 ). The link between transcription and recombination was first established in studies of the recombination hotspot HOT1, which corresponds to the transcription initiation site of the 35S ribosomal RNA precursor and an enhancer of RNA polymerase I transcription (KEIL and ROEDER 1984 ; VOEKEL-MEIMAN et al. 1987 ; VOELKEL-MEIMAN and ROEDER 1990 ; ZEHFUS et al. 1990 ). A stimulatory effect of RNA polymerase II transcription on mitotic recombination was subsequently documented using gal10 chromosomal direct repeats constructed by integrating a plasmid at GAL10 (THOMAS and ROTHSTEIN 1989A , THOMAS and ROTHSTEIN 1989B ). Although recombination leading to plasmid loss was found to be stimulated when GAL10 was constitutively expressed, gene conversion was not elevated, indicating that transcription in this system primarily stimulated SSA rather than repair via the DSB repair model. A second demonstration of a relationship between RNA polymerase II transcription and mitotic recombination was provided by studies of recombination between Ty retrotransposons. Both deletional recombination events involving the Ty long terminal repeats (LTRs) as well as gene conversions between Ty's were stimulated when one of the elements was highly transcribed (NEVO-CASPI and KUPIEC 1994 ). Although the increased level of Ty recombination was shown to be dependent on Rad52p and Rad1p and independent of Rad50p, the high frequency of LTR/LTR recombination prevented determination of the requirements for other Rad proteins in transcription-associated gene conversion events (NEVO-CASPI and KUPIEC 1994 , NEVO-CASPI and KUPIEC 1996 ).

A recent study from this laboratory systematically examined RNA polymerase II transcription-stimulated mitotic recombination between lys2 repeats (SAXE et al. 2000 ). Significantly, transcription was found to stimulate recombination between substrates positioned on nonhomologous chromosomes as well as substrates positioned as direct repeats. With either type of repeat, both crossover and gene conversion events were elevated by transcription. In the current study, substrates positioned on nonhomologous chromosomes have been used to determine the genetic requirements for spontaneous vs. transcription-stimulated crossovers and gene conversions. These studies reveal distinct genetic requirements for spontaneous vs. transcription-stimulated recombination and provide novel insight into the roles of recombination proteins in the RAD52 epistasis group.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Media and growth conditions:
All yeast strains were grown at 30° unless otherwise indicated. Yeast strains were grown nonselectively in YEP medium (1% yeast extract, 2% Bacto-peptone with 2% agar for plates) supplemented with 2% dextrose (YEPD) or 2% glycerol/2% ethanol (YEPGE). For selective growth, SC medium (SHERMAN 1991 ) missing the relevant nutrient was supplemented with 2% dextrose (SCD) or 2% glycerol/2% galactose/2% ethanol (SCGGE). Ura^- derivatives of yeast strains were selected on SCD plates containing 1 g/liter of 5-fluoroorotic acid (5-FOA; BOEKE et al. 1987 ). Sensitivity of yeast transformants to methylmethanesulfonate (MMS) was determined using YEPD plates containing 0.016% MMS (Kodak). Kanamycin-resistant transformants were selected by growing cells overnight on YEPD and then replica plating to YEPD containing 200 mg/liter of geneticin (Sigma).

Plasmid constructions:
Plasmids pSR183 and pSR234 contain the lys23' allele fused to the LYS2 (pLYS) and the GAL1-10 (pGAL) promoters, respectively. In both plasmids the URA3 gene is positioned at the 3' end of the lys23' allele and is transcribed in the same direction as the LYS2 sequences. Plasmids pSR184 and pSR235 are identical to plasmids pSR183 and pSR234, respectively, but URA3 and LYS2 sequences are transcribed in opposite directions. Plasmid pSR136 contains the rad52::hisG-URA3-hisG allele and was constructed by ligating a 2-kb EcoRI-BamHI RAD52 fragment from plasmid YRp7-RAD52-A4-Sal (SCHILD et al. 1983 ) to EcoRI-BamHI-digested pUC9. The 3.8-kb BamHI-BglII hisG-URA3-hisG cassette from plasmid pNKY51 (ALANI et al. 1987 ) was then inserted into the BglII site within the RAD52 sequences. Plasmid pRDK713 was constructed by inserting 25 bp (5'-GATCTGTCCCTTACTAGCTAGGTAG-3' on the coding strand) into the unique BglII site within LYS2 (A. SHOEMAKER and R. D. KOLODNER, unpublished results). The 25-bp insertion causes a frameshift mutation and creates stop codons in all three reading frames.

Yeast strain constructions:
All yeast strains used in this study are isogenic derivatives of SJR195 (MAT ade2-101_oc his3200 ura3Nco) and were constructed using a lithium acetate transformation protocol (GIETZ and SCHIESTL 1995 ). SJR733 and SJR734 are Gal80⁺ and Gal80^- strains, respectively, that contain the pLYS-lys25'/pGAL-lys23' crossover-only (C/O) recombination substrates (Fig 1; for construction details see SAXE et al. 2000 ). SJR1305 and SJR1306 are Ura^- Leu^- derivatives of SJR733 and SJR734, respectively, constructed by first selecting a Ura^- derivative on 5-FOA medium and then introducing the leu2-R allele by two-step allele replacement using KpnI-digested pJH189 (LICHTEN et al. 1987 ).

View larger version (27K): In this window In a new window Download PPT slide   Figure 1. The recombination substrates. Open boxes depict LYS2 sequences; asterisks depict the lys2-oligo allele; solid and hatched boxes depict pLYS and pGAL, respectively; open and shaded lines depict chromosome II and V sequences, respectively; and circles depict centromeres. C/O, crossover; GC, gene conversion. (A) Crossover-only substrates, C/O system, are shown. Because experiments were done in haploid strains, recombinants containing only one translocation chromosome would be inviable due to a deficiency in genetic material. (B) Gene conversion-only substrates with donor highly transcribed, GCD system, are shown. (C) Gene conversion-only substrates with recipient highly transcribed, GCR system, are shown. In the GCD and GCR systems, a crossover between chromosomes II and V will yield dicentric and acentric translocation chromosomes, resulting in cell death.

Yeast strains SJR1599 and SJR1583 (Gal80⁺ and Gal80^-, respectively) contain the GCD (gene conversion with donor highly transcribed) gene conversion substrates with the pGAL donor and pLYS recipient alleles and were constructed as follows. First SJR1503 and SJR1504, Lys⁺ derivatives of SJR357 (MAT ade2-101_oc his3200 ura3Nco lys2Bgl) and SJR358 (MAT ade2-101_oc his3200 ura3Nco lys2Bgl gal80::HIS3), respectively, were made by transformation with EcoRV-digested pDP6 (FLEIG et al. 1986 ). SJR1516 and SJR1517, leu2-R derivatives of SJR1503 and SJR1504, respectively, were then made by two-step allele replacement using KpnI-digested pJH189 (LICHTEN et al. 1987 ). The LYS2 allele of SJR1516 and SJR1517 was replaced with the lys2-oligo allele by two-step allele replacement using XhoI-digested pRDK713. The pGAL-lys23' allele was next introduced at the URA3 locus by transformation with SmaI-digested pSR235. Following selection of Ura⁺ transformants, integration of a single copy of the plasmid at the URA3 locus was confirmed by Southern blot analysis or PCR. Finally, Ura^- derivatives were selected on 5-FOA medium.

Yeast strains SJR1584 and SJR1582 (Gal80⁺ and Gal80^-, respectively) contain the GCR (gene conversion substrates with recipient highly transcribed) gene conversion substrates with the pLYS-lys23' donor and pGAL-lys2-oligo recipient alleles. These strains were constructed as described above, except for starting with strains containing the pGAL-LYS2 allele (for a description see SAXE et al. 2000 ) and using SmaI-digested pSR184 to introduce the pLYS-lys23' allele.

rad50, rad55, and rad59 derivatives of SJR1305, SJR1306, SJR1599, SJR1583, SJR1584, and SJR1582 were constructed using a PCR-based gene disruption methodology (WACH et al. 1994 ). PCR fragments containing the bacterial kan gene or the yeast LEU2 gene flanked by 60 bp of appropriate yeast sequences were obtained using either pFA6-kanMX2 or pUC7-LEU2 as a template, respectively. Following transformation, geneticin-resistant or Leu⁺ yeast transformants were selected as appropriate and the disruption was confirmed by MMS sensitivity and by PCR.

rad51, rad52, rad54, and rad57 derivatives of SJR1305, SJR1306, SJR1599, SJR1583, SJR1584, and SJR1582 were constructed by transformation with BamHI-digested pSR464 (rad51::URA3; ABOUSSEKHRA et al. 1992 ), BamHI-EcoRI-digested pSR136 (rad52::hisG-URA3-hisG), BglII-digested pSR465 (rad54::URA3; obtained from L. Symington), and SacI-digested pSR478 (rad57::LEU2; obtained from D. Schild), respectively. Following selection of Ura⁺ or Leu⁺ transformants as appropriate, disruptions were confirmed by MMS sensitivity and PCR. A complete list of the wild-type and rad mutant strains used to measure recombination rates is given in Table 1.

Analysis of recombinants:
Approximately 10 recombinants from each strain were analyzed to confirm the nature of the underlying recombination event. For the C/O substrates, presence of the lys25'3' allele in Lys⁺ recombinants was considered diagnostic of a true reciprocal exchange, and PCR was used to detect this allele. Under high-transcription conditions, all recombinants were generated by a reciprocal exchange event; under low-transcription conditions, at least 8/10 recombinants were generated by a reciprocal exchange event. In contrast to the results obtained with the other mutants, a significant difference in the nature of the recombinants under low- vs. high-transcription conditions was detected in the rad54 strain. In this strain only 56% (14/25) of the recombinants under low-transcription conditions result from a reciprocal exchange event, while 90% (28/31) of the recombinants under high-transcription conditions result from a reciprocal exchange event (P = 0.008 by contingency chi-square analysis). Recombinants generated by a reciprocal exchange event also could be detected in rad52 strains, as 2/6 low-transcription recombinants and 3/3 high-transcription recombinants resulted from a reciprocal exchange event. Because a disruption rather than a deletion allele was used, it is possible that there is residual Rad52p activity in these strains.

In the GCD and GCR gene conversion systems, phenotypic analysis was used to determine which promoter the LYS2 allele was fused to. Strains with the pLYS-LYS2 allele grow on lysine-deficient medium containing either glucose or galactose as a carbon source, while strains containing the pGAL-LYS2 allele grow only on galactose-containing medium. A gene conversion event between the GCD substrates should result in the LYS2 allele fused to pLYS whereas either a crossover event or a break-induced replication event should result in the LYS2 allele fused to pGAL (see Fig 1). In strains containing the GCR substrates, the opposite promoter configuration would be expected. With the GCD strains, results were consistent with a gene conversion mechanism in all recombinants under both low- and high-transcription conditions. In the strains containing the GCR system, under both low- and high-transcription conditions, the results were consistent with a gene conversion event generating at least 8/10 of the recombinants analyzed.

Determination of recombination rates:
Recombination rates were determined by the method of the median (LEA and COULSON 1949 ). Yeast strains were grown on YEPD plates for 3 days, independent colonies were inoculated into 5 ml of YEPGE, and cultures were grown for 3 days on a roller drum. Cells were washed with 5 ml of sterile H₂O and resuspended in 1 ml of sterile H₂O. Aliquots (100 µl) of appropriate dilutions were plated onto YEPD plates to determine the number of viable cells in each culture and onto SCGGE-Lys plates to determine the total number of recombinants in each culture. Colonies on YEPD plates were counted after 2 days and those on SCGGE-Lys plates after 3 days. The average number of viable cells and the median number of recombinants for each yeast strain were determined from a minimum of 12 cultures. Recombination rates are considered to be statistically different if the 95% confidence intervals do not overlap (DIXON and MASSEY 1969 ).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The recombination systems:
The recombination systems used in this study consist of RNA polymerase II-transcribed substrates positioned on nonhomologous chromosomes in haploid strains (Fig 1). In all systems, one substrate was fused to the constitutive low-level LYS2 promoter (pLYS) while the other substrate was fused to the highly inducible GAL1-10 promoter (pGAL). The crossover-only recombination substrates (C/O substrates) consist of 5' and 3' truncated lys2 alleles, with the promoter configuration being the one that was reported previously to yield the highest level of transcription-stimulated recombination (SAXE et al. 2000 ). The gene conversion-only recombination substrates consist of full-length and truncated lys2 alleles, with the alleles positioned in opposite orientations relative to the centromeres of their chromosomes to preclude the production of viable crossover Lys⁺ recombinants. Because only the full-length lys2 allele (lys2-oligo) can become prototrophic in this conversion-specific assay, this allele will hereafter be referred to as the recipient allele and the truncated lys2 allele as the donor allele. Fusion of these substrates to different promoters allows the genetic requirements for transcription-stimulated gene conversion to be analyzed when the donor (GCD substrates) vs. the recipient (GCR substrates) allele is highly transcribed.

Transcriptional activity of pGAL was regulated by growing isogenic Gal80⁺ and Gal80^- strains under noninducing, nonrepressing conditions (glycerol/ethanol as carbon sources in the absence of galactose). Under these conditions the Gal80^- strains highly transcribe the recombination substrate fused to pGAL (high-transcription conditions) while the Gal80⁺ strains, which contain the negative regulatory protein Gal80p, transcribe the recombination substrate that is fused to pGAL at a very low level (low-transcription conditions). The transcriptional activity of pGAL has been shown to increase 1000-fold under high-transcription conditions relative to low-transcription conditions (SAXE et al. 2000 ). Spontaneous and transcription-stimulated crossovers and gene conversions were examined in wild-type (RAD), rad50, rad51, rad52, rad54, rad55, rad57, and rad59 strains grown at 30°. Physical and phenotypic analyses of recombinants were consistent with the vast majority of crossovers corresponding to reciprocal exchange events and the vast majority of gene conversions resulting from the transfer of genetic information from the truncated to the full-length allele (see MATERIALS AND METHODS for details).

Genetic requirements for spontaneous and transcription-stimulated crossovers:
The crossover rate data for the RAD and rad mutant strains containing the C/O system are presented in Table 2. Under low-transcription conditions the crossover rate is reduced 10-fold in the rad51, rad52, and rad54 strains relative to the RAD strain. A less severe reduction is observed in rad55, rad57, and rad59 strains, with the crossover rate decreasing only 3-fold relative to the RAD strain. In contrast to the other mutants, a rad50 strain exhibits a hyperrecombinational phenotype under low-transcription conditions, with the crossover rate increasing 3.5-fold.

In the RAD strains, the crossover rate under high-transcription conditions is elevated 13-fold relative to the rate under low-transcription conditions. In contrast, the crossover rates in the rad51, rad52, rad55, and rad57 strains under high-transcription conditions do not increase significantly over the rates observed under low-transcription conditions. Although the rad54 strain exhibits a small (2.3-fold) increase in the crossover rate under high-transcription conditions, this increase is much less than the 13-fold increase observed in the RAD strain. The rad59 strain also exhibits a significant increase in crossover rate under high-transcription conditions, with crossovers increasing 24-fold. In this case the magnitude of the increase is greater than that observed in the RAD strain, suggesting either that transcription-stimulated crossovers are less dependent than spontaneous crossovers on RAD59 or that transcription partially compensates for loss of RAD59. In the rad50 mutant under high-transcription conditions, the crossover rate increases 4.6-fold relative to low-transcription conditions. The recombination rate under high-transcription conditions in the rad50 mutant is not statistically different from the recombination rate under high-transcription conditions in the RAD50 control strain.

Genetic requirements for spontaneous and transcription-stimulated gene conversions when the donor allele is highly transcribed (GCD system):
The spontaneous and transcription-stimulated gene conversion rates in RAD and rad mutant strains containing the GCD substrates are presented in Table 3. The relative phenotypes of the mutants under low-transcription conditions with respect to gene conversions are similar to those with respect to crossovers, with rad51, rad52, and rad54 mutants exhibiting the most severe phenotypes (at least a 30-fold decrease in recombination rate); rad55, rad57, and rad59 exhibiting weaker phenotypes (an 2-fold decrease); and the rad50 mutant exhibiting a hyperrecombinational phenotype (a 17-fold increase). We were unable to obtain an accurate estimate of the gene conversion rate in rad51, rad52, and rad54 strains because the Lys⁺ rates are not above the background reversion rate of the lys2-oligo allele (see Table 3 legend).

Under high-transcription conditions the gene conversion rate in the RAD strain increases 6.8-fold relative to the rate under low-transcription conditions. Similarly, rad51, rad54, and rad59 strains exhibit significant increases in gene conversions under high-transcription conditions relative to low-transcription conditions. With respect to rad51 mutants, it should be noted that transcription has a different impact on gene conversions vs. crossovers, with transcription stimulating only gene conversion events. In contrast to the rad51, rad54, and rad59 mutants, rad55 and rad57 strains exhibit a significant 2-fold decrease in gene conversion rate under high-transcription conditions relative to low-transcription conditions. No increase in the gene conversion rate in the rad52 strain under high-transcription conditions relative to low-transcription conditions is evident. In a rad50 mutant, a slight but significant rate increase under high-transcription relative to low-transcription conditions is observed. This increase is consistent with an additive relationship between transcription and disruption of RAD50 on gene conversions, with disruption of RAD50 resulting in a 17-fold increase in gene conversions, transcription resulting in a 6.8-fold increase in gene conversions, and both transcription and disruption of RAD50 resulting in a 22-fold increase in gene conversions.

Genetic requirements for spontaneous and transcription-stimulated gene conversions when the recipient allele is highly transcribed (GCR system):
Spontaneous and transcription-stimulated gene conversion rates in the RAD and rad mutant strains containing the GCR system are presented in Table 4. Again, as in the C/O and the GCD systems, rad51, rad52, and rad54 mutants exhibit the most severe phenotypes (at least an 18-fold decrease in recombination rate); rad55, rad57, and rad59 exhibit weaker phenotypes (a 2- to 4-fold decrease); and a rad50 mutant exhibits a hyperrecombinational phenotype (an 8.4-fold increase).

Under high-transcription conditions the Lys⁺ gene conversion rate in the RAD strain increases 18-fold relative to the rate under low-transcription conditions. rad55, rad57, and rad59 strains exhibit a transcription-associated increase in gene conversions comparable to that observed in the RAD strain. The increase is particularly striking in the rad55 and rad57 mutants and directly contrasts with the results obtained with the GCD system where high levels of transcription inhibited recombination. Although the rad51 and rad54 strains also exhibit a significant increase in gene conversions under high-transcription conditions (5.5- and 3.5-fold, respectively), the increase is less than that observed in the RAD, rad55, rad57, and rad59 strains. The Lys⁺ rate in the rad52 strain under high-transcription conditions does not increase significantly over the rate in the rad52 strain under low-transcription conditions and likely reflects reversion of the lys2-oligo allele. In the rad50 mutant, under high-transcription conditions, gene conversions increase 3.6-fold relative to low-transcription conditions. As with the GCD system, the data obtained with the GCR system are consistent with an additive relationship between transcription and disruption of RAD50.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The genetic requirements for spontaneous and transcription-stimulated mitotic crossover and gene conversion have been determined in order to investigate the basis for transcription-stimulated recombination and to further explore the in vivo roles of the recombination proteins. The recombination systems used here have several features that should be noted. First, because each system detects only crossover or only gene conversion events, a differential impact of transcription on crossovers vs. gene conversions can be discerned. Second, the positioning of the recombination substrates on nonhomologous chromosomes avoids some of the mechanistic ambiguities associated with other types of recombination systems, such as direct and inverted repeats (reviewed by PAQUES and HABER 1999 ; see also BARTSCH et al. 2000 ; MALAGON and AGUILERA 2001 ). Indeed, physical analyses indicated that the majority of recombinants obtained with the C/O system corresponded to true reciprocal exchange events, consistent with these events arising via the resolution of a Holliday junction intermediate rather than by a nonreciprocal mechanism such as BIR. Similarly, recombinants generated in the GCD and GCR systems correspond to gene conversion events, which could, in principle, occur via either the DSB repair or the SDSA model. In the discussion that follows, the recipient and donor alleles in gene conversion events are defined according to current models of recombination. The recipient thus is assumed to be the molecule within which the initiating DSB occurs and the nucleoprotein filament is formed, whereas the donor is assumed to be the molecule that is invaded by the nucleoprotein filament. Although this donor/recipient distinction appears to be correct in the majority of gene conversion events, transformation experiments have demonstrated that a broken DNA molecule can also donate information at a low frequency (ROITGRUND et al. 1993 ). Finally, in the systems used here recombination occurs between chromosomes rather than between a plasmid and a chromosome. Plasmid/chromosome vs. chromosome/chromosome recombination could proceed by different mechanisms and plasmid vs. chromosome chromatin structure may differ, either of which could alter genetic requirements in plasmid/chromosome systems (SUGAWARA et al. 1995 ).

A summary of the recombination rate data obtained with the C/O, GCD, and GCR substrates in the wild-type and RAD52 epistasis group mutants is presented graphically in Fig 2 and the results are discussed in detail below. The major observations are summarized as follows: (i) the effects of RAD50 disruption and transcription on gene conversion were additive; (ii) transcription stimulated gene conversions, but failed to stimulate crossovers in rad51 mutants; (iii) like spontaneous events, transcription-associated events were completely dependent on Rad52p; (iv) transcription stimulated all types of recombination in rad54 mutants, although not to the same extent as in wild-type strains; (v) Rad55p/Rad57p were required for transcription-stimulated crossover and for gene conversion when the donor allele was highly transcribed, but were not required for gene conversion when the recipient allele was highly transcribed; and (vi) transcription stimulated all recombination in rad59 mutants to at least the same extent as in wild-type strains.

View larger version (42K): In this window In a new window Download PPT slide   Figure 2. Relative recombination rates in the wild-type and RAD52 epistasis group mutants containing the C/O system, GCD system, or GCR system. Rates in each substrate system are normalized to the rate of the wild-type, low-transcription strain. The data obtained in the rad57 mutants are not shown, but are similar to that obtained in the rad55 mutants. Solid bars: wild-type strains, data obtained under low-transcription conditions. Shaded bars: mutant strains, data obtained under low-transcription conditions. Black hatched bars: wild-type strains, data obtained under high-transcription conditions. Gray hatched bars: mutant strains, data obtained under high-transcription conditions. Open bars: included with the GCR/GCD data, the reversion rate of the pGAL-lys2-oligo allele in a rad52, high-transcription background normalized to the recombination rate in the wild-type, low-transcription strain. Error bars depict 95% confidence intervals. The width of the solid bars depicts the 95% confidence interval for the wild-type, low-transcription strains.

Spontaneous gene conversions and crossovers in wild type vs. RAD52 epistasis group mutants:
Gene conversions occurred at a rate three- to sixfold higher than crossovers in the wild-type strain under low-transcription conditions. This difference is consistent with the general observation that gene conversions are favored in mitosis (ESPOSITO 1978 ; HABER and HEARN 1985 ; KUPIEC and PETES 1988 ), but also may reflect the difference in the length of substrate overlap in the C/O vs. the GCD/GCR systems (1.9 vs. 3.3 kb, respectively; JINKS-ROBERTSON et al. 1993 ). The relative phenotypes of the RAD52 epistasis group mutants under low-transcription conditions were similar in the C/O and GCD/GCR systems, with rad51, rad52, and rad54 mutants exhibiting the most severe recombination defects; rad55, rad57, and rad59 mutants exhibiting weaker defects; and rad50 mutants exhibiting hyperrecombinational phenotypes. As seen in previous studies, rad55 and rad57 mutants exhibited similar phenotypes (LOVETT and MORTIMER 1987 ; HAYS et al. 1995 ; JOHNSON and SYMINGTON 1995 ; RATTRAY and SYMINGTON 1995 ; SUGAWARA et al. 1995 ; IVANOV et al. 1996 ; BARTSCH et al. 2000 ; SIGNON et al. 2001 ), which is consistent with these proteins functioning as a complex (SUNG 1997B ). Although previous studies generally have found rad55 and rad57 mutants to be strongly cold sensitive for defects in mitotic recombination (LOVETT and MORTIMER 1987 ; HAYS et al. 1995 ; JOHNSON and SYMINGTON 1995 ; LIEFSHITZ et al. 1995 ; RATTRAY and SYMINGTON 1995 ; BARTSCH et al. 2000 ), disruption of RAD55 or RAD57 in our strains resulted in little or no decrease in recombination at either 30° or 23° (data not shown). It should be noted that a similar lack of cold sensitivity was reported when spontaneous gene conversion between nonhomologous chromosomes was assayed (LIEFSHITZ et al. 1995 ). The phenotypes of rad55/rad57 mutants thus could be related to the type of assay system used or could reflect strain background differences.

Many diverse recombination systems have been used to study the roles of Rad proteins, but there have been few systematic studies that allow direct comparisons between mutants. The most comprehensive study of this sort examined the roles of the RAD52 epistasis group genes in mitotic gene conversion and crossover between inverted repeats (IRs; RATTRAY and SYMINGTON 1994 , RATTRAY and SYMINGTON 1995 ; BAI and SYMINGTON 1996 ). In the IR system rad52 mutants exhibited the most severe recombination defects; rad54, rad55, and rad57 mutants exhibited less severe defects; and rad51, rad59, and rad50 mutants exhibited the weakest defects. This is in stark contrast to our system, where rad51 mutants had a greater recombination defect than did rad55/rad57 mutants and where rad50 mutants were hyperrecombinational. In addition, gene conversion was more dependent on RAD51 than was crossover in the IR system. Although we cannot differentiate roles for Rad51p in spontaneous gene conversion vs. crossover in our system (gene conversion rates were not above the background reversion rates), this protein seems to be required more for crossover than for gene conversion under high-transcription conditions (see below). Consistent with our results, studies using a gap repair assay also suggest that crossovers are dependent on RAD51 (BARTSCH et al. 2000 ). The conflicting results obtained with the IR system likely reflect the different mechanisms by which recombinants are generated. Specifically, a recent study has indicated that recombination in the IR system occurs predominantly by a RAD51-independent pathway involving break-induced replication followed by single-strand annealing (KANG and SYMINGTON 2000 ).

rad50 mutants generally are defective in sister chromatid recombination, but exhibit hyperrecombinational phenotypes with respect to recombination between nonsisters (MALONE and ESPOSITO 1981 ; GOTTLIEB et al. 1989 ; MALONE et al. 1990 ). The conflicting results obtained with the rad50 mutants in our system vs. the IR system thus likely reflect the role of Rad50p in specifically promoting sister chromatid interactions (reviewed by PAQUES and HABER 1999 ). Alternatively, the hyperrecombinational phenotype of rad50 mutants could be a consequence of the shorter 3' single-stranded tails in rad50 mutants (see PAQUES and HABER 1999 ). In recombination assays in which conversion of only one of two mutant sites generates a selectable recombinant, shorter tails would be expected to result in the generation of more selectable recombinants. Our results are more consistent with the idea that the hyperrecombinational phenotype of rad50 mutants reflects the role of Rad50p in promoting sister chromatid interactions, as rad50 mutants containing the C/O system (which has no mutant sites that need to undergo gene conversion) also exhibited a hyperrecombinational phenotype.

The relative roles of the RAD52 epistasis group genes in spontaneous gene conversions between lys2 heteroalleles positioned on nonhomologous chromosomes has been examined previously (LIEFSHITZ et al. 1995 ; JABLONOVICH et al. 1999 ). While the low-transcription results reported here agree well with the earlier results, it should be noted that our results extend the previous analysis by examining spontaneous crossovers as well. Interestingly, our results as well as those of JABLONOVICH et al. 1999 suggest a role for Rad59p in recombination between nonhomologous chromosomes in haploids, whereas recombination between homologous chromosomes in diploids has been reported to increase in a rad59/rad59 mutant (BAI and SYMINGTON 1996 ). These conflicting results may reflect differences between recombination in haploids vs. diploids (e.g., ploidy, MAT status, or cell-cycle effects), differences between allelic vs. ectopic recombination, or strain background differences.

Transcription-stimulated gene conversions and crossovers in wild-type strains:
The crossover and gene conversion rates in the RAD strains increased 7- to 17-fold under high-transcription conditions. The weak asymmetry reported here in the GCD vs. GCR systems is in stark contrast to results obtained in our previous study, which indicated that transcription stimulates recombination only when the recipient allele is highly transcribed (SAXE et al. 2000 ). Much of the asymmetry observed in the earlier study appears to be due to a transcription-associated bias in the mismatch repair system (J. FREEDMAN and S. JINKS-ROBERTSON, unpublished data). The weak asymmetry that does persist, however, is consistent with the idea that one effect of transcription is to increase the number of recombination-initiating lesions. Recent experiments examining plasmid-chromosome gene conversion or recombination between direct/ inverted repeats are also consistent with transcription increasing the number of initiating lesions (S. GONZÁLEZ-BARRERA, M. GARCÍA-RUBIO and A. AGUILERA, personal communication). It should be noted, however, that if this were the only effect of transcription, high transcription of the donor allele in our system would be expected to have little, if any, effect on the overall gene conversion rate. The fact that transcription of the donor has such a large effect implicates additional effects of transcription (see discussion of rad mutants below).

Transcription-stimulated gene conversions and crossovers in RAD52 epistasis group mutants:
Several general models have been proposed to account for the stimulatory effect of transcription on mitotic recombination (reviewed by AGUILERA 2001 ). The first model proposes that transcription stimulates recombination by increasing the number of recombination-initiating lesions. Such lesions could result from enhanced susceptibility of transcriptionally active DNA to nucleases or other damaging agents, transcription-associated changes in supercoiling, transcription elongation blocks, or transcriptional interference with replication fork passage. Consistent with the notion that high levels of transcription enhance DNA damage, transcription also has been found to stimulate mutation rates in yeast (DATTA and JINKS-ROBERTSON 1995 ; MOREY et al. 2000 ). It should be noted that if transcription stimulates mitotic recombination solely by increasing the number of recombination-initiating lesions, one would expect the magnitude of the transcription-stimulated recombination to be the same in wild-type and rad mutant strains. A second general model for transcription-stimulated recombination suggests that the disruption of chromatin structure during transcription increases the efficiency of one or more individual steps in the recombination process. According to this model, one might predict that high levels of transcription would relax the genetic requirements for one or more Rad proteins. A previous study examining HO-induced mating-type switching led to the suggestion that Rad51p, Rad54p, Rad55p, and Rad57p indeed specifically play a role in facilitating strand invasion into otherwise inaccessible chromosomal donor sequences (SUGAWARA et al. 1995 ). The results of a subsequent study have demonstrated, however, that gap repair is similarly dependent on RAD51, RAD52, RAD57, and RAD59 when the donor is plasmid borne vs. chromosomal (BARTSCH et al. 2000 ). Finally, recombination can occur by multiple pathways with different genetic requirements, raising the possibility that transcription may stimulate only one of several pathways. This leads to the prediction that the genetic requirements for spontaneous vs. transcription-stimulated recombination should differ in a manner reflecting the pathway(s) used. As is evident in Fig 2, the effects of transcription on recombination varied widely depending on the assay system used and the member of the RAD52 epistasis group that was eliminated. The results obtained in each rad mutant will be discussed individually below. Taken together, the data do not conform to the simplistic predictions made above if one assumes that transcription only increases the number of recombination-initiating lesions or that transcription functionally substitutes for only some of the Rad proteins. Instead, it is likely that transcription impacts recombination at several different levels.

RAD50 encodes a protein that plays a role in promoting sister chromatid interactions and in processing the ends of double-strand breaks (reviewed by PAQUES and HABER 1999 ). Although the crossover rate in the rad50 mutant increased under high-transcription conditions, it did not differ significantly from the rate in the wild-type strain under these same conditions. With respect to gene conversions, however, there appeared to be an additive relationship between disruption of RAD50 and high levels of transcription. These additive relationships suggest that RAD50 disruption vs. high levels of transcription increase recombination by different and nonoverlapping mechanisms.

RAD51 encodes the functional homolog of the RecA strand-exchange protein (SHINOHARA et al. 1992 ; OGAWA et al. 1993 ; SUNG 1994 ). Although transcription stimulated gene conversions in rad51 mutants, it had no effect on the crossover rate. This disparity may reflect the different mechanisms by which reciprocal crossover vs. gene conversion events can arise (see PAQUES and HABER 1999 ). While gene conversions can be generated by mechanisms that do not involve the formation of a Holliday junction (e.g., SDSA), reciprocal crossovers require the formation, stabilization, and resolution of a Holliday junction(s). We suggest that the stringent requirement for Rad51p for crossover under high-transcription conditions may reflect a role for Rad51p in stabilizing the single-end invasion intermediate, thus allowing the other end of the DSB to be engaged and a Holliday junction to be formed. It is possible that the stimulation of gene conversions in rad51 mutants reflects the lack of a requirement for Rad51p for SDSA or reflects a role of transcription in facilitating the pairing and strand-exchange reactions during SDSA. Although the genetic requirements for SDSA have not been reported, they would be expected to be similar to those for BIR, which is RAD51 independent (MALKOVA et al. 1996 ; KRAUS et al. 2001 ; SIGNON et al. 2001 ). It should be noted that transcription of the donor allele resulted in a greater stimulation of gene conversions than did transcription of the recipient allele in rad51 mutants relative to wild-type strains. It is possible that transcription of the donor may give single-stranded character to the invaded duplex, thus increasing the efficiency of gene conversions arising via a RAD51-independent, RAD52-dependent strand-annealing pathway (KANG and SYMINGTON 2000 ).

Previous studies have shown that rad52 mutants exhibit the most severe recombination defects (reviewed by PAQUES and HABER 1999 ), which is thought to reflect the requirement for Rad52p in all pathways of recombination in yeast. Consistent with its participation in multiple pathways, Rad52p has strand-annealing activity in vitro (MORTENSEN et al. 1996 ; SUGIYAMA et al. 1998 ) and, like Rad55p/Rad57p, it stimulates the strand exchange activity of Rad51p when Rpa is present (SUNG 1997A ; NEW et al. 1998 ; SHINOHARA and OGAWA 1998 ; SONG and SUNG 2000 ). Transcription-stimulated crossovers and gene conversions were dependent on RAD52, which may reflect a requirement for Rad52p-mediated strand annealing in DSB repair as well as in SDSA. In DSB repair the protein may be required for capture of the second end of the DSB, which corresponds to a strand-annealing rather than a strand-invasion reaction. The Rad52p homolog, Rad59p, also has strand-annealing activity and recently was reported to stimulate the annealing activity of Rad52p (BAI and SYMINGTON 1996 ; PETUKHOVA et al. 1999A ; DAVIS and SYMINGTON 2001 ). Under high-transcription conditions, both crossovers and gene conversions were stimulated in rad59 mutants to at least the same extent as in the wild-type strains, consistent with the idea of transcription increasing the occurrence of recombination-initiating lesions. Alternatively, it is possible that transcription facilitates the Rad52p-mediated strand-annealing reactions during DSB repair and SDSA, thereby alleviating the requirement for Rad59p in this process.

RAD54 encodes a protein homologous to the Swi2p/Mot1p family of chromatin remodeling helicases (EISEN et al. 1995 ). In vitro, Rad54p has DNA remodeling activity and has been shown to stimulate homologous pairing as well as heteroduplex extension (PETUKHOVA et al. 1998 , PETUKHOVA et al. 1999B ; MAZIN et al. 2000 ; VAN KOMEN et al. 2000 ; SOLINGER and HEYER 2001 ). Transcription significantly stimulated both crossovers and gene conversions in rad54 mutants, although not to the same extent as observed in wild-type strains. The stimulation of recombination in rad54 mutants may reflect a transcription-associated change in chromatin structure or supercoiling, either of which could partially relieve the requirement for Rad54p in creating a favorable chromosomal context within which the homology search as well as later heteroduplex extension can occur. In rad54 mutants, it should be noted that when the donor allele was highly transcribed the high-transcription recombination rate/low-transcription recombination rate was at least 50% of the wild-type ratio whereas when the recipient allele was highly transcribed the high-transcription recombination rate/low-transcription recombination rate was only 20% of the wild-type ratio. This result is consistent with the suggestion that Rad54p plays a role in strand separation/chromatin remodeling of donor sequences.

The products of the RAD55 and RAD57 genes form a stable heterodimer that stimulates the strand-exchange activity of Rad51p when Rpa is present, suggesting a role for Rad55p/Rad57p in displacing Rpa and facilitating the loading of Rad51p onto single-stranded tails during nucleoprotein filament formation (SUNG 1997B ). Transcription of the recipient allele (GCR system) stimulated gene conversions in rad55/rad57 mutants to the same extent as in the wild-type strain, consistent with the idea of transcription increasing the occurrence of recombination-initiating lesions. Alternatively, it is possible that the association of Rad51p with the RNA polymerase II holoenzyme complex (MALDONADO et al. 1996 ) facilitates targeting of Rad51p to the recipient allele and thereby alleviates the requirement for Rad55p/Rad57p to appropriately load Rad51p. Finally, it is possible that the transcription machinery may participate in the displacement of proteins associated with the recipient locus, thereby facilitating nucleoprotein filament formation in the absence of Rad55p/Rad57p. In striking contrast to the stimulatory effect of transcription observed with the GCR substrates, high transcription of the donor allele failed to stimulate gene conversions in rad55/rad57 mutants. Although Rad55p/Rad57p has been implicated only in the initial nucleoprotein filament formation, our data suggest that the complex may remain associated with the Rad51p nucleoprotein filament and could play a subsequent role in displacing proteins (e.g., the transcription machinery) associated with the donor duplex molecule, thereby facilitating strand invasion and/or heteroduplex extension. As observed with the GCD substrates, transcription did not stimulate crossovers, which again might reflect a role of Rad55p/Rad57p in displacing the transcription machinery, thereby allowing the other end of the DSB to be engaged and a Holliday junction to be formed.

In summary, the use of highly transcribed recombination substrates has revealed clear differences between rad mutants that exhibit similar phenotypes in spontaneous recombination assays. The experiments reported here demonstrate that the genetic requirements for spontaneous vs. transcription-stimulated crossovers and gene conversions are different, which may reflect the distinct mechanisms for generating these alternative products of recombination. Furthermore, with respect to transcription-stimulated gene conversions, the genetic requirements differ depending on whether the recipient vs. the donor allele is highly transcribed. Finally, the relationship between transcription and recombination appears to be complex, with transcription stimulating recombination not only by increasing the number of recombination-initiating lesions, but also by facilitating subsequent steps of the recombination process.

	ACKNOWLEDGMENTS

We thank Alex Shoemaker and Richard Kolodner for the generous gift of the plasmid containing the lys2-oligo allele without which many of these experiments would not have been possible. We thank Patrick Sung for helpful discussions and members of the S.J.-R. laboratory for valuable input throughout the course of this work and for comments on the manuscript. This work was supported by National Institutes of Health grant GM-38464 to S.J.-R. J.A.F. was partially supported by the Emory University Graduate Division of Biological and Biomedical Sciences.

Manuscript received April 5, 2002; Accepted for publication May 10, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ABOUSSEKHRA, A., R. CHANET, A. ADJIRI, and F. FABRE, 1992  Semidominant suppressors of Srs2 helicase mutations of Saccharomyces cerevisiae map in the RAD51 gene, whose sequence predicts a protein with similarities to procaryotic RecA proteins. Mol. Cell. Biol. 12:3224-3234.[Abstract/Free Full Text]

AGUILERA, A., 2001  The connection between transcription and genomic instability. EMBO J. 21:195-201.[Medline]

ALANI, E., L. CAO, and N. KLECKNER, 1987  A method for gene disruption that allows repeated use of URA3 selection in the construction of multiply disrupted yeast strains. Genetics 116:541-545.[Abstract/Free Full Text]

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]

BARTSCH, S., L. E. KANG, and L. S. SYMINGTON, 2000  RAD51 is required for the repair of plasmid double-stranded DNA gaps from either plasmid or chromosomal templates. Mol. Cell. Biol. 20:1194-1205.[Abstract/Free Full Text]

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

DATTA, A. and S. JINKS-ROBERTSON, 1995  Association of increased spontaneous mutation rates with high levels of transcription in yeast. Science 268:1616-1619.[Abstract/Free Full Text]

DAVIS, A. P. and L. S. SYMINGTON, 2001  The yeast recombinational repair protein Rad59 interacts with Rad52 and stimulates single-strand annealing. Genetics 159:515-525.[Abstract/Free Full Text]

DIXON, W. J., and F. J. MASSEY, JR., 1969 Introduction to Statistical Analysis. McGraw-Hill, New York.

EISEN, J. A., K. S. SWEDER, and P. C. HANAWALT, 1995  Evolution of the SNF2 family of proteins: subfamilies with distinct sequences and functions. Nucleic Acids Res. 23:2715-2723.[Abstract/Free Full Text]

ESPOSITO, M. S., 1978  Evidence that spontaneous mitotic recombination occurs at the two-strand stage. Proc. Natl. Acad. Sci. USA 75:4436-4440.[Abstract/Free Full Text]

ESSERS, J., A. B. HOUTSMULLER, L. VAN VEELEN, C. PAULUSMA, and A. L. NIGG et al., 2002  Nuclear dynamics of RAD52 group homologous recombination proteins in response to DNA damage. EMBO J. 21:2030-2037.[Medline]

FLEIG, U. N., R. D. PRIDMORE, and P. PHILIPPSEN, 1986  Construction of LYS2 cartridges for use in genetic manipulations of Saccharomyces cerevisiae.. Gene 46:237-245.[Medline]

GIETZ, R. D. and R. H. SCHIESTL, 1995  Transforming yeast with DNA. Methods Mol. Cell. Biol. 5:255-269.

GOTTLIEB, S., J. WAGSTAFF, and R. E. ESPOSITO, 1989  Evidence for two pathways of meiotic intrachromosomal recombination in yeast. Proc. Natl. Acad. Sci. USA 86:7072-7076.[Abstract/Free Full Text]

HABER, J. E. and M. HEARN, 1985  RAD52-independent mitotic gene conversion in Saccharomyces cerevisiae frequently results in chromosomal loss. Genetics 111:7-22.[Abstract/Free Full Text]

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]

HOEIJMAKERS, J. H., 2001  Genome maintenance mechanisms for preventing cancer. Nature 411:366-374.[Medline]

IVANOV, E. L., N. SUGAWARA, C. I. WHITE, F. FABRE, and J. E. HABER, 1994  Mutations in XRS2 and RAD50 delay but do not prevent mating-type switching in Saccharomyces cerevisiae.. Mol. Cell. Biol. 14:3414-3425.[Abstract/Free Full Text]

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]

JABLONOVICH, Z., B. LIEFSHITZ, R. STEINLAUF, and M. KUPIEC, 1999  Characterization of the role played by the RAD59 gene of Saccharomyces cerevisiae in ectopic recombination. Curr. Genet. 36:13-20.[Medline]

JINKS-ROBERTSON, S., M. MICHELITCH, and S. RAMCHARAN, 1993  Substrate length requirements for efficient mitotic recombination in Saccharomyces cerevisiae.. Mol. Cell. Biol. 13:3937-3950.[Abstract/Free Full Text]

JOHNSON, R. D. and L. S. SYMINGTON, 1995  Functional differences and interactions among the putative RecA homologs Rad51, Rad55, and Rad57. Mol. Cell. Biol. 15:4843-4850.[Abstract]

KANG, L. E. and L. S. SYMINGTON, 2000  Aberrant double-strand break repair in rad51 mutants of Saccharomyces cerevisiae.. Mol. Cell. Biol. 20:9162-9172.[Abstract/Free Full Text]

KEIL, R. L. and G. S. ROEDER, 1984  Cis-acting, recombination-stimulating activity in a fragment of the ribosomal DNA of S. cerevisiae.. Cell 39:377-386.[Medline]

KRAUS, E., W. Y. LEUNG, and J. E. HABER, 2001  Break-induced replication: a review and an example in budding yeast. Proc. Natl. Acad. Sci. USA 98:8255-8262.[Abstract/Free Full Text]

KREJCI, L., J. DAMBORSKY, B. THOMSEN, M. DUNO, and C. BENDIXEN, 2001  Molecular dissection of interactions between Rad51 and members of the recombination-repair group. Mol. Cell. Biol. 21:966-976.[Abstract/Free Full Text]

KUPIEC, M. and T. D. PETES, 1988  Allelic and ectopic recombination between Ty elements in yeast. Genetics 119:549-559.[Abstract/Free Full Text]

KUZMINOV, A., 1999  Recombinational repair of DNA damage in Escherichia coli and bacteriophage . Microbiol. Mol. Biol. Rev. 63:751-813.[Abstract/Free Full Text]

LEA, D. E. and C. A. COULSON, 1949  The distribution of the numbers of mutants in bacterial populations. J. Genet. 49:264-285.

LICHTEN, M., R. H. BORTS, and J. E. HABER, 1987  Meiotic gene conversion and crossing over between dispersed homologous sequences occur frequently in Saccharomyces cerevisiae.. Genetics 115:233-246.[Abstract/Free Full Text]

LIEFSHITZ, B., A. PARKET, R. MAYA, and M. KUPIEC, 1995  The role of DNA repair genes in recombination between repeated sequences in yeast. Genetics 140:1199-1211.[Abstract]

LOVETT, S. T. and R. K. MORTIMER, 1987  Characterization of null mutants of the RAD55 gene of Saccharomyces cerevisiae: effects of temperature, osmotic strength and mating type. Genetics 116:547-553.[Abstract/Free Full Text]

MALAGON, F. and A. AGUILERA, 2001  Yeast spt6-140 mutation, affecting chromatin and transcription, preferentially increases recombination in which Rad51p-mediated strand exchange is dispensable. Genetics 158:597-611.[Abstract/Free Full Text]

MALDONADO, E., R. SHIEKHATTAR, M. SHELDON, H. CHO, and R. DRAPKIN et al., 1996  A human RNA polymerase II complex associated with SRB and DNA-repair proteins. Nature 381:86-89.[Medline]

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.