We have previously shown that recombination between 400-bp substrates containing only 4-bp differences, when present in an inverted repeat orientation, is suppressed by >20-fold in wild-type strains of S. cerevisiae. Among the genes involved in this suppression were three genes involved in mismatch repair—MSH2, MSH3, and MSH6—and one in nucleotide excision repair, RAD1. We now report the involvement of these genes in interchromosomal recombination occurring via crossovers using these same short substrates. In these experiments, recombination was stimulated by a double-strand break generated by the HO endonuclease and can occur between completely identical (homologous) substrates or between nonidentical (homeologous) substrates. In addition, a unique feature of this system is that recombining DNA strands can be given a choice of either type of substrate. We find that interchromosomal crossover recombination with these short substrates is severely inhibited in the absence of MSH2, MSH3, or RAD1 and is relatively insensitive to the presence of mismatches. We propose that crossover recombination with these short substrates requires the products of MSH2, MSH3, and RAD1 and that these proteins have functions in recombination in addition to the removal of terminal nonhomology. We further propose that the observed insensitivity to homeology is a result of the difference in recombinational mechanism and/or the timing of the observed recombination events. These results are in contrast with those obtained using longer substrates and may be particularly relevant to recombination events between the abundant short repeated sequences that characterize the genomes of higher eukaryotes.

---

MISMATCHES in DNA heteroduplex originate from replicative error or from recombination between similar but nonidentical (homeologous) sequences. Several homologs of the Escherichia coli MutS gene are involved in recognizing mismatches in the yeast Saccharomyces cerevisiae (CROUSE 1998; HARFE and JINKS-ROBERTSON 2000b). Three of these yeast MutS homologs—Msh2, Msh3, and Msh6—are required for recognition of mismatches arising from nuclear mitotic processes. Synergistic effects of disruption of MSH3 and MSH6 suggest that Msh2 forms a heterodimer with either Msh3 or Msh6 (JOHNSON et al. 1996b; MARSISCHKY et al. 1996; GREENE and JINKS-ROBERTSON 1997), a result subsequently confirmed biochemically (HABRAKEN et al. 1996; IACCARINO et al. 1996). The Msh2/Msh3 or Msh2/Msh6 heterodimer complexes with a heterodimer of MutL homologs; the heterodimer of Pms1 and Mlh1 is thought to be the primary partner for the Msh2/Msh3 or Msh2/Msh6 heterodimer (PROLLA et al. 1994a,b), but Mlh1 will interact with Mlh2 or Mlh3, resulting in differing activities of the protein complexes (FLORES-ROZAS and KOLODNER 1998; WANG et al. 1999; HARFE et al. 2000; HARFE and JINKS-ROBERTSON 2000b). Recognition specificity of the mismatch repair (MMR) complex is dictated by its MutS homolog components. Small and large loops resulting from replicative error or homeologous recombination are recognized by Msh2 and Msh3, presumably acting as a heterodimer (JOHNSON et al. 1996b; MARSISCHKY et al. 1996; GREENE and JINKS-ROBERTSON 1997; SIA et al. 1997; EVANS and ALANI 2000; NICHOLSON et al. 2000); this heterodimer is also involved in suppression of homeologous recombination even when the recombination heteroduplex contains only base–base mismatches (NICHOLSON et al. 2000). In contrast, the Msh2/Msh6 heterodimer in the MMR complex mediates only recognition of base–base mismatches (MARSISCHKY et al. 1996; DATTA et al. 1997; EARLEY and CROUSE 1998; CHEN and JINKS-ROBERTSON 1999) and small loops (JOHNSON et al. 1996b; MARSISCHKY et al. 1996; GREENE and JINKS-ROBERTSON 1997; NICHOLSON et al. 2000) resulting from either replicative error or homeologous recombination, but demonstrates no recognition of loops of 4 bp or larger (SIA et al. 1997; NICHOLSON et al. 2000).

In addition to its role in recognition of mismatches resulting from replicative error or homeologous recombination, the Msh2/Msh3 heterodimer is necessary for removal of terminal nonhomology during recombination (PâQUES and HABER 1997; SUGAWARA et al. 1997; COLAIÁCOVO et al. 1999). During double-strand break (DSB) repair, a 5' 3' exonuclease resects the 5' strand of each DNA end, leaving the recombinogenic 3' strand to detect and invade homologous DNA sequences (SZOSTAK et al. 1983; SUN et al. 1991). If the terminal 3' sequence is not homologous to the invaded donor DNA sequence, this terminal nonhomology must be removed to complete recombination. To remove terminal nonhomology, the Msh2/Msh3 heterodimer works in conjunction with a heterodimer of the nucleotide excision repair (NER) proteins Rad1 and Rad10, apparently to facilitate endonucleolytic cleavage of the junction between single-strand nonhomologous DNA and homologous duplex DNA by the Rad1/Rad10 complex (FISHMAN-LOBELL and HABER 1992; IVANOV and HABER 1995; SAPARBAEV et al. 1996; PâQUES and HABER 1997; SUGAWARA et al. 1997; EVANS et al. 2000) and thereby remove the single-strand DNA flap (SUNG et al. 1993; TOMKINSON et al. 1993; BARDWELL et al. 1994). Msh2 has been shown to interact directly with Rad1 and Rad10 (BERTRAND et al. 1998), and because the two heterodimers clearly function very closely in the same pathway, we will refer to them in that pathway as the MSH/RAD complex. The function of Msh2 in removal of terminal nonhomology is separate from its function in mismatch repair, as mutants in MSH2 that are defective in mismatch repair but competent for removal of terminal nonhomology have been identified (STUDAMIRE et al. 1999). Rad1 and Rad10, along with Msh2, Msh3, and Msh6, were shown to be involved in the suppression of homeologous recombination occurring by a sister-chromatid conversion process (CHEN and JINKS-ROBERTSON 1998; NICHOLSON et al. 2000). In these experiments, homologous recombination was increased in the absence of either Msh2 or Msh3 and, to a much lesser extent, by loss of Rad1 or Rad10. The failure to remove terminal nonhomology would not be expected to increase recombination rates, so this result suggests a larger role for the MSH/RAD complex than simple removal of terminal nonhomology during recombination.

In this study we examine the roles of Msh2, Msh3, Msh6, and Rad1 in crossover recombination between the same recombination substrates used previously, but this time located on nonhomologous chromosomes. These substrates contain 400 bp of homologous sequences. We find substantial differences in the effects of all of the studied genes compared to our previous inverted repeat assay where recombination is thought to occur via gene conversion (NICHOLSON et al. 2000). In addition, we find that recombination in our interchromosomal assay is markedly less sensitive to mismatches than was recombination in the inverted repeat assay. These differences can be explained by the different recombinational pathways in the two systems and/or by differences in the timing of the lesions initiating recombination.

Media and growth conditions:

All incubations were done at 30°. Nonselective (YPD) medium contained 2% dextrose, 1% yeast extract, and 2% bacto-peptone, as well as 2.5% agar for plates. Synthetic dextrose (SD) minimal medium contained 0.17% yeast nitrogen base, 0.5% ammonium sulfate, 2% dextrose, and 2.5% agar for plates, except for SD–Ura plates used to select recombinants, which contained 0.85% agarose instead of agar. SD–Ura, SD–Leu, and SD–Lys plates contained SD medium supplemented with a uracil-, leucine-, or lysine-deficient amino acid mix, respectively. Ura^– segregants were identified on plates of SD medium supplemented with a complete amino acid mix and 0.1% 5-fluoroorotic acid (5-FOA) (BOEKE et al. 1984). Cells to be assayed for recombination were grown initially in SD–Leu medium. Galactose induction was performed in the same medium with the dextrose substituted by 2% ethanol, 3% glycerol, and 2% galactose (SGEG–Leu).

Plasmid constructions:

Plasmids were constructed to place portions of the URA3 gene or the URA3/lacZ gene fusion, as well as intron splice recognition sequences and intronic recombination substrates, at each of several chromosomal loci. The recombination substrates were derived from a 350-bp fragment of chicken ß-tubulin isoform 2 cDNA (Cß2) (DATTA et al. 1996).

pK2:5U3–Cß2HO was used to place a recombination construct consisting of the promoter and 5' portion of the URA3 gene, with an artificial intron containing the Cß2–HO recombination substrate at the LYS2 locus. pK2:5U3–Cß2HO contains a 4.9-kb EcoRV fragment of the LYS2 gene from pDP6 (FLEIG et al. 1986) cloned into the 3.9-kb PvuII fragment of pRS306 (SIKORSKI and HIETER 1989). A 1.5-kb XmnI/EcoRV fragment of pRS316 (SIKORSKI and HIETER 1989), followed by the artificial intron from pUC-AI (YOSHIMATSU and NAGAWA 1989), was cloned into the BamHI site of the LYS2 gene in the transcriptionally opposite orientation. The BamHI site of the artificial intron contained an 350-bp Cß2 fragment amplified by PCR using primers 5'-GGGGATCCACTCCACAAAGTAG-3' and 5'-CGGTCGACAGATCTGGCCACCATGAGCGGGTGA-3', with pSR257 (DATTA et al. 1996) as a template. A 30-bp HO endonuclease recognition site (Figure 2) was inserted into the Cß2 sequence to form CB2–HO using the Stratagene (La Jolla, CA) ExSite PCR-based site-directed mutagenesis kit and primers 5'-GCTTCGTCAGCGGGGCGAAGCCCG-3' and 5'-TTATACTGTTGCGGAAAGCTGAAACTAAAAGCCGGCAGCCAGCAGTACCGAGCCC-3'.

View larger version (19K): In this window In a new window Download PPT slide   FIGURE 2.— Recombination substrates used in this study. (A) Recombination substrates were derived from a 350-bp fragment of the coding sequence of chicken ß-tubulin. Int(HO) had a 30-bp HO endonuclease cut site inserted at the SacII site in Cß2. Int(ns) had nucleotide substitutions introduced so that recombination with the Int substrate produces base–base mismatches in the recombination heteroduplex. Int(HOnc) had the HO cut site modified, preventing incision by HO endonuclease. However, the recombining 3'-ends produced by cleavage of Int(HO) do not require processing of terminal nonhomology when recombining with the Int(HOnc) substrate. (B) The original HO cut site, used in Int(HO), is 30 bp and produces a single-strand overhang when cut by HO endonuclease. The noncutting HO site contains a duplication of the overhang sequence, interrupted by an SphI site.

 
pL2:Cß2–3U3LZ was used to place the Cß2 recombination substrate joined to the 3'-end of a URA3/lacZ gene at the LEU2 locus. Recombination to this substrate resulted in blue Ura⁺ colonies in the presence of X-gal. pL2:Cß2–3U3LZ was generated as follows: A PvuII fragment of pUC18URA3 (ROTHSTEIN 1990), lacking the polylinker sites, was cloned into the PvuII sites of pRS305 (SIKORSKI and HIETER 1989). The recombination cassette was placed between the KpnI and EcoRI sites of the LEU2 gene in this plasmid in reverse transcriptional orientation. The cassette consisted of the same artificial intron construct with the cß2 sequences as above (without the HO site) joined to the EcoRV site within the URA3 gene and the 3'-end of the URA3 coding sequence joined in frame to codon 9 of the E. coli lacZ gene, followed by the yeast CYC1 terminator sequence from p416 MET25 (MUMBERG et al. 1994).

pU3:Cß2–3U3 targets the Cß2 recombination substrate to the URA3 locus and contains the 3'-end of the URA3 gene so that recombination to this sequence results in formation of white Ura⁺ colonies. It was constructed as follows: The LEU2 gene from pRS305 (SIKORSKI and HIETER 1989) was cloned between the KasI and AatII sites of pUC18URA3 (ROTHSTEIN 1990) from which the polylinker sites had been removed. Between the PstI and ApaI sites of the URA3 gene, in transcriptionally reverse orientation, a MscI/Bsu36I fragment of pL2:Cß2–3U3LZ containing the cß2 containing artificial intron joined to the 3'-end of the URA3 gene was placed.

Several plasmids derived from pL2:Cß2–3U3LZ contain modified introns. pL2:Cß2–HOnc–3U3LZ contains a modified HO substrate, which cannot be cut by the HO endonuclease, but for which an invading end that has been cut by the HO endonuclease will find perfect homology (see Figure 2b). The modified HO site was introduced into the Cß2 intron sequence of pK2:5U3–Cß2HO using the ExSite procedure from Stratagene and primers 5'-GCTTCGTCAGCGGGGCGAAGCCCG-3' and 5'-TTATACTGTTGCATGCTGTTGCGGAAAGCTGAAACT-3'. The modified intron was then moved into pL2:Cß2–3U3LZ. pL2:Cß2–ns–3U3LZ contains four nucleotide substitutions in the recombination substrate, which will result in base–base mismatches during recombination, and was constructed by replacing the cß2 intron sequence with that of the intron sequence of pAB92 (NICHOLSON et al. 2000).

The plasmid used to disrupt the native MAT locus, pACN143, contains the MAT locus of plasmid pJH526 disrupted by URA3, obtained from Jim Haber but originally from Jeff Strathern's lab. The URA3 gene was replaced by a PvuII fragment containing the HisG–URA3–Kan^R–HisG cassette from pHUKH4 (EARLEY and CROUSE 1996).

Strain constructions:

All strains were derived from SJR328 (MAT ade2-101 his3200 ura3-Nhe lys2RV::hisG leu2-R). Disruption of the native MAT locus was established by transforming cells with BsrGI/EcoRI-digested pACN143. Following the initial selection for Ura⁺ transformants, Ura^– segregants were screened by PCR. The plasmids containing the recombination constructs were targeted in a two-step replacement process to the LEU2, LYS2, or URA, loci on chromosome III, II, or V, respectively, by digestion with KasI, BstXI, or NsiI.

Following introduction of the recombination constructs into yeast, individual MMR, or NER genes were disrupted using a one-step disruption plasmid containing the HisG–URA3–HisG cassette. All transformants were selected on SD–Ura plates, and then spontaneous deletion of the URA3 gene from the HisG–URA3–HisG cassette was selected on 5-FOA medium. MSH2 was disrupted by transformation with AatII/XbaI-digested pmsh2 (EARLEY and CROUSE 1998), MSH3 by transformation with AflII/MscI-digested pmsh3 (EARLEY and CROUSE 1998), MSH6 by transformation with EcoRI/SacI-digested Msh6pHUH (KRAMER et al. 1996), and RAD1 by transformation with EcoRI/SalI-digested pR1.6 (HIGGINS et al. 1983). Gene disruption was confirmed by PCR and phenotype analysis. Strains used in this study are listed in Table 1.

View this table: In this window In a new window   TABLE 1 S. cerevisiae strains used in this study

 

Fluctuation analysis:

To assay recombination induced by incision with HO endonuclease, the plasmid pJH727 (originally from Anne Plessis; obtained from Jim Haber) containing a LEU2 marker and the HO endonuclease under a galactose-inducible promoter was introduced into the wild-type or gene-disrupted strains. For each strain, 36 (or more) independent cultures were grown in 5 ml of SD–Leu media for 22 hr. Approximately 2 x 10⁷ cells were then transferred to SGEG–Leu and grown for 15 hr. At this point, cells were washed in H₂O and appropriate dilutions were plated onto YPD medium to determine the total number of cells or onto SD–Ura agarose medium to determine the number of recombinants. YPD plates were incubated for 3 days prior to counting colonies, and SD-Ura plates were incubated for 5 days. Recombination frequencies were determined and error limits were calculated as described (SPELL and JINKS-ROBERTSON 2004).

Determination of Lac⁺ and Lac^– colonies:

Lac⁺ colonies were determined by overlaying chloroform-treated colonies with X-gal in low-melting agarose as previously described (DUTTWEILER 1996).

Pulsed-field gel analysis:

Yeast plugs were analyzed on a 1% agarose gel using a Bio-Rad (Hercules, CA) CHEF-DR III apparatus at 6 V/cm for 28 hr with a ramped switch time of 20–80 sec and an angle of 120°. After electrophoresis, the gel was stained with Vistra Green (Amersham Biosciences). DNA in the gel was nicked and denatured, first by treatment with 0.15 J/cm² UV and then by treatment with 0.5 N NaOH for 30 min followed by neutralization in HCl. DNA was transferred to a nylon membrane using a TurboBlotter (Schleicher & Schuell, Keene, NH) according to the manufacturer's directions followed by hybridization (CHURCH and GILBERT 1984) with a 1.3-kb AflII/NcoI fragment of the LYS2 gene.

Southern blot analysis of HO cutting:

After digestion with the appropriate restriction enzymes, genomic DNA was electrophoresed on a 1% agarose gel. DNA in the gel was transferred and hybridized as described above, omitting the UV nicking. The blot was hybridized with an 816-bp AflII/SpeI fragment of the LYS2 gene.

Construction of an assay for analysis of reciprocal exchange events:

We describe here the development and use of a novel recombination assay, illustrated in Figure 1. In the assay in this study, recombination is stimulated by a DSB made by the HO endonuclease at a defined location in a recombination substrate. The resulting recombinogenic 3' DNA ends are allowed to recombine with a substrate containing complete identity, or with a substrate that is identical only after the removal of short terminal nonhomology. Recombination with substrates producing defined mismatches in the heteroduplex DNA was also examined. By placing differing substrates at separate chromosomal locations, a unique competitive recombination assay has also been created. Thus, we can measure the degree to which a substrate that will form mismatches in the recombination heteroduplex intermediate is discriminated against when there is a substrate present that is fully identical after any terminal nonhomology is processed.

View larger version (23K): In this window In a new window Download PPT slide   FIGURE 1.— A competitive ectoptic recombination assay. (A) Recombination is stimulated by a double-strand break and proceeds ectopically by reciprocal crossover to one of two possible chromosomal locations. As described in the text, recombination occurs within intervening intronic sequences in the URA3 gene; subsequent RNA processing allows a functional gene product to be produced. The locus chosen is distinguished by presence or absence of blue colony color. (B) A schematic of recombination bringing the two parts of the URA3 gene together, separated by an intron, which is subsequently spliced out to give a functional URA3 mRNA.

 
The recombination assay described here makes use of an artificial intron recombination assay that has been extensively characterized (DATTA et al. 1996; CHEN and JINKS-ROBERTSON 1998, 1999; NICHOLSON et al. 2000). Recombination substrates, consisting of DNA sequence derived from chicken ß-tubulin isoform 2 cDNA—either unmodified or modified so that base–base mismatches [Int(ns)] (NICHOLSON et al. 2000) will form in the recombinational heteroduplex when these substrates recombine—were placed within the intron. Although previous studies with this system used an artificial intron placed in the HIS3 gene, this work placed the artificial intron in the URA3 gene. The URA3 promoter and coding sequence 5' of the EcoRV site in the URA3 gene, as well as the 5' intron splice donor and a recombination substrate, were placed at the LYS2 locus, disrupting the LYS2 gene. Another recombination substrate, the intron splice acceptor, and the URA3 sequence 3' of the EcoRV site were placed at either, or both, the URA3 and LEU2 loci, disrupting the URA3 and/or LEU2 genes, respectively. Recombination between the cß2-derived substrates proceeded ectopically, resulting in a reciprocal translocation that brought 5' and 3' URA3 sequences into physical alignment and allowed the production of a functional URA3 gene product.

Spontaneous ectopic crossover recombination occurred at extremely low rates (data not shown). To increase recombination frequencies to measurable levels and to have a defined initiation site, a 30-bp MATa HO endonuclease recognition site (NICKOLOFF et al. 1990) was placed in the recombination substrate located at the LYS2 locus (see Figure 2) on chromosome II. The placement of the HO site in the substrate resulted in 156 nt of homology with the recombining substrate on the 5' side of the HO break and 255 nt of homology on the 3' side of the HO break. Growth of the strains in galactose-containing media, which induced expression of the HO endonuclease from a plasmid containing the HO gene under a GAL promoter, resulted in greatly increased levels of Ura⁺ recombinants. Growth of these strains in glucose-containing SD–Leu media, which repressed production of HO endonuclease, resulted in an extremely low frequency of production of Ura⁺ colonies (data not shown).

To distinguish between recombination with the substrates on chromosomes III and V, the lacZ gene from E. coli was fused in frame to the 3' URA3 coding sequence that was placed at the LEU2 locus on chromosome III (Figure 1). The lacZ gene produces ß-galactosidase, resulting in blue pigmentation of colonies exposed to X-gal, so recombination events targeted to the LEU2 locus were measured as blue colonies on SD–Ura plates (see MATERIALS AND METHODS).

As noted previously, earlier assays utilizing the artificial intron had placed it in the HIS3 gene rather than in the URA3 gene (DATTA et al. 1996). Placement of the artificial intron and recombination substrates in the URA3 gene was necessitated by an instability of the HIS3 mRNA (HERRICK and JACOBSON 1992), which resulted in a lack of colony formation when production of the His3 protein was challenged by both an elongated mRNA resulting from the fusion of HIS3 to lacZ and a requirement for intron splicing (results not shown). A fortunate consequence of this change was that the Ura3/LacZ fusion protein was produced robustly and the native URA3 promoter could be used in the assay, rather than the GAL1-10 promoter that had been necessary in assays utilizing HIS3. Fusion of the lacZ gene to URA3 did not affect the survival or growth of the resultant colonies on SD–Ura plates (data not shown).

To determine the nature of the recombination events stimulated by the HO endonuclease, individual white and blue recombinants were analyzed by pulsed-field gel electrophoresis. The resulting gel was blotted and hybridized with a probe specific for the 5'-end of the recombining gene. As shown in Figure 3, the chromosomal pattern of the recombinants typically differs from that of the parent exactly as predicted. Three of four of the white recombinants and all five of the blue recombinants have a chromosomal pattern consistent with reciprocal translocation, revealed both by the stained bands and by hybridization with the 5' probe. One of the recombinants (W1) has a chromosomal banding pattern like that of the parent and may represent a combination of homologous and nonhomologous recombination events like that seen in mammalian cells (RICHARDSON and JASIN 2000) (see DISCUSSION).

View larger version (80K): In this window In a new window Download PPT slide   FIGURE 3.— Pulsed-field gel of white and blue recombinants. (A) A pulsed-field gel of the parent strain, GCY1249 (P), four white recombinants (W1–4), and five blue recombinants (B1–5) was run as described in MATERIALS AND METHODS and stained with Vistra Green. The arrows on the left indicate the positions of the predicted wild-type chromosomes, and the arrows on the right indicate the expected positions of the recombinant chromosomes for reciprocal translocations. White colonies should have chromosome II/V and V/II translocations, and blue colonies should have chromosome II/III and III/II translocations. (A II/V translocation would have the left end of chromosome II and the right end of chromosome V, for example.) (B) Hybridization of the pulsed-field gel in A with the LYS2 probe described in MATERIALS AND METHODS, which should hybridize with the 5'-end of the recombining URA3 gene. In the parent strain, this should be to chromosome II, and in the recombinant strains, it should be to the II/V chromosome in white colonies and to the II/III chromosome in blue colonies.

 

Effect of MMR and RAD1 on recombination:

In our previous study using the inverted repeat assay (NICHOLSON et al. 2000), we had found that, with homologous substrates, deleting either MSH2 or MSH3 increased recombination approximately twofold, deleting RAD1 had a smaller positive effect, and deleting MSH6 consistently gave a slight decrease in recombination. However, the effects of these gene deletions were strikingly different with the present recombination substrates. Deletion of MSH2 or MSH3 decreased recombination by 4-fold and deletion of RAD1 decreased recombination by 10-fold, whereas deletion of MSH6 did not affect recombination (Figure 4). The relative effect of the various deletions was similar for both recombination substrates, as seen in Figure 4A.

View larger version (21K): In this window In a new window Download PPT slide   FIGURE 4.— The effects of mutations in MMR and RAD1 on recombination compared to wild-type strains. (A) Recombination frequencies to the ura3::Int-URA3 () or to the leu2::Int-URA3-LacZ () substrates are shown for different genetic backgrounds. (B) Recombination frequencies to the leu2::Int-URA3-LacZ () or to the leu2::Int(HOnc)-URA3-Lac () substrates. (C) Recombination frequencies to the leu2::Int-URA3-LacZ () or to the leu2::Int(ns)-URA3-LacZ () substrates.

 

Effect of terminal nonhomology on recombination:

When the Int(HO) recombination substrate recombines with the Int substrate, the HO recognition site sequences (KOSTRIKEN et al. 1983; NICKOLOFF et al. 1990) at the ends of the DSB are not homologous to the Int sequence. Recombination between a cleaved Int(HO) substrate and an Int substrate requires processing of 10 and 24 nt of terminal nonhomology from the two recombinogenic 3'-ends. It had previously been shown that the MSH/RAD complex was required for removal of 30 nt or more of terminal nonhomology during gene conversion but had no effect on terminal nonhomology as short as 10 nt (PâQUES and HABER 1997). It was still a formal possibility that the reduction of recombination in the absence of MSH2, MSH3, or RAD1 was due to less efficient processing of terminal nonhomology, thereby reducing the overall level of recombination. To determine the effect of this small amount of terminal nonhomology on recombination, an inactivated HO site was placed into the Int sequence, generating the Int(HOnc) substrate (Figure 2). Because we expected recombination to be sensitive to any type of base mismatch (NICHOLSON et al. 2000), we had to inactivate the HO site in a way that would not generate any mismatches with the recombining HO ends of the 5' substrate. By duplicating the 4-bp overhang of the cut HO site and inserting DNA between this duplication, the substrate was made not susceptible to cleavage by HO endonuclease. In addition, as the recombining DNA ends of a cleaved Int(HO) substrate should encounter no terminal nonhomology in the Int(HOnc) substrate, processing of terminal nonhomology should not be required.

Wild-type strains containing the Int(HOnc) substrate displayed recombination frequencies comparable to those with the Int substrate (Figure 4B). Similarly, deletion of MSH6 had little effect on recombination. Recombination in strains that were deleted for MSH2 or MSH3 was at similar levels; only in the case of rad1 strains was this level significantly higher with the HOnc substrate than with the original recombination substrate. This result indicates that the presence of terminal nonhomology was not responsible for the differing effect of MSH2, MSH3, and RAD1 on recombination in this system compared to the inverted substrates used previously.

Effect of defined mismatches on recombination frequencies:

As a cleaved Int(HO) substrate will require the same processing of terminal nonhomology to recombine with the Int substrate or the Int(ns) substrate, these substrates can be compared directly to determine the effect of homeology on recombination proceeding by a reciprocal exchange pathway. The Int(ns) substrate has four evenly spaced nucleotide substitutions compared to the Int substrate; recombination frequencies of these substrates are summarized in Figure 4C.

Comparing the Int(ns) substrate to the Int substrate, when each is located at the LEU2 locus, we observe a 4-fold reduction in recombination to the Int(ns) substrate in wild-type strains. This compares to a 22-fold reduction with the same substrates in the inverted repeat cassette assay (NICHOLSON et al. 2000). Recombination of homeologous substrates in an msh6 background is equal to that of the homologous substrates. However, recombination with the Int(ns) substrate in msh2, msh3, and rad1 strains continues to be strongly decreased and is essentially the same as with the homologous substrates.

Competitive crossover recombination:

The relatively small effect of mismatches on recombination in the assay in this study could perhaps be explained by the fact that, once a DSB has occurred, the cell is forced to repair its DNA by recombination or it will die. Even if a recombination intermediate with mismatches were dissolved by MMR, the same broken ends might be able to participate in additional recombination attempts. Thus, if recombinogenic ends were given a choice of homologous or homeologous substrates, one might see a greater effect due to mismatches. By placing identical or differing recombination substrates at both the URA3 and the LEU2 locus, as illustrated in Figure 1, we can examine which of the substrates competes most effectively for recombinogenic 3' DNA ends resulting from the HO-endonuclease-induced cleavage of Int(HO). These results are illustrated in Figure 5.

View larger version (43K): In this window In a new window Download PPT slide   FIGURE 5.— Effects of MMR and RAD1 on competitive recombination. (A) Recombination frequencies in the strains with ura3::Int-URA3 and leu2::Int-URA3-LacZ () and recombination frequencies in the strains with ura3::Int-URA3 and leu2::Int(ns)-URA3-LacZ () are shown for different genetic backgrounds. (B) The percentage of blue colonies in the strains with ura3::Int-URA3 and leu2::Int-URA3-LacZ () or in the strains with ura3::Int-URA3 and leu2::Int(ns)-URA3-LacZ ().

 
When identical substrates were available for recombination at both the URA3 and the LEU2 locus, the number of total recombinants was not significantly increased over the recombinants from strains with only one choice of recombination substrate. The distribution of recombinants between the two potential recombination substrates could be observed by determining blue vs. white colonies and 40% of the colonies were blue. This distribution is consistent with the results obtained with strains containing the individual substrates (Figure 4), where recombination with the substrate on chromosome V, which would give white colonies, was always slightly greater than that with the substrate on chromosome III, which would give blue colonies.

The effect of gene deletions on recombination frequencies for the homologous competitive substrates was as expected. The relative effects of these gene deletions can be seen most clearly in Figure 5A. The choice of homologous substrates, as assayed by the color of the resulting colonies, was also unaffected by gene deletion, with one possible exception (Figure 5B). In rad1 strains, the percentage of blue colonies appeared to be significantly higher than in other strains; this was an unexpected result because RAD1 did not appear to have a differential effect when either locus was assayed independently (Figure 4A).

When both homologous and homeologous substrates were available for recombination, the overall frequencies were similar to those obtained with two homologous substrates, as were the effects of various gene deletions, as seen in Figure 5A. In wild-type strains, the percentage of recombination events with the homeologous substrate, which would lead to blue colonies, was suppressed compared to the homologous substrate, but less than twofold. This is even less than the effect seen in the noncompetitive strains and indicates that there is no greater discrimination between homologous vs. homeologous substrates when choice is possible. The discrimination against homeologous sequences observed in the wild-type strain was not seen in any of the mutant strains, as indicated by the ratio of blue to white colonies.

Physical analysis of HO cutting:

In all cases, the levels of crossover recombination that we observed were very low. Was this low rate of recombination a reflection of very inefficient cutting by HO, or an alternative means of repair of the DSB? The most likely pathways of alternative repair would be gene conversion using the sister chromatid in late S or G₂ cells or gene conversion using the ectopic recombination substrate. We had no method to distinguish gene conversion using the sister chromatid from DNA that was not cut by HO, but gene conversion using the HOnc substrate would result in an added SphI restriction site, which could be assayed by Southern blot. Therefore we used cells containing the HOnc substrate to monitor HO cutting and ectopic gene conversion,

The results of this assay are shown in Figure 6. DNA digested by a combination of BsaI and HpaI and hybridized with the indicated LYS2 probe should result in one 8.2-kb band; if this region is cut by HO, the cut band would be 2.4 kb. As shown in lane 1, before induction of HO by galactose, there is no 2.4-kb band, but a 2.4-kb band is visible after 4 and 15 hr of induction in lanes 2 and 3. The HO site is cut inefficiently, as after 15 hr of induction, it represents only 4% of the uncut DNA. When induction is stopped and cells are allowed to grow for an additional 4 hr, the HO-cut fragment entirely disappears (lane 4). Note that the amount of crossover recombination is so low in these cells that in unselected cells the crossover product is not observed.

View larger version (74K): In this window In a new window Download PPT slide   FIGURE 6.— Physical analysis of HO cleavage. GCY1849 cells taken just before HO induction (lanes 1 and 5), after a 4-hr induction in galactose (lanes 2 and 6), 15 hr after induction (lanes 3 and 7), or after a 15-hr induction followed by 4 hr of growth in YPD (lanes 4 and 8) were harvested and genomic DNA was prepared. DNA was digested with BsaI and HpaI (lanes 1–4) or with BsaI, HpaI, and SphI (lanes 5–8), electrophoresed, blotted, and hybridized as described in MATERIALS AND METHODS. Relevant bands on the Southern are indicated by arrows, and the diagram below illustrates the position of the 5' URA3 exon, the HO site in the recombination substrate, and the expected sizes of restriction digests. The short horizontal line above the diagram represents the position of the hybridization probe. Note that a gene conversion event with the HOnc substrate would result in an SphI site at the position of the HO cut site.

 
The same DNA used in lanes 1–4 was digested in addition with SphI. The length of this fragment in uninduced cells is 7.3 kb, as shown in lane 5. Gene conversion of the HO cut using the HOnc substrate would result in an SphI site at the location of the original HO site, resulting in a 2.4-kb band. Lane 8 has a band of 2.4 kb representing 4% of the uncut DNA. All of this band must be due to cutting by SphI and thus a result of gene conversion using the HOnc substrate, as the same DNA when not digested by SphI shows no band of this size (lane 4).

Use of recombination substrates within artificial introns to study ectopic crossover recombination:

Use of artificial introns has several inherent advantages in recombination studies. Assays in which recombination must produce a functional gene product place constraints on recombination that can be avoided by allowing recombination to occur between recombination substrates within an artificial intron. Furthermore, the sequence of individual recombination substrates can be readily manipulated. By using site-directed mutagenesis to place an HO endonuclease recognition site in a recombination substrate, the exact location of the initiating DSB was defined. Previously generated substrates that form defined mismatches in the recombination heteroduplex (NICHOLSON et al. 2000) were used to determine the effects of mismatches during chromosomal ectopic recombination.

We were particularly interested in studying crossover recombination in these substrates both to contrast the results with those previously obtained via intrachromosomal gene conversion (NICHOLSON et al. 2000) and to observe recombination with relatively short substrates. Previous work had shown that HO-stimulated ectopic recombination involving long stretches of homology preferentially proceeded via gene conversion and that crossing over was not detected by Southern blots for total homology <1.6 kb (INBAR et al. 2000). There are many instances of repeated sequences of several hundred bases in eukaryotic chromosomes, such as solo LTR sequences in yeast (KIM et al. 1998) and Alu sequences in mammalian cells (LANDER et al. 2001); gene conversion with these sequences is unlikely to have any major phenotypic effect, whereas crossovers would lead to major chromosomal rearrangements. In addition, most of these elements show sequence divergence, and we were interested in the impact of such sequence divergence on recombination.

By examining the chromosomal structure of individual recombinants using pulsed-field electrophoresis, we were able to confirm that most of the observed events were clearly reciprocal translocations. One recombinant in Figure 3 (W1) did not appear to be rearranged and may represent a gene conversion event followed by nonhomologous end joining as observed previously in the mouse (RICHARDSON and JASIN 2000), but has not been further analyzed. Because the strains are haploid, other aberrant types of recombination that would be nonreciprocal in nature would not be compatible with viability.

Physical analysis of HO cutting revealed that, as expected, at any one time only a small fraction of genomes containing the HO site was cleaved by HO endonuclease. Unlike the situation with a longer HO site that is very efficiently cleaved, this inefficient cleavage would allow repair of the HO site from the sister chromatid, if it existed at the time of cleavage, since it would be unlikely that both copies of the HO site would be cut. However, repair of the HO break from the sister would restore a cleavable site, allowing for the possibility of recleavage during the period of induction. The fraction of cleaved HO sites at any one time was on the order of 4%, and this value was orders of magnitude higher than the fraction of cells with a selectable crossover event (3.3 x 10^–5 for the HOnc substrate). Gene conversion using the ectopic HOnc substrate, unlike gene conversion using the sister chromatid, would yield a product that could not be subsequently recleaved by HO (but would contain an SphI site). We found that 4 hr after terminating induction of HO, 4% of unselected cells showed such a gene conversion, as measured by the acquisition of an SphI site. Although the percentage of cells showing a gene conversion to the HOnc site cannot be directly compared to the percentage of cells at any one time showing a cleaved HO site, it is clear that the vast majority of interactions of the HO site with the HOnc substrate result in gene conversions rather than crossovers. Thus the value of our assay is that we can observe specific events that, although composing a very small fraction of total events, represent those that are most significant for the genome.

A mild preference for the URA3 locus over the LEU2 locus is detected:

Comparing the strains containing the recombination substrate at either the URA3 or the LEU2 locus, recombination occurs at consistently higher frequencies at the URA3 locus, although the confidence intervals of the two measurements overlap. When the same recombination substrate is available at both URA3 and LEU2 loci, recombination occurs at the URA3 locus 60% of the time, giving white colonies and supporting the preference seen in the recombination frequency measurements. The preference for the URA3 locus over the LEU2 locus may be due to LEU2's location on the right arm of chromosome III, for which there is a generalized suppression of recombination, allowing the cell to distinguish between donor sequences for mating-type switching (HABER 1998). This recombination suppression is most effective at the telomere, becoming less effective nearer the centromere. The LEU2 locus is located relatively near the centromere and so should be only mildly suppressed in recombination, which is what we observe. It is interesting that this preference seems to be abolished, or even reversed, in rad1 strains.

Terminal nonhomology and ectopic recombination:

The HO endonuclease creates an overhanging, staggered cut (NICKOLOFF et al. 1990) so that the terminal nonhomology in this assay with a minimal HO site is 24 nt on one side of the HO cut and 10 nt on the other. As successful formation of reciprocally translocated chromosomes is necessary for recovery of most of the recombinants in this assay, both recombinogenic 3'-ends must complete recombination. The ectopic recombination assay described here makes it possible to directly compare recombination that requires processing of small amounts of terminal nonhomology with recombination not requiring such processing. To examine the effect of this terminal nonhomology on recombination, a recombination substrate [Int(HOnc)] containing perfect homology to the ends of the HO cut site was created. Comparing recombination frequencies of the Int and Int(HOnc) substrates, when each was the only available recombination substrate, reveals essentially no change in wild-type or msh6 strains. Because the proteins necessary for processing of the terminal nonhomology are functional in these backgrounds, this result indicates that in normal circumstances the presence of small amounts of terminal nonhomology presents no obstacle for the recombination system. Recombination to the Int(HOnc) substrate in msh2 and msh3 strains is not significantly increased when compared to recombination to the Int substrate, consistent with previous results examining recombination with short nonhomology (PâQUES and HABER 1997). The level of recombination in rad1 strains is similar to that of the msh2 and msh3 strains, but about threefold higher than with the Int(HO) substrates. This result suggests that some processing of short terminal nonhomology could be due to Rad1/Rad10 without the involvement of Msh2/Msh3. In general, these results show that the short length of terminal nonhomology has little effect on recombination in this system.

The role of MSH/RAD in crossover recombination:

If the terminal nonhomology present in this assay system is not responsible for the strong negative effects of deletion of msh2, msh3, or rad1 on recombination, then what accounts for the marked difference in the behavior of these genes on recombination in the reciprocal translocations observed here compared to the inverted repeat assay? Using a plasmid gap repair assay, SYMINGTON et al. (2000) found that noncrossover repair, likely using a type of synthesis-dependent strand annealing (SDSA) mechanism, was slightly increased over that of wild type in a rad1 background, whereas repair accompanied by crossovers was decreased by several-fold in a rad1 background. The explanation for this finding was that Rad1 was involved in cleaving the D-loop formed by strand invasion during the process of recombination and that this cleavage would stabilize the intermediate and help form a Holliday junction that could be resolved via a crossover. The absence of this cleavage would tend to result in displacement of the invading (elongated) single strand that would then pair with a single-strand extension at the other end of the DSB, resulting in a gene conversion event without crossover.

Our results are consistent with these previous results. In the assay in this study, almost all of the selectable events that we observe are crossovers and are strongly dependent on Rad1, as well as on Msh2 and Msh3. Moreover, as shown above, this dependence is not due to the presence of terminal nonhomology. Therefore, we hypothesize that to stabilize the Holliday junction required for a crossover, Rad1 would need to cleave the D-loop; in the absence of this cleavage, crossovers would be severely repressed. The work with the plasmid gap repair assay did not address the role of Rad1 vs. that of the complete MSH/RAD complex (SYMINGTON et al. 2000). Our current work suggests that the entire complex is needed for the D-loop cleavage, because crossover recombination is suppressed to essentially the same extent in msh2, msh3, or rad1 strains when there is no terminal nonhomology. A possible model for crossover recombination is illustrated in Figure 7A. How, then, can one reconcile these results with previous work that has shown no effect of Msh2 or Rad1 on crossing over in HO-mediated recombination (COLAIÁCOVO et al. 1999)? That work used relatively long stretches of homology, but later work using variable lengths of homology showed an effect of Msh2 and Rad1 on short substrates (IRA and Haber 2002). As illustrated in Figure 7A, given the position of the HO site in the recombination substrate, the D-loop created during recombination could be no longer than 200 bp for crossovers to be formed. Such a relatively short D-loop might be particularly unstable in the absence of cleavage by MSH/RAD. Alternatively, because the D-loop region is so short, the initial primer extension after strand invasion might usually proceed past the region of homology, leaving long nonhomologous tails that would have to be processed by MSH/RAD. Regardless of the explanation, it is clear from the previous work cited above that with longer regions of homology, the dependence on the MSH/RAD complex is abolished and so there would be no dependence of normal allelic recombination on MSH/RAD.

View larger version (26K): In this window In a new window Download PPT slide   FIGURE 7.— Recombination models. As explained in the text, these models are based in part on previous models (SYMINGTON et al. 2000). In the diagrams, the only sequences with any homology in the orientation depicted are the two homeologous intron sequences, represented as a red bar and a blue bar. The dotted boxes represent regions of heteroduplex with mismatched bases. (A) A possible mechanism for interchromosomal recombination in which a DSB in one intron sequence recombines with a homeologous sequence on another chromosome. After the HO DSB, the 3'-end invades the other chromosome, and the invading end is stabilized by the nucleolytic cleavage of the D-loop by the MSH/RAD complex. The stabilized Holliday junction is cleaved such that a crossover results. (B) A possible mechanism for the intrachromosomal inverted repeat assay that is RAD1 independent and sensitive to mismatches. The recombining DNAs represent sequences on one chromatid and the inverted sister chromatid (CHEN and JINKS-ROBERTSON 1998). The colored arrows represent the relative orientation of the intron sequences. After a DSB is formed, it is resected to form a gap and one 3'-end invades the homeologous sequence on the sister chromatid. DNA synthesis occurs, but the synthesized strand is extruded and reanneals with the homeologous sequence on the other side of the gap in the original chromatid, after which the gap is presumably filled by DNA synthesis.

 
Recombination with the inverted repeat assay likely proceeded via a gene conversion event involving the sister chromatid (CHEN and JINKS-ROBERTSON 1998), as illustrated in Figure 7B. As predicted by a SDSA mechanism, this recombination did not require Rad1, and in fact was slightly inhibited by Rad1 (NICHOLSON et al. 2000). Note here that, although the recombination substrates are the same size, the orientation of the substrates requires that the length of DNA synthesized after strand invasion is >1.5 kb (DATTA et al. 1996; NICHOLSON et al. 2000), which is substantially longer than in the case of the ectopic crossovers reported here. Although the recombination substrates themselves are identical, the predicted structures of the recombination intermediates are very different in the two cases, as are the genetic requirements for recombination.

The effects of mismatches on recombination:

The assay in this study is markedly less sensitive to mismatches than the inverted repeat assay. A substrate that would create four base–base mismatches in the recombinant heteroduplex caused a 22-fold decrease in recombination in the inverted repeat assay, whereas the same substrate reduced recombination by only 4-fold with the current assay (Figure 4). Given a choice of homologous and homeologous substrates, recombination to the homeologous substrate dropped by <2-fold, as measured most directly by the ratio of blue to white colonies. Although previous results with ectopic gene conversion had shown that the search for homology was a dynamic process (INBAR and KUPIEC 1999), our results indicate that for most of the crossovers that we observed there is no second chance for recombination to proceed. Otherwise, we would have expected to see a greater effect of homeology in a competitive situation, as those rejected recombination intermediates would have had another chance for recombination with a homologous substrate.

Recombination in the two assays is likely to be mechanistically different, and that difference could be responsible for the varying effects of mismatches on recombination. The model for each type of recombination illustrated in Figure 7 accounts for both the difference in response to the MSH/RAD complex and the evidence that the inverted repeat recombination is via gene conversion whereas the interchromosomal recombination is a crossover event. Both of these models have invasion of a single strand end that is not identical to the template strand, but then they differ in subsequent steps. The gene conversion–SDSA model requires that the extruded strand pairs with the broken chromosome at a region of mismatch whereas in the other model the D-loop must pair with the other end of the DSB.

It is instructive to compare our results with older work that examined the spontaneous recombination of homeologous SAM genes in yeast (BAILIS and ROTHSTEIN 1990). SAM1 and SAM2 are located on different chromosomes in relative orientations such that only gene conversions between the two genes are viable. Although completely homologous genes were not available for controls, it was estimated that the sequence differences between the two genes were responsible for an 8- to 23-fold decrease in recombination and that this recombinational suppression was partly relieved by a pms1 mutation (the only MMR gene examined) (BAILIS and ROTHSTEIN 1990). Mutations in RAD1 did not affect recombination, but unlike the inverted repeat assay, relief of suppression of recombination by pms1 was dependent on RAD1 function. These results stand in contrast to those presented here and are consistent with the different mode of recombination being responsible for the difference in sensitivity to mismatches. However, a study of ectopic recombination between the PMA1 and PMA2 genes showed that although recombination was severely repressed by sequence homeology, the ratio of gene conversion to crossover did not vary over a wide range of sequence homology (HARRIS et al. 1993). Therefore, rather than the mode of recombination, it might be the short D-loop that would be inherently less available for mismatch scanning.

Another possible explanation for the difference in the effect of mismatches in the two systems concerns the timing of the recombination events. The experiments mentioned above, as well as our previous inverted repeat assay, relied on spontaneous recombination. It is likely that most recombination events resulted from nicks that failed to be repaired before they were replicated, leading to double-strand ends (KUZMINOV 2001). In that case, the replication apparatus would already be at the double-strand end prior to recombination. Given the association of MMR proteins with replication proteins (JOHNSON et al. 1996a; UMAR et al. 1996; CHEN et al. 1999; HARFE and JINKS-ROBERTSON 2000a; BOWERS et al. 2001), the MMR apparatus would be immediately available to monitor such recombination. In contrast, DSBs created by HO endonuclease could occur at any time in the cell cycle and outside the context of replication.

Because most recombination events that we observe are dependent on proteins in the MSH/RAD complex, recombination frequencies cannot be used to examine the effects that this complex has on mismatches that occur during recombination. However, the use of competitive recombination substrates revealed that in msh2 strains there was no preference in choice of homologous over homeologous substrates. Given the loss of preference in msh6 strains, this finding was expected. However, essentially the same result was observed in msh3 and rad1 strains. Because neither Msh3 nor Rad1 would be expected to be involved in the recognition step of base–base mispairs, this result implies that the MSH/RAD complex is involved in a later step of mismatch rejection. If this complex is not present, then even though a mismatched heteroduplex is recognized, recombination will still proceed. In the inverted repeat assay, some, but not all, of the mismatch rejection was due to the MSH/RAD complex (NICHOLSON et al. 2000), whereas in the interchromosomal assay here, essentially all mismatch rejection appears to involve this complex.

More work will be required to understand why recombination between nonidentical sequences is sometimes blocked by mismatch repair proteins and in other circumstances seems to be relatively insensitive to sequence differences. Part of the answer may be that the current assay, like many others, detects what would normally be extremely rare events. Spontaneously occurring DSBs in diploid organisms would almost always be repaired from the sister chromatid if available or from the homolog if the sister chromatid were unavailable. Because of the proximity of sister chromatids, these events would be targeted to identical sequences. However, artificial DSBs are usually so efficiently induced that all instances of the cleavage site are cut, thereby making repair off of a sister chromatid impossible. There is thus no normal repair pathway in a haploid organism, and the cell is forced to recombine in alternative ways. In the assay in this study, most HO sites remain uncleaved and repair could take place off of the sister chromatid, but these events would not be detected. It remains to be determined how other similar types of induced, compared to spontaneous, recombination events occur in terms of sensitivity to regulation by mismatch repair proteins.

We are grateful for the technical assistance provided by Virginia Bowen, Amanda McCook, Lestell Alfred, and William Bowers. We thank Regan Gealy for providing probes and assistance with hybridization and Rachelle Spell for helpful discussions and review of this manuscript. We are especially grateful to Sue Jinks-Robertson for many fruitful interactions throughout the course of this work and preparation of the manuscript. Grant CA54050 from the National Institutes of Health supported this work.

¹ Present address: Centers for Disease Control and Prevention, Atlanta, GA 30333.

BAILIS, A. M., and R. ROTHSTEIN, 1990 A defect in mismatch repair in Saccharomyces cerevisiae stimulates ectopic recombination between homeologous genes by an excision repair dependent process. Genetics 126: 535–547.[Abstract]

BARDWELL, A. J., L. BARDWELL, A. E. TOMKINSON and E. C. FRIEDBERG, 1994 Specific cleavage of model recombination and repair intermediates by the yeast Rad1-Rad10 DNA endonuclease. Science 265: 2082–2085.[Abstract/Free Full Text]

BERTRAND, P., D. X. TISHKOFF, N. FILOSI, R. DASGUPTA and R. D. KOLODNER, 1998 Physical interaction between components of DNA mismatch repair and nucleotide excision repair. Proc. Natl. Acad. Sci. USA 95: 14278–14283.[Abstract/Free Full Text]

BOEKE, J. D., F. LACROUTE and G. R. FINK, 1984 A positive selection for mutants lacking orotidine-5'-phosphate decarboxylase activity in yeast: 5-fluoro-orotic acid resistance. Mol. Gen. Genet. 197: 345–346.[CrossRef][Medline]

BOWERS, J., P. T. TRAN, A. JOSHI, R. M. LISKAY and E. ALANI, 2001 MSH-MLH complexes formed at a DNA mismatch are disrupted by the PCNA sliding clamp. J. Mol. Biol. 306: 957–968.[CrossRef][Medline]

CHEN, C., B. J. MERRILL, P. J. LAU, C. HOLM and R. D. KOLODNER, 1999 Saccharomyces cerevisiae pol30 (proliferating cell nuclear antigen) mutations impair replication fidelity and mismatch repair. Mol. Cell. Biol. 19: 7801–7815.[Abstract/Free Full Text]

CHEN, W. L., and S. JINKS-ROBERTSON, 1998 Mismatch repair proteins regulate heteroduplex formation during mitotic recombination in yeast. Mol. Cell. Biol. 18: 6525–6537.[Abstract/Free Full Text]

CHEN, W., and S. JINKS-ROBERTSON, 1999 The role of the mismatch repair machinery in regulating mitotic and meiotic recombination between diverged sequences in yeast. Genetics 151: 1299–1313.[Abstract/Free Full Text]

CHURCH, G. M., and W. GILBERT, 1984 Genomic sequencing. Proc. Natl. Acad. Sci. USA 81: 1991–1995.[Abstract/Free Full Text]

COLAIÁCOVO, M. P., F. PâQUES and J. E. HABER, 1999 Removal of one nonhomologous DNA end during gene conversion by a RAD1- and MSH2-independent pathway. Genetics 151: 1409–1423.[Abstract/Free Full Text]

CROUSE, G. F., 1998 Mismatch repair systems in Saccharomyces cerevisiae, pp. 411–448 in DNA Damage and Repair, Vol. 1: DNA Repair in Prokaryotes and Lower Eukaryotes, edited by J. A. NICKOLOFF and M. F. HOEKSTRA. Humana Press, Totowa, NJ.

DATTA, A., A. ADJIRI, L. NEW, G. F. CROUSE and S. JINKS-ROBERTSON, 1996 Mitotic crossovers between diverged sequences are regulated by mismatch repair proteins in yeast. Mol. Cell. Biol. 16: 1085–1093.[Abstract]

DATTA, A., M. HENDRIX, M. LIPSITCH and S. JINKS-ROBERTSON, 1997 Dual roles for DNA sequence identity and the mismatch repair system in the regulation of mitotic crossing-over in yeast. Proc. Natl. Acad. Sci. USA 94: 9757–9762.[Abstract/Free Full Text]

DUTTWEILER, H. M., 1996 A highly sensitive and non-lethal ß-galactosidase plate assay for yeast. Trends Genet. 12: 340–341.[CrossRef][Medline]

EARLEY, M. C., and G. F. CROUSE, 1996 Selectable cassettes for simplified construction of yeast gene disruption vectors. Gene 169: 111–113.[CrossRef][Medline]

EARLEY, M. C., and G. F. CROUSE, 1998 The role of mismatch repair in the prevention of base pair mutations in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 95: 15487–15491.[Abstract/Free Full Text]

EVANS, E., and E. ALANI, 2000 Roles for mismatch repair factors in regulating genetic recombination. Mol. Cell. Biol. 20: 7839–7844.[Free Full Text]

EVANS, E., N. SUGAWARA, J. E. HABER and E. ALANI, 2000 The Saccharomyces cerevisiae Msh2 mismatch repair protein localizes to recombination intermediates in vivo. Mol. Cell 5: 789–799.[CrossRef][Medline]

FISHMAN-LOBELL, J., and J. E. HABER, 1992 Removal of nonhomologous DNA ends in double-strand break recombination: the role of the yeast ultraviolet repair gene RAD1. Science 258: 480–484.[Abstract/Free Full Text]

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.[CrossRef][Medline]

FLORES-ROZAS, H., and R. D. KOLODNER, 1998 The Saccharomyces cerevisiae MLH3 gene functions in MSH3-dependent suppression of frameshift mutations. Proc. Natl. Acad. Sci. USA 95: 12404–12409.[Abstract/Free Full Text]

GREENE, C. N., and S. JINKS-ROBERTSON, 1997 Frameshift intermediates in homopolymer runs are removed efficiently by yeast mismatch repair proteins. Mol. Cell. Biol. 17: 2844–2850.[Abstract]

HABER, J. E., 1998 A locus control region regulates yeast recombination. Trends Genet. 14: 317–321.[CrossRef][Medline]

HABRAKEN, Y., P. SUNG, L. PRAKASH and S. PRAKASH, 1996 Binding of insertion/deletion DNA mismatches by the heterodimer of yeast mismatch repair proteins MSH2 and MSH3. Curr. Biol. 6: 1185–1187.[CrossRef][Medline]

HARFE, B. D., and S. JINKS-ROBERTSON, 2000a DNA mismatch repair and genetic instability. Annu. Rev. Genet. 34: 359–399.[CrossRef][Medline]

HARFE, B. D., and S. JINKS-ROBERTSON, 2000b Mismatch repair proteins and mitotic genome stability. Mutat. Res. 451: 151–167.[Medline]

HARFE, B. D., B. K. MINESINGER and S. JINKS-ROBERTSON, 2000 Discrete in vivo roles for the MutL homologs Mlh2p and Mlh3p in the removal of frameshift intermediates in budding yeast. Curr. Biol. 10: 145–148.[CrossRef][Medline]

HARRIS, S., K. S. RUDNICKI and J. E. HABER, 1993 Gene conversions and crossing over during homologous and homeologous ectopic recombination in Saccharomyces cerevisiae. Genetics 135: 5–16.[Abstract]

HERRICK, D., and A. JACOBSON, 1992 A coding region segment is necessary, but not sufficient for rapid decay of the HIS3 mRNA in yeast. Gene 114: 35–41.[CrossRef][Medline]

HIGGINS, D. R., S. PRAKASH, P. REYNOLDS and L. PRAKASH, 1983 Molecular cloning and characterization of the RAD1 gene of Saccharomyces cerevisiae. Gene 26: 119–126.[CrossRef][Medline]

IACCARINO, I., F. PALOMBO, J. DRUMMOND, N. F. TOTTY, J. J. HSUAN et al., 1996 MSH6, a Saccharomyces cerevisiae protein that binds to mismatches as a heterodimer with MSH2. Curr. Biol. 6: 484–486.[CrossRef][Medline]

INBAR, O., and M. KUPIEC, 1999 Homology search and choice of homologous partner during mitotic recombination. Mol. Cell. Biol. 19: 4134–4142.[Abstract/Free Full Text]

INBAR, O., B. LIEFSHITZ, G. BITAN and M. KUPIEC, 2000 The relationship between homology length and crossing over during the repair of a broken chromosome. J. Biol. Chem. 275: 30833–30838.[Abstract/Free Full Text]

IRA, G., and J. E. HABER, 2002 Characterization of RAD51-independent break-induced replication that acts preferentially with short homologous sequences. Mol. Cell. Biol. 22: 6384–6392.[Abstract/Free Full Text]

IVANOV, E. L., and J. E. HABER, 1995 RAD1 and RAD10, but not other excision repair genes, are required for double-strand break-induced recombination in Saccharomyces cerevisiae. Mol. Cell. Biol. 15: 2245–2251.[Abstract]

JOHNSON, R. E., G. K. KOVVALI, S. N. GUZDER, N. S. AMIN, C. HOLM et al., 1996a Evidence for involvement of yeast proliferating cell nuclear antigen in DNA mismatch repair. J. Biol. Chem. 271: 27987–27990.[Abstract/Free Full Text]

JOHNSON, R. E., G. K. KOVVALI, L. PRAKASH and S. PRAKASH, 1996b Requirement of the yeast MSH3 and MSH6 genes for MSH2-dependent genomic stability. J. Biol. Chem. 271: 7285–7288.[Abstract/Free Full Text]

KIM, J. M., S. VANGURI, J. D. BOEKE, A. GABRIEL and D. F. VOYTAS, 1998 Transposable elements and genome organization: a comprehensive survey of retrotransposons revealed by the complete Saccharomyces cerevisiae genome sequence. Genome Res. 8: 464–478.[Abstract/Free Full Text]

KOSTRIKEN, R., J. N. STRATHERN, A. J. KLAR, J. B. HICKS and F. HEFFRON, 1983 A site-specific endonuclease essential for mating-type switching in Saccharomyces cerevisiae. Cell 35: 167–174.[CrossRef][Medline]

KRAMER, W., B. FARTMANN and E. C. RINGBECK, 1996 Transcription of mutS- and mutL-homologous genes in Saccharomyces cerevisiae during the cell cycle. Mol. Gen. Genet. 252: 275–283.[Medline]

KUZMINOV, A., 2001 DNA replication meets genetic exchange: chromosomal damage and its repair by homologous recombination. Proc. Natl. Acad. Sci. USA 98: 8461–8468.[Abstract/Free Full Text]

LANDER, E. S., L. M. LINTON, B. BIRREN, C. NUSBAUM, M. C. ZODY et al., 2001 Initial sequencing and analysis of the human genome. Nature 409: 860–921.[CrossRef][Medline]

MARSISCHKY, G. T., N. FILOSI, M. F. KANE and R. KOLODNER, 1996 Redundancy of Saccharomyces cerevisiae MSH3 and MSH6 in MSH2-dependent mismatch repair. Genes Dev. 10: 407–420.[Abstract/Free Full Text]

MUMBERG, D., R. MüLLER and M. FUNK, 1994 Regulatable promoters of Saccharomyces cerevisiae: comparison of transcriptional activity and their use for heterologous expression. Nucleic Acids Res. 22: 5767–5768.[Free Full Text]

NICHOLSON, A., M. HENDRIX, S. JINKS-ROBERTSON and G. F. CROUSE, 2000 Regulation of mitotic homeologous recombination in yeast: functions of mismatch repair and nucleotide excision repair genes. Genetics 154: 133–146.[Abstract/Free Full Text]

NICKOLOFF, J. A., J. D. SINGER and F. HEFFRON, 1990 In vivo analysis of the Saccharomyces cerevisiae HO nuclease recognition site by site-directed mutagenesis. Mol. Cell. Biol. 10: 1174–1179.[Abstract/Free Full Text]

PâQUES, F., and J. E. HABER, 1997 Two pathways for removal of nonhomologous DNA ends during double-strand break repair in Saccharomyces cerevisiae. Mol. Cell. Biol. 17: 6765–6771.[Abstract]

PROLLA, T. A., D.-M. CHRISTIE and R. M. LISKAY, 1994a Dual requirement in yeast DNA mismatch repair for MLH1 and PMS1, two homologs of the bacterial mutL gene. Mol. Cell. Biol. 14: 407–415.[Abstract/Free Full Text]

PROLLA, T. A., Q. PANG, E. ALANI, R. D. KOLODNER and R. M. LISKAY, 1994b MLH1, PMS1, and MSH2 interactions during the initiation of DNA mismatch repair in yeast. Science 265: 1091–1093.[Abstract/Free Full Text]

RICHARDSON, C., and M. JASIN, 2000 Coupled homologous and nonhomologous repair of a double-strand break preserves genomic integrity in mammalian cells. Mol. Cell. Biol. 20: 9068–9075.[Abstract/Free Full Text]

ROTHSTEIN, R., 1990 Targeting, disruption, replacement, and allele rescue: integrative DNA transformation in yeast. Methods Enzymol. 194: 281–301.

SAPARBAEV, M., L. PRAKASH and S. PRAKASH, 1996 Requirement of mismatch repair genes MSH2 and MSH3 in the RAD1–RAD10 pathway of mitotic recombination in Saccharomyces cerevisiae. Genetics 142: 727–736.[Abstract]

SIA, E. A., R. J. KOKOSKA, M. DOMINSKA, P. GREENWELL and T. D. PETES, 1997 Microsatellite instability in yeast: dependence on repeat unit size and DNA mismatch repair genes. Mol. Cell. Biol. 17: 2851–2858.[Abstract]

SIKORSKI, R. S., and P. HIETER, 1989 A system of shuttle vectors and yeast host strains designed for efficient manipulation of DNA in Saccharomyces cerevisiae. Genetics 122: 19–27.[Abstract/Free Full Text]

SPELL, R. M., and S. JINKS-ROBERTSON, 2004 Determination of mitotic recombination rates by fluctuation analysis in Saccharomyces cerevisiae. Methods Mol. Biol. 262: 3–12.[Medline]

STUDAMIRE, B., G. PRICE, N. SUGAWARA, J. E. HABER and E. ALANI, 1999 Separation-of-function mutations in Saccharomyces cerevisiae MSH2 that confer mismatch repair defects but do not affect nonhomologous-tail removal during recombination. Mol. Cell. Biol. 19: 7558–7567.[Abstract/Free Full Text]

SUGAWARA, N., F. PâQUES, M. COLAIÁCOVO and J. E. HABER, 1997 Role of Saccharomyces cerevisiae Msh2 and Msh3 repair proteins in double-strand break-induced recombination. Proc. Natl. Acad. Sci. USA 94: 9214–9219.[Abstract/Free Full Text]

SUN, H., D. TRECO and J. W. SZOSTAK, 1991 Extensive 3'-overhanging, single-stranded DNA associated with the meiosis-specific double-strand breaks at the ARG4 recombination initiation site. Cell 64: 1155–1161.[CrossRef][Medline]

SUNG, P., P. REYNOLDS, L. PRAKASH and S. PRAKASH, 1993 Purification and characterization of the Saccharomyces cerevisiae RAD1/RAD10 endonuclease. J. Biol. Chem. 268: 26391–26399.[Abstract/Free Full Text]

SYMINGTON, L. S., L. E. KANG and S. MOREAU, 2000 Alteration of gene conversion tract length and associated crossing over during plasmid gap repair in nuclease-deficient strains of Saccharomyces cerevisiae. Nucleic Acids Res. 28: 4649–4656.[Abstract/Free Full Text]

SZOSTAK, J. W., T. L. ORR-WEAVER, R. J. ROTHSTEIN and F. W. STAHL, 1983 The double-strand-break repair model for recombination. Cell 33: 25&#