A Test of the Double-Strand Break Repair Model for Meiotic Recombination in Saccharomyces Cerevisiae

We tested predictions of the double-strand break repair (DSBR) model for meiotic recombination by examining the segregation patterns of small palindromic insertions, which frequently escape mismatch repair when in heteroduplex DNA. The palindromes flanked a well characterized DSB site at the ARG4 locus. The ``canonical'''' DSBR model, in which only 5'' ends are degraded and resolution of the four-stranded intermediate is by Holliday junction resolvase, predicts that hDNA will frequently occur on both participating chromatids in a single event. Tetrads reflecting this configuration of hDNA were rare. In addition, a class of tetrads not predicted by the canonical DSBR model was identified. This class represented events that produced hDNA in a ``trans'''' configuration, on opposite strands of the same duplex on the two sides of the DSB site. Whereas most classes of convertant tetrads had typical frequencies of associated crossovers, tetrads with trans hDNA were parental for flanking markers. Modified versions of the DSBR model, including one that uses a topoisomerase to resolve the canonical DSBR intermediate, are supported by these data.     

Identification of New Genes Required for Meiotic Recombination in Saccharomyces Cerevisiae

Mutants defective in meiotic recombination were isolated from a disomic haploid strain of Saccharomyces cerevisiae by examining recombination within the leu2 and his4 heteroalleles located on chromosome III. The mutants were classified into two new complementation groups (MRE2 and MRE11) and eight previously identified groups, which include SPO11, HOP1, REC114, MRE4/MEK1 and genes in the RAD52 epistasis group. All of the mutants, in which the mutations in the new complementation groups are homozygous and diploid, can undergo premeiotic DNA synthesis and produce spores. The spores are, however, not viable. The mre2 and mre11 mutants produce viable spores in a spo13 background, in which meiosis I is bypassed, suggesting that these mutants are blocked at an early step in meiotic recombination. The mre2 mutant does not exhibit any unusual phenotype during mitosis and it is, thus, considered to have a mutation in a meiosis-specific gene. By contrast, the mre11 mutant is sensitive to damage to DNA by methyl methanesulfonate and exhibits a hyperrecombination phenotype in mitosis. Among six alleles of HOP1 that were isolated, an unusual pattern of intragenic complementation was observed.     

Analysis of Meiotic Recombination Pathways in the Yeast Saccharomyces Cerevisiae

In the yeast, Saccharomyces cerevisiae, several genes appear to act early in meiotic recombination. HOP1 and RED1 have been classified as such early genes. The data in this paper demonstrate that neither a red1 nor a hop1 mutation can rescue the inviable spores produced by a rad52 spo13 strain; this phenotype helps to distinguish these two genes from other early meiotic recombination genes such as SPO11, REC104, or MEI4. In contrast, either a red1 or a hop1 mutation can rescue a rad50S spo13 strain; this phenotype is similar to that conferred by mutations in the other early recombination genes (e.g., REC104). These two different results can be explained because the data presented here indicate that a rad50S mutation does not diminish meiotic intrachromosomal recombination, similar to the mutant phenotypes conferred by red1 or hop1. Of course, RED1 and HOP1 do act in the normal meiotic interchromosomal recombination pathway; they reduce interchromosomal recombination to ~10% of normal levels. We demonstrate that a mutation in a gene (REC104) required for initiation of exchange is completely epistatic to a mutation in RED1. Finally, mutations in either HOP1 or RED1 reduce the number of double-strand breaks observed at the HIS2 meiotic recombination hotspot.     

Isolation of Com1, a New Gene Required to Complete Meiotic Double-Strand Break-Induced Recombination in Saccharomyces Cerevisiae

We have designed a screen to isolate mutants defective during a specific part of meiotic prophase I of the yeast Saccharomyces cerevisiae. Genes required for the repair of meiotic double-strand breaks or for the separation of recombined chromosomes are targets of this mutant hunt. The specificity is achieved by selecting for mutants that produce viable spores when recombination and reductional segregation are prevented by mutations in SPO11 and SPO13 genes, but fail to yield viable spores during a normal Rec(+) meiosis. We have identified and characterized a mutation com1-1, which blocks processing of meiotic double-strand breaks and which interferes with synaptonemal complex formation, homologous pairing and, as a consequence, spore viability after induction of meiotic recombination. The COM1/SAE2 gene was cloned by complementation, and the deletion mutant has a phenotype similar to com1-1. com1/sae2 mutants closely resemble the phenotype of rad50S, as assayed by phase-contrast microscopy for spore formation, physical and genetic analysis of recombination, fluorescence in situ hybridization to quantify homologous pairing and immunofluorescence and electron microscopy to determine the capability to synapse axial elements.     

The Sep1 Mutant of Saccharomyces Cerevisiae Arrests in Pachytene and Is Deficient in Meiotic Recombination

Strand exchange protein 1 (Sep1) from Saccharomyces cerevisiae promotes homologous pairing of DNA in vitro and sep1 mutants display pleiotropic phenotypes in both vegetative and meiotic cells. In this study, we examined in detail the ability of the sep1 mutant to progress through meiosis I prophase and to undergo meiotic recombination. In meiotic return-to-growth experiments, commitment to meiotic recombination began at the same time in wild type and mutant; however, recombinants accumulated at decreased rates in the mutant. Gene conversion eventually reached nearly wild-type levels, whereas crossing over reached 15-50% of wild type. In an assay of intrachromosomal pop-out recombination, the sep1, dmc1 and rad51 single mutations had only small effects; however, pop-out recombination was virtually eliminated in the sep1 dmc1 and sep1 rad51 double mutants, providing evidence for multiple recombination pathways. Analysis of meiotic recombination intermediates indicates that the sep1 mutant is deficient in meiotic double-strand break repair. In a physical assay, the formation of mature reciprocal recombinants in the sep1 mutant was delayed relative to wild type and ultimately reached only 50% of the wild-type level. Electron microscopic analysis of meiotic nuclear spreads indicates that the sep1δ mutant arrests in pachytene, with apparently normal synaptonemal complex. This arrest is RAD9-independent. We hypothesize that the Sep1 protein participates directly in meiotic recombination and that other strand exchange enzymes, acting in parallel recombination pathways, are able to substitute partially for the absence of the Sep1 protein.     

Genetic Analysis of a Meiotic Recombination Hotspot on Chromosome III of Saccharomyces Cerevisiae

In a previous study, we analyzed meiotic recombination events that occurred in the 22-kb region (LEU2 to CEN3) of chromosome III of Saccharomyces cerevisiae. We found one region with an enhanced level of crossovers (a hotspot) and one region with a depressed level of crossovers. In this study, we show that about one-third of the crossovers that occur between LEU2 and CEN3 are initiated in a 1.3-kb region located approximately 6 kb from the centromere. Both crossovers and gene conversion events are initiated at this site. Events initiated at this position can be resolved as crossovers in regions located either centromere-distally or centromere-proximally from the initiation site.     

Competition between Adjacent Meiotic Recombination Hotspots in the Yeast Saccharomyces Cerevisiae

In a wild-type strain of Saccharomyces cerevisiae, a hotspot for meiotic recombination is located upstream of the HIS4 gene. An insertion of a 49-bp telomeric sequence into the coding region of HIS4 strongly stimulates meiotic recombination and the local formation of meiosis-specific double-strand DNA breaks (DSBs). When strains are constructed in which both hotspots are heterozygous, hotspot activity is substantially less when the hotspots are on the same chromosome than when they are on opposite chromosomes.     

Analysis of a Recombination Hotspot for Gene Conversion Occurring at the His2 Gene of Saccharomyces Cerevisiae

The properties of gene conversion as measured in fungi that generate asci containing all the products of meiosis imply that meiotic recombination initiates at specific sites. The HIS2 gene of Saccharomyces cerevisiae displays a high frequency of gene conversion, indicating that it is a recombination hotspot. The HIS2 gene was cloned and sequenced, and the cloned DNA was used to make several different types of alterations in the yeast chromosome by transformation; these alterations were used to determine the location of the sequences necessary for the high levels of meiotic conversion observed at HIS2. Previous work indicated that the gene conversion polarity gradient is high at the 3' end of the gene, and that the promoter of the gene is not necessary for the high frequency of conversion observed. Data presented here suggest that at least some of the sequences necessary for high levels of conversion at HIS2 are located over 700 bp downstream of the end of the coding region, extend over (at least) several hundred base pairs, and may be quite complex, perhaps involving chromatin structure. Additional data indicate that multiple single base heterologies within a 1-kb interval contribute little to the frequency of gene conversion. This contrasts with other reports about the role of heterologies at the MAT locus.     

Rdh54, a Rad54 Homologue in Saccharomyces Cerevisiae, Is Required for Mitotic Diploid-Specific Recombination and Repair and for Meiosis

Most mitotic recombination and repair genes of Saccharomyces cerevisiae show no specificity of action for the genome ploidy. We describe here a novel repair and recombination gene that is specific for recombination and repair between homologous chromosomes. The RDH54 gene is homologous to the RAD54 gene, but rdh54 mutants do not show sensitivity to methyl methanesulfonate at concentrations that sensitize a rad54 mutant. However, the rdh54 null mutation enhances the methyl methanesulfonate sensitivity of a rad54 mutant and single rdh54 mutants are sensitive to prolonged exposure at high concentrations of methyl methanesulfonate. The RDH54 gene is required for recombination, but only in a diploid. We present evidence showing that the RDH54 gene is required for interhomologue gene conversion but not intrachromosomal gene conversion. The rdh54 mutation confers diploid-specific lethalities and reduced growth in various mutant backgrounds. These phenotypes are due to attempted recombination. The RDH54 gene is also required for meiosis as homozygous mutant diploids show very poor sporulation and reduced spore viability. The role of the RDH54 gene in mitotic repair and in meiosis and the pathway in which it acts are discussed.     

Interaction between Mismatch Repair and Genetic Recombination in Saccharomyces Cerevisiae

The yeast Saccharomyces cerevisiae encodes a set of genes that show strong amino acid sequence similarity to MutS and MutL, proteins required for mismatch repair in Escherichia coli. We examined the role of MSH2 and PMS1, yeast homologs of mutS and mutL, respectively, in the repair of base pair mismatches formed during meiotic recombination. By using specifically marked HIS4 and ARG4 alleles, we showed that msh2 mutants displayed a severe defect in the repair of all base pair mismatches as well as 1-, 2- and 4-bp insertion/deletion mispairs. The msh2 and pms1 phenotypes were indistinguishable, suggesting that the wild-type gene products act in the same repair pathway. A comparison of gene conversion events in wild-type and msh2 mutants indicated that mismatch repair plays an important role in genetic recombination. (1) Tetrad analysis at five different loci revealed that, in msh2 mutants, the majority of aberrant segregants displayed a sectored phenotype, consistent with a failure to repair mismatches created during heteroduplex formation. In wild type, base pair mismatches were almost exclusively repaired toward conversion rather than restoration. (2) In msh2 strains 10-19% of the aberrant tetrads were Ab4:4. (3) Polarity gradients at HIS4 and ARG4 were nearly abolished in msh2 mutants. The frequency of gene conversion at the 3' end of these genes was increased and was nearly the frequency observed at the 5' end. (4) Co-conversion studies were consistent with mismatch repair acting to regulate heteroduplex DNA tract length. We favor a model proposing that recombination events occur through the formation and resolution of heteroduplex intermediates and that mismatch repair proteins specifically interact with recombination enzymes to regulate the length of symmetric heteroduplex DNA.     

Meiotic Nondisjunction and Recombination of Chromosome III and Homologous Fragments in Saccharomyces Cerevisiae

A yeast strain, in which nondisjunction of chromosome III at the first-meiotic division could be assayed, was constructed. Using chromosome fragmentation plasmids, chromosomal fragments (CFs) were derived in isogenic strains from six sites along chromosome III and one site on chromosome VII. Whereas the presence of the CFs derived from chromosome III increased considerably the meiosis I nondisjunction of that chromosome, the CF derived from chromosome VII had no effect on chromosome III segregation. The effects of the chromosome III-derived fragments were not linearly related to fragment length. Two regions, one of 12 kb in size located at the left end of the chromosome, and the other of 5 kb, located at the center of the right arm, were found to have profound effects on chromosome III nondisjunction. Most disomics arising from meioses in strains containing chromosome III CFs did not contain the CF; thus it appears that the two chromosome III homologs had segregated away from the CF. Among the disomics, recombination between the homologous chromosomes III was lower than expected from the genetic distance, while recombination between one of the chromosomes III and the fragment was frequent. We suggest that there are sites along the chromosome that are more involved than others in the pairing of homologous chromosomes and that the pairing between fragment and homologs involves recombination among these latter elements.     

Unrepaired Heteroduplex DNA in Saccharomyces Cerevisiae Is Decreased in Rad1 Rad52-Independent Recombination

A direct repeat recombination assay between SUP4 heteroalleles detects unrepaired heteroduplex DNA (hDNA) as sectored colonies. The frequency of unrepaired heteroduplex is dependent on the mismatch and is highest in a construct that generates C:C or G:G mispairs and lowest in one that generates T:G or C:A mispairs. In addition, unrepaired hDNA increases for all mismatches tested in pms1 mismatch repair-deficient strains. These results support the notion that hDNA is formed across the SUP4 repeats during the recombination event and is then subject to mismatch repair. The effects of various repair and recombination defective mutations on this assay were examined. Unrepaired heteroduplex increases significantly only in rad52 mutant strains. In addition, direct repeat recombination is reduced 2-fold in rad52 mutant strains, while in rad51, rad54, rad55 and rad57 mutants direct repeat recombination is increased 3-4-fold. Mutations in the excision repair gene, RAD1, do not affect the frequency of direct repeat recombination. However, the level of unrepaired heteroduplex is slightly decreased in rad1 mutant strains. Similar to previous studies, rad1 rad52 double mutants show a synergistic reduction in direct repeat recombination (35-fold). Interestingly, unrepaired heteroduplex is reduced 4-fold in the double mutants. Experiments with shortened repeats suggest that the reduction in unrepaired heteroduplex is due to decreased hDNA tract length in the double mutant strain.     

MEIOB Targets Single-Strand DNA and Is Necessary for Meiotic Recombination

Meiotic recombination is a mandatory process for sexual reproduction. We identified a protein specifically implicated in meiotic homologous recombination that we named: meiosis specific with OB domain (MEIOB). This protein is conserved among metazoan species and contains single-strand DNA binding sites similar to those of RPA1. Our studies in vitro revealed that both recombinant and endogenous MEIOB can be retained on single-strand DNA. Those in vivo demonstrated the specific expression of Meiob in early meiotic germ cells and the co-localization of MEIOB protein with RPA on chromosome axes. MEIOB localization in Dmc1 ^−/− spermatocytes indicated that it accumulates on resected DNA. Homologous Meiob deletion in mice caused infertility in both sexes, due to a meiotic arrest at a zygotene/pachytene-like stage. DNA double strand break repair and homologous chromosome synapsis were impaired in Meiob ^−/− meiocytes. Interestingly MEIOB appeared to be dispensable for the initial loading of recombinases but was required to maintain a proper number of RAD51 and DMC1 foci beyond the zygotene stage. In light of these findings, we propose that RPA and this new single-strand DNA binding protein MEIOB, are essential to ensure the proper stabilization of recombinases which is required for successful homology search and meiotic recombination.     

Analysis of Mitotic and Meiotic Defects in Saccharomyces Cerevisiae Srs2 DNA Helicase Mutants

The hyper-gene conversion srs2-101 mutation of the SRS2 DNA helicase gene of Saccharomyces cerevisiae has been reported to suppress the UV sensitivity of rad18 mutants. New alleles of SRS2 were recovered using this suppressor phenotype. The alleles have been characterized with respect to suppression of rad18 UV sensitivity, hyperrecombination, reduction of meiotic viability, and definition of the mutational change within the SRS2 gene. Variability in the degree of rad18 suppression and hyperrecombination were found. The alleles that showed the severest effects were found to be missense mutations within the consensus domains of the DNA helicase family of proteins. The effect of mutations in domains I (ATP-binding) and V (proposed DNA binding) are reported. Some alleles of SRS2 reduce spore viability to 50% of wild-type levels. This phenotype is not bypassed by spo13 mutation. Although the srs2 homozygous diploids strains undergo normal commitment to meiotic recombination, this event is delayed by several hours in the mutant strains and the strains appear to stall in the progression from meiosis I to meiosis II.     

Physical Lengths of Meiotic and Mitotic Gene Conversion Tracts in Saccharomyces Cerevisiae

Physical lengths of gene conversion tracts for meiotic and mitotic conversions were examined, using the same diploid yeast strain in all experiments. This strain is heterozygous for a mutation in the URA3 gene as well as closely linked restriction site markers. In cells that had a gene conversion event at the URA3 locus, it was determined by Southern analysis which of the flanking heterozygous restriction sites had co-converted. It was found that mitotic conversion tracts were longer on the average than meiotic tracts. About half of the tracts generated by spontaneous mitotic gene conversion included heterozygous markers 4.2 kb apart; none of the meiotic conversions included these markers. Stimulation of mitotic gene conversion by ultraviolet light or methylmethanesulfonate had no obvious effect on the size or distribution of the tracts. Almost all conversion tracts were continuous.     

Isolation of Mutants Defective in Early Steps of Meiotic Recombination in the Yeast Saccharomyces Cerevisiae

Using a selection based upon the ability of early Rec- mutations (e.g., rad50) to rescue the meiotic lethality of a rad52 spo13 strain, we have isolated 177 mutants. Analysis of 56 of these has generated alleles of the known Rec genes SPO11, ME14 and MER1, as well as defining five new genes: REC102, REC104, REC107, REC113 and REC114. Mutations in all of the new genes appear to specifically affect meiosis; they do not have any detectable mitotic phenotype. Mutations in REC102, REC104 and REC107 reduce meiotic recombination several hundred fold. No alleles of RED1 or HOP1 were isolated, consistent with the proposal that these genes may be primarily involved with chromosome pairing and not exchange.     

Me14, a Yeast Gene Required for Meiotic Recombination

Mutants at the MEI4 locus were detected in a search for mutants defective in meiotic gene conversion. mei4 mutants exhibit decreased sporulation and produce inviable spores. The spore inviability phenotype is rescued by a spo13 mutation, which causes cells to bypass the meiosis I division. The MEI4 gene has been cloned from a yeast genomic library by complementation of the recombination defect and has been mapped to chromosome V near gln3. Strains carrying a deletion/insertion mutation of the MEI4 gene display no meiotically induced gene conversion but normal mitotic conversion frequencies. Both meiotic interchromosomal and intrachromosomal crossing over are completely abolished in mei4 strains. The mei4 mutation is able to rescue the spore-inviability phenotype of spo13 and 52 strains (i.e., mei4 spo13 rad52 mutants produce viable spores), indicating that MEI4 acts before RAD52 in the meiotic recombination pathway.     

Genetic and Physical Analysis of Double-Strand Break Repair and Recombination in Saccharomyces Cerevisiae

We have investigated HO endonuclease-induced double-strand break (DSB) recombination and repair in a LACZ duplication plasmid in yeast. A 117-bp MATa fragment, embedded in one copy of LACZ, served as a site for initiation of a DSB when HO endonuclease was expressed. The DSB could be repaired using wild-type sequences located on a second, promoterless, copy of LACZ on the same plasmid. In contrast to normal mating-type switching, crossing-over associated with gene conversion occurred at least 50% of the time. The proportion of conversion events accompanied by exchange was greater when the two copies of LACZ were in direct orientation (80%), than when inverted (50%). In addition, the fraction of plasmids lost was significantly greater in the inverted orientation. The kinetics of appearance of intermediates and final products were also monitored. The repair of the DSB is slow, requiring at least an hour from the detection of the HO-cut fragments to completion of repair. Surprisingly, the appearance of the two reciprocal products of crossing over did not occur with the same kinetics. For example, when the two LACZ sequences were in the direct orientation, the HO-induced formation of a large circular deletion product was not accompanied by the appearance of a small circular reciprocal product. We suggest that these differences may reflect two kinetically separable processes, one involving only one cut end and the other resulting from the concerted participation of both ends of the DSB.     

Multiple Pathways for Homologous Recombination in Saccharomyces Cerevisiae

The genes in the RAD52 epistasis group of Saccharomyces cerevisiae are necessary for most mitotic and meiotic recombination events. Using an intrachromosomal inverted-repeat assay, we previously demonstrated that mitotic recombination of this substrate is dependent upon the RAD52 gene. In the present study the requirement for other genes in this epistasis group for recombination of inverted repeats has been analyzed, and double and triple mutant strains were examined for their epistatic relationships. The majority of recombination events are mediated by a RAD51-dependent pathway, where the RAD54, RAD55 and RAD57 genes function downstream of RAD51. Cells mutated in RAD55 or RAD57 as well as double mutants are cold-sensitive for inverted-repeat recombination, whereas a rad51 rad55 rad57 triple mutant is not. The RAD1 gene is not required for inverted-repeat recombination but is able to process spontaneous DNA lesions to produce recombinant products in the absence of RAD51. Furthermore, there is still considerably more recombination in rad1 rad51 mutants than in rad52 mutants, indicating the presence of another, as yet unidentified, recombination pathway.