Yeast Meiotic Mutants Proficient for the Induction of Ectopic Recombination

Corresponding author: JoAnne Engebrecht, Department of Pharmacological Sciences, State University of New York, Stony Brook, NY 11794-8651, joanne{at}pharm.som.sunysb.edu (E-mail).

Communicating editor: S. JINKS-ROBERTSON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

A screen was designed to identify Saccharomyces cerevisiae mutants that were defective in meiosis yet proficient for meiotic ectopic recombination in the return-to-growth protocol. Seven mutants alleles were isolated; two are important for chromosome synapsis (RED1, MEK1) and five function independently of recombination (SPO14, GSG1, SPOT8/MUM2, 3, 4). Similar to the spoT8-1 mutant, mum2 deletion strains do not undergo premeiotic DNA synthesis, arrest prior to the first meiotic division and fail to sporulate. Surprisingly, although DNA replication does not occur, mum2 mutants are induced for high levels of ectopic recombination. gsg1 diploids are reduced in their ability to complete premeiotic DNA synthesis and the meiotic divisions, and a small percentage of cells produce spores. mum3 mutants sporulate poorly and the spores produced are inviable. Finally, mum4-1 mutants produce inviable spores. The meiotic/sporulation defects of gsg1, mum2, and mum3 are not relieved by spo11 or spo13 mutations, indicating that the mutant defects are not dependent on the initiation of recombination or completion of both meiotic divisions. In contrast, the spore inviability of the mum4-1 mutant is rescued by the spo13 mutation. The mum4-1 spo13 mutant undergoes a single, predominantly equational division, suggesting that MUM4 functions at or prior to the first meiotic division. Although recombination is variably affected in the gsg1 and mum mutants, we hypothesize that these mutants define genes important for aspects of meiosis not directly related to recombination.

MEIOSIS enables diploid organisms to reproduce sexually by generating haploid gametes through two successive divisions. At meiosis I, the reductional division, homologous chromosomes disjoin from each other. At meiosis II, as at mitosis, sister chromatids separate and move to opposite poles. The meiotic divisions, in turn, are tightly linked to gamete differentiation. Fusion of gametes at fertilization restores the diploid chromosome number and initiates zygotic development.

Once meiosis is initiated, DNA replication occurs. Several differences have been observed between premeiotic and mitotic DNA synthesis. For instance, in Triturus vulgaris, Mus musculus, and Saccharomyces cerevisiae, the length of premeiotic S phase is significantly longer than mitotic S (CALLAN 1972 ; KOFMAN-ALFARO and CHANDLEY 1970 ; WILLIAMSON et al. 1983 ). The basis for the difference between premeiotic and mitotic DNA synthesis is unknown; in fact, studies in S. cerevisiae have shown that the DNA replication rate and the origins of replication used are the same in both replication modes ( JOHNSTON et al. 1982 ; COLLINS and NEWLON 1994 ).

Premeiotic DNA synthesis is followed by a lengthy prophase in which homologous chromosomes pair and synapse and high levels of recombination occur. Chromosome pairing refers to the presynaptic alignment of the homologous chromosomes; this is temporally and genetically distinct from chromosome synapsis (LOIDL et al. 1994 ; WEINER and KLECKNER 1994 ). Chromosome synapsis is the intimate association of homologous chromosomes in the context of the synaptonemal complex (SC). The SC is a meiosis-specific structure that is elaborated along the lengths of the chromosomes during prophase (reviewed in VON WETTSTEIN et al. 1984 ). Although the relationship between chromosome pairing, synapsis, and recombination has not been precisely delineated, there is convincing evidence that recombination is initiated before the SC is fully formed; however, mature recombinants do not appear until after chromosome synapsis is complete (PADMORE et al. 1991 ; GOYON and LICHTEN 1993 ). Reciprocal crossing over occurring in the context of the SC establishes physical connections between homologous chromosomes (ENGEBRECHT et al. 1990 ). These connections, termed chiasmata, ensure proper spindle attachment and alignment on the metaphase plate (reviewed in HAWLEY 1987 ). At the end of prophase, the SC breaks down; however, the connections between homologous chromosomes are presumably maintained until anaphase when homologous chromosomes segregate to opposite poles.

At meiosis II, sister chromatids segregate from each other. The second meiotic division differs from mitosis, most basically, in that it follows meiosis I and is not directly preceded by a round of DNA replication. This requires that the mitotic cell-cycle controls that tightly regulate the ordered occurrence of S phase and chromosome segregation be significantly altered in meiosis (MCCARROLL and ESPOSITO 1994 ). Finally, the two meiotic divisions must be temporally controlled and, in turn, coordinated with gamete differentiation.

Genetic analysis of meiosis in the yeast S. cerevisiae has significantly advanced our understanding of meiotic chromosome behavior (reviewed in ROEDER 1995 ; KLECKNER 1996 ). Meiotic mutants that perturb premeiotic DNA synthesis without affecting mitotic S phase have been isolated (mei1, mei2,3, ROTH 1973 ; spo7, spo9, ESPOSITO and ESPOSITO 1974A ; spoT1-spoT11, TSUBOI 1983 ); however, the functions of the corresponding gene products have not been elucidated. Consequently, little is known about the unique aspects of premeiotic DNA synthesis. In contrast, there are a large number of meiotic mutants that define genes required for chromosome synapsis and recombination (reviewed in KUPIEC et al. 1997 ). These mutants have been identified in genetic screens that look for the production of inviable meiotic products or defects in meiotic recombination by using specially marked strains (reviewed in LOIDL et al. 1997 ). Molecular and cytological analyses of the corresponding gene products have shown that they encode structural proteins of the SC and components of the recombination machinery (reviewed in KUPIEC et al. 1997 ). Recently, a number of mutants have been identified that arrest prior to the meiosis I division (ndt80, XU et al. 1995 ; com1/sae2, PRINZ et al. 1997 ; MCKEE and KLECKNER 1997A ; sae1, sae3, MCKEE and KLECKNER 1997B ). Except for ndt80, the meiotic arrest observed in these mutants is dependent on the initiation of recombination, suggesting that these mutations define genes important for the completion of meiotic recombination.

We initiated a mutant hunt to isolate recombination-proficient meiotic yeast mutants by examining the induction of recombination between an endogenous locus and an artificial insert placed at a nonhomologous site in the genome (ectopic recombination). Previous studies in yeast have suggested that ectopic recombination is mechanistically similar to allelic recombination (JINKS-ROBERTSON and PETES 1986 ; LICHTEN et al. 1987 ; HABER et al. 1991 ; STEELE et al. 1991 ), although the frequency of ectopic recombination is influenced by the locations of the recombining sequences (GOLDMAN and LICHTEN 1996 ). We reasoned that a subset of meiotic mutants proficient for the induction of ectopic recombination would define genes required for processes other than recombination and chromosome synapsis such as premeiotic DNA synthesis, chromosome pairing and coordination of meiosis and spore formation. In addition to isolating alleles of previously characterized genes that function independently of recombination (SPO14, HONIGBERG et al. 1992 ; ROSE et al. 1995 ) and are important for chromosome synapsis (RED1, ROCKMILL and ROEDER 1990 ; SMITH and ROEDER 1997 ; MEK1, ROCKMILL and ROEDER 1991 ; LEEM and OGAWA 1992 ), mutations in GSG1 (KAYTOR and LIVINGSTON 1995 ), SPOT8/MUM2 (TSUBOI 1983 ), MUM3, and MUM4 (MUddled Meiosis) were identified and characterized.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Yeast strains and genetic manipulations:
Yeast media were prepared and genetic methods were carried out as described by ROSE et al. 1990 . YPAD is YEPD medium supplemented with 100 µM adenine. YPAD medium containing 0.02% MMS (methyl methanesulfonic acid; Sigma) was used to determine MMS sensitivity of candidate mutants. Yeast strains were transformed using the lithium acetate procedure of ITO et al. 1983 . All integrative transformants were verified by Southern blot analysis (SOUTHERN 1975 ). All gene disruptions were performed by the one-step gene replacement described by ROTHSTEIN 1991 .

The genotypes of yeast strains used in this study are listed in Table 1. JB128 was obtained from JAYA BHARGAVA, BR2171-7B from BETH ROCKMILL, and MTM-964 from CRAIG GIROUX (Wayne State University, Detroit).

spo11::TRP1 (Y344) and spo13::URA3 (Y686) derivatives of strain Y315 were made by transforming the haploid parents with pGB324, obtained from CRAIG GIROUX, and pNKY58 (ALANI et al. 1987 ), respectively, and then mating the transformants. Strains Y366, Y369, Y693, and Y1069 for the physical recombination assay were constructed by introducing XbaI-Bgl II digested pJH118 and pJH119 (BORTS et al. 1986 ) into the corresponding haploids and mating the transformants.

Plasmid constructions:
The original GSG1-complementing plasmid was designated pME556. The gsg1::LYS2 allele was constructed in two steps. First, the BamHI-PvuII fragment from pME556 was inserted into HindIII-BamHI digested pHSS6 after the HindIII end was filled in using the Klenow fragment of DNA polymerase I. The resulting plasmid, pME699, was restricted with KpnI, followed by removal of the single-stranded ends with T4 DNA polymerase, and XbaI; the PvuII-XbaI LYS2 fragment from pDP6 (FLEIG et al. 1986 ) was then inserted to create pME704. This plasmid was digested with Not I prior to transformation into yeast.

The original MUM2-complementing plasmid was designated pME759. A 3.5-kb Bgl II-Sal I fragment from pME759 was subcloned into the BamHI-Sal I sites of pUN55 (ELLEDGE and DAVIS 1988 ), generating pME897. A 2.5-kb HindIII-Sal I fragment from pME759 was ligated to HindIII-Sal I digested pUN105 (ELLEDGE and DAVIS 1988 ) to create pME898. MUM2 was moved into a TRP1 CEN vector, pUN15 (ELLEDGE and DAVIS 1988 ), by inserting the 2.5-kb Not I-Sal I fragment from pME898 into the corresponding sites of pUN15; the resulting plasmid was designated pME1141. To create the mum2::LEU2 allele in pME917, pME897 was digested with EcoRI and ClaI, followed by filling in the ends with the Klenow fragment of DNA polymerase I, and the HpaI-Nar I LEU2 fragment from YEp351 (HILL et al. 1986 ), whose ends had been filled in with the Klenow fragment of DNA polymerase I, was inserted. The mum2::LYS2 allele in pME1026 was constructed by inserting the PvuII-XbaI LYS2 fragment from pDP6, whose ends had been filled in with the Klenow fragment of DNA polymerase I, into the EcoRI and ClaI sites of pME897, whose ends had also been filled in with the Klenow fragment of DNA polymerase I. Plasmids ME917 and ME1026 were digested with PvuII to allow substitution of the MUM2 locus in yeast.

The original MUM3-complementing plasmid contained a 11.5-kb Sau3A partial yeast genomic fragment at the BamHI site of YCp50 and was designated pKR588. A 5-kb ClaI fragment that retained complementing activity was inserted into the ClaI site of pHSS6 (SEIFERT et al. 1986 ) to generate pKR718. This plasmid was subjected to transposon mutagenesis (SEIFERT et al. 1986 ) and a plasmid containing a transposon marked with the LEU2 gene inserted at the 5'-end of the MUM3 ORF, pKR765, was isolated. Plasmid KR765 was digested with Not I prior to transformation into yeast to allow substitution of the MUM3 locus. The mum3::LYS2 disruption allele was created in two steps. First, the 1.7-kb EcoRI-Sal I fragment from pKR765 was inserted into the corresponding sites in Bluescript SK+ (Stratagene) to create pME1133. The LYS2 gene on a XbaI fragment, whose ends had been filled in with the Klenow fragment of DNA polymerase I, was inserted at the Bgl II site, whose ends had also been filled in with the Klenow fragment of DNA polymerase I, at position 851 of the MUM3 ORF to create pME1138. Plasmid ME1138 was digested with ApaI and Not I prior to transformation into yeast.

Mutant screen protocol:
Spores from JB128 were treated with 10 mM DTT and 1 mg/ml Zymolyase 100T for approximately three hours to remove the ascus cell wall from greater than 90% of the tetrads. The spores were collected, washed with water, resuspended in 1 ml of 0.01% Nonidet P-40 and sonicated on ice for 30 sec using a sonicator equipped with a microtip. The haploid spores were plated for single colonies onto solid YPAD medium and the plates immediately exposed to UV using a 8 watt long wave ultraviolet light source (350–400 nm). Exposure to UV was calibrated to obtain approximately 50% viability (5 sec of exposure at a distance of 12 inches from the source). Approximately 50,000 mutagenized colonies were replica-plated to sporulation medium to induce sporulation and after 3 to 4 days replica-plated to YPAD medium and immediately exposed to ether vapors as described by ROCKMILL and ROEDER 1988 . Briefly, the bottom of the YPAD plates were inverted over glass petri lids containing Whatman #1 filters soaked with ether (approximately 0.75 ml) for 15 min in a fume hood. Ether was reapplied to the filter in the lids and the cells were exposed for an additional 15 min. The bottom of the plates were then removed and placed ajar on their own lids for 30–60 min to allow the ether to dissipate. The plates were incubated overnight at 30° and the colonies were scored for growth the next day. Ether-sensitive mutants were streaked for single colonies and retested for ether-sensitivity after being induced for sporulation. Candidate mutants were analyzed by phase-contrast microscopy and/or tetrad dissection; 108 mutants were recovered. Ten strains were sensitive to MMS and not further characterized. The remaining mutants were assayed for commitment to meiotic ectopic recombination; twelve mutants displayed at least 50% of wild-type levels of ectopic recombination. These mutants were analyzed for meiotic progression as described below; five mutants appeared to undergo the meiotic divisions at normal levels and with normal kinetics yet failed to produce spores and were not further characterized.

Doubling times:
Overnight cultures were diluted 1:200 in YPAD at a starting density of ~5 x 10⁵ cells/ml and grown with aeration at 30°. After 4 hr of growth, aliquots were removed every 2 hr and fixed with 3.7% formaldehyde. Each sample was counted three times in a hemocytometer and the mean cell density determined. Doubling times were calculated between two points in mid-log with the following formula: elapsed time divided by [log (fold increase)/log 2].

Flow cytometric analysis:
Cells were grown to saturation in YPAD medium, collected by centrifugation, washed once, and resuspended in 2% potassium acetate medium at a cell density of 1–2 x 10⁷ cells/ml. Cells were prepared for flow cytometric analysis as described (SAZER and SHERWOOD 1990 ; with modifications: http://flosun.salk.edu/fcm/protocols/ycc.html). At 5-hr intervals, 5 ml samples were collected, washed with dH₂O, resuspended in 70% ethanol and incubated for at least 12 hr at 4°. The cells were washed and resuspended in 50 mM sodium citrate containing 0.1 mg/ml RNase A and incubated for 8 hr at 34°. One volume of 50 mM sodium citrate containing 20 µg/ml propidium iodide was added and incubated overnight at 4°. The cells were sonicated prior to flow cytometry, which was performed on a Becton Dickinson FACScan. At least 8000 cells were analyzed per sample.

ß-galactosidase assays:
Assays were performed as described by HOLLINGSWORTH and PONTE 1997 using the HOP1-LacZ fusion constructed by VERSHON et al. 1992 .

Analysis of the meiotic divisions:
Cells were grown and sporu-lated as described above. Aliquots from duplicate cultures were removed and fixed with 3.7% formaldehyde at the indicated times after transfer to sporulation medium. After 1 hr, the fixed cells were washed and resuspended in SHA (1 M sorbitol, 0.1 M HEPES, 5 mM NaN₃) buffer. The cells were stained with 4',6-diamidino-2-phenyl-indole (DAPI) and analyzed by fluorescence microscopy as described (ROSE et al. 1995 ). A minimum of 600 cells were examined at each time point.

Recombination frequencies:
Yeast strains were grown to saturation in YPAD, and aliquots of each culture were plated onto YPAD medium and synthetic media lacking the appropriate amino acids to determine the mitotic prototroph frequencies. The values are the median frequency ± the standard deviation.

Meiotic recombination frequencies were determined by taking 1.5 ml from each saturated culture, washing once with dH₂O, resuspending the cells in 10 ml of 2% potassium acetate and incubating them at 30° with aeration. Aliquots of the sporulating cultures were removed after 15 hr for return-to-growth experiments or after 2 days (completion of sporulation) for plating onto YPAD and synthetic media lacking the appropriate amino acids. Prototroph formation was assayed in triplicate, and the mean meiotic frequencies ± the standard deviation were determined. The percent sporulation was monitored after 2 days in all cultures to ensure that meiosis had progressed as expected.

Physical recombination assay:
Strains Y366, Y369, Y693, and Y1069 were used to perform the physical recombination assay (BORTS et al. 1986 ). Sporulation, DNA isolation and analysis were performed as described (SYM et al. 1993 ). Cells were allowed to grow for 24 hr in medium lacking leucine after induction of meiosis for the return-to-growth samples. Audio-radiography was quantitated on a Bio-Rad densitometer. The relative levels of the recombinant 16.2-kb band were calculated as follows: The volume of the recombinant band was divided by the volume of the corresponding parental band (P₁). The wild-type number was set at 1.0 and the level of mutant recombinants expressed as a fraction or percentage of wild type.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Isolation of mutants defective in meiosis:
Yeast mutants defective in meiosis were isolated using a variation of a screen described by ROCKMILL and ROEDER 1988 . Meiotic defects in DNA synthesis, chromosome pairing, synapsis, recombination or segregation should result in either failure to complete meiosis or the random segregation of chromosomes resulting in lethality due to aneuploidy. Therefore, yeast mutants defective in meiosis should either fail to sporulate or produce inviable spores. However, mutants that fail to sporulate could also define genes required for entry into meiosis or for packaging the meiotic products into spores. To eliminate these latter classes from consideration, ROCKMILL and ROEDER 1988 only examined mutants that produced inviable spores.

To identify novel genes required for meiosis, we screened for mutants that failed to complete sporulation or produced inviable spores, but were proficient for the induction of meiotic ectopic recombination using a specially marked strain. We reasoned that a subset of mutants identified in this screen would define genes important for aspects of meiosis other than recombination, while at the same time eliminate any candidate mutant that failed to enter meiosis. In addition, we examined the meiotic divisions in strains that failed to sporulate to eliminate mutants that completed meiosis normally yet were unable to package the meiotic products into spores.

The homothallic (HO) strain used for the mutagenesis, JB128, contains a mutant allele of the URA3 gene (ura3-1) at its normal location on chromosome V and a different mutant allele (ura3-Stu) at the HIS4 locus on chromosome III (BHARGAVA et al. 1992 ); ectopic recombination between these sequences results in the production of Ura prototrophs and provides an assay for meiotic recombination ( JINKS-ROBERTSON and PETES 1986 ; LICHTEN et al. 1987 ; GOLDMAN and LICHTEN 1996 ). Induction of meiosis in this strain resulted in an approximately 50-fold increase in Ura prototrophs (7 x 10^-5) over mitotic levels (Table 2).

Spores from JB128 were mutagenized with ultraviolet light to 50% survival. The spores were allowed to germinate, undergo mating-type switching, and mate to form diploid colonies homozygous for any induced mutations. These colonies were sporulated and then exposed to ether vapors. Vegetative cells are more sensitive to ether vapors than spores (DAWES and HARDIE 1974 ); therefore mutants that fail to sporulate or produce inviable spores are ether-sensitive. The sporulation defect was verified by phase-contrast microscopy and/or tetrad dissection. Of approximately 50,000 colonies screened, 108 mutants were recovered.

Many mutants defective in DNA repair produce inviable spores (e.g., the RAD50 epistasis group; GAME 1983 ). These mutants are sensitive to gamma rays and MMS because they are unable to repair the DNA damage induced by these agents. To eliminate these genes from consideration, candidate mutants were tested for sensitivity to MMS; 98 strains that displayed wild-type resistance to MMS were further characterized.

Because the mutants we isolated either failed to sporulate or produced inviable spores, induction of meiotic ectopic recombination was assayed by examining Ura prototrophy in return-to-growth experiments (ESPOSITO and ESPOSITO 1974B ). These experiments are performed by permitting yeast cells to initiate meiosis and then returning them to vegetative growth medium prior to the completion of the first meiotic division; the resulting cells are referred to as meiototic cells (ESPOSITO and ESPOSITO 1974B ). This procedure induces meiotic levels of recombination; however, the mutants are returned to vegetative growth prior to the manifestation of the meiotic-lethal phenotype. Twelve mutants were induced to at least 50% of wild-type levels for meiotic recombination and were selected for further study.

To eliminate mutants that were defective in packaging the meiotic products into spores, we analyzed the meiotic divisions in candidate mutants that failed to sporulate by staining meiotic cultures with DAPI and examining the cells by fluorescence microscopy. Only mutants that displayed a defect in either the number of cells that were able to undergo the meiotic divisions or appeared to arrest prior to the completion of meiosis I or meiosis II were further characterized. Seven recombination-proficient mutants defective in meiosis were recovered. Figure 1 summarizes the strategy and results of the mutant screen.

View larger version (84K): In this window In a new window Download PPT slide   Figure 1. Strategy and summary of mutant screen. The flow chart indicates the number of mutants at each stage of analysis. Meiototic refers to the return-to-growth protocol.

Complementation analysis of recovered mutants:
To determine if the meiotic phenotypes of these strains were due to single mutations, sporulated cultures from the mutants were mated to spores from a wild-type HO strain, BR2171-7B and the heterozygotes selected for on appropriate medium. Although the mutants sporulated poorly and/or produced inviable spores, the ability to mate large numbers of cells allowed for the selection of the rare spores that were produced and viable. Multiple heterozygotes from each cross were sporulated and tetrads dissected. In all cases, the heterozygotes sporulated well, indicating that the mutations were recessive; however, spore viability varied among the different heterozygotes derived from the same mutant. This is most likely due to aneuploidy in the mutant spores that were crossed to the wild-type strain. The individual diploid spore colonies from crosses that exhibited high spore viability were sporulated and tested for ether sensitivity. A minimum of 10 four-spore-viable tetrads were examined from each cross and in every tetrad ether-sensitivity segregated 2:2, indicating that single mutations are responsible for the sporulation defects in these mutants.

To determine whether the mutations in these strains were allelic with previously identified genes, plasmid rescue experiments were performed by transforming the mutant strains with centromere plasmids carrying known meiotic genes that one might expect to find in this screen. We tested the ability of HOP1 (HOLLINGSWORTH and BYERS 1989 ), RED1 (ROCKMILL and ROEDER 1988 ), MEK1 (ROCKMILL and ROEDER 1991 ; LEEM and OGAWA 1992 ), DMC1 (BISHOP et al. 1992 ), ISC10 (KOBAYASHI et al. 1993 ), MSH4 (ROSS-MACDONALD and ROEDER 1995 ), MSH5 (HOLLINGSWORTH et al. 1995 ), ZIP1 (SYM et al. 1993 ), and ZIP2 (P. CHUA and G. S. ROEDER, personal communication) to rescue the meiotic-lethal phenotype of the mutant strains. Strains harboring mutations in all of these genes either produce no or inviable spores and are induced to varying extents for meiotic intragenic recombination. One mutant, Y17, was rescued by RED1 and another mutant, Y207, was rescued by MEK1; linkage studies confirmed that the corresponding mutations are alleles of RED1 and MEK1 (data not shown) and consequently were not further characterized. The wild-type gene responsible for the mutant phenotype of strain KR1-20A was identified; restriction map and complementation analyses demonstrated that it is the SPO14 gene (HONIGBERG et al. 1992 ). Characterization of this allele, spo14-3, has been described (ROSE et al. 1995 ).

Haploid segregants of each of the remaining mutants were generated by crossing sporulated cultures of each mutant to a haploid strain and selecting heterozygotes on the appropriate medium. The resulting diploids were sporulated and tetrads dissected. HO segregants were sporulated directly and subjected to ether tests to determine if they harbored the mutation. Pairwise crosses were performed with the remaining haploid segregants and the resulting diploids were sporulated and subjected to ether tests to determine which haploids carried the mutation. Complementation tests were then performed with haploid mutant segregants; in all cases, the mutations complemented each other, indicating that four complementation groups are represented. We have designated the genes responsible for the mutant phenotypes MUM1, 2, 3, 4 (for MUddled Meiosis). The phenotypes of the mutants isolated in this screen are shown in Table 2. These mutants are proficient for meiototic ectopic recombination and either fail to sporulate or produce spores with a reduced frequency and display varying degrees of spore inviability.

mum1-1 is an allele of GSG1:
MUM1 was cloned from a yeast genomic library in a centromere plasmid (ROSE et al. 1987 ) by complementation of the ether-sensitivity of strain Y 264. Five Ura⁺ transformants, of approximately 5000 screened, produced spores at high frequency. Spontaneous mitotic Ura^- segregants no longer produced spores. The same plasmid (pME556) was recovered from all transformants and shown to be responsible for the rescue of the meiotic defect. Subclone analysis localized the complementing activity to a 3-kb BamHI-XbaI fragment. Strains carrying deletion alleles of these sequences (see below) were mated to a strain containing the original mutant allele, mum1-1; the resultant diploid failed to sporulate, indicating that the sequences represent the MUM1 gene.

Sequence analysis revealed that the MUM1 gene is identical to the chromosome IV ORF, YDR108w, identified in the S. cerevisiae genome project. The 2094-bp open reading frame encodes a 77-kDa protein with no homology to any known proteins. In the course of this study, MUM1 was independently identified and named GSG1 (KAYTOR and LIVINGSTON 1995 ). Consequently, we have designated our original UV-induced allele gsg1-2 and will hereafter refer to MUM1 as GSG1.

KAYTOR and LIVINGSTON 1995 reported that a homozygous GSG1 deletion strain has no noticeable growth defect but fails to sporulate. Furthermore, premeiotic DNA synthesis is delayed and reduced in gsg1 mutants and the sporulation defect is not rescued by the introduction of the spo13 mutation (KAYTOR and LIVINGSTON 1995 ). However, meiotic progression and recombination were not monitored in these gsg1 strains.

We constructed a deletion allele of GSG1 marked with the LYS2 gene in vitro using the cloned gene (pME704; see MATERIALS AND METHODS). Similar to what was observed by KAYTOR and LIVINGSTON 1995 , strains homozygous for gsg1::LYS2 grew with rates similar to an isogenic wild-type strain (doubling times: gsg1, 98 min; wild type, 99 min). While the ability of gsg1 strains to sporulate was greatly impaired (3.2% for gsg1 vs. 65.1% sporulation for the isogenic wild-type strain; Table 2), spores were produced and viable (75% for gsg1 vs. 96% spore viability in the isogenic wild-type strain; Table 2). Consistent with the analysis of our original mutant isolate, the gsg1 deletion strain was induced for meiotic ectopic recombination (Table 2). The difference observed in the ability of gsg1 deletion mutants to sporulate between this and the previous study (KAYTOR and LIVINGSTON 1995 ) most likely reflects strain background differences.

mum2-1 is an allele of SPOT8:
MUM2 was cloned from a yeast genomic library in a centromere plasmid (ROSE et al. 1987 ) by complementation of the sporulation defect of strain Y 279. One Ura⁺ transformant of approximately 5000 screened produced spores at high frequency. Spontaneous mitotic Ura^- segregants no longer produced spores. A single plasmid (pME759) was recovered and shown to be responsible for the rescue of the meiotic defect. Sequences derived from this plasmid were subcloned and targeted for integration into the yeast genome (ROTHSTEIN 1991 ). Spores from a HO strain marked with URA3 at these sequences were mated to a sporulated culture from a HO mum2-1 strain and heterozygotes selected. The resulting diploid was sporulated and tetrads dissected. All 15 four-spore-viable tetrads displayed a parental configuration for Ura prototrophy and ether-sensitivity, indicating that the marked sequences are tightly linked to the mum2-1 mutation.

Subclone and sequence analysis demonstrated that the MUM2 gene corresponds to YBR057c on the right arm of chromosome II (FELDMANN et al. 1994 ). This location places MUM2 in proximity to the spoT8-1 mutation on the genetic map. The spoT8-1 strain was isolated as a sporulation-deficient mutant that fails to undergo premeiotic DNA synthesis (TSUBOI 1983 ). Similar to spoT8-1, analysis of premeiotic DNA synthesis revealed that mum2 mutants failed to replicate their DNA during meiosis (see below). Therefore, we tested whether the MUM2 gene on a CEN plasmid could rescue the sporulation defect of the spoT8-1 temperature sensitive mutant, MTM-964. Approximately 50% of the cells produced spores at the restrictive temperature in the spoT8-1 strain harboring the MUM2 plasmid when induced for meiosis compared with no spores in the same strain harboring the vector alone. Taken together, these data indicate that MUM2 is the gene responsible for the mutant phenotype of the spoT8-1 strain. As spoT does not conform to standard nomenclature we have adopted MUM2 as the locus designation.

The MUM2 gene contains a 1098-bp ORF that encodes a 41-kD protein. The C terminus of the protein is predicted to form a coiled coil based on the Coils Program (LUPAS et al. 1991 ).

MUM2 deletions marked with LEU2 (pME917) and LYS2 (pME1026) were constructed in vitro using the cloned gene. A strain homozygous for the mum2::LEU2 deletion grew at rates similar to an isogenic wild-type strain (doubling times: mum2, 103 min; wild type, 99 min). Unlike the original UV induced allele, mum2 deletion strains did not produce any spores (Table 2), suggesting that mum2-1 is a leaky allele. In addition, while ectopic recombination was stimulated to wild-type levels in the mum2-1 strain, ectopic recombination was reduced approximately threefold relative to wild type in the corresponding mum2 deletion strains (Table 2).

Identification of the MUM3 gene:
MUM3 was cloned from a yeast genomic library in a centromere plasmid (ROSE et al. 1987 ) by complementation of the sporulation defect of strain YKR150. One Ura⁺ transformant of approximately 10,000 screened produced spores at high frequency. Spontaneous mitotic Ura^- segregants no longer produced spores. A single plasmid (pME588) was recovered and shown to be responsible for the rescue of the sporulation defect. Sequences derived from this plasmid were subcloned and targeted for integration into the yeast genome (ROTHSTEIN 1991 ). Spores from a HO strain marked with URA3 at these sequences were mated to spores from a HO mum3-1 strain and heterozygotes selected. The resulting diploid was sporulated and tetrads dissected. Of 17 four-spore-viable tetrads, all displayed a parental configuration for Ura prototrophy and ether-sensitivity, indicating that the marked sequences are tightly linked to the mum3-1 mutation.

Subclone and sequence analysis demonstrated that the MUM3 gene corresponds to YOR298w on chromosome XV. YOR298w does not correspond to any previously identified gene; the sequence does not display similarity to proteins in the Genbank Database.

Disruptions of the MUM3 gene marked with LEU2 (pME765) and LYS2 (pME1138) were constructed. Strains homozygous for the mum3 disruption alleles grew at rates similar to an isogenic wild-type strain (doubling times: mum3, 95 min; wild type, 99 min). When transferred to sporulation medium, mum3 mutants sporulated poorly and produced inviable spores (Table 2). In addition, ectopic recombination was stimulated to the same extent as wild type in strains carrying the different mum3 alleles (Table 2).

Complementation tests with mum4-1:
Several genomic libraries have been screened for sequences that complement the spore inviability of mum4-1 mutants; however, to date, no complementing plasmids have been identified. Therefore, we determined if the spore inviability defect of a mum4-1 mutant could be rescued with plasmids that contain additional yeast genes known to be important for meiosis. Centromere plasmids containing MER1, MER2, REC102, XRS2, MEI4, SPO11, SPO13, and SPO12 failed to rescue the spore inviability of mum4-1, indicating that mum4-1 is not an allele of these genes. Although the phenotype of a mum4-1 mutant suggests it is not an allele of REC103, REC104, REC114 (MALONE et al. 1991 ), MRE2, MRE11 (AJIMURA et al. 1993 ), COM1/SAE2 (PRINZ et al. 1997 ; MCKEE and KLECKNER 1997A ), SAE1, SAE3 (MCKEE and KLECKNER 1997B ), NDT80 (XU et al. 1995 ), NDJ1/TAM1 (CONRAD et al. 1997 ; CHUA and ROEDER 1997 ), or MEI5 (GIROUX et al. 1993 ), we can not presently rule out this possibility.

Premeiotic DNA synthesis:
To determine if gsg1 or the mum mutants undergo premeiotic DNA synthesis, DNA content of yeast cells induced for sporulation was analyzed by flow cytometry. Wild-type yeast cells entered meiosis with a G1 content of DNA, corresponding to the 2N peak (Figure 2). After approximately 10 hr in sporulation medium in this strain background, there was an increase in the 4N peak, corresponding to cells that had undergone premeiotic DNA synthesis. The number of cells with a 4N content of DNA continued to increase until 20 hr after transfer to sporulation medium. As shown in Figure 2, premeiotic DNA synthesis was variably affected in the mutants. In comparison to wild type, only a small percentage of gsg1 cells underwent premeiotic DNA synthesis. Consistent with the analysis of spoT8-1, mum2 deletion mutants failed to replicate their DNA. The number of mum4-1 cells that underwent premeiotic DNA synthesis was slightly reduced, while mum3 mutants replicated their DNA as well as wild type.

View larger version (27K): In this window In a new window Download PPT slide   Figure 2. Premeiotic DNA synthesis in wild-type and homozygous gsg1::LYS2, mum2::LYS2, mum3::LEU2-765, and mum4-1 mutants. Cellular DNA content was assayed by flow cytometry at the indicated times after induction of sporulation. Left and right peaks of each histogram represent cells with 2N and 4N DNA content, respectively.

The failure of the mum2 mutant to undergo premeiotic DNA synthesis raised the possibility that meiosis is not initiated in these strains. To eliminate this possibility, we performed two experiments. First, we analyzed DNA content and induction of ectopic recombination from the same culture; ectopic recombination was induced even though DNA replication did not occur (data not shown). Second, we examined the expression of the HOP1 gene in a mum2 mutant. HOP1 encodes a meiosis-specific component of the SC; HOP1 is not expressed in vegetative cells but is specifically induced in prophase of meiosis I (HOLLINGSWORTH et al. 1990 ; SMITH and ROEDER 1997 ; VERSHON et al. 1992 ). Wild-type and mum2 strains harboring a HOP1-LacZ fusion on a 2-µ plasmid (VERSHON et al. 1992 ) were assayed for ß-galactosidase activity in vegetative cells and upon transfer to sporulation medium. As expected, no activity was detected in vegetative cells; however, ß-galactosidase activity was induced in both wild-type and mum2 cells transferred to sporulation medium (data not shown). Taken together, these results indicate that meiosis is induced in mum2 mutants.

Meiotic progression:
To examine the meiotic divisions in gsg1 and mum mutants, diploids were induced to undergo meiosis and at various time-points, cells were fixed, stained with the DNA-specific dye, DAPI, and examined by fluorescense microscopy. Spore formation was simultaneously monitored by phase contrast microscopy. At 10 hr after transfer to sporulation medium, binucleate cells were observed in wild-type cells, representing the completion of the first meiotic division. The level of binucleated cells peaked at 13 hr and then began to decline. Tri- and tetranucleated cells, representing the completion of the second meiotic division, were observed at 13 hr after transfer to sporulation medium and the levels continued to rise throughout the experiment. At 18 hr, tetrads were observed by phase contrast microscopy (Figure 3). In gsg1 mutants induced for meiosis, there was a reduction in the number of cells that progressed through the first meiotic division, a further reduction in the number of cells able to progress through the second meiotic division and only a small fraction formed mature asci (Figure 3A). In contrast, only uninucleated cells were observed in mum2 mutants, indicating that mum2 cells arrest prior to the meiosis I division (Figure 3B). The timing and levels of the meiotic divisions appeared to occur normally in mum3 mutants although there was a reduction in the number of spores that were packaged (Figure 3C). Finally, mum4-1 mutants were delayed approximately 3 hr relative to wild type for the meiosis I division. The second meiotic division and spore formation appeared to occur with normal kinetics in mum4-1 but at a reduced frequency relative to wild type (Figure 3D).

View larger version (33K): In this window In a new window Download PPT slide   Figure 3. Meiotic progression in homozygous gsg1::LYS2 (A), mum2::LYS2 (B), mum3::LEU2-765 (C) and mum4-1 (D) mutants. The left graph of each panel shows the average % of cells from two independent cultures completing meiosis I, monitored by fluorescence microscopy, in mutant (closed circles) and the isogenic wild-type strain (opened circles) plotted against time (Hours) in sporulation medium. The middle graph shows the sum of the average % of cells from two independent cultures completing meiosis I and meiosis II, monitored by fluorescence microscopy, in mutant (closed circles) and the isogenic wild-type strain (opened circles) plotted against time. The graph on the right shows the average % of asci, monitored by phase microscopy in mutant (closed circles) and wild type (open circles) plotted against time. At least 600 nuclei were counted at each time point.

Epistasis analysis:
To further define the stage at which these genes function, epistasis analysis was undertaken with meiotic mutations that define genes required for recombination and one of the meiotic divisions. spo11 mutants sporulate; however, the spores produced are inviable due to severe defects in recombination (KLAPHOLZ et al. 1985 ). A subset of meiotic mutants fail to sporulate due to a cell cycle checkpoint that monitors recombination (LYDALL et al. 1996 ). These mutants are able to sporulate in the presence of a spo11 mutation since meiotic recombination is not initiated (i.e., dmc1; BISHOP et al. 1992 ). The sporulation defect of gsg1, mum2, or mum3 mutants was not relieved by introduction of a spo11 mutation (Table 3). In addition, the meiotic divisions in the double mutants look similar to the single mutants (Table 3). These results suggest that GSG1, MUM2, and MUM3 function is not dependent on the initiation of meiotic recombination.

spo13 mutants perform only a single division resulting in the production of two diploid spores (KLAPHOLZ and ESPOSITO 1980 ); in our strain background, this single division is predominately equational (ENGEBRECHT and ROEDER 1989 ). Many mutants defective in processes necessary for the proper completion of the first reductional division produce viable spores in a spo13 background (i.e., spo11, rad50, red1, hop1; reviewed in PETES et al. 1991 ). Analysis of gsg1 spo13 and mum3 spo13 strains revealed that these mutants sporulated as inefficiently as gsg1 SPO13 and mum3 SPO13 strains and viability among the rare spores was not improved (Table 3). The introduction of a spo13 mutation did not allow mum2 mutants to divide or sporulate (Table 3), indicating that in the absence of MUM2 neither of the meiotic divisions can occur.

In contrast to the other double mutants, mum4-1 spo13 mutants produced viable spores (Table 3). Analysis of the products of the mum4-1 spo13 meiosis revealed that they primarily undergo equational divisions (see below; Table 6), indicating that MUM4 functions at or before the first meiotic division.

Meiotic recombination:
We examined meiotic recombination using a number of different assays in strains harboring gsg1, mum2, and mum3 alleles. The effect of a gsg1, mum2, or mum3 mutation on allelic intragenic recombination was examined in diploid strains that carry heteroalleles at the HIS4 and LEU2 loci; His and Leu prototrophic recombinants result primarily from gene conversion (FOGEL and HURST 1967 ). Wild-type and mutant cells were transferred to sporulation medium and then returned to vegetative medium in return-to-growth experiments. As shown in Table 4, induction of meiotic prototroph formation in gsg1 mutants was close to the wild-type levels (a two- to threefold decrease is observed) and occurred at wild-type levels in mum3 mutants.

The mitotic frequency of His and Leu prototrophs was significantly lower in the diploid mum2 Y 792 strain compared to the isogenic wild-type strain. Meiotic prototroph formation in mum2 mutants was reduced 18-fold at the HIS4 locus and 406-fold at the LEU2 locus compared to the isogenic wild-type strain (Table 4) at 15 hr after transfer to sporulation medium. The decrease in meiotic recombination in these strains was not simply due to the reduced basal levels, since the fold induction was also reduced [His prototrophs: 25-fold (wt) vs. 6-fold (mum2) over the mitotic frequency; Leu prototrophs: 492-fold (wt) vs. 5-fold (mum2) over the mitotic frequency]. Thus, allelic recombination is perturbed in mum2 mutants.

The effect of a gsg1, mum2, or mum3 mutation on induction of meiotic intrachromosomal recombination was measured in spo13 haploid strains disomic for chromosome III using the assay developed by HOLLINGSWORTH and BYERS 1989 . One of the chromosome III homologs contains a duplication of 11.4 kb of DNA between HIS4 and LEU2; these repeats flank inserted vector DNA including the URA3 and CYH2 genes. Cells that have lost the CYH2 marker as a result of reciprocal recombination between the duplicated elements can be selected on medium containing cycloheximide, due to a recessive mutation conferring cycloheximide-resistance at the CYH2 locus on chromosome VII. As spo13 does not rescue the sporulation defect of these mutants, induction of intrachromosomal recombination was also measured in return-to-growth experiments. In this assay, the gsg1::LYS2, mum2::LYS2, and mum3::LYS2 strains displayed a level of recombination that was indistinguishable from wild type (Table 5).

Reciprocal recombination in the gsg1, mum2, and mum3 mutants was examined by monitoring the physical exchange of DNA molecules using the system developed by BORTS et al. 1986 . In this assay, diploids carrying restriction site polymorphisms generate novel restriction fragments as a result of reciprocal exchange (Figure 4A). Genomic DNA was isolated from wild-type and mutant cells before entry into meiosis and 24 hr after induction of meiosis. DNA samples were then digested and analyzed by Southern blot hybridization using a probe that detects both parental and recombinant fragments. The recombinant 29-kb fragment is not resolved from the 34-kb parental fragment; consequently, the appearance of recombined DNA is monitored by the appearance of the 16.2-kb fragment (BORTS et al. 1986 ; Figure 4). A meiotic recombinant band was present in wild-type and mum3 cells at similar levels (1.0 vs. 0.93 for wild type and mum3, respectively); however, reciprocal recombinants were reduced in gsg1 cells (0.4) and not detectable in mum2 cells at this time in meiosis (Figure 4B, lane 2).

View larger version (46K): In this window In a new window Download PPT slide   Figure 4. Reciprocal recombination in wild-type, gsg1::LYS2, mum2::LYS2, and mum3::LYS2 mutants. (A) Diagramatic representation of parental and recombinant chromosomes in strains used for the physical detection of exchange (BORTS et al. 1986 ). The Bgl II (BglII) sites and predicted sizes of the Bgl II fragments are indicated. MAT (open rectangles), URA3 (black rectangles), leu2 (hatched rectangles), and MATa (grey rectangles) are shown. (B) Autoradiograph of a Southern blot. P1 (34 kb) and P2 (11.2 kb) indicate the position of the parental bands and R (16.2 kb) indicates the recombinant band (the 29 kb recombinant band is not resolved from the parental 34 kb fragment). Lane 1: DNA isolated from vegetative cells; lane 2: DNA isolated from meiotic cells and lane 3: DNA isolated from cells that had been induced to undergo meiosis and then returned to vegetative growth for 24 hrs before DNA isolation.

Meiotic intragenic and intrachromosomal recombination in these mutants had been measured in return-to-growth assays (see above). Therefore, DNA was also isolated from cells that had been induced in meiosis and returned to mitotic growth before DNA isolation. This experiment revealed that physical recombinants were observed at 70% and 83% of the wild-type level upon return to mitotic growth in gsg1 and mum3 strains, respectively (Figure 4B, lane 3). A recombinant band was detected in DNA isolated from mum2 cells that had been allowed to return to growth at approximately 10% of the wild-type level (Figure 4B, lane 3).

Analysis of meiotic recombination in mum4-1 mutants was made possible by the high spore viability exhibited by mum4-1 strains in the presence of the spo13 mutation. Congenic spo13 strains homozygous or heterozygous for mum4-1 were used to assay gene conversion at the TRP1 locus by random spore analysis and crossing over on chromosome III by dyad dissection. Mitotic and meiotic levels of Trp prototrophs were determined and the results presented in Table 6. Similar to ectopic recombination between the URA3 alleles (Table 2), allelic recombination at the TRP1 locus occurred at the wild-type level in mum4-1 mutants.

Reciprocal crossing over on chromosome III occurred at close to wild-type levels in mum4-1 mutants (Table 6). The map distances for wild-type and mum4-1 homozygous diploids were 37 cM and 32 cM for the LEU2-MAT interval and 16 cM and 16 cM for the LEU2-HIS4 interval, respectively. In addition, analysis of the segregation pattern of the centromere linked marker, TRP1, indicated that mum4-1 spo13 strains primarily undergo equational divisions. Given the spo13 rescue of mum4-1, these data suggest that MUM4 acts at or before the first meiotic division but independently of recombination.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In most organisms, recombination is essential for proper chromosome segregation at the meiosis I division (reviewed in HAWLEY 1987 ). In S. cerevisiae, genetic screens and selections have been designed to identify meiotic mutants that, for the most part, perturb meiotic recombination. Consequently, most of these mutants define genes important for meiotic recombination and chromosome synapsis. To identify other processes important for meiosis, we initiated a mutant hunt to identify genes required for meiosis that function independently of recombination. The screen was designed to isolate mutants that either failed to sporulate or produced inviable spores, but were proficient for the induction of meiotic ectopic recombination. Of the seven mutants isolated, five appear to define genes important for processes distinct from genetic recombination (SPO14, HONIGBERG et al. 1992 ; ROSE et al. 1995 ; GSG1, MUM2, MUM3, MUM4, Figure 5A). The other two mutants we identified carry alleles of MEK1 (ROCKMILL and ROEDER 1991 ; LEEM and OGAWA 1992 ) and RED1 (ROCKMILL and ROEDER 1988 ). RED1 encodes a component of the SC (SMITH and ROEDER 1997 ) and MEK1 encodes a protein kinase that displays genetic interactions with RED1 and HOP1 (ROCKMILL and ROEDER 1991 ; HOLLINGSWORTH and PONTE 1997 ), another SC component (HOLLINGSWORTH et al. 1990 ). Strains harboring red1 and mek1 mutations display substantial levels of meiotic recombination; therefore it is not surprising that alleles of these genes were isolated.

View larger version (21K): In this window In a new window Download PPT slide   Figure 5. The effect of mutations of genes isolated in this screen on the major landmarks of meiosis. (A) Summary of mutant phenotypes. The chart shows the phenotypes of gsg1, mum2, mum3, and mum4-1 with respect to premeiotic DNA synthesis, recombination, spo11 and spo13 epistasis, the meiotic divisions, sporulation and spore viability. ND = not determined; intra = intrachromosomal; NA = not applicable. (B) Based on genetic and epistasis analyses, mum2 mutants perturb the early events of chromosome pairing and premeiotic DNA replication. In contrast, gsg1 and mum3 mutants affect multiple processes independent of recombination, while the mum4-1 mutant affects a process at or before the first meiotic division.

Recently, genetic screens to identify mutants whose meiotic arrest is dependent on the initiation of recombination have been designed and genes important for completion of recombination identified (MCKEE and KLECKNER 1997A ; PRINZ et al. 1997 ; MCKEE and KLECKNER 1997B ). Similar to this class of mutants, gsg1, mum2, and mum3 mutants fail to sporulate or sporulate very poorly; however, in contrast to these mutants, their defect in meiotic progression is not relieved by the introduction of the spo11 mutation. Spo11p catalyzes meiosis-specific double strand breaks, the initiators of recombination (KEENEY et al. 1997 ). The failure of the sporulation defect in these mutants to be rescued by the introduction of a spo11 mutation suggests that these mutations define genes important for processes independently of recombination (Figure 5).

Mutants specifically defective in premeiotic DNA synthesis have been previously isolated in yeast (ROTH 1973 ; ESPOSITO and ESPOSITO 1974A ; TSUBOI 1983 ); however, the corresponding gene products have either not been identified or remain uncharacterized. One of these, spoT8-1, is allelic with mum2. Genetic analysis of spoT8-1 demonstrated that these mutants fail to undergo premeiotic DNA synthesis, the meiotic divisions and spore formation (TSUBOI 1983 ).

Strains harboring mum2 null alleles are induced for meiosis yet fail to replicate their DNA and arrest prior to the first meiotic division. Recombination between chromosomes is reduced in both vegetative and meiotic cells, while intrachromosomal recombination occurs at wild-type levels in mum2 deletion strains. Furthermore, in the assays examined, allelic recombination is more perturbed than ectopic recombination although we cannot eliminate the possibility that this is due to allele or strain differences. Interestingly, the observed induction of recombination in mum2 mutants indicates that the initiation of meiotic recombination is not dependent on the completion of premeiotic DNA synthesis (Figure 5). However, failure to undergo premeiotic DNA synthesis does prevent the meiotic divisions from occurring.

mum2 mutants grow at a rate indistinguishable from wild type, suggesting that mitotic DNA synthesis occurs normally. However, the decrease in mitotic recombination between homologous chromosomes in mum2 mutants indicates that MUM2 functions in vegetative cells. The ability to replicate DNA during the mitotic cycle and undergo normal levels of intrachromosomal recombination imply that mum2 mutants are not defective in either DNA replication or the recombination machinery per se and suggest that MUM2 functions in a process that affects both premeiotic DNA synthesis and recombination between chromosomes.

Cytological analysis of yeast chromosomes have revealed that homologous chromosomes are paired before meiosis; chromosome pairing is lost during premeiotic DNA synthesis and reestablished after the completion of DNA replication (WEINER and KLECKNER 1994 ). WEINER and KLECKNER 1994 argue that mitotic and meiotic pairing are mechanistically similar. As mum2 mutants are perturbed for mitotic interchromosomal recombination as well as premeiotic DNA synthesis and meiotic interchromosomal recombination, we favor the hypothesis that MUM2 is important for both chromosome pairing and premeiotic DNA synthesis. However, our genetic data cannot distinguish between a direct or indirect role for MUM2 in either of these processes. Examination of chromosome behavior in mum2 mutants and cytological analysis of Mum2p should help clarify the role of this protein in the early events of chromosome pairing and premeiotic DNA synthesis.

In contrast to mum2, strains harboring mutations in GSG1 and MUM3 do not display an uniform arrest but affect multiple processes required for the proper formation of viable spores. Furthermore, epistasis analysis suggests that the GSG1 and MUM3 gene products function independently of recombination and a single meiotic division.

During the course of these experiments, GSG1 was independently identified as a secondary mutation in a strain carrying a suppressor of rad52 (KAYTOR and LIVINGSTON 1995 ). RAD52 is important for DNA repair and both mitotic and meiotic recombination (GAME 1983 ), consequently, rad52 mutants produce inviable spores. KAYTOR and LIVINGSTON 1995 reported that gsg1 mutants are delayed for premeiotic DNA synthesis and do not sporulate. In addition, the GSG1 gene is expressed in a similar manner to RAD52 in that GSG1 RNA is present in vegetative cells and RNA levels increase approximately twofold during meiosis (KAYTOR and LIVINGSTON 1995 ). As both GSG1 and RAD52 are required for meiosis and display a similar expression pattern, KAYTOR and LIVINGSTON 1995 suggested that there is a functional link between these two genes. However, our epistasis studies and analysis of recombination in gsg1 mutants makes it unlikely that GSG1 plays a direct role in meiotic recombination.

Genetic analysis of the mum4-1 mutant suggests that MUM4 is important for the successful completion of the first meiotic division. mum4-1 mutants are rescued by the introduction of the spo13 mutation and the double mutant undergoes a primarily equational division. Until the corresponding gene has been identified, we can not rule out the possibility that mum4-1 is an allele of a previously characterized gene. However, the phenotype of this mutant allele is distinct from mutations in other genes that function at or before the first meiotic division. Most notably, at all loci examined, mum4-1 mutants appear to undergo wild-type levels of meiotic recombination. Although we cannot exclude the possibility that the mum4-1 mutation is a leaky allele and that strains that carry a deletion of the gene will display effects on recombination, it seems likely that MUM4 may mediate a process distinct from recombination that is required for proper chromosome segregation at the meiosis I division.

Given the spectrum of the phenotypes of the mutants isolated, it is not surprising that this screen was not saturated. We did not isolate multiple alleles of any of the genes we identified nor alleles of genes that we predict should have been isolated. For instance, zip1 (SYM et al. 1993 ), zip2 (CHUA and ROEDER, personal communication), msh4 (ROSS-MACDONALD and ROEDER 1995 ), and msh5 (HOLLINGSWORTH et al. 1995 ) mutants display close to wild-type levels of intragenic recombination and fail to sporulate (zip1 and zip2) or reduce the number of viable spores (msh4 and msh5) but were not isolated in this screen.

While this screen did not isolate mutations in a specific aspect of meiosis, analysis of the mutants identified suggests that early events in meiotic chromosome metabolism, independent of the initiation of meiotic recombination, play key roles in meiosis. Furthermore, processes are occurring parallel to the initiation of meiotic recombination and the reductional division that are important for the production of viable meiotic products. Molecular and cytological analyses of the genes and corresponding gene products identified in this screen should help define the roles they play in meiosis.

	ACKNOWLEDGMENTS

We are indebted to J. POLYAKOVA and T. CALHOUN for their contribution to this project in its initial stages. We thank N. HOLLINGSWORTH, S. RUDGE, J. TRIMMER, and S. STRICKLAND for helpful discussions and comments on the manuscript. J.E. was a recipient of American Cancer Society Junior Faculty Research Award during this study. K.R. was supported by National Institutes of Health Training Grant 5T32GM07518 and by a Hoffmann-LaRoche Scholarship from the Institute for Cell and Developmental Biology. This work was supported by National Institutes of Health Grant GM4863903 to J.E.

Manuscript received August 12, 1997; Accepted for publication October 10, 1997.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

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

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