A RAD 9 - Dependent Checkpoint Blocks Meiosis of cdcl 3 Yeast Cells

Mutations in CDC13 have previously been found to cause cell cycle arrest of Saccharomyces cerevisiae at a stage in G2 immediately preceding the mitotic division. We show here that cdcl3 blocks the meiotic pathway at a stage that follows DNA replication, but in this case the spindle has not yet formed nor have the chromosomes undergone synapsis or recombination. This arrest is alleviated by rad9, thus implicating the same checkpoint function that delays mitotic progression when chromosomal lesions are present. An assessment of the spores produced upon alleviation of the meiotic arrest by rad9 reveals that the absence of recombination in strains bearing cdcl3 alone is attributable to the RAD9-mediated arrest rather than to other effects of cdcl3 lesions. We have tested the possibility that this checkpoint function is important in regulating meiotic progression to permit resolution of recombinational intermediates during ongoing meiosis and have found no evidence that rad9 alters the execution of functions that might depend upon such regulation. We consider the possible role of other checkpoints in yeast meiosis. I T has recently been suggested that the consistent states of arrest displayed by various temperaturesensitive mutants in the yeast cell division cycle reflect checkpoint functions which monitor the efficacy of preceding steps and, if warranted, delay progression to provide time for correction of the relevant defect (HARTWELL and WEINERT 1989). A notable example is the induction of mitotic arrest by mutation in the DNA ligase encoded by CDC9. Deficiency for the enzyme results in the accumulation of chromosomal lesions that trigger a checkpoint in G2 under the control of RAD9 (HARTWELL and WEINERT 1989). Cells doubly mutant for cdc9 and rad9 fail to undergo cell cycle arrest but rapidly become inviable, probably because of their failure to pause for repair of the cdc9dependent chromosomal lesions. A similar pattern of RADPdependent mitotic arrest is caused by temperature-sensitivity for cdcl3 (WEINERT and HARTWELL 1988), which is also known to cause elevated mitotic recombination in a manner indicative of accumulated chromosomal lesions (HARTWELL and SMITH 1985). In the present work, we have explored the roles of the functions encoded by CDC13 and RAD9 in meiosis. Genetic dissection of the meiotic pathway has previously shown that many cdc functions, which initially were defined by their essential roles in mitosis, are also required during meiosis. A comprehensive survey by SIMCHEN (1974) demonstrated that meiotic cells require every CDC locus he tested except those that mediate processes absent from sporulation, such as budding and cytokinesis. Subsequent phenotypic analysis has shown, however, that many of these mutations lead to arrests of the meiotic pathway in states Genetics 131: 55-63 (May, 1992) that are cytologically distinct from those in mitosis. For example, cdc28 is required for START in mitosis, causing arrest before bud emergence or entry into Sphase (HARTWELL 1973; HARTWELL et al. 1973). In meiosis, the first cdc28 arrest point falls during pachytene, following DNA replication, synaptonemal complex formation and commitment to meiotic levels of recombination (SHUSTER and BYERS 1989). A mutation causing mitotic arrest at a slightly later stage is cdc7, permitting bud emergence and spindle formation but failing in the initiation of DNA synthesis (CULOTTI and HARTWELL 197 1; HARTWELL 1973; BYERS and GOETSCH 1974). In meiosis, cdc7 is permissive for DNA synthesis but causes a reversible arrest prior to synapsis and spindle formation (SCHILD and BYERS 1978). Mutants defective in DNA replication per se during mitotic growth similarly block spindle elongation in mitosis while preventing both synapsis and spindle formation in meiosis (SCHILD and BYERS 1978). These and other distinctions between the genetic requirements for mitosis and for meiosis may be expected to reveal differences between fundamental regulatory mechanisms that coordinate analogous events in either mode of division. We have therefore sought to establish the meiotic role of the RAD9-mediated checkpoint as revealed by the behavior of sporulating cells mutant for cdcl3. A priori, two patterns of behavior seemed possible. The first was that RAD9 might play a key role in meiosis because the special patterns of DNA replication, meiotic gene conversion, and reciprocal recombination may involve extensive disruptions of chromosomal DNA (STERN and HOTTA 56 L. Weber and B. Byers 1973, 1976; RESNICK et al. 1984). A checkpoint might then be required to ensure the restoration of normal DNA structure before progression into division. We note in this regard that other functions that were originally identified by their roles in radiation repair have already been shown to play crucial roles in meiosis [RAD50 (GAME and MORTIMER 1974), RAD52 (GAME and MORTIMER 1974), and RAD57 (GAME et al. 1980)], so one might expect a meiotic role for RAD9 as well. Alternatively, one might imagine that these and various meiosis-specific recombinational functions, such as HOP1 (HOLLINGSWORTH and BYERS 1989; HOLLINGSWORTH, GOETSCH and BYERS 1990), MER1 (ENGEBRECHT and ROEDER 1989,1990), MER2 (ENGEBRECHT, HIRSCH and ROEDER 1990), RED1 (ROCKMILL and ROEDER 1988, 1990; THOMPSON and ROEDER 1989), SPOl l (KLAPHOLZ, WADDELL and EsPOSITO 1985), andMEl4 (MENEES and ROEDER 1989), a re so prevalent in meiosis that a checkpoint for repair of any lesions that may result from exogenous agents would be dispensable. T h e present work reveals not only that CDCl3 plays an essential role in meiosis but also that the arrest attendant upon temperature-sensitivity for cdcl3 is alleviated by rad9. Unlike the case in mitosis, however, the RAD9-mediated meiotic checkpoint does not cause arrest at the stage of spindle elongation but blocks progression at an earlier stage, before the spindle has formed or the chromosomes have undergone synapsis. MATERIALS AND METHODS Strains: The genotypes of the strains used in these experiments are described in Table 1. Standard procedures for strain construction and genetic analysis were employed (MORTIMER and HAWTHORNE 1969). LW3201, LW3202, LW3203 and LW3204 are isogenic strains generated by crossing spore clones from a diploid of 7845-8-4 and 39854-lb, both of which had been backcrossed into the A364a background five or more times. LW3205 is a haploid strain from this same parent diploid. LW3000 is a diploid of 428 and 131-20. LW3001, LW3002 and LW3003 were independently generated from strain LW3000 by brief X-irradiation followed by screening for homozygosis of cdcl3-I and maintenance of other markers as in SCHILD and BYERS (1978). The genotypes of these latter three strains are therefore identical to LW3000 with the exception of the homozygosed chromosomal segment that includes cdc l3 I . Strains LW3302 through LW3306 were generated by crossing spore clones from a diploid of 7845-8-4 and 131-20 and thus retain the hybrid background of A364a and 131-20. Strain LW390 1 was generated by crossing spore clones from a diploid of LW3205 and 415-1829 to 131-20, thus also retaining the hybrid background of A364a and 131-20. Phenotypes attributed to c d c l 3 and rad9 were confirmed by their alleviation in strains transformed with the corresponding wild-type genes. Genetic procedures: Materials and methods used in this study have been described, including liquid and solid media (HARTWELL 1967; WOOD 1982), sporulation procedures (SCHILD and BYERS 1978), induction of mitotic recombination by X-irradiation (SCHILD and BYERS 1978), assays of viability and commitment to recombination (SHUSTER and BYERS 1989), and the assay for haploidization (SIMCHEN 1974; HOLLINCSWORTH and BYERS 1989). Plasmids YEP24CDC13+ (L. HARTWELL, unpublished) and YCp5O-RAD9+ (WEINERT and HARTWELL 1990) were provided by L. HARTCytological procedures: Electron microscopy preparation and procedures were as described in BYERS and GOETSCH (1 99 1). Serial sections were examined with a Philips EM300 electron microscope. Fluorescence microscopy procedures were as described in ADAMS and PRINCLE (1 984) and BAUM et al . (1988). DNA was stained with 4,6-diamidino-2-phenylindole (DAPI) (1 .O rg/ml, Sigma). Stained cells were examined with a Nikon Microphot microscope equipped with epifluorescence optics and filters. Flow cytometry analysis was as described in HUTTER and EIPEL (1979) using propidium iodide to stain DNA (Sigma). A Becton Dickinson FACSCAN flow cytometer and the software packages CELLFIT and LYSYS were used to collect and analyze these data. The RFIT algorithm was used to estimate the percentage of the population in GO + G1, S, and G2 + M phases for each time point. The cumulative percentage of cells that had entered into S phase, graphed as “premeiotic S phase” (see Figure 2), corresponds to the sum of the S and G2 + M values. WELL.