A homolog of male sex-determining factor SRY cooperates with a transposon-derived CENP-B protein to control sex-specific directed recombination

Swi5 Affects Directionality of Mating-Type Switching.

Swi5 affects mating-type switching (28, 29), but its role in donor choice and its relationship to other factors involved in this process have not been studied. We determined the effects of single and double mutants on mating-type switching in a homothallic (h⁹⁰) strain using an iodine-staining assay (Fig. 2 A and B). Cells that switch efficiently to the opposite mating type form colonies with a homogenous distribution of M and P cells and can thereby readily mate and sporulate, resulting in a dark iodine staining. In contrast, cells that switch their mating type with a lower efficiency form lightly stained colonies due to poor mating and sporulation within the colony. As expected, a swi5Δ mutant produced lightly stained colonies with a sectored/mottled staining pattern (Fig. 2A) (28, 29). Consistent with Swi2 and Swi5 acting together as components of the RPC, swi2 swi5 double mutants showed mating-type switching defects similar to those displayed by the individual single mutants (Fig. 2B). To confirm that reduced iodine staining in swi5Δ was due to defective utilization of one or both donor loci, we analyzed mat1 genetic content by multiplex PCR using mating-type allele-specific (i.e., mat1-P or mat1-M) primers. Wild-type strains efficiently use both donors, as indicated by the almost equal intensity bands representing mat1-P and mat1-M (Fig. 2 A and B). However, swi5Δ or swi2Δ swi5Δ cells were predominantly of the P mating type (Fig. 2B), implying that mat2-P is the preferred donor in these mutants.

Loss of Swi6 compromises the spreading of Swi2 from the SRE to the surrounding sequences. As a result, swi6 mutants predominantly use mat3-M as a donor, resulting in reduced iodine staining (5, 18, 28) (Fig. 2A). However, cells containing a swi2 swi6 double-mutant combination show elevated heterologous switching and iodine staining (18). This is believed to be caused by the random utilization of donors in cells lacking both RPC and heterochromatin structure at the mat locus. swi5 swi6 double mutants showed a significant increase in switching to the opposite mating type, compared with swi5 or swi6 single mutants (Fig. 2A). The effects of swi5Δ on mating-type switching were confirmed in h⁰⁹ background cells, in which the contents of the donor cassettes have been swapped (5) (Fig. 2C). Because cells select a donor on the basis of the location of the silent cassette, switching in h⁰⁹ cells culminates in the replacement of mat1 information with the same allele (5). Deletion of swi5 alone or in combination with the swi6 mutant had a predicted effect on donor choice, consistent with the results from experiments using h⁹⁰ cells (Fig. 2C). Thus, Swi5 is a critical factor involved in the nonrandom utilization of donor mating-type loci, irrespective of the mating-type configuration.

Mc Localizes to M-Specific Genes, Retrotransposon LTRs, and Noncoding RNA Loci.

Because the spreading of Swi2 and Swi5 occurs specifically in mat1-M cells, addressing the mechanism by which cell type regulates the spread of the RPC is critical for understanding the directionality of switching. We therefore studied the roles of mating-type genes, in particular mat1-M, in the process of mating-type switching. Of the two genes in the mat1-M cassette, Mi is induced specifically during meiotic induction (10). However, Mc is expressed during vegetative growth (10) and might have a role in the directionality of switching. Like its SRY homolog (26), Mc belongs to the family of HMG-box proteins (24, 30) that modulate expression of target loci (31). As a first step toward addressing the possible functions of Mc in mating-type switching, we constructed a strain expressing a myc epitope-tagged Mc (Mc-myc). The tagged Mc is expressed under the control of native regulatory elements. The stable mat1-M (mat1Msmt0) (32) cells expressing Mc-myc mated efficiently with opposite mating-type (mat1-P) cells (Fig. S1), suggesting that the tagged protein is functional.

We used the strain expressing Mc-myc to perform genome-wide localization of Mc. Chromatin immunoprecipitation coupled with DNA microarray (ChIP-chip) analyses showed that Mc binds to specific sites on all three chromosomes, although no major peaks mapped to the mat locus (Fig. 3A). Mc binding occurred preferentially in intergenic regions, in particular in between divergently transcribed loci. Mc binding was detected upstream of several genes that are expressed in an M cell-type–specific manner (Fig. 3 and Fig. S2). These included genes such as mfm2 and mam1 that are involved in the production and secretion of M factor, respectively (33, 34), and mam3, which encodes the M cell-type–specific agglutinin that allows specific aggregation between cells of opposite mating type (35). Interestingly, Mc was found to be enriched at the promoter region of scm3 (Fig. 3B), which encodes a chaperone involved in loading of the centromere-specific histone H3 variant, CENP-A (36–38). Mc likely induces the expression of this locus because a previous study has shown that scm3 is expressed at higher levels in M cells, compared with P cells (35). Surprisingly, we also found Mc binding upstream of the map2 and map4 genes (Fig. S2) that express specifically in P cells (35, 39), raising the possibility that Mc regulates expression of these loci.

These analyses also detected prominent Mc peaks near several loci that show no obvious induction in M cells (Fig. 3 and Fig. S2). In some of these cases, Mc localization corresponds to sites of noncoding RNA (ncRNA) production or antisense transcripts (Fig. 3A). Several Mc peaks correspond to WTF (With Tf) elements on chromosome 3 and solo LTRs dispersed across the genome (Fig. 3). Mc localization was detected only at a subset of LTRs (Fig. S3), suggesting that specific features of these LTRs or their chromosomal context influence Mc targeting to these loci.

mat1-Mc Is Highly Enriched at swi2 and swi5 Promoter Regions.

Remarkably, we found that Mc is highly enriched at the promoter regions of swi2 and swi5 (Fig. 3 A and B). This is significant because differential expression of Swi2 and Swi5 in M and P cells may help explain differences in the spreading of the RPC at the mat locus. To confirm this result, we performed conventional ChIP assays to detect Mc binding at these loci. Mc could be readily detected at both swi2 and swi5 loci (Fig. S4). Results of ChIP and ChIP-chip further demonstrated that Mc levels are considerably higher at swi2 than at swi5 (Fig. 3 and Fig. S4). These findings suggest a model in which mat1-Mc modulates levels of Swi2 and Swi5 proteins to promote differential spreading of the RPC observed at the mat2/3 interval of M and P cells.

Swi2 and Swi5 Are Differentially Expressed in M and P Cells.

We next investigated whether Swi2 and Swi5 are differentially expressed in M and P mating-type cells. For this purpose we constructed stable M (mat1Msmt0) (32) or stable P (mat1PΔ17) (40) strains expressing myc-tagged Swi2 or Swi5. Western blot analyses with anti-myc antibody revealed that Swi2 is expressed at higher levels in M cells, compared with P cells (Fig. 4A). In view of a previous report of comparable protein levels (18), the differential expression of Swi2 in M and P cells was further confirmed by using independent strains. Further analyses showed that Swi5 levels were also slightly higher in M cells than in P cells, although the difference is smaller compared with Swi2 (Fig. 4B). We next investigated whether Mc is responsible for the differential expression of the RPC components, in particular the pronounced difference in Swi2 levels between M and P cells. Mutation in mat1-Mc caused reduction in Swi2 and Swi5 levels (Fig. 4 A and B), although the effect on Swi5 was less robust. This result is consistent with the relatively lower levels of Mc detected at the swi5 locus, compared with the swi2 locus (Fig. 3B and Fig. S4). Taken together with results presented above, our analyses demonstrate that Mc regulates the levels of RPC components. Moreover, the failure of the RPC to spread from the high-affinity SRE recruitment site near mat3 to mat2 is most likely attributable to lower Swi2 and Swi5 levels in P cells, compared with higher levels of these proteins in M cells.

mat1-Mc Affects the Spreading of the RPC Across the Silent Mating-Type Region.

We next sought to determine if mat1-Mc affects spreading of the RPC. For this purpose, we compared the distribution of the RPC across the silent mat region of strains that carry either the stable mat1-M wild-type allele or the mat1-Mc^– mutant allele (mat1Msmt0). ChIP was performed to map the binding of Swi2-myc across the entire mat2/3 interval, as described previously (11, 18). As expected (18), Swi2 was found throughout the 20-kb silent mating-type domain in wild-type mat1-M cells, but its localization was attenuated and restricted largely to the region containing the SRE in mat1-Mc^– mutant cells (Fig. 5). Defects in Swi2 expression and RPC spreading observed in mat1-Mc^– mutant cells could be reversed by the ectopic expression of Mc from an inducible promoter (Fig. S5). Together, these results suggest that Mc, involved in proper expression of RPC components, affects the spreading of the RPC at the mat2/3 region. As a control, we also tested the effects of mutation in mat1-Pc. mat1-Pc^– mutant cells showed Swi2 distribution patterns similar to those in wild-type mat1-P cells (Fig. S6), with restricted RPC localization at the SRE (18). Therefore, the differential spreading of the RPC in P and M cells is not due to mat1-Pc inhibiting the spreading of this complex. Rather, mat1-Mc appears to actively promote RPC spreading by modulating levels of Swi2 and Swi5 proteins.

CENP-B Homolog Abp1 Facilitates Mc Localization at swi2 and LTRs.

Abp1 regulates donor choice and is required for RPC spreading across the silent mat region (29). Because our previous work has shown that Abp1 localizes to gene promoters and LTRs (41), we wondered if loss of Abp1 impacts Mc binding to these loci. ChIP-chip analyses were performed to compare Mc distribution in wild-type and abp1Δ cells. Remarkably, abp1Δ abolished Mc localization at LTRs and swi2 (Fig. 6A). Abp1 is localized at the swi2 promoter (Fig. 6A), in addition to its localization at LTRs (41), and hence likely directly facilitates Mc localization at these loci. Conventional ChIP confirmed Abp1 involvement in Mc localization at swi2 (Fig. 6B). Loss of Abp1 had no major impact on Mc distributions at swi5 (Fig. S7). Similarly, Mc binding at the map2 locus was not adversely affected, but abp1Δ showed considerable reduction of Mc enrichment at the promoter regions of scm3, mam3, and mfm2 (Fig. S7). We next asked if Abp1 affects expression of Swi2 protein. Western analysis was performed to assay Swi2 levels in stable M (mat1Msmt0) cells carrying either wild-type or abp1Δ mutant alleles. abp1Δ resulted in a severe reduction in Swi2 levels (Fig. 6C). Together with previous data (29), these results suggest that Abp1 affects spreading of the RPC and directionality of mating-type switching at least in part by affecting Mc-mediated increase in Swi2 levels in M cells.

Yeast Strain.

Strains containing swi5Δ, Swi5-myc, and Mc-myc alleles were constructed by using a PCR-based module method (47). The construction of mat1-Mc^– or mat1-Pc^– mutant strains is shown in Fig. S8. swi2Δ, Swi2-myc, and abp1Δ have been described previously (18, 41). Mutant swi6 strains used in this study contain the missense swi6-115 (W269R) allele that causes severe reduction in Swi6 protein levels (48).

Mating-Type Switching Assays.

Iodine-staining procedure was used to estimate the efficiency of switching as described previously (9, 49). Switching efficiency was also examined by a quantitative PCR assay. In this assay, DNA prepared from wild-type and mutant cells is used to perform multiplex PCR to determine the genetic content at the mat1 locus, as described previously (18). PCR fragments corresponding to mat1-P and mat1-M alleles were resolved on polyacrylamide gels and quantified using Phosphoimager.

ChIP and ChIP-chip.

ChIP and ChIP-chip using myc antibody (A-14; Santa Cruz) or affinity-purified rabbit Mc antibody raised against full-length recombinant protein (this study) were performed as described previously (13). Enrichment values were corrected for nonspecific signals from control samples with the same antibodies used to perform ChIP. ChIP analysis was carried out essentially as previously described (11, 18). Microarray data are available at the National Center for Biotechnology Information‘s Gene Expression Omnibus repository under accession number GSE32996.