SUMl - 1 : A Suppressor of Silencing Defects in Saccharomyces cerevisiae

The repression of transcription of the silent mating-type locus HMRa in the yeast Saccharomyces cerevisiae requires the four SIR proteins, histone H4 and a flanking site designated HMR-E. The SUMl-1 mutation alleviated the need for many of these components in transcriptional repression. In the absence of each of the SIR proteins, SUMl-1 restored repression in MATa strains; thus, SUMl-1 appeared to bypass the need for the SIR genes in repression of HMRa. Repression was not specific to the genes normally present at HMR, since the T R P l gene placed at H M R was repressed by SUMl-1 in a sir3 strain. Therefore, like the mechanism of silencing normally used at HMR, silencing by SUMl1 was gene-nonspecific. SUMl-1 suppressed point mutations in histone H4, but failed to suppress strongly a deletion mutation in histone H4. Similarly, SUMl-1 suppressed mutations in the three known elements of HMR-E, but was unable to suppress a deletion of HMR-E. These epistasis analyses implied that the functions required for repression at HMR can be ordered, with the SIR genes and silencer elements acting upstream of SUMl-1. SUMl-1 itself may function at the level of chromatin in the assembly of inactive DNA at the silent mating-type loci. T RANSCRIPTIONAL repression plays an essential role in the control of gene expression. The well-studied cases of transcription repression range from those specific to a particular gene to those that exert a regional effect. Specific repression involves the selective inactivation of particular promoters and often involves sequence-specific binding proteins; SV40 T antigen, for example, binds to the SV40 early promoter and blocks transcription (TJIAN 1981). Regional repression results in the inactivation of a domain of DNA. A striking example of regional repression is X chromosome inactivation in female utherian mammals, in which many of the genes on one X chromosome become transcriptionally quiescent (reviewed in GARTLER and RICCS 1983). Although changes in chromatin structure, methylation, and timing of replication are correlated with regional repression, the underlying mechanism(s) responsible for the establishment and maintenance of any repressed region remain elusive. A second example of regional transcriptional repression occurs at the silent mating-type loci in the yeast Saccharomyces cerevisiae. The mating-type genes in this yeast are located at three loci on chromosome ZZZ. Two of the copies of the genes, specifically those at the HML and HMR loci, are transcriptionally silent. The third copy, at the MAT locus, is transcriptionally active and determines the mating-type of the cell. MATa cells exhibit the characteristics of the a cell type, including production of the a-factor mating pheromone and the ability to mate with a cells. Similarly, MATa cells produce the a-factor mating pheroGenetics 129: 685-696 (November, 1991) mone and are able to mate with a cells. Cells of the a and a cell type mate to form the ala diploid, which is unable to mate and sporulates under appropriate nutritional conditions. In most laboratory strains, a transcriptionally inactive copy of a information is present at the HML locus (HMLa), and a transcriptionally inactive copy of a information is found at HMR (HMRa). If HMLa and HMRa are improperly expressed in a haploid cell, the cell assumes the nonmating phenotype of an ala diploid (reviewed in HERThe position effect leading to repression of HML and HMR requires several proteins and four flanking DNA sequences. The SIR proteins are necessary for the repression of the silent mating-type loci since mutations in any of the four SIR genes result in derepression of HML and HMR [RINE and HERSKOWITZ (1 987) and references therein]. Thus, in a formal genetic sense, the SIR proteins act as repressors of HML and HMR. Histone H4 is also required for the repression of HML and HMR. Deletions of the highlyconserved N-terminal domain of histone H4 result in a loss of repression at both loci, and point mutations in the N-terminal domain result in an HML-specific loss of repression (KAYNE et al. 1988; JOHNSON et al. 1990; MECEE et al. 1990; PARK and SZOSTAK, 1990). The phenotypes associated with mutations in the two genes encoding histone H4 (HHFI and HHF2) are recessive. Thus, the mechanism of repression is likely to require, at least in part, the assembly of a particular chromatin structure at HML and HMR. The ARDl and NATl (AAAI) genes, which appear to encode SKOWITZ 1988, 1989). 686 P. Laurenson and J. Rine components of an N-terminal protein acetyl transferase, are essential for the repression of HML, and less important for the repression of HMR (WHITEWAY et a/. 1987; LEE, LIN and SMITH 1988; MULLEN et al. 1989; LEE, LIN and SMITH 1989). The physiologically relevant substrate for this enzyme is unknown. Sites flanking HML and HMR, designated E and I, are necessary in cis for transcriptional repression (ABRAHAM et al. 1984; FELDMAN, HICKS and BROACH 1984). The HMR-E, HML-E and HML-I sites are known as “silencers” because of the relative positionand orientation-independence of their abilities to repress nearby genes (BRAND et al. 1985; MAHONEY and BROACH 1989). The available evidence indicates that the transcriptional repression of the silent mating-type loci reflects a SIR-dependent inactivation of that region of DNA. When placed at HML or HMR, other genes, such as URA3 and TRPI, are transcriptionally repressed in a SIR-dependent manner (BRAND et al. 1985; MAHONEY and BROACH 1989). Thus, repression is promoternonspecific. Repression is also RNA polymerase-nonspecific because tRNA genes, transcribed by RNA polymerase 111, are transcriptionally repressed when present at HMR and HML (SCHNELL and RINE 1986; MAHONEY and BROACH 1989). The regional inactivation of the silent mating-type loci also affects the activity of the site-specific endonuclease HO. Whereas the HO recognition sequence at MAT is cleaved by HO in vivo, the same sequence at HML and HMR is resistant to cleavage in wild-type strains (STRATHERN et al. 1982). However, cleavage of the HO recognition sequence occurs at HML in a sir2 strain (KLAR, STRATHERN and ABRAHAM 1984), demonstrating that resistance to HO cleavage is SIR-dependent. Therefore, HML and HMR appear to be in a state that renders them inaccessible to a variety of proteins that interact with DNA. Further evidence suggests that HML and HMR are silenced because they form an inactive chromatin structure (see DISCUSSION). The SUM1 gene product may participate in the inactivation of the silent mating-type loci. The SUMII allele was isolated as a suppressor of sir2 (marl-I; KLAR et al. 1985). SUMI-1 suppresses null alleles of SIR2 and SIR3 (KLAR et al. 1985); thus, SUMI-1 is unique among suppressors of SIR mutations (SCHNELL et al. 1989; LIN et al. 1990) in its ability to suppress mutations in more than one gene. SUMI-I restores mating to sir2 strains by restoring the transcriptional repression of HML and HMR (LIVI, HICKS and KLAR 1990). In this paper, we describe the further characterization of the SUMI-I mutation. Our results suggest that the role of SUMI-I is intimately involved with that of histone H4, that SUMI-I is dependent on the HMR-E silencer for repression of HMRa, and that SUMI-I is epistatic to several other mutations affecting silencing. MATERIALS AND METHODS Yeast strains, media and genetic methods: The genotypes of the yeast strains described in this paper are listed in Table 1. All yeast genetic manipulations were performed as described (ROSE, WINSTON and HIETER 1989). Casamino acid plates were prepared as previously described (AXELROD and RINE 1991), as was minimal sporulation medium supplemented with 25 rg/ml zinc acetate (PILLUS and RINE 1989) and all other media used (ROSE, WINSTON and HIETER 1989). Mating-type tests of patches of cells were performed as described (RINE and HERSKOWITZ 1987), using the tester strains DBY 1034 (a) and JRY2516 (a) unless otherwise indicated. Yeast transformations were performed using the lithium acetate (ITO et al. 1983) or polyethylene-glycolinduced methods (KLEBE et al. 1983). The mating proficiency of plasmid-bearing strains was determined for at least three transformants for each strain. Transformants were spread as patches on casamino acid plates, which selected for the maintenance of URA3-containing plasmids. After 2 days of growth, the mating-type of plasmid-bearing cells was determined using a ura3 mating-type tester strain (TD4), so that only plasmid-bearing diploids were able to grow. Strain constructions: The strains described in Table 2 were isogenic to a MATa SUMI-I strain URY2456) derived from a cross between a MATa SUMI-I strain (K697; provided by A. KLAR) and YM256. SIR mutant strains isogenic to JRY2456 were constructed by a combination of gene disruptions, MAT interconversions and genetic crosses. SIR gene disruptions were performed by one-step gene replacement (ROTHSTEIN 1983) using a sirZ::HIS3 allele (pJR531; KIMMERLY and RINE 1987), a sir3A::LEUZ allele (D330; SHORE, SQUIRE and NASMYTH 1984), and a sir4M::HIS3 allele (pRS229; SCHNELL 1987). Gene disruptions were confirmed by DNA blot hybridization (MANIATIS, FRITSCH and SAMBROOK 1982) and by outcrossing the strains to a wildtype strain and evaluating the pattern of segregation. The MATa SUMI-I strain JRY2456 was converted to MATa by transforming the strain with a plasmid containing the HO gene under control of the GAL10 promoter (pGAL-HO; HERSKOWITZ and JENSEN 1991), growing transformants overnight in liquid galactose-containing medium, and plating for single colonies on solid YPD medium. Colonies with the MATa genotype were detected by a mating-type test, and the genotype was confirmed by subsequent crosses. T o test the ability of SUMI-I to suppress point mutations in the N-terminal domain of histone H4, MATa was disrupted in the strains LJY4051, LJY412I and LJY421I using a fragment from the matk:LEUZ disruption plasmid GRmatAa1::LEUZ (a gift from J. STRATHERN). The strains, which previously mated weakly as a’s, became weak bimaters. Bimaters mate to both a and a mating-type tester strains, and these bimaters mated poorly to either tester strain. The matA::LEUZ strains were transformed with a URA3 plasmid containing MATa (pJR157) and mated to a matal SUMI-I strain URY3245) which was itself transformed with a plasmid expressing a1 and URA3 (pJR760). The plasmids allowed the diploids to sporulate, and segregants were streaked to 5-fluoro-orotic acid-containing medium to select for cells that had lost both plasmids (BOEKE, LACROUTE and FINK 1984). The mating genotype of the segregants was subsequently scored. Determination of the SUMI genotype: In contrast to a previous report (KLAR et al. 1985), in the experiments Suppression of Silencing Defects 687