New SNF Genes , GAL 11 and GRRl Affect SUC 2 Expression in Saccharomyces cerevisiae Laura

To identify new genes required for derepression of the SUCZ (invertase) gene in Saccharomyces cereuisiae, we have isolated mutants with defects in raffinose utilization. In addition to mutations in SUCP and previously identified SNF genes, we recovered recessive mutations that define four new complementation groups, designated snj7 through snfl0. These mutations cause defects in the derepression of SUCP in response to glucose limitation. We also recovered five alleles of gall I and showed that a gall I null mutation decreases SUCZ expression to 30% of the wild-type level. Finally, one of the mutants carries a grrl allele that converts SUCZ from a glucose-repressible gene to a glucose-inducible gene. M ANY genes in Saccharomyces cerevisiae are regulated in response to glucose availability. Yeast cells prefer to use glucose as a carbon source and when grown in glucose, repress expression of the genes involved in utilizing alternate carbon sources. T o understand the regulatory mechanism(s) responsible for glucose repression, we have focused on understanding the control of SUCP (invertase) gene expression. The SUCB gene provides a convenient model system because it is regulated solely by glucose repression and is not induced by the substrates sucrose or raffinose. SUCP encodes both secreted and intracellular forms of invertase via two mRNAs (PERLMAN and HALVORSON 198 1 ; CARLSON and BOTSTEIN 1982). The secreted enzyme is the physiologically important isozyme, and its expression is regulated by glucose repression at the RNA level. The low level constitutive expression of the intracellular invertase is not relevant to this study. Genes required for the derepression of SUCB in response to glucose starvation have been identified in previous mutant searches (CARLSON, OSMOND and BOTSTEIN 198 1 ; NEIGEBORN and CARLSON 1984). The SNF (sucrose nonfermenting) genes fall into three groups of functionally related genes: SNFl and SNF4; SNFP, SNFS and SNF6; and SNF3. These groups are distinguishable on the basis of phenotype and patterns of interaction with extragenic suppressors (NEIGEBORN, RUBIN and CARLSON 1986; ESTRUCH and CARLSON 1990b). Mutations in S N F l (also known as CAT1 and C C R l ) and SNF4 (CAT3) prevent expression of many glucose-repressible genes (CIRIACY 1977; CARLSON, OSMOND and BOTSTEIN 198 1 ; ENTIAN and ZIMMERMANN 1982; NEIGEBORN and CARISON 1984; Genetics 159: 675-684 (November, 1991) SCHULLER and ENTIAN 1987; SCHULLER and ENTIAN 1988). SNFl encodes a protein-serine/threonine kinase (CELENZA and CARLSON 1986). SNF4 encodes a protein that is physically associated with the SNFl kinase and is required for maximal SNFl kinase activity (CELENZA and CARLSON 1989; CELENZA, ENG and CARLSON 1989; FIELDS and SONG 1989). The SNF2, SNFS and SNF6 genes affect not only glucose-repressible genes (NEIGEBORN and CARLSON 1984; ESTRUCH and CARLSON 1990b), but also expression of acid phosphatase (ABRAMS, NEIGEBORN and CARLSON 1986), cell type-specific genes (LAURENT, TREITEL and CARLSON 1990), and T y elements (HAPPEL, SWANSON and WINSTON 199 1). Also snj2 and snf5 mutations cause constitutive expression of protease B; a leaky snf6 allele had no effect (MOEHLE and JONES 1990). Mutations in the three genes were reported to cause singular phenotypes with respect to glucose transport (BISSON 1988). Thus, these genes affect expression of a variety of differently regulated genes, and it seems unlikely that they convey specific regulatory signals. Recent evidence implicates SNF2 and SNFS in transcriptional activation: DNA-bound LexASNF2 and LexA-SNF5 fusion proteins activate transcription from a nearby promoter (LAURENT, REITEL and CARLSON 1990, 1991). The SNF2, SNF5, and SNFG proteins appear to function interdependently because activation by LexA-SNF2 is dependent on SNF5 and SNF6, and activation by LexA-SNF5 is dependent on SNF2 and SNFG (LAURENT, TREITEL and CARLSON 1990, 1991). SNF2 and SNFS are the same as TYE3 and TYE4, respectively (CIRIACY and WILLIAMSON 1981) (M. CIRIACY, personal communication). 676 L. G. Vallier and M. Carlson SNF3 encodes a high-affinity glucose (and fructose) transporter (BISSON et al. 1987; CELENZA, MARSHALLCARLSON and CARISON 1988) but does not affect invertase expression (NEIGEBORN et al. 1986; MARSHALL-CARLSON et ai. 1990). Mutants were recovered because they are defective in growth on raffinose, which requires ability to transport the low levels of fructose released by the action of secreted invertase. Genes required for full derepression of SUC2 have also been identified by other genetic approaches. The MSNl gene was isolated as a multicopy suppressor that restored growth on raffinose in a snfl-ts mutant. Mutations in MSNl cause a few-fold decrease in invertase expression in an otherwise wild-type background (ESTRUCH and CARLSON 1990a). Genes required for glucose repression of SUC2 include HXK2 (ZIMMERMANN and SCHEEL 1977; ENTIAN and MECKE 1982; MA and BOTSTEIN 1986), REG1 (HEX2) (MATSUMOTO, YOSHIMATSU and OSHIMA 1983; NIEDERACHER and ENTIAN 1987), SSN6 (CUCS) (SCHULTZ and CARLSON 1987; TRUMBLY 1988), TUPl (WILLIAMS and TRUMBLY 1990 and references therein), MZGl (NEHLIN and RONNE 1990), GRRl (BAILEY and WOODWORD 1984), RGRl (SAKAI et al. 1990), and CZDl (NEIGEBORN and CARLSON 1987). HXK2 encodes hexokinase PI1 and may function early in the signaling pathway (ENTIAN et al. 1985; MA et al. 1989). The MZGl product is a zinc-finger protein that binds to SUC2 DNA (NEHLIN and RONNE 1990). SSN6 encodes a nuclear protein containing the TPR sequence motif (SCHULTZ, MARSHALL-CARLSON and CARLSON 1990), and TUPl encodes a protein with homology to the B-subunit of transducin (WILLIAMS and TRUMBLY 1990). The mechanism of action of these gene products in glucose repression is not understood. A limitation in our efforts to unravel the regulatory pathway for glucose repression is that, most likely, not all of the relevant genes have yet been identified. For example, the distribution of snf alleles obtained in previous studies suggests that some SNF genes remain to be identified: only one or two snj2, snf4, snf5 and snf6 alleles were isolated (Table 1). We therefore undertook another mutant search in the hopes of recovering snf mutations at new loci. In this search we wished to optimize the recovery of mutants that were only partially impaired in invertase expression. Such mutants could carry either “leaky” mutations or null mutations that caused only a partial defect; in either case, the genes might be of interest. To detect mild growth defects, we screened mutagenized colonies for defective growth on medium containing raffinose plus antimycin A. Our strains use raffinose less efficiently than sucrose, and antimycin A increases the dependence on raffinose by blocking respiration. TABLE 1 Distribution of snfalleles isolated