INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

In Saccharomyces cerevisiae the Ssn6 (Cyc8)-Tup1 complex represses transcription of genes regulated by glucose, cell type, oxygen, DNA damage, and other signals (27, 34, 45, 46, 48, 52, 54, 56, 57, 61). Ssn6-Tup1 is recruited to these promoters by specific DNA-binding proteins, including 2-Mcm1, a1-2, Mig1, Mig2, Rox1, and Rgt1 (1, 27, 31, 39, 47, 51), and mediates repression by interacting with chromatin (43) and/or the general transcriptional machinery (20, 21, 41). In this work we have focused on the role of Mig1 in regulating repression by Ssn6-Tup1 in response to the glucose signal.

Mig1 is a Cys₂-His₂ zinc finger protein (36) that binds to the promoters of SUC, GAL, MAL, and other glucose-repressible genes; mutation of Mig1 or its binding sites partially relieves glucose repression (15, 17, 22, 25, 35, 36, 44, 53, 55). A LexA-Mig1 fusion protein represses transcription of a CYC1-lacZ reporter containing lexA operators. Such repression requires Ssn6-Tup1 and occurs only in glucose-grown cells (47, 51). Mig1 is differentially phosphorylated in response to glucose availability (11, 47), and the localization of Mig1 to the nucleus requires glucose (11). In contrast, no difference in modification was detected for Ssn6 or Tup1 (46, 57), and Ssn6 resides in the nucleus regardless of glucose availability (46). These findings strongly suggest that the recruitment of Ssn6-Tup1 to a promoter by Mig1 is regulated by glucose. However, it remains possible that other mechanisms also contribute to regulation of repression by the Mig1-Ssn6-Tup1 complex. Here we present evidence that LexA-Mig1 confers glucose-regulated repression to a promoter that is not otherwise glucose repressed, thereby excluding any requirement for other promoter-bound glucose-regulated factors. We also show that repression by LexA-Ssn6 is not glucose regulated, indicating that the repressor function of the Ssn6-Tup1 complex is not directly regulated by the glucose signal.

The differential phosphorylation of Mig1 in response to glucose suggests that phosphorylation controls its activity in repression, and genetic evidence implicates the Snf1 (Cat1) protein kinase in regulating Mig1. The Snf1 kinase is activated by glucose starvation and is required for expression of glucose-repressed genes (9, 23, 58, 59). Mig1 is thought to function downstream from Snf1 in the pathway, because a mig1 mutation suppresses the snf1 mutant defects in SUC2 and GAL1 expression (25, 53). Thus, Snf1 appears to inhibit repression by Mig1. Snf1 also inhibits the function of a hybrid Mig1-VP16 activator in the absence of glucose (37). Deletion analysis of Mig1 defined regions that both inhibit repression by Mig1 in the absence of glucose and confer inhibition of Mig1-VP16 by Snf1 (37). Finally, mutation of SNF1 causes constitutive nuclear localization of Mig1 (11).

In this study, we have examined the role of the Snf1 protein kinase in regulating Mig1 function. We show that Snf1 is required for the phosphorylation of Mig1 in vivo and that the two proteins interact in the two-hybrid system and coimmunoprecipitate. We present evidence that Mig1 is phosphorylated in vitro in a Snf1-dependent reaction. Finally, we show that mutation of four putative Snf1 recognition sites in Mig1 eliminates most of the differential phosphorylation of Mig1 in response to glucose and improves the two-hybrid interaction with Snf1.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Strains and genetic methods. The S. cerevisiae strains used are listed in Table 1. The Escherichia coli strains used for propagation of plasmid DNA were XL1-Blue and DH5. Standard genetic methods were used, and yeast cultures were grown in synthetic complete (SC) medium lacking appropriate supplements to maintain selection for plasmids (42).

Oligonucleotides. Oligonucleotides used for PCR are as follows, with serine-to-alanine conversions underlined: OL-H1, 5'-ACTACCATAGCCATGGGCGGCCGCCAAAGCCCATATCCAATG-3'; OL-L1, 5'-TCGAGTGCTGTATATAAAACCAGTGGTTATATGTACAGTACC-3'; OL-L2, 5'-TCGAGGTACTGTACATATAACCACTGGTTTTATATACAGCAC-3'; OL-S1, 5'-AATAGCCATAGTGGCGCTAGACTGAAACTGAAC-3'; OL-S2, 5'-ATATTACCAGGTCCGCGAGCTTTAACGGATTTTCAA-3'; OL-S3, 5'-CAGTTGAAGAGACCAGCTGCTGTTTTAAGTTTGAAC-3'; and OL-S4, 5'-ATGCTAAGTAGAGCTGCTGCTGGTACGAATTTGCAC-3'.

Plasmids. To construct pHA-Mig1, the SmaI-KpnI fragment from pMIG1 (36) was cloned into the cognate sites of pKB174, a derivative of pRS426 lacking the NotI site (2). The resulting plasmid was subjected to site-directed mutagenesis with OL-H1 to introduce a NotI site 3' to the initiating ATG of Mig1. The resulting DNA was digested with NotI and ligated to a NotI fragment from pGTEP encoding a triple-hemagglutinin (HA) epitope tag (50). pHA-Mig1 partially complements a mig1 mutation.

View larger version (26K): [in this window] [in a new window]   FIG. 1.   (A) Structures of Mig1 and Mig1Z. Stippled boxes represent the two zinc fingers. The two regions that are involved in the inhibition of Mig1 function are shown as hatched boxes, and the solid box represents the C-terminal 24 amino acids required for repression (37). Serine residues that are potentially phosphorylated by Snf1 are marked. (B) Potential Snf1 recognition sites. The consensus Snf1 recognition sequence (10) is shown. , hydrophobic residue. Preferred residues: position 5, L > F = I = M > V; position +4, L > I > F > M > V. Potential Snf1 recognition sites in Mig1 are shown and are numbered according to the position of the phosphorylatable serine. The alanine substitutions resulting from site-directed mutagenesis are indicated, and the mutated sites are designated by an asterisk.

Invertase and -galactosidase assays. Invertase activity was assayed as previously described (7, 16) and expressed as micromoles of glucose released per minute per 100 mg of cells (dry weight). -Galactosidase activity was assayed in permeabilized cells (42) and expressed in Miller units (32) or was assayed in protein extracts (8) and expressed as units per milligram of protein (3).

Immunoblot analysis. Cells were grown to mid-log phase in selective SC medium containing 5% glucose (repressed) and derepressed by a shift to 0.05% glucose for 1 h. Cells were collected by centrifugation for 2 min and frozen immediately at 70°C without washing. For Fig. 2C, cells were collected by rapid filtration onto a 0.8-µm-pore-size filter (Micron Separations), and the cell cake was scraped off into methanol at 80°C. Protein extracts were prepared as described previously (8). Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and analyzed by immunoblotting. Primary antibodies were polyclonal LexA antibody (a gift of J. Kamens and R. Brent, Massachusetts General Hospital, Boston) or monoclonal HA antibody (Boehringer Mannheim Biochemical). Antibodies were detected by enhanced chemiluminescence with ECL or ECL Plus reagents (Amersham).

Coimmunoprecipitation assays. Preparation of protein extracts and immunoprecipitation procedures were essentially as described previously (8). The extraction buffer was 50 mM HEPES (pH 7.5)-150 mM NaCl-0.1% Triton X-100-1 mM dithiothreitol-10% glycerol, containing 1 or 2 mM phenylmethylsulfonyl fluoride and complete protease inhibitor cocktail (Boehringer Mannheim). rProtein A immobilized on Sepharose beads (RepliGen) was added to protein lysates, which were rotated for 20 min and then cleared by centrifugation at 12,000 rpm for 10 min. Anti-HA antibody was added, and samples were mixed for 30 min and cleared by centrifugation for 5 min at 10,000 rpm. The supernatant was mixed with immobilized rProtein A for 1.5 h. The beads were collected by brief centrifugation and washed four times with 1 ml of extraction buffer without protease inhibitor cocktail. The entire procedure was done at 4°C or on ice.

Immune complex kinase assays. Preparation of protein extracts and immunoprecipitation were as described above. Beads were then washed in kinase buffer (50 mM Tris-HCl [pH 7.5], 10 mM MgCl₂, 1 mM dithiothreitol, 0.1% Triton X-100) and resuspended in 20 µl of kinase buffer. The kinase reaction was initiated by the addition of 20 µCi of [-³²P]ATP (3,000 Ci/mmol; NEN). Reaction mixtures were incubated at room temperature for 30 min, and reactions were terminated by the addition of 30 µl of 2× sample buffer. Proteins were separated by SDS-PAGE. After electrophoresis, the gel was stained, washed extensively in destaining solution containing 10 mM sodium pyrophosphate, dried, and exposed to film at 70°C with an intensifying screen.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Glucose-regulated repression by LexA-Mig1. Previous work showed that LexA-Mig1 represses transcription of a lexAop-CYC1-lacZ reporter only in glucose-grown cells and that repression depends on Ssn6-Tup1 (47, 51) (Table 2). These findings suggested that recruitment of Ssn6-Tup1 by Mig1 is regulated by the glucose signal. However, the CYC1 promoter responds to glucose, and it remained possible that other factors bound to this reporter contribute to the regulation of repression. To address this issue, we tested the ability of LexA-Mig1 to repress transcription of a reporter driven by the LEU2 UAS and HIS3 promoter, with no or one lexA operator 5' to the UAS. In glucose-grown cells, LexA-Mig1 repressed LEU2-HIS3-lacZ expression 11-fold, whereas in raffinose-grown cells, LexA-Mig1 did not repress transcription better than LexA₈₇ alone (Table 2); levels of the LexA-Mig1 protein were comparable (data not shown). Thus, glucose-regulated repression by LexA-Mig1 is not promoter specific or dependent on other glucose-regulated factors bound at the reporter.

Effect of snf1 on LexA-Mig1Z function. Genetic evidence suggests that the Snf1 protein kinase inhibits repression by Mig1 during glucose limitation (25, 37, 53). To further examine the role of Snf1 in regulating Mig1 function, we used LexA-Mig1Z, which was stably expressed in a snf1 mutant; for unknown reasons, LexA-Mig1 was not detectable. In wild-type cells, LexA-Mig1Z conferred glucose-dependent, Ssn6-dependent repression of a reporter (Table 2 and data not shown); thus, the zinc finger domain is dispensable for regulated repression.

Requirement for the Snf1 protein kinase in phosphorylation of Mig1 in vivo. Previously, we showed that LexA-Mig1 is differentially phosphorylated in response to glucose availability (47) (Fig. 2A). To test whether the Snf1 protein kinase is required for this phosphorylation, we used immunoblot analysis to examine tagged Mig1 proteins in a snf1 mutant. In wild-type cells, LexA-Mig1Z was also differentially modified, similarly to LexA-Mig1 (Fig. 2A). In snf1 mutant cells, however, the major band migrated at the position predicted for the unmodified protein in both glucose-repressed and derepressed cells (Fig. 2A). We next examined the HA-Mig1 protein expressed from its own promoter. HA-Mig1 was differentially modified in wild-type cells but not in snf1 mutants (Fig. 2B). In control experiments, a snf1 mutation did not alter the phosphorylation of an unrelated protein encoded by SFH1 (4) (data not shown). Thus, the Snf1 protein kinase is required for the phosphorylation of Mig1.

View larger version (89K): [in this window] [in a new window]   FIG. 2.   Immunoblot analysis of Mig1 fusion proteins. Cultures were grown in selective SC medium plus 5% glucose (repressed) (lanes R). Mid-log-phase cultures were derepressed by a shift to SC medium plus 0.05% glucose for 1 h (lanes D). Extracts were prepared, and proteins were separated by SDS-PAGE in 7.5% polyacrylamide and subjected to immunoblot analysis with anti-LexA (A, C, and D) or anti-HA (B). (A and B) Protein extracts (25 and 50 µg for wild-type [WT] and snf1-K84R strains, respectively) were prepared from strains MCY829 and MCY2692 transformed by pLexA-Mig1 or pLexA-Mig1Z (A) and pHA-Mig1 (B). (C) Protein extracts (25 µg) were prepared from strain YM4738 expressing LexA-Mig1 from vector pJH106. Cells were collected by rapid membrane filtration (Filt.) or by centrifugation for 2 min (Cent.) (see Materials and Methods). (D) Protein extracts (25 µg for the wild type and 50 µg for reg1 and hxk2 strains) were prepared from strains FY250, MCY829, MCY3278, and MCY3541 transformed with pLexA-Mig1. The position of the 66-kDa size marker is indicated.

Phosphorylation of Mig1 in reg1 and hxk2 mutants. We next examined mutants in which Snf1 is active in glucose-grown cells. If differential phosphorylation of Mig1 reflects the functional status of Snf1, then mutations that affect the regulation of Snf1 should also influence Mig1 phosphorylation. The REG1 gene encodes a targeting subunit that directs the function of protein phosphatase 1 in the glucose response (49). Mutation of REG1 relieves glucose repression of Snf1-dependent genes (24) and causes the Snf1 protein kinase complex to assume an active conformation even in glucose-grown cells (23, 30). Mutation of HXK2, encoding hexokinase PII, causes similar phenotypes (23, 24).

Two-hybrid interaction between Mig1 and both wild-type and kinase-dead Snf1 proteins. The preceding data show that Snf1 is required for phosphorylation of Mig1 but do not address whether Snf1 phosphorylates Mig1 directly or controls the phosphorylation of Mig1 by another kinase. To assess the interaction of Snf1 with Mig1 in vivo, we used the two-hybrid system (14). In glucose-grown cells, LexA-Snf1 did not interact significantly with GAD-Mig1Z, but LexA-Snf1K84R interacted strongly (Table 3). The mutant Snf1K84R protein contains a substitution of Arg for the conserved Lys84 in the ATP binding site and exhibits no catalytic activity (8). After cells were shifted to 0.05% glucose for 3 h, -galactosidase activity could be detected for both combinations. When cells were grown in raffinose, no significant interaction was detected (data not shown), consistent with evidence that Mig1 is cytoplasmic under derepressing conditions (11). These data support the view that Mig1 is a substrate of Snf1 in vivo and suggest that for the wild-type Snf1, glucose deprivation transiently enhances interaction with Mig1.

Coimmunoprecipitation of Snf1 and Mig1. To obtain biochemical evidence for the interaction of Snf1 and Mig1 in vivo, we tested LexA-Snf1 for coimmunoprecipitation with HA-Mig1, expressed from the MIG1 promoter (Fig. 3). Whole-cell extracts were prepared from cells expressing both proteins, and HA-Mig1 was immunoprecipitated with monoclonal anti-HA antibody. Immunoblot analysis of the precipitate showed that LexA-Snf1 coprecipitated; it was expected that only a small fraction of the LexA-Snf1 would be associated with Mig1. In control experiments, LexA-Snf1 was not detected when the extract contained HA instead of HA-Mig1. Similar results were obtained with LexA-Snf1K84R (Fig. 3); the mutant protein did not coprecipitate better than the wild-type LexA-Snf1, probably because the two-hybrid system and coimmunoprecipitation are not comparable assays.

View larger version (59K): [in this window] [in a new window]   FIG. 3.   Coimmunoprecipitation of LexA-Snf1 and LexA-Snf1K84R with HA-Mig1. Strains MCY3573 and MCY3640 were transformed with plasmids expressing LexA-Snf1 and LexA-Snf1K84R, respectively, and either HA-Mig1 (from the MIG1 promoter) or HA (from the ADH1 promoter). Protein extracts were prepared from cells grown in glucose, and proteins (200 µg) were immunoprecipitated (IP) with anti ()-HA antibody, separated by SDS-PAGE in 7.5% polyacrylamide, and immunoblotted with -LexA (upper panel). The input protein (25 µg) is also shown (middle panel); an arrow indicates the position of the Snf1 fusion. The same immunoblot was reprobed with -HA to confirm the precipitation of HA-Mig1 (lower panel). No band was detected at the position of LexA-Snf1 when HA-Mig1 was immunoprecipitated from extracts lacking LexA-Snf1. WT, wild type.

Snf1-dependent phosphorylation of Mig1 in vitro. We next addressed the ability of Snf1 to phosphorylate Mig1 in vitro. Extracts were prepared from cells expressing HA-Snf1 and LexA-Mig1 from the ADH1 promoter, and HA-Snf1 was immunoprecipitated with anti-HA. The immune complexes were resuspended in kinase assay buffer and incubated with [-³²P]ATP. The proteins were separated by gel electrophoresis, and the phosphorylated products were visualized by autoradiography (Fig. 4A). In addition to the products usually detected in such assays, including Snf1, Sip1, and Gal83 (60), a phosphorylated protein corresponding to LexA-Mig1 was detected (Fig. 4A, lane 1). This product was absent in assays of extracts containing only the LexA moiety (expressed from the parental vector) (Fig. 4A, lane 3). Control experiments with the kinase-dead HA-Snf1K84R mutant protein confirmed that the kinase activity detected in this assay was dependent on Snf1 (Fig. 4A, lane 2), and no phosphorylated LexA-Mig1 was detected even upon overexposure (Fig. 4A, lanes 4 and 5). In an independent experiment, we similarly detected phosphorylation of LexA-Mig1 in immune complex assays of the wild-type HA-Snf1, but not the mutant kinase, and also showed that no phosphorylation was detected in controls with untagged Snf1 expressed at the same level (data not shown). Thus, LexA-Mig1 is phosphorylated in vitro in a Snf1-dependent reaction.

View larger version (55K): [in this window] [in a new window]   FIG. 4.   Snf1-dependent phosphorylation of Mig1 in vitro. Strain FY250 (wild type [WT]) was transformed with plasmids expressing HA-Snf1 or HA-Snf1K84R and LexA-Mig1 or the LexA moiety (from the parental vector). Protein extracts were prepared from cells grown in selective SC medium plus 2% glucose; previous studies have shown that the Snf1 kinase is activated during preparation of the extracts (13, 58). (A and B) Proteins (200 µg) were immunoprecipitated with anti-HA antibody. (A) Half of the immunoprecipitate was resuspended in kinase assay buffer and incubated with [-32P]ATP. Proteins were then separated by SDS-PAGE in 8% polyacrylamide, and the gel was subjected to autoradiography to detect phosphorylated products. The left panel (lanes 1 to 3) shows a 3.5-h exposure, and the right panel (lanes 4 and 5) shows a 15-h exposure of lanes 1 and 2. (B) The remaining half of the sample was subjected to SDS-PAGE and immunoblot analysis with anti-HA to confirm the immunoprecipitation of HA-Snf1 and HA-Snf1K84R. We could not reproducibly detect LexA-Mig1 by immunoblot analysis of the immunoprecipitates. (C) Input proteins (10 µg) were analyzed by immunoblot analysis to verify the expression of LexA-Mig1. The phosphorylated LexA-Mig1 product detected in panel A corresponds to the lower band in this panel. Numbers on the right of each panel are molecular masses in kilodaltons.

Mutation of putative Snf1 phosphorylation sites in Mig1. We identified potential Snf1 phosphorylation sites in Mig1 based on their similarity to the consensus substrate recognition sequence (10), which contains an arginine at position 3 and hydrophobic residues at positions 5 and +4 relative to the phosphorylated serine (Fig. 1B). This consensus sequence was determined by assaying the ability of purified Snf1 kinase to phosphorylate variants of a synthetic peptide that is recognized by the mammalian Snf1 homolog, AMP-activated protein kinase (5, 33).

View larger version (33K): [in this window] [in a new window]   FIG. 5.   Analysis of mutant LexA-Mig1 fusion proteins. (A) Strain MCY3640 was transformed with plasmids derived from the vector pJH106. Cultures were grown selectively in SC medium plus 5% glucose (repressed) (lanes R) and were derepressed by a shift to SC medium plus 0.05% glucose for 1 h (lanes D), and protein extracts were prepared. Proteins (25 µg for cells expressing wild-type LexA-Mig1 [WT] and 50 µg for cells expressing mutant LexA-Mig1 proteins) were separated by SDS-PAGE in 7.5% polyacrylamide and subjected to immunoblot analysis with anti-LexA antibody. The position of the 66-kDa size marker is indicated. (B) Transformants of strains MCY3640 (mig1) (shaded bars) and YM4738 (mig1 mig2) (open bars) expressing the indicated LexA-Mig1 protein were grown selectively in SC medium plus 2% raffinose and 0.05% glucose. Derepression of SUC2 was monitored by assaying invertase activity for 4 to 11 transformants. Standard errors were <13%. Immunoblot analysis confirmed that mutant proteins were present at lower levels than wild-type LexA-Mig1 in strain YM4738, consistent with panel A. Similar assays of transformants shifted from 5 to 0.05% glucose for 1 h did not reveal any delay in derepression for strains expressing mutant proteins.

mig1 mig2

mig1 mig2

mig1 mig2

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Previous evidence suggested that Mig1 recruits the Ssn6-Tup1 corepressor to glucose-repressed promoters in response to the glucose signal. Here we have further examined the role of Mig1 in regulating repression. First, we show that LexA-Mig1 confers glucose-regulated repression to a LEU2-HIS3-lacZ reporter, thereby excluding any requirement for other regulatory factors specific to glucose-regulated reporters. Similar studies of LexA-Mig1Z further indicate that the zinc finger region is not required for regulated repression. Second, we show that repression by LexA-Ssn6 is not regulated by glucose. These experiments substantiate the model that regulation is achieved by the regulated recruitment of Ssn6-Tup1 by Mig1.

The differential phosphorylation of Mig1 in response to glucose suggested that phosphorylation regulates its repressor function (11, 47), and genetic evidence indicated that the Snf1 protein kinase inhibits Mig1 function during glucose starvation (25, 37, 53). Here we present evidence that Snf1 regulates the phosphorylation of Mig1. We show that modification of Mig1 is dramatically reduced in a snf1 mutant, indicating that Snf1 is required for the phosphorylation of Mig1. Consistent with these observations, Snf1 kinase activity increases in glucose-limited cells (58, 59). Conversely, in glucose-grown hxk2 and reg1 mutants, which are defective in glucose inhibition of the Snf1 kinase activity (23) and glucose repression of Snf1-dependent genes (24), the migration patterns of LexA-Mig1 resemble that of the derepressed wild type. These data strongly suggest that the regulation of Snf1 kinase activity is coupled to the regulation of Mig1 modification, with the caveat that hxk2 and reg1 may also affect Mig1 by other mechanisms. During the preparation of this paper, the Snf1-dependent phosphorylation of a Mig1-VP16 protein containing the N-terminal two-thirds of Mig1 (residues 1 to 351) was reported; however, this truncated Mig1 fusion differs from the full-length Mig1 proteins examined here in that it is not phosphorylated in glucose-grown cells (38).

Several lines of genetic and biochemical evidence support the view that Snf1 phosphorylates Mig1 in vivo. First, Snf1 and Mig1 interact in the two-hybrid system. Moreover, the kinase-dead mutant Snf1K84R gives a stronger signal than wild-type Snf1, and mutant Mig1 proteins with Ser-to-Ala substitutions in consensus Snf1 recognition sites give a stronger signal than wild-type Mig1. A shift to low glucose causes an increase in interaction between wild-type Snf1 and Mig1, presumably transient because no interaction was detected in cells grown in raffinose. Second, Snf1 coimmunoprecipitates with Mig1 from cell extracts. Third, mutation of all four putative Snf1 recognition sites eliminates most of the differential phosphorylation of Mig1 in response to glucose. Finally, functional assays of the mutant LexA-Mig1 proteins revealed defects of up to 2.8-fold in release of repression of SUC2, and the magnitude is most likely underestimated due to the reduced levels of the mutant proteins. Studies of Mig1-VP16 similarly showed that mutation of serines 278, 310, and 311 affects its phosphorylation and reduces the Snf1 dependence of its activation function 3.8-fold, although protein levels were not reported (38).

These studies of the relationship of Snf1 and Mig1 in vivo are further supported by in vitro evidence that immunoprecipitated Snf1 kinase phosphorylates coprecipitated Lex-Mig1. This reaction was dependent on Snf1 activity, and no phosphorylation was detected in immune complex assays of Snf1K84R. The simple interpretation is that Mig1 is phosphorylated by Snf1, but a more complicated scenario, in which Snf1 is still intimately involved in the phosphorylation of Mig1, cannot be excluded. It is possible that Snf1 phosphorylates and activates an associated Snf1-dependent kinase, which then phosphorylates Mig1; however, this model is difficult to reconcile with the effects of mutations in Mig1 on its phosphorylation and two-hybrid interaction with Snf1.

Although most of the phosphorylation of Mig1 in vivo depends on the Snf1 kinase, Snf1 may not be directly responsible for all of the phosphorylation events. Mutation of all four Snf1 consensus recognition sites did not reduce phosphorylation of Mig1 nearly as substantially as the snf1 mutation. Mig1 may contain other Snf1 recognition sites, unrelated to the defined consensus, and/or Snf1 may regulate the phosphorylation of Mig1 by another protein kinase. Consistent with this view, analysis of Mig1-VP16 identified a Snf1-dependent phosphorylation site at serine 108, which does not resemble a Snf1 site (38).

Phosphorylation of Mig1 could regulate the recruitment of Ssn6-Tup1 to a promoter by affecting any of several steps: binding of Mig1 to the promoter, interaction of Mig1 with the Ssn6-Tup1 corepressor, or localization of Mig1 to the nucleus. It is unlikely that DNA binding is regulated, because regulated repression was achieved by LexA-Mig1 bound to lexA operators. The possibility that phosphorylation disrupts the interaction of Mig1 with Ssn6-Tup1 has not been addressed, but this cannot be the only mechanism, because Snf1 affects activation by Mig1-VP16 (37) and affects localization of Mig1 in an ssn6 mutant (11). Evidence that the differential localization of Mig1 is Snf1 dependent and correlates with its differential phosphorylation (11) strongly suggests that phosphorylation functions as a regulatory signal for localization.