Media and growth assays:

Rich (YPD), YPGlycerol, minimal (SD), synthetic complete (SC), 5-fluoroorotic acid (5-FOA), presporulation, sporulation, and defined inositol media used in all experiments except Figures 5B and 7 were as described (Shirra and Arndt 1999). For Figures 5B and 7, inositol media were prepared from a yeast nitrogen base containing ammonium sulfate but lacking inositol (QBIOgene) and supplemented with the appropriate nutrients; inositol was added to 100 μm where indicated. Rapamycin (Sigma, St. Louis) was added to a final concentration of 200 ng/ml from a 1 mg/ml stock solution in 90% ethanol/10% Tween 20. Growth assays were performed by growing cells overnight at 30° in the appropriate media, washing twice with water, and resuspending to a concentration of 1 × 10⁸ cells/ml. Two microliters of 10-fold serially diluted cultures were spotted on the relevant solid media and incubated at 30° for the indicated times.

Yeast strains:

The S. cerevisiae strains used in these experiments are listed in Table 1. All strains are congenic with FY2, a GAL2⁺ derivative of S288C (Winston et al. 1995). FY strains were obtained from Fred Winston; KY and PY strains were generated by genetic crosses or integrative transformations. The INO1-lacZ reporter plasmid pJH334, which contains 0.5 kb of INO1 promoter and 0.4 kb of INO1 coding sequence ligated in frame to the lacZ sequence, was integrated at URA3 as described (Lopes et al. 1991). A strain containing sit4Δ::HIS3 was constructed by PCR-mediated one-step gene disruption (Ausubel et al. 2004). A disruption of INO4 by LEU2 was created in the FY2 background as described (Ambroziak and Henry 1994). Disruption of INO2 was achieved by PCR-mediated amplification of genomic DNA from an ino2::TRP1 strain (gift of S. A. Henry; Nikoloff and Henry 1994), followed by a one-step gene replacement. Strains containing multiple disruptions with the same auxotrophic marker were obtained from genetic crosses and confirmed by PCR.

Isolation and cloning of sit4-731:

The genetic screen in which sit4-731 was identified has been described (Shirra et al. 2001); sit4-731 was previously named sup731. Several plasmids containing sequences from different chromosomal regions were capable of complementing the Ino⁺ phenotype of PY731 (snf1Δ sup731); however, only plasmids containing SIT4 sequences fully complemented the Gly⁻ phenotype (inability to grow on media containing glycerol as the carbon source) of SNF1 sit4-731 strains. We confirmed that the sup731 mutation was in the SIT4 gene by linkage analysis using a genetic cross between PY731 and a SNF1 SIT4 strain containing URA3 sequences integrated at STP4, the gene immediately adjacent to SIT4. To follow the snf1Δ10 mutation, we examined the ability of segregants to utilize sucrose as a carbon source (Suc phenotype); snf1 mutants, regardless of the sit4 allele, are Suc⁻. No Suc⁻, Ino⁺, Ura⁺ spores were obtained from 25 tetrads. To identify the mutation in sit4-731, plasmid pDA9 (see below) was digested with XbaI, and the 10.1-kb vector fragment was transformed into PY731 for gap repair (Ausubel et al. 2004), creating pLM1. The entire SIT4 coding region in pLM1 was sequenced on both strands to identify the single point mutation.

Plasmids:

Standard procedures were used to construct plasmids and isolate DNAs (Ausubel et al. 2004). pDA9, one of the original complementing plasmids obtained from a YCp50-based yeast genomic library (Rose et al. 1987), contains genomic sequence from chromosome IV, 364652–371650 (numbering per Saccharomyces Genome Database, http://www.yeastgenome.org). Subclones containing SIT4 sequences were constructed by inserting the 2.1-kb PvuII-BstEII (blunted) fragment from pDA9 into the SmaI sites of pRS316 and pRS426 (Sikorski and Hieter 1989), creating pPS83 and pPS86, respectively. Plasmids encoding Sit4-TT (true truncation) were obtained by first subcloning the 1.5-kb BamHI-SphI fragment from pPS83 into pUC19 to generate a plasmid with a single XbaI site in SIT4, pPS92. The oligonucleotide PSO68 (5′-CGAGTTAGGGAGGGCATGCCGTCGTGTTACTAGTAGAACCCGTAAACTTGAGTAATTTGT-3′) contains DNA sequence upstream of the codon affected by the sit4-731 point mutation (boldface A) placed adjacent to the native stop codon of SIT4 (underlined), which therefore removes DNA encoding the C-terminal 184 amino acids. The native SphI site downstream of the stop codon was included in PSO68 to facilitate subcloning. A PCR fragment amplified from pLM1 with PSO37 (5′-CATGCATATCGAGTCGAAGCCC-3′) and PSO68 was digested with XbaI and SphI and subcloned into the same sites in pPS92, to create pMM2. To introduce this truncated SIT4 sequence into a yeast expression plasmid, the 0.97-kb BamHI-SphI fragment from pMM2 was subcloned into the same sites in pPS83 and pPS86, creating pMM19 and pMM21, respectively. To introduce an immunoreactive tag at the N terminus of Sit4, the triple HA1 epitope (Tyers et al. 1992) from pMR2307 (gift from M. C. Schmidt) was amplified with primers that incorporated flanking XbaI sites and subcloned into the XbaI site of pPS92, to create pMM1. Next, the 1.6-kb EcoRI (blunted)-SphI fragment from pMM1 was subcloned into the BamHI (blunted)-SphI backbone from pPS83 and pPS86 to create pMM25 and pMM17, respectively. Subclones containing untagged and HA-tagged versions of sit4-731 were derived by substituting the 0.6-kb BsiWI-NruI sequence from pLM1 (see previous section) for the identical sequence in pPS83, pPS86, pMM25, and pMM17, generating pSR1, pSR5, pSR3, and pSR7, respectively. The presence of the sit4-731 mutation was confirmed by sequence analysis. Finally, the 0.8-kb BsiWI-XhoI fragment from pMM19 was subcloned into the same sites in pMM25 and pMM17 to create HA-tagged versions of Sit4-TT encoded by pSR9 and pSR11, respectively. Plasmid maps are available upon request. Plasmids encoding wild type and histone H3 derivatives were obtained from S. Y. Dent (Edmondson et al. 2002).

RNA analysis:

Cells grown at 30° as described in the figure legends were harvested by centrifugation, washed with water, and frozen. RNA isolation, Northern analysis, and Northern blot probe preparation have been described (Shirra and Arndt 1999). S1 nuclease protection assays were performed using oligonucleotides KA124 (DED1) and PSO87 (INO1), 5′-GTAGTCTTGAACAGTGGGCGTTACATCGAAGCGGCCACTAGCTGTCTTCGTAACTACAGCATTTTCGTAGCTGATGAACC-3′, labeled with ³²P at their 5′-ends by polynucleotide kinase (New England Biolabs, Beverly, MA) (Spencer and Arndt 2002). Quantitation was achieved using a Fuji phosphorimager and Image Gauge software (Fuji).

β-Galactosidase assays:

Extract preparation, β-galactosidase assays, and unit calculations were performed as described (Rose and Botstein 1983). A₄₂₀/min readings were obtained using the Power WaveX 340 96-well microtiter plate reader and KC4 software (Bio-tek Instruments, Winooski, VT), taking readings every 30 sec for 90 or 120 min.

Immunoblot analysis:

PY502 was transformed with pPS83, pMM25, pPS86, pMM17, pSR1, pSR3, pSR5, pSR7, pMM19, pSR9, pMM21, and pSR11 and grown in SC-Ura medium to a cell density of 3–4 × 10⁷ cells/ml. Cells were harvested and washed in lysis buffer containing 20 mm HEPES-KOH pH 7.4, 100 mm sodium acetate, 2 mm magnesium acetate, 10% glycerol, 1 mm DTT with protease inhibitors (2.2 μm pepstatin, 0.67 μm leupeptin, 2.0 μg/ml chymotrypsin, 2.2 mm benzamidine, 1.0 mm phenylmethylsulfonyl fluoride). Cells were lysed in the presence of glass beads and lysis buffer by six pulses of 30 sec each using a Mini-BeadBeater 8 (Biospec Products, Bartlesville, OK), and extracts were clarified by centrifugation. Extract concentrations were determined by Bradford assay using crystalline bovine serum albumin as a standard, and 100 μg of total extract were loaded on an SDS-15% polyacrylamide gel. Proteins were transferred to nitrocellulose and immunoblotted as described (Harlow and Lane 1988). Western blots were probed with anti-HA1 12CA5 mouse monoclonal antibody (Boehringer Mannheim, Indianapolis) at a 1:3000 dilution and with sheep anti-mouse horseradish-peroxidase-coupled secondary antibody (Amersham Pharmacia) at a 1:5000 dilution, and the immunoblots were developed with a chemiluminescent detection system (Pierce, Rockford, IL).

Chromatin immunoprecipitations:

Strains were grown at 30° in 400 ml of +Ino medium (200 μm inositol) to a cell density of 1–2 × 10⁷ cells/ml. One-half of the culture was immediately processed for crosslinking as described below. The remaining 200 ml of each culture was harvested, washed once with water, once with −Ino medium (0 μm inositol), resuspended in 200 ml of −Ino medium, and allowed to grow an additional 4 hr at 30°. Crosslinking of the chromatin was achieved as described (Komarnitsky et al. 2000). Glass beads were added to cells thawed in 1 ml of FA lysis buffer (Komarnitsky et al. 2000) with 0.5% SDS, and the mixture was vortexed for 15 min, using intervals of 30 sec vortexing and 30 sec rest on ice. Chromatin was pelleted by ultracentrifugation, washed, and resuspended in 1.5 ml of FA lysis buffer. The chromatin was sonicated to an average size of ∼300 bp (range 100–850 bp) with a Misonix 3000 sonicator (Farmingdale, NY). Sonication was performed using cycles of 20 sec sonication with 1 min rest; the first two cycles were at output of 2, followed by 15 cycles at output of 4. Chromatin was chilled during the entire sonication procedure. The extracted chromatin was adjusted to 4.2 ml with FA lysis buffer, and cell debris was pelleted by ultracentrifugation. TBP was immunoprecipitated from 450 μl of chromatin, which had been adjusted to 275 mm NaCl, by overnight incubation with 2 μl of anti-TBP antibody (gift from Greg Prelich), followed by 2 hr incubation with 30 μl of a 50% slurry of protein A sepharose beads (Amersham Pharmacia) in FA lysis buffer with 275 mm NaCl. Immunoprecipitated chromatin was washed essentially as described (Komarnitsky et al. 2000). Chromatin was eluted in 250 μl of 50 mm Tris-HCl, pH 7.5, 10 mm EDTA, 1% SDS at 65° for 15 min and an additional 250 μl of TE was added to the elution. To reverse the crosslinking, the samples were incubated for 1 hr at 42° and overnight at 65° in the presence of 400 μg Pronase (Roche). Immunoprecipitated DNA that was extracted with phenol:chloroform:isoamyl alcohol (25:24:1) and precipitated with ethanol was resuspended in one-tenth the original volume (45 μl) with TE. Input DNA was treated to reverse the crosslinking by the same procedure as above and resuspended in its original volume. Four microliters of 1:100 and 1:200 dilutions of input DNA and 2 and 4 μl of precipitated DNA were amplified by quantitative PCR in a 10-μl reaction with 2.5 pmol each of primers flanking the INO1 TATA region (5′-GGCTAAATGCGGCATGTGAAAAGT-3′ and 5′-CGGAGGTGATTGGAGCAATATTATC-3′), and subtelomeric primers from chromosome VI-R (Vogelauer et al. 2000), 100 μm dNTPs, 0.1 μl [α-³²P]dATP, and 0.5 units Platinum Taq (Invitrogen, San Diego). Amplification was performed for 22 cycles of 94° for 30 sec, 55° for 30 sec, and 72° for 45 sec. Samples were separated on a 6% native polyacrylamide gel and quantitated by phosphorimaging.

Identification of sit4 as a suppressor of a snf1Δ mutation:

In an earlier report, we described the identification of mutations in the fatty acid biosynthesis pathway as suppressors of the inositol auxotrophy (Ino⁻ phenotype) caused by deletion of the SNF1 gene (Shirra et al. 2001). An additional suppressor mutation, sup731, was identified in this screen but the SUP731 gene proved difficult to identify, because plasmids containing at least five different chromosomal regions complemented or partially complemented the suppression phenotype. We subsequently discovered that strains containing the sup731 mutation by itself grow poorly on media containing glycerol as the carbon source. This phenotype was masked by the snf1Δ mutation in the original suppressor strain because Snf1 is also required for growth on glycerol media. We found that only plasmids containing SIT4 sequences could complement the sup731 mutation for both growth on glycerol media and suppression of snf1Δ. Further genetic analysis confirmed that the SUP731 gene is allelic to SIT4 (see materials and methods), and we renamed the mutation sit4-731. As shown in Figure 1A, the sit4-731 mutation suppresses the growth defect of a snf1Δ strain on medium lacking inositol. We also constructed a strain containing a complete deletion of the sit4 coding region. Deletion of SIT4 causes slow growth but not inviability in our strain background (S288C), indicating the presence of an SSD1-v allele (Sutton et al. 1991). The sit4Δ mutation also partially suppresses the inositol auxotrophy of snf1Δ strains. It should be noted that strains containing deletions in both genes grow very slowly even in the presence of inositol, and this severe growth defect must be taken into account when visualizing the suppression on solid media.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.—

Mutations in SIT4 suppress the inositol auxotrophy of snf1Δ strains. (A) Equal numbers of cells in 10-fold serial dilutions were spotted onto synthetic media lacking or containing 200 μm inositol and allowed to grow for 5 days at 30°. Note that strains containing sit4 mutations are less healthy. The following strains were examined: PY165, PY133, PY731, and PY515. (B) Northern blot analysis of INO1 mRNA levels. Samples were grown to early log phase in 200 μm inositol (+) or washed into medium lacking inositol and allowed to grow an additional 4 hr (−). The following strains were examined: PY165, PY133, PY523, PY859, PY499, and PY515. (C) β-Galactosidase assays of INO1-lacZ expression. Cells were grown to early log phase in medium containing 200 μm inositol (0 hr), washed in medium lacking inositol, and harvested at the indicated times. Results presented are the mean average from three colonies assayed at two extract concentrations. Error bars represent one standard deviation of the mean. The following strains were used: PY177, PY539, PY520, PY518, PY498, and PY536.

Transcription of the INO1 gene is required for cell growth in the absence of inositol and is repressed by the presence of inositol (reviewed in Henry and Patton-Vogt 1998). We asked if the suppression of snf1Δ by sit4 mutations correlates with an increase in INO1 transcription when inositol is removed from the medium. Relative to a wild-type strain, INO1 mRNA levels are reduced in a snf1Δ strain under derepressing conditions (−inositol), as reported previously (Lo et al. 2001; Shirra et al. 2001) (Figure 1B). In agreement with the growth suppression observed on −Ino medium, the sit4-731 mutation restores INO1 mRNA levels to approximately wild-type levels in a snf1Δ strain. A suppressive effect of the sit4Δ mutation on snf1Δ was not observed in this assay, possibly because the growth defect of this double-mutant strain slows the kinetics of INO1 induction (see below). In the strain containing the sit4-731 mutation on its own, INO1 transcript levels are higher than wild-type levels under derepressing conditions, suggesting that Sit4 negatively regulates INO1 mRNA levels even in the presence of a functional Snf1 protein.

We asked whether the transcriptional effects of the snf1 and sit4 mutations occurred through the INO1 promoter by assaying expression of an INO1-lacZ reporter gene integrated in the yeast genome. Consistent with previous reports that Snf1 targets factors involved in transcription initiation (Kuchin et al. 2000; Lo et al. 2001), expression of the INO1-lacZ reporter gene is reduced in the snf1Δ strain relative to the wild-type strain (Figure 1C). Similar to the Northern assay results, the sit4-731 mutation suppresses the expression defect of the snf1Δ mutant. In this experiment, which provides a more sensitive measure of INO1 promoter activity and which was conducted over a 6-hr time course, suppression of snf1Δ by sit4Δ was observed. This assay also revealed that mutations in SIT4 partially derepress INO1 transcription under conditions of high inositol, further suggesting a repressive role for Sit4 in INO1 regulation. The results from these expression studies suggest that Snf1 and Sit4 have opposing effects on transcription initiation from the INO1 promoter.

The sit4-731 gene contains a nonsense mutation and expresses a small amount of full-length protein:

We noted that strains containing the sit4Δ allele grow much more slowly than strains containing the original suppressor allele, sit4-731, even on rich media. As shown in the tetrad dissections in Figure 2A, colonies derived from sit4Δ spores are smaller than those derived from sit4-731 spores even after 3 additional days of incubation at 30°. This observation prompted us to investigate the nature of the sit4-731 mutation. To our surprise, the coding region of the sit4-731 gene contains a single-base-pair change, a G-to-T change at nucleotide 382, which introduces a stop codon at the position normally encoding amino acid 128 (Figure 2B). This mutation is predicted to remove approximately two-thirds of the protein, including a number of amino acids important for catalysis, on the basis of an alignment with the crystal structure of a related phosphatase, rabbit PP1 (Goldberg et al. 1995). To ask whether the first 127 amino acids of Sit4 could constitute a functional protein, we generated a SIT4 derivative in which the DNA sequences downstream of the sit4-731 point mutation had been removed. We termed the product of this derivative Sit4-TT, for Sit4 true truncation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.—

Characterization of the sit4-731 allele. (A, left) Representative tetratype tetrads derived from a cross between a snf1Δ strain (PY130) and a sit4-731 strain (PY859) grown on YPD for 3 days. (A, right) Representative tetratype tetrads derived from sporulation of diploid PY504 grown on YPD for 6 days. Open circles indicate SIT4 SNF1 progeny, circles with a vertical line indicate snf1Δ SIT4 progeny, circles with a horizontal line indicate SNF1 sit4 progeny, and circles with crossed lines indicate snf1 sit4 double-mutant progeny. (B) Diagram of the Sit4-731 and Sit4-TT proteins. sit4-731 contains a G → T mutation that introduces a stop codon at amino acid 128; the box labeled “??” represents the potential readthrough product. The sit4-TT mutation was created by removing the DNA sequences that would encode amino acids 128–311 (see materials and methods). Solid lines indicate the position of residues that are important for metal chelation, and dashed lines indicate the position of additional residues that are important for the active site in the crystal structure of a phosphatase with a similar catalytic domain, rabbit PP-1. (C) Western blot analysis of extracts of PY502 expressing untagged or HA₃-tagged forms of Sit4, Sit4-731, or Sit4-TT from CEN/ARS (C/A) or 2μ plasmids (see materials and methods). Arrows indicate the predicted positions of the 344- or 160-amino-acid HA₃-tagged proteins. (D) Equal numbers of cells in 10-fold serial dilutions were spotted onto YPD media lacking or containing 200 ng/ml rapamycin and allowed to grow for 3 days at 30°. The following strains were examined: PY165, PY133, PY523, PY859, PY538, and PY502.

We first evaluated the expression levels of the Sit4, Sit4-731, and Sit4-TT proteins. For these studies, we generated and expressed in yeast HA-epitope-tagged and untagged versions of Sit4, Sit4-731, and Sit4-TT from low (CEN/ARS=C/A) or high-copy-number (2μ) plasmids and compared protein levels by immunoblot analysis with anti-HA antibody. As expected, none of the extracts derived from transformants expressing untagged proteins produced immunoreactive bands with the mobilities expected of the Sit4 proteins (Figure 2C, odd-numbered lanes). Notably, the level of HA-tagged wild-type Sit4 expressed from a 2μ plasmid was similar to the level expressed from a C/A plasmid, suggesting the presence of a mechanism that limits the expression of Sit4 (Figure 2C, lane 4), as also suggested by Douville et al. (2004). A protein of the expected size for Sit4-TT was only weakly detected when expressed from a 2μ plasmid (Figure 2C, lane 12). Unfortunately, this reduced expression level of Sit4-TT complicated our ability to analyze its properties in vivo. However, a likely explanation for the effects of Sit4-731 emerged from this study. In extracts from cells expressing HA-tagged Sit4-731, a protein comigrating with Sit4-TT was observed, as expected. However, a considerable amount of protein that comigrated with full-length Sit4 was also detected (Figure 2C, lanes 6 and 8). In fact, 2μ plasmids expressing wild-type Sit4 and Sit4-731 produced comparable levels of full-length protein (Figure 2C; compare lanes 4 and 8). Furthermore, expression of Sit4-731 from a 2μ plasmid can complement the Ino⁺ phenotype of a snf1Δ sit4-731 double mutant, suggesting that the full-length protein is functional (data not shown). Analysis of the sequence surrounding the mutation in sit4-731 showed that it contains bases compatible with translational readthrough. In one assay, when the stop codon UAG was followed by a G, as it is in sit4-731, the greatest amount of readthough was detected (Bonetti et al. 1995). In addition, the A found in the −2 position relative to the stop codon is also a strong determinant of stop codon readthrough (Tork et al. 2004). We suggest that the chromosomal sit4-731 gene expresses a small amount of full-length Sit4 by translational readthrough and that this amount of Sit4 is sufficient to support better growth compared to the complete absence of Sit4. However, this full-length protein is not sufficient for normal INO1 regulation.

The sit4-731 mutation does not cause sensitivity to rapamycin:

Strains lacking Sit4 are hypersensitive to the antifungal drug rapamycin (Cutler et al. 2001). To extend our comparison of the sit4Δ and sit4-731 mutations, we asked whether sit4-731 strains are also sensitive to this drug. At a concentration of rapamycin that almost completely prevents growth of the sit4Δ strain, strains containing the sit4-731 mutation grow even better than the wild-type control strain (Figure 2D). This finding confirms that the sit4-731 mutation is distinct from a sit4 null allele and indicates that the truncated or full-length sit4-731 products behave differently from wild-type Sit4 in their ability to confer resistance to rapamycin. Indeed, with respect to rapamycin resistance, the sit4-731 mutation exhibits a dominant phenotype (data not shown). Previously, snf1Δ strains were reported to be more resistant to rapamycin than wild-type strains (Bertram et al. 2002). We did not observe this phenotype, possibly because our wild-type strains are more rapamycin resistant than those reported elsewhere (Heitman et al. 1991), most likely as a result of differences in genetic background or auxotrophic markers (Beck et al. 1999). However, the absence of Snf1 increases resistance to rapamycin in both sit4 mutant strains. Therefore, similar to the genetic relationship we have uncovered for INO1 regulation, Snf1 and Sit4 have opposing effects on growth in the presence of rapamycin.

Strains containing mutations in SIT4 exhibit an Spt⁻ phenotype:

Other genetic studies suggested a role for Sit4 in the regulation of RNA pol II (Arndt et al. 1989). We therefore asked whether sit4 mutant strains exhibit phenotypes associated with general defects in RNA pol II transcription. In particular, we tested whether sit4Δ and sit4-731 mutations confer an Spt⁻ (suppressor of Ty) phenotype. Promoter insertion mutations caused by the integration of the retrotransposon Ty frequently interfere with transcription of the adjacent gene. For example, the presence of a single Ty LTR, or solo δ-element, within the promoter of the HIS4 gene alters transcription initiation and causes a His⁻ phenotype in wild-type strains (Winston and Sudarsanam 1998). Mutations in several classes of generally acting transcription factors suppress the his4-912δ mutation and other Ty insertion mutations and therefore confer an Spt⁻ phenotype (Winston and Sudarsanam 1998). As shown in Figure 3, both the sit4-731 and the sit4Δ mutations suppress the his4-912δ mutation and restore growth on media lacking histidine. In conjunction with our suppression data, these results suggest that sit4 plays a general role in transcriptional regulation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.—

Mutations in SIT4 cause an Spt⁻ phenotype. Equal numbers of cells in 10-fold serial dilutions were spotted onto minimal media containing or lacking histidine and allowed to grow for 5 days at 30°. (Bottom) Diagram of the his4-912δ locus, which contains a Ty δ-element (square with a solid triangle), and the predicted position of the RNA transcripts in Spt⁺ or Spt⁻ strains. The following strains were examined: FY653, PY667, and PY688.

Association of TBP with the INO1 promoter is regulated by Snf1 and Sit4:

Spt⁻ phenotypes have been associated with defects in TBP and proteins required for the recruitment of TBP to promoters (Eisenmann et al. 1989; Dudley et al. 1999; Bhaumik and Green 2001). On the basis of these observations as well as our genetic results that indicated a role for Snf1 in regulating INO1 transcription through an effect on TBP function (Shirra and Arndt 1999; Shirra et al. 2001), we decided to investigate the requirement for Snf1 and Sit4 in the recruitment of TBP to the INO1 promoter. However, since relatively little is known about the molecular mechanism of INO1 activation, we first asked if binding of TBP to the INO1 promoter correlates with gene activation and requires the functions of previously described regulators of INO1 transcription, Opi1, Ino2, and Ino4. Using chromatin immunoprecipitation (ChIP) assays, we found that TBP association with the promoter is not detectable in wild-type strains grown in the presence of inositol (+Ino conditions) and significantly increases upon gene activation (−Ino conditions) (Figure 4, A and B). Compared to the wild-type strain, TBP occupancy is elevated in an opi1Δ strain in the presence or absence of inositol, consistent with the pattern of INO1 transcription in opi1 mutants (White et al. 1991; Ashburner and Lopes 1995; Shirra and Arndt 1999). In contrast, TBP association with the INO1 promoter is strongly dependent on Ino2-Ino4. In the absence of these transactivators, INO1 is not transcribed, and TBP is not detected at the promoter in the presence or absence of inositol. Therefore, through either direct or indirect mechanisms, Opi1 represses and Ino2-Ino4 activates INO1 transcription by regulating TBP association with the promoter.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.—

ChIP analysis of TBP binding to the INO1 promoter. (A and C) Cells were grown to early log phase in 200 μm inositol (+Ino) or washed into medium lacking inositol and allowed to grow an additional 4 hr (−Ino). An anti-TBP antibody was used to immunoprecipitate crosslinked chromatin. Two dilutions of input DNA (I) and two amounts of precipitated DNA (P) were amplified with primers surrounding the INO1 TATA sequence and multiplexed with primers from a subtelomeric sequence of chromosome VI as a control (TELVI). Representative experiments are shown. The following strains were used: PY165, PY528, PY635, PY133, and PY523. (B and D) Quantitation of ChIP analysis. The amount of radiolabeled INO1 DNA amplified by quantitative PCR from input and precipitated samples was normalized to the TELVI levels in each lane. The relative TBP occupancy was determined by dividing the average of the two normalized precipitated samples by the average of the two normalized input samples. The mean relative occupancy from three independent ChIP experiments is presented. Error bars represent one standard deviation from the mean.

Having shown that TBP occupancy correlates with INO1 activation, we next asked if mutations in SNF1 and SIT4 also affect TBP recruitment. As shown in Figure 4, C and D, under inducing conditions promoter association of TBP in snf1Δ strains is reduced relative to wild-type strains. The sit4-731 mutation suppresses this effect to restore wild-type levels of TBP association. These ChIP data are in good agreement with the results of our transcriptional assays and suggest that Snf1 and Sit4 regulate INO1 transcription, at least in part, by modulating TBP recruitment. In addition, these results are consistent with in vitro data showing that recruitment of TBP to preinitiation complexes is reduced in extracts lacking Snf1 (Young et al. 2002). We note, however, that the defect in INO1 activation in snf1Δ cells is slightly greater than the decrease in TBP occupancy at the INO1 promoter. This observation may indicate that Snf1 also regulates INO1 expression at steps that follow preinitiation complex assembly.

Phosphorylation of histone H3 at serine 10 is not required for suppression of snf1Δ by sit4-731:

To study further the mechanism by which sit4-731 suppresses the requirement for SNF1 in INO1 transcription, we investigated possible targets of the Sit4 protein phosphatase. The Snf1 protein kinase has been shown to phosphorylate serine 10 within the amino-terminal tail of histone H3 (S10), and this phosphorylation is associated with an increase in INO1 transcription under derepressing conditions (Lo et al. 2001). One explanation for our results is that Sit4 represses INO1 transcription by dephosphorylating S10 and that otherwise low levels of phosphorylated S10, arising from a kinase other than Snf1, are elevated in the snf1Δ sit4-731 double-mutant strain. We therefore asked if a serine at position 10 in histone H3 was required for suppression of snf1Δ by sit4-731. Strains in which both copies of the histone H3 and H4 genes had been deleted (hht1-hhf1Δ hht2-hhf2Δ) but were kept alive by a plasmid-borne copy of HHT1-HHF1 were transformed with empty vector, a plasmid encoding wild-type histones H3 and H4 (H3-WT), or a plasmid encoding histone H3 mutated to an alanine at position 10 (H3-S10A) and wild-type histone H4. The URA3-marked HHT1-HHF1 plasmid was removed from the transformants by purification on medium containing 5-FOA. The 5-FOA plates were then replica printed to media containing or lacking inositol. We saw no differences in the ability of SNF1 SIT4 or snf1Δ sit4-731 strains to grow in the absence of inositol, regardless of the histone H3 sequence, indicating that phosphorylation of S10 plays little, if any, role in the suppression (Figure 5A). Furthermore, we observed no suppression of the Ino⁻ phenotype of snf1Δ strains by histone H3 derivatives containing aspartic acid or glutamic acid at position 10, substitutions that might mimic the effects of phosphorylation (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.—

Phosphorylation of histone H3 at serine 10 is not required for suppression of snf1Δ by sit4-731. SNF1 SIT4 (FY1716) or snf1Δ sit4-731 (PY636) strains deleted for both copies of the histone H3 and H4 genes, but containing a URA3-marked plasmid with HHT1-HHF1, were transformed with plasmids containing no histone genes (vector; pRS314), wild-type histone H4 and H3 (H3-WT; pWZ414-F13), or wild-type histone H4 and histone H3 with an alanine substitution at serine 10 (H3-S10A; pRS414-59). Transformants selected on SC-Trp medium were purified on medium containing 5-FOA to select against the URA3-marked plasmid. (A) Strains grown on 5-FOA medium were replica plated to synthetic medium lacking inositol or containing 200 μm inositol and allowed to grow for 1 day (SNF1 SIT4) or 2 days (snf1Δ sit4-731) at 30°. Tryptophan was omitted to maintain selection for the histone plasmids. (B) S1 nuclease protection analysis of the strains described in A. Cells were grown to early log phase in medium containing 100 μm inositol (+Ino) or washed into medium lacking inositol and allowed to grow an additional 4 hr (−Ino). INO1 mRNA levels were normalized to DED1 mRNA levels, and the results from a representative experiment are shown.

We also measured the effect of the histone H3-S10A substitution on INO1 transcription in our strain background. To be consistent with earlier studies that examined a role for histone H3-S10 phosphorylation in INO1 activation, we performed S1 nuclease protection assays and used media and conditions similar to those previously described (Lo et al. 2001; Clements et al. 2003). Our results indicate no significant requirement for histone H3-S10 phosphorylation for INO1 transcription in our strains, as similar levels of INO1 transcript were detected in the histone H3-WT and H3-S10A strains independent of the SNF1 and SIT4 mutations (Figure 5B). Identical results were obtained when RNA levels were measured over a time course of induction by S1 or Northern analysis (data not shown). To determine whether the lack of an effect of the S10A substitution was due to a gene conversion event that occurred during plasmid shuffling and replaced the histone H3-S10A gene sequence with wild type, we recovered the plasmids from the 5-FOA^R strains used in our experiments and sequenced the histone H3 genes. This analysis revealed that no sequence alterations had taken place in the construction of the strains. Therefore, the discrepancies between our data and those previously published remain unexplained. However, it is apparent that suppression of snf1Δ by sit4-731 does not require the ability to phosphorylate serine 10 of histone H3.

Analysis of Opi1 as a potential target of Sit4:

Another potential substrate for Sit4 is the phosphoprotein Opi1. Phosphorylation of Opi1 by Pkc1 inhibits its function (Sreenivas et al. 2001), and we postulated that a reduction in Sit4 phosphatase activity might lead to elevated levels of the inactive, phosphorylated form of Opi1, thus allowing INO1 transcription. In addition, we were intrigued by our observations that exogenous expression of Opi1 partially complements sit4-731 (data not shown) and that sit4Δ strains express the INO1-lacZ reporter gene in the presence of inositol (Figure 1C), a phenotype reminiscent of opi1 mutants. If Opi1 is the target of Sit4 relevant to INO1 regulation, sit4Δ might be epistatic to opi1Δ, such that the phenotype of the double mutant would be similar to a single mutant. Using INO1-lacZ expression as a sensitive measure of INO1 promoter activity, we found that the sit4Δ opi1Δ strains exhibited significantly greater levels of β-galactosidase expression than strains containing either single mutation alone (Figure 6).This additive effect on INO1 promoter activity in the absence of both proteins suggests that Sit4 and Opi1 act through different pathways to effect promoter regulation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.—

Opi1 is unlikely to be the only target of Sit4. β-Galactosidase assays of INO1-lacZ expression. Cells were processed and β-galactosidase activity was quantitated as in Figure 1C. The following strains were used: PY177, PY533, PY498, and PY526.

Mutations in SIT4 suppress the effects of rapamycin on INO1 transcription:

Recent work by Loewen et al. (2004) has shown that the binding of Opi1 to PA helps tether Opi1 to the endoplasmic reticulum in media lacking inositol. These interactions inhibit the nuclear import of Opi1, prevent transcriptional repression by Opi1, and allow INO1 transcription. Intriguingly, PA has also been shown to activate the TOR-signaling pathway in mammalian cells, most likely through a direct interaction with the mTOR kinase (Fang et al. 2001). Because we have shown a connection between a component of the TOR-signaling pathway, namely Sit4, and INO1 transcription, we asked whether TOR kinase activity is required for full INO1 expression. We reasoned that the high PA levels present in cells grown in media lacking inositol (Loewen et al. 2004) may activate TOR signaling and thus inactivate Sit4, which we have shown here is a negative regulator of INO1 transcription.

To test whether the TOR pathway stimulates INO1 transcription, we inhibited the yeast TOR kinases with rapamycin and analyzed the effects of this treatment on INO1 induction. Cells were grown in inositol-containing medium (100 μm) to early log phase and subsequently incubated for an additional 30 min in the presence or absence of rapamycin. Cells were then shifted to inducing media (−Ino) with or without rapamycin, and INO1 mRNA was detected by Northern blot analysis. Consistent with our hypothesis that TOR may play a positive role in INO1 transcription, we found that rapamycin reproducibly inhibits INO1 induction in wild-type cells approximately twofold (Figure 7; compare lanes 4 and 8). Mutations in SIT4 relieve the inhibition by rapamycin (compare lanes 16 and 24 to lane 8). This result suggests that TOR may activate INO1 expression by inhibiting Sit4. However, our results also suggest that Sit4 regulates INO1 expression by an additional TOR-independent mechanism. In the presence or absence of rapamycin, the sit4Δ and sit4-731 mutations lead to INO1 transcript levels significantly higher than those observed for wild-type strains grown in the absence of rapamycin. Therefore, the effects of the sit4 mutations appear additive to the effect of simply omitting rapamycin from the media.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.—

The TOR pathway inhibits INO1 induction. Northern blot analysis of INO1 transcription. Cell cultures were grown to early log phase in medium containing 100 μm inositol, and samples (−30 min) were harvested from strains prior to rapamycin treatment. The cultures were divided and incubated with (0 min) or without (data not shown) 200 ng/ml rapamycin for an additional 30 min. Samples were washed into medium lacking inositol, in the presence or absence of rapamycin, and incubated for the indicated times (30, 60, or 90 min). The following strains were examined: PY165, PY859, and PY502.

The results in Figure 7 further highlight distinctions between the sit4Δ and sit4-731 mutations. When grown in the presence of rapamycin, the sit4-731 mutant induces INO1 expression to particularly high levels, a result that correlates with the increased resistance of this strain to rapamycin (see Figure 2D). The sit4Δ cells have derepressed levels of INO1 mRNA when compared to wild-type cells but reduced levels when compared to sit4-731 cells (compare lane 24 with lanes 8 and 16 in Figure 7). Whether this difference between the mutant strains is due to a toxic effect of rapamycin on sit4Δ cells (see Figure 2D) or to altered activity of the sit4-731 gene products remains to be determined. Another possible explanation for the residual INO1 repression in the rapamycin-treated sit4Δ cells or for the enhanced INO1 expression in the rapamycin-treated sit4-731 cells is that TOR-signaling activates INO1 expression by inhibiting Sit4 as well as by affecting the activity of a second, unidentified regulatory factor. Future work will be needed to distinguish between these possibilities.

Differential rapamycin sensitivity of sit4 alleles:

We identified a novel allele of SIT4, sit4-731, by selecting for suppressors of the inositol auxotrophy of a snf1Δ strain. The sit4-731 mutation introduces a premature stop codon in the SIT4 gene and gives rise to both truncated and full-length products, the latter presumably arising from translational readthrough. Unlike sit4Δ cells, which are strongly sensitive to rapamycin, sit4-731 cells exhibit enhanced resistance to this antifungal drug when compared to wild-type cells. While increased resistance to rapamycin was unexpected for a sit4 mutant, it is not unprecedented. A strain containing another sit4 mutation, sit4-51, which is synthetically lethal with defects in proteasomal proteolysis, also exhibits significant drug resistance (Singer et al. 2003). Although our sit4-731 allele is recessive for suppression of snf1Δ and behaves similarly to the sit4Δ allele in other assays, it behaves dominantly to SIT4 with respect to rapamycin resistance (data not shown). Perhaps the truncated Sit4 protein or a mutated form of the full-length protein, generated by nonsense suppression, interferes with transmission of the rapamycin signal or otherwise renders wild-type Sit4 insensitive to the effects of rapamycin.

Convergence of the Snf1 and Sit4 pathways on transcriptional regulation:

Even though sit4-731 and sit4Δ cells exhibit different sensitivities to rapamycin, deletion of SNF1 enhances growth of both sit4 mutant strains in the presence of the drug. This result is consistent with other work that has implicated Snf1 as a negative regulator of TOR-dependent processes (Bertram et al. 2002; Bolster et al. 2002; Kimura et al. 2003). The functions of the TOR and Snf1 kinases were previously shown to converge on the regulation of the transcription factors Gln3 and Msn2. Starvation of cells for either glucose or nitrogen causes nuclear accumulation of Gln3 (Bertram et al. 2002). Snf1, activated by growth in low glucose, promotes nuclear localization of Gln3, while TOR promotes its cytoplasmic localization under nitrogen-rich conditions. The two kinases appear to phosphorylate and regulate the cellular localization of Gln3 independently. Exposure of cells to rapamycin or glucose depletion also leads to the nuclear accumulation of Msn2 (Beck and Hall 1999). Under these conditions, activation of Snf1 by removal of its negative regulator, Reg1, prevents nuclear localization of Msn2 (Mayordomo et al. 2002). Therefore, the TOR and Snf1 kinases can coordinately regulate transcription by controlling the cellular localization of transcription factors, albeit in different ways. Interestingly, Sit4 regulates the cellular localization of Gln3 but not Msn2 (Beck and Hall 1999).

In this study, we show that treatment of cells with rapamycin reduces expression of the INO1 gene and that mutations in SIT4 alleviate this effect. These findings suggest that one function of TOR may be to inhibit the repressing effect of Sit4 on INO1 transcription. The current data do not distinguish between models in which the TOR and Snf1 pathways modulate the activity of the same or different targets that are important for INO1 transcription. Recent reports indicate that the cellular localization of Opi1 changes upon INO1 induction (Loewen et al. 2004), while Ino2 and Ino4 appear to reside constitutively in the nucleus (Brickner and Walter 2004). Whether the TOR or Snf1 pathways impact the localization of these or any other regulators of INO1 transcription remains to be determined. However, it is important to note that sit4 mutations have a greater effect on INO1 expression than would be predicted if Sit4 were acting solely through the TOR pathway (Figure 7). Sit4 appears to work through multiple pathways to repress INO1 transcription.

Sit4 as a regulator of RNA pol II transcription:

The results of our ChIP assays indicate that Sit4 and Snf1 govern promoter association of TBP in a manner consistent with their effects on INO1 transcription. Deletion of SNF1 reduces INO1 transcription and reduces TBP occupancy at the promoter. The sit4-731 mutation suppresses both of these effects. We have also shown that two different sit4 mutations, sit4-731 and sit4Δ, cause Spt⁻ phenotypes, further supporting a connection between Sit4 and the RNA pol II general transcription machinery. Previous reports suggest a potentially broad role for Sit4 in transcriptional regulation. A sit4 mutation was first identified in a screen for negative regulators of HIS4 transcription (Arndt et al. 1989), and cell cycle defects of sit4 cells are associated with reduced expression of CLN1 and CLN2 (Fernandez-Sarabia et al. 1992). Furthermore, Sit4 interacts genetically and, indirectly, physically with RNA pol II. Certain sit4 mutations cause lethality when combined with mutations in RPB1 or RPB2 (Arndt et al. 1989). More recently, Sit4 has been shown to dephosphorylate a subunit of Elongator (Jablonowski et al. 2004), a histone acetyltransferase complex that associates with elongating RNA pol II (Gilbert et al. 2004). Like Elongator mutants, sit4Δ strains are sensitive to the base analog 6-azauracil, a phenotype often associated with defects in transcription elongation (Jablonowski et al. 2001). However, elp3 mutants are weak inositol auxotrophs (Wittschieben et al. 2000), whereas we find that Sit4 is a negative regulator of INO1. Importantly, while our results strongly support a role for Sit4 in regulating preinitiation complex assembly, they do not rule out the possibility that Sit4 also affects INO1 expression at the level of transcript elongation.

Potential targets for Snf1 and Sit4 that are relevant to INO1 regulation:

The identity of the direct target(s) of the Sit4 protein phosphatase that is important for INO1 transcription remains elusive. In prior work, we uncovered a role for the Snf1 substrate Acc1, acetyl coenzyme A carboxylase, in INO1 transcription (Shirra et al. 2001). We found that reduced Acc1 activity could suppress the inositol auxotrophy of snf1Δ cells. Curiously, two recent studies found Sit4 and Acc1 in the same protein complexes in high-throughput affinity purification analyses (Gavin et al. 2002; Ho et al. 2002). This physical association between Sit4 and Acc1 suggests the possibility that the sit4-731 mutation suppresses the snf1Δ mutation by lowering Acc1 activity in a manner similar to ACC1 mutations (Shirra et al. 2001). However, on rich media the sit4-731 mutation does not enhance the growth defect associated with conditional depletion of the essential Acc1 protein, arguing against an additive reduction in Acc1 activity (M. Shirra and K. Arndt, unpublished observations). Whether Sit4 regulates the expression of specific genes, such as INO1, by influencing Acc1 activity remains to be determined.

Another Snf1 target, the serine at position 10 (S10) of histone H3, which was shown previously to contribute to INO1 regulation (Lo et al. 2001), is also not involved in the suppression of snf1Δ by sit4-731. In addition, our results suggest that phosphorylation of histone H3-S10 is not a major contributor to INO1 activation, even in our wild-type strains. The basis for the discrepancy between our results and those previously published (Lo et al. 2001) is not clear, but may relate to slight differences in strain backgrounds or procedures used for cell growth and induction. However, induction of another gene regulated by the Snf1 kinase, SUC2, has also been found to be unaffected by the histone H3 S10A substitution (Geng and Laurent 2004).

Finally, we investigated Opi1 as a potential Sit4 target. PA levels can regulate both TOR kinase and Opi1 repressor activity (Fang et al. 2001; Loewen et al. 2004). Because the activity of Opi1 can be inhibited by phosphorylation (Sreenivas et al. 2001), we asked if Sit4 worked solely through Opi1, in which case, the sit4Δ opi1Δ double mutants should derepress INO1 transcription to the same degree as the opi1Δ single mutant. Instead, we found that deletion of SIT4 resulted in additional derepression of the INO1-lacZ reporter gene compared to the effect of deleting OPI1 alone. This result is consistent with previous reports of Opi1-independent repression of INO1 (Graves and Henry 2000). Furthermore, we did not expect that Opi1 would be the only target of Sit4 because Sit4 influences the transcription of genes that are not known to be regulated by Opi1 (Arndt et al. 1989; Fernandez-Sarabia et al. 1992). We hypothesize that Sit4 may regulate the phosphorylation state of a protein with a more general role in RNA pol II transcription or chromatin function.

Relevance of our studies in yeast to understanding cell growth in mammalian cells:

Elucidating the functions of Snf1 and Sit4 is important to understanding not only how yeast cells respond to different environments but also how human cell proliferation is controlled. In human cells, the most closely related homolog to Snf1 is the AMP-activated protein kinase, AMPK, which has been termed the “fuel gauge” of the cell (Hardie and Carling 1997). In addition to monitoring the energy status of the cell through the AMP-to-ATP ratio, AMPK can be inhibited by leptin, a hormone that plays a role in appetite suppression (Andersson et al. 2004; Minokoshi et al. 2004). Three upstream kinases, Elm1, Pak1, and Tos3, have overlapping functions in phosphorylating and activating Snf1 (Hong et al. 2003; Sutherland et al. 2003). In mammalian cells, a kinase closely related to these Snf1-activating kinases is LKB1, which has been shown to be the major activator of AMPK (Woods et al. 2003; Lizcano et al. 2004; Shaw et al. 2004). A mutation in LKB1 has been linked to Peutz-Jeghers cancer susceptibility syndrome (reviewed in Yoo et al. 2002; Boudeau et al. 2003), suggesting a connection between cell metabolism and cell proliferation in cancer via AMPK (Kyriakis 2003). Therefore, identification of the functions and targets of Snf1 in yeast may provide new insights into the regulation of AMPK and its roles in human cell proliferation.