Western blotting.

Yeast cells were inoculated to an A₆₀₀ of 0.1 and grown in YPD medium to an A₆₀₀ of 1.0. Four A₆₀₀ units were harvested and immediately boiled in SDS sample buffer. Denatured proteins were separated on a 15% denaturing polyacrylamide gel and Western blotting with anti-histone H3 polyclonal antibody (ab1791; Abcam) at a dilution of 1:1000, anti-acetyl histone H3 (Lys14) polyclonal antibody (acH3K14; 07-353, Upstate Biotechnology) at a dilution of 1:500, and anti-hyperacetylated histone H4 polyclonal antibody (acH4K5,8,12,16; 06-946; Upstate Biotechnology) at a dilution of 1:1,000 was carried out as described previously (32). Western blot assays were quantified with NIH ImageJ software (W. S. Rasband, ImageJ, U.S. National Institutes of Health, http://imagej.nih.gov/ij/).

Real-time RT-PCR.

Real-time reverse transcription (RT)-PCR with the following primers was performed as described previously (32): ACT1, 5′-TATGTGTAAAGCCGGTTTTGC-3′ and 5′-GACAATACCGTGTTCAATTGGG-3′; SUC2, 5′-TCAATGACAAACGAAACTAGCG-3′ and 5′-GGCATCTTTTTCATCGTACCAC-3′; GAL4, 5′-ACTGTCTTCTATCGAACAAGCATG-3′ and 5′-TCCCAGTTGTTCTTCAGACACTT-3′; REG2, 5′-ATGTGAGAAGCTCGAACAACACT-3′ and 5′-TCTCCAAAAAAGGTTCTCCAATC-3′; MAL33, 5′-CAAGTATGCATGCGACTATTGTC-3′ and 5′-GTTGCTGATATGTGCAATCGAA-3′; ACS1, 5′-CTCTGCCGTACAATCATCAAAAC-3′ and 5′-CCGAAGTCAAATGTTCATACTCAT-3′; CAT8, 5′-TCATTAGAACTCTGGGTTCACAAG-3′ and 5′-GTCTTGATCATTGCTTCCTTCAC-3′; RPS11B, 5′-TGTCCACTGAATTAACTGTTCAATC-3′ and 5′-CATCTCTTGGTTCTCTTGGAAGTC-3′; RPS22B, 5′-AGCTGATGCTTTGAATGCCA-3′ and 5′-TTCGCCAATGTAACCATGCT-3′; HHT1/2, 5′-GAAGCCTCACAGATATAAGCCAG-3′ and 5′-ATCTTGAGCGATTTCTCTGACC-3′; HHF1/2, 5′-CCAAGCGTCACAGAAAGATTCTA-3′ and 5′-ACCAGAAATACGCTTGACACCA-3′; HTA1/2, 5′-CGGTGGTAAAGGTGGTAAAGC-3′ and 5′-TGGAGCACCAGAACCAATTC-3′; HTB1/2, 5′-CAAAGTTTTGAAGCAAACTCACCC-3′ and 5′-GCCAATTTAGAAGCTTCAGTAGC-3′). Primer pairs for the canonical histones were designed so that they measure the expression of both genes for that particular histone (HTA1 and HTA2, HTB1 and HTB2, HHT1 and HHT2, and HHF1 and HHF2).

ChIP assays.

In vivo chromatin cross-linking and immunoprecipitation were performed essentially as described previously (32). Immunoprecipitation was performed with the following antibodies: for anti-histone H3 (Abcam ab1791), anti-acetyl-histone H3 (Lys14) (Upstate Cell Signaling Solutions 07-353), and anti-hyperacetylated histone H4 (Penta) (Upstate Cell Signaling Solutions 06-946). The primers used for real-time PCR are as follows: POL1, 5′-TCCTGACAAAGAAGGCAATAGAAG-3′ and 5′-TAAAACACCCTGATCCACCTCTG-3′; INO1, 5′-ACGAGCTGCTCACCAAGTACAG-3′ and 5′-TGAACACGTAGTCTTGAACAGTGG-3′; UBC8, 5′-TTGGAGGTTACATGTAGAGTTGCC-3′ and 5′-CCCGATGCGATATCAATGTTAG-3′; INT6, 5′-GGAAGCGAAATTCAAGGGTTG-3′ and 5′-TTGCATGATTGTGCATACGC-3′; SUC2, 5′-CGACTTTTTTTTTTTGGATTTCG-3′ and 5′-GCACAAGAACAAGAGAATGTTTTG-3′; REG2, 5′-ATGTGAGAAGCTCGAACAACACT-3′ and 5′-TCTCCAAAAAAGGTTCTCCAATC-3′; HXK1, 5′-CATTTAGGTCCAAAGAAACCACA-3′ and 5′-GGTCTCGCTGTCAACTGTAAACA-3′; GAL4, 5′-ACTGTCTTCTATCGAACAAGCATG-3′ and 5′-TCCCAGTTGTTCTTCAGACACTT-3′; MAL31, 5′-AAAAAAGACAGGAACGACTCACAC-3′ and 5′-ACTTTTCTTACCTTGCTCCTCCA-3′; MAL33, 5′-CAAGTATGCATGCGACTATTGTC-3′ and 5′-GTTGCTGATATGTGCAATCGAA-3′; HHT2, 5′-CTTCTTGTGACCGCAGTTGTATAT-3′ and 5′-AAAGGGGAAGAACAGTTGGAAG-3′; HHF2, 5′-GGCTGTTGTTATTCGGCTAGAT-3′ and 5′-CTTTCACCGCCTCATCCTAATAT-3′; ACT1, 5′-CTCTTGTATTCTTCCTTCCCCT TTC-3′ and 5′-ATGGTGCAAGCGCTAGAACATAC-3′; ADH1, 5′-AATCCCACGGTAAGTTGGAATAC-3′ and 5′-AAGCGTGCAAGTCAGTGTGAC-3′; RPS11B, 5′-CTTTACGTTTCCCTTCTCTGCTG-3′ and 5′-TTTCCCTGGCTTGATACGTTTC-3′), RPS22B, 5′-GCCCATGTGTTGGAGGGAAGG-3′ and 5′-ATCAGCTAAAACGGAAGAGCGAG-3′; MDN1, 5′-GTACCTCTGAGTTAATTTGACATCG-3′ and 5′-GGACATTCTCAGCAGATTATAACGG-3′; RNY1, 5′-CGGGAATCATGAGTCCTTATG-3′ and 5′-AGCACCCCACTCAGTATAGCA-3′).

Acetyl-CoA and malonyl-CoA assays.

Cells were grown in YPD medium to an A₆₀₀ of 1.0. Sodium azide was added to a final concentration of 10 mM, and 3 × 10⁸ cells were harvested by centrifugation and lysed in 200 μl of 10% perchloric acid with prechilled glass beads. The lysate was neutralized with 10 M KOH to pH 7.5. Acetyl-CoA and malonyl-CoA were assayed by enzyme-linked immunosorbent assay (ELISA) (Antibodies-online GmbH; ABIN457115 and ABIN366452).

Immunoprecipitation.

Immunoprecipitation of acetylated proteins was performed as described previously (32).

snf1Δ mutant cells display globally decreased untargeted acetylation of chromatin histones.

To test whether SNF1 regulates histone acetylation globally or only at specific loci, we used chromatin immunoprecipitation (ChIP) to evaluate the occupancy of histone H3 acetylated at lysine 14 (acH3K14), as well as hyperacetylated histone H4 (acH4K5,8,12,16), in the promoter regions of SNF1-dependent genes (Fig. 2A) and SNF1-independent loci (Fig. 2B). The SNF1-dependent genes included INO1, UBC8, SUC2, REG2, HXK1, GAL4, MAL31, MAL33, HHT2, and HHF2. The SNF1-independent loci included promoters of ACT1, ADH1, RPS11B, MDN1, and RNY1 and the intergenic region in chromosome VI (INT6; coordinates 261412 to 261538). Expression of INO1 and UBC8 is upregulated in the rpd3Δ mutant and downregulated in the snf1Δ mutant (44, 45). SUC2, REG2, HXK1, GAL4, MAL31, and MAL33 are repressed by glucose in a Mig1p/Mig2p-dependent manner (46). The intergenic region was selected since it is 11 kb distant from the nearest coding region. We used anti-H3 antibody, which recognizes the C-terminal region of H3, which is not posttranslationally modified. The ChIP signal obtained with this antibody thus represents the total H3 occupancy and can be used to calculate the histone acetylation levels per nucleosome content (47). To account for differences in nucleosome density at different genomic loci, we corrected the acH3 and acH4 occupancies for histone H3 content and generated values that represent acetylation per nucleosome (Fig. 2). The acetylation of histones H3 and H4 per nucleosome in promoters of SNF1-dependent genes was 1.2 to 1.9 and 1.3 to 2.1 times lower in snf1Δ mutant cells than in wild-type cells, respectively. The histone H3 and H4 acetylation in snf1Δ mutant cells was increased 1.1- to 1.9- and 1.1- to 2.1-fold by the introduction of the tetO₇-ACC1 allele, respectively (Fig. 2A). The acetylation of histones H3 and H4 per nucleosome in promoters of SNF1-independent loci was 1.2 to 1.9 and 1.4 to 1.8 times lower in snf1Δ mutant cells than in wild-type cells, respectively. The histone H3 and H4 acetylation in promoters of SNF1-independent genes of snf1Δ mutant cells was increased 1.1- to 2.8- and 1.2- to 1.8-fold by introducing the tetO₇-ACC1 allele, respectively (Fig. 2B).

Open in a separate window

Fig 2

snf1Δ mutant cells display decreased untargeted acetylation of chromatin histones. The indicated strains were grown at 30°C in YPD medium to an A₆₀₀ of 1.0. ChIP experiments were performed with antibodies against total histone H3 (H3), histone H3 acetylated at lysine 14 (acH3), and hyperacetylated histone H4 (acH4). Occupancies of H3, acH3, and acH4 were determined in the promoter regions of SNF1-dependent genes (A) and SNF1-independent loci (B). Acetylation per nucleosome was calculated as ratios of acH3 to total H3 and acH4 to total H3. The experiments were repeated three times, and the results are shown as means ± standard deviations. Values that are statistically significantly different (P < 0.05) from the wild-type (WT) values are indicated by an asterisk. Values that are statistically significantly different (P < 0.05) from each other are indicated by a bracket and an asterisk.

To test whether the decreased acetylation of chromatin histones found in promoters and intergenic regions of snf1Δ mutant cells is due to altered H3S10 phosphorylation, we used ChIP to evaluate the occupancy of acetylated histones also in the H3S10A mutant. The acetylation of histones H3 and H4 per nucleosome in promoters and intergenic regions differed less than 3% and 6% between H3S10A mutant and wild-type cells, respectively (Fig. 3A and ​andB).B). This indicates that phosphorylation of histone H3 at serine 10 is not responsible for the decreased acetylation of histones H3 and H4 at the tested loci in snf1Δ mutant cells. The decreased acetylation of both histones H3 and H4 in snf1Δ mutant cells and suppression by the tetO₇-ACC1 allele at all SNF1-dependent loci, as well as SNF1-independent loci, suggest that inactivation of SNF1 decreases the acetylation of chromatin histones in a global untargeted manner and are in agreement with the decreased acetylation of bulk histones measured by Western blot analysis (Fig. 1C).

Open in a separate window

Fig 3

Phosphorylation of histone H3 at serine 10 does not affect the acetylation of chromatin histones. The indicated strains were grown at 30°C in YPD medium to an A₆₀₀ of 1.0. ChIP experiments were performed with antibodies against total histone H3 (H3), histone H3 acetylated at lysine 14 (acH3), and hyperacetylated histone H4 (acH4). Occupancies of H3, acH3, and acH4 were determined in the promoter regions of SNF1-dependent genes (A) and SNF1-independent loci (B). Acetylation per nucleosome was calculated as ratios of AcH3 to total H3 and acH4 to total H3. The experiments were repeated three times, and the results are shown as means ± standard deviations. The acetylation per nucleosome in wild-type (WT) and H3S10A cells was not statistically significantly different at any of the loci tested.

Inactivation of SNF1 leads to hypoacetylation of nonhistone proteins.

In addition to histones, acetylation has been shown to regulate other proteins (38, 48, 49). To test whether the decreased cellular level of acetyl-CoA in snf1Δ mutant cells affects the acetylation of nonhistone proteins, we assayed the acetylation of Sip2p and Swi4p by immunoprecipitation with anti-pan-acetyl-lysine antibody (48). Sip2p is one of the regulatory β subunits of the SNF1 complex. Sip2p is acetylated by NuA4 (49), and the elevated level of acetyl-CoA in tetO₇-ACC1 cells results in increased Sip2p acetylation (32). Swi4p binds Swi6p to form transcriptional complex SBF (50). The acetylation of Swi4p regulates the interaction with Swi6p and thus the SBF-dependent gene transcription (38). Our results show that the acetylation of both Sip2p and Swi4p is decreased in snf1Δ mutant cells in comparison with wild-type cells (Fig. 4) and suggests that in addition to histones, SNF1 is important also for the normal acetylation of nonhistone proteins.

Open in a separate window

Fig 4

snf1Δ mutant cells display reduced acetylation of nonchromatin proteins. Wild-type (WT) and snf1Δ mutant cells were grown in YPD medium at 30°C to an A₆₀₀ of 1.0. Acetylated proteins were immunoprecipitated (IP) with anti-acetyllysine antibody or control antibody (normal mouse IgG) and analyzed by Western blotting with antibodies against the tandem affinity purification (TAP) tag. Typical results from three independent experiments are shown. The intensity of each immunoprecipitated band was quantified by densitometry and expressed as a percentage of the wild-type value ± the standard deviation. The asterisks indicate that the values were statistically significantly different (P < 0.05) from the wild-type value.

SNF1 genetically interacts with HATs and HDACs.

If SNF1 promotes histone acetylation and plays a role in establishing the global level of histone acetylation, then inactivation of histone acetyltransferases (HATs) in snf1Δ mutant cells should result in a synthetic growth defect, while inactivation of histone deacetylases (HDACs) should improve the growth rate and fitness of snf1Δ mutant cells. To test this hypothesis, we introduced the yng2Δ, gcn5Δ, hda1Δ, and rpd3Δ mutations into snf1Δ mutant cells. Yng2p is a subunit of NuA4 and piccolo NuA4 complexes and regulates the HAT activity of Esa1p. Gcn5p is the catalytic subunit of the related HATs SAGA, ADA, and SLIK. Rpd3p is the catalytic subunit of the class I Sin3/Rpd3 HDAC, and Hda1p is the catalytic subunit of the class II HDA1 HDAC complex. SAGA and NuA4 are the major HATs in yeast, responsible for most of the HAT activity, while the Sin3/Rpd3 and Hda1 complexes account for most of the HDAC activity (51). Introduction of the hda1Δ or rpd3Δ mutation into snf1Δ mutant cells did not affect the growth rate at 30°C; however, it partially suppressed their temperature sensitivity (Fig. 5A). On the other hand, both the yng2Δ and gcn5Δ mutations produced a negative synthetic phenotype with the snf1Δ mutation (Fig. 5B). The genetic interactions between the snf1Δ mutation and mutations in the major yeast HDACs and HATs thus indicate that SNF1 is involved in the control of the global level of histone acetylation.

Open in a separate window

Fig 5

The snf1Δ mutation displays genetic interactions with mutations in HATs and HDACs. (A) Inactivation of HDA1 or RPD3 partially suppresses the temperature sensitivity of snf1Δ mutant cells. Tenfold serial dilutions of the indicated strains were spotted onto YPD plates and grown for 3 days at 30 or 35°C. (B) Strains with YNG2 or GCN5 deleted display a synthetic growth defect with the snf1Δ mutation. Tenfold serial dilutions of the indicated strains were spotted onto YPD plates and grown for 3 days at 30°C. Typical results from three independent experiments are shown. WT, wild type.

Reduced ACC1 expression suppresses the stress sensitivity of snf1Δ mutant cells.

In addition to its importance in the regulation of metabolism, SNF1 was also implicated in responses to a variety of stresses (52–55). The cellular response to many stresses requires rapid reprogramming of the transcriptional machinery (56, 57). Since active transcription generally correlates with increased acetylation of promoter histones (58, 59), we reasoned that a reduced supply of acetyl-CoA and globally decreased histone acetylation in snf1Δ mutant cells would interfere with the appropriate stress response. To test whether the histone hypoacetylation of snf1Δ mutant cells is partially responsible for stress sensitivity, we measured the growth rates of the snf1Δ and snf1Δ tetO₇-ACC1 mutant strains under different stress conditions. The snf1Δ mutant grows somewhat slower than the wild type at 30°C but is not able to grow at 35°C. The introduction of the tetO₇-ACC1 allele into snf1Δ mutant cells improves their growth rate at 30°C only slightly but clearly suppresses the growth defect of snf1Δ mutant cells at 35°C, at pH 7.5, or in the presence of 50 mM hydroxyurea (Fig. 6). This suppression was abolished by the introduction of the gcn5Δ or yng2Δ mutation, indicating that the activities of SAGA and NuA4 are required for suppression of the stress sensitivity of snf1Δ mutant cells by the tetO₇-ACC1 allele. The tetO₇-ACC1 allele in snf1Δ mutant cells increased the acetyl-CoA level above the wild-type level (Fig. 1A) but failed to completely suppress the growth defect of snf1Δ mutant cells (Fig. 6). This result was expected, since SNF1 phosphorylates and regulates a number of other proteins. Altogether, the results suggest that the sensitivity of snf1Δ mutant cells to different stresses is suppressed by the increased level of acetyl-CoA brought about by the reduced activity of Acc1p and is not caused by reduced H3S10 phosphorylation (Fig. 6).

Open in a separate window

Fig 6

Reduced ACC1 expression partially suppresses the stress sensitivity of snf1Δ mutant cells in a HAT-dependent manner. Tenfold serial dilutions of the indicated strains were spotted onto YPD plates and grown for 2 days at 30 or 35°C, grown for 6 days on YPD plates at pH 7.5, or grown for 3 days on YPD plates containing 50 mM hydroxyurea. Typical results from three independent experiments are shown. WT, wild type.

Derepression of ACS1 elevates the acetyl-CoA level and histone acetylation and suppresses the snf1Δ mutation.

One of the key substrates of SNF1 is the repressor Mig1p. To explore whether the hda1Δ or rpd3Δ mutation would synergize with the mig1Δ mutation to suppress snf1Δ, we constructed the corresponding strains and measured their growth rates. Introduction of the mig1Δ mutation partially suppressed the growth defect of snf1Δ mutant cells; however, introduction of the hda1Δ or rpd3Δ mutation into a snf1Δ mig1Δ mutant strain did not provide additional suppression under any of the conditions tested (Fig. 7A). This unexpected result could be explained by assuming that mig1Δ suppresses snf1Δ by increasing the level of acetyl-CoA. We hypothesized that this could occur by derepression of ACS1. ACS1 and ACS2 encode two yeast acetyl-CoA synthetases that catalyze the ligation of acetate and CoA to yield acetyl-CoA. On glucose-containing laboratory media, ACS2 is essential and ACS1 is repressed (28, 60). Upon glucose exhaustion, the expression of ACS1 is activated by transcriptional factor Cat8p (61). In the presence of glucose, the expression of CAT8 is repressed by Mig1p/Mig2p (46). We hypothesized that the deletion of MIG1 would increase the expression of CAT8 and consequently ACS1 (Fig. 7B). In this model, the suppression of snf1Δ by mig1Δ would require ACS1. To test this possibility, we introduced the acs1Δ mutation into the snf1Δ mig1Δ mutant strain and measured its growth rate. Indeed, the full suppression of snf1Δ by mig1Δ required ACS1 (Fig. 7C). This observation is in agreement with the increased expression of the CAT8 and ACS1 genes in snf1Δ mig1Δ mutant cells in comparison with snf1Δ mutant cells (Fig. 7D). The K_m of acetyl-CoA synthetase 1 (Acs1p) for acetate is about 30-fold lower than that of Acs2p (60). Therefore, even a small increase in ACS1 expression would be expected to have a significant impact on the cellular level of acetyl-CoA. Measurement of acetyl-CoA levels showed that introduction of the mig1Δ mutation into snf1Δ mutant cells results in an increase in the acetyl-CoA level, which depends on ACS1 (Fig. 7E). These results are consistent with the proposed model of the role of Mig1p in the regulation of ACS1 expression and acetyl-CoA homeostasis (Fig. 7). In wild-type cells grown in the presence of glucose, ACS1 is repressed and therefore the acs1Δ mutation does not affect the cellular level of acetyl-CoA. However, ACS1 is derepressed by the introduction of the mig1Δ mutation (Fig. 7D) and Acs1p contributes to the cellular pool of acetyl-CoA (compare wild-type and mig1Δ mutant cells in Fig. 7E). Inactivation of ACS1 thus affects the acetyl-CoA level in mig1Δ mutant cells but not in wild-type cells. These results are consistent with the acetylation of bulk histones. In snf1Δ mutant cells, histone hypoacetylation was partially suppressed by the mig1Δ mutation, and this suppression was abrogated by the acs1Δ mutation (Fig. 7F). Cumulatively, these results provide another piece of evidence that many of the snf1Δ mutant phenotypes can be suppressed by elevating the cellular level of acetyl-CoA.

Open in a separate window

Fig 7

Inactivation of MIG1 requires ACS1 to partially suppress the stress sensitivity and histone hypoacetylation of snf1Δ mutant cells. (A) The hda1Δ or rpd3Δ mutation does not synergize with the mig1Δ mutation in the suppression of the snf1Δ mutation. Tenfold serial dilutions of the indicated strains were spotted onto YPD plates and grown for 2 days at 30 or 35°C, grown for 6 days on YPD plates at pH 7.5, or grown for 3 days on YPD plates containing 50 mM hydroxyurea. Typical results from three independent experiments are shown. (B) Diagram indicating regulation of Acs1p by Mig1p and Cat8p. (C) Suppression of the snf1Δ mutant's slow-growth phenotype by the mig1Δ mutation requires ACS1. Tenfold serial dilutions of the indicated strains were spotted onto YPD plates and grown for 2 days at 30°C. Typical results from three independent experiments are shown. (D) Inactivation of MIG1 increases the expression of CAT8 and ACS1 in snf1Δ mutant cells. The indicated strains were grown in YPD medium to an A₆₀₀ of 1.0 at 30°C. Total RNA was isolated and assayed for CAT8 and ACS1 transcripts by real-time RT-PCR. The results were normalized to ACT1 RNA and expressed relative to the value for the wild-type (WT) strain. (E) Inactivation of MIG1 increases acetyl-CoA levels in snf1Δ mutant cells in an ACS1-dependent manner. Acetyl-CoA was assayed in cell lysates of the indicated strains by ELISA. The concentration of acetyl-CoA in wild-type cells is 1.8 nmol/10⁷ cells. (F) Inactivation of MIG1 increases the acetylation of histones in snf1Δ mutant cells in an ACS1-dependent manner. The indicated strains were grown in YPD medium to an A₆₀₀ of 1.0 at 30°C and analyzed by Western blotting with antibodies against histone H3 acetylated at lysine 14 (acH3), hyperacetylated histone H4 (acH4), and total histone H3. The experiment was performed three times, and representative results are shown. (D and E) The experiments were repeated three times, and the results are shown as means ± standard deviations. Values that are statistically significantly different (P < 0.05) from the wild-type values are indicated by an asterisk. Values that are statistically significantly different (P < 0.05) from each other are indicated by a bracket and an asterisk.

Acetyl-CoA homeostasis contributes to the regulation of SNF1-dependent genes.

To test the contribution of acetyl-CoA homeostasis and global histone acetylation to the transcriptional regulation of SNF1-dependent genes, we used published data sets from DNA microarray experiments and compared set of genes upregulated upon the inactivation of SNF1 (62) with set of genes upregulated in a strain expressing histone H4 with mutations that change the acetylatable lysine residues into nonacetylatable arginine residues within the N-terminal tail (H4-K5R/K8R/K12R/K16R) (63). The first data set was generated with cells that express the snf1^as allele instead of wild-type SNF1. The snf1^as allele carries a mutation that renders mutant cells sensitive to the inhibitor mMe-PP1 (62). snf1^as mutant cells were grown in SC medium with 3% glycerol and harvested 60 min after the addition of mMe-PP1 (62). We compared the set of genes upregulated at least 2-fold in this experiment with the set of genes upregulated at least 2-fold in histone H4 mutant H4-K5R/K8R/K12R/K16R (63) and found a significant overlap (Fig. 8A). More than 45% of the genes upregulated upon snf1^as inactivation are upregulated in the H4-K5R/K8R/K12R/K16R mutant. As indicated by the χ² test (P < 0.001), the correlation between the two data sets is much higher than that expected in an overlap of two random data sets of the same size (Fig. 8A, values in parentheses). We conclude that the mechanism responsible for upregulation of the genes upon the inactivation of snf1^as involves globally decreased histone acetylation. The notion that SNF1 affects histone acetylation is supported by a recent study that showed decreased histone acetylation in the promoters of ADH2, POX1, and ACS1 upon the inhibition of snf1^as (64).

Open in a separate window

Fig 8

Acetyl-CoA homeostasis and histone acetylation contribute to the regulation of SNF1-dependent genes. (A) Inactivation of SNF1 results in the increased expression of genes that are also upregulated in a nonacetylatable histone H4 mutant (H4-K5R/K8R/K12R/K16R). The Venn diagram is based on previously published data sets (62, 63) and shows the extent of overlap between genes upregulated at least 2-fold upon the inactivation of Snf1p in the snf1^as mutant strain and genes upregulated at least 2-fold in the H4-K5R/K8R/K12R/K16R strain. The values in parentheses are theoretical values calculated for the chance overlap of two random data sets of the corresponding sizes. (B to F) The indicated strains were grown in YPD medium to an A₆₀₀ of 1.0 at 30°C. Total RNA was isolated and assayed for ACT1, REG2, GAL4, SUC2, MAL33, HHT1/2, HHF1/2, HTA1/2, and HTB1/2 transcripts by real-time RT-PCR. The results were normalized to ACT1 RNA and expressed relative to the value for the wild-type (WT) strain. (G) The snf1Δ mutant displays a synthetic growth defect with the SWI4-RR mutation. Tenfold serial dilutions of the indicated strains were spotted onto YPD plates and grown for 2 days at 30°C. Typical results from three independent experiments are shown. (B to F) The experiments were repeated three times, and the results are shown as means ± standard deviations. Values that are statistically significantly different (P < 0.05) from the wild-type values are indicated by an asterisk. Values that are statistically significantly different (P < 0.05) from each other are indicated by a bracket and an asterisk.

To test the contribution of acetyl-CoA homeostasis to the transcriptional regulation of genes that require SNF1 for full expression (45, 62), we measured the expression of two groups of genes, glucose-repressed and non-glucose-repressed genes, in wild-type and snf1Δ, tetO₇-ACC1, and snf1Δ tetO₇-ACC1 mutant cells. The genes in the first group include REG2, GAL4, SUC2, and MAL33 and are repressed by glucose in a Mig1p/Mig2p-dependent manner; REG2 and GAL4 are repressed by Mig1p, and SUC2 and MAL33 are repressed redundantly by Mig1p/Mig2p (46). The genes in the second group include the histone genes HHT1/2, HHF1/2, HTA1/2, and HTB1/2. Similarly to INO1, the expression of the histone genes is not repressed by glucose but requires SNF1 (45). The mRNA levels for the individual genes in snf1Δ mutant cells were reduced to 14 to 79% of the wild-type levels, and the defect was partially suppressed by the tetO₇-ACC1 allele (Fig. 8B and ​andC).C). The H3S10 phosphorylation did not affect the expression of the genes tested, since their expression did not differ in wild-type cells and the H3S10A mutant (Fig. 8D and ​andE).E). Similar result was reported previously for SUC2; the H3S10A mutation did not significantly affect the SUC2 mRNA level and histone H3 acetylation in the SUC2 promoter (20). These results suggest that the decreased cellular level of acetyl-CoA and histone hypoacetylation contribute to the altered transcriptional regulation in snf1Δ mutant cells.

Recently, the G₁/S-specific transcriptional factor Swi4p was shown to be acetylated (38). Swi4p binds Swi6p to form transcriptional complex SBF. The acetylation of Swi4p regulates the interaction with Swi6p and thus SBF-dependent gene transcription. The SBF complex is involved in the regulation of transcription of the histone genes (65). Since the acetylation of Swi4p is affected in snf1Δ mutant cells (Fig. 4), we wanted to determine whether decreased acetylation of Swi4p affects the expression of the histone genes. The expression of all four histone genes was reduced in a strain that expresses Swi4p with mutations that change two acetylatable lysine residues into nonacetylatable arginine residues (Swi4p-RR) (38) (Fig. 8F). The Swi4p mutation that changes the two acetylatable lysines into glutamine residues (Swi4p-QQ), which mimics constitutive acetylation, did not affect the expression of the histone genes. Introduction of the snf1Δ mutation into the Swi4p-RR strain resulted in additive decreases in the expression of the histone genes (Fig. 8F). Interestingly, unlike Swi4p-QQ, the Swi4p-RR mutation displayed a synthetic growth defect with the snf1Δ mutation (Fig. 8G). These results suggest that the snf1Δ mutation affects the acetylation of histones, as well as Swi4p, both of which are required for normal expression of the histone genes.

HATs are required to mediate the effect of an increased acetyl-CoA level.

Since the regulation of SUC2 transcription is relatively well understood (20), we used it to analyze in more detail the role of acetyl-CoA homeostasis. While the tetO₇-ACC1 allele does not significantly change the extent or kinetics of SUC2 induction, it allows weak expression in the snf1Δ mutant (Fig. 9A and ​andB).B). The effect is relatively small, but only partial suppression would be expected because the Mig1p-mediated repression is not relieved in the snf1Δ mutant. Since SNF1 is required for normal acetyl-CoA homeostasis and a normal level of untargeted histone acetylation when cells are grown in YPD medium (Fig. 1 and ​and2),2), we analyzed SUC2 regulation under repressing conditions in the mig1Δ mutant background (Fig. 9C). The major known roles of SNF1 in the regulation of SUC2 expression are phosphorylation of Mig1p and relief of the suppression mediated by the Mig1p-Tup1p/Ssn6p complex (11, 14). The snf1Δ mutation reduced SUC2 expression similarly in both the MIG1 (Fig. 8B) and mig1Δ (Fig. 9C) mutant strains. However, while the tetO₇-ACC1 allele suppressed this effect only partially in MIG1 mutant cells (Fig. 8B), the suppression was more robust in mig1Δ mutant cells (compare the expression in mig1Δ mutant cells with the expression in mig1Δ snf1Δ tetO₇-ACC1 cells; Fig. 9C). This result suggests that in the absence of the Mig1p repressor, the defect in SUC2 expression due to the snf1Δ mutation can be completely suppressed by introducing the tetO₇-ACC1 allele and increasing the cellular level of acetyl-CoA and histone acetylation. To test whether HATs are required to mediate the effect of an increased acetyl-CoA concentration, we determined SUC2 expression in the mig1Δ gcn5Δ and mig1Δ yng2Δ mutants. As expected, the gcn5Δ and yng2Δ mutations significantly decreased SUC2 expression in the mig1Δ mutant strain and the effect was additive with the snf1Δ mutation. Interestingly, the tetO₇-ACC1 allele did not significantly increase expression in the mig1Δ gcn5Δ and mig1Δ yng2Δ mutants, suggesting that the effect of the increased cellular level of acetyl-CoA on SUC2 expression requires the activities of SAGA and NuA4 (Fig. 9C). Cumulatively, these results suggest that in addition to relieving the repression mediated by the Mig1p-Tup1p/Ssn6p complex, SNF1 also supports SUC2 expression by maintaining a normal level of histone acetylation.

Open in a separate window

Fig 9

Loss of SUC2 expression in snf1Δ mutant cells is partially suppressed by the tetO₇-ACC1 allele. (A) The indicated strains were grown in YPD medium to an A₆₀₀ of 1.0 at 30°C. The cells were harvested, washed, and resuspended and grown in YEP medium containing 0.05% glucose. Total RNA was isolated and assayed for ACT1 and SUC2 transcripts by real-time RT-PCR. The results were normalized to ACT1 RNA and expressed relative to the value for the wild-type (WT) strain at 0 min. The experiments were repeated three times, and a typical time course is shown. (B) Closeup view of the values for the snf1Δ and snf1Δ tetO₇-ACC1 mutants. (C) The indicated strains were grown in YPD medium to an A₆₀₀ of 1.0 at 30°C. Total RNA was isolated and assayed for ACT1 and SUC2 transcripts by real-time RT-PCR. The results were normalized to ACT1 RNA and expressed relative to the value for the mig1Δ mutant strain. The experiments were repeated three times, and the results are shown as means ± standard deviations. Values that are statistically significantly different (P < 0.05) from those of the mig1Δ mutant strain are indicated by an asterisk, and values that are statistically significantly different (P < 0.05) from each other are indicated by a bracket and an asterisk.

We thank B. Andrews, K. M. Arndt, M. Carlson, M. H. Kuo, L. F. Pemberton, M. C. Schmidt, D. J. Stillman, T. Tsukiyama, F. Winston, and H. Zhu for strains and plasmids and members of the Vancura lab and I. Vancurova for helpful comments.

This work was supported by grant GM106324 from the National Institutes of Health to A.V.