Evidence that the Elongation Factor TFIIS Plays a Role in Transcription Initiation at GAL1 in Saccharomyces cerevisiae

Yeast strains and plasmids.All S. cerevisiae strains (Table 1) are isogenic to a GAL2 ⁺ derivative of S288C (55). To epitope tag TFIIS, three copies of a sequence encoding the Myc epitope were integrated at the 5′ end of the DST1 gene (encoding TFIIS) (43). The 3×Myc-TFIIS fusion was fully functional in vivo, as assayed by several phenotypes, including growth on medium containing galactose, caffeine, or mycophenolic acid. The carboxy-terminally tagged TFIIS-9×Myc allele described previously (40) has impaired function, as indicated by the inability of the fusion protein to grow in a manner similar to that of the wild-type protein on medium containing galactose (at 37°C) or mycophenolic acid (data not shown); this finding is likely to explain the differences between earlier data and data from this work. The dst1Δ::KANMX allele was constructed by one-step PCR-mediated disruption, which replaced the entire ORF with the KANMX4 cassette (7). Plasmid SGP4, containing the three consensus Gal4-binding sites, was described previously (5).

ChIP.All ChIP experiments were carried out as previously described (27). Antibodies used in these experiments included mouse monoclonal antibody 12CA5 against the hemagglutinin (HA) epitope (1 μl for Spt20 and 0.5 μl for Rpb3), anti-Med19 rabbit polyclonal antibody (1 μl; gift from R. Kornberg), anti-Myc rabbit polyclonal antibody A14 (5 μl; Santa Cruz), anti-Gal4 DNA-binding domain antibody (1 μl; Santa Cruz), anti-TBP antibody (5 μl) (22), anti-Spt16 antibody (2 μl; gift from T. Formosa), anti-TFIIS antibody (1 μl; gift from C. Kane), anti-Rpb3 antibody (1 μl; NeoClone Biotechnology), and monoclonal antibody 8WG16 against the C-terminal domain of Rpb1 (2 μl; Covance).

Immunoprecipitations were carried out with either FA lysis buffer (150 mM NaCl, 50 mM HEPES-KOH [pH 7.5], 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, 1 mM phenylmethylsulfonyl fluoride) containing 150 mM NaCl and no sodium dodecyl sulfate (for antibodies to Myc, HA-Rpb3, Spt16, Rpb3, and TFIIS and for 8WG16) or FA lysis buffer containing 300 mM NaCl and 0.1% sodium dodecyl sulfate (for HA-Spt20, Med19, Gal4, and TBP). Dilutions of input DNA (1/50 and 1/100) and immunoprecipitated DNA (1/2 and 1/4) were subjected to quantitative PCR by the incorporation of [α-³²P]dATP. The products were separated on a 6% nondenaturing polyacrylamide gel, and quantification was carried out by PhosphorImager (Molecular Dynamics) analysis. The percent immunoprecipitation was calculated for each sample, and values for all samples were normalized to the value for a control PCR product amplified in each reaction. The primers for this control PCR product amplify a region of chromosome V that is not contained in any ORFs (22).

The GAL1, GAL2, and GAL3 primer sets (27) were described previously, and the sequences are available upon request. The relative positions of the primer sites are as follows: GAL1-UAS, −536 to −276; TATA, −190 to +54; 5′, +590 to +877; and 3′, +1330 to +1657; GAL2-UAS, −464 to −195; GAL3-UAS, −420 to −157; ARG1-UAS, −441 to −213; TATA, −176 to +64; and ORF, +230 to +474; AHP1-UAS, −530 to −348; TATA, −151 to +80; and ORF, +267 to +486; TEF1-UAS, −459 to −259; promoter (Prom.), −149 to +49; and ORF, +813 to +1043; PDC1-Prom., −485 to −118; 5′, +48 to +424; middle of ORF (Mid.), +411 to +799; and 3′, +1226 to +1602; and PMA1-UAS, −623 to −390; Prom., −370 to −90; 5′, +584 to +807; 3′, +2018 to +2290; and 3′ untranslated region, +529 to +742 (relative to stop codon).

All values are relative to that for the translation start site (ATG, +1) except where noted otherwise. All experiments were carried out at least three times, and standard error bars are shown on the graphs. Western analysis revealed no differences in protein levels determined by ChIP between any of the mutants and the wild type.

Northern analysis.RNA isolation and Northern hybridization experiments were carried out as described previously (3, 50). Northern hybridization analysis was conducted with probes to GAL1 and SCR1. At least three independent Northern analysis experiments were carried out.

TFIIS is present at the UAS of Gal4-activated genes.Although TFIIS has been shown to be required for the proper induction of GAL1 transcription, it was not known whether this requirement occurs during initiation or elongation in vivo (54). To better define the role for TFIIS in transcriptional regulation at Gal4-activated genes, we performed ChIP analysis of TFIIS over the GAL1 gene (Fig. 1A). Surprisingly, we observed that when cells were grown at 30°C under inducing conditions (galactose-grown cultures), TFIIS was present over the entire GAL1 locus, with the highest level of association over the UAS (Fig. 1B and C). This localization was dependent on transcription because TFIIS was not present at GAL1 under noninducing conditions (raffinose-grown cultures) (Fig. 1B). Similar results were observed in ChIP experiments with either amino-terminally Myc-tagged TFIIS or untagged TFIIS and anti-TFIIS antisera (Fig. 1B). A previous study did not find TFIIS present at GAL1 under these same conditions (40). However, the epitope-tagged version of TFIIS used in those experiments (nine copies of the Myc epitope at the carboxy-terminal end) has impaired function because it causes a mutant phenotype, sensitivity to mycophenolic acid, likely explaining the different results (D. Prather, E. Larschan, D. Pokholok, R. Young, and F. Winston, unpublished data). Our results obtained with both a functional, epitope-tagged version of TFIIS and wild-type TFIIS showed that TFIIS is physically associated with the GAL1 UAS and coding region.

• Open in new tab
• Download powerpoint

FIG. 1.

TFIIS is recruited to both the UAS and coding regions of GAL1 upon induction by galactose. (A) Schematic representation of the GAL1 gene. The Gal4-binding sites are represented as black boxes (−460 to −330, where the A of the ATG is +1), the TATA box is represented by a gray box (−140), and the transcription initiation site is denoted by the arrow (−60). Regions amplified in the ChIP PCRs are noted below the gene. (B) ChIP analysis of TFIIS at GAL1. Strains containing 3×Myc-TFIIS, wild-type TFIIS, or dst1Δ were grown in YP medium-2% raffinose at 30°C to 1 × 10⁷ to 2 × 10⁷ cells/ml and induced for 20 min with 2% galactose. Cells were cross-linked before galactose addition (raffinose sample) or after galactose addition (galactose sample). In the left three panels, TFIIS was immunoprecipitated from chromatin with anti-Myc polyclonal antibody (Ab) A14 against the 3×Myc epitope. In the right two panels, TFIIS was immunoprecipitated with anti-TFIIS antibody. PCRs were conducted with two dilutions to ensure linearity. IN, input chromatin; IP, immunoprecipitated chromatin. (C) Quantitation of 3×Myc-TFIIS ChIPresults at 30°C. Strains containing 3×Myc-TFIIS or wild-type TFIIS were cross-linked before (raffinose) or after (galactose) the addition of galactose. With anti-Myc antibody A14, the percent immunoprecipitation of each GAL1 region was calculated and normalized to the percent immunoprecipitation of a noncoding region of chromosome V. Each value represents the ratio reported as the mean and standard error from at least three independent experiments. In this experiment, wild-type TFIIS was included as a negative control, as it is not recognized by the anti-Myc antibody. (D) Quantitation of TFIIS ChIP results at 30°C. Strains containing wild-type TFIIS were cross-linked after the addition of galactose. With the anti-TFIIS antibody, the percent immunoprecipitation of each GAL1 region was calculated and normalized to the percent immunoprecipitation of a noncoding region of chromosome V. Each value represents the ratio reported as the mean and standard error from at least three independent experiments. (E) Distributions of TFIIS, Pol II, and Spt16 across GAL1. Antibodies to Spt16 and Pol II (antibody 8WG16 against the C-terminal domain of Rpb1) were used to immunoprecipitate these proteins from the same wild-type chromatin samples as those used in panel B. The relative distributions of TFIIS, Pol II, and Spt16 are shown; the highest level of enrichment for each protein was arbitrarily set to 1. (F) TFIIS is recruited to the GAL2 and GAL3 UAS regions. Strains containing 3×Myc-TFIIS or wild-type TFIIS were grown in YP medium-2% raffinose at 30°C to 1 × 10⁷ to 2 × 10⁷ cells/ml and induced for 20 min with 2% galactose. Cells were cross-linked after galactose induction. TFIIS was immunoprecipitated from chromatin with anti-Myc polyclonal antibody A14 against the 3×Myc epitope. PCRs amplified the UAS regions of either GAL2 or GAL3 (see Materials and Methods).

To compare the distribution of TFIIS across GAL1 to that of another well-studied elongation factor, we assayed the association of Spt16 with GAL1. Previous studies showed that Spt16 and other elongation factors, including Spt6 and the Paf1 complex, are generally present at significantly higher levels over coding regions than over promoters (21, 33, 40, 45). As expected, our ChIP experiments showed that Spt16 associated at a higher level with the GAL1 ORF than with the promoter region (Fig. 1E). These results are in contrast to those obtained for TFIIS, which showed a higher level of association with the GAL1 UAS than with the coding region. Thus, the distribution of TFIIS is distinct from those of other elongation factors or initiation factors, such as SAGA and Mediator, which are present only over promoter regions (27, 28). This pattern of TFIIS localization suggests that TFIIS may play roles in both transcription initiation and elongation at GAL1.

We next investigated whether TFIIS is specifically recruited to the GAL1 UAS or whether it is also recruited to other galactose-inducible promoters in S. cerevisiae. We found a significant level of association with the GAL2 UAS (3-fold) but at most only very weak enrichment at the GAL3 UAS (1.3-fold) (Fig. 1F). These levels of TFIIS association correlate with the level of SAGA occupancy at the GAL2 UAS and the GAL3 UAS (27) and suggest a general role for TFIIS in Gal4-activated transcription.

TFIIS is required for the rapid induction of GAL1 transcription.TFIIS was previously shown to be required for full induction of the GAL1 gene in synthetic complete (SC) medium (54). Since we used different growth conditions (yeast extract-peptone [YP] medium with galactose rather than SC medium with galactose), we performed experiments to measure the requirement for TFIIS during GAL1 induction. When strains were grown at 30°C, GAL1 mRNA levels were clearly reduced in the dst1Δ mutant, particularly at early time points (Fig. 2A and B). At later time points, GAL1 mRNA levels in the dst1Δ mutant were approximately 60% wild-type levels. When strains were grown at a higher temperature, 37°C, we found a significantly more severe defect in the dst1Δ mutant over the entire 2-h time course of the experiment (Fig. 2C). These results suggest that TFIIS plays an important role in the robust and rapid induction of GAL1, both under optimal growth conditions and at higher temperatures.

• Open in new tab
• Download powerpoint

FIG. 2.

TFIIS is necessary for the rapid induction of GAL1 transcription. (A) Northern analysis of GAL1 mRNA levels in response to galactose induction. Wild-type and dst1Δ strains were grown in YP medium-2% raffinose at 30 or 37°C to 1 × 10⁷ to 2 × 10⁷ cells/ml and induced for 20 min with 2% galactose at either 30 or 37°C. RNA was isolated at different times after galactose addition. SCR1 served as a loading control. (B) Quantitation of Northern analysis at 30°C. Levels of GAL1 mRNA were quantitated and normalized to the level of SCR1 mRNA. The level of GAL1 mRNA at 120 min in the wild-type strain at 30°C was set to a value of 100, and all other values are shown relative to this time point. Each value represents the mean and standard deviation from at least three independent experiments. (C) Quantitation of Northern analysis at 37°C. Northern analysis was performed as described for panel B.

TFIIS is required for TBP and Pol II association with the GAL1 promoter.Our observations that TFIIS is necessary for normal levels of GAL1 transcription and is physically associated with the GAL1 UAS suggest that TFIIS may be important for initiation at GAL1. To determine whether TFIIS functions to promote transcription initiation, we performed ChIP experiments to measure the levels of TBP and Pol II at the GAL1 promoter. We observed that the levels of TBP and Pol II at the GAL1 promoter were significantly reduced in the dst1Δ mutant, with TBP being recruited at only 34% and Pol II being recruited at 37% the levels found in a wild-type strain (Fig. 3A). Previous studies suggested that some elongation mutants, such as spt6 and spt16, have an indirect effect on TBP recruitment at promoters due to aberrant recruitment of TBP to cryptic TATA elements (19, 33). This possibility seems unlikely for the dst1Δ mutant for two reasons. First, unlike spt6 and spt16 mutations, dst1Δ mutations do not cause transcription to initiate from cryptic TATA elements (data not shown). Second, unlike Spt6 and Spt16, TFIIS shows a high level of physical association with the GAL1 UAS. Thus, the most likely explanation for the reduced levels of TBP and Pol II at GAL1 in the dst1Δ mutant is that TFIIS functions directly in transcription initiation, likely at the level of the formation of the preinitiation complex (PIC).

• Open in new tab
• Download powerpoint

FIG. 3.

TFIIS is required for the association of TBP and Pol II with GAL1. (A) TFIIS is required for the association of TBP and Pol II with GAL1 at 30°C. ChIP experiments were conducted with wild-type and dst1Δ strains containing Spt20-3×HA or Rpb3-HA. Strains were grown in YP medium-2% raffinose at 30°C to 1 × 10⁷ to 2 × 10⁷ cells/ml and induced for 20 min with 2% galactose. Primers were directed to the GAL1 TATA for TBP and Pol II and to the GAL1 UAS for Gal4, SAGA (Spt20-3×HA), and Mediator (Med19). Percent immunoprecipitation was calculated relative to that for the control region. Values shown represent the mean and standard error from at least three independent experiments. (B) SAGA and Mediator are required for TFIIS association with the GAL1 UAS. ChIP experiments were conducted with wild-type, spt20Δ, and med15Δ (gal11Δ) strains containing 3×Myc-TFIIS. Strains were grown in YP medium-2% raffinose at 30°C and induced for 20 min with galactose. Cells were cross-linked after the addition of galactose, and 3×Myc-TFIIS was immunoprecipitated with anti-Myc antibody A14. Primers were directed to the GAL1 UAS region, and percent immunoprecipitation was calculated relative to that for the control region. Values shown represent the mean and standard error from at least three independent experiments.

To define further the role for TFIIS in initiation at GAL1, we compared wild-type and dst1Δ strains for the recruitment of three factors known to act during GAL1 transcription initiation: Gal4, SAGA (Spt20), and Mediator (Med19/Rox3). Gal4 is bound to the GAL1 UAS even under noninducing conditions (11, 42), while SAGA and Mediator are recruited early after induction by galactose (5, 6, 8, 27, 28). Mediator was also recently shown to be recruited to the GAL1 UAS independently of Pol II, suggesting a role for Mediator in transcription initiation prior to its association with Pol II (25, 28). In our experiments, Gal4 and SAGA were recruited to wild-type levels in dst1Δ cells, whereas the association of Mediator was only modestly affected (Fig. 3A). These results show that under these conditions, TFIIS is required for normal levels of TBP and Pol II association with the GAL1 promoter, yet it is not significantly required for the association of Gal4 or SAGA and has only a modest effect on Mediator association.

Because SAGA and Mediator were recruited to significant levels in the absence of TFIIS, we tested whether these factors are required for TFIIS association with the GAL1 promoter. To do this, we measured the level of TFIIS association in an spt20Δ mutant, in which the integrity of the SAGA complex is abolished (47, 56), and a med15Δ (gal11Δ) mutant, in which the recruitment of Mediator to the GAL1 promoter is eliminated (8, 28, 35, 39, 48). We found that TFIIS association was completely dependent upon both SAGA and Mediator because in the absence of either, TFIIS was not detectable at GAL1 (Fig. 3B). These results suggest that the association of TFIIS is dependent upon both SAGA and Mediator.

Higher temperatures increase the requirement for TFIIS in initiation at GAL1.Since TFIIS has been implicated as playing a role in the stress response (32, 40), we also measured its level of association with GAL1 when cells were grown at an elevated temperature, 37°C (Fig. 4A). Under these conditions, TFIIS was still associated with the GAL1 UAS, although the levels were apparently lower than those seen at 30°C (compare Fig. 4A to Fig. 1C). However, the association of TFIIS with the GAL1 coding region was no longer detected. These results raise the possibility that TFIIS plays a more limited role in GAL1 transcription at 37°C than at 30°C, possibly confined to initiation.

• Open in new tab
• Download powerpoint

FIG. 4.

Higher temperatures increase the requirement for TFIIS for initiation at GAL1. (A) Levels of TFIIS at GAL1 at high temperatures. Strains containing 3×Myc-TFIIS or wild-type TFIIS were grown in YP medium-2% raffinose at 37°C to 1 × 10⁷ to 2 × 10⁷ cells/ml and induced for 20 min with 2% galactose at 37°C. ChIP experiments were performed as described in the legend to Fig. 1 with antibody A14. (B) TFIIS is partially required for SAGA and Mediator recruitment and becomes essential for the association of TBP and Pol II with GAL1 at 37°C. ChIP experiments were conducted with wild-type and dst1Δ strains containing Spt20-3×HA or Rpb3-HA. Strains were grown in YP medium-2% raffinose at 30°C to 1 × 10⁷ to 2 × 10⁷ cells/ml and induced for 20 min with 2% galactose. Primers used were directed to the GAL1 TATA for TBP and Rpb3-HA and to the GAL1 UAS for Gal4, SAGA (Spt20-3×HA), and Mediator (Med19). Percent immunoprecipitation was calculated relative to that for the control region. Values shown represent the mean and standard error from at least three independent experiments.

Because the overall level of TFIIS associated with GAL1 was decreased at 37°C, we reexamined the effects of dst1Δ on the recruitment of Gal4, SAGA, Mediator, TBP, and Pol II at this temperature to determine whether this decreased level was functionally significant. Under these conditions, the association of Gal4 with the GAL1 UAS was not affected (Fig. 4B). However, the levels of both SAGA and Mediator associated with the GAL1 locus were reduced although not completely abolished (Fig. 4B). Furthermore, the recruitment of TBP and Pol II to the GAL1 promoter became completely dependent on TFIIS at 37°C because the levels of association of both of these factors were at background levels. This increased dependence on TFIIS for TBP and Pol II association compared to the effect observed at 30°C likely was caused by the defects in SAGA and Mediator recruitment seen in dst1Δ cells at this temperature. These results could have been due to an increased requirement for TFIIS in stabilizing SAGA and Mediator at 37°C or to some other indirect effect of TFIIS at a more stressful temperature. Thus, at 37°C, the loss of TFIIS caused additional defects at the GAL1 promoter, consistent with the more severe transcriptional defects that we observed at high temperatures (Fig. 2).

TFIIS can be recruited to Gal4-binding sites independently of Pol II.The localization of TFIIS to the GAL1 UAS and the defects in TBP and Pol II recruitment in the dst1Δ mutant suggest a role for TFIIS in transcription initiation. However, these results do not rule out the possibility that TFIIS is recruited to the promoter by its well-characterized association with Pol II. To gain additional evidence of a role for TFIIS in initiation, we tested whether it can be recruited to a UAS in the absence of Pol II. To do this, we used a plasmid containing three consensus Gal4-binding sites and no other promoter elements (5). Previous results showed that these three Gal4-binding sites are sufficient to recruit Gal4 and SAGA; however, TBP and Pol II do not associate with this plasmid (5). Our ChIP results (Fig. 5) showed that TFIIS was recruited to the three Gal4-binding sites, while Pol II did not associate with this plasmid, providing strong evidence that TFIIS can be recruited to genes in a Pol II-independent fashion. These results are consistent with the results obtained at GAL1 (Fig. 3) because both SAGA and Mediator were also recruited in this experiment and the recruitment of all of these factors was dependent upon Gal4 (Fig. 5). We conclude that TFIIS is recruited to isolated Gal4-binding sites by the Gal4 activator and that this recruitment can occur independently of Pol II.

• Open in new tab
• Download powerpoint

FIG. 5.

TFIIS can be recruited to Gal4-binding sites independently of Pol II. (A) ChIP experiments were conducted with wild-type and gal4Δ strains containing plasmid SGP4 with three consensus Gal4-binding sites. Strains were grown in SC raffinose medium without Ura at 37°C to 1 × 10⁷ to 2 × 10⁷ cells/ml and induced for 20 min with 2% galactose at 37°C. Cells were cross-linked after the addition of galactose. Immunoprecipitation was performed with anti-HA antibody 12CA5 (Spt20-3×HA), anti-Myc antibody A14 (3×Myc-TFIIS), and antibodies to Gal4, Med19, and Rpb3. Primers used amplified a region of the plasmid containing the three consensus Gal4-binding sites. PCRs were conducted with two dilutions to ensure linearity. These experiments were conducted at 37°C, and similar results were seen for strains grown at 30°C (data not shown). (B) Quantitation of recruitment to the plasmid containing the Gal4-binding sites shown in panel A. Percent immunoprecipitation was calculated relative to that for the control region. Values shown represent the mean and standard error from three independent experiments.

TFIIS is recruited to the coding regions of several constitutive and inducible genes.To examine the association of TFIIS with other genes not regulated by Gal4, we performed several additional ChIP experiments. First, we examined two genes that have been shown to be SAGA dependent, ARG1 and AHP1 (41, 49). ARG1 is regulated by Gcn4 and is induced by the addition of sulfometuron methyl, which mimics nutrient starvation. AHP1 is regulated by osmolarity and is induced by the addition of 0.4 M NaCl. When we performed ChIP analyses of TFIIS at these two genes, we found that the association of TFIIS with these genes increased under activating conditions (Fig. 6A). However, in contrast to the association of TFIIS with GAL1, we found high levels of TFIIS associated with the coding regions but very little associated with the promoters.

• Open in new tab
• Download powerpoint

FIG. 6.

TFIIS is recruited to several inducible and constitutive genes even under nonstress conditions. (A) TFIIS is recruited to the inducible genes ARG1 and AHP1. For ARG1 induction, strains containing 3×Myc-TFIIS were grown in SC medium lacking isoleucine and valine at 30°C to 10⁷ cells/ml. Half of the culture was induced with 0.6 μg of sulfometuron methyl (SM)/ml for 2 h, and the other half was left untreated. For AHP1 induction, strains containing 3×Myc-TFIIS were grown in YP glucose (YPD) medium at 30°C to 1 × 10⁷ to 2 × 10⁷ cells/ml. Half of the culture was induced with 0.4 M NaCl for 5 min, and the other half was left untreated. ChIP experiments were performed with antibody A14, and the percent immunoprecipitation was calculated relative to that for the control region. Values shown represent the mean and standard error from three independent experiments. (B) TFIIS is recruited to constitutively expressed genes under nonstress conditions. Strains containing 3×Myc-TFIIS were grown in YPD medium at 30°C to 1 × 10⁷ to 2 × 10⁷ cells/ml, and cells were cross-linked. TFIIS localization at the TEF1, PDC1, and PMA1 genes was analyzed by ChIP with antibody A14, and the percent immunoprecipitation was calculated relative to that for the control region. Values shown represent the mean and standard error from three independent experiments.

We extended our analysis of TFIIS to three constitutively expressed genes; TEF1, PMA1, and PDC1. Similar to what we found at ARG1 and AHP1, TFIIS was present at higher levels over the coding regions of all three of these genes than over the promoter regions (Fig. 6B). A previous study showed a similar distribution of TFIIS at TEF1 but suggested that TFIIS was recruited only under stressful conditions, such as 14°C (40). However, we found that even when cells were grown in rich media at 30°C, TFIIS was recruited to each of the genes examined here. This difference in results likely was due to the different epitope tags used in these studies (as discussed above). We also analyzed the association of TFIIS with TEF1 and PMA1 by using anti-TFIIS antibody and obtained results very similar to those obtained with Myc-tagged TFIIS (data not shown). These results suggest that TFIIS is cotranscriptionally recruited to the coding regions of most genes even under nonstressful conditions. In contrast, TFIIS appears to play an additional, more specialized role in initiation at Gal4-regulated genes.