Site-specific cross-linking of TBP in vivo and in vitro reveals a direct functional interaction with the SAGA subunit Spt3

Incorporation of the photo-cross-linker ρ-benzoyl-phenylalanine (BPA) into TBP

The photoreactive, nonnatural amino acid BPA was incorporated on the surface of S. cerevisiae TBP using the using an Escherichia coli nonsense suppressor tRNA/tRNA synthetase system (Supplemental Fig. 1; Chin et al. 2003; Chen et al. 2007). The TBP gene with 13 Myc epitope tags fused at the N terminus and under the control of the strong ADH1 promoter was cloned to a high-copy 2μ vector. TBP was mutagenized to individually substitute the codons of 61 different exposed surface residues with the amber codon TAG (Supplemental Table 1). Each plasmid bearing a TAG codon in the TBP ORF was transformed into a yeast TBP shuffle strain along with a plasmid expressing both an E. coli tRNA^Tyr amber suppressor tRNA and the E. coli tyrosyl-tRNA synthetase variant selected to charge the suppressor tRNA with BPA (Chen et al. 2007). TBP residues that could be substituted by BPA to generate a functional protein were identified by a plasmid shuffle assay on media containing BPA and the drug 5-FOA to select against the plasmid-borne wild-type copy of TBP. About one-third of these substitutions gave rise to viable yeast, and the growth of nearly all of these viable strains required the addition of BPA to the media (Supplemental Table 1). The TBP mutant strains that were viable in the absence of BPA showed very slow growth, and Western blot analysis showed about 20-fold lower levels of TBP when BPA was omitted. Interestingly, insertion of tyrosine at each of the 61 positions in TBP gave viable cells, suggesting that it is the incorporation of BPA rather than the loss of the wild-type residue that results in lethality. Together, our results show that BPA can be incorporated at 24 surface-exposed positions on TBP to give rise to viable cells.

TBP cross-linking in the PIC

Nuclear extracts were prepared from each of the viable strains containing the BPA-substituted TBP, and all supported levels of Pol II transcription in vitro that were comparable to those observed with wild-type nuclear extracts (data not shown). Next, the nuclear extracts and the activator Gal4-VP16 were used to form PICs on immobilized DNA templates bearing the HIS4 promoter and a single Gal4-binding site. The PICs were washed to remove nonspecifically bound proteins, and one-half of the reaction was irradiated with UV light for 5 min while the other half received no UV treatment. Samples were analyzed by SDS-PAGE and Western blot (Fig. 1A). Myc-tagged TBP migrates at ∼55 kDa, and no cross-linking was observed using extracts with wild-type TBP (Fig. 1A, lanes 1,2). In contrast, BPA substituted at 10 different positions on TBP gave numerous cross-links that were apparent as UV-dependent shifts in the mobility of TBP (Fig. 1A, lanes 3–22). The remaining 14 viable TBP variants showed no reproducible cross-linking (Supplemental Table 1). The altered mobility of cross-linked TBP was used to estimate the size of the cross-linked polypeptides, which ranged from 17 to >190 kDa (Supplemental Table 2). As expected, positions that are close to each other in the TBP structure most often cross-link to proteins of the same mobility (for example, compare cross-linked proteins in variants K97 and L87) (Fig. 1A,B). Although several substituted positions of TBP were observed to cross-link to multiple polypeptides, from Figure 1 we cannot determine if TBP is contacting multiple proteins simultaneously or if the PICs represent a mixture of complexes in which the TBP makes alternative interactions (see below). Polypeptides that have been implicated previously in PIC formation and that have molecular weights similar to the predicted sizes of the cross-linking candidates were individually tested for cross-linking to TBP by Flag epitope-tagging the potential candidate gene and testing for a change in mobility of the cross-linked protein. Supplemental Table 3 lists the proteins that were successfully identified by this method as well as other factors that were tested but did not appear to be cross-linked based on the mobility shift assay.

TBP position S184 cross-links to Ncb2

TBP with BPA substituted at position S184 cross-links to a polypeptide with an observed mobility of 17 kDa. The NC2 complex that is known to interact with TBP has two subunits, Ncb1 and Ncb2, and the latter has a predicted size of 16.6 kDa. To test whether Ncb2 cross-links to TBP at position 184, the NCB2 gene was triple Flag tagged in this strain, and the cross-linking experiment was repeated. Figure 2 (lanes 2,4) shows that the predicted 17-kDa cross-linked polypeptide shifts mobility in the Flag-tagged strain to a new estimated size of 22 kDa, the size difference predicted to be conferred by the addition of the triple Flag tag (Chen et al. 2007). This shifted cross-linking product is also detected by an antibody to the Flag epitope tag on Ncb2, confirming that this cross-linked complex contains both TBP and Ncb2 (Fig. 2, lanes 5,6). Using similar criteria, we showed that BPA at TBP position S61 also cross-links to Ncb2 (Supplemental Table 3; data not shown). The cross-linking of TBP to Ncb2 at these positions is in agreement with the X-ray structure of hTBP and hNC2, which shows that these residues in human TBP are close to the NC2β subunit of the NC2 complex (Kamada et al. 2001). These findings demonstrate that our cross-linking method detects specific and biologically relevant interactions.

TBP cross-links to Mot1 at positions S184, D130, and S61

Several TBP derivatives exhibit very slow mobility upon cross-linking, suggesting that they may be cross-linked to Mot1, the largest known TBP-interacting factor at 210 kDa (Auble et al. 1994). To determine whether Mot1 was the cross-linked factor, Mot1 was Flag-tagged in these TBP–BPA strains, and the cross-linking was repeated. Due to the large size of Mot1, a triple Flag tag on Mot1 does not produce a visible mobility shift. However, probing the Western blot with anti-Flag showed a UV-dependent shift in the mobility of Mot1-Flag to a mobility identical to that of the cross-linked TBP derivatives (Fig. 3A, cf. lanes 1–6 and 7–12). This confirms that Mot1 is one of the large polypeptides cross-linked at TBP positions 61, 130, and 184. These three positions of TBP also cross-link to at least one other large polypeptide that has not yet been identified (Fig. 3A). The cross-linked Mot1 at TBP position S184 displays a faster mobility than those at TBP positions 61 and 130 (Fig. 3A), and this is likely due to cross-linking at a different position within Mot1 giving rise to a branched polypeptide with altered mobility compared with cross-linking at other positions. TBP S184 has the strongest relative cross-linking to Mot1 compared with cross-linking at positions D130 and S61. We also found that Mot1 cross-links to TBP at positions T153 and L87 (Fig. 1A; data not shown).

Figure 3B summarizes the positions on the TBP surface that cross-link to Mot1 on immobilized templates (red residues). Previously, it was shown that mutation of three positions on the TBP surface disrupts interaction with Mot1 in an in vitro binding assay (Fig. 3B, blue residues) (Auble and Hahn 1993). Our cross-linking results significantly extend the known surface of TBP interacting with Mot1 across the top and both legs of TBP.

TBP cross-links to Mot1 in vivo

Since the cross-linker BPA is incorporated in vivo, we tested whether this can be used as a probe for protein–protein interactions in living yeast cells. For this experiment, we used the TBP strain substituted with BPA at position S184 and a Flag tag on Mot1, as TBP S184 gives rise to strong, reproducible in vitro cross-links with several polypeptides, including Mot1. The strain was grown in minimal media and the culture was irradiated with UV light to activate the cross-linker. Cells were then collected by centrifugation and disrupted to generate whole-cell extracts. Extracts from irradiated and nonirradiated cells were separated by SDS-PAGE and were probed by Western blotting. Upon irradiation, TBP position S184 cross-links to several polypeptides, including a high-molecular-weight protein (Fig. 3C, lane 2). This high-molecular-weight polypeptide is also recognized by the anti-Flag antibody (Fig. 3C, lane 4), confirming that Mot1-Flag cross-links to TBP position S184 in vivo. While this cross-linking is seen on the immobilized promoter as well as in vivo, other cross-links are also observed when the cells are irradiated, potentially reflecting the interactions made by TBP at other types of Pol II promoters or in association with Pol I or Pol III factors. Ncb2, which cross-links to TBP at position 184 on the immobilized promoter, does not visibly cross-link in vivo, suggesting that that a greater fraction of TBP in growing cells is interacting with Mot1 compared with NC2 (Fig. 3C; data not shown).

TBP helix 2 contacts TFIIA

The structure of the yeast TBP–TFIIA–DNA complex reveals that the two protein factors interact primarily through the N-terminal leg and stirrup of TBP (Geiger et al. 1996; Tan et al. 1996). However, the TFIIA used in these structures lacked a 160-amino-acid internal segment (residues 60–220) of Toa1, the large TFIIA subunit. Mutations on TBP helix H2, located on the saddle surface, result in loss of TFIIA interaction and suggested that TFIIA makes additional interactions with TBP that were not observed in the X-ray structure (Buratowski and Zhou 1992; Bryant et al. 1996; Geiger et al. 1996; Tang et al. 1996). We found that position D130, located in TBP helix H2, cross-links to a protein with an estimated size of 49 kDa, similar to that of Toa1. To determine whether this cross-linked polypeptide was the large TFIIA subunit, Toa1 was Flag tagged in this strain and the cross-linking experiment was repeated. In this new strain, Toa1 shows a UV-dependent mobility shift at the exact mobility observed with cross-linked TBP (Fig. 4, lanes 2,4). This result directly demonstrates that wild-type TFIIA contacts TBP through interactions with TBP helix 2.

TBP interacts directly with Spt3 and Spt8 in the PIC

The Spt3 and Spt8 subunits of SAGA interact genetically with TBP (Eisenmann et al. 1992, 1994; Laprade et al. 2007), and Spt8 has been shown to interact biochemically with TBP in purified systems (Sterner et al. 1999; Warfield et al. 2004; Sermwittayawong and Tan 2006). However, to date there has been no biochemical evidence to show a direct interaction between Spt3 and TBP. The HIS4 promoter used for our cross-linking studies is SAGA-dependent and TFIID-independent (Fishburn et al. 2005) and therefore provides an ideal system to test TBP–SAGA interactions within the PIC.

TBP position R79 cross-links to several proteins ranging in size from 54 kDa to >191 kDa (Fig. 5A, lane 2). This position is near the surface of the TFIIA four-helix bundle in the TBP–TFIIA–DNA complex, which has been shown to cross-link to Spt8 in PICs (Warfield et al. 2004). To determine if one of these cross-linked polypeptides was Spt8, SPT8 was Flag tagged in this strain and the cross-linking was repeated. This resulted in a mobility shift of the p70 cross-linked polypeptide by ∼5 kDa, conclusively identifying p70 as Spt8 (Fig. 5A, lane 4). BPA positioned at S184 cross-links to a protein with an estimated size of 43 kDa, the size of Spt3 (Fig. 5B, lane 2). Residue S184 is also close to the TBP residue G174, which, when mutated, genetically interacts with Spt3 (Fig. 5E; Eisenmann et al. 1992). SPT3 was Flag tagged in this strain and the cross-linking was repeated. A mobility shift of the p43 cross-linked polypeptide was observed upon Flag fusion, identifying this polypeptide as Spt3 (Fig. 5B, lanes 2,4). Cross-linking experiments were also performed with extracts from a strain bearing a deletion of SPT3 and with BPA at TBP position S184. As predicted, TBP no longer cross-links to the p43 polypeptide, again confirming the identity of Spt3 as the cross-linked p43 factor (data not shown). Together, our analysis provides direct evidence for specific interactions between key subunits of SAGA and TBP in the PIC.

Our results show that TBP contacts at least three different proteins at position S184: Ncb2, Mot1, and Spt3. Since previous work has shown that SAGA is recruited to promoters by activators (Bhaumik and Green 2001; Larschan and Winston 2001), we investigated whether any of these cross-links were activator-dependent. The cross-linking experiment was repeated as before except that the activator Gal4-VP16 was omitted. Our results show that cross-linking of TBP to Mot1 and Ncb2 is not dependent on the presence of the activator (Fig. 5C, lane 2) while cross-linking to Spt3 is detected only when the activator is present (Fig. 5C, lane 4). Western analysis of these reactions, probing for the SAGA subunits Ada1 and Spt3, showed that SAGA is recruited to the immobilized promoter only in the presence of activator (data not shown). Our results suggest that Mot1 and Ncb2 contact TBP in complexes that are separate from PICs that are formed in the presence or absence of activator. This is consistent with the observation that TBP is present in approximately fivefold excess over other general transcription factors, and Pol II is bound to the immobilized promoter (Ranish 1999). In order to confirm that the cross-links between TBP and Mot1, Ncb2, and Spt3 are not dependent on each other, we repeated the cross-linking experiment using extracts that were immunodepleted of SAGA via a Flag tag on Spt3. As shown in Supplemental Figure 2B, this results in a 75% depletion of Spt3 (lane 4) as compared with mock-depleted extract (lanes 1–3), and the TBP cross-link to Spt3 is correspondingly reduced without reduction in the cross-linking to Mot1 or Ncb2 (Supplemental Fig. 2A). Together, our results show that the cross-links between TBP, Mot1, and Ncb2 are not dependent on Spt3, and suggest that Mot1, NC2, and SAGA are present in separate complexes with TBP.

TBP cross-links to Spt3 in vivo

To determine whether the contact between TBP and Spt3 occurs efficiently in vivo, cells containing BPA at TBP S184 and a Flag tag on Spt3 were grown in galactose media. The cells were irradiated, whole-cell extracts were prepared, and cross-linked proteins were assayed by Western blot. As seen in Figure 5D (left panel), probing with the anti-Myc antibody reveals a TBP-cross-linked polypeptide with an estimated size of 43 kDa. This cross-linking product is also visible when the Western blot is probed with anti-Flag antibody, which detects Spt3-Flag (Fig. 5D, right panel), indicating that the cross-linked product contains both TBP and Spt3. This experiment demonstrates that the TBP–Spt3 interaction occurs efficiently in vivo, confirming the interaction observed in PICs.

TBP cross-linking is not DNA-dependent

Our studies above showed that TBP contacts several factors on immobilized promoter templates. To determine whether these proteins interact with TBP only at the promoter, we performed cross-linking in extracts in the absence of any exogenously added promoter DNA. We tested for TBP cross-linking to Mot1 and Toa1 at TBP position D130 and to Spt3 at TBP position S184 by irradiating yeast nuclear extracts and analysis by Western blot. Supplemental Figure 3 shows that TBP cross-links to Flag-tagged Mot1, Toa1, and Spt3 in extracts. This indicates that a fraction of TBP is associated with these factors in extracts before binding to promoter DNA.

Functional interactions of TBP with Spt3

To investigate the significance of the Spt3–TBP interaction, mutations were generated in surface-exposed residues of TBP located near the site of Spt3 cross-linking and near residue G174, previously shown to genetically interact with TBP (Fig. 5E; Eisenmann et al. 1992). Several mutations near TBP residue 184 at positions Y185, E186, L189, and F190 were lethal. In contrast, other nearby residues at TBP positions A176, F177, G180, T181, F182, S183, I194, and R196 were viable and were tested for growth phenotypes consistent with a SAGA defect (Fig. 6A; Supplemental Table 4). Mutations at positions F177, F182, and R196 all displayed a slow growth phenotype in liquid culture and a cold-sensitive phenotype at 18°C, similar to that of an spt3Δ mutant. These three mutations also showed slow growth on YP GAL (+anti-mycin) with mutations F177E and R196E showing the strongest phenotypes, which were similar to, but not as severe as, a complete SPT3 deletion. Mutations F177E, R196E, and spt3Δ also showed a strong growth phenotype on plates containing the drug sulfometuron methyl (SM), which mimics amino acid starvation and requires induction of Gcn4-regulated genes for growth. In contrast, mutations near the site of Spt8 cross-linking at residue R79 (G77, C78, L80, D81, L82, A100, A117, S118, K120, and F148) were all viable and had no detectable growth phenotype in our assays, comparable to the lack of growth phenotype observed in an spt8Δ strain. Mutation of residues L80 and K120 was lethal.

A likely reason for the growth phenotypes of mutations F177, F182, and R196 is that these TBP mutations disrupt the physical interaction between TBP and Spt3. We were unable to test for disruption of this protein–protein interaction using the cross-linking assay since the combination of BPA insertion at S184 and the individual TBP mutations is lethal. As an alternative assay, we measured the association of TBP and SAGA in extracts made from strains containing either of the TBP mutations or a deletion of the SPT3 gene. Previous analysis found that SAGA purified using the TAP-tag method from whole-cell extracts contains very low levels of TBP (Laprade et al. 2007). However, we found that a single-step affinity purification of SAGA via a triple Flag tag on the SAGA subunit Spt7 retains associated TBP (Fig. 6B). In contrast, an spt3 deletion eliminates ∼90% of the SAGA-associated TBP. Similarly, the three SPT15 mutations (SPT15 is the gene encoding TBP) F177E, F182E, and R196E all had lower levels of SAGA-associated TBP with levels ranging from 15%–22% of wild type (Fig. 6B). This demonstrates that these mutations in TBP reduce the TBP–Spt3 interaction, resulting in a phenotype similar to an SPT3 deletion. In contrast to these results, an SPT8 deletion mutation did not significantly disrupt SAGA–TBP coprecipitation, consistent with the lack of growth phenotype in this strain (Fig. 6C). Radical mutant substitutions around the TBP–Spt8 interaction site, including C78E and A100E, also lacked a TBP interaction defect (Fig. 6C).

To confirm that the defect in Spt3–TBP interactions could lead to a transcription activation defect, in vitro transcription assays were performed using nuclear extracts from cells with wild-type TBP or TBP bearing the radical mutations F177, F182, or R196 or from cells containing an SPT3 deletion. Single-round transcription reactions were performed using the SAGA-dependent HIS4 core promoter. Activation is nearly lost in the spt3Δ mutant extract and was severely reduced in extracts from the three TBP mutations (Fig. 6D). The activation levels in these three strains were ∼10% of wild-type levels. As expected, adding recombinant TBP to the transcription reactions suppresses the activation defect caused by the TBP mutations, but not in the spt3Δ extract.

To check for activation defects in vivo, Gal4 and Gcn4-dependent transcription was measured in the mutant strains. Wild-type yeast, spt3Δ cells, or the strains bearing the TBP mutations were induced with galactose or with SM. RNA was harvested and primer extension was used to quantitate transcripts from the SAGA-independent gene ACT1, and the SAGA-dependent genes GAL1, GAL10, and HIS4. Activated transcription from the GAL10, GAL1, or the HIS4 genes is impaired 50-fold, eightfold, and sixfold, respectively, in the Δspt3 strain (Fig. 7A,B). Similarly, activated transcription of all three genes is also reduced by the TBP mutations. The reduction in expression ranges from 16-fold to 2.5-fold for GAL1,10 genes and sixfold to 1.6-fold for HIS4, while the expression of the SAGA-independent gene ACT1 is unaffected.

In order to determine whether the reduction in activated transcription caused by the TBP mutations reflects a defect in TBP recruitment to promoters in vivo, we performed ChIP assays to measure the level of TBP occupancy at the GAL1 promoter. Upon galactose induction, SAGA is recruited to the GAL1 UAS, and recruitment is reduced about twofold in the Δspt3 strain (Fig. 7C). It is well established that the SAGA complex remains intact in the absence of the Spt3 subunit (Wu and Winston 2002). The strains containing the three TBP mutants also show SAGA recruitment to the GAL UAS, at a level equal to or greater than that observed in the Δspt3 strain. In contrast, upon galactose induction, TBP recruitment to the GAL1 promoter was abolished in the Δspt3 strain as well as in the TBP mutants F177E and R196E. Reduced TBP recruitment was observed in the TBP mutant F182E. Together, our results confirm the functional significance of the direct Spt3–TBP interaction in SAGA-dependent transcription observed in our cross-linking assay.

Construction of plasmids

TBP was cloned into the LEU2 2μ plasmid pRS425 (Christianson et al. 1992). The ADH1 promoter was cloned upstream of the TBP ORF. A 13× Myc tag was inserted at the TBP N terminus after the first codon using the QuikChange method (Stratagene) (Geiser et al. 2001), yielding the plasmid pNM10. QuikChange was used to substitute 61 different codons in the ORF with the amber codon TAG in order to create the site-specific mutations that were subsequently confirmed by sequencing. pLH157 (TRP1) containing the E. coli nonsense suppressor tRNA/tRNA synthetase system was a gift from Linda Warfield (Chen et al. 2007).

Yeast strains and media

All strains are derived from BY4705 (Brachmann et al. 1998). In the TBP shuffle strain, the entire gene encoding TBP (SPT15) was deleted using the KanMX marker. The amber-substituted TBP plasmids and pLH157 were cotransformed into this shuffle strain. To prepare solid media containing BPA, 10 mL of 50 mM BPA (BACHEM) dissolved in 1N HCl was added to 1 L of media along with 10 mL of 1N NaOH to a final BPA concentration of 0.5 mM. The transformants were streaked onto −Trp −Leu +BPA medium. TBP residues that could be substituted by BPA to generate a functional protein were identified by a plasmid shuffle assay on media containing BPA and the drug 5-FOA to select against the wild-type URA3 plasmid-borne copy of SPT15. Strains were Flag-tagged at genomic loci using standard methods, and SPT3 and SPT8 were deleted using the hygromycin marker. Anti-mycin A (1 μg/mL; Sigma) was used where indicated.

Photo-cross-linking reactions, immunoprecipitations, and transcription

Nuclear extracts were made from the TBP amber mutants strains as described previously (Ranish et al. 1999), except that 0.5 mM BPA was added to the media during cell growth. The extracts were used to form PICs as described previously using immobilized HIS4 promoter DNA (Ranish et al. 1999). The complexes were washed with and resuspended in 100 μL of transcription buffer. UV-treated samples were then placed in a Spectrolinker XL-1500a UV cross-linker (Spectronics Corp.) and irradiated twice at energy setting 7500. Samples were then resuspended in LDS Sample Buffer and run on a 4%–12% BIS-Tris gel (Invitrogen) in MOPS buffer for 80–120 min at 200 V. To detect photo-cross-linking, TBP was visualized by Western analysis using commercially available mouse monoclonal anti-Myc antibodies (1:2000; Covance), and Flag-tagged proteins were visualized using M2-Flag antibody (1:2000; Sigma) using the Odyssey infrared imaging system (LiCor Biosciences). For immunodepletion of SAGA from nuclear extracts that were used in cross-linking, we prepared Protein G Dynabeads (Invitrogen) conjugated to the M2-anti-Flag antibody. Five milligrams of nuclear extract were depleted using 20 μL of Protein G Dynabeads that had been conjugated to 1 μg of M2 antibody. Depletions were carried out for 2 h at room temperature with flicking to resuspend the beads every 15 min. Mock-depleted extracts were incubated with Protein G Dynabeads that were not conjugated to antibody. For in vivo cross-linking, cells were grown in minimal media with glucose or galactose to OD600 ∼1.0, harvested and resuspended in 10 mL of minimal media, and poured into an 8-cm square Petri dish. Cells were irradiated as described above for PICs, washed in extraction buffer containing in 100 mM Tris (pH 8.0), 250 mM ammonium sulfate, 10% glycerol, 1 mM EDTA, 0.1% Triton X-100, 1 mM DTT with protease inhibitors, and resuspended in 1 mL of extraction buffer in a 2-mL screwcap tube with ∼400 μL of glass beads. Cells were disrupted using a Mini Bead Beater (Biospec Products) for 5 min to prepare whole-cell extracts. Extracts were spun at 4°C, 20,000g in an Eppendorf 5417C microfuge for 20 min, and frozen on dry ice. To demonstrate DNA-independent cross-linking to TBP, nuclear extracts were prepared and 50 μg of extract were irradiated with UV light. Ten micrograms of the irradiated as well as nonirradiated extract was separated by SDS-PAGE and analyzed by Western blotting. For SAGA immunoprecipitations, 0.5 mg of extract was prepared as above without UV irradiation and incubated with 20 μL of M2-Flag beads (Sigma) for 1–2 h at 4°C. The beads were washed two to three times as above in extraction buffer, and SAGA was eluted with 200 ng/μL Flag peptide (Sigma) in 60 μL of TBS. In vitro transcription assays were carried out as described previously (Ranish et al. 1999), and used plasmid pSH515 with the HIS4 promoter and an upstream Gal4-binding site along with yeast nuclear extract and recombinant activator Gal4-VP16.

RNA extraction and primer extension

For induction of the GAL genes, 100 mL of cells was grown at 30°C to OD600 ∼0.8–1.0 in YPD and shifted to 2% galactose for 5 h. For induction of Gcn4-dependent genes, cells were grown in −isoleucine, −valine minimal media, and sulfometuron methyl dissolved in DMSO (Sigma) was added to a final concentration of 5 μg/mL for 30 min. RNA was extracted using hot acid phenol and chloroform. Twenty micrograms to 40 μg of RNA were used per 10-μL primer extension reaction and incubated with radioactive end-labeled primers to recognize transcripts from ACT1, GAL1, GAL10, and HIS4. MMLV-RT (Invitrogen) was used for the reverse transcription, which was performed as described previously for in vitro transcription (Ranish et al. 1999). Reactions were run on an 8% sequencing gel and visualized by phosphorimager. Transcript levels were quantitated using ImageQuant software.

ChIP

ChIP assays were performed as described previously (McConnell et al. 2004), with a few modifications. For GAL1, 10 induction, strains were cultured in rich media with 2% glucose and then shifted to galactose for 5 h. Chromatin was solubilized by sonication using a Bioruptor (Diagenode, Inc.) for 30 min total (30 sec ON/30 sec OFF) with samples cooled on ice after 15 min. Chromatin containing ∼500 mg of protein per reaction was incubated overnight at 4°C with 6 μg of the following antibodies: for TBP immunoprecipitationss, anti-Myc 9E10 (Covance); for Spt7 immunoprecipitationss, anti-M2-Flag (Sigma). Twenty microliters of Dynabeads Protein G (Dynal) were washed with FA buffer + 5 mg/mL IgG-free BSA, then added to each overnight reaction and allowed to bind for 90 min at 4°C. Beads were washed as described (McConnell et al. 2004) and eluted with 50 μL of 20 mM Tris (pH 8.0), 100 mM NaCl, 20 mM EDTA, 1% SDS for 15 min at 65°C. Elutions were repeated and combined. One-hundred microliters of TE were added and cross-linking was reversed as described, then DNA was purified with a PCR purification kit (Invitrogen) into a 50-μL final volume. Ten percent of the chromatin input was processed to reverse cross-links along with immunoprecipitation samples. qPCR was carried out in 15-μL reaction volumes using Power SYBR PCR master mix (Applied Biosystems), 300 nM each forward and reverse primers, and 1.5 μL of 1:4 diluted DNA. PCR program: 2 min at 50°C, 10 min at 95°C, 40× (15 sec at 95°C, 1 min at 60°C). Samples were run in duplicate, and relative amounts of DNA were calculated using a standard curve generated from serial dilutions of yeast genomic DNA using POL1 primers. Experiments were performed in biological duplicate. Spt7 occupancy was determined by amplifying a region surrounding the GAL1 UAS, while TBP occupancy was determined by amplifying a region surrounding the GAL1 TATA. Primers in a Pol1 ORF were used as a nontranscribed region control.

The following primers were used to amplify the indicated sequences: GAL1-UAS-FP, AGTAATACGCTTAACTGCTC ATTGCT; GAL1-UAS-RP, ACGCACGGAGGAGAGTCTTC; GAL1-TATA-FP, GCGATTAGTTTTTTAGCCTTATTTCTG; GAL1-TATA-RP, GGTTATGCAGCTTTTCCATTTATATAT CT; POL1-ORF-FP, TTTCTGCTGAGGTGTCTTATAGAAT TCA; POL1-ORF-RP, GCTTTGGGCCCATGCAT.

The ratios of GAL1 UAS or TATA to control signals (Pol I) in immunoprecipitation samples were normalized for the corresponding ratios for Input (IN) samples, and the resulting values were plotted in histograms as “occupancy.” Three independent experiments were performed.