Rsc4 Connects the Chromatin Remodeler RSC to RNA Polymerases^‡

Media, plasmids, and strains.

Yeast strains are described in Table 1. Three hemagglutinin (HA)- or thirteen Myc-tagged strains were constructed as described by Longtine et al. (29). The Escherichia coli strains used were DH5α [endA1 hsdR17 supE44 thi-1 recA1 gyrA relA1 Δ(lacZYA-argF)U169 deoR (φ80d lacZΔ M15)] and BL21 [F⁻ ompT hsdS(r_B⁻ m_B⁻)]. Bacteria were grown in LB medium supplemented with kanamycin (50 μg/ml) or ampicillin (100 μg/ml) for, respectively, pENTR or pDEST20 and pDEST17 plasmids.

Yeast genetic techniques and media have been described by Sherman (46). Standard molecular biology techniques were as described by Sambrook et al. (42). Two-hybrid interactions were tested in strain Y190 (18) as described previously (14, 53).

Immunoprecipitations.

Whole-cell extracts were prepared from 100 ml of cells growing exponentially in yeast extract-peptone-dextrose (YPD) medium. Cells were collected at an optical density at 600 nm of 0.6 to 0.8, washed twice with water and twice with lysis buffer (50 mM HEPES [pH 7.5], 100 mM NaCl, 20% glycerol, 1 mM dithiothreitol [DTT], 0.5 mM EDTA, 0.05% NP-40) supplemented with a protease inhibitor cocktail (Complete; Roche) and 1 mM phenylmethylsulfonyl fluoride, and resuspended in 0.5 ml of the same buffer. Lysis was performed in the presence of glass beads (0.2 ml, 425 to 600 μm) by vortexing for 30 min at 4°C. The cell debris were eliminated by centrifugation (15 min at 4°C at 18,000 × g, twice). The supernatant was collected and stored at −80°C.

During the immunoprecipitation procedure, the incubations and washes were performed at 4°C with agitation. A total of 2 × 10⁷ anti-mouse immunoglobulin G (IgG)-magnetic beads (Dynal M450) were washed with 0.1% bovine serum albumin in phosphate-buffered saline (PBS; 137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 2 mM KH₂PO₄) and preincubated for 30 min in 0.1% bovine serum albumin in PBS. The 12CA5 anti-HA antibody (70 ng/μl), 9E10 anti-Myc antibody (30 ng/μl), or 8WG16 antibody (200 ng/μl) were added for 1 h, after which the beads were washed three times for 5 min and twice shortly with incubation buffer (the same as lysis buffer). The protein extracts (15 μg/μl, 100 μl) were incubated with the beads for 3 h, after which the beads were washed three times (5 min) with incubation buffer. The affinity-purified proteins were released from the beads by boiling them for 10 min. Eluted proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and analyzed by Western blotting with 12CA5 anti-HA, 9E10 anti-Myc, anti-Rsc4, anti-Rsc6 (B. Cairns, unpublished material), or 8WG16 anti-CTD of Rpb1 (Covance).

In vitro interaction studies.

To overexpress Rpb5, Rsc4, Rsc4Δ4, Rsc4-C93, and Rsc4-C93-Δ4 proteins, the RPB5 open reading frame (ORF), the RSC4 ORF, or the RSC4 C-terminal domain, as well as the RSC4 ORF or the RSC4 C-terminal domain lacking codons for the four last amino acids, were amplified by PCR of genomic DNA and cloned in the pENTR/D-TOPO vector (Invitrogen). The oligonucleotide sequences are available on request. The resulting pENTR plasmids were verified by DNA sequence analysis. The plasmids for the glutathione S-transferase (GST) fusion proteins Rsc4, Rsc4Δ4, Rsc4-C93, and Rsc4-C93-Δ4 and the hexahistidine (His₆)-Rpb5 fusion protein expression were obtained by an LR Gateway reaction between the pENTR plasmids and the pDEST20 or pDEST17 plasmids (Invitrogen).

Rsc4, Rsc4Δ4, Rsc4-C93, and Rsc4-C93-Δ4 were expressed as GST fusion proteins in High Five insect cells by using Bac-to-Bac baculovirus expression system (Invitrogen). Cell extracts were prepared as described previously (12) except that sonication was performed in PBS (pH 7.4) containing 20% glycerol, 5 mM 2-mercaptoethanol, and protease inhibitor cocktail (Complete). Then, 1-ml lysates (6 to 8 mg of total protein) were diluted to final volume of 20 ml in loading buffer (PBS [pH 7.4], 1 mM phenylmethylsulfonyl fluoride, 10 mM DTT, 0.1% Triton X-100). and loaded onto 0.5 ml of glutathione-Sepharose 4 Fast Flow columns (Amersham Biosciences) equilibrated with the same buffer plus 1% Triton X-100. The columns were washed sequentially with 15 ml of (i) loading buffer; (ii) PBS, 1 M NaCl, and 10 mM DTT; and (iii) PBS plus 10 mM DTT. GST fusion proteins were eluted with 5 ml of elution buffer (50 mM Tris-HCl [pH 8.0], 50 mM NaCl, 10 mM DTT, 20 mM reduced glutathione, 5% glycerol), and 0.5-ml fractions were collected. Peak fractions were pooled, dialyzed against dialysis buffer (10 mM HEPES [pH 7.5], 50 mM NaCl, 10 mM DTT, 10% glycerol), frozen in small aliquots in liquid nitrogen, and stored at −80°C. SDS-PAGE, followed by Coomassie brilliant blue staining, showed that GST fusion proteins were purified to near homogeneity. GST alone and GST-Anc1(Taf14) were purified from E. coli BL21 transformed with recombinant plasmids (M. Kabani, unpublished material).

A total of 0.3 nmol of GST-Rsc4, GST-Rsc4Δ4, GST-Rsc4-C93, or GST-Rsc4-C93-Δ4 or as controls GST or GST-Anc1(Taf14) were incubated with 50 μl of glutathione-Sepharose 4 Fast Flow (50% slurry) in GST-binding buffer (20 mM HEPES, [pH 6.8], 100 mM KCl, 5 mM MgCl₂, 0.1% NP-40, 2% glycerol, 1 mM DTT, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride) and rotated for 1 h at 4°C in a total volume of 900 μl. A 40-μl aliquot of extracts of E. coli BL21 expressing His₆-Rpb5 (140 μg of total protein) was added, and reaction mixtures were rotated for 2 h at 4°C. By Western blotting with anti-Rpb5 antibodies (21) compared to serial dilutions of homogeneous yeast Pol III preparation, we estimated that this amount of the extract contained ca. 1.4 nmol of His₆-Rpb5 protein. Reactions were centrifuged for 2 min at 14,000 rpm, the supernatant was removed, and the pellet was washed three times with 500 μl of GST-binding buffer. The pellet fractions were boiled for 5 min in 50 μl of SDS-PAGE sample buffer. The supernatant and pellet fractions (10 μl) were then resolved by SDS-PAGE (12% polyacrylamide) and analyzed by Western blotting with polyclonal anti-GST antibodies (Sigma) and anti-Rpb5 antibodies (21).

Microarray analysis.

Gene expression was monitored with DNA microarray manufactured by the Service de Génomique Fonctionnelle (CEA/Evry, France) as described previously (13) except that an indirect cDNA labeling protocol of the targets was used (adapted from P. Brown [http://cmgm.stanford.edu/pbrown/protocols/aadUTPCouplingProcedure.htm]; see Materials and Methods in the supplemental material).

DNase I analysis of chromatin structure.

Preparation of yeast nuclei and chromatin analysis with DNase I were performed as described previously (16). Yeast strains MW4023 and MW4019 were grown in YPD medium at 25°C and then incubated for 6 h at 30°C or 37°C. The rpb1-1 mutant (strain D439-2a, the RY260 isolate obtained by backcrosses with the YPH499 wild-type strain (3), and the wild-type strain YPH499 were grown in YPD medium at 25°C and then shifted to 37°C for 1 h. Next, 250-ml portions of the cell cultures at an optical density at 600 nm of 0.8 were collected and used for nucleus preparation. Nuclear pellets were further digested with different concentrations of DNase I (Invitrogen). Typically, samples corresponding to 0.5, 1, 2, and 4 U of DNase I per ml were purified and selected for secondary digestion with the appropriate restriction enzyme: SphI for the DUT1 gene, BanII for the SMX3 gene, and SpeI for the HTA3 gene. Further, DNA was fractionated on a 1.2% agarose gel, blotted onto nylon membrane, and hybridized to the corresponding radioactively labeled probe. To prepare the probes, PCR fragments obtained by amplification on yeast genomic DNA were labeled by using a Prime-It II random primer labeling kit (Stratagene). The DUT1 PCR fragment corresponds to the region from 222 to 526 bp downstream from the initiation codon, the SMX3 fragment covers the region from 390 to 643 bp downstream from the ATG, and the HTA3 fragment corresponds to the region from 331 to 798 bp downstream from the initiation codon.

Restriction enzyme accessibility assay.

Restriction enzyme accessibility assays were performed essentially as described previously (40, 52). Cell nuclei prepared as described for DNase I analysis were digested with a saturating quantity of enzyme at 37°C for 1 h. A control without enzyme was included to monitor endonuclease activity. Purified DNA samples were quantified by SYBR Green real-time PCR using primers specific to the promoter region encompassing the corresponding restriction enzyme site. Additional control with primers to the region lacking restriction enzyme site was included. We expressed the accessibility as a percentage of the digestion on the promoter region normalized to the control region.

Interaction of the RSC complex with a shared subunit of RNA polymerases.

Previously, we screened the RNA polymerase subunits for interacting protein domains by using the two-hybrid system. The work centered on the interactions between RNA polymerase subunits and did not address the protein contacts between enzyme subunits and polypeptides belonging to other factors or complexes (14). We now investigate the significance of an unpublished interaction between Rpb5, one of the five subunits shared by the three nuclear RNA polymerases (45), and Rsc4, a subunit of the RSC remodeling complex (22). Using Rpb5 as a bait, two fragments were selected (one of which was found twice) that encoded the C terminus of Rsc4 subunit of the RSC chromatin remodeling complex. The interaction domains of the prey fusions spanned the C terminus of Rsc4 and began at either amino acid 533 or amino acid 558 and extended to the end of the protein at position 625, covering the last 93 or 68 amino acids of the subunit, respectively. The interaction of Rsc4 with a common polymerase subunit suggested that RSC could interact with all three RNA polymerases, a possibility we explored further.

To test this hypothesis, Sth1 was tagged at its C terminus by the insertion of a sequence encoding 13 Myc epitopes at the very end of the STH1 chromosomal ORF (29) in a wild-type strain (see Table 1 for a description of the strains used in the present study) or in strains in which the chromosomally expressed largest subunit of Pol I, A190, was tagged with three HA epitopes at its C terminus. The addition of the Myc or HA tags to these essential proteins did not affect the growth of the strain, indicating that the tag did not affect function. The Sth1-Myc expressing strain, in which the RNA polymerase subunits were not tagged, was used as a control for nonspecific adsorption. When A190-HA was immunoprecipitated with anti-HA antibodies (12CA5) in an RPA190-HA STH1-MYC strain, an anti-Myc reacting band was revealed (Fig. 1A, lane 2). No such band was observed when the immunoprecipitations (IP) were performed in strains in which Sth1 or A190 were not tagged (Fig. 1A, lanes 1 and 9, respectively), indicating that no anti-Myc reacting material was immunoprecipitated nonspecifically by 12CA5 antibody and that Sth1-Myc was not adsorbed nonspecifically to 12CA5 IgG magnetic beads. Similarly, the Myc-tagged Rsc4 subunit of RSC complex could be detected when A190-HA was immunoprecipitated from an RPA190-HA rsc4-MYC strain (Fig. S1, lane 2, in the supplemental material) but not from strains in which Rsc4-Myc or A190-HA were absent (Fig. S1, lanes 1 and 3, respectively, in the supplemental material). These observations suggest that Pol I and RSC interact specifically.

We next tested the interaction between Pol III and RSC. Using the anti-HA antibodies, Sth1-Myc coimmunoprecipitated with C160-HA in a RPC160-HA STH1-MYC strain (Fig. 1A, lane 6) but not when Sth1 or C160 were not tagged (Fig. 1A, lanes 5 and 9). We also tested the coimmunoprecipitation of chromosomally expressed Rsc4-Myc or Rsc1-Myc fusions with C160 tagged with three HA epitopes at its N terminus (HA-C160), expressed from a centromeric plasmid under the control of its own promoter. In this strain, the wild-type chromosomal copy of RPC160 was deleted. HA-C160, but not untagged C160, was immunoprecipitated by the anti-HA antibody (Fig. S1A, lanes 4 to 7, in the supplemental material). Strikingly, both Rsc1-Myc and Rsc4-Myc coimmunoprecipitated with HA-C160 but not with untagged Pol III. Furthermore, Rsc4-Myc and HA-C160 still coprecipitated when DNase was added during the incubation, indicating that their association was not dependent on the presence of DNA (data not shown). The observation that Sth1, Rsc1, and Rsc4 coimmunoprecipitated with C160 strongly suggests that RSC complex can be associated with Pol III.

To test for an interaction between Pol II and RSC, we immunoprecipitated Pol II with 8WG16 antibodies directed against the Rpb1 CTD in the STH1-MYC strain. Sth1-Myc coimmunoprecipitated with Rpb1 when 8WG16 antibodies were bound to the IgG magnetic beads (Fig. 1B, lane 1). However, Sth1-Myc or Rpb1 was not retained by the IgG beads alone (Fig. 1B, lane 2), suggesting that RSC specifically associates with Pol II. However, we note that only ∼0.1% of total cellular RSC was immunoprecipitated under these conditions. Altogether, these coimmunoprecipitation experiments indicate that the three RNA polymerases interact with the RSC chromatin remodeling complex in extracts.

rsc4 mutants.

A strain bearing the rsc4-MYC allele exhibited a slight growth defect on YPD-rich medium at 30°C (Fig. 2A) and was unable to grow at 37 and 16°C (Fig. 2A and data not shown), suggesting that the presence of the Myc epitopes interfered with an essential function of the Rsc4 C terminus. In contrast, the addition of Myc epitopes to Sth1 or Rsc1 had no deleterious effect. To directly test for an essential function for the C terminus, a diploid strain was constructed in which one of the RSC4 alleles lacked the last 68 amino acids (rsc4-Δ68::Kan^r). Tetrad dissection and complementation analysis indicated that Rsc4 C terminus is essential, in keeping with the sequence conservation of the last 40 amino acids of Rsc4 in various yeast species (data not shown).

Next, the C terminus of Rsc4 was progressively deleted by inserting stop codons in the chromosomal copy of RSC4 gene. We found that deleting the last amino acid had no effect and that removal of five amino acids or more was lethal. Interestingly, deletion of three or four amino acids (rsc4-Δ3; rsc4-Δ4) led to a growth defect at various temperatures and a strong thermosensitive phenotype at 37°C (Fig. 2A), providing alleles useful for further study.

To investigate the role of Rsc4 C terminus on the structural integrity of the RSC complex, RSC was immunoprecipitated via chromosomally expressed Sth1-Myc in a wild-type or rsc4-Δ4 strain. The proteins immunoprecipitated with anti-Myc antibodies were analyzed by Western blotting and revealed by using anti-Myc, anti-Rsc4, or anti-Rsc6 antibodies. Rsc4-Δ4 was coimmunoprecipitated with RSC irrespective of the temperature at which the mutant was grown (25°C or 6 h at 30 or 37°C; data not shown and Fig. 2B). The abundance of Rsc4 relative to that of Sth1-Myc or Rsc6 was not affected by the mutation. Neither Rsc4-Δ4 nor Rsc6 was immunoprecipitated if Sth1 was not tagged. Rsc4-Δ4 could also be immunoprecipitated in similar experiments performed with a Rsc8-Myc-tagged strain grown at 30 or 37°C (data not shown). In addition, the Rsc4-Myc mutant protein coimmunoprecipitated with RSC (data not shown). We concluded that the growth defect of the mutant strains did not result from an assembly or stability defect of the mutant RSC complex.

We next tested whether the rsc4-Δ4 mutation impaired the interaction between the RNA polymerases and RSC. For that purpose, we performed the immunoprecipitation experiments as described above but in a rsc4-Δ4 background, immunoprecipitating either A190-HA, C160-HA, or Rpb1 in strains expressing either tagged or untagged Sth1. We found that Sth1-Myc coimmunoprecipitated Pol I or Pol II in rsc4-Δ4 strains (Fig. 1A, compare lanes 2 and 4; Fig. 1B, compare lanes 1 and 3). However, Sth1-Myc did not coimmunoprecipitate the C160-HA subunit of Pol III in the rsc4-Δ4 background (Fig. 1A, compare lanes 6 and 8).

rsc4 mutations impair interaction between Rsc4 C terminus and Rpb5.

To investigate whether the interaction between Rsc4 and Rpb5 was direct and to analyze the effect of the Rsc4-Δ4 mutation on the interaction with Rpb5, we performed GST pull-down assays. Rsc4, Rsc4-Δ4, Rsc4-C93 (corresponding to the fragment selected in the two-hybrid screen), and Rsc4-C93-Δ4 were produced as GST fusions in insect cells. The purified fusion proteins (0.3 nmol) were bound to glutathione-Sepharose beads and then incubated with a whole-cell extract of E. coli expressing Rpb5 (1.4 nmol). Unbound Rpb5 was washed away, and Rpb5 bound to the affinity beads was released by boiling and revealed by Western blotting (Fig. 3A). The beads alone, GST, or GST fused to Anc1, a subunit common to the general transcription factors TFIID and TFIIF, were used as a negative control in the pull-down experiments and did not retain Rpb5 (Fig. 3A, lanes 1 to 5, 10, and 11). In contrast, GST-Rsc4-C93 and GST-Rsc4 bound Rpb5, indicating that the interaction between the two proteins was direct (Fig. 3A, lanes 6, 7, 12, and 13). Moreover, the deletion of the last four amino acids of Rsc4-C93 markedly diminished the interaction with Rpb5 (Fig. 3A, lanes 8 and 9). However, the mutation did not impair the interaction of the complete Rsc4 protein with Rpb5 (Fig. 3A, lanes 14 and 15), suggesting that at least another segment of Rsc4 besides its C terminus contributed to the interaction with the RNA polymerase subunit.

We next tested whether the Rsc4-Δ4 mutant protein interacted with Rpb5 in vivo. We fused the Gal4 activation domain to fragments of Rsc4 beginning at amino acid 533 and ending at the last amino acid of the protein (G_AD-Rsc4-C93) or three or four amino acids before the C terminus (G_AD-Rsc4-C93-Δ3 and G_AD-Rsc4-C93-Δ4, respectively). The complete Rsc4 subunit did not interact with Rpb5 in the two-hybrid system and was thus not used. Such a situation, where a domain of a protein but not the complete polypeptide interact, is not uncommon since we found previously that Rpb5 interacts in the two-hybrid system with C-terminal fragments of A190, Rpb1, and C160 but not with the complete RNA polymerase large subunits (14), even though these interactions were confirmed by the yeast Pol II crystal structure (10). Strikingly, when tested in the two-hybrid system, the deletion of three or four C-terminal amino acids led to a complete loss of the interaction between Rsc4-C93 and Rpb5 (Fig. 3B) even though the fusion proteins were produced at the same level as G_AD-Rsc4-C93 (Fig. 3C). The addition of 13 Myc epitopes also abolished the interaction between Rpb5 and Rsc4 C terminus in the two-hybrid system (data not shown), indicating a close correlation between the observed loss of interaction in vivo and the growth phenotypes of the mutants.

To further document the physiological importance of the interaction between Rpb5 and Rsc4, we tested for synthetic phenotypes between rsc4-Δ4 and three rpb5 thermosensitive alleles (Fig. 4). This was done by transforming the rsc4-Δ4 rpb5 YCB1 strain, in which the rpb5 deletion was complemented by the wild-type RPB5 allele borne on a URA3 plasmid, or with TRP1 centromeric plasmids bearing various RPB5 alleles. Two of these, rpb5-H147R and rpb5-R200E, changed conserved residues (Zaros et al., unpublished data). The other conditional mutation was a chimera in which the hinge region, linking the N and C termini of Rpb5 (amino acids 120 to 146) was substituted by a mutant version of the human protein that changed H146 to K (rpb5-Chi7K). All rpb5 mutants were able to grow at 25°C, as was rsc4-Δ4. Strikingly, while rpb5-H147R and rpb5-Chi7K were unable to grow at 25°C in the presence of rsc4-Δ4, no synthetic phenotype was apparent when rsc4-Δ4 was combined with rpb5-R200E. Whereas the GST pull-down and two-hybrid experiments indicated that the interaction between Rpb5 and Rsc4 C terminus was direct and required its last four amino acids, the synthetic lethality experiments with specific RPB5 alleles supported the idea of a specific functional interaction.

Effect of rsc4 mutations on Pol II and Pol III transcription.

The effect of rsc4 C-terminal mutations on Pol II transcription was investigated by using DNA microarray hybridization. RNAs were extracted from a wild-type or an rsc4-Δ4 mutant strain growing exponentially on YPD-rich medium at 25°C and then shifted to 30 or 37°C for 6 h. Two independent RNA preparations were analyzed by hybridization to microarrays. The mean expression ratios in the mutant versus the wild-type strain were computed. At the permissive temperature of 30°C, of the 4,396 genes that were analyzed (see Materials and Methods), 204 were induced and 124 were repressed twofold or more in the mutant (Table 2; see list in the supplemental material). At 37°C, of 4,330 analyzed genes, 338 genes were induced and 168 genes were repressed twofold or more. Thus, Rsc4 appears to control the transcription of ca. 11.7% of all yeast genes.

Genes misregulated in the rsc4 mutant partitioned into various Munich Information Center for Protein Sequences (MIPS) functional categories including metabolism, cell wall organization and biogenesis, cell cycle, stress response, transport or amino acid biosynthesis, etc. (Fig. 5; see also the supplemental material). Repressed genes included ribosomal proteins, proteins involved in rRNA processing, and genes involved in pheromone response, and genes from these functional categories were induced. Previous transcriptome analysis of RSC3 and RSC30 mutants indicated a bias toward similar functional categories even though the genes affected were different (1; see Discussion).

Since RSC and Pol III coimmunoprecipitated, we also investigated the effect of an rsc4 thermosensitive mutation on transcription by Pol III in vivo. Total RNAs from a wild-type or rsc4-MYC strain were extracted, separated by gel electrophoresis, and blotted onto a membrane. The amounts of U6 snRNA, RPR1 RNA, RNase P RNA, and two different tRNAs (tRNA₃^Leu and tRNA₂^Lys) and their short-lived precursors were quantified by Northern blot analysis and phosphorimaging (Fig. 6 and data not shown). Since the U6 RNA transcript is stable, we analyzed the accumulation of the unstable Maxi-U6 derivative in which 59 nucleotides were inserted (5, 32). Although transferring a wild-type strain to 37°C had no effect on transcription of Maxi-U6, the presence of the rsc4-MYC mutation decreased its transcription threefold (Fig. 6A).

At the permissive temperature, the RPR1 precursor was already 1.8-fold less abundant in the mutant strain compared to the wild-type (Fig. 6B). After 7 h of incubation at 37°C, transcription of RPR1 RNA was further diminished threefold in the rsc4-MYC mutant strain but only twofold in the wild-type strain, showing that its transcription was also impaired by the mutation. A diminution of RPR1 transcription was also seen in the rsc4-Δ4 mutant (data not shown). Accordingly, it was found previously that RPR1 transcription is particularly sensitive to Pol III machinery defects (27). In contrast, the transcription of two short-lived tRNA precursors (tRNA₃^Leu and tRNA₂^Lys ) was not affected (data not shown). The amount of mature tRNA₃^Leu did not change and was used as a load control (Fig. 6B). This latter observation indicates that altered RPR1 and SNR6 does not result from a general Pol III transcription defect due to the temperature shift. Altogether, these results indicated that the synthesis of at least some Pol III transcripts is dependent on the integrity of Rsc4.

Effect of rsc4-Δ4 mutation on RSC recruitment and chromatin structure.

RSC recruitment to promoters of several Pol II promoters and to several Pol III genes was investigated by using chromatin immunoprecipitation analysis. Myc tags were fused to the Sth1 subunit of RSC in RSC4 or rsc4-Δ4 yeast strains (Table 1). Nonspecific DNA-binding of the tagged protein was evaluated by control PCR experiments with primers specific for POL1 coding region which was previously shown to be devoid of RSC (36). We analyzed RSC occupancy of the promoter regions of Pol II genes repressed at 30 and 37°C (BAR1 and DUT1) or only at 37°C (SMX3 and STE2) or induced (DDR2, OYE3, GPH1, and CDC26) in the rsc4-Δ4 strain. RSC occupancy of the 35S rRNA gene promoter (Pol I-transcribed gene) and Pol III-transcribed genes [SNR6, RPR1, and tF(GAA)P2] was also analyzed. HTA3 promoter region was used as positive control since RSC binding to this region was previously observed by chromatin immunoprecipitation analysis (36) and since HTA3 transcription was not affected in the rsc4-Δ4 mutant. The recruitment of RSC complex was readily detected at most of the promoter regions tested (Fig. S2 in the supplemental material). However, no significant difference was observed in RSC occupancy between the wild-type and rsc4 mutant strains at either permissive or restrictive temperatures. These results indicated that the observed effect of rsc4-Δ4 mutation on transcription was not the consequence of decreased RSC recruitment to target promoter regions.

Even though RSC recruitment was normal in the rsc4-Δ4 mutant, chromatin remodeling could still be impaired if the mutation affected RSC function. To examine this possibility, we analyzed the chromatin structure of the DUT1 gene, the transcription of which is repressed 3.6-fold at 30°C in rsc4-Δ4 mutant and further reduced 6.1-fold at 37°C relative to the wild type. Yeast nuclei were isolated from RSC4 and rsc4-Δ4 strains grown at 30°C or shifted to 37°C for 6 h, treated with various concentrations of DNase I, and the DUT1 nucleosomal organization was assessed by indirect end labeling. Well-positioned nucleosomes could be identified on the DUT1 coding region and upstream from the ATG initiation codon in the wild-type strain (Fig. 7A), and a strong hypersensitive region was located close to the potential TATA box in the 5′ noncoding region (15). Strikingly, the hypersensitivity of this region was markedly decreased in the mutant at 30 and 37°C, in line with the repressive effect of the mutation on DUT1 transcription at both temperatures.

To confirm the decreased accessibility of DUT1 TATA box region, we analyzed the chromatin structure by the restriction enzyme accessibility assay and real-time PCR analysis (40). The MseI site (positions −70 to −67) located on the potential DUT1 TATA box was chosen for this purpose. Nuclei from RSC4 and rsc4-Δ4 strains grown at 30°C or shifted to 37°C for 6 h were digested with excess MseI enzyme, and purified DNA was subjected to real-time PCR with the primer pair specific to the region from positions −205 to −25 of the DUT1 promoter that included the MseI site. The amount of PCR product generated was inversely proportional to the level of digestion occurring across the region amplified by the primers. As a control of DNA recovery, the region from positions +126 to +206 of DUT1 lacking an MseI site was also amplified. The MseI accessibility was expressed as the percentage of digestion on the promoter region normalized to the control region. We observed that the DUT1 promoter region was significantly less accessible to MseI in the rsc4-Δ4 mutant at 30 and 37°C (Fig. 7B), a finding in agreement with our DNase I accessibility results. Thus, two independent approaches revealed a decreased accessibility of the DUT1 promoter region in the rsc4-Δ4 background.

We also analyzed the nucleosomal organization of SMX3 gene, the transcription of which was repressed less than twofold at 30°C and fourfold at 37°C in the rsc4-Δ4 background. DNase I mapping of the SMX3 region indicated the presence of several positioned nucleosomes with strong hypersensitive sites between nucleosomes −2 and −1 and nucleosomes −1 and +1 in the promoter region and between nucleosomes +3 and +4 in the terminator region (Fig. 7C). At the permissive temperature, the chromatin organization in the rsc4-Δ4 mutant was very similar to that of the wild type. In contrast, at the restrictive temperature, the DNase I accessibility of the hypersensitive sites between promoter nucleosomes was strongly reduced.

As a control, we analyzed the chromatin structure of HTA3 gene since it is one of the genes that is the most strongly enriched by the RSC chromatin immunoprecipitation assay (36) and since its transcription was not impaired by the rsc4-Δ4 mutation. As expected, the chromatin structure of HTA3 is not altered in the rsc4-Δ4 background either at 30 or at 37°C (Fig. S3 in the supplemental material). We also analyzed the chromatin structure of GPH1 and CDC26, two genes that were induced (threefold at 30°C and sevenfold at 37°C for GPH1; fourfold at 30 and 37°C for CDC26) in the rsc4-Δ4 mutant. Their chromatin structure was unchanged in the mutant irrespective of the temperature (data not shown). These data suggested a role for Rsc4 in chromatin remodeling that correlated with the efficiency of transcription of the RSC-responsive DUT1 and SMX3 genes.

Transcription inhibition does not alter DUT1 or SMX3 chromatin organization.

Since both transcription and chromatin organization were altered in the rsc4-Δ4 mutant, we could not distinguish whether altered chromatin organization was the cause or the consequence of diminished transcription at DUT1 and SMX3. We thus analyzed chromatin organization at these two genes in the rpb1-1 mutant that rapidly stops transcription after a temperature shift at 37°C (37). We measured the amount of DUT1 and SMX3 RNA in the wild-type and rpb1-1 strains by reverse transcription-PCR and found that the transcription of the two genes was not significantly altered at 25°C, whereas it was repressed 12.1-fold for DUT1 and 4.2-fold for SMX3 at 37°C. Strikingly, arresting transcription did not change the nuclease sensitivity of the two genes (Fig. 8), indicating that changes in chromatin organization at DUT1 and SMX3 are the consequence of altered remodeling by RSC in the rsc4-Δ4 mutant and not the indirect consequence of weaker transcription.