The Yeast RNA Polymerase II-associated Factor Iwr1p Is Involved in the Basal and Regulated Transcription of Specific Genes*

RNA polymerase II (RNA pol II) is a multisubunit enzyme that requires many auxiliary factors for its activity. Over the years, these factors have been identified using both biochemical and genetic approaches. Recently, the systematic characterization of protein complexes by tandem affinity purification and mass spectroscopy has allowed the identification of new components of well established complexes, including the RNA pol II holoenzyme. Using this approach, a novel and highly conserved factor, Iwr1p, that physically interacts with most of the RNA pol II subunits has been described in yeast. Here we show that Iwr1p genetically interacts with components of the basal transcription machinery and plays a role in both basal and regulated transcription. We report that mutation of the IWR1 gene is able to bypass the otherwise essential requirement for the transcriptional regulator negative cofactor 2, which occurs with different components of the basal transcription machinery, including TFIIA and subunits of the mediator complex. Deletion of the IWR1 gene leads to an altered expression of specific genes, including phosphate-responsive genes and SUC2. Our results show that Iwr1p is a nucleocytoplasmic shuttling protein and suggest that Iwr1p acts early in the formation of the pre-initiation complex by mediating the interaction of certain activators with the basal transcription apparatus.

RNA polymerase II (RNA pol II) is a multisubunit enzyme that requires many auxiliary factors for its activity. Over the years, these factors have been identified using both biochemical and genetic approaches. Recently, the systematic characterization of protein complexes by tandem affinity purification and mass spectroscopy has allowed the identification of new components of well established complexes, including the RNA pol II holoenzyme. Using this approach, a novel and highly conserved factor, Iwr1p, that physically interacts with most of the RNA pol II subunits has been described in yeast. Here we show that Iwr1p genetically interacts with components of the basal transcription machinery and plays a role in both basal and regulated transcription. We report that mutation of the IWR1 gene is able to bypass the otherwise essential requirement for the transcriptional regulator negative cofactor 2, which occurs with different components of the basal transcription machinery, including TFIIA and subunits of the mediator complex. Deletion of the IWR1 gene leads to an altered expression of specific genes, including phosphate-responsive genes and SUC2. Our results show that Iwr1p is a nucleocytoplasmic shuttling protein and suggest that Iwr1p acts early in the formation of the pre-initiation complex by mediating the interaction of certain activators with the basal transcription apparatus.
In eukaryotes, RNA polymerase II (RNA pol II) 3 is the enzyme that synthesizes all the mRNA precursors. This enzyme is highly conserved among species and, in yeast, consists of 12 subunits, namely Rpb1 to Rpb12 (2). The RNA pol II holoenzyme activity is regulated by several associated proteins (3)(4)(5), and its composition changes with the stage of transcription (5). Therefore, identification of the individual components of the RNA pol II complex is essential for its functional characterization. In recent years, "tandem affinity purification" (TAP) methodology and mass spectrometry have been used for the systematic genome-wide characterization of protein complexes (1,6). This approach has not only led to the identification of new complexes, but also to the identification of new components of previously well established complexes. This is the case of Iwr1p (Interacting with RNA pol II; YDL115c). Two works on the proteome-wide purification of yeast protein complexes report a variety of interactions between Iwr1p and the subunits of RNA pol II (1,6). Interactions among Iwr1p and Rpb1p, Rpb2p, Rpb3p, Rpb4p, Rpb5p, Rpb7p, and Rpb10p were identified in both analyses. Together, these results suggest that Iwr1p is a component of the RNA pol II holoenzyme complex.
The promoter of the genes transcribed by the RNA pol II contains DNA sequences that are the binding sites of specific transcription factors that regulate their expression in response to environmental and internal changes. Thus, when yeast cells are starved of inorganic phosphate (P i ), Pho4p in cooperation with Pho2p, activates the expression of more than 20 different genes. These include the high-affinity transporter of inorganic orthophosphate PHO84 and the acid phosphatases PHO5, PHO11, and PHO12 (7). Other transcription factors bind to specific sequences in the promoter to repress transcription of the target genes. This is the case of Mig1p, which acts as repressor of different genes, including the SUC2 and GAL genes (8).
In addition to the sequence-specific DNA-binding proteins, there are a variety of accessory factors that act in a global manner. They include chromatin-modifying enzymes and proteins that positive or negatively affect the formation of active transcription initiation complexes between RNA pol II and general transcription factors. Transcriptional repressor negative cofactor 2 (NC2) is one example of this second type of regulatory factor. It consists of two subunits (NC2␣ and NC2␤; Bur6p/Ydr1p in yeast), which form a stable complex via histone-fold domains (9 -11). In yeast, NC2 interacts directly with the TATA box-binding protein, blocks its association with TFIIA and TFIIB (11,12), and is able to repress transcription by RNA pol II in vitro (13). However, other studies indicate that the NC2 complex activates transcription at some promoters (14,15).
We have previously reported that defects in the transcriptional regulator NC2 can be suppressed by defects in a variety of components of the basal transcription machinery, including the two largest subunits of the RNA pol II, the TFIIH components Tfb1p and Ssl1p, and most of the subunits of the mediator complex (16). Here we report that deletion of the gene encoding the RNA pol II-associated factor Iwr1p is also able to bypass the essential requirement of the NC2-component Bur6p. Our results show that Iwr1p shuttles between the nucleus and the cytoplasm in a Xpo1p-dependent manner, and is able to activate transcription when artificially recruited to a promoter. We report that the deletion of IWR1 positively or negatively affects the basal or induced expression of genes regulated through different pathways. The genetic interactions of the iwr1 mutation with components of the transcription machinery and its effects on the recruitment of RNA pol II suggest that Iwr1p acts early in the gene expression process.

EXPERIMENTAL PROCEDURES
Yeast Strains and Genetic Methods-All the strains used in this study are listed under supplemental Table S1. They were cultured using standard methods. For the growth assays, yeast cultures were diluted to the same A 600 and serial dilutions (1:10) were spotted onto YPD, YPGal, or selective plates, and incubated at various temperatures. 5Ј-Fluoroorotic acid containing plates were prepared by adding 1 g/liter of 5Ј-fluoroorotic acid to synthetic complete medium. Doxycycline was used at a concentration of 10 mg/liter. P i -depleted medium was prepared according to Ref. 17. The P i -free medium was a synthetic complete medium prepared using yeast nitrogen base without phosphate.
Strains containing the P GAL10 -BUR6 and/or P GAL10 -YDR1 alleles were constructed by replacing the wild-type promoter with the GAL10 promoter (including 3 copies of the HA epitope). Deletions of nonessential genes were obtained by substituting the coding sequences by the his5 or KAN R markers. Two different deletions were constructed for IWR1. In the iwr1-⌬1 allele, the sequence from position ϩ58 (from the ATG) to the STOP codon (position ϩ1130) was removed. In the iwr1-⌬2 allele, the entire open reading frame was deleted. No differences were observed between both strains in terms of growth or acid phosphatase activity (results not shown). A Myc, HA, TAP, or GFP tag was added to the C terminus of Iwr1p by introducing the transforming cassette into the genomic copy of the IWR1 gene. In all cases, the PCR-based method described in Ref. 18 was used. The strain containing the P tetO -IWR1 allele was constructed by replacing the wild-type promoter by the tetO promoter using the PCR-based method as described in Ref. 19.
Plasmids-To construct plasmid pIWR1-GFP, genomic DNA from the strain containing the IWR1::GFP::Kan R allele was digested with XbaI and SphI, and ligated into the YCplac33 vector. Plasmids with the right insert were selected in LB plates containing ampicillin and kanamycin. pIWR1-⌬NES-GFP was constructed by PCR amplification of pIWR1-GFP with partially overlapping oligonucleotides where the nucleotides encoding the amino acid sequence IIEHIDLAL had been deleted. The PCR product was digested with DpnI and used to transform Escherichia coli. The deletion and absence of further modifications in the resulting plasmid were checked by sequencing the entire IWR1 open reading frame. The plasmid used to express the LexA-Iwr1p fusion (pLexA-IWR1) was obtained by PCRmediated generation of BamHI and SalI sites immediately upstream and downstream, respectively, of the IWR1 coding sequence. The resulting fragment was ligated into the BamHI and SalI sites of pBM116 (20). pLexA-IWR1 was introduced into the TAT-7 yeast strain (lexAop-HIS3, lexAop-LacZ). The Pho4p overexpression plasmid was constructed by PCR amplification of the genomic PHO4 locus and by cloning the restriction fragment in the YEplac181 multicopy plasmid.
Fluorescence Microscopy-Yeast grown in synthetic complete (SC) medium at the early exponential phase were used to visualize GFP fusions. Aliquots (1.2 l) of the cultures were put onto microscope slides and covered with 18 ϫ 18-mm coverslips. Cells were then viewed using a Zeiss Axioskop II fluorescence microscope. Images were scanned with a SPOT digital camera (Diagnostic Instruments Inc.) and processed using the Adobe PhotoShop software version 8.0.1.
Preparation and Northern Analysis of RNA-For SUC2 analysis, yeast were grown in YPD (with 2% dextrose) at the early exponential phase, and then washed and transferred to YPD containing 0.05% dextrose at the indicated times. For GAL1 induction, cells exponentially growing in YPD were washed and transferred to YPGal media and incubated for 4 h. For PHO5 and PHO84 analyses, cells were grown at the exponential phase in P i -depleted or P i -free medium supplemented with P i to 13.4 mM. Then, cells were washed and transferred to P i -depleted or P i -free medium containing different P i concentrations, and incubated at the indicated times. Alkaline shock was performed according to Ref. 21. After these treatments, cells were harvested, washed with water, and frozen at Ϫ80°C. RNA extraction, analysis, and probes have been previously described elsewhere (22).
Chromatin Analysis-Chromatin digestion with ClaI restriction nuclease was performed as described (25). The chromatin immunoprecipitation (ChIP) analysis was essentially performed as previously described (26). 8WG16 (anti-Rpb1 antibody) from Neoclone (Madison, WI), 12CA5 (anti-HA antibody) from Roche, and 9E10 (anti-Myc antibody) from Roche were used. Co-immunoprecipitated DNA was analyzed in triplicate by quantitative PCR using the DNAEngine (Bio-Rad). Oligonucleotides TTGGTCACCTTA-CTTGGCAAGGCAT and TGGTAATCTCGAATTTGCTT-GCTCT were used to amplify the PHO5 promoter. Sequences of other oligonucleotides are available upon request.
Whole Genome Transcriptional Analysis-Strains used for the whole genome transcription analysis were grown in YPD at the early exponential phase. cDNA synthesis and labeling, filter hybridization, and quantification/normalization of hybridization signals were performed as described (27). RNA was obtained from three independent cultures and analyzed for each strain. Original macroarray data were submitted to the GEO data base, where they have been assigned accession number GSE17303.
Western Blot Analysis-Protein extracts were prepared by trichloroacetic acid precipitation of exponentially growing yeast cells grown in 2% YPGal, and from yeast cells grown in this medium and then transferred to YPD for 5 h. Bur6-HA and Ydr1-HA were detected with the monoclonal anti-HA antibody (Sigma) and chemiluminescence visualization (ECL Advanced; Amersham Biosciences), according to the manufacturer's instructions.

Mutations in IWR1 Suppress Depletion of the Transcriptional
Repressor NC2-To gain further insight into the function of NC2, we genetically screened for suppressors of the NC2 defect (16). For this screening, we used a yeast strain in which the genes encoding both NC2 components (YDR1 and BUR6) were placed under control of the GAL10 promoter. Transposon insertion mutants were generated in this strain using a yeast genomic library that had been mutagenized by insertion of an mTn3-lacZ/LEU2 transposon (28). Suppressor mutants were selected for their ability to grow on dextrose-containing plates at room temperature, and transposon insertion sites were determined. In this screen, we identified a variety of components of the basal transcription machinery involved in the control of transcription initiation, like Mediator components NUT2, MED7, and SRB2, the TFIIH component TFB1, and the RNA pol II subunit RPB7 (16). Another gene identified in this screen, IWR1, is the subject of the present work.
In our screen, we isolated 5 independent clones where transposon insertions were mapped at different locations in either the IWR1 open reading frame or the 5Ј flank of this gene, suggesting that the suppression was a result of the complete loss of the Iwr1p function. This point was confirmed by deleting the IWR1 open reading frame in strains carrying the P GAL10 -YDR1 or the P GAL10 -BUR6 alleles. As observed in Fig. 1A, the deletion of IWR1 partially suppressed the growth defects caused by depletion of both NC2 components. Because the genetic screen was performed in a strain in which expressions of both BUR6 and YDR1 genes are under the control of the GAL10 promoter, the ability to grow in media containing dextrose as the sole carbon source may be due to a defective repression of the GAL10 promoter in the iwr1 mutant. To check this possibility, we analyzed the levels of Bur6p and Ydr1p proteins in the iwr1 mutant by making use of the three copies of the HA epitope added to the N-terminal end of the proteins while substituting the BUR6 and YDR1 promoters by GAL10 (18). In the Western analysis shown in Fig. 1B, we compared the amounts of Bur6p and Ydr1p (expressed from the GAL10 promoter) in both the iwr1 mutant and the isogenic wild-type strain before and after a 5-h incubation in dextrose. Deletion of IWR1 reduces the amount of Bur6p-HA and Ydr1p-HA in both galactose and dextrose ( Fig. 1B), suggesting that this mutation causes defects in the basal and induced expression from the GAL10 promoter. In fact, Northern blot analysis of GAL10 did not show any constitutive expression in dextrose, but revealed a reduced induction in galactose of this gene in the iwr1 mutant strain (Fig. 1C). Therefore, suppression of NC2 depletion by ⌬iwr1 is not a consequence of the defective repression of the GAL10 promoter in dextrose. Then, we wondered whether the deletion of IWR1 could bypass the absence of the otherwise essential NC2-component Bur6p. For this purpose, we deleted the IWR1 gene from a strain carrying a chromosomal deletion of BUR6 and the wildtype BUR6 gene on an URA3/CEN plasmid. The resulting cells were able to grow on 5Ј-fluoroorotic acid-containing plates (Fig. 1D), indicating that deletion of IWR1 is able to relieve the requirement for the essential function of NC2␣, as occurs with mutations in other components of the basal transcription machinery (16, 29 -31).

Transcriptional Defects Caused by Mutations in IWR1 and BUR6 Are Partially Compensated in the Double Mutant Strain-
To assess the effect of the deletion of IWR1 on gene expression when NC2 was defective, we analyzed the expression of genes whose transcription is affected in NC2 mutants (16,32). The FIGURE 1. Deletion of IWR1 suppresses the depletion of the NC2 components Bur6p and Ydr1p. A, the IWR1 gene was deleted in strains carrying P GAL10 -BUR6 or P GAL10 -YDR1 alleles. Wild-type (FY86) and mutant cells were spotted onto YPGal (galactose) or YPD (dextrose) plates and incubated for 3 days. B, suppression of NC2 depletion by ⌬iwr1 was not due to a defective repression of the P GAL10 -YDR1 and P GAL10 -BUR6 in dextrose. The total protein extracts either obtained from cells exponentially growing in galactose (Gal) or transferred to dextrose for 5 h (Dext) were analyzed on 8% SDS-PAGE gel. Bur6-HA and Ydr1-HA were detected with mouse monoclonal anti-HA. Ponceau staining of the Western blot is shown as a loading control. C, effect of the ⌬iwr1 mutation in the basal and induced levels of the GAL10 gene. Total RNA was prepared from the wild-type or ⌬iwr1 mutant strain growing in glucose and after incubation for the indicated times (minutes) in galactose containing medium. The application and transfer of equal amounts of RNA were verified by ethidium bromide staining. D, deletion of the IWR1 gene bypasses the requirement of Bur6p. A deletion of IWR1 was introduced in a strain carrying a chromosomal deletion of BUR6 and the wild-type BUR6 gene on an URA3/ CEN plasmid. The resulting strain (⌬iwr1) and the parental strain (IWR1) were streaked onto a 5Ј-fluoroorotic acid plate and incubated for 5 days at 30°C.
induction of the GAL1 gene by galactose has been reported to be abolished in a bur6-1 mutant strain (32). As Fig. 2A indicates, the transcriptional induction of GAL1 was impaired in both the bur6-1 and ⌬iwr1 strains, but was partially restored in the bur6-1 ⌬iwr1 double mutant. On the other hand, depletion of the NC2-component Ydr1p increased the constitutive level of HSP26 (16). As seen in Fig. 2B, the bur6-1 mutation also increased the basal expression of the HSP26 gene. A weak constitutive expression of HSP26 was also observed in the ⌬iwr1 mutant strain. Interestingly, the constitutive expression of HSP26 was not observed in the bur6-1 ⌬iwr1 double mutant strain (Fig. 2B). Therefore, deletion of IWR1 was able to partially compensate some transcriptional defects caused by mutations in NC2.
Transcriptional Activation by a LexA-Iwr1p Fusion-The nature of the suppressors of NC2 requirement that we and others have identified suggests that Iwr1p could play a positive role in transcription initiation (16,29,33). To investigate this possibility, we expressed a fusion protein that includes the DNA binding domain of LexA and the entire Iwr1p protein (LexA-Iwr1p) in a strain carrying the HIS3 and lacZ reporter genes whose promoters include LexA binding sites. As Fig. 3 illustrates, the expression of LexA-Iwr1p has a positive effect on HIS3 and lacZ transcription, as indicated by the ability of the strain carrying the LexA-Iwr1p fusion, but not the LexA binding domain alone, to grow in the presence of 3-aminotriazole ( Fig. 3A) and by the increased ␤-galactosidase activity caused by the presence of LexA-Iwr1p (Fig. 3B). Therefore, our results show that the binding of Iwr1p to the promoter could have a positive effect on transcription.
Iwr1p Shuttles between the Nucleus and Cytoplasm in an Xpo1pdependent Manner-The isolation of iwr1 mutations as suppressors of the transcriptional repressor NC2, together with the reported physical interactions between Iwr1p and different RNA pol II subunits (1,6), suggest that at least some Iwr1p functions occur in the nucleus. To localize Iwr1p in the cell, we added a GFP tag to the C terminus of Iwr1p and cloned the gene fusion into a centromeric plasmid. The functionality of the Iwr1p-GFP fusion protein was confirmed by its ability to complement the slow growth phenotype of a ⌬iwr1 mutant strain (results not shown). As Fig. 4A depicts, Iwr1p-GFP was distributed throughout the entire cell. Analysis of the Iwr1p amino acid sequence using the NetNES 1.1 server predicted a possible leucine-rich nuclear export signal (NES) between positions 169 and 177 (IIEHIDLAL), suggesting that Iwr1p could shuttle between the nucleus and cytoplasm in an Xpo1p-dependent manner. To verify this possibility, we examined the localization of Iwr1p in xpo1-1 cells. Fig. 4B illustrates how Iwr1p-GFP was detected mainly in the nucleus following a shift to 37°C, a condition under which the Xpo1-dependent protein export was blocked. No difference in the localization of Iwr1p-GFP was observed between 30 and 37°C when the fusion protein was expressed in a wild-type strain (results not shown). This result indicates that Iwr1p is a shuttling protein, and suggests the presence of a NES recognized by Xpo1p. Deleting this sequence from Iwr1p-GFP led to the nuclear accumulation of Iwr1p-GFP (Fig. 4C), suggesting that this sequence is required for the nuclear export of Iwr1p. Deletion of this sequence only slightly affected the ability of the fusion protein to complement the growth defect caused by the iwr1 null mutation (results not shown).
Transcriptional Effects of the IWR1 Deletion-We analyzed the genome-wide transcriptional effects caused by the deletion of IWR1 by comparing the transcriptional profile of a ⌬iwr1 strain with the isogenic wild-type strain when cells were exponentially grown in YPD medium. A total of 179 genes showed at FIGURE 2. Deletion of the IWR1 gene alleviates the transcriptional defects caused by mutations in NC2 components. A, exponentially growing cells in YPD medium were washed and transferred to YPGal, and incubated at 30°C for 4 h. B, yeast cells were grown in YPD medium at the early exponential phase. Total RNA was prepared from the indicated strains and analyzed by Northern blot. The RNA pol III transcript SCR1 was used as a loading control.  Procedures") were grown at the mid-exponential phase in SC medium with dextrose as the carbon source at 30°C. B, xpo1-1 mutant cells expressing Iwr1p-GFP were grown at mid-exponential phase at 25°C and heat shocked at 37°C for 30 min. C, nuclear accumulation of a ⌬NES-Iwr1p-GFP. ⌬iwr1::HIS3 mutant cells expressing Iwr1p-GFP or Iwr1p-(⌬NES)-GFP (see "Experimental Procedures") were grown in SC medium with dextrose as the carbon source at the mid-exponential phase. DIC, differential interference contrast. OCTOBER 16, 2009 • VOLUME 284 • NUMBER 42 least 2-fold changes in their expression levels in the ⌬iwr1 strain, as compared with the wild-type control (see supplemental Table S2). Depletion of IWR1 resulted in increased transcript levels of 74 genes, whereas a decreased expression was noted for 105 genes. To determine whether the deletion of IWR1 affects specific metabolic pathways, we analyzed the distribution in functional categories of those genes whose expression levels were altered in the iwr1 mutant strain (supplemental Table S3). We noticed that the products of 15 of the 74 genes that were up-regulated in the iwr1 mutant were localized in the mitochondria membrane. Of these, 10 genes encoded components involved in oxidative phosphorylation, including 7 genes encoding ATP synthase subunits. The possibility of an increased expression of respiratory genes in the iwr1 mutant resulting in an augmented respiratory activity was analyzed by comparing the oxygen consumption of the iwr1 mutant with the wild-type when cells were grown in rich glucose-containing medium. ⌬iwr1 deletion only brought about a modest increase (around 20%) in oxygen consumption (results not shown). The list of genes up-regulated in the iwr1 mutant includes 11 genes whose promoters contain binding sites for the Hap2p/3p/4p/5p complex (GDH1, ATP16, ATP17, COX13, ATP19, ATP15, ATP3, ATP7, QCR7, COX7, and PET9 (SGD) (34)). Interestingly, HAP4 itself was one of the up-regulated genes in the iwr1 mutant. We also noticed that there were 6 genes whose expression might be regulated by Gcn4p (GDH1, ARG1, HSP78, PCL5, ARG5,6, and TRP5 (SGD) (34,35)). Among the down-regulated genes in the iwr1 mutant strain, the only significant functional category found corresponded to acid phosphatase activity, and included PHO3, PHO5, PHO11, and PHO12 genes.

Control of Transcription by Iwr1p in Yeast
Iwr1p Is Required for the Basal but Not the Induced Transcription of the ARG1 and ATP16 Genes-We further studied two well characterized genes whose expression increased in the iwr1 mutant (supplemental Table S2). The ARG1 gene, encoding arginosuccinate synthetase, is controlled by two regulatory pathways. It is induced under amino acid starvation through the Gcn4p transcription factor (36) and repressed by arginine through the ArgR/Mcm1 complex (37,38). The expression of the ARG1 gene can be induced by the addition of sulfometuronmethyl, an inhibitor of isoleucine/valine biosynthesis to the growth medium (39). Fig. 5A shows that the basal expression of ARG1 increased in the ⌬iwr1 mutant. Both the basal and induced expressions of ARG1 were Gcn4p-dependent, but Iwr1p was completely dispensable for the induction of ARG1 by sulfometuron (Fig. 5A). The deletion of IWR1 moderately increased the basal expression of ARG1 in both the presence and absence of Gcn4p.
The ATP16 gene encodes one of the Saccharomyces cerevisiae ATP synthase subunits (40). Transcription of genes encoding the electron transport chain components is regulated by the heme-activated protein complex Hap2/3/4/5p by binding to the CCAAT box at the upstream activation sequence (41). The expression of ATP16 was upregulated in nonfermentative carbon sources, like ethanol, in a Hap4p-dependent manner (Fig. 5B). As with ARG1, the deletion of IWR1 did not affect the induction of ATP16, but moderately increased its basal expression in both the presence and absence of Hap4p (Fig. 5B).
Iwr1p Is Involved in the Expression of the SUC2 and Phosphate Responsive Genes-Many of the aerobic genes up-regulated in the IWR1 mutant (supplemental Table S2) are subjected to glucose repression. Because our genome wide transcription analysis was done with cells exponentially grown in YPD, the increased expression of genes involved in the respiration in this medium could be caused by an impaired catabolite repression in the iwr1 mutant. To check this possibility, we analyzed the expression of SUC2, a well characterized glucoserepressible gene. Fig. 6A shows that the iwr1 mutant cells grown in glucose-containing media did not transcribe SUC2, which suggests that the constitutive expression of glucose-repressed genes was not the cause of the increased level of the respiratory gene transcripts observed in the iwr1 mutant. The expression of SUC2 was induced by lowering the amount of glucose in the growing medium. Under such conditions, we observed reduced levels of the SUC2 transcript in the iwr1 mutant strain compared with the wild-type (Fig. 6A). We also note the almost complete absence of SUC2 mRNA after a 60-min incubation under low glucose conditions. We found that deletion of the MIG1 gene, which encodes a repressor of the SUC2 expression (8), avoided the transient absence of the SUC2 transcript observed after 60 min incubation. However, it did not significantly increase the reduced levels of SUC2 in the iwr1 mutant (Fig. 6B).
Among the transcripts whose expression was reduced in the iwr1 mutant, we identified several phosphate responsive genes (supplemental Table S2). These genes are induced when cells are starved for P i (42), and their expression is regulated by transcription factors Pho4p and Pho2p (43). Fig. 6C shows how the ⌬iwr1 mutation impaired induction of the PHO5 gene under low P i conditions. The defective induction of the PHO5 gene in the iwr1 mutant was not caused by a defective localization of Pho4p because the entrance of this transcription factor into the nucleus under low P i conditions was not affected by the iwr1 mutation (result not shown). To check whether Iwr1p nucleocytoplasmic shuttling was required for activation of the PHO5 gene, we analyzed the induction of the acid phosphatase by low P i conditions in a strain carrying a version of Iwr1p from which the NES had been deleted. As observed in supplemental Fig. S1, FIGURE 5. Increased basal transcription of the ARG1 and ATP16 genes in the ⌬iwr1 mutant. A, yeast strains were cultured in SC medium lacking Ile and Val, and treated (ϩ) or not (Ϫ) with sulfometuron (SM) for 2 h to induce Gcn4p synthesis by starvation of Ile/Val, and then subjected to Northern blot analysis. Two different exposures of the film (long and short) are shown. B, yeast strains were grown overnight in YPD (D) or YPEthanol (E). Total RNA was prepared from the indicated strains, and analyzed by Northern blot. The RNA pol III transcript SCR1 was used as a loading control.
the induction of acid phosphatase in this strain was similar to that observed in the wild-type strain. Phosphate responsive genes respond differently to intermediate P i conditions. Thus, a low expression of PHO5 occurs at intermediate P i concentrations, whereas PHO84 is significantly induced (44,45). When we analyzed the effect of the IWR1 deletion on expression of PHO84, we observed increased PHO84 mRNA levels as the P i concentrations lowered in both the wild-type and ⌬iwr1 mutant strains, although the level of transcripts was always lower in the mutant (Fig. 6C). The induction of PHO84 observed in the ⌬iwr1 mutant under low P i conditions was dependent on Pho4p as no expression was observed in a ⌬iwr1⌬pho4 double mutant strain (Fig. 6D). Pho4p has also been found to be involved in induction of PHO84 by alkaline pH (21). As Fig. 6E shows, this induction was also negatively affected by the absence of Iwr1p.
The Requirement of Iwr1p for the Induction of PHO5 Depends on the Induction Conditions-Korber et al. (46) reported that the histone chaperone Asf1p is essential for PHO5 induction under submaximal inducing conditions, but not in P i -free medium. We analyzed the effect of IWR1 deletion on the induction of PHO5 in both the P i -depleted medium and the complete absence of P i . We were unable to detect any increase in secreted phosphatase activity in the iwr1 mutant in the P i -depleted medium (Fig. 7A). In the complete absence of P i (P i -free), however, induction of PHO5 was delayed in the iwr1 mutant, but the levels of acid phosphatase activity after overnight incubation were similar to those shown by the wild-type strain (Fig. 7A).
Deletion of the negative regulator Pho80p leads to nuclear concentration of Pho4p and to the full induction of PHO5, even under high P i conditions (47,48). Fig. 7B shows that Iwr1p is not required for the activation of PHO5 in the absence of Pho80p, as indicated by the high level of secreted acid phosphatase activity noted in the ⌬pho80 ⌬iwr1 double mutant strain. Finally, we found that the requirement of Iwr1p for the induction of PHO5 is alleviated by overexpression of Pho4p, as indicated by the increased levels of secreted acid phosphatase activ-ity shown by the iwr1 mutant strain that was transformed with a multicopy plasmid carrying the PHO4 gene (Fig. 7C).
Iwr1p Is Required for Normal Chromatin Remodeling Kinetics at the PHO5 Promoter-Among the phosphate responsive genes, PHO5 is a well studied example of a gene whose promoter undergoes extensive remodeling at the chromatin level during the induction process. Under high P i conditions, four positioned nucleosomes are associated with the PHO5 promoter region (49,50). These four nucleosomes are evicted during P i starvation, and the process can be followed by measuring the accessibility of a ClaI restriction site located within nuclesome Ϫ2 (Fig. 8A). As observed in Fig. 8B, whereas the ClaI site in the wild-type strain was accessible in the phosphate-depleted medium after a 4-h incubation, the level of chromatin opening in the iwr1 at this time was similar to that observed under high phosphate conditions, suggesting that the absence of Iwr1p delays the kinetics of PHO5 induction at the chromatin opening level. Next, we checked the recruitment of the Pho4p activator in a ⌬iwr1 mutant. Previous reports indicate the requirement of chromatin remodeling for Pho4p binding to the nucleosomeoccluded UASp2 under low P i conditions (17,60,61). Accordingly by ChIP analysis, we found that the recruitment of Pho4p upon P i depletion is delayed when Iwr1p is absent (Fig. 8C). As expected, the impairment in chromatin remodeling observed in the ⌬iwr1 mutant also delays the recruitment of Rpb1p to the PHO5 promoter (Fig. 8D).
Defects in the Induction of SUC2 in the iwr1 Mutant Are Suppressed in the Absence of H2A.Z-Synthetic genetic interactions between ⌬iwr1 and deletions in components of the SWR complex have been identified by synthetic genetic arrays (1). The SWR complex assembles the variant histone H2A.Z/Htz1p into chromatin (51). Because the deletion of both the IWR1 and HTZ1 genes results in synthetic lethality (Fig. 9A), we constructed a yeast strain carrying a deletion of HTZ1 and IWR1 under control of the repressible tetO promoter. As observed, doxycycline addition severely impaired the growth of P tetO -IWR1 ⌬htz1 double mutants (Fig. 9B). We then analyzed the effect of the ⌬htz1 mutation on the defects caused by depletion  SUC2 and PHO genes in the ⌬iwr1 mutant. A and B, yeast strains were grown in YPD (2% dextrose) at the early exponential phase, and then washed and transferred to YPD containing 0.05% dextrose for the indicated times (in minutes). In B, two different exposures (long and short) of the film are shown. C, yeast strains were grown in P i -depleted medium containing the indicated amounts of KH 2 PO 4 . D, yeast cells were grown in P i -depleted medium either without P i (low phosphate) or with 13.4 mM KH 2 PO 4 . E, yeast cells were grown in YPD at pH 6.4 (n.i.) and exposed for the indicated times (in minutes) to pH 7.6. Total RNA was prepared from the indicated strains and analyzed by Northern blot. The RNA pol III transcript SCR1 was used as a loading control. OCTOBER 16, 2009 • VOLUME 284 • NUMBER 42 of Iwr1p on the transcription of SUC2. As observed in Fig. 9C, the induction of SUC2 in the P tetO -IWR1 strain followed the same pattern as that observed in the ⌬iwr1 strain (Fig. 6, A and  B). However, some differences in the transcript levels were noted, which were likely due to the residual expression of IWR1 in the P tetO -IWR1 strain, even in the presence of doxycycline. As with the ⌬iwr1 strain, an almost complete disappearance of the SUC2 mRNA was noted after an initial induction. Interestingly, the presence of the SUC2 transcript at intermediate induction times was restored by the deletion of HTZ1, and the kinetics of induction in the double mutant P tetO -IWR1 ⌬htz1 was similar to that shown by the wild-type strain, although the transcript levels were lower in the double mutant at all the times analyzed (Fig. 9C).

Control of Transcription by Iwr1p in Yeast
Iwr1p Was Not Found Preferentially Associated with the Chromatin of the PHO5, SUC2, and ARG1 Promoters-To check whether Iwr1p was recruited together with RNA pol II to the promoter of the genes whose transcription was Iwr1p-dependent, we performed ChIP assays using Iwr1p proteins tagged in different ways (Myc, HA, or TAP). We measured the amount of Iwr1p present at the PHO5 and SUC2 promoters under repressed (high P i or high glucose, respectively) and induced (low P i or low glucose, respectively) conditions. We also analyzed the presence of Iwr1p in the promoter of the ARG1 gene, whose basal expression was increased in the ⌬iwr1 mutant (Fig. 5A). The levels of Iwr1p associated with the PHO5 and SUC2 promoters did not vary significantly during their induction, and were similar to the level of Iwr1p found in a nontranscribed, intergenic region (supplemental Fig. S2). In addition, no enrichment in the Iwr1p content (in relation to the intergenic region) was observed at the different DNA sequences along the PHO5 locus or in the promoter of the ARG1 gene (supplemental Fig. S2). Therefore, our results suggest that Iwr1p is not associated with the RNA pol II when this enzyme is recruited to the promoter of the active genes.

DISCUSSION
Iwr1p was originally identified as a protein that co-purifies with almost every subunit of RNA pol II (1, 6). Iwr1p is highly conserved with a homolog, CG10528, in Drosophila melanogaster, that partly co-localizes with phosphorylated, actively transcribing RNAPII on polytene chromosomes (1). In this report, we present data that further supports the relationship between Iwr1p and the basal transcription machinery. We identified IWR1 in a genetic screen for mutations that suppress the requirement for NC2, a general transcriptional repressor that operates by interacting with the core promoter and components of the basal transcription machinery, like TATA boxbinding protein (reviewed in Ref. 52). Different suppressors of the NC2 function have been identified (13,16,29,31,33). To date, the common feature of each NC2 suppressor reported is that they are components of the basal transcription machinery that play putative positive roles in transcription initiation. Our finding that deletion of IWR1 can bypass the requirement of the NC2 function (Fig. 1) through compensation of the transcriptional defects (Fig. 2) is in agreement with a general role of Iwr1p in transcription initiation. This role is also supported by the ability of Iwr1p to activate transcription when fused to a heterologous DNA binding domain (Fig. 3), which occurs for several subunits of the RNA pol II holoenzyme (53,54).
Unlike other general transcription factors, the localization of Iwr1p is not exclusively nuclear. In this work, we demonstrate that Iwr1p shuttles between the nucleus and cytoplasm and can be detected throughout the cell at the steady state (Fig. 4). Iwr1p accumulates in the nuclei in the xpo1-1 mutant cells and contains a leucine-rich sequence (IIEHID-LAL) that likely acts as a NES. The capability to shuttle across the nuclear envelope is expected for those proteins that play FIGURE 7. A different requirement of Iwr1p depending on the PHO5 inducing conditions. A, yeast cells were grown in P i -depleted medium containing 13.4 mM KH 2 PO 4 at the early exponential phase (0 h), and then washed and transferred to P i -depleted or P i -free medium without added P i . B, yeast cells were grown in P i -depleted medium containing 13.4 mM KH 2 PO 4 to the early exponential phase. C, the wild-type or ⌬iwr1 mutant cells transformed with YEplac181 (vector) or YEplac181 carrying the PHO4 gene (PHO4) were grown in P i -depleted medium containing 13.4 mM KH 2 PO 4 at the early exponential phase (0 h). They were then washed and transferred to the P i -free medium. Error bars indicate the S. D. of three independent experiments. both nuclear and cytosolic roles. However, a dynamic distribution is not required for those proteins whose function takes place exclusively in the nuclear compartment, as with the typical general transcription factors, suggesting that Iwr1p could have additional functions in cells that require shuttling. Thus, Iwr1p could participate in one or more signal transduction pathways ending in the nucleus. However, even though this possibility may be in agreement with the transcription profile shown by the ⌬iwr1 mutant, we have not been able to find any condition that modifies the cellular distribution of Iwr1p (results not shown). In addition, we found that shuttling is not required for induction of the PHO5 gene by low P i conditions (supplemental Fig. S1).
Although IWR1 is not an essential gene, deletion of IWR1 moderately impairs cell growth under all the tested conditions (see Fig. 1A; results not shown). Growth impairment is likely related with the defects in gene transcription shown by the iwr1 mutant. We performed whole genome profiling to further investigate the cellular role of Iwr1p. The most striking finding of this analysis was the increased expression of genes encoding mito-  . Deletion of HTZ1 partially suppresses the defect in SUC2 induction shown by the ⌬iwr1 mutant strain. A, synthetic lethality between ⌬iwr1 and ⌬htz1 mutations. The indicated strains carrying the IWR1 gene in URA3/CEN plasmids were streaked in SC-ura and 5Ј-fluoroorotic acid (FOA) plates and incubated for 3 days at 30°C. B, wild-type and mutant strains were spotted onto YPD plates or YPD plates supplemented with 10 mg/liter of doxycycline and incubated for 3 days at 30°C. C, yeast strains were grown in YPD (2% dextrose) at the early exponential phase, and transferred to YPD (2% dextrose) containing 10 mg/liter of doxycycline and incubated for 4 h at 30°C. Then, cells were washed and transferred to YPD (0.05% dextrose) containing 10 mg/liter of doxycycline and incubated for the indicated times (in minutes). Total RNA was prepared from the indicated strains and analyzed by Northern blot. The RNA pol III transcript SCR1 was used as a loading control. chondrial proteins in the iwr1 mutant, particularly those involved in oxidative phosphorylation (see supplemental Table  S2 and Fig. 5). Although most of the genes involved in aerobic respiration are subjected to glucose repression, the transcription profile of the ⌬iwr1 mutant does not show an up-regulation of other glucose repressible genes such as SUC2. An increased transcription of the nuclear genes encoding the components of oxidative phosphorylation has been reported for the mediator component Med5p (Nut1p) (55). Therefore, it is possible that the basic transcription machinery includes several components to coordinate the nuclear gene expression with the respiratory requirements.
In addition to the genes encoding oxidative phosphorylation components, the expression of genes regulated by amino acids (ARG1), carbon source (SUC2 and GAL10), or phosphorus starvation (PHO5 and PHO84) is also altered in the ⌬iwr1 strain. However, the effect of the deletion of IWR1 on the expression of these genes is very different. Deletion of IWR1 increases the basal expression of ARG1 and ATP16, but not their induction. On the other hand, Iwr1p is involved in the induction of SUC2, GAL1, PHO5, and PHO84 genes, suggesting that the requirement of Iwr1p is gene-specific.
The role of Iwr1p as a target of specific transcription factors could explain the striking kinetics of SUC2 induction observed in the iwr1 mutant (see Fig. 6). Geng and Laurent (56) reported a biphasic pattern for the transcriptional induction of SUC2, the first phase of which is partly mediated by the general stress response transcription factors Msn2p and Msn4p. Our results are compatible with an effect of Iwr1p in delaying the second phase of the induction without affecting the early Msn2p/4pdependent induction. In agreement with this model, we found that induction of HSP26, which depends on Msn2p/4p (57), is not affected by the iwr1 mutation (results not shown). Moreover, our results involve Mig1p in the late phase of the SUC2 induction because the transient disappearance of the SUC2 transcript was not observed in the iwr1 mig1 double mutant (see Fig. 6). This suggests that Iwr1p participates in eliminating the repression exerted by Mig1p. The transient disappearance of the SUC2 transcript is also suppressed by deletion of the HTZ1 gene encoding the histone H2A.Z variant (see Fig. 9C). Notably, the repression mediated by Mig1p requires the corepressor Ssn6p-Tup1p (58), and Tup1p has been recently involved in the deposition of histone H2A.Z in the promoter of the GAL1 and SUC2 genes (59). It is, therefore, tempting to speculate that the Mig1p-Ssn6p-Tup1p complex participates in the deposition of H2A.Z in the SUC2 promoter, which could have a negative effect on the late induction of SUC2, in which Iwr1p is necessary to remove this repression. Future studies examining the presence of all these proteins on the SUC2 promoter at different times during the activation process are required to test this model.
The most striking effect of deleting of IWR1 is observed in the induction of the phosphate responsive genes under P i starvation conditions. Our results show that the absence of Iwr1p impairs the normal induction of PHO5 and PHO84 (see Figs. 6 and 7). Different possibilities may be considered to account for the impaired induction of the PHO5 and PHO84 genes in the iwr1 mutant: (i) defects in the PHO signal transduction, e.g. caused by the slower growth rate and, therefore, by the slower exhaustion of intracellular phosphate pools; (ii) a defect in chromatin remodeling, e.g. through the lack of cofactors that are normally recruited via Iwr1p; (iii) a defective recruitment of the transcriptional apparatus to the PHO5 promoter; or (iv) the defective activity of RNA pol II lacking Iwr1p to initiate transcription once bound to the PHO5 sequences. The lack of effect of the iwr1 mutation on the kinetics of the Pho4p nuclear translocation upon P i depletion (results not shown), or the impairment in the induction of PHO84 by alkaline pH (see Fig. 6E), does not support the possible delay in phosphate depletion caused by the slower growth rate of the iwr1 mutant. The response of PHO84 to alkaline pH is also dependent on Pho4p, and exposure to alkali has been suggested to actually generate a situation of phosphate starvation that could be responsible for the induction of some of the PHO genes (21). Under alkaline pH conditions, a significant induction of PHO84 occurs at times as short as 25 min as opposed to the reduced growth rate of the iwr1 mutant being the cause of the defective induction of PHO84 and other PHO genes. On the other hand, our results show a delay in the chromatin remodeling of the PHO5 promoter in the absence of Iwr1p (see Fig. 8B). The failure in chromatin opening is likely to be responsible for the delay in Pho4p and RNA pol II recruitment to the PHO5 promoter observed in the iwr1 mutant (see Fig. 8, C and D).
Together, our findings suggest a role of Iwr1p prior to preinitiation complex formation that would be dependent on the amount of Pho4p recruited to the PHO gene promoter. Thus, the higher requirement of Iwr1p for the induction of PHO5 than for the induction of PHO84 may be explained by the higher accessibility of Pho4p to the PHO84 promoter (44,45). We find that the requirement of Iwr1p for the induction of PHO5 can be partially alleviated by increasing the amount of Pho4p in "reciprocal accordance" with an earlier study in which increased cofactor dependence correlates with decreased Pho4p occupancy at the PHO5 promoter (60). Even though Iwr1p is partially or totally dispensable under full activation conditions (such as P i -free medium or in the pho80 mutant), it is absolutely essential under intermediate activating conditions (such as the P i -depleted medium). These findings support the strong influence of Iwr1p on the Pho4p signaling pathway to the PHO5 promoter, and indicate that once the RNA pol II lacking Iwr1p is bound to the promoter, it is able to transcribe the PHO5 gene normally.
Despite our many efforts, we have not been able to find any specific recruitment of Iwr1p to the chromatin of its target genes (see supplemental Fig. S2). It is, therefore, possible that the association of Iwr1p with RNA pol II is restricted to the enzyme that is not bound to DNA. Thus, Iwr1p could participate in regulation of the recruitment of RNA pol II to specific promoters, but would leave the enzyme prior to the binding of the enzyme to the DNA. The interaction of Iwr1p with the polymerase that is not bound to chromatin would be in agreement with the fact that only a minor fraction of the total Rpb3p subunit could be recovered when a Iwr1-TAP protein fusion was pulled down using IgG-Sepharose beads. 4 Further studies will be needed to elucidate the molecular mechanism through which Iwr1p participates in the transcriptional regulation of its target genes.