Negative cofactor 2 (NC2) has been described as an essential and evolutionarily conserved transcriptional repressor, although in vitro and in vivo experiments suggest that it can function as both a positive and a negative effector of transcription. NC2 operates by interacting with the core promoter and components of the basal transcription machinery, like the TATA-binding protein (TBP). In this work, we have isolated mutants that suppress the growth defect caused by the depletion of NC2. We have identified mutations affecting components of three different complexes involved in the control of basal transcription: the mediator, TFIIH, and RNA pol II itself. Mutations in RNA pol II include both overexpression of truncated forms of the two largest subunits (Rpb1 and Rpb2) and reduced levels of these proteins. Suppression of NC2 depletion was also observed by reducing the amounts of the mediator essential components Nut2 and Med7, as well as by deleting any of the nonessential mediator components, except Med2, Med3, and Gal11 subunits. Interestingly, the Med2/Med3/Gal11 triad forms a submodule within the mediator tail. Our results support the existence of different components within the basic transcription complexes that antagonistically interact with the NC2 repressor and suggest that the correct balance between the activities of specific positive and negative components is essential for cell growth.
IN eukaryotic cells, transcription initiation of genes encoding mRNA involves the assembly of a functional preinitiation complex containing general transcription factors (GTFs) and RNA pol II (Roeder 1998). GTFs include TATA-binding protein (TBP), which plays a central role in the assembly complex. Several factors modulate binding of TBP to DNA. The TBP-associated factors (TAFs) facilitate binding of TBP specifically to TATA-less promoters (Martinez et al. 1995; Oelgeschlager et al. 1996), whereas the general transcription factor TFIIA accelerates and stabilizes binding of TBP to TATA boxes (Yokomori et al. 1994; Weideman et al. 1997; Stewart and Stargell 2001).
Transcription initiation is tightly controlled by the interplay between positive and negative factors (Narlikar et al. 2002; Orphanides and Reinberg 2002). Negative regulation is generally associated with promoter inaccessibility due to chromatin structure (Struhl 1999). However, other mechanisms of repression operate through the core promoter and general transcription factor interactions (Lee and Young 1998). An example of this type of repressor is the negative cofactor 2 (NC2), also known as Dr1-DRAP. NC2 was initially purified from human cell extracts as an activity that inhibits basal TATA-dependent transcription in vitro (Meisterernst and Roeder 1991; Inostroza et al. 1992). NC2 consists of two subunits, NC2α (DRAP1) and NC2β (Dr1), which form a stable complex via histone fold domains (Goppelt et al. 1996; Mermelstein et al. 1996; Kamada et al. 2001). In yeast, a homologous complex exists (Bur6/NC2α and Ydr1/NC2β), and it is required for cell growth (Gadbois et al. 1997; Kim et al. 1997; Prelich 1997). Several lines of evidence have suggested that NC2 functions as an inhibitor of pol II transcription. In vitro, the NC2 complex interacts with TBP and blocks its association with TFIIA and TFIIB (Goppelt et al. 1996; Mermelstein et al. 1996) and, in yeast, a defective TFIIA can suppress the essential role for NC2 (Xie et al. 2000). However, other experiments suggest that NC2 could also positively affect gene transcription (Geisberg et al. 2001; Chitikila et al. 2002). The role of NC2 could depend on the nature of the promoter; it has been shown that Bur6 is able to selectively repress basal transcription from some promoters and to stimulate activated transcription from others (Cang and Prelich 2002).
In addition to gene-specific activators and the RNA pol II machinery, transcriptional activation requires the participation of additional proteins termed co-activators (Biddick and Young 2005). Co-activators can act through the modification of the chromatin structure or by interacting with the RNA pol II and the GTFs. This second class of co-activators includes a multiprotein complex known as mediator. Mediator was originally identified as an adaptor required for activator-dependent stimulation of RNA pol II transcription (Kelleher et al. 1990; Flanagan et al. 1991). In addition to its role in activator-dependent transcription, acting as an interface between gene-specific regulatory factors and the general transcription machinery, there is also evidence that mediator is required for basal transcription (Biddick and Young 2005). The mediator subunits form three functionally and physical distinct modules and an additional subgroup of Srb proteins (the Srb8-11 module), which is variably present in different mediator preparations (Borggrefe et al. 2002). The head module is thought to interact with the C-terminal domain (CTD) of RNA pol II (Lee and Kim 1998). The middle module interacts with the CTD of the RNA pol II, TFIIE, and the Srb8-11 module (Kang et al. 2001). Finally, the tail module does not seem to contact pol II, and it has been implicated in interactions with gene-specific activators (Lewis and Reinberg 2003). Recently, expression-profiling studies have revealed the existence of several antagonistic submodules within the nonessential mediator subunits (van de Peppel et al. 2005). One of these submodules is composed by the mediator tail subunits Med2, Med3, and Gal11. Deletion of any of these components results in similar changes in the global expression profile (mainly decreased transcript levels), suggesting a positive role for this submodule in transcription regulation (van de Peppel et al. 2005). In addition, it has been suggested that the Med2/Med3/Gal11 triad may promote the recruitment of TBP independently of the rest of mediator (Zhang et al. 2004).
The genetic interactions between components of the NC2 repressor and mediator are a paradigm for the complicated network of regulators required to adjust gene expression according to the cell's necessities. Specifically, defects in NC2 components can compensate for the global transcriptional defects caused by mutations in the mediator components MED17/SRB4 and MED22/SRB6 (Gadbois et al. 1997). On the other hand, mutations in MED16/SIN4 can bypass the requirement for NC2 (Kim et al. 2000; Lemaire et al. 2000). The observation that mutations in GAL11 could not suppress the cold-sensitive phenotype shown by ydr1 mutant strains (Kim et al. 2000) and that mutations in MED3, MED2, and RGR1 were unable to bypass the NC2 requirement (Lemaire et al. 2000) led to the interpretation that the suppression was due to a specific genetic interaction between NC2 and SIN4, and not a consequence of the opposite effects of the mediator and NC2 activities on gene expression.
Here we report 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 Tfb1 and Ssl1, and most of the subunits of the mediator complex. The fact that mutations in other subunits of RNApol II, TFIIH, and the mediator exacerbate the growth defects observed in NC2 mutants suggests the existence of submodules within the components of these basic transcription complexes that antagonistically interact with the NC2 repressor.
MATERIALS AND METHODS
Yeast strains and genetic methods:
All strains used in this study are listed in Table 1 and were cultured using standard methods. For growth assays, yeast cultures were diluted to the same OD600 and serial dilutions (1:10) were spotted onto YPD, YPGalactose (YPGal), or selective plates and incubated at various temperatures. 5′-Fluoroorotic acid (5′-FOA)-containing plates were prepared by adding 1g/liter of 5′ fluoroorotic acid to synthetic complete medium. Doxycycline-containing plates were prepared by adding 1, 5, or 10 μg/ml of doxycycline to YPD, YPGal, or selective plates.
Strains containing the PGAL10-BUR6, PGAL10-YDR1, PGAL10-MOT1, PtetO-NUT2, PtetO-MED7, PtetO-RPB7, or PtetO-TFB1 alleles were constructed by replacing the wild-type promoter with the GAL10 [including three copies of the hemagglutinin (HA) epitope] or tetO promoter, using the PCR-based method as described (Longtine et al. 1998). To construct the YDR1 shuffle strain, the BY4743 diploid strain (EUROSCARF) was transformed with a plasmid containing the wild-type YDR1 gene cloned into the pRS316 vector (URA3/CEN). In this strain, one of the YDR1 wild-type alleles was disrupted with the Schizosaccharomyces pombe his5+ gene using a PCR-based method (Longtine et al. 1998). Finally, diploids were sporulated and segregants carrying a genomic disruption of the YDR1 gene were selected.
Plasmids carrying truncated forms of the RPB2 gene were generated by subcloning different restriction fragments from pRP212 (RPB2-CEN) (Scafe et al. 1990) into the YEplac181 vector (2 μm/LEU2). All these fragments start at the XbaI site located at position −872 relative to the ATG codon and end at the positions indicated in Figure 2. The plasmid overexpressing the 3′ truncation of the RPB1 was constructed by subcloning a HindIII–XbaI fragment from plasmid pPR112 (RPB1-CEN) (Nonet et al. 1987) into YEplac181. This plasmid was digested with XbaI and BamHI (in the vector polylinker) and ligated to an XbaI–BglII fragment generated by PCR amplification of pFA6a-13Myc-His3MX6 (Longtine et al. 1998). Accordingly, a 13xMyc epitope was introduced in frame at the XbaI site of RPB1 (position +2532).
Plasmids YEplac195-RPB12, pRS425-RPB9, pRS426-SIN4, pRS316-NUT2, and pRS316-YDR1 were constructed by subcloning restriction fragments obtained by PCR amplification of genomic DNA.
Whole-genome transcriptional analysis:
Strains used for whole-genome transcription analysis were grown in synthetic complete medium lacking leucine (SC−leu) with 2% galactose as the carbon source at 30° to an OD600 of 0.5 and then transferred to a SC−leu containing 2% dextrose for 4 hr at the same temperature. Isolation of total RNA, cDNA synthesis and labeling, filter hybridization, and quantification/normalization of hybridization signals were performed as described (Garcia-Martinez et al. 2004). Data Excel files are available at http://scsie.uv.es/chipsdna/chipsdna-e.html#datos.
Transposon insertion suppression screen:
The PGAL10-BUR6 PGAL10-YDR1 double-mutant strain was transformed with a yeast genomic library mutagenized by the insertion of an mTn3-lacZ/LEU2 transposon (Burns et al. 1994). Suppressor mutants were selected in synthetic complete medium lacking leucine containing 2% dextrose as the sole carbon source at room temperature. After 8 days of incubation, 74 colonies were selected. From these colonies, genomic DNA was isolated and, in 17 of them, sites of transposon insertion were identified by the “vectorette” PCR rescue protocol developed by C. Friddle (http://genome-www.stanford.edu/group/botlab).
Western blot analysis:
Protein extracts were prepared by trichloroacetic acid precipitation of exponentially growing yeast cells grown in SC−leu medium containing 2% galactose and from yeast cells grown in this medium and then transferred to SC−leu medium containing 2% dextrose for 5 hr. Bur6-HA and Ydr1-HA were detected with the monoclonal anti-HA antibody (Sigma, St. Louis) and chemiluminescence visualization (ECL Advanced; Amersham, Buckinghamshire, UK) according to the manufacturer's instructions.
Overexpression of a truncated allele of RNA pol II subunit RPB2 suppresses the growth defect of NC2- and Mot1-depleted cells:
In a previous report, we showed that DBP5/RAT8 genetically interacts with several genes encoding factors that are involved in the control of transcriptional initiation. These interactions include the synthetic lethality between mutations in the BUR6 gene and the dbp5-2 mutant allele (Estruch and Cole 2003). Recently, we found that overexpression of a truncated form of the second largest subunit of the RNA pol II (Rpb2) can suppress the synthetic lethality of the bur6-1 dbp5-2 double mutant (F. Estruch, L. Peiró-Chova and C. Cole, unpublished results). The truncated RPB2 allele contains the coding sequence for the first 379 amino acids (total length of 1224 residues) and was named rpb2t. This finding prompted us to determine whether overexpression of rpb2t was also able to suppress the growth defect shown by bur6 single mutants. To analyze the suppression of the bur6 mutation by rpb2t, we used a strain in which the BUR6 promoter was replaced by the GAL10 promoter (PGAL10-BUR6), rendering cells unable to grow in glucose, since the BUR6 gene is essential for cell growth (Kim et al. 1997). As shown in Figure 1A, overexpression of rpb2t allowed the PGAL10-BUR6 strain to grow in dextrose. The same effect was observed in a strain in which the gene encoding the second, also essential, component of the NC2 repressor, YDR1, was placed under the control of the GAL10 promoter (Figure 1A) or in a strain carrying both PGAL10-BUR6 and PGAL10-YDR1 genes (result not shown). However, when we transformed cells carrying a chromosomal deletion of BUR6 and the wild-type BUR6 gene on an URA3/CEN plasmid with YEp-rpb2t, the resulting cells were unable to grow on 5′-FOA-containing plates (results not shown), indicating that overexpression of rpb2t is unable to relieve the requirement for the essential function of NC2α. The overexpression of rpb2t was also unable to suppress the growth defect of a thermosensitive bur6 strain at 37° (results not shown). Together, these results suggest that a threshold amount of wild-type NC2 is required for the rpb2t-mediated suppression of lethality.
Like NC2, Mot1p appears to repress transcription by acting through DNA-bound TBP (Pugh 2000; Pereira et al. 2003). We therefore investigated whether overexpression of rpb2t also suppresses the growth defect of Mot1-depleted cells. Figure 1B shows that when MOT1 was placed under the control of the GAL10 promoter, the strain was unable to grow on plates containing dextrose as the sole carbon source. However, when this strain was transformed with YEp-rpb2t, growth on dextrose-containing plates was observed.
Depletion of NC2 can be bypassed by the overexpression of a variety of truncations in RPB2:
In the genetic screen where rpb2t was identified, we isolated six independent clones harboring plasmids carrying this allele and no other truncations in the RPB2 ORF (F. Estruch, L. Peiró-Chova and C. Cole, unpublished results). To determine whether the suppressor activity was specific for this truncation, we cloned different restriction fragments of the RPB2 gene in the multicopy vector YEplac181, and the resulting plasmids were used to transform the strain carrying the PGAL10-YDR1 allele. As shown in Figure 2, multiple truncated versions of RPB2, including as few as 240 amino acids (from a total length of 1224 aminoacids), are able to suppress the growth defect of the PGAL10-YDR1 allele in dextrose, but truncated versions containing only the first 25 or 119 amino acids cannot. The longest truncation contains 943 amino acids and could suppress the growth defect, whereas overexpression of the full-length RPB2 gene could not (Figure 2). We did not observe suppression when the rpb2t allele was cloned in a centromeric plasmid (results not shown).
The suppression of the NC2 defect by overexpression of rpb2t is not due to the squelching of Rpb9p, Rpb12p, or Sin4p by the truncated protein:
The N-terminal region of Rpb2p forms the domains of RNA pol II known as the “protrusion” and the “lobe” and includes regions known to interact with Rpb12p and Rpb9p (Cramer et al. 2001). On the other hand, mutations in the mediator component Sin4p bypass the requirement for NC2 (Lemaire et al. 2000). We analyzed the possibility that suppression of the NC2 defect by Rpb2t could be due to functional depletion of one of these proteins by simultaneously overexpressing rpb2t, and RPB9, RPB12, or SIN4 in the PGAL10-YDR1 strain. We examined whether these strains could grow in media containing dextrose as the carbon source. In all cases, double transformants were still able to grow in dextrose (results not shown), suggesting that the squelching of Rpb9, Rpb12, or Sin4 is not responsible for, or at least is not the only cause of, the suppression of the growth defect associated with the reduced activity of NC2 by overexpression of rpb2t.
Overexpression of rpb2t counterbalances the transcriptional defects caused by the reduced activity of NC2:
As mentioned above, depletion of either of the two components of NC2 stops cell growth. We have analyzed the genomewide transcriptional effects caused by the depletion of NC2 by comparing the transcriptional profile of a PGAL10-YDR1 strain with the isogenic wild-type strain after a 4-hr incubation in dextrose-containing medium (complete data sets at http://scsie.uv.es/chipsdna/chipsdna-e.html#datos). A total of 509 genes showed at least twofold change in their expression levels in the PGAL10-YDR1 strain, as compared to the wild-type control. Depletion of YDR1 results mainly in increased transcript levels (414 genes vs. only 95 genes exhibiting decreased expression), suggesting that NC2 has a mostly negative effect on gene expression. To assess the effect of the overexpression of rpb2t on gene expression in the PGAL10-YDR1 strain, we compared the ratios of expression (relative to the isogenic wild-type strain) in the PGAL10-YDR1 strain transformed with YEp-rpb2t or with the empty plasmid YEplac181. Figure 3 shows that, when rpb2t is overexpressed, most of the genes (88%) whose expression is increased or decreased by more than twofold in the PGAL10-YDR1 strain (compared to the wild type) are restored to expression levels similar to those observed in the wild-type control. The overexpression of rpb2t in a wild-type strain has a less pronounced effect on gene expression, with only 94 transcripts displaying expression changes by a factor of twofold or more (see complete data sets). Therefore, overexpression of rpb2t compensates for the overall effect on gene expression caused by the depletion of NC2.
Isolation of mutants that suppress the growth defect caused by depletion of NC2:
To gain further insight into the function of NC2, we performed a genetic screen for suppressors of the NC2 defect. In this screen, we used a yeast strain in which the genes encoding both NC2 components (YDR1 and BUR6) were placed under the control of the GAL10 promoter. Transposon insertion mutants were generated in this strain using a yeast genomic library mutagenized by the insertion of an mTn3-lacZ/LEU2 transposon (Burns et al. 1994). Suppressor mutants were selected by their ability to grow on dextrose-containing plates at room temperature, and sites of transposon insertion were determined. A total of 17 suppressors, which were named snd (suppressor of NC2 depletion), were identified. Ten of these mutations were mapped to genes encoding previously characterized components of the basic transcription machinery (Table 2) and are analyzed in this work.
The suppression of NC2 depletion by snd mutations is not due to the constitutive expression of the GAL10 promoter:
Since the genetic screen was performed in a strain in which the GAL10 promoter controlled the expression of both BUR6 and YDR1, the ability to grow in media containing dextrose as the sole carbon source may be due to the defective repression of the GAL10 promoter in the snd mutants. To check this possibility, first we analyzed whether the mutants isolated in our screen were able to suppress the growth defect caused by the depletion of an unrelated protein expressed under the control of the GAL10 promoter. We used a strain in which the ROT1 gene, encoding an ER-localized gene involved in actin cytoskeleton dynamics (J. C. Igual, personal communication), was placed under the control of the GAL10 promoter. Accordingly, this strain is unable to grow in glucose, since the ROT1 gene is essential for cell growth. None of the mutations isolated in our screen conferred the ability of the PGAL10-ROT1 strain to grow in dextrose (results not shown). Next, we analyzed the levels of Bur6 and Ydr1 proteins in the snd mutants, taking advantage of the three copies of the HA epitope added to the N-terminal end of the proteins during the substitution of the BUR6 and YDR1 promoters by GAL10 (Longtine et al. 1998). In the Western analysis presented in Figure 4, we compare the amounts of Bur6p and Ydr1p (expressed from the GAL10 promoter) in the snd mutants and the strain overexpressing rpb2t with the amount observed in the isogenic SND wild-type strain before and after a 5-hr incubation in dextrose. None of the mutants showed increased levels of the Bur6 or Ydr1 proteins, indicating that the suppression is not a consequence of the defective repression of the GAL10 promoter in dextrose. We also noted that in the snd1-2/nut2-2 and snd2-1/tfb1-1 mutants (in a PGAL10-BUR6 PGAL10-YDR1 background), the levels of Bur6 and Ydr1 proteins in galactose were lower than in the control strain, suggesting that these mutations could be defective in the induction of the GAL10 promoter. The defective induction of the GAL10 promoter could be the reason for the marked slow-growth phenotype observed for the snd1-2/nut2 and the snd2-1/tfb1 mutant in galactose (results not shown). We found that these strains grew even better in dextrose than in galactose. The remaining mutations did not significantly impair the growth in galactose of the PGAL10-BUR6 PGAL10-YDR1 mutant strain (results not shown).
snd mutants can bypass the requirement for Bur6 and Ydr1:
To assess if defects in the genes isolated in our screen were able to bypass the otherwise essential requirement of NC2, we introduced mutant alleles of these genes in a strain carrying a chromosomal deletion of BUR6 and the wild-type BUR6 gene on an URA3/CEN plasmid. Defects in the essential genes (NUT2, MED7, TFB1, and RPB7) were created by replacing their promoters with the doxycycline-repressible tetO promoter, whereas a complete deletion was used for the nonessential SRB2 gene. The ability to suppress the need for BUR6 was tested by streaking the strains onto 5′-FOA plates (for Δsrb2) or 5′-FOA plates containing doxycycline (for PtetO-NUT2, PtetO-MED7, PtetO-RPB7, and PtetO-TFB1). As positive controls, we introduced a deletion in the previously characterized suppressor sin4 (Lemaire et al. 2000), and a centromeric plasmid carrying the wild-type BUR6 using LEU2 as the selection marker. As observed in Figure 5A, whereas loss of the pBUR6/URA3 plasmid prevents growth of the unmodified control strain, deletion of SRB2 and depletion of Nut2, Med7, Tfb1, or Rpb7 allow, to different extents, cell growth in the absence of Bur6. Similarly, the requirement of Ydr1 could also be bypassed by mutating components of the Mediator (SRB2), TFIIH (TFB1), and the RNA pol II (RPB7), revealing that defects in these components of the basic transcription machinery can bypass the essential requirement for NC2 (Figure 5B). Moreover, these results further support the previous results showing that the suppression caused by the transposon insertion mutants in the NC2-depletion strains is not a consequence of their effect on the GAL10 promoter (Figure 4).
Different mutations in mediator subunits suppress depletion of NC2:
The mutated locus identified most often in the suppressor screen corresponds to the NUT2/MED10 gene (Table 2). Eight independent mutants were isolated and six different points of transposon insertion were mapped. Since NUT2 is an essential gene, it is surprising that two of the insertions (snd1-2 and snd1-9/snd1-50) are located in the coding region at positions that are likely to produce a nonfunctional Nut2 product. These mutants showed extremely poor growth in both galactose and dextrose (result not shown) and we failed to obtain single nut2-1/snd1-2 by backcrossing the PGAL10-BUR6 PGAL10-YDR1 nut2-1∷mTn-lacZ/LEU2 triple mutant with a wild-type strain, suggesting that strains carrying the nut2-1 are viable only when the activity of NC2 is compromised. To check if the otherwise essential NUT2 becomes dispensable when NC2 is depleted, we used a yeast strain in which the wild-type NUT2 promoter was substituted by the doxycycline-repressible tetO promoter. Although it has been described that this change of promoter causes a constitutive slow-growth phenotype (Mnaimneh et al. 2004), we could detect only a slight impairment of cell growth at all doxycycline concentrations tested (Figure 6A). When we introduced the PGAL10-YDR1 allele replacing the wild-type YDR1 gene in this strain, growth of the double-mutant PtetO-NUT2 PGAL10-YDR1 in the absence of deoxycycline was indistinguishable from growth observed for the PGAL10-YDR1 single-mutant strain on both galactose and dextrose (results not shown). However, repression of the tetO promoter by deoxycycline has a positive effect on the growth of the Ptet0-NUT2 PGAL10-YDR1 double mutant in dextrose, and this effect is stronger at higher deoxyclycline concentrations (Figure 6A).
The isolation of mutations in three different mediator subunits as suppressors of the NC2-dependent growth defects prompted us to investigate whether mutations in other mediator components could have the same effect. We introduced the PGAL10-YDR1 or the PGAL10-BUR6 allele in each nonessential mediator subunit mutant and analyzed the ability to suppress the depletion of NC2 by evaluating the ability to grow in media containing dextrose as the carbon source. Figure 6B shows that mutations in all the nonessential mediator subunits, except in MED2, MED3, and GAL11, are able to suppress, to different extents, the depletion of Bur6 and Ydr1. The ability to suppress the NC2 depletion likely also involves the ability to bypass the absence of NC2. Thus, deletion of the NUT1 and SRB10 genes, but not of MED3, was able to bypass the Bur6 requirement (results not shown). Mutations in med2, med3, and gal11 not only are unable to suppress the NC2 growth defect in dextrose, but also exacerbate it, abolishing the reduced level of growth shown by the PGAL10-YDR1 and PGAL10-YDR1 mutant strains in dextrose (Figure 6B). Interestingly, mutations in the three subunits that negatively effect growth of NC2-depleted cells are components of the mediator tail, and their mutations show similar expression profiles, very different from that observed for the other nonessential tail subunit Sin4 (van de Peppel et al. 2005). Deletion of SIN4 has a clear effect as a suppressor of NC2-depleted cell growth (Figure 6B).
The antagonistic relationship between the Gal11/Med2/Med3 triad and the other nonessential mediator subunits with NC2 was further investigated by combining mutations with opposite effects on NC2-depleted strains. Figure 6C shows that the compensatory effect of SRB2 deletion on Ydr1 depletion is completely abolished by the mutation in the mediator subunit Med3. Moreover, the mutation in the triad component MED2 also avoids the positive effect on the growth of the PGAL10-YDR1 mutant strain in dextrose caused by the depletion of the TFIIH component Tfb1 or the RNA pol II component Rpb7 (Figure 6C).
Defects on the RNA pol II subunits suppress the NC2 depletion:
The transposon insertional mutagenesis approach also identified rpb7 as a suppressor of NC2 depletion defects (Table 2). Although RPB7 is an essential gene, the insertion of the transposon in the snd3-1 mutant occurs 129 nucleotides after the stop codon, and it is likely to cause a reduction of RPB7 expression. The reduced Rpb7 level caused by the snd3-1 would be in agreement with the fact that a reduced amount of Rpb7 can bypass the essential requirement of Bur6 (Figure 5). As described above, the overexpression of truncated forms of Rpb2 also suppresses the growth defects caused by the depletion of NC2. These results suggest that, in general, defects in RNA pol II could compensate for the depletion of NC2. To further analyze this, we subcloned a C-terminal truncated form of the largest subunit of RNA pol II, Rpb1, in the multicopy vector YEplac181. The truncated form (referred to as rpb1t) contains the N-terminal 844 amino acids (from a total length of 1733 residues). As observed in Figure 7A, overexpression of rpb1t has a positive effect on the growth of the PGAL10-YDR1 strain in dextrose-containing media, although weaker than the overexpresson of rpb2t. However, deletion of the nonessential subunit Rpb9 is unable to suppress the growth defect of the PGAL10-YDR1 strain in dextrose, having a negative effect on growth (Figure 7A). As observed for rpb2t, overexpression of rpb1t cannot bypass the essential requirement for Bur6 (result not shown).
In the experiments overexpressing Rpb1 and Rpb2 truncations, cells also contain the wild-type genes. To check if a reduction of the amount of Rpb1 or Rpb2 is also able to compensate for the depletion of NC2, we used strains in which the promoters of the RPB1 or RPB2 have been substituted by the regulatable tetO promoter. Figure 7B shows that the substitution of RPB1 and RPB2 wild-type promoters with the regulatable tetO promoters has detectable effects on cell growth only at high concentrations of doxycycline. The presumed reduction of RPB1 and RPB2 gene expression alleviates the requirement for NC2 (Figure 7B) without increasing the Ydr1 protein levels in dextrose (results not shown). Unexpectedly, the suppression was stronger when low concentrations of doxycycline were used. Since the repression of the tetO promoter is proportional to the doxycycline concentration (Gari et al. 1997), this result could indicate that suppression of NC2 depletion requires a moderate reduction of the Rpb1 and Rpb2 amounts, but very low levels of these proteins are unable to suppress the NC2-depletion defect, even if they can still support cell grow (Figure 7B).
Mutations in TFIIH and TFIIB, but not in SAGA and TFIID components, suppress the depletion of NC2:
The transposon mutagenesis screen also identified TFB1 as a target of mutations that are able to suppress the NC2 depletion defect (see Table 2). The TFB1 gene encodes an essential component of TFIIH (Gileadi et al. 1992). The transposon insertion was mapped to the C terminus of the TFB1 open reading frame, suggesting that the suppression is consequence of the production of a truncated form of Tfb1p. This mutation severely impairs growth, especially when galactose is used as the carbon source (results not shown). The slow growth in galactose could be related to the defective induction of the GAL promoter (Figure 4). To assess if mutations in other TFIIH components are also able to suppress the NC2 depletion defects, we introduced the PGAL10-YDR1 alleles in strains carrying mutations in the TFIIH components SSL1 and RAD3 (Feaver et al. 1993). Figure 8 shows that the ssl1-1 mutation, but not rad3-20, is able to suppress the growth defect caused by the depletion of Ydr1. Suppression is not a consequence of an increased level of Ydr1 protein in dextrose in the ssl1-1 mutant (result not shown).
During the course of a related study, we tagged TFIIB (encoded by the SUA7 gene) with the myc epitope in a PGAL10-YDR1 mutant strain. Surprisingly, the tagged TFIIB is able to suppress the PGAL10-YDR1 defect in dextrose-containing media (Figure 8), suggesting that the introduction of the myc epitope results in a partially defective TFIIB and that defects in TFIIB can suppress the depletion of NC2.
Finally, we checked whether mutations in other functions involved in transcription initiation, not identified in our screening, were able to suppress the NC2 depletion defect. Accordingly, we introduced the PGAL10-YDR1 allele in a strain carrying a complete deletion of the GCN5 or ADA2 gene, or a thermosentitive allele of TAF1, encoding a TFIID subunit. As observed in Figure 8, none of these mutations are able to suppress the growth defect of the PGAL10-YDR1 in dextrose.
In previous studies, we established a genetic interaction between the NC2-component Bur6 and mRNA export. More specifically, mutations in BUR6 were observed to be synthetic lethal with dbp5-2, yra1-1 or sub2-85 mutant alleles (Estruch and Cole 2003; F. Estruch, L. Peiró-Chova and C. Cole, unpublished results). To gain more information about the causes of the genetic relationships between BUR6 and genes encoding mRNA export factors, we selected for genes that, upon overexpression, could bypass the synthetic lethality of the bur6-1 and dbp5-2 mutations. This genetic screen identified a truncated version of Rpb2 (Rpb2t) as a multicopy suppressor of both the bur6-1 dbp5-2 and the mot1-301 dbp5-2 double-mutant strain (F. Estruch, L. Peiró-Chova and C. Cole, unpublished results). Interestingly, overexpression of different truncated forms of Rpb2 acted as a suppressor of the growth defect caused by a limiting amount of NC2 or Mot1, like that provided by the expression of the PGAL10-BUR6, PGAL10-YDR1, or PGAL10-MOT1 alleles when dextrose was used as the carbon source (Figure 1). It seems likely that the presence of a truncated Rpb2t causes a reduction in transcriptional activity, as a result of the nonfunctional RNA pol II holoenzymes (those containing the truncated Rpb2t) competing with the functional ones. According to this premise, our results suggest that, although NC2 could have both positive and negative roles in transcription (Prelich 1997; Geisberg et al. 2001; Cang and Prelich 2002), an essential function of NC2 results from its role as a transcriptional repressor. The microarray experiments support a largely negative effect of NC2 on transcription, since depletion in the NC2-component Ydr1 results in increased transcript levels for 414 genes, whereas only 95 genes exhibit decreased expression. Experiments comparing the transcription profile of cells depleted in Ydr1 and the same cells overexpressing Rpb2t indicate that overexpression of Rpb2t could counteract the global transcriptional effects of the depletion of NC2. Accordingly, nearly 90% of the genes whose expression is increased or decreased by more than twofold by the depletion of Ydr1 (as compared to the wild type) show levels of expression close to the wild-type value when rpb2t is overexpressed (Figure 3). The fact that overexpression of rpb2t causes an increase in the level of expression of 77 of the 95 genes whose transcription is negatively affected by the depletion of NC2 suggests that the effect of NC2 on these genes is mostly indirect, further supporting that the major and essential function of NC2 in vivo results from its role as transcriptional repressor.
As mentioned above, the most likely mechanism for the NC2 suppression by the overexpression of rpb2t and rpb1t is the competition between functional and nonfunctional RNA pol II holoenzymes. The suppressor activity observed for fragments that include as few as 240 amino acids, representing <20% of the full-length protein, is surprising. This result suggests that the N-terminal portion of Rpb2 can adopt structure(s) able to interact with other holoenzyme components. The N terminus of Rpb2 includes sequences that interact with Rbp12 and Rpb9 in the RNA pol II holoenzyme (Cramer et al. 2001), but the individual sequestering of any of these subunits does not seem to be the reason for the suppression. The fact that a reduction in transcriptional activity can compensate for the requirement of NC2 is also supported by the suppression observed when Rpb1 or Rpb2 are depleted (Figure 7B). However, our results show that the highest level of suppression is obtained when the reduction in the amount of these proteins is moderate. On the other hand, a defective RNA pol II activity, like that resulting from the absence of the nonessential subunit Rpb9, does not suppress, but exacerbates the growth defects observed for the PGAL10-YDR1 strain (Figure 7A). Taken together, these results reveal the necessity of a fine balance between specific positive and negative transcription functions to support cell growth.
Taking advantage of the inability to grow on dextrose observed for the PGAL10-YDR1 PGAL10-BUR6 double-mutant strain, we selected for other mutations that suppress the requirement for NC2. Most of the mutations identified in our screening correspond to components of three complexes with basic roles in transcription: mediator, TFIIH, and RNA pol II (Table 2). The remaining mutations correspond to genes of unknown function that are being characterized (L. Peiró and F. Estruch, unpublished results). The mechanism of suppression by these mutations is not related to a partial derepression of the GAL10 promoter, as shown by the analysis of the amount of HA-Bur6 and HA-Ydr1 proteins expressed by these mutants when dextrose was used as the carbon source (Figure 4). In addition, we have shown that defects in all the genes identified in our screen are able to bypass the essential requirement of Bur6 and Ydr1, validating the use of the PGAL10-YDR1 and PGAL10-BUR6 strains as tools for identifying genes functionally related to NC2.
Previous studies identified mutations in the mediator component SIN4 and in the general transcription factor TFIIA as suppressors of ydr1 and bur6 mutations (Kim et al. 2000; Lemaire et al. 2000; Xie et al. 2000). Sin4, together with Rgr1, Med2, Gal11, and Med3, forms the mediator tail module (Chadick and Asturias 2005). The observations that mutations in GAL11 could not suppress the cold-sensitive phenotype shown by ydr1 mutant strains (Kim et al. 2000) and that mutations in MED3, MED2, and RGR1 were unable to bypass the NC2 requirement (Lemaire et al. 2000) led to the interpretation that the suppression was due to a specific genetic interaction between NC2 and SIN4, and not derived from the alteration of the tail module structure. Our screen identified mutations in the mediator components NUT2, MED7, and SRB2 as suppressors of NC2 depletion, and we have found that these mutations can also bypass the absence of Bur6. By themselves, the screening results suggest an involvement of the mediator as a whole in a transcriptional process that counterbalances the function of NC2, arguing against a specific genetic relationship between NC2 and a particular mediator subunit. Moreover, when we extended our analysis to all the nonessential subunits of the mediator, we found that 11 of the 14 subunits tested were able to suppress, to different extents, the depletion of NC2. Importantly, the only three subunits for which mutations fail to suppress the defect caused by the depletion of NC2 (Med2, Med3, and Gal11) form a distinctive group within the mediator tail module. Analysis of the expression profiles shows that these three subunits have a positive role in gene expression (van de Peppel et al. 2005) and, therefore, one could expect that this positive role in transcription would make mutations in MED2, MED3, and GAL11 candidates for suppressing the depletion of NC2. As mentioned above, Med2, Med3, and Gal11 are part of the mediator tail module and do not seem to contact pol II, but interact with gene-specific activators (Lewis and Reinberg 2003), raising the possibility that the Med2/Med3/Gal11 triad would not be directly involved in basal transcription. The possibility raised by Cang and Prelich (2002) that NC2 acts by selectively repressing basal transcription but stimulates activated transcription could explain why the deletion of these positive components further impairs the growth defect of NC2 mutant strains (Figure 6B) and the results of the epistasis analysis (Figure 6C). Moreover, the observation that the Med2/Med3/Gal11 triad is required for the suppression caused by a reduction in the activity of TFIIH or RNA pol II (Figure 6C) could indicate a specific interaction between NC2 and these mediator components.
In addition to mutations in genes encoding RNA pol II and mediator components, our screening (and additional experiments included in this work) has identified mutations in the TFIIH components Tfb1 and Ssl1 as suppressors of the depletion of NC2. We were, however, unable to observe suppression by mutating genes encoding SAGA and TFIID components (Gcn5, Ada2, and Taf1), although we cannot exclude that mutations in other components of these complexes could act as suppressors.
It has been suggested that NC2 could function as both a positive and a negative effector of transcription (Geisberg et al. 2001; Chitikila et al. 2002). The nature of some suppressors identified in this work strongly suggests that the growth arrest of cells depleted in NC2 results from an excessive rate of basal transcription and that suppressors would function by reducing it. Taken together, our results suggest that NC2 participates in a continuous and dynamic competition with positive factors, likely TFIIA and TFIIB. This competition would control the extent of the RNA pol II recruitment to the promoter and the level of gene expression.
We thank A. Aguilera, P. Alepuz, A. Bailis, S. Buratowski, C. Cole, G. Prelich, M. Hampsey, D. Reinberg, R. Sendra, and R. Young for providing plasmids and yeast strains; L. Yenush for critically reading the manuscript; and J. Garcia-Martinez for microarray data analysis. This work was supported by grants from the Spanish Ministry of Education and Science (BFU2004-00069/BMC) and Generalitat Valenciana (GV04B-113 and GVACOMP2006-027) to F.E. L.P-C. is a recipient of a fellowship from the Formación de Personal Investigador program of the Spanish Ministry of Education and Science.
Communicating editor: M. Hampsey
- Received October 13, 2006.
- Accepted February 12, 2007.
- Copyright © 2007 by the Genetics Society of America