PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of molcellbPermissionsJournals.ASM.orgJournalMCB ArticleJournal InfoAuthorsReviewers
 
Mol Cell Biol. 2005 December; 25(24): 11171–11183.
PMCID: PMC1316967

Activator Gcn4p and Cyc8p/Tup1p Are Interdependent for Promoter Occupancy at ARG1 In Vivo

Abstract

The Cyc8p/Tup1p complex mediates repression of diverse genes in Saccharomyces cerevisiae and is recruited by DNA binding proteins specific for the different sets of repressed genes. By screening the yeast deletion library, we identified Cyc8p as a coactivator for Gcn4p, a transcriptional activator of amino acid biosynthetic genes. Deletion of CYC8 confers sensitivity to an inhibitor of isoleucine/valine biosynthesis and impairs activation of Gcn4p-dependent reporters and authentic amino acid biosynthetic target genes. Deletion of TUP1 produces similar but less severe activation defects in vivo. Although expression of Gcn4p is unaffected by deletion of CYC8, chromatin immunoprecipitation assays reveal a strong defect in binding of Gcn4p at the target genes ARG1 and ARG4 in cyc8Δ cells and to a lesser extent in tup1Δ cells. The defects in Gcn4p binding and transcriptional activation in cyc8Δ cells cannot be overcome by Gcn4p overexpression but are partially suppressed in tup1Δ cells. The impairment of Gcn4p binding in cyc8Δ and tup1Δ cells is severe enough to reduce recruitment of SAGA, Srb mediator, TATA binding protein, and RNA polymerase II to the ARG1 and ARG4 promoters, accounting for impaired transcriptional activation of these genes in both mutants. Cyc8p and Tup1p are recruited to the ARG1 and ARG4 promoters, consistent with a direct role for this complex in stimulating Gcn4p occupancy of the upstream activation sequence (UAS). Interestingly, Gcn4p also stimulates binding of Cyc8p/Tup1p at the 3′ ends of these genes, raising the possibility that Cyc8p/Tup1p influences transcription elongation. Our findings reveal a novel coactivator function for Cyc8p/Tup1p at the level of activator binding and suggest that Gcn4p may enhance its own binding to the UAS by recruiting Cyc8p/Tup1p.

Cyc8 and Tup1 are evolutionary conserved proteins that function in a complex to mediate repression of diverse sets of genes in Saccharomyces cerevisiae, including those expressed only in haploids or in MATα cells or in response to DNA damage, osmotic stress, or hypoxic (low-oxygen) conditions (42). The Cyc8p/Tup1p complex is targeted to promoters by DNA binding proteins specific for the different classes of repressed genes (42). Whereas Cyc8p is crucial for recruitment of Cyc8p/Tup1p by the Mig1 and Rox1 repressors (45, 46), Tup1p makes contact with the α2 repressor (19). It is thought that Tup1p provides the main repressor functions of the complex, because tethering a LexA-Tup1 fusion protein to a promoter led to transcriptional repression in cyc8 mutant cells (46), whereas repression by LexA-Cyc8p was dependent on Tup1p (4).

It has been proposed that Cyc8p/Tup1p can repress transcription by organizing nucleosomes into repressive chromatin structures (5, 12, 26). This function could be mediated by direct interaction of Tup1p with the N-terminal tails of histones H3 and H4 (11, 18) and nonhistone chromosomal protein Nhp6 (24). Other findings suggest that repression by Tup1p occurs by deacetylation of histone N-terminal tails at the promoter (1, 7, 8), which might involve direct recruitment of histone deacetylases by Cyc8p/Tup1p (6, 48, 49). A third repression mechanism was suggested by observations that Cyc8p/Tup1p-mediated repression is partially impaired by mutations in subunits of Srb mediator (henceforth referred to as mediator) (reference 25 and references therein). The mediator is a multisubunit cofactor that forms a holoenzyme complex with RNA polymerase II (Pol II) and other general transcription factors and is implicated in activation as well as repression (31). There is evidence that mediator subunits Srb7p, Srb10p, and Hrs1p/Med3p/Pgd1p are targets of Tup1 in the mediator. As mediator is required for optimal recruitment of TATA binding protein (TBP) and Pol II by various activators (23, 27, 39), interaction of Cyc8p/Tup1p with mediator subunits could interfere with assembly of the preinitiation complex (PIC).

It appears that none of these repression mechanisms alone can account for Tup1p-mediated repression. Mutations in H3, H4, and Nhp6p increase expression of only a subset of Cyc8p/Tup1p target genes, and repression of certain promoters in wild-type (WT) cells occurs in the absence of positioned nucleosomes (42). In addition, repression of ANB1 (30) and RNR3 (53) was unaffected by mutations in TUP1 and ISW2, respectively, that eliminate nucleosome positioning at these genes. In fact, nucleosome positioning, histone deacetylation, and mediator subunits make overlapping contributions to Tup1p-mediated repression of RNR3, as successive inactivation of each mechanism led to stepwise derepression of this gene (53).

Interestingly, Tup1p/Cyc8p can participate in overcoming its own repressive function at certain target genes. Thus, phosphorylation of transcription factor Sko1p during osmotic stress overcomes the repressing function of the promoter-bound Sko1p-Cyc8p/Tup1p complex, leading to Tup1p-dependent recruitment of SAGA and SWI/SNF and attendant transcription of Sko1p target genes. As deletion of TUP1 did not reduce transcription of these genes under inducing conditions, it appears that recruitment of SWI/SNF and SAGA is required primarily to overcome the repressor function of Cyc8p/Tup1p (38). Similarly, Cyc8p/Tup1p remains bound at the GAL1 promoter under galactose-inducing conditions and Cyc8p recruits SAGA to the promoter via Cti6p/Rxt1p (35). Again, deletion of CYC8 does not impair galactose induction of GAL1, suggesting that Cti6p/Rxt1p-dependent recruitment of SAGA serves to counteract Cyc8p/Tup1p repression at GAL1.

There is also evidence that Cyc8p can function as a conventional coactivator at CYC1 (52) and SUC2 (34), where gene induction is impaired in cyc8 mutant cells. Cyc8p/Tup1p plays a dual role at CIT2, with Tup1p mediating repression and Cyc8p supporting activation by Rtg3p. Indeed, Cyc8p/Tup1p was recruited by Rtg3p to the CIT2 promoter in vitro, and the Rtg3p activation domain can directly interact with Cyc8p (4). Cyc8p and, to a lesser extent, Tup1p are also required for activation of FRE2 by Aft1p. Cyc8p interacts directly with Aft1p in vitro and is recruited by Aft1p to the FRE2 promoter in vivo, where it mediates nucleosome remodeling of the promoter in conjunction with Nhp6p (13).

We have been analyzing the coactivator requirements for Gcn4p, a transcriptional activator of amino acid biosynthetic genes in yeast (33) that is induced at the translational level by starvation for any amino acid (16). The Gcn4p activation domain interacts specifically in vitro with mediator, SAGA, and the ATP-dependent chromatin remodeling complexes SWI/SNF and RSC (10, 32, 43, 47), and Gcn4p recruits all four coactivators to target genes in living cells (21, 22, 43, 44). Furthermore, mutations in various subunits of these coactivators confer sensitivity to inhibitors of amino acid biosynthetic enzymes (Gcn phenotype) and reduce transcriptional activation of one or more Gcn4p-dependent target genes or reporters in vivo (39; reference 43 and references therein). Recent findings indicate that SAGA, SWI/SNF, and mediator arrive simultaneously at Gcn4p target promoters under inducing conditions but are highly interdependent for their recruitment by Gcn4p (15, 40). These coactivators stimulate both PIC assembly and one or more steps in transcription elongation at the Gcn4p target gene ARG1 (15, 39).

In this report, we show that Cyc8p/Tup1p is a coactivator for Gcn4p at multiple target genes in vivo, and we present evidence that Cyc8p/Tup1p functions to enhance Gcn4p binding to the upstream activation sequence (UAS) elements at ARG1 and ARG4. Interestingly, Gcn4p recruits Cyc8p/Tup1p to both genes, which may provide a positive-feedback mechanism to maintain high-level Gcn4p binding at the UAS elements of these genes in vivo.

MATERIALS AND METHODS

Yeast strains and plasmids.

All strains and plasmids used in this study are listed in Tables Tables11 and and2,2, respectively. The WT parent strains BY4741 (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0) and BY4742 (MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0) and deletion derivatives thereof were described previously (14) and purchased from Research Genetics. The presence of all reported deletion alleles was confirmed by PCR amplification or complementation of mutant phenotypes by plasmid-borne wild-type genes (43). myc-tagged strains were constructed as previously described (43). The presence of the myc-tagged alleles was verified by colony PCR and Western blot analysis using anti-Myc antibodies.

TABLE 1.
Yeast strains used in this study
TABLE 2.
Plasmids used in this study

Plasmids p1208, p2382, pHYC2, pKN7, pSK1, pHQ1239, pHQ1240, and pHQ1303 were described previously (9, 17, 32, 43, 50), and the empty vector employed throughout was the URA3 CEN4 plasmid YCp50 (36). The BamHI-MluI fragment from pHQ1239 containing GCN4-myc13 was subcloned into the high-copy-number URA3 vector YEplac195 to produce pHQ1293. Plasmids pME2126 and pME2129 were described previously (37).

Biochemical methods.

The reporter gene assays were performed as described previously (43). For Western analysis, whole-cell extracts (WCEs) were prepared as described previously (41) and analyzed using monoclonal anti-myc (Roche) and polyclonal anti-Gcd6p (3) antibodies. Northern analysis was carried out as described previously (43), with the following modification. QuikHyb hybridization solution (Stratagene) was used for prehybridization and hybridization as described by the vendor. After hybridization, the membranes were washed twice with 2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) buffer and 0.1% sodium dodecyl sulfate (SDS) at room temperature for 15 min and once with 0.1× SSC buffer-0.1% SDS at 60°C for 30 min. The washed membranes were subjected to phosphorimaging analysis for quantifying the signals and also to autoradiography. The chromatin immunoprecipitation (ChIP) experiments were conducted as described previously, using the same primers described there (39, 40, 43, 50). The coimmunoprecipitation and glutathione S-transferase (GST) pulldown assays shown in the supplemental material were conducted essentially as described previously (51).

RESULTS

Cyc8p is broadly required for wild-type activation of Gcn4p target genes.

We screened the library of haploid deletion strains produced by the Saccharomyces Genome Deletion Project (14) for mutants with increased sensitivity to sulfometuron methyl (SM), an inhibitor of isoleucine and valine biosynthesis, to identify new mutants defective in transcriptional activation by Gcn4p. The cyc8Δ strain was found to be highly sensitive to SM (SMs), although it was less sensitive than the isogenic gcn4Δ mutant in the deletion library. The tup1Δ strain is also SMs but less so than the cyc8Δ mutant (Fig. (Fig.1A).1A). We verified by PCR analysis of genomic DNA that these two mutants contain the reported deletions. We also determined that the cyc8Δ/cyc8Δ and tup1Δ/tup1Δ mutants in the homozygous diploid deletion library exhibit SMs phenotypes comparable to those observed for the corresponding haploid mutants (data not shown). Finally, the SMs phenotype of the haploid cyc8Δ mutant was diminished by introducing a plasmid-borne CYC8 allele into this strain (data not shown). We conclude that inactivation of CYC8 or TUP1 confers sensitivity to SM.

FIG. 1.
Deletion of CYC8 confers sensitivity to an inhibitor of isoleucine/valine biosynthesis and impairs induction of Gcn4p-dependent reporter genes. (A) The WT strain (BY4741) and isogenic gcn4Δ (249), cyc8Δ (7161), and tup1Δ (7198) ...

Consistent with its strong SMs phenotype, the cyc8Δ mutant showed a dramatic defect in activation of a Gcn4p-dependent lacZ reporter containing two Gcn4p binding sites upstream of the CYC1 core promoter (UASGCRE-CYC1-lacZ) (Fig. 1B and C). This reporter is induced more than 10-fold in the WT strain when Gcn4p is induced by SM. Induction of the reporter was nearly eliminated in the cyc8Δ mutant, which exhibits only slightly higher reporter expression than does the gcn4Δ mutant under inducing conditions (4% versus 1% of WT, respectively). By contrast, the cyc8Δ mutant showed >80% of WT UASGCRE-CYC1-lacZ expression under noninducing conditions, even though deleting GCN4 eliminates most of the reporter expression in nonstarved cells (Fig. 1B and C). Thus, Cyc8p is required primarily to support the high-level transcription of UASGCRE-CYC1-lacZ evoked by induced levels of Gcn4p produced under starvation conditions.

Cyc8p is also required for efficient activation of a HIS3-GUS reporter containing the intact HIS3 promoter, as induction of this reporter occurred at only ~30% of the WT level in cyc8Δ cells (Fig. 1D and E). Most of the HIS3-GUS expression under noninducing conditions is independent of Gcn4p. Interestingly, cyc8Δ led to an approximately sevenfold derepression of HIS3-GUS expression under noninducing conditions. These findings suggest that Cyc8p is required to repress basal HIS3 promoter function under noninducing conditions even though it is necessary for strong activation of this promoter by induced levels of Gcn4p.

Consistent with its weaker SMs phenotype, the tup1Δ mutant showed only a modest defect in activation of the UASGCRE-CYC1-lacZ reporter and no defect in HIS3-GUS expression under inducing conditions. However, tup1Δ derepressed the UASGCRE-CYC1-lacZ and HIS3-GUS reporters by approximately threefold and fivefold, respectively, under noninducing conditions (Fig. 1B to E). It seems that Tup1p plays only a negative role in regulating the transcription of these reporter genes, being required to maintain low-level expression under noninducing conditions.

Somewhat different results were obtained when we compared the effects of cyc8Δ and tup1Δ mutations on mRNAs produced from the chromosomal Gcn4p target genes ARG1 and ARG4. Northern analysis of the steady-state levels of these mRNAs relative to the Pol III transcript scR1 indicated that cyc8Δ impairs ARG1 and ARG4 expression under both inducing and noninducing conditions (Fig. 2A and B). Similar results were obtained for ARG1 and ARG4 and extended to include the Gcn4p targets HIS4 and TRP3 in cyc8Δ cells, using the Pol II transcript ACT1 as an internal control (Fig. (Fig.2C).2C). Thus, cyc8Δ impairs transcriptional activation by Gcn4p at multiple authentic target genes. Interestingly, tup1Δ led to reductions in the induced levels of ARG1 and ARG4 mRNAs comparable to those observed for cyc8Δ (Fig. 2A and B), suggesting that Tup1p plays a positive role in activation of these chromosomal target genes even though it is dispensable for activation of the plasmid-borne reporters. Furthermore, tup1Δ did not elicit significant derepression of ARG1 or ARG4 mRNAs under noninducing conditions (Fig. 2A and B).

FIG. 2.
Deletions of CYC8 and TUP1 impair induction of Gcn4p target genes. Total RNA was isolated from strains described in the legend for Fig. Fig.1,1, grown under the same inducing (I) and noninducing (N) conditions described there and subjected to ...

The results presented thus far suggest that Cyc8p is broadly required for wild-type activation by Gcn4p, whereas Tup1p contributes more selectively at specific Gcn4p-dependent promoters. Tup1p and Cyc8p also contribute to repression of plasmid-borne reporter genes under noninducing conditions. It is unclear whether the distinct effects of cyc8Δ and tup1Δ on expression of reporter genes versus authentic target genes reflect differences in the promoter sequences involved or, rather, an atypical chromatin structure of the plasmid-borne reporters.

Cyc8p is required for efficient binding of Gcn4p to target genes in vivo.

To determine whether the activation defects observed with cyc8Δ and tup1Δ mutants result from reduced binding of Gcn4p at its target genes, we first analyzed the effects of cyc8Δ on GCN4 expression. As shown in Fig. Fig.2D,2D, we observed no difference in GCN4 mRNA levels between WT and cyc8Δ cells under inducing or noninducing conditions. We next measured the steady-state expression of a myc-tagged form of Gcn4p by Western analysis. The GCN4-myc allele is nearly indistinguishable from untagged GCN4 in complementing the SMs phenotype of the gcn4Δ mutant (data not shown). Western blot analysis revealed little difference in the induced levels of myc-Gcn4p between transformants of gcn4Δ and cyc8Δ gcn4Δ strains harboring GCN4-myc on single-copy (s.c.) or high-copy-number (h.c.) plasmids (Fig. (Fig.3A).3A). It is noteworthy that the myc-Gcn4p level in cyc8Δ cells containing h.c. GCN4-myc is considerably higher than that in gcn4Δ cells containing s.c. GCN4-myc, (Fig. (Fig.3A,3A, cf. lanes 11 and 12 with lanes 5 and 6) yet overexpressing Gcn4p from an h.c. plasmid did not suppress the SMs phenotype of the cyc8Δ mutant. Thus, the activation defect in cyc8Δ cells does not appear to result from reduced accumulation of Gcn4p in starved cells.

FIG. 3.
Deletion of CYC8 does not reduce myc-Gcn4p abundance, and the SMs phenotype of cyc8Δ cells is not overcome by overexpressing Gcn4p. Western analysis of myc-Gcn4p expression. (A) gcn4Δ (249) and cyc8Δ gcn4Δ (KS4) strains ...

We next asked whether binding of Gcn4p to its target promoters is impaired by the cyc8Δ mutation by conducting ChIP analysis. Remarkably, after 2 h of induction by SM, binding to the ARG1 UAS of native Gcn4p was greatly reduced in the cyc8Δ mutant to a level only slightly above the nonspecific background signal measured for the gcn4Δ strain (Fig. 4A and B, vector lanes). Moreover, overexpressing functional hemagglutinin (HA)-tagged Gcn4p increased binding at ARG1 by ~2.2-fold in WT cells but had no effect on Gcn4p binding in the cyc8Δ mutant (Fig. 4A and B). Binding of Gcn4p at the ARG1 UAS increased between 30 min and 2 h of induction with SM in WT cells, and cyc8Δ impaired Gcn4p binding at all three time points assayed in this time course (Fig. (Fig.4C).4C). The cyc8Δ mutation also impaired UAS binding of myc-tagged Gcn4p, whether expressed in gcn4Δ cells from an s.c. or an h.c. plasmid (Fig. (Fig.4D).4D). Finally, cyc8Δ impaired Gcn4p binding to the UAS at ARG4 to roughly the same extent observed at ARG1 in a manner unaffected by overexpressing HA-Gcn4p (Fig. (Fig.4E4E).

FIG. 4.
Deletion of CYC8 and TUP1 impair binding of Gcn4p to the UAS elements at ARG1 and ARG4. Strains BY4741 (WT), 249 (gcn4Δ), 7161 (cyc8Δ), and 7189 (tup1Δ) were transformed with empty vector (YCp50), s.c. GCN4-HA plasmid p2382, or ...

The tup1Δ mutation had a somewhat different effect on Gcn4p binding at its target genes. Whereas tup1Δ strongly reduced Gcn4p binding at the ARG1 and ARG4 UAS elements (Fig. 4F and G), this defect was overcome by overexpressing HA-Gcn4p from an h.c. plasmid (Fig. (Fig.4F).4F). Thus, Tup1p may be less critical than Cyc8p for efficient binding of Gcn4p at its target genes.

Recruitment of coactivators, TBP, and Pol II by Gcn4p is strongly impaired in cyc8Δ cells.

We wished to determine whether the reduced UAS occupancy by Gcn4p in cyc8Δ cells observed with ChIP assays is sufficient to impair recruitment of SAGA, Srb mediator, TBP, and Pol II to the promoter after 2 h of induction. The ChIP analyses in Fig. Fig.55 show that cyc8Δ reduces recruitment of myc-tagged subunits of SAGA (Spt7p) (Fig. 5A and B) and mediator (Srb6p) (Fig. 5C and D) to the UAS elements at ARG1 and ARG4 to nearly the same extent as does gcn4Δ, even in cells overexpressing HA-Gcn4p from a high-copy-number plasmid. The recruitment of myc-tagged forms of TBP and the Rpb3p subunit of Pol II to the ARG1 promoter also was impaired by cyc8Δ (Fig. 5E and F). The reduction of TBP recruitment to the promoter and of Pol II occupancy in the ARG1 open reading frame (ORF) were evident in the cyc8Δ mutant immediately on induction of Gcn4p in parallel with decreased UAS binding of Gcn4p at each time point analyzed (Fig. 6A to C). We showed previously that the myc-tagged alleles of SPT7, SRB6, SPT15/TBP1, and RPB3 are functional for transcriptional activation by Gcn4p (39, 43), and we verified here that cyc8Δ and tup1Δ do not reduce steady-state expression of myc-Spt7p, myc-Srb6p, myc-TBP, or myc-Rpb3p (see Fig. S1 in the supplemental material). Thus, the reduced UAS occupancy of Gcn4p in cyc8Δ cells is associated with substantial defects in coactivator recruitment and PIC assembly by Gcn4p.

FIG. 5.
Recruitment of SAGA (myc-Spt7p), Srb mediator (myc-Srb6p), TBP (myc-TBP1), and Pol II (myc-Rpb3p) by Gcn4p is strongly dependent on Cyc8p and moderately dependent on Tup1p. ChIP analysis was conducted on the following strains harboring empty vector (Ycp50) ...
FIG. 6.
Kinetic ChIP analysis. The WT strain (BY4741) and cyc8Δ strain (7161) bearing h.c. plasmid GCN4-HA pHQ1239 were cultured in synthetic complete medium lacking Ile and Val containing 1 μg/ml SM for the indicated times and subjected to ChIP ...

We also observed impaired recruitment of coactivators and PIC assembly at Gcn4p target genes in the tup1Δ mutant. Recruitment of myc-Spt7p and myc-Srb6p to the UAS elements at ARG1 and ARG4 and of myc-TBP and myc-Rpb3p binding to the ARG1 promoter were all reduced in tup1Δ cells. With the possible exception of myc-Rpb3p, these recruitment defects were suppressed by overexpressing HA-Gcn4p, such that the levels of factor recruitment in tup1Δ cells bearing h.c. GCN4-HA were comparable to the levels in TUP1 cells expressing Gcn4p at the native level (Fig. 5A to F). These results are consistent with the fact that overexpressing Gcn4p greatly reduced the SMs phenotype of the tup1Δ mutant (Fig. (Fig.3B)3B) and restored WT Gcn4p binding to the ARG1 UAS in tup1Δ cells (Fig. (Fig.4F4F).

Gcn4p recruits Cyc8p and Tup1p to target genes in vivo.

In an effort to determine whether Cyc8p and Tup1p function in a direct manner to promote Gcn4p binding to the UAS, we asked whether Cyc8p and Tup1p are present at ARG1 and ARG4 under inducing conditions by conducting ChIP analysis of strains expressing myc-tagged forms of these proteins. We verified that the CYC8-myc and TUP1-myc alleles conferred wild-type growth on SM plates (data not shown) and thus appeared to be functional in supporting transcriptional activation by Gcn4p. The ChIP results shown in Fig. 7B and C reveal Gcn4p-dependent binding of myc-Cyc8p and Tup1p to the ARG1 UAS after 30 min or 1 h of induction, in parallel with the increased binding of Gcn4p at the UAS observed at these time points (Fig. (Fig.7A).7A). In the strain containing native levels of Gcn4p (GCN4/vector), binding of both myc-Cyc8p and Tup1p to the UAS under inducing conditions was only twofold or less above the background level observed for the gcn4Δ strain. However, much higher levels of myc-Cyc8p and Tup1p recruitment to the UAS occurred in strains overexpressing HA-Gcn4p (GCN4/h.c. GCN4-HA). Similar levels of Gcn4p-dependent binding of myc-Cyc8p and myc-Tup1p were observed for the ARG4 UAS (data not shown). Western analysis showed that gcn4Δ does not reduce the steady-state levels of myc-Cyc8p or myc-Tup1p (see Fig. S1 in the supplemental material). Hence, we conclude that Gcn4p recruits Tup1p and Cyc8p to the ARG1 and ARG4 UAS elements under inducing conditions, consistent with the idea that these proteins function directly as coactivators at these Gcn4p target genes.

FIG. 7.
Cyc8p/Tup1p is associated with the UAS and coding regions of Gcn4p target genes ARG1 and ARG4 under inducing conditions. Strains KS6 (WT) and KS32 (gcn4Δ) carrying empty vector (YCp50) or h.c. GCN4-HA plasmid pHQ1239 were cultured for the indicated ...

Surprisingly, we observed relatively high levels of myc-Cyc8p binding at the 3′ ends of the ARG1 and ARG4 ORFs in both the WT strain and the strain overexpressing HA-Gcn4p (Fig. 7D and F). Thus, there appears to be greater Gcn4p-dependent Cyc8p binding to the ORF than to the UAS (cf. Fig. 7B and D). By contrast, Tup1p association with the ARG1 ORF was more similar in magnitude to that observed at the UAS (cf. Fig. 7C and E), and comparable findings were obtained for Tup1p at ARG4 (Fig. (Fig.7G7G and data not shown). These results raise the possibility that Cyc8p and Tup1p influence transcription elongation in the ORF in addition to promoting Gcn4p binding to the UAS elements at these genes.

DISCUSSION

The results of this study show that Cyc8p and, to a lesser extent, Tup1p are required for high-level activation of Gcn4p target genes in vivo. Deletion of CYC8 confers sensitivity to SM and impairs transcriptional induction of two Gcn4p-dependent reporters (UASGCRE-CYC1-lacZ and HIS3-GUS) and authentic Gcn4p target genes (ARG1, ARG4, HIS4, and TRP3). ChIP analysis revealed that cyc8Δ cells exhibited diminished recruitment of SAGA and Srb mediator to the UAS, reduced promoter occupancy of TBP and Pol II, and decreased Pol II association with the 3′ ends of the ORFs at ARG1 and ARG4. The tup1Δ mutant showed similar defects in coactivator recruitment, PIC assembly, and transcriptional activation of authentic target genes but was less sensitive to SM and did not significantly impair induction of the reporter genes. Thus, it appears that Cyc8p is more broadly required than Tup1p for transcriptional activation by Gcn4p.

It is possible that all of the defects in transcriptional activation of Gcn4p-dependent promoters observed with cyc8Δ and tup1Δ cells can be attributed to the decreased Gcn4p binding to the UAS elements at its target genes in these mutants. This decrease in UAS occupancy by Gcn4p cannot be attributed to reduced expression of GCN4 mRNA or decreased steady-state levels of Gcn4p protein in cyc8Δ or tup1Δ cells. No other mutations in subunits of SAGA, SWI/SNF, Srb mediator, or RSC that impair transcriptional activation by Gcn4p have been found to reduce Gcn4p binding at ARG1 or ARG4 (15, 39). Thus, reduced Gcn4p promoter occupancy in cells containing WT levels of Gcn4p is a unique phenotype of cyc8Δ and tup1Δ mutants.

One way to account for the role of Cyc8p and Tup1p in Gcn4p binding to UAS elements is to propose that they are required for efficient nuclear import of Gcn4p. For example, expression of one or more factors involved in nuclear import of Gcn4p could be reduced in cyc8Δ and tup1Δ cells. It was shown previously that a functional green fluorescent protein (GFP)-Gcn4 fusion protein is localized primarily in the nucleus independently of amino acid abundance (37). Thus, nuclear localization of Gcn4p is probably not regulated by amino acid levels. We have compared levels of localization of this GFP-Gcn4p fusion for cyc8Δ versus WT cells and found only a small decrease in the percentage of cells (from 95% to 79%) containing GFP-Gcn4p that was localized exclusively in the nucleus in cyc8 δ cells (see Fig. S2 in the supplemental material). Thus, we believe that Cyc8p/Tup1p has no significant role in nuclear localization of Gcn4p.

A direct role for Cyc8p/Tup1p in promoting UAS binding by Gcn4p is suggested by our finding that both proteins are recruited by Gcn4p to its target genes in vivo. This raises the possibility that the Cyc8p/Tup1p complex modifies the chromatin structure of the UAS to increase accessibility of Gcn4p to its binding sites, perhaps by positioning nucleosomes in an organized array (5, 12, 26, 53) or by recruiting histone deacetylases to reduce histone acetylation (1, 6-8, 48, 49). Another possibility is that Cyc8p/Tup1p impedes degradation of UAS-bound Gcn4p. This idea is prompted by the finding that Srb10p, a component of the mediator, phosphorylates the Gcn4p activation domain and targets the protein for rapid degradation by the proteosome (2, 20, 29). Hence, recruitment of the mediator by Gcn4p could accelerate the degradation of UAS-bound Gcn4p without affecting the turnover of unbound Gcn4p (2). According to this last model, Cyc8p/Tup1p recruited to the UAS by Gcn4p would decrease the rate of Srb10p-mediated degradation of UAS-bound Gcn4p. To account for the fact that total steady-state levels of Gcn4p are not reduced by cyc8Δ, it would be necessary to stipulate that UAS-bound Gcn4p represents only a small fraction of the total cellular pool of the protein and that Gcn4p degraded at the UAS in cyc8Δ cells is quickly replenished by new synthesis.

Whatever the mechanism involved, it seems that Cyc8p plays a more important role than Tup1p in promoting Gcn4p binding to UAS elements. Thus, in addition to the less severe activation defects in tup1Δ cells versus cyc8Δ cells, we found that overexpressing Gcn4p can alleviate the Gcn4p binding defect in tup1Δ cells but not in cyc8Δ cells. Presumably, Gcn4p binding to the UAS can be driven to nearly WT levels in tup1Δ cells by increasing the cellular concentration of Gcn4p, but the impediment to UAS binding is too great to overcome by mass action in cyc8Δ cells.

It is interesting that high-level UAS occupancy of Gcn4p is dependent on Cyc8p and Tup1p, while at the same time, Gcn4p recruits Cyc8p and Tup1p. Thus, Gcn4p and Cyc8p/Tup1p are interdependent for high-level binding at ARG1 and ARG4. Although the reduction in UAS occupancy by Gcn4p in cyc8Δ cells is substantial, it is not complete (e.g., Fig. Fig.4C).4C). Indeed, the SM sensitivity and reductions in reporter and target gene transcription are less severe in cyc8Δ cells than in gcn4Δ cells. Hence, it seems that Gcn4p can bind to a UAS element lacking associated Cyc8p, albeit with reduced efficiency. By subsequently recruiting Cyc8p/Tup1p, Gcn4p may achieve a higher level of UAS occupancy, thereby creating a positive-feedback loop (Fig. (Fig.8).8). We cannot rule out the possibility, however, that Cyc8p/Tup1p binds at ARG1 independently of Gcn4p at a level below the detection limit of the ChIP assay and then promotes binding of Gcn4p to the unoccupied UAS. Once bound to the UAS, Gcn4p would then recruit Cyc8p/Tup1p to the higher levels observed under inducing conditions with our ChIP assays.

FIG. 8.
Model for interdependent binding of activator Gcn4p and Cyc8p/Tup1p at ARG1 and ARG4 in vivo. Cyc8p/Tup1p functions to stimulate Gcn4p binding to the UAS elements at ARG1and ARG4. Interestingly, Gcn4p recruits Cyc8p/Tup1p to the promoter and ORF regions ...

We have done several experiments to determine whether Gcn4p recruits Cyc8p/Tup1p to target genes by direct binding to these cofactors but have uncovered no evidence to support this possibility. First, in cells where myc-tagged Gcn4p was induced by SM, we observed no coimmunoprecipitation of Tup1p with myc-Gcn4p from cell extracts, whereas Tup1p coimmunoprecipitated with myc-Cyc8p under the same conditions (see Fig. S3 in the supplemental material). We also failed to observe a specific interaction of GST-Gcn4p with myc-Tup1p or myc-Cyc8p in yeast WCEs under conditions where robust binding of GST-Gcn4p to the mediator was readily detected (see Fig. S3 in the supplemental material). We presume that Gcn4p recruits Cyc8p/Tup1p to the UAS and ORF of its target genes in a manner dependent on other coactivators or histone modifications, and future experiments will address the mechanism of this indirect recruitment, particularly as it applies to the coding sequences.

Most previous studies of Cyc8p and Tup1p have underscored their functions as corepressors (42), but the complex can also function in activation. As noted above, Cyc8p/Tup1p remains bound at promoters regulated by Sko1p (38) and at GAL1 (35) under inducing conditions and participates in recruitment of coactivators to overcome its own repressing functions. In the Cyc8p-dependent induction of CIT2 by Rtg3p (4), it is unknown whether Cyc8p stimulates UAS occupancy by Rtg3p or provides a coactivator function for UAS-bound Rtg3p. The exact step in activation of CYC1 by Hap1p that is dependent on Tup1p/Cyc8p is also unknown (52). Regarding Cyc8p-dependent activation of FRE2 by Aft1p, Cyc8p binds in vitro to the DNA binding domain of Aft1p, and Aft1p binding at FRE2 was apparently reduced by a small amount in cyc8Δ cells. Nevertheless, activation by a LexA-Aft1p fusion from LexA binding sites was strongly Cyc8p dependent, pointing to a role for Cyc8p in the activation function of Aft1p rather than in Aft1p promoter binding (13). Thus, the prominent role of Cyc8p in stimulating Gcn4p binding to the UAS elements at ARG1 and ARG4 described here appears to be unique among the yeast activators studied thus far.

We obtained one indication that Cyc8p/Tup1p also enhances the ability of Gcn4p to recruit coactivators. In strains overexpressing Gcn4p from the h.c. GCN4-HA plasmid, there was essentially the same level of Gcn4p binding to the ARG1 UAS in WT and tup1Δ cells (Fig. (Fig.4F);4F); however, the tup1Δ cells showed a relatively lower level of SAGA and mediator recruitment to the UAS and also less myc-Rpb3p recruitment to the promoter (Fig. 5A, C, and F). Hence, Cyc8p/Tup1p may play a dual role at Gcn4p target genes, increasing the efficiency of coactivator recruitment by Gcn4p in addition to stimulating Gcn4p binding to the UAS.

Supplementary Material

[Supplemental material]

Acknowledgments

We thank Gerhard Braus for GFP-GCN4 plasmids, Joe Reese for antibodies against TBP and Tup1p, Sang-Yeob Yeo for help with fluorescence microscopy, and Fan Zhang and other members of our laboratory for protocols and many helpful suggestions.

Footnotes

Supplemental material for this article may be found at http://mcb.asm.org/.

REFERENCES

1. Bone, J. R., and S. Y. Roth. 2001. Recruitment of the yeast Tup1p-Ssn6p repressor is associated with localized decreases in histone acetylation. J. Biol. Chem. 276:1808-1813. [PubMed]
2. Chi, Y., M. J. Huddleston, X. Zhang, R. A. Young, R. S. Annan, S. A. Carr, and R. J. Deshaies. 2001. Negative regulation of Gcn4 and Msn2 transcription factors by Srb10 cyclin-dependent kinase. Genes Dev. 15:1078-1092. [PubMed]
3. Cigan, A. M., M. Foiani, E. M. Hannig, and A. G. Hinnebusch. 1991. Complex formation by positive and negative translational regulators of GCN4. Mol. Cell. Biol. 11:3217-3228. [PMC free article] [PubMed]
4. Conlan, R. S., N. Gounalaki, P. Hatzis, and D. Tzamarias. 1999. The Tup1-Cyc8 protein complex can shift from a transcriptional co-repressor to a transcriptional co-activator. J. Biol. Chem. 274:205-210. [PubMed]
5. Cooper, J. P., S. Y. Roth, and R. T. Simpson. 1994. The global transcriptional regulators, SSN6 and TUP1, play distinct roles in the establishment of a repressive chromatin structure. Genes Dev. 8:1400-1410. [PubMed]
6. Davie, J. K., D. G. Edmondson, C. B. Coco, and S. Y. Dent. 2003. Tup1-Ssn6 interacts with multiple class I histone deacetylases in vivo. J. Biol. Chem. 278:50158-50162. [PubMed]
7. Davie, J. K., R. J. Trumbly, and S. Y. Dent. 2002. Histone-dependent association of Tup1-Ssn6 with repressed genes in vivo. Mol. Cell. Biol. 22:693-703. [PMC free article] [PubMed]
8. Deckert, J., and K. Struhl. 2001. Histone acetylation at promoters is differentially affected by specific activators and repressors. Mol. Cell. Biol. 21:2726-2735. [PMC free article] [PubMed]
9. Drysdale, C. M., E. Dueñas, B. M. Jackson, U. Reusser, G. H. Braus, and A. G. Hinnebusch. 1995. The transcriptional activator GCN4 contains multiple activation domains that are critically dependent on hydrophobic amino acids. Mol. Cell. Biol. 15:1220-1233. [PMC free article] [PubMed]
10. Drysdale, C. M., B. M. Jackson, R. McVeigh, E. R. Klebanow, Y. Bai, T. Kokubo, M. Swanson, Y. Nakatani, P. A. Weil, and A. G. Hinnebusch. 1998. The Gcn4p activation domain interacts specifically in vitro with RNA polymerase II holoenzyme, TFIID, and the Adap-Gcn5p coactivator complex. Mol. Cell. Biol. 18:1711-1724. [PMC free article] [PubMed]
11. Edmondson, D. G., M. M. Smith, and S. Y. Roth. 1996. Repression domain of the yeast global repressor Tup1 interacts directly with histones H3 and H4. Genes Dev. 10:1247-1259. [PubMed]
12. Fleming, A. B., and S. Pennings. 2001. Antagonistic remodelling by Swi-Snf and Tup1-Ssn6 of an extensive chromatin region forms the background for FLO1 gene regulation. EMBO J. 20:5219-5231. [PubMed]
13. Fragiadakis, G. S., D. Tzamarias, and D. Alexandraki. 2004. Nhp6 facilitates Aft1 binding and Ssn6 recruitment, both essential for FRE2 transcriptional activation. EMBO J. 23:333-342. [PubMed]
14. Giaever, G., A. M. Chu, L. Ni, C. Connelly, L. Riles, S. Veronneau, S. Dow, A. Lucau-Danila, K. Anderson, B. Andre, A. P. Arkin, A. Astromoff, M. El-Bakkoury, R. Bangham, R. Benito, S. Brachat, S. Campanaro, M. Curtiss, K. Davis, A. Deutschbauer, K. D. Entian, P. Flaherty, F. Foury, D. J. Garfinkel, M. Gerstein, D. Gotte, U. Guldener, J. H. Hegemann, S. Hempel, Z. Herman, D. F. Jaramillo, D. E. Kelly, S. L. Kelly, P. Kotter, D. LaBonte, D. C. Lamb, N. Lan, H. Liang, H. Liao, L. Liu, C. Luo, M. Lussier, R. Mao, P. Menard, S. L. Ooi, J. L. Revuelta, C. J. Roberts, M. Rose, P. Ross-Macdonald, B. Scherens, G. Schimmack, B. Shafer, D. D. Shoemaker, S. Sookhai-Mahadeo, R. K. Storms, J. N. Strathern, G. Valle, M. Voet, G. Volckaert, C. Y. Wang, T. R. Ward, J. Wilhelmy, E. A. Winzeler, Y. Yang, G. Yen, E. Youngman, K. Yu, H. Bussey, J. D. Boeke, M. Snyder, P. Philippsen, R. W. Davis, and M. Johnston. 2002. Functional profiling of the Saccharomyces cerevisiae genome. Nature 418:387-391. [PubMed]
15. Govind, C. K., S. Yoon, H. Qiu, S. Govind, and A. G. Hinnebusch. 2005. Simultaneous recruitment of coactivators by Gcn4p stimulates multiple steps of transcription in vivo. Mol. Cell. Biol. 25:5626-5638. [PMC free article] [PubMed]
16. Hinnebusch, A. G. 1996. Translational control of GCN4: gene-specific regulation by phosphorylation of eIF2, p. 199-244. In J. W. B. Hershey, M. B. Mathews, and N. Sonenberg (ed.), Translational control. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
17. Hinnebusch, A. G., G. Lucchini, and G. R. Fink. 1985. A synthetic HIS4 regulatory element confers general amino acid control on the cytochrome c gene CYC1 of yeast. Proc. Natl. Acad. Sci. USA 82:498-502. [PubMed]
18. Huang, L., W. Zhang, and S. Y. Roth. 1997. Amino termini of histones H3 and H4 are required for a1-α2 repression in yeast. Mol. Cell. Biol. 17:6555-6562. [PMC free article] [PubMed]
19. Komachi, K., and A. D. Johnson. 1997. Residues in the WD repeats of Tup1 required for interaction with α2. Mol. Cell. Biol. 17:6023-6028. [PMC free article] [PubMed]
20. Kornitzer, D., B. Raboy, R. G. Kulka, and G. R. Fink. 1994. Regulated degradation of the transcription factor Gcn4. EMBO J. 13:6021-6030. [PubMed]
21. Kuo, M. H., E. vom Bauer, K. Struhl, and C. D. Allis. 2000. Gcn4 activator targets Gcn5 histone acetyltransferase to specific promoters independently of transcription. Mol. Cell 6:1309-1320. [PubMed]
22. Kuo, M. H., J. Zhou, P. Jambeck, M. E. Churchill, and C. D. Allis. 1998. Histone acetyltransferase activity of yeast Gcn5p is required for the activation of target genes in vivo. Genes Dev. 12:627-639. [PubMed]
23. Kuras, L., and K. Struhl. 1999. Binding of TBP to promoters in vivo is stimulated by activators and requires Pol II holoenzyme. Nature 399:609-613. [PubMed]
24. Laser, H., C. Bongards, J. Schuller, S. Heck, N. Johnsson, and N. Lehming. 2000. A new screen for protein interactions reveals that the Saccharomyces cerevisiae high mobility group proteins Nhp6A/B are involved in the regulation of the GAL1 promoter. Proc. Natl. Acad. Sci. USA 97:13732-13737. [PubMed]
25. Lee, M., S. Chatterjee, and K. Struhl. 2000. Genetic analysis of the role of Pol II holoenzyme components in repression by the Cyc8-Tup1 corepressor in yeast. Genetics 155:1535-1542. [PubMed]
26. Li, B., and J. C. Reese. 2001. Ssn6-Tup1 regulates RNR3 by positioning nucleosomes and affecting the chromatin structure at the upstream repression sequence. J. Biol. Chem. 276:33788-33797. [PubMed]
27. Li, X., A. Virbasius, X. Zhu, and M. R. Green. 1999. Enhancement of TBP binding by activators and general transcription factors. Nature 399:605-609. [PubMed]
28. Longtine, M. S., A. McKenzie III, D. J. Demarini, N. G. Shah, A. Wach, A. Brachat, P. Philippsen, and J. R. Pringle. 1998. Additional modules for versatile and economical PCR-based gene deletion and modification in Saccharomyces cerevisiae. Yeast 14:953-961. [PubMed]
29. Meimoun, A., T. Holtzman, Z. Weissman, H. J. McBride, D. J. Stillman, G. R. Fink, and D. Kornitzer. 2000. Degradation of the transcription factor Gcn4 requires the kinase Pho85 and the SCFCDC4 ubiquitin-ligase complex. Mol. Biol. Cell 11:915-927. [PMC free article] [PubMed]
30. Mennella, T. A., L. G. Klinkenberg, and R. S. Zitomer. 2003. Recruitment of Tup1-Ssn6 by yeast hypoxic genes and chromatin-independent exclusion of TATA binding protein. Eukaryot. Cell 2:1288-1303. [PMC free article] [PubMed]
31. Myers, L. C., and R. D. Kornberg. 2000. Mediator of transcriptional regulation. Annu. Rev. Biochem. 69:729-749. [PubMed]
32. Natarajan, K., B. M. Jackson, E. Rhee, and A. G. Hinnebusch. 1998. yTAFII61 has a general role in RNA polymerase II transcription and is required by Gcn4p to recruit the SAGA coactivator complex. Mol. Cell 2:683-692. [PubMed]
33. Natarajan, K., M. R. Meyer, B. M. Jackson, D. Slade, C. Roberts, A. G. Hinnebusch, and M. J. Marton. 2001. Transcriptional profiling shows that Gcn4p is a master regulator of gene expression during amino acid starvation in yeast. Mol. Cell. Biol. 21:4347-4368. [PMC free article] [PubMed]
34. Ozcan, S., L. G. Vallier, J. S. Flick, M. Carlson, and M. Johnston. 1997. Expression of the SUC2 gene of Saccharomyces cerevisiae is induced by low levels of glucose. Yeast 13:127-137. [PubMed]
35. Papamichos-Chronakis, M., T. Petrakis, E. Ktistaki, I. Topalidou, and D. Tzamarias. 2002. Cti6, a PHD domain protein, bridges the Cyc8-Tup1 corepressor and the SAGA coactivator to overcome repression at GAL1. Mol. Cell 9:1297-1305. [PubMed]
36. Parent, S. A., C. M. Fenimore, and K. A. Bostian. 1985. Vector systems for the expression, analysis and cloning of DNA sequences in S. cerevisiae. Yeast 1:83-138. [PubMed]
37. Pries, R., K. Bomeke, S. Irniger, O. Grundmann, and G. H. Braus. 2002. Amino acid-dependent Gcn4p stability regulation occurs exclusively in the yeast nucleus. Eukaryot. Cell 1:663-672. [PMC free article] [PubMed]
38. Proft, M., and K. Struhl. 2002. Hog1 kinase converts the Sko1-Cyc8-Tup1 repressor complex into an activator that recruits SAGA and SWI/SNF in response to osmotic stress. Mol. Cell 9:1307-1317. [PubMed]
39. Qiu, H., C. Hu, S. Yoon, K. Natarajan, M. J. Swanson, and A. G. Hinnebusch. 2004. An array of coactivators is required for optimal recruitment of TATA binding protein and RNA polymerase II by promoter-bound Gcn4p. Mol. Cell. Biol. 24:4104-4117. [PMC free article] [PubMed]
40. Qiu, H., C. Hu, F. Zhang, G. J. Hwang, M. J. Swanson, C. Boonchird, and A. G. Hinnebusch. 2005. Interdependent recruitment of SAGA and Srb mediator by transcriptional activator Gcn4p. Mol. Cell. Biol. 25:3461-3474. [PMC free article] [PubMed]
41. Reid, G. A., and G. Schatz. 1982. Import of proteins into mitochondria. J. Biol. Chem. 257:13062-13067. [PubMed]
42. Smith, R. L., and A. D. Johnson. 2000. Turning genes off by Ssn6-Tup1: a conserved system of transcriptional repression in eukaryotes. Trends Biochem. Sci. 25:325-330. [PubMed]
43. Swanson, M. J., H. Qiu, L. Sumibcay, A. Krueger, S.-J. Kim, K. Natarajan, S. Yoon, and A. G. Hinnebusch. 2003. A multiplicity of coactivators is required by Gcn4p at individual promoters in vivo. Mol. Cell. Biol. 23:2800-2820. [PMC free article] [PubMed]
44. Syntichaki, P., I. Topalidou, and G. Thireos. 2000. The Gcn5 bromodomain co-ordinates nucleosome remodeling. Nature 404:414-417. [PubMed]
45. Tzamarias, D., and K. Struhl. 1995. Distinct TPR motifs of Cyc8 are involved in recruiting the Cyc8-Tup1 corepressor complex to differentially regulated promoters. Genes Dev. 9:821-831. [PubMed]
46. Tzamarias, D., and K. Struhl. 1994. Functional dissection of the yeast Cyc8-Tup1 transcriptional co-repressor complex. Nature 369:758-761. [PubMed]
47. Utley, R. T., K. Ikeda, P. A. Grant, J. Cote, D. J. Steger, A. Eberharter, S. John, and J. L. Workman. 1998. Transcriptional activators direct histone acetyltransferase complexes to nucleosomes. Nature 394:498-502. [PubMed]
48. Watson, A. D., D. G. Edmondson, J. R. Bone, Y. Mukai, Y. Yu, W. Du, D. J. Stillman, and S. Y. Roth. 2000. Ssn6-Tup1 interacts with class I histone deacetylases required for repression. Genes Dev. 14:2737-2744. [PubMed]
49. Wu, J., N. Suka, M. Carlson, and M. Grunstein. 2001. TUP1 utilizes histone H3/H2B-specific HDA1 deacetylase to repress gene activity in yeast. Mol. Cell 7:117-126. [PubMed]
50. Yoon, S., H. Qiu, M. J. Swanson, and A. G. Hinnebusch. 2003. Recruitment of SWI/SNF by Gcn4p does not require Snf2p or Gcn5p but depends strongly on SWI/SNF integrity, SRB mediator, and SAGA. Mol. Cell. Biol. 23:8829-8845. [PMC free article] [PubMed]
51. Zhang, F., L. Sumibcay, A. G. Hinnebusch, and M. J. Swanson. 2004. A triad of subunits from the Gal11/tail domain of Srb mediator is an in vivo target of transcriptional activator Gcn4p. Mol. Cell. Biol. 24:6871-6886. [PMC free article] [PubMed]
52. Zhang, L., and L. Guarente. 1994. Evidence that TUP1/SSN6 has a positive effect on the activity of the yeast activator HAP1. Genetics 136:813-817. [PubMed]
53. Zhang, Z., and J. C. Reese. 2004. Ssn6-Tup1 requires the ISW2 complex to position nucleosomes in Saccharomyces cerevisiae. EMBO J. 23:2246-2257. [PubMed]

Articles from Molecular and Cellular Biology are provided here courtesy of American Society for Microbiology (ASM)