In Saccharomyces cerevisiae, optimal utilization of various compounds as a nitrogen source is mediated by a complex transcriptional network. The zinc cluster protein Dal81 is a general activator of nitrogen metabolic genes, including those for γ-aminobutyrate (GABA). In contrast, Uga3 (another zinc cluster protein) is an activator restricted to the control of genes involved in utilization of GABA. Uga3 binds to DNA elements found in the promoters of target genes and increases their expression in the presence of GABA. Dal81 appears to act as a coactivator since the DNA-binding activity of this factor is dispensable but its mode of action is not known. In this study, we have mapped a regulatory, as well as an activating, region for Uga3. A LexA–Uga3 chimeric protein activates a lexA reporter in a GABA- and Dal81-dependent manner. Activation by Uga3 requires the SAGA complex as well as Gal11, a component of mediator. ChIP analysis revealed that Uga3 is weakly bound to target promoters. The presence of GABA enhances binding of Uga3 and allows recruitment of Dal81 and Gal11 to target genes. Recruitment of Gal11 is prevented in the absence of Dal81. Importantly, Dal81 by itself is a potent activator when tethered to DNA and its activity depends on SAGA and Gal11 but not Uga3. Overexpression of Uga3 bypasses the requirement for Dal81 but not for SAGA or Gal11. Thus, under artificial conditions, both Dal81 and Uga3 can activate transcription independently of each other. However, under physiological conditions, both factors cooperate by targeting common coactivators.
MANY unicellular organisms can use various nitrogen-containing compounds as a nitrogen source. In Saccharomyces cerevisiae, the favored sources of nitrogen are ammonium, glutamine, and asparagine. S. cerevisiae can also grow on nonpreferred nitrogen sources, including proline, urea, ornithine, allantoin, and γ-aminobutyrate (GABA). These nonpreferred nitrogen sources are converted into glutamate and glutamine. Nonpreferred nitrogen sources induce the derepression of many genes, which are involved in the utilization of these compounds and are unexpressed when a preferred nitrogen source is present (Godardet al. 2007). In S. cerevisiae, four GATA proteins are responsible for regulating the expression of these nitrogen catabolic genes. Gln3 and Gat1 are activators that induce the transcription of nitrogen catabolic genes in the presence of a nonfavored nitrogen source, while Dal80 and Deh1 are transcriptional repressors (Wonget al. 2008).
The amino acid derivative GABA serves as a nonpreferred source of nitrogen in S. cerevisiae (Magasanik and Kaiser 2002). GABA can be degraded into succinate semialdehyde through the action of the GABA transaminase (encoded by the UGA1 gene). This enzymatic reaction requires α-ketoglutarate or pyruvate, which is converted into glutamate or alanine, respectively. Succinate semialdehyde can be further degraded into succinate by the enzyme succinate semialdehyde dehydrogenase (encoded by UGA2), simultaneously catalyzing the production of NADPH from NADP (Colemanet al. 2001). Succinate can then be fed into the Krebs cycle for further energy production (Godardet al. 2007). Another important protein involved in the utilization of GABA is the permease Uga4, which mediates the specific uptake of GABA in S. cerevisiae. GABA is also able to enter S. cerevisiae nonspecifically through the proline permease Put4 and through the general amino acid permease Gap1 (Andréet al. 1993). Genome-wide studies have shown that expression of only five genes is induced by GABA, including the UGA1 and UGA4 genes (Godardet al. 2007). Expression of these two genes is regulated in part by GATA factors described above (Talibiet al. 1995; Marzluf 1997). In addition, the transcriptional activation of UGA1 and UGA4 depends on Dal81 (also called Uga35) and Uga3 (André 1990; Visserset al. 1990; Coornaertet al. 1991; Talibiet al. 1995).
Dal81 and Uga3 are members of the family of zinc cluster proteins that form a major class of transcriptional regulators in S. cerevisiae (MacPhersonet al. 2006). Zinc cluster proteins are typically composed of three domains: a DNA-binding domain containing a highly conserved cysteine-rich cluster, a regulatory domain, and an activation domain. The DNA-binding domain of most members of the zinc cluster protein family is situated in the N terminus of the proteins (MacPhersonet al. 2006). The regulatory domain, also known as the middle homology region, is a region of approximately 80 amino acids that appears to play an important role in the regulation of the activity of zinc cluster proteins (Schjerling and Holmberg 1996). The activation domain of zinc cluster proteins is typically at the C terminus of the proteins and composed of acidic amino acid residues (Schjerling and Holmberg 1996). A number of zinc cluster proteins are activated by binding of small ligands and are the functional analogs of nuclear receptors found in metazoans (Naar and Thakur 2009).
Dal81 has been shown to be required for the transcriptional activation of a significantly large number of genes, whose functions include the catabolism of nitrogen sources such as urea, allantoin, arginine, and GABA. Dal81 is also required for the activation of AGP1, a wide-range specificity amino acid permease expressed in the presence of amino acids in the external environment of S. cerevisiae (Abdel-Sateret al. 2004). Dal81 has been shown to be necessary for the induction of UGA1 and UGA4 in the presence of GABA (André 1990; Talibiet al. 1995). GABA-dependent activation of UGA1 and UGA4 is also mediated by Uga3, which binds to upstream activating sequences (UASGABA) found in the promoters of these genes (André 1990; Visserset al. 1990; Talibiet al. 1995).
Interestingly, Bricmontet al. (1991) have shown that a yeast strain expressing a mutant of Dal81 lacking the cysteine-rich zinc cluster motif did not have reduced levels of urea amidolyase, an enzyme encoded by the Dal81-regulated gene DUR1,2. Furthermore, this strain did not have detectable growth defects on medium containing GABA as the sole nitrogen source (Bricmontet al. 1991). It therefore appears that the zinc cluster motif of Dal81 is not required for the activation of at least some of its target genes. A similar observation was made for TamA, a protein found in the filamentous fungus Aspergillus nidulans, which is also required for the activation of genes involved in the catabolism of nitrogen sources such as GABA (Daviset al. 1996; Smallet al. 2001).
Besides gene-specific activators and RNA polymerase, various complexes termed coactivators have been shown to play an important role in transcriptional activation. Mediator is a multiprotein complex that bridges gene regulators to general transcription factors and RNA polymerase II (Biddick and Young 2005; Casamassimi and Napoli 2007). The complex can be subdivided into three regions: the tail subcomplex that includes Gal11 and Sin4, the middle region, and the head region. Mediator produces its effect by directly interacting with the C-terminal domain of the largest subunit of RNA polymerase II. Another coactivator is the SAGA (Spt–Ada–Gcn5–acetyltranferase) complex, which possesses different enzymatic activities (Baker and Grant 2007). For example, the subunit Gcn5 is responsible for acetylation of various substrates including histone H3. Chromatin remodeling can be performed by the SWI/SNF complex, which can displace nucleosomes in an ATP-dependent manner (Mohrmann and Verrijzer 2005; Smith and Peterson 2005). SWI2 encodes the catalytic subunit of this complex. RSC (remodel the structure of chromatin) is another example of an ATP-dependent chromatin remodeling complex (Mohrmann and Verrijzer 2005).
In this study, we show that GABA enhances DNA binding of Uga3 and recruitment of Dal81 and Gal11. Under artificial conditions, both Dal81 and Uga3 activate transcription independently of each other, a process mediated by SAGA and Gal11. Thus, under physiological conditions, Dal81 and Uga3 cooperate by targeting common coactivators for transcriptional activation.
Materials and methods
Strains used in this study are listed in Table 1. The wild-type yeast strains used were BY4741 (MATa his3Δ1, leu2Δ0, met15Δ0, ura3Δ0), BY4742 (MATα his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0) (Brachmannet al. 1998), and YPH499 (MATa his3Δ200, leu2Δ1, lys2-801, trp1Δ1, ade2-101, ura3-52) (Sikorski and Hieter 1989). Deletion strains were obtained from Invitrogen/Research Genetics (Huntsville, AL) (Winzeleret al. 1999). The open reading frame (ORF) of UGA3 was tagged at its natural chromosomal location with a triple Myc epitope in strains BY4741 or BY4742 according to Schneideret al. (1995). Uga3 was tagged at the N terminus by transforming strain BY4741 with the PCR product obtained using plasmid pMPY-3XMYC as the template (Schneideret al. 1995) and the oligonucleotides CATGTATGGATGCCAAGAAAACAAAGTTTTTTAAAGTGAGGT ATGAGGAACAAAAGCTGGAG and CCCATGCTTCGAATATTTCAATTTCAGCTTCTCCACGCCATAATTTAGGGCGAATTGGGTACC. The nucleotides in boldface type correspond to the initiatorcodon of the tagged ORF. After transformation, colonies were selected on plates lacking uracil, and homologous recombination was verified by PCR. Cells were then grown overnight in YPD medium (Adamset al. 1997) to allow internal recombination between the two regions coding for the epitomes. Ura3− cells were selected on plates containing 5-fluoroorotic acid (Schneideret al. 1995). Tagging of DAL81 was performed as described above using the oligos TGTTTAGACGAGCGGCAGAACGACAGGCAGCCATACTATCAA ATGAGGGAACAAAAGCTGGAG and CTTCGTAGGCGATGCGGCATTATCAGCTGGTGATTGGTGAGGGTCTAGGGCGAATTGGGTACC and plasmid pMPY-3XHA as the template.
Strain BY4742 Δuga3Δtrp1 (MATα his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0, trp1Δ::hisG, uga3Δ::kanMX4) was produced by gene disruption of the TRP1 gene in BY4742 Δuga3, as described by Alaniet al. (1987). Specifically, strain BY4742 Δuga3 was transformed with a 4.7-kb fragment produced from the digestion of plasmid pNKY1009 (Alaniet al. 1987) with the restriction enzymes EcoRI and BglII. The extremities of this 4.7-kb fragment were homologous to a portion of the TRP1 gene and the middle section of the fragment was composed of the URA3 gene flanked by two direct repeats of Salmonella hisG DNA. This fragment was inserted into the TRP1 locus by homologous recombination and the URA3 gene was eliminated by selecting for Ura3− strains on plates containing 5-fluoroorotic acid.
Strain BMASY1 Δdal81Δuga3Δtrp1 (MATa his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0, trp1Δ::his, dal81Δ::kanMX4, uga3Δ::kanMX4) was produced by first mating BY4742 Δuga3Δtrp1 with BY4741 Δdal81 followed by sporulation and tetrad dissection. Spore BMASY1 Δdal81Δuga3Δtrp1 was confirmed as containing disruptions of UGA3 and DAL81 by Southern blot analysis (data not shown). Other strains carrying a deletion of DAL81 were obtained by amplifying the LEU2 marker using pRS405 (Sikorski and Hieter 1989) as a template and the oligonucleotides GCCTGTTTAGACGAGCGGCAGAACGACAGGCAGCCATACTATCAAAGATTGTACTGAGAGTGCAC and TAGTTATTACCGTATTCCATTTTTACTTATGTGCTATTATTTATACTGTGCGGTATTTCACACCG. The PCR product contains sequences homologous to the 5′ and 3′ regions of the DAL81 gene.
Expression vectors are described in Table 2. Plasmid p414MET25-3HA-UGA3 was constructed in two steps. The UGA3 ORF was first amplified by PCR, using genomic DNA from strain YPH499 as the template, with oligonucleotides CGGGATCCATGAATTATGGCGTGGAGAA and GGAATTCACGCATAGTCAGGAACATCGTATGGGTATCAGGCAAAATTAATATTT. The resulting product was digested with the restriction enzymes EcoRI and BamHI and subcloned into the low-copy plasmid p414MET25 (TRP1 selective marker) (Mumberget al. 1994) digested with the same enzymes, yielding the expression vector p414MET25-UGA3. The region coding for three HA epitopes was then amplified by PCR, using plasmid pMPY-3XHA (Schneideret al. 1995) as the template, with oligonucleotides GAAGATCTCTGCAGATGTACCCATACGATGTTCCT and GAAGATCTAGCAGCGTAATCTGGAACG. This PCR product was digested with the restriction enzyme BglII and subcloned into p414MET25-UGA3 digested with BamHI, thereby producing p414MET25-3HA-UGA3.
Plasmid p414MET25-3HA-UGA3(Δ124–300) was constructed by amplifying a region of the UGA3 ORF by PCR, using p414MET25-3HA-UGA3 as the template, with oligonucleotides GGGATTGCAGCTGTACAACTTCCCGATCGAACAATAC and TTACATGCGTACACGCGTTT. The resulting product was digested with the restriction enzyme BsrG1 and subcloned into p414MET25-3HA-UGA3 digested with the same enzyme.
Plasmid p414MET25-3HA-UGA3(Δ124–350) was constructed in two steps. A region of the UGA3 ORF was first amplified by PCR, using p414MET25-UGA3 as the template, with oligonucleotides ATCATTGTACAACGTCACAAATTTGACGAG and TTACATGCGTACACGCGTTT. The resulting product was digested with BsrG1 and subcloned into p414MET25-UGA3. The obtained plasmid was afterward digested with BtgI and NsiI, and the smaller fragment was subcloned into p414MET25-3HA-UGA3 digested with the same enzymes.
Plasmid p414MET25-3HA-UGA3(Δ518–528) was constructed by amplifying a region of the UGA3 ORF by PCR, using p414MET25-3HA-UGA3 as the template, with oligonucleotides GGAATTCATCAAATCATGTGGACCAGGTC and GACGATGCATGCTGAACTACAGATACTGAGA. The resulting product was digested with the restriction enzymes NsiI and EcoRI, and subcloned into p414MET25-3HA-UGA3 digested with the same enzymes.
Plasmid p414MET25-lexA-UGA3(78-528) was constructed by amplifying the region coding for the DNA-binding domain of LexA by PCR, using plasmid pEG202 (Ausubelet al. 1997) as the template, with oligonucleotides CCGGGATCCATGAAAGCGTTAACGGCCA and GCTATGGCATGCGGCGGGAATTCCAGCCAGT. The resulting product was digested with BamHI and SphI and subcloned into p414MET25-UGA3 digested with the same enzymes.
Plasmid p423MET25-lexA-HA-DAL81 was constructed in two steps. The DAL81 ORF was amplified by PCR using genomic DNA from strain BY4741 HA-DAL81 (Akacheet al. 2004) as the template and oligonucleotides AGCTAATCGTCGACTTACAGAGGGGTTTCCCTTG and CCATCGATGAGGGAACAAAAGCTGGA. The resulting product was digested with SalI and ClaI and subcloned into p423MET25 (Mumberget al. 1994) to yield p423MET25-3HA-DAL81. The region coding for the DNA-binding domain of LexA was amplified by PCR, using plasmid pEG202 (Ausubelet al. 1997) as the template, with oligonucleotides CCATCGATGAAAGCGTTAACGGCCAG and CCATCGATAACGGGAATTCCAGCCAGTCGC. The resulting product was digested with ClaI and subcloned into p423MET25-3HA-DAL81 digested with the same enzyme.
Plasmid p423MET25-3HA-DAL81 was constructed by amplifying the DAL81 ORF by PCR using genomic DNA from strain BY4742 HA-DAL81 as template with oligonucleotides AGCTAATCGTCGACTTACAGAGGGGTTTCCCTTG and CCATCGATGAGGGAACAAAAGCTGGA. The resulting product was digested with SalI and ClaI and subcloned into the high-copy vector p423MET25 (Mumberget al. 1994) cut with ClaI and XhoI.
Plasmid p423MET25-3HA-DAL81ΔZn was constructed by using plasmid p423MET25-3HA-DAL81 as a template for production of single-stranded DNA and oligonucleotide AGCAATACTGAAGGTAGATCCCATCAGATT for site-directed mutagenesis. This resulted in the deletion of DAL81 sequences encoding aa 150–179.
Plasmid p413MET25-3HA-UGA3 was constructed by subcloning a 1.5-kb SpeI–EcoRI fragment (containing the UGA3 ORF) from plasmid p414MET25-3HA-UGA3 into p413MET25 cut with the same enzymes (Mumberget al. 1994).
The reporter pUASGABA-lacZ is a high-copy plasmid containing a URA3 marker that was constructed by inserting the double-stranded oligonucleotides TCGAAAAGCCGCGGGCGGGATTGTA and AATCCCGCCCGCGGCTTT in front of a minimal CYC1 promoter driving lacZ transcription, as previously described (“UGA1-WT” in Noël and Turcotte 1998). The pSH18.34 reporter is a high-copy plasmid containing a URA3 marker with eight lexA-binding sites in front of a minimal GAL1 promoter driving lacZ transcription (Ausubelet al. 1997).
Yeast strains were transformed with reporters, and expression vectors when appropriate, and were grown on selective medium lacking uracil and/or tryptophan and/or histidine, depending on the selective marker contained on the plasmids. Transformed colonies were grown overnight in YPD medium (Adamset al. 1997). Cells were then diluted in SC medium (Adamset al. 1997) (minus the amino acids used for selection) at 0.008% each, and 2% glucose. For assays shown in Figures 3 and 6, drop-out media were used. The activity of the reporters was assayed in the absence and in the presence of 0.1% GABA in the growth medium. The β-galactosidase assays were performed with permeabilized cells (Guarente 1983). Results were obtained from at least two independent experiments done with duplicate or triplicate samples.
Production of whole cell extracts for immunoblot analysis
Yeast cultures were grown overnight in YPD (Adamset al. 1997). Cells containing expression vectors and reporter plasmids were diluted in 100 ml of SD medium (Adamset al. 1997) lacking ammonium sulfate, supplemented with adenine, leucine, lysine, and histidine at 0.008%, as well as 2% glucose and 0.1% proline; cells expressing chromosomally tagged proteins were diluted in 300 ml SD medium (Adamset al. 1997) lacking ammonium sulfate, supplemented with adenine, leucine, lysine, histidine, tryptophan, and uridine at 0.004%, as well as 0.01% drop-out medium supplement, 2% glucose, and 0.1% proline. All cultures were grown to an OD600 of ∼0.7. Cells were then pelleted, washed in ice-cold water, and resuspended in an equal volume of ice-cold IP-1 buffer (15 mm Tris-HCl pH 7.6, 150 mm NaCl, 1% Triton X-100, 10 mm pyrophosphate, 2 mm dithiothreitol (DTT), 1 mm phenylmethylsulfonyl fluoride (PMSF), 1 μg/ml pepstatin, 1 μg/ml leupeptin), as modified from Mamnunet al. (2002). An equal volume of cold glass beads was then added and the cells were lysed by vortexing four times for 1 min at 4°C. The lysate was then separated from the debris by centrifugation and aliquots were boiled in 1.5× Laemmli buffer. The samples (containing 30 μg of proteins) were then run on a 7.5% SDS–polyacrylamide gel, transferred to a PVDF membrane, and analyzed by immunoblotting with 8 μg of anti-HA antibody (clone 12Ca5, Roche Applied Science).
Chromatin immunoprecipitation (ChIP) assay
Strains expressing chromosomally tagged proteins (Table 1) were grown as described in the previous section. ChIP assays were performed as described (Larochelleet al. 2006) except that quantitation was performed by qPCR. Enrichments were calculated over an untagged strain and normalized with the signal obtained with ARN1 as an internal control using the 2−ΔΔCT method (Livak and Schmittgen 2001). Oligonucleotides used for ChIP were ATTCGCGCTATCTCGATTTC and CACCGCACCAATGGATAAAC for the UGA1 promoter (−501 bp to −251 bp relative to the ATG) and TGCACCCATAAAAGCAGGTGT and GAGAGCTATCGAATGTTTCCTC for the ARN1 promoter (−260 bp to −86 bp relative to the ATG).
Analysis of the functional domains of Uga3
Previous studies have shown that the DNA-binding domain of Uga3 is located at the N terminus (Noël and Turcotte 1998; Idiculaet al. 2002). For example, a purified Uga3 polypeptide (aa 1–124) binds to UASGABA in vitro (Noël and Turcotte 1998). However, very little information about the other domains of Uga3 is available and sequence alignments of zinc cluster proteins failed to identify a middle homology region in Uga3 (Schjerling and Holmberg 1996). To map the functional domains of this regulator, the UGA3 ORF was first amplified by PCR and was subcloned into the plasmid p414MET25, a low-copy yeast expression vector under the control of the repressible MET25 promoter (Mumberget al. 1994). The coding region of three HA epitopes was afterward inserted in-frame with the UGA3 ORF, allowing the expression of Uga3 tagged at its N terminus with three HA epitopes. We also generated vectors expressing Uga3 derivatives with C-terminal or internal truncations and a summary of key mutants is presented in Figure 1A.
Uga3 and mutants (Table 2) were expressed in a Δuga3 strain and their activity assayed with the pUASGABA-lacZ reporter containing a single binding site for Uga3 (UASGABA) inserted upstream of a minimal CYC1 promoter driving expression of lacZ (Noël and Turcotte 1998). Since this reporter contains only a UASGABA, it allows specific monitoring of the activity due to Uga3 (or derivatives). Plasmid-expressed HA-Uga3 was able to induce high levels of reporter gene activity in a GABA-dependent manner (Figure 1B) while a reporter lacking the UASGABA gave background activity (data not shown and Noël and Turcotte 1998). Truncations of 100 or 200 aa at the C terminus of Uga3 completely abolished transcriptional activation (Figure 1A). Interestingly, an Uga3 mutant HA-Uga3(Δ518–528) carrying a very short (10 aa) deletion at the C terminus, also failed to activate the reporter even in the presence of GABA (Figure 1B). These results indicate that an activation domain is located at the C terminus of Uga3.
A mutant carrying an internal deletion [HA-Uga3(Δ124–300)] showed constitutive activity albeit at reduced levels when compared to wild-type Uga3. Addition of GABA slightly induced (less than twofold) reporter activity. A mutant with a larger internal deletion, HA-Uga3(Δ124–350), failed to activate the reporter. We assayed the relative expression of full-length HA-Uga3 and the three mutant forms of Uga3 described above, to test that the results of the β-galactosidase assays were not due to Uga3 derivatives not being expressed. Whole-cell extracts from BY4742 Δuga3Δtrp1 transformed with each of the four expression vectors were used in immunoblot analysis, using an anti-HA primary antibody (Figure 1C). Despite the fact that there is variation in protein levels, all of the Uga3 mutant proteins were clearly expressed. Furthermore, HA-Uga3(Δ124–300) induced levels of β-galactosidase activity lower than HA-Uga3 but the mutant protein was also expressed at lower levels than HA-Uga3. Our functional analysis of Uga3 shows that this factor has an activation domain located at its C terminus while a regulatory region is found in its “central” portion.
The Dal81 zinc cluster domain is not required for activation of a UASGABA reporter
In line with studies on a native UGA1 promoter (Bricmontet al. 1991), results show that removal of DAL81 greatly reduced activity of the UASGABA reporter (Figure 2, left part). Introduction of a high copy expression vector for HA-Dal81 into a Δdal81 strain increased promoter activity at levels higher than a wild-type strain, presumably because DAL81 is overexpressed under these conditions. We also expressed a truncated version of Dal81 lacking amino acids 150 to 179 encompassing all 6 cysteines residues of the zinc cluster domain. Many studies on zinc cluster proteins predict that this truncated protein would be defective for binding to DNA (MacPhersonet al. 2006). However, the truncated Dal81 protein was as efficient as wild-type Dal81 for activation of the UASGABA reporter suggesting it does not need to directly contact DNA for activation of a Uga3-responsive promoter (Figure 2, right part).
Activity of LexA-Uga3 chimeric protein is dependent on the presence of GABA and Dal81
We were then interested in determining if the GABA-induced activation of Uga3 requires a specific DNA-binding domain or can be transferred to a heterologous DNA-binding domain. To this end, we expressed a chimeric protein consisting of the LexA DNA binding protein fused to amino acids 78 to 528 of Uga3. Interestingly, this chimeric protein activated expression of a lacZ reporter bearing lexA sites in a GABA-dependent manner (Figure 3A) while expression of LexA by itself gave background reporter activity (data not shown). A LexA-Uga3 fusion protein lacking N-terminus of Uga3 [LexA-Uga3(300-528)] was constitutively active, in analogy to the Uga3 Δ124-300 mutant described previously. Taken together, these results suggest that a GABA-responsive domain could be present between aa 78 and 300 of Uga3.
As observed with wild-type Uga3, activation by LexA-Uga3(78-528) was abolished in the absence of DAL81 (Figure 3A). Thus, Dal81 dependence of Uga3 for transcriptional activation by Uga3 can be transferred to a chimeric protein bearing a heterologous DNA-binding domain. Moreover, activation was as efficient as wild-type Dal81 when a truncated Dal81 lacking the zinc cluster region was co-expressed along with LexA-Uga3(78-528) (Figure 3B). Taken together, these results raises the possibility that the role of Dal81 is not (only) to facilitate DNA-binding of regulators such as Stp1 (Boban and Ljungdahl 2007), but also to directly contribute to transcriptional activation (see below).
Cofactors involved in Uga3 transcriptional activation
We were interested in determining the possible requirement of Uga3 for coactivators other than Dal81. To this end, the UASGABA reporter was introduced into various strains carrying deletions of genes encoding components of SAGA (ADA1, GCN5, SPT3, SPT20), mediator (GAL11, MED5, SIN4), ATP-dependent chromatin remodeling complexes (RSC2, SWI2), and β-galactosidase activity assayed. Deletion of genes encoding various components of the SAGA complex (ADA1, GCN5, SPT3, SPT20) resulted in significantly decreased activation by Uga3 (Figure 4). A strain lacking GAL11 showed 30% β-galactosidase activity relative to a wild-type strain but no marked effect was observed when deleting SIN4 or MED5 encoding other components of mediator. Finally, deletion of RSC2 or SWI2 had little effect on activation by Uga3. In summary, the SAGA complex as well as Gal11 modulate the transcriptional activity of Uga3.
Dal81 and Gal11 are recruited to the UGA1 promoter in the presence of GABA
Dal81 was shown to enhance activation of the factor Stp1 by facilitating its binding to target promoters (Boban and Ljungdahl 2007). To test if a similar mechanism is responsible for the requirement of Dal81 for Uga3 activation, we performed ChIP experiments using Uga3 tagged at its natural chromosomal location with three Myc epitopes. Results show that Myc-Uga3 is very weakly bound (1.6-fold enrichment) to the promoter of the target gene UGA1, as assayed with proline as the nitrogen source (Figure 5, column 1). In the presence of GABA, binding of Myc-Uga3 is enhanced approximately fourfold. Deletion of DAL81 decreased but, importantly, did not abolish binding of Uga3 to the UGA1 promoter. We also assayed the recruitment of Dal81 at the UGA1 promoter by ChIP. No enrichment of HA-Dal81 at UGA1 was detected in the absence of GABA while addition of the inducer allowed recruitment of this factor (Figure 5, column 3), in agreement with another study performed with the UGA4 promoter (Cardilloet al. 2010). As shown above, Gal11, a component of the mediator complex, is required for full activation by Uga3. In agreement with these observations, ChIP analysis showed that Gal11 is recruited to the UGA1 promoter in the presence of GABA (Figure 5, column 4). Interestingly, removal of DAL81 prevented recruitment of Gal11 (Figure 5, column 5). Similar results were obtained with UGA4, another target gene of Uga3 (data not shown). Taken together, these results suggest that Dal81 facilitates binding of Uga3 and recruitment of Gal11 at the UGA1 and UGA4 promoters.
Dal81 acts as a transcriptional activator when tethered to DNA
We wanted to test if the effect of Dal81 is strictly restricted to helping Uga3 bind to DNA or if the effect of Dal81 is mediated via recruitment of coactivators. We first tested to see if Dal81 by itself has transcriptional activation properties; to this end, we fused the DNA-binding domain of LexA to Dal81 and measured the activity of a reporter containing lexA binding sites. Strong activity of the reporter was observed when expressing LexA-Dal81 in the absence of GABA but not LexA alone (Figure 6). Addition of GABA did not increase activity of LexA-Dal81, in agreement with the fact that Dal81 is a general coactivator of nitrogen-regulated genes. Activation by LexA-Dal81 was approximately 2.5-fold stronger when performing the assay in a strain lacking UGA3 while addition of GABA had no effect. Reduced activity of LexA-Dal81 in the presence of wild-type Uga3 may be explained by the fact that, under these artificial conditions, Uga3 may compete with Lex-Dal81 for cofactors. These results provide additional evidence that the mechanism of action of Dal81 is not only exerted by favoring binding of transcription factors to target sites but that Dal81 also contributes directly to transcriptional activation. Activation by LexA-Dal81 was measured in various strains carrying deletions of genes encoding coactivators, as described for Uga3. As observed with Uga3, deleting SPT20, SPT3, and GAL11 (but not GCN5) significantly reduced activation by LexA-Dal81 (Figure 6). In contrast to Uga3, removal of SIN4 or RSC2 did reduce (approximately 2-fold) the activation by LexA-Dal81. These differences could be explained by the fact that the assay was performed in different promoter contexts. Taken together, our data clearly show that Dal81 can directly activate transcription regardless of the presence of Uga3 and that SPT20, SPT3, and GAL11 are important for transcriptional activation by native Uga3/Dal81 and LexA-Dal81.
Overexpression of Uga3 bypasses the requirement for Dal81
Two models could explain the results described above. GABA-bound Uga3 allows recruitment of Dal81 that, in turn, would bring coactivators such as SAGA and Gal11 to target promoters. Alternatively, both Uga3 and Dal81 would be responsible for recruitment of SAGA and Gal11. To distinguish between these two possibilities, we took advantage of our observation that overexpression of Uga3 bypasses the requirement for Dal81 (Figure 7). As expected, activation by Uga3 expressed from its native promoter is strongly dependent on Dal81 (Figure 7, columns 1 and 2). However, introduction of an expression vector for Uga3 in wild-type yeast cells results in a DAL81-independent activation of the reporter (Figure 7, columns 3 and 4). Western blot analysis showed that, under these conditions, Uga3 was overexpressed by a factor of at least 20 (data not shown). We then performed β-galactosidase assays with Uga3 overexpressed in strains lacking both DAL81 and genes encoding other coactivators. β-Galactosidase activity was greatly reduced in strains Δdal81Δspt20 and Δdal81Δspt3 and approximately 3-fold in strain Δdal81Δgal11 (Figure 7, columns 5–7). However, with MED5, a slightly stronger effect was observed with overexpressed Uga3 when compared with native Uga3/Dal81 or LexA-Dal81 (Figure 7, column 8). Taken together, these results show that the role of Uga3 is not restricted to DNA binding and the recruitment of Dal81 but that it also directly participates in transcriptional activation via SAGA and Gal11.
In this study, we characterized the zinc cluster proteins Dal81 and Uga3, a general positive regulator of nitrogen utilization genes and an activator of genes for GABA catabolism, respectively. The mutational analysis done on Uga3 provided new information on the functional domains of this zinc cluster protein (Figure 1). Previous comparative genome analysis did not identify a regulatory domain (also named the middle homology region) in Uga3 (Schjerling and Holmberg 1996). However, our experiments showed that a mutant lacking the region corresponding to amino acids 124–300 was constitutively active. These results were further supported by the chimeric protein LexA-Uga3(300–528), which functioned in a similar constitutive manner (Figure 3A). Analogous observations were made with the zinc cluster proteins Leu3 and Hap1, which become constitutively active when their regulatory domains are deleted (Pfeiferet al. 1989; Zhouet al. 1990). It therefore appears that Uga3 possesses such a regulatory domain, located between the DNA-binding domain and the activation domain. The reason why comparative genome analysis failed to identify a regulatory domain in Uga3 may be due to the fact that this domain is not actually highly conserved among zinc cluster proteins. It is possible that the regulatory domain of Uga3 diverged from the other zinc cluster proteins to accommodate for its function in the response to GABA.
Our mutational analysis also demonstrated that the region of Uga3 encompassing amino acid residues 300–350 is necessary for the activity of Uga3 as a transcriptional activator. The reason this domain is necessary for the function of Uga3 is, however, not clear. It is possible that an important part of the activation domain of Uga3 could be between amino acids 300 and 350. The Uga3 truncation mutant lacking the last 10 amino acid residues of the full-length protein was also unable to induce the activity of the pUASGABA-lacZ reporter in the presence of GABA. The C terminus of Uga3 is required for the activity of this transcription factor but does not show homology to known domains involved in transcription.
Besides Dal81, Uga3 requires the SAGA complex as well as the Gal11 component of mediator for transcriptional activation (Figure 4). For example, removal of GAL11 resulted in reduced (3.3-fold) reporter activity. Similarly, activation of transcription by the zinc cluster protein Pdr1 was shown to be mediated by direct interaction with the component Gal11 of the mediator complex (Thakuret al. 2008). Although we observed reduced binding of Uga3 at the UGA1 (and UGA4) promoter in cells lacking DAL81, there was still significant binding of this factor to target genes, as assayed by ChIP (Figure 5). In contrast to our observations, removal of Dal81 almost completely abolished binding of the transcription factor Stp1 to target promoters (Boban and Ljungdahl 2007).
It is possible that Dal81 has two, nonexclusive, modes of action: (1) one operating by facilitating binding of transcription factors to DNA and (2) another by contacting coactivators for a more efficient formation of an initiation complex at target promoters. The observation that Dal81, when tethered to DNA, acts as a strong and Uga3-independent transcriptional activator (Figure 6) supports (at least in part) the second mechanism of action of Dal81. A Dal81 mutant lacking the cysteine-rich region and, as a result, most likely defective in DNA binding, was fully functional as a coactivator for Uga3 as well as LexA-Uga3(78–528) (Figure 3). In the latter case, both factors act in concert in spite of the fact that the truncated Dal81 (Dal81ΔZn) cannot bind DNA and that LexA-Uga3 lacks the natural Uga3 DNA binding domain. These results, as well as previous observations (Bricmontet al. 1991), strongly suggest that Dal81 does not need to directly contact DNA but may associate with Uga3 to form a stable complex at target promoters, such as UGA1. We failed to observe such an interaction using co-immunoprecipitation assays perhaps because this interaction is weak (M.-A. Sylvain and B. Turcotte, unpublished results). However, in support of a Dal81–Uga3 interaction, Dal81 interacts with Dal82, an activator of allantoin catabolic genes, in a two-hybrid assay (Scottet al. 2000). A parallel can also be made between Dal81 and its A. nidulans homolog TamA. In analogy to Dal81, TamA activates transcription when tethered to DNA (Smallet al. 1999). As shown in a two-hybrid assay, TamA interacts with the transcriptional activator LeuB (Polotniankaet al. 2004). LeuB, a homolog of S. cerevisiae Leu3, and TamA are both involved in controlling the expression of the gdhA gene encoding NADP-linked glutamate dehydrogenase. Moreover, mutating a critical cysteine residue in the zinc cluster domain of TamA has no effect on its functionality (Daviset al. 1996). Thus, results suggest that the DNA-binding domain of Dal81 and TamA are dispensable in spite of the fact that the primary amino acid sequence of the cysteine-rich region of these factors corresponds to a bona fide zinc cluster. Dal81 and TamA DNA-binding activity may be required for activation at a subset of the target genes, yet to be identified.
Importantly, our results show that the role of Uga3 is not restricted to recruiting Dal81. This was demonstrated by showing that overexpressed Uga3 can activate transcription at wild-type levels even in the absence of Dal81 (Figure 7). Moreover, both LexA-Dal81 and overexpressed Uga3 mainly rely on the same set of coactivators (SAGA and Gal11) for transcriptional activation, suggesting a cooperative mode of action for these factors (Figures 6 and 7). As stated above, binding of Stp1 to DNA is more dependent on Dal81 than Uga3. However, it is also possible that a mechanism similar to Uga3 operates for Stp1. Recruitment of Dal81 would increase formation of a transcription complex at target promoters of Stp1, the only difference being that Stp1 DNA affinity would be lower than Uga3. The yeast transcriptional activator Gcn4 also requires SAGA and mediator for activity. Recruitment of these factors is mediated by Gcn4 itself (Qiuet al. 2005). In contrast, our genetic studies and ChIP assays suggest that both Uga3 and Dal81 can independently recruit Gal11 and SAGA. It will be interesting to determine if Dal81 also relies on SAGA and Gal11 at other nitrogen metabolic genes.
What is the mechanism of activation of Uga3 by GABA? Activity of a number of zinc cluster proteins is controlled by intermediary metabolites (for a review, see Sellick and Reece 2005). For example, the activity of the zinc cluster proteins Leu3 and Lys14 is controlled by intermediary metabolites of leucine and lysine biosynthesis, respectively (Felleret al. 1994; Felleret al. 1999; Szeet al. 1992). In contrast, our unpublished results showed that activation by Uga3 is not dependent on the presence of GABA metabolites since no reduction of reporter activity is observed in strains Δuga1 and Δuga2, which are unable to catabolize GABA. We hypothesize that Uga3 is directly activated by GABA in a manner analogous to the zinc cluster protein Put3, an activator of proline utilization genes that is directly activated by binding to its ligand proline (Sellick and Reece 2003). Similarly, the zinc cluster protein Pdr1, an activator of drug resistance genes, is activated by direct binding of diverse drugs and xenobiotics (Thakuret al. 2008). Attempts to show direct binding of GABA to purified Uga3 failed since overexpression of this factor from yeast or bacterial cells did not yield sufficient purified Uga3 for binding studies.
Our unpublished results suggest that Uga3 is found in the form of a homodimer, even in the absence of GABA. In addition, a large-scale localization study showed that, under noninducing conditions, Uga3 is found in the nucleus (Huhet al. 2003). Thus, Uga3 is most probably found as a homodimer in the nucleus regardless of the presence of GABA (see proposed model in Figure 8, top). However, our ChIP assays (Figure 5) demonstrated that Uga3 is only weakly bound to the UGA1 promoter. As stated above, we hypothesize that GABA binds directly to Uga3 resulting in a conformational change that would favor binding of Uga3 to UGA1 (Figure 8, bottom). This conformational change would also permit Uga3 to interact with Dal81. Both Uga3 and Dal81 can then interact with coactivators such as Gal11 and SAGA to allow the formation of a pre-initiation complex. The absence of Dal81 results in reduced surface for interaction with coactivators leading to greatly diminished transcriptional activation by Uga3. Conversely, overexpression of Uga3 bypasses the requirement of Dal81 because higher levels of Uga3 would increase its binding to target promoters for recruitment of Gal11 and SAGA. In conclusion, our studies show that Dal81 and Uga3 act in concert by targeting common components of the transcriptional machinery.
We thank Dr. A. Hinnebusch for providing the Myc-Gal11 strain. We are grateful to Cindy Ong for her contribution to the analysis of the functional domains of Uga3. We acknowledge the help of Dr. François Robert and Louise Laramée for qPCR analysis. We also thank Anne-Marie Sdicu, from the laboratory of Dr. Howard Bussey (Department of Biology, McGill University) for tetrad dissection. We thank Drs. John White, Martine Raymond, and Geoffrey Hendy for critical reading of the manuscript. The advice provided by Drs. Marc Larochelle, Bassel Akache, and Sarah MacPherson was greatly appreciated. This work was supported by a grant from the Natural Sciences and Engineering Research Council of Canada (NSERC) to B.T. (grant 184053-09). M.-A. S. was supported by a an undergraduate research award from NSERC and studentship from the Fonds de la Recherche en Santé du Québec.
- Received December 22, 2010.
- Accepted April 5, 2011.
- Copyright © 2011 by the Genetics Society of America