Previous Article | Next Article ![]()
Eukaryotic Cell, October 2003, p. 876-885, Vol. 2, No. 5
1535-9778/03/$08.00+0 DOI: 10.1128/EC.2.5.876-885.2003
Copyright © 2003, American Society for Microbiology. All Rights Reserved.
Claudia Wagner,1 Sven Krappmann,1 Fritz Thoma,2 and Gerhard H. Braus1*
Institute of Microbiology and Genetics, Georg-August-University, D-37077 Göttingen, Germany,1 Institute of Cell Biology, ETH Hönggerberg, CH-8093 Zürich, Switzerland2
Received 13 January 2003/ Accepted 4 August 2003
|
|
|---|
|
|
|---|
Swi/Snf was purified as a 2-MDa protein complex that is composed of the Swi1, Swi2/Snf2, Swi3, Snf5, and Snf6 proteins plus five additional polypeptides (7). Swi/Snf can bind to nucleosomes and DNA, thereby creating loops in nucleosomal arrays or naked DNA, respectively, bringing distant sites into close proximity (4). In an ATP-dependent fashion it can reposition nucleosomes in cis on the same DNA molecule (47), with the SWI2/SNF2 gene encoding the DNA-dependent ATPase activity (33). Transcriptome analyses with a swi2/snf2 mutant strain have revealed that Swi/Snf controls transcription of only 6% of all Saccharomyces cerevisiae genes and that the control is exerted at the level of individual promoters rather than over chromosomal domains (39). Swi/Snf thereby both activates and represses transcription of different target genes. Recruitment of Swi/Snf to specific promoters by DNA-binding regulatory proteins, as well as targeting of the complex by the general transcription machinery, has been suggested (48).
Two high-molecular-mass Ada-Gcn5 complexes (0.8 and 1.8 MDa) have been biochemically isolated from S. cerevisiae and shown to be able to acetylate nucleosomes both in vitro and in vivo at specific lysine residues of histones H3 and H4 (14). Both complexes share Gcn5p, Ada2p, and Ada3p, whereas the larger one additionally contains Spt proteins (Spt20p, Spt3p, Spt8p, and Spt7p) and is called SAGA (for Spt/ADa/Gcn5 acetyltransferase). Gcn5p comprises the histone acetyltransferase (HAT) activity to acetylate histones in promoter regions in a manner that is correlated with Gcn5p-dependent transcriptional activation and HAT activity in vitro (23). SAGA interacts with both TATA-binding protein and acidic transcriptional activators such as the herpesvirus VP16 activation domain and yeast Gcn4p, suggesting that the complex might also have a transcriptional adaptor function for some promoters (15).
HIS7 is a typical housekeeping gene of S. cerevisiae, and its gene product is necessary for the biosynthesis of the amino acid histidine as well as purines (21). Its expression has previously been shown to be activated by two major stimuli: the lack of amino acid availability and limitation of external purines (37). Starvation for amino acids triggers increased expression of Gcn4 protein that in turn activates transcription of amino acid biosynthetic genes (general control of amino acid biosynthesis) (6, 18). Upon starvation, two Gcn4p-recognition elements (GCREs) within the HIS7 promoter are targeted by the transcription factor, thereby mediating a sixfold increase in HIS7 expression (Fig. 1) (37). The absence of purines causes a twofold increase in HIS7 expression and is mediated by the heterodimeric transcription factor Bas1/Bas2p that shares a common binding site with Gcn4p, the TATA-proximal GCRE2 (Fig. 1). This activation requires both factors Bas1p and Bas2p (37). Nevertheless, both activation pathways act independently of each other and, moreover, are additive upon simultaneous amino acid and purine limitation (37). Located 416 bp upstream of the HIS7 gene is the ARO4 gene, encoding an enzyme that is required for the biosynthesis of aromatic amino acids (22). Both transcriptional units are separated by a positioned nucleosome in between them (45).
![]() View larger version (17K): [in a new window] |
FIG. 1. Diagram of the yeast HIS7 promoter. Two binding sites for Gcn4p, GCRE1 and GCRE2, are functional parts of the promoter. GCRE2 additionally functions as recognition element for the heterodimeric transcription factor Bas1p/Bas2p. TA at position -120 reflects the putative TATA element. The arrow at -87 indicates the initiator element of the major transcriptional start site. The positions are relative to the translational start codon AUG, which is shown as +1.
|
|
|
|---|
|
View this table: [in a new window] |
TABLE 1. Yeast strains and plasmids used in this study
|
Plasmids. The plasmids used in the present study are listed in Table 1. Plasmid pME1236, carrying the gcn5::KANR deletion cassette, was created by replacing the GCN5 coding sequence with the kanamycin resistance cassette by a PCR-based three-step cloning strategy, with plasmids pME1234 and pME1235 as intermediates. The other plasmids used here have been described previously.
Genomic chromatin preparation and nuclease digestions. These methods have been described previously (42). Biodyne B nylon membranes (Pall, Dreieich, Germany) were used for Southern transfer. Probes were labeled by the random primer method (11).
Indirect end labeling. Chromosomal DNA from the nuclease digestion was digested with XbaI and MluI and fractionated on 1.2% agarose gels. The fractionated DNA was blotted on a Hybond-N nylon membrane by the alkaline blotting method and then hybridized with a radioactively labeled 250-bp PCR amplicon, generated with oligonucleotides HIS7-CHR1 (5'-GAGATTAAAGAAATTGTCAGA-3') and HIS7-CHR2 (5'-CAAGTATTGAGGAGAAATGGTA-3'), with annealing just downstream of the XbaI site. A DNA ladder consisting of multiples of 256 bp was used for calibration (43).
RNA analysis. Total RNA from S. cerevisiae was isolated according to the method of Cross and Tinkelenberg (9). For Northern hybridization analysis, 20 µg of total RNAs were separated on a formaldehyde-agarose gel and transferred to a positively charged nylon membrane (Biodyne B; Pall) by capillary blotting. Hybridization with specific DNA probes was performed after 32P labeling of the respective DNA fragments with the Prime-It II DNA labeling kit from Stratagene. The HIS7 probe was generated with oligonucleotides HIS-OL1 (5'-GTGGTAACCTACAGTCACTAACC-3') and HIS-OL2 (5'-CCGATCGATACTTTATCAGCACC-3'), and the ACT probe was generated with oligonucleotides ACT-OL1 (5'-GCTGCTTTGGTTATTGATAACGG-3') and ACT-OL2 (5'-CACTTGTGGTGAACGATAGATGG-3'). Signal intensities were visualized by autoradiography and quantified by using a BAS-1500 Phospho-Imaging scanner (Fuji).
ß-Galactosidase assay. ß-Galactosidase activities were determined by using permeabilized yeast cells and the fluorogenic substrate 4-methylumbelliferyl-ß-D-galactoside (MUF) as described previously (21).
|
|
|---|
![]() View larger version (130K): [in a new window] |
FIG. 2. Chromatin structure at the ARO4-HIS7 intergenic region. Mnase protection experiments are shown with chromatin from strains FY1353 (wild-type GCN4) or RH1381 (gcn4 ). Cells were cultivated in the absence of exogenous adenine. Control Mnase digests of naked DNA are shown in lanes 3, 4, 8, 9, and 10. Lanes M on the far left and right show a 256-bp DNA-ladder (42). Black ramps at the bottom indicate increasing amounts of Mnase and exactly represent 20 and 40 U for lanes 1 and 2; 20, 40, and 60 U for lanes 5, 6, and 7; 4 and 8 U for lanes 3 and 4; and 4, 8, and 12 U for lanes 8, 9, and 10, respectively. The diagrams on the left and right show the location of the open reading frames of ARO4 and HIS7. Protected regions are shown as black ovals N1-N3 (here referred to as positioned nucleosomes). The binding sites for Gcn4p and Bas1p/Bas2p are indicated at the bottom. The site of HIS7 transcriptional initiation is indicated by Itr. pA, poly(A) addition site of the ARO4 mRNA; Z/S, Zaret/Sherman element required for efficient ARO4 mRNA 3'-end processing.
|
strain. However, they display the same Mnase-sensitive promoter region flanked by nucleosomes on both sides, as shown for cells with low Gcn4p levels (Fig. 2, lanes 5 to 10). A positioned nucleosome N4 in the 3' end of the ARO4 open reading frame is absent in cells with low Gcn4p levels but present in Gcn4p-lacking cells. Gcn4p overexpression releases the ordered chromatin structure at the intergenic region. Cells starving for amino acids increase expression of Gcn4p, the master regulator of the general control system of amino acid biosynthesis. For experimental purposes, amino acid starvation is often induced by adding 3-AT or 5-methyltryptophan to the growth culture. 3-AT and 5-methyltryptophan are amino acid analogs acting as false feedback inhibitors of the histidine or tryptophan biosynthetic pathways, respectively. The general control network can also be activated by using a constitutively highly expressed GCN4 allele (27). For Gcn4p-dependent HIS7 activation both induction methods were shown to increase its transcription (21, 37).
Basing our studies on the HIS7 promoter chromatin structure of the basally expressed gene as described above (for low Gcn4p, Bas1/Bas2p activation caused by absence of adenine), we now addressed the question of whether activation of HIS7 expression mediated by Gcn4p overexpression affects the HIS7 promoter chromatin structure.
To answer this question, we investigated the chromatin structure with a strain that harbors a highly expressed GCN4 allele and therefore constitutively expressing the general control system (27). Cells were cultivated in the presence of adenine so that the Gcn4p activation of HIS7 is not affected by Bas1/Bas2p. In striking contrast to the ordered HIS7 promoter chromatin structure described for the basal expression (Fig. 2, lanes 1 and 2), which is also observed for growth with 0.3 mM adenine (Fig. 3, lanes 1 and 2), several Mnase-sensitive bands in close proximity appear throughout the entire ARO4-HIS7 intergenic region as a result of Gcn4p overexpression (Fig. 3, lanes 5 and 6). Unambiguous localization of positioned nucleosomes that protect the DNA from the nuclease activity in that region is no longer feasible. We conclude that Gcn4p overexpression alters the chromatin structure at the ARO4-HIS7 intergenic region specifically and at the same time causes increased HIS7 and ARO4 transcription. We conclude that the HIS7 activation by high Gcn4p levels is associated with severe changes of the local ARO4-HIS7 chromatin structure, whereas the regulation of HIS7 transcription mediated by Bas1p/Bas2p seems not to be associated with any remodeling of nucleosomes.
![]() View larger version (70K): [in a new window] |
FIG. 3. Chromatin structure of the ARO4-HIS7 intergenic region at low and high intracellular Gcn4p levels. Yeast cells were grown in the presence of 0.3 mM adenine to prevent Bas1p/Bas2p-mediated HIS7 transcription. The Mnase protection pattern as described in Fig. 2 for cells with basal HIS7 expression levels of strain FY1353 can be seen in lanes 1 and 2. Strain FY1353[p238] expresses constitutively high amounts of Gcn4p and yields a protection pattern as shown in lanes 5 and 6. In lanes 3 and 4, restriction digests with Sau3A and XbaI/TaqI show DNA fragments precisely located within the HIS7 promoter. The diagram on the left and the black ramps at the bottom are as described for Fig. 2.
|
genetic background. Cells were cultivated in the absence of adenine in the growth medium to induce Bas1/2p-dependent HIS7 transcription. The chromatin mappings of strain RH2029 (bas1
and bas2
), which lacks any Bas1/Bas2 protein, revealed a distribution of Mnase-sensitive and Mnase-protected sites identical to the BAS1/2 wild-type strain RH1381 (Fig. 4, lanes 1 to 3 and 4 to 6). The absence of Bas1/2p does not cause changes at the HIS7 promoter chromatin structure. We conclude that the heterodimeric complex Bas1/2p is not required to keep the HIS7 promoter and its cis-acting elements accessible. Overexpression of BAS1 is toxic for cell viability. We analyzed HIS7 transcription and its promoter chromatin structure after a 2-h burst of BAS1 expression to test whether toxicity correlates with a changed nucleosome structure. High BAS1 expression was achieved with a high-copy-number plasmid, pCB159 (Table 1), carrying BAS1 under the control of the GAL1 promoter, induced by shifting cells to medium with galactose as single carbon source. Surprisingly, HIS7 transcription is not only not affected by high Bas1p levels but in fact nearly prevented under these circumstances (data not shown). The HIS7 chromatin structure during BAS1 overexpression remains unaffected and is identical to that of zero or low Bas1p levels, shown in lanes 10 to 12 of Fig. 4. Thus, Bas1p, in conjunction with Bas2p, neither requires nor triggers chromatin remodeling to fulfill its function at the HIS7 promoter.
![]() View larger version (92K): [in a new window] |
FIG. 4. Chromatin structure of the ARO4-HIS7 intergenic region at various intracellular Bas protein levels. Yeast cells were grown in the absence of adenine. Lanes: 1 to 3, chromatin of strain RH2029; 4 to 6, chromatin of strain RH1381; 10 to 12, chromatin of strain RH1381 with the BAS1 overexpression plasmid pCB159 (44). On this high-copy-number plasmid the BAS1 open reading frame is fused to the GAL1 promoter. After shifting from glucose to galactose medium, BAS1 expression is strongly induced (44). The location of the promoter region is labeled by Southern hybridizations of chromosomal DNA digested with either XbaI/TaqI or Sau3A.
|
background. We analyzed whether the ordered chromatin structure of the basal expressed HIS7 promoter or the Gcn4p-dependent change in chromatin structure is Swi/Snf dependent. Experiments with the snf2
mutant strain grown in the presence of low Gcn4p levels show a HIS7 promoter chromatin structure similar to the wild-type strain grown under basal gene expression conditions (Fig. 5 A, compare lanes 1 and 2 with lanes 5 and 6). The positions of sensitive, as well as protected, regions within the intergenic region are similar for the snf2 mutant and the wild-type SNF2 strains. This suggests that the Swi/Snf complex is not involved in the establishment and maintenance of the defined HIS7 promoter chromatin structure at low Gcn4p-levels (low Gcn4p, adenine limitation). However, the basal transcription of HIS7 is affected by the Swi/Snf complex, since in comparison to wild-type cells the snf2
strain showed a twofold increase in HIS7 mRNA levels (Fig. 6, lanes 1 and 5). The mechanism of Swi/Snf action remains unsolved here. One possibility is that it actually binds as a repressor to the promoter or, alternatively, this might be the result of an indirect effect. However, the transition in chromatin structure during high Gcn4p expression as described above is prevented in the snf2
background. In the absence of SNF2, the chromatin of the HIS7 promoter remains in its ordered structure of the basal transcribed promoter (Fig. 5 B, lanes 1, 2, 7, and 8). Therefore, the Swi/Snf complex is essential for the nucleosomal change upon Gcn4p-overexpression. Consistent with this influence on chromatin, Gcn4p-dependent activation of HIS7 transcription is almost absent in the snf2
background (Fig. 6, lanes 5 and 6).
![]() View larger version (75K): [in a new window] |
FIG. 5. Chromatin structure of the ARO4/HIS7 intergenic region of strains deficient in nucleosome modifying activity. Yeast strains FY1360 (snf2 ), FY1354 (gcn5 ), or FY1352 (snf2 /gcn5 ) were cultivated without exogenous adenine. Experiments were performed with strains expressing either low (A) or high (B) amounts of Gcn4p from wild-type GCN4 (wt) or the additional CGCN4 (constitutively expressed GCN4) allele of p238, respectively.
|
![]() View larger version (49K): [in a new window] |
FIG. 6. HIS7 mRNA levels in yeast cells expressing either a low or a high amount of Gcn4 protein after cultivation without exogenous adenine. Deletions of components required for chromatin modification are as indicated. Low Gcn4p amounts are expressed from the wild-type GCN4 gene (-), whereas high Gcn4p amounts are expressed from plasmid p238 (+) for strains FY1353 (wt), FY1360 (snf2D), FY1354 (gcn5D), and FY1352 (snf2 /gcn5D). Four independent RNA isolations were hybridized twice in Northern experiments and then equalized to ACT1 mRNA levels, resulting in the mean values given in the graph.
|
We then tested whether GCN5 is involved in the Gcn4p-mediated HIS7 transcription. Northern hybridization with total RNA of gcn5
mutant cells revealed that Gcn4p activation of HIS7 transcription is similar to that of the GCN5 wild-type strain (Fig. 6, lanes 3 and 4). Therefore, Gcn4p-dependent HIS7 activation seems to be independent of Gcn5p-containing HAT complexes. However, since the double mutation gcn5
snf2
led to a much stronger reduction of HIS7 transcription than
snf2 by itself does, a synergistic contribution of both activities seems obvious. Chromatin mappings Furthermore, showed that the remodeling of the nucleosomes upon Gcn4p overexpression as described above also takes place in gcn5
mutant cells (Fig. 5B, lanes 5 and 6). Gcn4p-dependent HIS7 gene expression and its associated nucleosome remodeling do not require functional Gcn5p-containing HAT complex. In contrast, the Bas1p/Bas2p-dependent basal HIS7 transcription is halved in a gcn5
background (Fig. 6, lanes 1 and 3). The nucleosome mappings, depicted in Fig. 5A (lanes 7 and 8), show no changes in the absence of Gcn5p.
Gcn5p is required for the Gcn4p-independent activation of the HIS7 promoter by adenine limitation.
Low amounts of Gcn4p are expressed in cells even during the absence of amino acid starvation. It is known that these low amounts are responsible for various promoters to maintain certain expression levels. To confirm that the Gcn5p-dependency of the basal HIS7 expression is mediated through Bas1p/Bas2p and not through low Gcn4p, we tested the role of GCN5 in the gcn4
background. Specific ß-galactosidase activities of cells harboring a translational his7-lacZ fusion instead of the wild-type HIS7 gene were measured to determine HIS7 expression. The activity of the reporter gene is strongly reduced in cells harboring the gcn5
background. In the presence of adenine when his7-lacZ expression is low anyway, deletion of GCN5 further reduces it to ca. 30% of that value (Fig. 7, lanes 1 and 3). Furthermore, any activation of his7-lacZ expression upon adenine limitation is prevented in a gcn5
background (Fig. 7, lane 4). This means that the Bas1p/Bas2p-dependent activation of his7-lacZ requires the HAT activity encoded from GCN5 to enable increased expression upon adenine limitation.
![]() View larger version (33K): [in a new window] |
FIG. 7. HIS7 promoter driving a lacZ reporter in cells defective in GCN4, as well as with defects in chromatin modification. The effects of external adenine on his7-lacZ expression for strains RH1616 (GCN5) and RH2570 (gcn5D) were measured as specific ß-galactosidase activity. The graphs give average values of three individual enzyme assays performed with crude extracts from four independent cultures. Strains were cultivated either in adenine-deficient medium (-Ade) or in medium containing 0.3 mM adenine (+Ade). Specific ß-galactosidase units correspond to nanomoles/(hour · milliliter · optical density at 546 nm)-1.
|
|
|
|---|
![]() View larger version (45K): [in a new window] |
FIG. 8. Model of the alternative mechanisms of nucleosome-dependent transcriptional activation of yeast HIS7 initiated by two different transcription factors. +ade/+aa, Growth in the presence of adenine and amino acids, that is, without any acid starvation; +ade/-aa, growth in presence of adenine but without amino acids, that is, with amino acid starvation; -ade/+aa, growth without adenine but in the presence of amino acids, that is, with purine starvation. Dark circles represent positioned nucleosomes with strong DNA-histone interactions.
|
or bas2
mutant strain (37). Changes of chromatin structure at promoters during gene activation are a common phenomenon previously reported for numerous genes as, e.g., PHO5, PHO8, SUC2, CHA1, HIS4, GAL10, and CUP1 (1, 8, 10, 12, 16, 26, 34). Activators that specifically increase gene expression comprise transcription factors of different DNA-binding motif classes, including the basic-helix-loop-helix activator Pho4p, the acidic Cys6-Zn cluster activators Gal4p or Cha4p, and the basic-leucine-zipper Gcn4p.
Yeast HIS7 is an example of a gene with a promoter chromatin structure that changes upon the activation by one but not the other physiological operating activator (Fig. 7). High Gcn4p levels change the chromatin structure within the ARO4-HIS7 intergenic region and at the same time increase HIS7 transcription. We assume that the simultaneous binding to GCRE1 and GCRE2 causes the observed nucleosome randomization or disruption at the locus. This would support the previously described synergistic nature of HIS7 transcription by the binding of Gcn4p to two binding sites within the HIS7 promoter (37). Different levels of Bas1/2p influence HIS7 transcription (BAS1 overexpression astonishingly even represses it) but do not affect the chromatin structure.
Gcn4p-dependent HIS7 chromatin remodeling demands for a functional Swi/Snf-complex. Genome-wide expression analysis revealed that ca. 6% of all yeast genes are affected twofold or more by the inactivation of Swi/Snf. The affected genes are subdivided into two groups: one with reduced amounts of transcript and the other with an increased transcript level (20, 39). Remodeling by the Swi/Snf-complex can in general have two consequences. One is that repressive nucleosomes are removed from the promoter region, thereby enforcing gene activity. The other is that nucleosomes are positioned to promoter elements, leading to the repression of gene activity.
An active role of the Swi/Snf complex in Gcn4p-mediated activation of the yeast HIS3 gene was previously described (29). However, a preferential accessibility for Mnase to the HIS3 promoter was shown to be a general property of the DNA sequence and not mediated by the Gcn4p-binding site (24). Moreover, several studies have described direct interactions between transcription factors and Swi/Snf that support the recruitment of the remodeling activity to the promoter by binding the DNA-bound activator (30, 46, 48).
The chromatin of the noninduced HIS7-promoter is an accessible open one, sustaining weak HIS7 expression (Fig. 7). This open structure is compatible with the Bas1p/Bas2p-mediated HIS7 activation during adenine limitation and does not require further remodeling. However, this open configuration does not seem to be suitable for high HIS7 expression at high Gcn4p levels. High HIS7 expression levels caused by Gcn4p overexpression result in the randomization or disruption of nucleosomes within the ARO4-HIS7 intergenic region. Mechanistically, this remodeling is coupled to the Swi/Snf complex.
The Bas1p/Bas2p-dependent adenine response on HIS7 expression depends on a functional SAGA/Ada (Gcn5p) complex. The HAT activity of the SAGA or Ada complexes in yeast is encoded by GCN5 and is necessary for the transcriptional activation of several genes (5), including both Gcn4p-regulated and Gcn4p-independent genes. Previous studies reported Gcn5p dependence for the Gcn4p activation of an artificial PHO5 promoter that also harbors a binding site for Gcn4p and was therefore inducible by amino acid starvation. Moreover, the same PHO5 promoter is activated upon phosphate limitation by the transcription factor Pho4p and yet independently of Gcn5p (41). Other genes that require a functional GCN5 gene for the Gcn4p-dependent transcriptional activation are HIS3, TRP3, and ILV1 (13). In contrast, the Gcn4p-dependent activation of the HIS4 and ARG4 genes has been shown to be strictly GCN5 independent (13). In common with these latter genes the Gcn4p response of the HIS7 promoter does not require GCN5. The Gcn4p-binding sites of the HIS4, ARG4, and HIS7 promoters nearly perfectly matches the consensus sequence 5'-TGACTC-3', whereas the HIS3, TRP3, and ILV1 promoters possess weak Gcn4p-binding sites (19, 38). Possibly, the requirement for Gcn5p increases with the decreasing strength of the respective Gcn4p recognition element.
However, since GCN5 is required for basal Bas1p/Bas2p-dependent HIS7 transcription, the situation is complex for this promoter, enabling individual responses to different stimuli by using different mechanisms. Altogether, we demonstrated that the transcriptional regulation of the HIS7 gene by two independent activation pathways uses different chromatin-modifying machineries. Gcn4p changes the nucleosomal distribution at the HIS7 promoter upon its overexpression. This process requires a functional Swi/Snf complex but no functional SAGA/Ada (Gcn5p) complex. Bas1p/Bas2p-dependent HIS7 activation, in contrast, requires a functional SAGA/Ada (Gcn5p) complex but is not associated with chromatin remodeling. Astonishingly, this is possible even though both activators share a common cis element of the HIS7 promoter. Chromatin immunoprecipitation studies shall further elucidate the recruitment of the cofactors to the HIS7 promoter chromatin and their relationship to the bound transcription factors at the different stages of HIS7 expression.
This work was supported by grants from the Deutsche Forschungsgemeinschaft, the Volkswagenstiftung, and the Fonds der Chemischen Industrie.
Present address: Present address: Institute of Cell Biochemistry and Clinical Neurobiology, UKE, D-22529 Hamburg, Germany. ![]()
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»