![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Department of Biochemistry and Molecular Biology, Louisiana State University Health Sciences Center, Shreveport, Louisiana 71130-3932
Received 19 June 2005/ Returned for modification 8 July 2005/ Accepted 15 July 2005
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
3 kb, beginning upstream of the promoter and extending through the transcribed region. A kinetic analysis reveals that histone H4 loses contact with DNA within 45 s of thermal upshift. Nucleosomal reassembly, prompted by temperature downshift, is also rapid, detectable within 60 s. Prior to their eviction, promoter-associated histones are transiently hyperacetylated, while those in the coding region are not. An upstream activation sequence mutation that weakens the binding of heat shock factor obviates domain-wide remodeling, while deletion of the TATA box that nearly abolishes transcription is permissive to 5'-end remodeling. The Swi/Snf complex is rapidly recruited to HSP82 upon heat shock. Nonetheless, domain-wide remodeling occurs efficiently in Swi/Snf mutants despite a sixfold reduction in transcription; it is also seen in gcn5
, set1, and paf1 mutants. Contrary to current models, we demonstrate that a high density of RNA polymerase (Pol) is insufficient to elicit histone displacement. This finding suggests that histone eviction is modulated by factors that are not linked to elongating Pol II. It further suggests that histone depletion plays a causal role in mediating vigorous transcription in vivo and is not merely a consequence of it. |
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Chromatin is intimately involved in the regulation of transcription, and there is much evidence to suggest that its remodeling, particularly over regulatory DNA sequences, plays a key role in both the activation and repression of genes. For example, the regulatory regions of genes poised for or actively engaged in transcription are hypersensitive to digestion by nucleases such as DNase I (for reviews, see references 21 and 30). Such nuclease-hypersensitive sites tend to contain hyperacetylated histones and be depleted of nucleosomes (9, 47, 49, 67). In budding yeast, the promoters of heat shock genes are assembled into constitutive DNase I-hypersensitive sites (22, 83). These sites contain both acetylated and methylated histones under noninducing conditions (16, 76; this study), suggesting that they are comprised of remodeled nucleosomes.
Surprisingly little is known of the chromatin structure of the coding regions of actively transcribed genes, although the great weight of evidence suggests that core histones, particularly H3 and H4, are present. Early studies of chick chromatin revealed that the transcription units of active genes are highly sensitive to digestion by DNase I (27, 89). These observations suggested that active genes might be associated with structurally altered, unfolded nucleosomes. Indeed, chromatin enriched in transcriptionally active sequences exhibits enhanced accessibility of histone H3 sulfhydryl groups and is selectively retained on organomercurial agarose columns (66). In addition, active transcription units may be assembled into nucleosomes that are depleted of H2A and H2B. This is suggested by the fact that nucleosome core particles deficient in H2A and H2B are strongly enriched for transcribed DNA sequences and are preferentially bound by RNA polymerase II (Pol II) (6). More direct evidence for H2A/H2B depletion has come from the finding that transcription of an in vitro-reconstituted mononucleosome results in the loss of one H2A/H2B dimer (41) and recent in vivo evidence that Pol II transcription of coding sequences induces H2A/H2B exchange with free histone pools in Physarum (84).
Actively transcribed chromatin is also enriched in hyperacetylated core histones. This has been shown by nuclease digestion (75), biochemical fractionation (35, 88), and chromatin immunoprecipitation (ChIP) assays (72). One role for this acetylation may be to unfold the nucleosomal fiber, rendering regulatory sites more accessible and transcriptional units more pliable (for a review, see reference 19). Moreover, actively transcribed regions are hypermethylated at Lys4 and Lys79 of H3 (62, 68, 72). These marks may promote the association of coactivators (69) and/or suppress the binding of corepressors (86). Interestingly, transcription-associated histone acetylation is restricted to the promoter regions of two inducible yeast genes, HIS3 and HO (44, 48), suggesting that the correlation between coding region histone modification and transcriptional activation might not always exist.
Cells adapt to stress by rapidly and selectively inducing the transcription of specific genes. The broadly conserved heat shock factor (HSF) activates heat shock protein genes in eukaryotic cells in response to thermal, chemical, or oxidative stress; exposure to heavy metals; and glucose starvation (33, 60, 91). Yeast HSF, encoded by the essential HSF1 gene, contains two potent activation domains (63, 80) and possesses the unusual property of being able to activate transcription in the absence of a number of critical coactivators and general transcription factors. These include TFIIA, TAF9 (TAFII17, a subunit of both TFIID and SAGA), the TFIIH kinase Kin28, the Med17 and Med22 (Srb4 and Srb6) subunits of Mediator (5, 13, 50, 59), and even the carboxyl-terminal domain (CTD) of Pol II (57). Thus, the mechanism by which HSF activates its target genes is of unusual interest.
Here, we have used ChIP to investigate the occupancy and modification state of core histones at HSF-regulated genes in Saccharomyces cerevisiae. In response to heat shock, promoters and coding regions of heat shock genes rapidly lose contact with all histone isoforms. In response to attenuating conditions, histones promptly rebind DNA. At HSP82, heat shock-induced histone loss is strictly dependent upon HSF, yet histone depletion within the gene’s promoter region occurs independently of elevated transcription. Domain-wide histone loss also occurs independently of the Swi/Snf ATP-dependent chromatin remodeling complex, despite severely impaired transcription, and is unaffected by mutations in prominent histone modification and transcription elongation complexes. While a high density of Pol II is seen at the induced HSP82 gene (in both Swi/Snf+ and Swi/Snf – contexts), we provide evidence that elongating Pol II is not of itself sufficient to elicit histone displacement.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
TATA (strain DSG118). The allele termed hsp82-P2, present in strain DSG101, bears a 2-bp substitution mutation in the high-affinity HSF site HSE1. The P2 mutation has been extensively characterized (24, 52, 56). The snf2 deletion was engineered into SLY101 via one-step transplacement with the KANr cassette (31), creating strain JHD204. The knockout was confirmed by genomic PCR. To create a swi1 chromosomal knockout, we crossed a swi1::LEU2 allele into strain EAS2001, creating strain DGY101. Note that EAS2001 contains the hsp82-2001 allele, whose basal and induced transcription is indistinguishable from that of HSP82+ (77).
|
|
|
|
|
|
|
ChIP analysis.
ChIP was performed essentially as previously described (76). Cells were cross-linked with 1% formaldehyde, then converted to spheroplasts with lyticase, and lysed using 1 volume of 0.5-mm glass beads for 30 min at 4°C on an Eppendorf 5432 mixer. Chromatin was sheared to a mean length of 0.4 kb with a Branson 250 sonifier equipped with a microtip (three 25-s pulses at constant power and an output setting of 22 W). Antibodies specific for the following epitopes were used: histone H3 (unacetylated isoform), diacetyl histone H3 (K9 and K14), tetra-acetyl histone H4 (K5, K8, K12, and K16), monoacetyl histone H2A (K7), and dimethyl histone H3 (K4) (all obtained from Upstate Biotechnology); trimethyl histone H3 (K4) (Abcam); Myc (monoclonal antibody [MAb] 9E10; Santa Cruz Biotechnology); and mouse Pol II CTD (a gift of David Bentley, University of Colorado Health Sciences Center) (71). IPs were achieved by adding generally either 3 µl of polyclonal antiserum or 5 µl of MAb to 300 µl of chromatin lysate. They were mixed on a nutator at 4°C overnight. Pansorbin cells (40 µl; Calbiochem) were then added, and the incubation was continued for an additional 3 h. For Pol II ChIPs, the 300-µl chromatin lysate was preincubated with 50 µl blocked protein A-Sepharose beads for 3 h at 4°C, and then 6 to 8 µl of Pol II antibody and 3 µl of 30% Sarkosyl were added to the clarified supernatant and permitted to incubate overnight at 4°C. A total of 50 µl of fresh protein A beads was then added, and the mixture was incubated at 4°C for 2 h. Beads were then washed as previously described (76).
Following washing and reversal of formaldehyde-induced cross-links, DNA was ethanol precipitated and dissolved in 30 µl Tris-EDTA (TE). Two microliters was then used as a template in a PCR. For input, 200 µl of soluble chromatin was ethanol precipitated and dissolved in TE, and cross-links were reversed. The chromatin was reprecipitated, and DNA was purified and dissolved in 150 µl TE. One microliter of this was used as template in a PCR. In addition to template DNA, the 50-µl reaction mixtures contained 2.5 mM MgCl2; 400 µM (each) dCTP, dGTP, dTTP, and dATP; and 1 µCi of [
-32P]dATP (250 Ci/mmol). After 2 min of denaturation at 93°C and addition of 1.25 U Taq DNA polymerase, the temperature was lowered to 60°C for 1 min, followed by 32 s at 72°C. Samples were then subjected to a program of 22 to 24 cycles, each consisting of 1 min at 93°C, 1 min at 60°C, and 32 s at 72°C. PCR products were precipitated, electrophoresed on 8% Tris-borate-EDTA polyacrylamide gels, dried, exposed to a Phosphor Screen, and quantified on a Storm 860 PhosphorImager (Molecular Dynamics) utilizing ImageQuant 5.2 software. Linearity of the PCR was tested by assaying threefold serial dilutions of 15 individual templates. The mean increase in multiplex PCR signal for the heavier loaded sample of each pair was 2.2 (see Fig. 8A).
|
To calculate the relative abundance of a given gene promoter or coding sequence present in an IP, we used the following formula: Qgene = IPgene/inputgene. For histone ChIPs, the abundance of each test promoter or coding region is expressed relative to that of the PHO5 promoter, which served as an internal recovery control (76). For the Snf2-Myc ChIP assay (see Fig. 7B), nonspecific background (the PCR signal arising from immunoprecipitation of chromatin isolated from a non-Myc-tagged strain) was subtracted from each specific IP signal prior to calculating Qgene. For Pol II ChIPs, abundance at each hsp82 region is expressed relative to its abundance at a neutral locus on chromosome V (SCR1, a 521-bp Pol III gene closely flanked by two non-heat-inducible Pol II genes, YERWdelta17 and YER137W-A). To accurately quantify Pol II abundance, we subtracted the signal arising from a mock ChIP (–Ab) from the corresponding ChIP signal (+Ab) (see Fig. 8A) prior to calculating Qgene.
Northern analysis. RNA was isolated, purified, electrophoresed, and blotted to Gene Screen as previously described (77). Blots were sequentially hybridized to HSP82- and SCR1-specific probes at 45°C and 55°C, respectively. The HSP82-specific probe was generated by 25 cycles of linear PCR using as a template an oligonucleotide corresponding to the transcript's unique 3' UTR (spanning +2167 to +2228 relative to ATG). The SCR1 probe was generated by 25 cycles of linear PCR amplification of the gene's region spanning +343 to +467 with the SCR1 PCR fragment as a template.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
Two further observations suggest that the apparent loss of histone occupancy at HSP genes reflects their physical displacement from DNA and not, for example, inefficient cross-linking caused by heat shock. First, the effect is highly specific to strongly activated heat shock genes. There was no decrease in histone cross-linking at the coamplified PHO5 promoter, nor was there any diminishment of histones within either the promoter or coding sequence of a less vigorously transcribed HSP gene (see Fig. 6B). Second, nonhistone proteins such as HSF, TATA-binding protein (TBP), and Pol II are readily detected within the promoters and coding regions of activated heat shock genes (see Fig. 8A) (76). Together with previous observations (16, 23), the results indicate that chromatin spanning entire heat shock gene domains is grossly altered upon heat shock, an alteration that likely involves the loss of histone octamers.
Chromatin remodeling is confined to a circumscribed domain encompassing the 5' flank, ORF, and 3' UTR of HSP82. To delimit the extent of heat shock-induced remodeling, we assayed the histone content of the two genes flanking HSP82. This was of interest, since it was unclear whether the dramatic reduction in histone abundance at HSP82 was locus-wide or circumscribed to the heat shock gene's domain. Using ChIP as above, we found that Myc-H4 was not significantly depleted within the coding region of either YAR1 or CIN2 upon heat shock (Fig. 3C). However, it was 50% reduced within the YAR1-HSP82 intergenic region (position –485 relative to HSP82). Therefore, while heat shock-induced histone displacement spanned a domain of nearly 3 kb, it did not spread into the closely flanking neighboring genes. This was true even though the coding region of one of them, CIN2, begins just 110 bp downstream of the 3'-mapped end of the HSP82 transcript (25). Indeed, histones were hypoacetylated within both YAR1 and CIN2 coding regions, concomitant with their heat shock-induced transcriptional repression (data not shown). This finding is consistent with previous observations correlating coding region histone hypoacetylation with transcriptional down-regulation (10, 46).
Kinetics of domain-wide remodeling are remarkably rapid. We next sought to determine the kinetics with which histones are displaced. To do so, we formaldehyde-cross-linked cells at various times following a rapid 30 to 39°C upshift. This approach afforded instantaneous snapshots of protein-DNA interactions, since formaldehyde rapidly induces cross-linking and simultaneously freezes all metabolic processes (39), including those induced by heat shock (Iqbal and Gross, unpublished). As shown in Fig. 4A, Myc-H4 dissociation at HSP82 was detectable within 45 s (14 to 28% loss) and was essentially complete by 10 min (80 to 88% loss). Notably, displacement occurs over the three assayed regions with similar, although not necessarily identical, kinetics. Histone loss within the promoter is detectable as early as 30 s of postthermal upshift, whereas comparable changes in the ORF and 3' UTR do not appear until 45 s. This difference might reflect distinct mechanisms and is further explored below. An additional and intriguing finding is that histone loss is transient. Histone-DNA contacts were reestablished 30 to 60 min into the heat shock (Fig. 4A), coincident with a decline in the rate of HSP82 transcription (data not shown).
To extend this analysis, we also examined the kinetics with which Myc-H4 reestablished contact with DNA upon a 39 to 30°C downshift (recovery). If histone loss reflects nucleosomal disassembly, its restoration may reflect de novo nucleosomal assembly, a process thought to be generally dependent on DNA replication. We found that Myc-H4 rapidly rebound to all three regions of HSP82 in response to temperature downshift, a condition that results in drastically reduced HSP82 transcription (77). This was detectable as early as 60 s and was essentially complete by 20 min (Fig. 4B; see also Fig. 3B). We conclude that histone H4 and, by extension, the H3-H4 tetramer itself reassociates with HSP82 with rapid kinetics over the entire length of the gene. Given the rapidity of the kinetics and the fact that the cell population is asynchronous, these striking findings argue against a replication-coupled process and instead suggest that nucleosome reassembly is replication independent (see Discussion).
A transient burst of histone acetylation at the HSP82 promoter precedes gene-wide histone displacement. The histone code hypothesis (38, 81) postulates that covalent modifications within histone tails serve as essential recognition sites for the binding of factors including, for example, ATP-dependent remodeling enzymes (2, 34) and histone chaperones (1). Thus, covalent modifications of histones might precede their dissociation from DNA, as has been recently shown for the activated PHO5 promoter (67). We assayed the abundance of both acetylated and methylated histones throughout the HSP82 gene during the first 60 s of heat shock. Consistent with a role for lysine acetylation, we found that the relative abundance of acetylated histones increases 40 to 50% during the first 45 s of heat shock (Fig. 4C). This enrichment was short lived and was restricted to the promoter. At 60 s following heat shock, there was a marked reduction in acetylated H2A and H4. As above, this change took place relative to total histone (calculated as isoform occupancy/Myc-H4 occupancy), suggesting that nucleosomes acetylated at H2A and H4 are preferentially targeted for eviction. Notably, the net abundance of promoter-associated trimethylated H3 precipitously decreased during the first minute, arguing against a role for H3 K4 methylation in domain-wide remodeling. We conclude that heat shock-induced acetylation of promoter-associated nucleosomes precedes, by a matter of seconds, their disassembly.
Remodeling of the hsp82 promoter and 5' UTR occurs independently of elevated transcription.
Demonstration of domain-wide histone displacement at HSF-regulated genes raises the question of what underlies this phenomenon. To test whether histone displacement at HSP82 is dependent upon transcription, we engineered a 19-bp substitution of its core promoter (Fig. 5A), encompassing the TATA box and coinciding with a region strongly protected from hydroxyl radical cleavage in chromatin (29). The mutant, termed TATA, exhibited a 23-fold reduction in heat shock-induced transcription (Fig. 5B). Despite this, partial displacement of each histone isoform, as well as of Myc-H4, took place over the gene's promoter in response to acute heat shock (Fig. 5C and D). By comparison, there was little change in histone abundance over the ORF or 3' UTR (Fig. 5D). Virtually identical results (data not shown) have been observed in a strain bearing a less crippling mutation in the hsp82 TATA box (a 2-bp substitution, termed T2, converting TATAAA to TAGCAA) and exhibiting a 12-fold reduction in induced transcription (52). We conclude that the dramatic remodeling which takes place at the 5' end of HSP82 is at least partially independent of elevated transcription.
Mutation of HSE1 blocks histone displacement.
We next asked whether HSF itself is critical to domain-wide remodeling. To do so, we assayed histone abundance at an hsp82 promoter mutant bearing a 2-bp mutation in HSE1, the high-affinity HSF binding site (Fig. 6A). This mutation, termed P2, significantly weakens HSF binding to the hsp82 promoter both in vivo and in vitro (24, 56) and drastically reduces noninduced expression. hsp82-P2 nonetheless activates transcription to a level which exceeded that of hsp82-TATA (Fig. 5B; see Fig. 9B). In contrast to the TATA box mutants, however, the upstream activation sequence (UAS) mutant showed little evidence of histone displacement upon heat shock (Fig. 6B). Moreover, despite the nearly 50-fold increase in transcription, neither the promoter nor the coding region of hsp82-P2 showed evidence of increased histone acetylation or methylation (Fig. 6C). Therefore, weakening of HSF's binding to the hsp82 promoter effectively eliminates heat shock-induced histone modification and displacement.
Swi/Snf is rapidly recruited to the activated HSP82 gene, is required for its vigorous transcription, yet is dispensable for domain-wide remodeling.
We next sought to test the contribution of Swi/Snf, given its critical role in the transcriptional activation of such inducible yeast genes as HO, SUC2, INO1, and PHO8 (for a review, see reference 90) and its demonstrated role in regulating transcription, particularly elongation, of the mammalian hsp70 gene (14, 17). As shown in Fig. 7A, heat shock-induced HSP82 expression was reduced >6 fold in cells bearing a deletion of either SNF2 or SWI1. Consistent with this, Myc-tagged Snf2 was rapidly recruited to the HSP82 promoter in response to heat shock (Fig. 7B), and its abundance positively correlates with that of HSF (32, 76). Swi/Snf was also detected within the HSP82 coding region, although with delayed kinetics and reduced occupancy. Recruitment of Swi/Snf suggests that the expression phenotypes associated with the snf2 and swi1 mutations are due to a direct role for Swi/Snf in regulating heat shock-induced transcription. Despite this, Myc-H4 was efficiently displaced from all three regions of the gene following acute heat shock of either mutant (Fig. 7C and D). This outcome is especially striking, given that histone depletion was not seen in the UAS mutant hsp82-P2 (Fig. 6B), whose rate of activated transcription is similar to that of the Swi/Snf mutants (see Fig. 9B, lane 13 versus lane 14). Therefore, even though HSP82 transcription is Swi/Snf dependent, remodeling of both promoter and coding region occurs efficiently in its absence.
Pol II density at HSP82 is unaffected by either a SNF2 gene deletion or a UAS promoter mutation.
To explore the basis for the contrasting chromatin phenotypes of the P2 and snf2 mutants, we investigated Pol II density within the hsp82 promoter, ORF, and 3' UTR of the wild-type strain and of the two isogenic mutants. If histone displacement is a consequence of elongating Pol II, then Pol II abundance should be high in cases where domain-wide histone displacement is observed (HSP82+) and low in cases where it is not (hsp82-P2). Indeed, ChIP revealed that Pol II density was high within all three regions of HSP82+ in both SNF2+ and snf2 cells following a 20-min heat shock (Fig. 8A and B). As expected, abundance of histone H4 negatively correlated with that of Pol II in the same chromatin samples (compare Fig. 8C, lanes 3 and 7, with 8A, lanes 7 and 15). These results are consistent with the idea that elongating Pol II underlies histone displacement and are in agreement with results obtained with the hsp82-TATA strain that showed a low level of induced transcription and no depletion of coding region histones. Interestingly, since Pol II density was not diminished in snf2 cells despite a sixfold reduction in transcription, this suggests that loss of Swi/Snf impairs not only the rate of initiation but also the rate of elongation (see Discussion).
Arguing against a correlation between Pol II density and histone loss, however, are results obtained with the P2 mutant. Despite the fact that activated transcription was significantly reduced (Fig. 5B and 9B), Pol II density at the heat shock gene was unaffected (Fig. 8A, compare lane 11 with lane 7; Fig. 8B). This suggests that similar to snf2, the P2 mutation affects the rate of both transcription initiation and elongation. In contrast to the case with snf2, however, the P2 mutation diminished Pol II density under non-heat-shocked conditions (Fig. 8B) and, more importantly, abrogated histone displacement following HS (Fig. 8C, lane 5; Fig. 6B). This latter observation indicates that a high density of Pol II within the hsp82 coding region is insufficient, by itself, to bring about histone displacement.
Nucleosomal disassembly is independent of the Gcn5 and Set1 histone-modifying enzymes as well as the Paf1 elongation complex. To investigate additional coactivators that might underlie domain-wide histone displacement (and thereby be recruited to HSP82+ but not to hsp82-P2), we constructed isogenic strains bearing deletions of GCN5, SET1, or PAF1. Gcn5, the catalytic subunit of SAGA, was of interest, since the recruitment of SAGA to a variety of yeast promoters has been demonstrated. Indeed, it is a primary target of both Gal4 and Gcn4 (8, 12, 26, 82), and HSF-mediated trans-activation of target genes in response to stress is thought to occur principally through the SAGA (as opposed to the TFIID) pathway (92). Moreover, the finding that promoter-associated H3 was transiently hyperacetylated at Lys9 and Lys14 in response to heat shock (Fig. 4C) is consistent with a role for Gcn5, given its preference for acetylating these residues in vivo (48). However, as shown in Fig. 9, deletion of GCN5 had no detectable effect on either HSP82 expression (Fig. 9B) or histone displacement (Fig. 9A). This suggests that SAGA's role in the activation of HSP82 may derive more from its ability to facilitate TBP binding to core promoters (18) than from its ability to acetylate chromatin.
We next examined the role of the Set1 histone methyltransferase, given its suggested role in Pol II elongation at actively expressed yeast genes. While the absence of enhanced H3 K4 methylation at the induced HSP82 gene (Fig. 1 and 4C) would seem to argue against a role for Set1, it is possible that this modification occurs over a discrete region of HSP82 that eluded detection by the primer pairs we employed. In particular, H3 K4 trimethylation has been reported to be maximal within the 5' coding regions of activated genes (62, 68); thus a comparable, albeit transient, modification may occur over the HSP82 coding region. However, as shown in Fig. 9, a set1 mutation had little impact on either the extent of histone displacement or transcriptional activation of HSP82, arguing against a role for the Set1 complex, COMPASS, in regulating domain-wide histone displacement.
Finally, we investigated whether the Paf1 complex played a role in domain-wide histone displacement at HSP82. Paf1 associates with Pol II at multiple stages during the transcription cycle (for a review, see reference 61) and has been detected within the coding regions of HSF-regulated genes, including HSP82 and SSA4 (65). Despite this, deletion of the PAF1 gene had no effect on histone displacement, although it did reduce activation approximately twofold (Fig. 9A and B).
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
It is noteworthy that an earlier study employing nuclease digestion of chromatin also showed transcription-dependent alterations in nucleosomal structure over the HSP82 coding region (51). However, these alterations, which consisted of enhanced DNase I sensitivity and half-nucleosomal cleavage periodicity, were present in nuclei obtained from both noninduced and heat-shocked cells bearing the wild-type HSP82 gene; full-nucleosome cleavage periodicity was seen at two transcriptionally quiescent hsp82 mutants (51). Therefore, structurally altered (and possibly unfolded) nucleosomes may exist within the coding region of basally transcribed HSP82. The ChIP experiments presented here suggest that such nucleosomes are rapidly displaced upon heat shock. In the earlier study, these displaced histones may have partially reassociated over the coding DNA during isolation of nuclei and subsequent DNase I digestion, thereby masking the dramatic alterations documented here.
There is considerable evidence that the coding regions of actively transcribed eukaryotic genes contain altered nucleosomes. These altered nucleosomes have variously been shown to be unfolded (66), to contain hyperacetylated and hypermethylated core histones (35, 62, 68, 88), and to be depleted in histones H2A and H2B (6, 41). It is thus possible that domain-wide loss of histone octamers is peculiar to yeast heat shock genes. However, this is unlikely. First, electron micrographs have been obtained that show loss of nucleosomes over the coding regions of yeast rRNA genes (15). Second, recently published ChIP analyses demonstrate partial loss of histones over the GAL1 and GAL10 coding regions, following galactose induction (45, 49, 73). Therefore, nucleosome depletion within gene coding regions may be a characteristic of vigorously transcribed genes. Indeed, as alluded to above, a recent analysis of green fluorescent protein-H3 distribution in Drosophila polytene chromosome spreads reveals that H3 is rapidly lost from heat shock gene loci upon heat shock. In contrast to yeast heat shock genes, however, H3 is efficiently replaced by a variant, H3.3, and such replacement occurs early during transcriptional induction (74).
Distinct mechanisms of histone displacement over promoter and coding region.
Evidence presented here suggests that domain-wide histone displacement at HSP82 occurs in two discrete steps: (i) 5'-end remodeling that is dependent on a stably bound activator and (ii) coding region remodeling that is dependent on elongating Pol II. A critical role for HSF in triggering domain-wide displacement is indicated by the phenotype of the hsp82-P2 allele, which bears a 2-bp mutation in HSE1 that significantly reduces HSF binding to the UAS (24, 56). At this allele, domain-wide histone loss is obviated. In addition, elongating RNA polymerase is required for disrupting coding region chromatin. This conclusion is based on the phenotype of the hsp82-TATA allele, whose promoter is efficiently remodeled upon heat shock, while its coding region is not. (Polymerase's role in histone displacement is more extensively discussed below.) Since activated transcription is more severely affected in hsp82-TATA than in hsp82-P2, disruption of promoter chromatin in the TATA mutant likely stems from activated HSF (and the factors it recruits). A similar distinction between proximal and distal chromatin remodeling has been previously shown for the human hsp70 gene (11). That HSF is in fact bound to the UAS of the hsp82-TATA gene has not been formally shown. However, chromatin footprinting analysis of the closely related hsp82-T2 allele reveals that promoter-associated DNase I hypersensitivity, a signature of DNA-bound HSF (28), is fully retained (52). Thus, it is probable that HSF constitutively binds the promoter region of hsp82-T2, and by extension, to that of hsp82-TATA.
An additional mechanistic difference between promoter versus coding region remodeling is that the former is characterized by a burst of histone acetylation that precedes nucleosomal disassembly by a matter of seconds. No comparable modification is seen over the coding region. It is tempting to speculate that these promoter-specific marks, which are especially prominent on H2A and H4, facilitate the recruitment of bromodomain-containing enzymatic complexes (2, 34) that ultimately catalyze domain-wide histone displacement. Given the in vitro preference of Esa1 to acetylate H2A and H4 (85), it is possible that this histone acetyltransferase marks promoter-associated nucleosomes for eviction. Acetylated octamers appear to be preferentially targeted, since acetylated H2A and H4 are disproportionately depleted (relative to that of Myc-H4) at time points subsequent to 45 s.
Elongating Pol II is necessary but not sufficient to trigger nucleosomal disassembly.
The requirement of Pol II elongation for coding region disruption suggests the possibility that these chromatin alterations are simply a consequence of high rates of transcription. In support of this idea is the finding that even though heat shock-induced HSP82 transcript levels are drastically reduced in a snf2 mutant, Pol II density remains unchanged. A high density of Pol II within the gene's coding region provides a plausible explanation for the efficient displacement of histones that is seen in response to heat shock in Swi/Snf mutants. Furthermore, histone-DNA contacts are lost at the coding regions of galactose-induced GAL genes, and such loss is both correlated with high rates of transcription and dependent on the presence of elongating Pol II (45, 49, 73).
Nevertheless, we have obtained evidence that RNA polymerase is not, of itself, sufficient to elicit histone eviction. This is the case, since Pol II density is as high at the heat shock-induced hsp82-P2 gene (which exhibits both reduced transcription and no detectable loss of histones) as at wild-type HSP82. Therefore, activities in addition to elongating Pol II must contribute to the disruption of coding region chromatin. We suggest that activated HSF, tightly bound to the HSP82+ promoter, triggers the recruitment of remodeling activities that efficiently disassociate histones over the length of the entire gene. Such histone-displacing activities, by definition, would fail to be recruited to hsp82-P2, where HSF's occupancy of the UAS is significantly reduced (24). Importantly, the finding that high Pol II density can coexist with nucleosomes over the length of a gene is not simply an oddity of the hsp82-P2 mutant. We see a similar coupling of high Pol II density with high histone occupancy at the induced SSA3 gene, whose promoter appears to be weakly bound by HSF (J. Zhao, J. Iqbal, and D. S. Gross, unpublished data). Given the absence of a correlation between Pol II abundance and nucleosomal disassembly, it is probable that coding region histone displacement plays a causal role in the vigorous transcription seen in genes such as HSP82, HSP26, SSA4, and GAL1 and is not merely a consequence of it.
A role for Swi/Snf in governing Pol II elongation rate in vivo.
An additional implication of these experiments is that Swi/Snf affects the rate of Pol II elongation at HSP82. This is based on the finding that, as discussed above, Pol II density within the HSP82 coding region is unaffected by an snf2 mutation despite a sixfold reduction in transcription. Moreover, Pol II abundance within the promoter is actually higher in heat-shocked snf2 cells than in SNF2+ cells, consistent with Pol II having an increased dwell time and/or impairment in promoter escape. In this respect, the role of yeast Swi/Snf is reminiscent of that demonstrated for mouse Swi/Snf (BRG1), which has been shown to mediate release of paused Pol II within the 5' UTR of the hsp70 heat shock gene (14). Given the striking similarity of the snf2 and P2 phenotypes, the apparent elongation defect of the snf2 mutant might be a consequence of impaired binding by HSF to the chromatinized UAS, whose default state is nucleosomal (28, 87). Thus, the role of Swi/Snf in governing the elongation rate could be indirect. It will be interesting to test whether other Swi/Snf-dependent genes similarly retain a high density of Pol II in the context of an inactivating Swi/Snf mutation.
Implications of domain-wide nucleosomal disassembly and reassembly.
Nucleosomes carry important epigenetic information (3, 79). This information exists in the pattern of covalent modifications of the constituent histones, as well as in the presence of histone variants. Normally, this information is preserved during such processes as replication, recombination, and transcription. An intriguing finding is that histone H4 rapidly reassociates with HSP82 in response to thermal downshift, detectable within 60 s and restoring its original level within 20 min; it also reassociates with DNA during continuous heat shocks of 
30 min. The rapidity of Myc-H4 deposition, presumably accompanying its reassembly into nucleosomes, implies that such a process occurs independently of DNA replication. Replication-independent assembly of the histone variant H3.3 has been previously shown to mark transcriptionally active genes in Drosophila (3), and the Swr1 complex mediates ATP-dependent, replication-independent exchange of the H2AZ variant in S. cerevisiae (43, 58). Our results suggest that if epigenetic information is preserved at activated yeast heat shock genes, then retention of as little as 20% of preexisting H3-H4 tetramers may suffice to seed the domain-wide formation of nucleosomes with a similar epigenetic profile. This residual 20% may also be necessary for tethering Swi/Snf to HSP82's promoter and coding region during heat shock (Fig. 7B). We did not examine the composition of nucleosomes reassembling heat shock gene domains following recovery from heat shock. It might be anticipated that their modification profile and histone variant composition will resemble those seen in the pre-heat-shocked state, although the possibility of such nucleosomes bearing deposition-specific marks cannot be ruled out (93).
Role of other coactivators in domain-wide nucleosomal disassembly. We have shown here that a number of prominent remodeling and modification enzymes, such as Swi/Snf, Gcn5, and Set1, as well as the Paf1 elongation complex, are dispensable for domain-wide histone displacement. In addition, in other experiments we have found that the histone H3/H4 binding and deposition protein, Asf1, is not involved in histone displacement at HSP82 (unpublished data). This contrasts with Asf1's pivotal role in histone displacement from the activated PHO5 and PHO8 promoters (1). Absence of a detectable role for any of these factors may due to their being functionally redundant or because other proteins (such as Esa1) mediate nucleosomal disassembly at yeast heat shock genes. Whatever their identity, our work suggests that such enzymatic activities are recruited by HSF to wild-type HSP82, yet fail to be recruited to the 2-bp HSE1 mutant hsp82-P2.
|
ACKNOWLEDGMENTS |
|---|
TATA strain, David Bentley for providing RNA Pol II antiserum, Bill Garrard for the gift of the hsp82-T2 strain, Ned Sekinger and Kevin Struhl for providing the myc-HHF2 expression vector, and Selena Kremer, Sri Balakrishnan, and Ricky DeBenedetti for helpful comments on the manuscript. D.S.G. wishes to dedicate this paper in memory of his father.
This work was supported by grants to D.S.G. from the National Science Foundation (MCB-0350190 and MCB-0450419), the National Institutes of Health (GM45842), and the Center for Excellence in Cancer Research at the Louisiana State University Health Sciences Center.
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2. Agalioti, T., G. Chen, and D. Thanos. 2002. Deciphering the transcriptional histone acetylation code for a human gene. Cell 111:381-392.[CrossRef][Medline]
3. Ahmad, K., and S. Henikoff. 2002. The histone variant H3.3 marks active chromatin by replication-independent nucleosome assembly. Mol. Cell 9:1191-1200.[CrossRef][Medline]
4. Amoros, M., and F. Estruch. 2001. Hsf1p and Msn2/4p cooperate in the expression of the Saccharomyces cerevisiae genes HSP26 and HSP104 in a gene- and stress type-dependent manner. Mol. Microbiol. 39:1523-1532.[CrossRef][Medline]
5. Apone, L. M., C. A. Virbasius, F. C. P. Holstege, J. Wang, R. A. Young, and M. R. Green. 1998. Broad, but not universal, transcriptional requirement for yTAFII17, a histone H3-like TAFII present in TFIID and SAGA. Mol. Cell 2:653-661.[CrossRef][Medline]
6. Baer, B. W., and D. Rhodes. 1983. Eukaryotic RNA polymerase II binds to the nucleosome cores of transcribed genes. Nature 301:482-488.[CrossRef][Medline]
7. Belotserkovskaya, R., A. Saunders, J. T. Lis, and D. Reinberg. 2004. Transcription through chromatin: understanding a complex FACT. Biochim. Biophys. Acta 1677:87-99.[Medline]
8. Bhaumik, S. R., T. Raha, D. P. Aiello, and M. R. Green. 2004. In vivo target of a transcriptional activator revealed by fluorescence resonance energy transfer. Genes Dev. 18:333-343.
9. Boeger, H., J. Griesenbeck, J. S. Strattan, and R. D. Kornberg. 2003. Nucleosomes unfold completely at a transcriptionally active promoter. Mol. Cell 11:1587-1598.[CrossRef][Medline]
10. Braunstein, M., A. B. Rose, S. G. Holmes, C. D. Allis, and J. R. Broach. 1993. Transcriptional silencing in yeast is associated with reduced nucleosome acetylation. Genes Dev. 7:592-604.
11. Brown, S. A., and R. E. Kingston. 1997. Disruption of downstream chromatin directed by a transcriptional activator. Genes Dev. 11:3116-3121.
12. Bryant, G. O., and M. Ptashne. 2003. Independent recruitment in vivo by Gal4 of two complexes required for transcription. Cell 11:1301-1309.
13. Chou, S., S. Chatterjee, M. Lee, and K. Struhl. 1999. Transcriptional activation in yeast cells lacking transcription factor IIA. Genetics 153:1573-1581.
14. Corey, L. L., C. S. Weirich, I. J. Benjamin, and R. E. Kingston. 2003. Localized recruitment of a chromatin remodeling activity by an activator in vivo drives transcriptional elongation. Genes Dev. 17:1392-1401.
15. Dammann, R., R. Lucchini, T. Koller, and J. M. Sogo. 1993. Chromatin structures and transcription of rDNA in yeast Saccharomyces cerevisiae. Nucleic Acids Res. 21:2331-2338.
16. Deckert, J., and K. Struhl. 2001. Histone acetylation at promoters is differentially affected by specific activators and repressors. Mol. Cell. Biol. 21:2726-2735.
17. de la Serna, I. L., K. A. Carlson, D. A. Hill, C. J. Guidi, R. O. Stephenson, S. Sif, R. E. Kingston, and A. N. Imbalzano. 2000. Mammalian SWI-SNF complexes contribute to activation of the hsp70 gene. Mol. Cell. Biol. 20:2839-2851.
18. Dudley, A. M., C. Rougeulle, and F. Winston. 1999. The Spt components of SAGA facilitate TBP binding to a promoter at a post-activator-binding step in vivo. Genes Dev. 13:2940-2945.
19. Eberharter, A., and P. B. Becker. 2002. Histone acetylation: a switch between repressive and permissive chromatin. EMBO Rep. 3:224-229.[CrossRef][Medline]
20. Eickbush, T. H., and E. N. Moudrianakis. 1978. The histone core complex: an octamer assembled by two sets of protein-protein interactions. Biochemistry 17:4955-4964.[CrossRef][Medline]
21. Elgin, S. C. R. 1988. The formation and function of DNase hypersensitive sites in the process of gene activation. J. Biol. Chem. 263:19259-19262.
22. Erkine, A. M., C. C. Adams, M. Gao, and D. S. Gross. 1995. Multiple protein-DNA interactions over the yeast HSC82 heat shock gene promoter. Nucleic Acids Res. 23:1822-1829.
23. Erkine, A. M., and D. S. Gross. 2003. Dynamic chromatin alterations triggered by natural and synthetic activation domains. J. Biol. Chem. 278:7755-7764.
24. Erkine, A. M., S. F. Magrogan, E. A. Sekinger, and D. S. Gross. 1999. Cooperative binding of heat shock factor to the yeast HSP82 promoter in vivo and in vitro. Mol. Cell. Biol. 19:1627-1639.
25. Farrelly, F. W., and D. B. Finkelstein. 1984. Complete sequence of heat shock-inducible HSP90 gene of Saccharomyces cerevisiae. J. Biol. Chem. 259:5745-5751.
26. Fishburn, J., N. Mohibullah, and S. Hahn. 2005. Function of a eukaryotic transcription activator during the transcription cycle. Mol. Cell 18:369-378.[CrossRef][Medline]
27. Garel, A., and R. Axel. 1976. Selective digestion of transcriptionally active ovalbumin genes from oviduct nuclei. Proc. Natl. Acad. Sci. USA 73:3966-3970.
28. Gross, D. S., C. C. Adams, S. Lee, and B. Stentz. 1993. A critical role for heat shock transcription factor in establishing a nucleosome-free region over the TATA-initiation site of the yeast HSP82 heat shock gene. EMBO J. 12:3931-3945.[Medline]
29. Gross, D. S., K. E. English, K. W. Collins, and S. Lee. 1990. Genomic footprinting of the yeast HSP82 promoter reveals marked distortion of the DNA helix and constitutive occupancy of heat shock and TATA elements. J. Mol. Biol. 216:611-631.[CrossRef][Medline]
30. Gross, D. S., and W. T. Garrard. 1988. Nuclease hypersensitive sites in chromatin. Annu. Rev. Biochem. 57:159-197.[CrossRef][Medline]
31. Guldener, U., S. Heck, T. Fiedler, J. Beinhauer, and J. H. Hegemann. 1996. A new efficient gene disruption cassette for repeated use in budding yeast. Nucleic Acids Res. 24:2519-2524.
32. Hahn, J.-S., Z. Hu, D. J. Thiele, and V. R. Iyer. 2004. Genome-wide analysis of the biology of stress responses through heat shock transcription factor. Mol. Cell. Biol. 24:5249-5256.
33. Hahn, J.-S., and D. J. Thiele. 2004. Activation of the Saccharomyces cerevisiae heat shock transcription factor under glucose starvation conditions by Snf1 protein kinase. J. Biol. Chem. 279:5169-5176.
34. Hassan, A., P. Prochasson, K. Neely, S. Galasinski, M. Chandy, M. Carrozza, and J. Workman. 2002. Function and selectivity of bromodomains in anchoring chromatin-modifying complexes to promoter nucleosomes. Cell 111:369-379.[CrossRef][Medline]
35. Hebbes, T. R., A. W. Thorne, and C. Crane-Robinson. 1988. A direct link between core histone acetylation and transcriptionally active chromatin. EMBO J. 7:1395-1402.[Medline]
36. Jackson, V. 1988. Deposition of newly synthesized histones: hybrid nucleosomes are not tandemly arranged on daughter DNA strands. Biochemistry 27:2109-2120.[CrossRef][Medline]
37. Jackson, V. 1990. In vivo studies on the dynamics of histone-DNA interaction: evidence for nucleosome dissolution during replication and transcription and a low level of dissolution independent of both. Biochemistry 29:719-731.[CrossRef][Medline]
38. Jenuwein, T., and C. D. Allis. 2001. Translating the histone code. Science 293:1074-1080.
39. Katan-Khaykovich, Y., and K. Struhl. 2002. Dynamics of global histone acetylation and deacetylation in vivo: rapid restoration of normal histone acetylation status upon removal of activators and repressors. Genes Dev. 16:743-752.
40. Keener, J., J. A. Dodd, D. Lalo, and M. Nomura. 1997. Histones H3 and H4 are components of upstream activation factor required for the high-level transcription of yeast rDNA by RNA polymerase I. Proc. Natl. Acad. Sci. USA 94:13458-13462.
41. Kireeva, M. L., W. Walter, V. Tchernajenko, V. Bondarenko, M. Kashlev, and V. M. Studitsky. 2002. Nucleosome remodeling induced by RNA polymerase II: loss of H2A/H2B dimer during transcription. Mol. Cell 9:541-552.[CrossRef][Medline]
42. Knop, M., K. Siegers, G. Pereira, W. Zachariae, B. Windsor, K. Nasmyth, and E. Schiebel. 1999. Epitope tagging of yeast genes using a PCR-based strategy: more tags and improved practical routines. Yeast 15:963-972.[CrossRef][Medline]
43. Kobor, M. S., S. Venkatasubrahmanyam, M. D. Meneghini, J. W. Gin, J. L. Jennings, A. J. Link, H. D. Madhani, and J. Rine. 2004. A protein complex containing the conserved Swi2/Snf2-related ATPase Swr1p deposits histone variant H2A.Z into euchromatin. PLoS Biol. 2:E131.[CrossRef][Medline]
44. Krebs, J. E., M.-H. Kuo, C. D. Allis, and C. L. Peterson. 1999. Cell cycle-regulated histone acetylation required for expression of the yeast HO gene. Genes Dev. 13:1412-1421.
45. Kristjuhan, A., and J. Q. Svejstrup. 2004. Evidence for distinct mechanisms facilitating transcript elongation through chromatin in vivo. EMBO J. 23:4243-4252.[CrossRef][Medline]
46. Kristjuhan, A., J. Walker, N. Suka, M. Grunstein, D. Roberts, B. R. Cairns, and J. Q. Svejstrup. 2002. Transcriptional inhibition of genes with severe histone H3 hypoacetylation in the coding region. Mol. Cell 10:925-933.[CrossRef][Medline]
47. Kuo, M.-H., J. E. Brownell, R. E. Sobel, T. A. Ranalli, R. G. Cook, D. G. Edmondson, S. Y. Roth, and C. D. Allis. 1996. Transcription-linked acetylation by Gcn5p of histones H3 and H4 at specific lysines. Nature 383:269-272.[CrossRef][Medline]
48. Kuo, M.-H., J. Zhou, P. Jambeck, M. E. A. 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.
49. Lee, C. K., Y. Shibata, B. Rao, B. D. Strahl, and J. D. Lieb. 2004. Evidence for nucleosome depletion at active regulatory regions genome-wide. Nat. Genet. 36:900-905.[CrossRef][Medline]
50. Lee, D., and J. T. Lis. 1998. Transcriptional activation independent of TFIIH kinase and the RNA polymerase II mediator in vivo. Nature 393:389-392.[CrossRef][Medline]
51. Lee, M. S., and W. T. Garrard. 1991. Transcription-induced nucleosome "splitting": an underlying structure of DNase I sensitive chromatin. EMBO J. 10:607-615.[Medline]
52. Lee, M. S., and W. T. Garrard. 1992. Uncoupling gene activity from chromatin structure: promoter mutations can inactivate transcription of the yeast HSP82 gene without eliminating nucleosome-free regions. Proc. Natl. Acad. Sci. USA 89:9166-9170.
53. Lee, S., and D. S. Gross. 1993. Conditional silencing: The HMRE mating-type silencer exerts a rapidly reversible position effect on the yeast HSP82 heat shock gene. Mol. Cell. Biol. 13:727-738.
54. Litt, M. D., M. Simpson, F. Recillas-Targa, M. N. Prioleau, and G. Felsenfeld. 2001. Transitions in histone acetylation reveal boundaries of three separately regulated neighboring loci. EMBO J. 20:2224-2235.[CrossRef]
