The Histone Deacetylase Inhibitor Trichostatin A Blocks Progesterone Receptor-mediated Transactivation of the Mouse Mammary Tumor Virus Promoter in Vivo *

Post-translational modifications of histones play an important role in modulating gene transcription within chromatin. We used the mouse mammary tumor virus (MMTV) promoter, which adopts an ordered nucleosomal structure, to investigate the impact of a specific inhibitor of histone deacetylase, trichostatin A (TSA), on progesterone receptor-activated transcription. TSA induced global histone hyperacetylation, and this effect occurred independently of the presence of hormone. Interestingly, chromatin immunoprecipitation analysis revealed no significant change in the level of acetylated histones associated with the MMTV promoter following high TSA treatment. In human breast cancer cells, in which the MMTV promoter adopts a constitutively “open” chromatin structure, treatment with TSA converted the MMTV promoter into a closed structure. Addition of hormone did not overcome this TSA-induced closure of the promoter chromatin. Furthermore, TSA treatment resulted in the eviction of the transcription factor nuclear factor-1 from the promoter and reduced progesterone receptor-induced transcription. Kinetic experiments revealed that a loss of chromatin-remodeling proteins was coincident with the decrease in MMTV transcriptional activity and the imposition of repressed chromatin architecture at the promoter. These results demonstrate that deacetylase inhibitor treatment at levels that induce global histone acetylation may leave specific regulatory regions relatively unaffected and that this treatment may lead to transcriptional inhibition by mechanisms that modify chromatin-remodeling proteins rather than by influencing histone acetylation of the local promoter chromatin structure.

In chromatin, DNA is arranged into arrays of nucleosomes that consist of 146 bp of DNA wrapped around two copies each of histone proteins H2A, H2B, H3, and H4 assembled as an octamer (1). Addition of a linker histone (H1) assembles the DNA into less well described "higher order" chromatin, leading to a fully condensed chromosome (2). Studies of chromatin structure have shown that the packaging of DNA within chromatin plays an important role in the regulation of gene expres-sion (3,4). Chromatin structure may affect transcriptional activation by blocking the access of trans-acting factors to their target sequences and/or the assembly of the basal transcriptional machinery to form the preinitiation complex (5,6).
However, chromatin is not static. Its dynamic nature is evidenced through post-translational modification of histone Nterminal tail domains, which are reversibly acetylated at ⑀-lysine residues due to an equilibrium between acetylation and deacetylation (4,7,8). The observation that transcriptional cofactors possess enzymatic activity, specifically acetyltransferase activity, provided a direct mechanism by which histones within the context of a specific promoter might be modified (9 -13). Although histone hyperacetylation is often associated with increased gene expression, for a number of genes, this is not the case (14). This raises the intriguing possibility that, like protein phosphorylation, histone acetylation may be a multifaceted post-translational modification with respect to gene expression (7). For example, studies have shown the inhibitory effects of histone hyperacetylation induced by deacetylase inhibitors on steroid-inducible genes such as ovalbumin (15), tyrosine aminotransferase (16), prolactin receptor (17), interleukin-2 (18), and mouse mammary tumor virus (MMTV) 1 (19,20) and, more recently, on vitamin D regulation of the osteocalcin gene (21).
The progesterone (PR), glucocorticoid, androgen, and mineralocorticoid receptors are members of the steroid hormone receptor family, a class of receptors that belong to a large nuclear hormone receptor superfamily of hormone-activated transcriptional regulators (22). Steroid hormone receptors regulate gene expression by binding specific DNA sequences in target genes termed hormone response elements (23). To study the effects of histone acetylation on steroid-induced transcriptional activation, we used the MMTV promoter as a model system because it assumes a defined chromatin structure in vivo. The stably integrated MMTV promoter reproducibly assembles into a phased array of six nucleosomes (A-F) (24). The region of the promoter occupied by the second nucleosome in the array, nucleosome B (Nuc-B), contains the hormone response elements to which steroid hormone receptors bind as well as target sites for other necessary transcription factors, including nuclear factor-1 (NF1) and octamer transcription factors (25). When stably transformed into T47D cells that express the PR but lack the glucocorticoid receptor (PR ϩ /gr Ϫ ; 2963.1 cells), the * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Utilizing the 2963.1 cells, we assessed the effect of low and high trichostatin A (TSA) concentrations on chromatin structure mediated by the PR within the context of the MMTV promoter. We demonstrate that PR-mediated transcription and chromatin remodeling were inhibited at the MMTV promoter when cells were subjected to high levels of TSA. Interestingly, under conditions where global histone hyperacetylation was observed, the levels of histone hyperacetylation at the MMTV promoter were not significantly affected by TSA treatment. However, expression of a variety of transcriptional co-regulators, including chromatin-remodeling proteins, was reduced. Thus, in this human breast cancer cell line, histone deacetylase inhibitors may repress the chromatin architecture of the MMTV promoter by mechanisms other than histone hyperacetylation of the proximal promoter.

MATERIALS AND METHODS
Cell Culture-2963.1 cells were derived from T47D cells by stable cotransfection of the chimeric bovine papilloma virus-based vector pJ83d, carrying the MMTV long terminal repeat (LTR) attached to the bacterial chloramphenicol acetyltransferase (CAT) gene (26). Cells were grown at 37°C with 5% CO 2 in Dulbecco's modified Eagle's medium (Invitrogen) containing 10% fetal bovine serum (BioWhittaker, Inc.). Cells were treated with R-TSA (Sigma) at either low (5 ng/ml) or high (100 ng/ml) concentrations for the times indicated in the figure legends.
CAT Assays-2963.1 cells were seeded in 90-mm dishes at 6 ϫ 10 5 cells/dish in triplicate and treated as indicated in the figure legends. Cells were harvested and lysed in 100 l of 0.25 M Tris-HCl (pH 7.8) by freezing and thawing three times. CAT activity was determined by a kinetic enzymatic assay with 5 g of cell lysate (27). Experiments were repeated at least three times.
Isolation of Histones and Acid/Urea Gel Analysis-Subconfluent cells were treated with TSA and hormone as indicated in the figure legends. Nuclei were isolated as described previously (28). Acid-soluble proteins were isolated from nuclei in 100 l of 0.4 N H 2 SO 4 at 4°C for 1 h. After centrifugation for 5 min at 14,000 rpm, histones were precipitated from the supernatant in 1 ml of acetone placed at Ϫ20°C overnight. Following centrifugation, the proteins were resuspended in 50 l of 0.9 N acetic acid and 25 l of 75% sucrose. Histone proteins (40 g) were electrophoresed using a 16% acid/urea-polyacrylamide gel as previously described (25). Histones were visualized after staining with Amido Black.
In Vivo Analysis of Restriction Enzyme Hypersensitivity-Subconfluent cells were treated with TSA and hormone as indicated in the figure legends. Nuclei were isolated and partially digested with SstI or AflII (150 and 2000 units/ml, respectively; New England Biolabs Inc.) for 15 min at 30°C as described previously (28). All samples were then digested to completion in vitro with HaeIII (New England Biolabs Inc.) to provide an internal control. Digestion fragments were detected by reiterative primer extension using 32 P-labeled primer MMTV-22 (5Ј-TCT-GGAAAGTGAAGGATAAGTGACGA-3Ј), which is specific to the MMTV promoter. The purified extended products were separated on 6% polyacrylamide gels and visualized by exposure to Hyperfilm (Amersham Biosciences) at Ϫ80°C and also quantitated using a PhosphorImager (ImageQuant software, Molecular Dynamics, Inc.).
In Vivo Analysis of Transcription Factor Binding-Subconfluent cells were treated as indicated in the figure legends, and isolated nuclei were partially digested with HaeIII (1000 units/ml) and exonuclease III (3000 units/ml; New England Biolabs Inc.) for 15 min at 30°C to detect stops corresponding to the 5Ј-boundaries of bound factors on the MMTV LTR (29). DNA was purified, digested with mung bean nuclease (Invitrogen) to remove single-stranded overhangs, purified again, and digested to completion with HaeIII. For each sample, 20 g of DNA was subjected to reiterative primer extension with the 32 P-labeled MMTV-CAT primer (5Ј-TTAGCTTCCTTAGCTCCTGAAAAT-3Ј). The purified extended products were separated on 6% polyacrylamide gels and exposed to Hyperfilm. Quantitation was performed using ImageQuant software.
In Vivo Analysis of a Transient MMTV Template-Subconfluent cells were transfected with a plasmid containing the MMTV LTR attached to the firefly luciferase gene (designated pLTR-Luc) (30). Using FuGENE 6 (Roche Molecular Biochemicals) according to the manufacturer's recommendations, 2963.1 cells in 150-mm dishes were transfected with 10 g of pLTR-Luc and 30 l of FuGENE 6 diluted in phenol red-free Opti-MEM (Invitrogen). After incubation for 5 h, the transfection mixture was removed from cells and replaced with growth medium plus additional treatments as indicated in the figure legends. Cells were harvested, and nuclei were isolated as described previously (28). Restriction enzyme hypersensitivity and transcription factor binding were analyzed as described above. To amplify digestion products specific to the transient MMTV template, the 32 P-labeled Luc-618 primer (5Ј-CCTTTCTTTATGTTTTTGGCG-3Ј) was used. For each sample, 10 g of DNA for restriction enzyme hypersensitivity assays and 3 g of DNA for transcription factor binding assays were subjected to reiterative primer extension.
Chromatin Immunoprecipitation (ChIP) Analysis-ChIP analysis was carried out using the ChIP assay kit from Upstate Biotechnologies, Inc. (catalog no. 17-295) with minor modifications of the protocol. 2963.1 cells (10 6 ) were plated onto 100-mm dishes and treated the next day as described in the figure legends. Ten micrograms of either anti-acetylated histone H4 (catalog no. 06-866, Upstate Biotechnology, Inc.) or anti-IB kinase-␣ (catalog number sc-7606, Santa Cruz Biotechnologies) antibody was added to tubes containing 1 ml of chromatin solution. Following incubation with antibody, 60 l of salmon sperm DNA/protein A-agarose was added to each tube, and the agarose-antibody complexes were then captured by centrifugation. After the beads were pelleted and washed, the chromatin was extracted, and the protein-DNA cross-links were reversed. Following purification, the DNA was subjected to PCR amplification using primers MMTV-22 and MMTV-344 (5Ј-TTAAGTAAGTTTTTGGTTACAAACT-3Ј) under the following conditions: 30 cycles, 55°C, 1.5 mM MgCl 2 , 0.2 mM dNTPs, 5 units of Taq polymerase, and 25 pmol of each primer. Eight microliters of each reaction was analyzed on 1.5% 1ϫ Tris borate/EDTA-agarose gels, and the bands were quantified using the Alpha-Imager documentation system (Alpha Innotech Corp.). The DNA was analyzed twice by PCR, and graphs were calculated based on these data.

High Levels of Trichostatin A Inhibit PR-dependent Activation of the MMTV Promoter-
To examine the relationship between histone acetylation and PR-mediated transcriptional activation at the MMTV promoter, we made use of a unique human breast cancer cell line model, T47D/2963.1 (referred to as 2963.1 cells) (26). These cells normally exhibit an open chromatin structure with the constitutive binding of transcription factors at Nuc-B within the MMTV promoter. We also made use of TSA, a potent and specific inhibitor of histone deacetylase, at both low (5 ng/ml) and high (100 ng/ml) concentrations to manipulate the acetylation status of the histones and to induce global histone hyperacetylation (32). The MMTV promoter exhibited a significant level of basal transcriptional activity as measured by CAT assays (Fig. 1A, q). However, transcriptional activation from the promoter was hormoneinducible after treatment with the synthetic progestin R5020 (Fig. 1A, E). Pretreatment of cells with low TSA had no effect on hormone-stimulated MMTV transcriptional activity (Fig.  1A, compare E and ‚). Furthermore, treatment with low TSA alone failed to either increase or inhibit MMTV basal transcriptional activity (Fig. 1A, compare q and OE). In contrast, pre-treatment with high TSA prior to adding R5020 inhibited hormone-induced transcription as seen in the absence of TSA (Fig.  1A, compare E and Ⅺ). Moreover, treatment with high TSA alone reduced the level of MMTV transcriptional activity to below basal activity (Fig. 1A, compare q and f). Our results demonstrate that high (but not low) TSA treatment inhibits the transcriptional activity of the MMTV promoter in 2963.1 cells.
High Levels of Trichostatin A Induce Global Histone Acetylation-To examine potential causes for the different effects of low and high TSA treatments, we initially investigated the global histone acetylation status. Core histones were isolated and analyzed by electrophoresis on acid/urea-polyacrylamide gels, which resolved multiple acetylated histone isoforms (Fig.  1B). The TSA-induced changes in acetylation were most readily observed by examining changes in histone H4 isoforms. In the absence of TSA, only the un-and monoacetylated forms of histone H4 were detected. Treatment with the hormone R5020 did not alter the histone acetylation profile (Fig. 1B, lanes 1 and 2). Treatment with low TSA decreased the level of the unacetylated form and increased the level of the mono-and diacetylated forms of histone H4 (Fig. 1B, lanes 1 and 5). In contrast, treatment with high TSA resulted in increased levels of di-, tri-, and tetraacetylated forms of histone H4 (Fig. 1B,  lanes 1 and 6). Increases in the acetylated forms of histones H2A, H2B, and H3 were also observed with high TSA treat-ments ( Fig. 1B, lanes 1 and 6). Addition of R5020 did not alter the histone acetylation patterns generated by treatment with low and high TSA alone (Fig. 1B, lanes 3-6). Thus, exposure to high concentrations of TSA results in increased histone acetylation patterns, correlating with the inactivation of the MMTV promoter within 2963.1 cells.
High TSA Leads to Hyposensitive Chromatin at Nuc-B on the MMTV LTR-Previous analysis of 2963.1 cells demonstrated that the promoter region encompassed by Nuc-B is constitutively open and hypersensitive to restriction enzyme endonucleases (26). To evaluate the effect of histone hyperacetylation on the chromatin architecture, we first examined the extent of restriction enzyme hypersensitivity using the enzyme SstI. The constitutive hypersensitivity to SstI characteristic of 2963.1 cells was reduced by treatment with high (but not low) concentrations of TSA ( Fig. 2A, lanes 1, 5, and 6), indicating that Nuc-B is converted to a closed chromatin structure. Moreover, this closed chromatin structure was maintained in cells pretreated with high (but not low) concentrations of TSA prior to agonist treatment ( Fig. 2A, lanes 3-6).
To confirm that the observed effects of inhibiting histone deacetylation were not restricted to the 3Ј-region of Nuc-B, we utilized the restriction enzyme AflII, which cleaves near the 5Ј-boundary of Nuc-B. As seen with SstI, the MMTV promoter remained hypersensitive to AflII digestion both in the absence  (Fig. 2B, lanes 7 and 8). Treatment of cells with high concentrations of TSA resulted in reduced cleavage by AflII, consistent with the now "closed" chromatin architecture of the promoter in the absence or presence of hormone (Fig. 2B,  lanes 8, 10, and 12). This closed chromatin structure was not observed when 2963.1 cells were treated with low concentrations of TSA. The MMTV promoter remained accessible to digestion with AflII at levels similar to untreated cells in the absence or presence of hormone (Fig. 2B, lanes 7, 9, and 11). These data demonstrate that, under conditions that lead to histone hyperacetylation, the constitutive hypersensitivity of Nuc-B in the MMTV promoter is lost.
High TSA Evicts Transcription Factor NF1 from the MMTV Promoter-Because an open chromatin structure is essential for transcription factors such as NF1 to bind in vivo, the above data predict that treatment with high concentrations of TSA would block transcription factor binding. This direct consequence of histone hyperacetylation on protein-DNA interaction was examined using in vivo footprinting assays (33). In 2963.1 cells, transcription factor NF1 bound to the promoter independently of hormone (Fig. 3, lanes 1 and 2), consistent with a constitutively open chromatin architecture. Treatment with high TSA inhibited NF1 binding to the MMTV promoter (Fig. 3,  compare lanes 1 and 6); subsequent hormone treatment after high TSA pretreatment did not restore NF1 binding to the MMTV promoter in 2963.1 cells (Fig. 3, compare lanes 2, 4, and  6). As predicted, treatment with low TSA had no effect on NF1 binding either in the absence or presence of hormone (Fig. 3,  lanes 3 and 5). Thus, treatment with histone deacetylase inhibitor results in a closed MMTV chromatin structure, blocking the binding of NF1, which is necessary for transcriptional activation.
High TSA Does Not Alter the Structure of Nuc-B on a Transient MMTV Template-High levels of TSA result in the loss of restriction enzyme hypersensitivity within the region associated with nucleosome B in the MMTV promoter. Because high TSA treatment modifies the acetylation status of histones at many locations within the cell, we examined the possibility that decreased digestion of the MMTV promoter was a result of increased SstI and AflII restriction sites now available elsewhere within the cell. To determine this, we analyzed the ability of SstI to digest a transiently introduced MMTV template under the same experimental conditions in which decreased digestion was observed on the integrated template. By introducing a transient MMTV template, we increased the number of target sites for the restriction enzyme SstI. It is also important to note that the MMTV promoter adopts an ordered nucleosomal structure only upon stable integration into chromatin, but not when transiently expressed. Under these circumstances, the MMTV template adopts a relatively loose and open conformation. In 2963.1 cells, the transient MMTV template pLTR-Luc (30) remained sensitive to SstI in the absence and presence of agonist treatment, resulting in 42 and 51% digestion, respectively (Fig. 4A, lanes 1 and 2). Treatment with high TSA did not inhibit SstI digestion of the transient MMTV template (Fig. 4A, lanes 1 and 4), the same conditions under which digestion of the stable MMTV promoter decreased ( Fig.  2A, lanes 1 and 6). In fact, digestion with SstI was slightly enhanced, resulting in 59% cutting. Furthermore, agonist treatment following high TSA pretreatment had no effect on SstI digestion (54% cutting) (Fig. 4A, lanes 3 and 4). Therefore, decreased SstI digestion of the integrated MMTV promoter within Nuc-B is due to altered chromatin structure (i.e. chromatin closure), as opposed to an increase in restriction sites or other factors that may affect the restriction enzyme SstI.
High TSA Does Not Affect NF1 Binding to a Transient MMTV Template-High levels of TSA also result in the loss of NF1 binding to the MMTV promoter within Nuc-B. To determine the effect of high TSA on the ability of NF1 to interact with its DNA-binding site, we analyzed a transient MMTV template by exonuclease III footprinting under the same conditions that NF1 binding was inhibited. As expected from the enzyme hypersensitivity results, NF1 bound to the transient MMTV promoter within Nuc-B in the absence or presence of R5020 (Fig. 4B, lanes 1 and 2). Treatment with high TSA did not alter NF1 binding (Fig. 4B, lanes 1 and 4). NF1 bound the transient MMTV template under the same conditions in which NF1 bound to the stable chromatin MMTV promoter (Fig. 3,  lanes 1 and 6). Addition of R5020 subsequent to high TSA pretreatment did not alter NF1 binding (Fig. 4B, lanes 1, 3, and  4). Consequently, in 2963.1 cells, binding of NF1 to the stable, nucleosome-associated MMTV promoter is inhibited due to the closed chromatin architecture induced by high TSA treatment. Our results demonstrate that NF1 is competent to bind DNA under these conditions.
High Levels of TSA Reduce the Levels of Various Regulatory Factors-In addition to promoting histone acetylation, TSA is known to result in hyperacetylation of other cellular proteins (34). To extend our analysis of the inhibitory effects of high TSA on MMTV activation, we examined the steady-state levels of cellular factors that may participate in MMTV activation by the PR. Our results demonstrated that the levels of both progesterone receptor isoforms (PR B and PR A ) were greatly reduced in response to high TSA treatments in the absence and presence of hormone relative to untreated cells (Fig. 5A, com-  1, 4, and 6). PR B levels were decreased by 95% by high TSA treatment both in the absence and presence of hormone. PR A levels were also reduced by 95% by high TSA treatment either alone or in the presence of hormone. However, treatment with R5020 alone or in combination with low TSA also resulted in decreased PR levels, although not to the extent as observed with high TSA (Fig. 5A). Treatment with R5020 reduced the levels of PR B by 85% and the levels of PR A by 65%. Treatment with low TSA resulted in decreased levels of PR, but to a lesser extent. PR B was reduced by 50% by low TSA alone and by 85% upon addition of R5020; PR A was reduced by 80% by low TSA treatment alone and by 40% upon addition of R50202. Although hormone addition was unable to overcome the effects of high TSA treatment, it appeared to antagonize gested in vivo with HaeIII and exonuclease III and then re-digested in vitro with HaeIII to serve as an internal loading control. The HaeIII product and the 5Ј-boundary corresponding to NF1 are indicated. the effects of the low TSA-induced decrease in PR A levels while enhancing the low TSA-induced decrease in PR B levels. In contrast to these results, expression of NF1 was largely unaffected by treatment with either low or high concentrations of TSA (Fig. 5A). Therefore, the loss of NF1 binding at Nuc-B is not a result of reduced NF1 expression in response to high TSA exposure.
We next examined expression of proteins that have been shown to be components of the histone deacetylase complex, the target of TSA inhibition (35)(36)(37)(38). Expression of HDAC1, mSin3a, and RbAp48 and RbAp46 (components of the mSin3a repressor complex) was unaffected by treatment with low or high TSA independent of hormone induction (Fig. 5B). The nuclear receptor corepressor NCoR was reduced upon treatment with high TSA (80% decrease) (Fig. 5B, compare lanes 1  and 6), which was not reversed by subsequent treatment with R5020 (70% decrease) (Fig. 5B, compare lanes 1, 4, and 6).
Given the inhibitory effects of high TSA on activation, we then examined the fate of factors involved in chromatin remodeling and transactivation of the promoter (31). Consistent with the activity data presented earlier, treatment with high (but not low) TSA reduced the levels of the BRG-1 protein by 90% (Fig. 5C, compare lanes 1, 5, and 6). Treatment with R5020 did not abrogate the effects of high TSA on BRG-1 levels (Fig. 5C,  compare lanes 1, 4, and 6). Similarly, the levels of NCoA1 were reduced in response to high (but not low) levels of TSA in the absence of hormone (Fig. 5C, compare lanes 1, 5, and 6), displaying an 80% reduction in protein levels. However, unlike BRG-1 levels, this reduction was reversed by treatment with R5020 (Fig. 5C, compare lanes 1, 4, and 6). Finally, expression of the acetylase p300 was also decreased by high TSA treatment both in the absence and presence of hormone, although the decrease was less when R5020 was present (Fig. 5C).
The previous protein profiling results demonstrate, not surprisingly, that high TSA treatment had diverse effects on protein levels, down-regulating a subset of cofactors, whereas others remained unaffected when the MMTV promoter was inactivated. In particular, the down-regulation of the PR, coactivators, and BRG-1 proteins at the same time that histones are fully acetylated would potentially represent a powerful model for repressing this promoter. However, the loss of corepressors as seen with NCoR might compensate for the loss of the cofactors. In the next series of experiments, we examined the temporal relationship between protein down-regulation, histone acetylation, and closing of the MMTV promoter.
Kinetics of Chromatin Closure, Transcription Factor Eviction, and Histone Hyperacetylation-Given that TSA inhibition of HDAC1 is known to be rapid (32), we next examined the kinetics of TSA exposure upon PR transactivation. As an initial assay, we documented the time course of histone acetylation within cells treated with high TSA (Fig. 6A). As observed previously, histones from untreated 2963.1 cells were primarily unacetylated, with the presence of un-and monoacetylated forms of histone H4 (Fig. 6A, lane 1). Treatment with high TSA for 1 h induced the global acetylation of histones and resulted in an increase in the mono-, di-, and triacetylated forms of histone H4 (Fig. 6A, lane 2). By 4 h of high TSA, there was a loss of the un-and monoacetylated forms of histone H4; moreover, the acetylated forms of histones H2A, H2B, and H3 were the prevalent species as well as tetraacetylated histone H4 (Fig. 6A, lane 3). Longer treatments with high TSA did not alter the hyperacetylation pattern established within 4 h (Fig.  6A, lanes 4 -6). This pattern of histone acetylation was confirmed using antibodies directed against acetylated histones H3 and H4 (data not shown). Therefore, our results demonstrate that global histone acetylation occurred rapidly in re-sponse to treatment with high TSA.
In the next series of experiments, we examined the chromatin architecture by restriction enzyme hypersensitivity assays in cells under a similar time course of TSA exposure used for histone acetylation as described above. Results are representative of repeated experiments. Untreated 2963.1 cells demonstrated the constitutive hypersensitivity characteristic of the open chromatin architecture (Fig. 6B, lane 1), as shown previously (Fig. 2). Treatment with high TSA for increasing lengths of time significantly reduced the extent of cleavage by SstI, consistent with the closure of chromatin structure associated with Nuc-B. Treatment with high TSA for as little as 1 h reduced cleavage by the restriction endonuclease SstI within Nuc-B by 15% (Fig. 6B, compare lanes 1 and 2). Chromatin cleavage was reduced by 50% after treatment with high TSA for 4 h (Fig. 6B, compare lanes 1 and 3). Moreover, maximal closure of MMTV chromatin occurred after 12 h of treatment with high TSA (Fig. 6B, compare lanes 1 and 5), reducing chromatin cleavage by 80%. Therefore, chromatin closure of the MMTV promoter occurs rapidly subsequent to high TSA treatment and correlates with global histone hyperacetylation.
The binding of NF1 is intimately linked with MMTV expression; and as shown earlier (Fig. 3), NF1 is lost from the promoter with high TSA. This leads to the prediction that the kinetics of NF1 loss should mirror the closure of the promoter seen earlier. Indeed, treatment of 2963.1 cells with high TSA for increasing periods of time resulted in the sequential loss of NF1 binding that was coincident with the closing of the promoter as well histone hyperacetylation (Fig. 6C, lanes 1-6).
Kinetics of TSA-induced Loss of Chromatin-remodeling Proteins-Because we noted that inhibition of PR activity occurred progressively in response to high TSA treatment, we investigated the expression profiles of various cellular factors under the same time course of TSA exposure. Examination of both the PR B and PR A isoforms revealed that receptor levels did not fall significantly until after 12 h of TSA exposure (Fig. 6D). In fact, treatment with high levels of TSA for 1 h appeared to increase the levels of both PR isoforms (Fig. 6D, lane 2). Furthermore, the steady-state levels of NF1 were unaffected by treatment with high TSA for increasing lengths of time. Therefore, eviction of NF1 and loss of chromatin hypersensitivity at the MMTV promoter cannot be attributed to the loss of PR protein expression in response to TSA exposure.
In contrast to the PR, the levels of the chromatin-modifying factor BRG-1 displayed a noticeable 70% decrease in expression after 4 h of treatment with TSA (Fig. 6D, compare lanes 1 (0 h)  and 3 (4 h)). Protein levels underwent a further decrease after exposure to TSA for 24 h, demonstrating an ϳ85-90% reduction (Fig. 6D). BRG-1 is known to function as part of a larger macromolecular complex of proteins to remodel chromatin. To ascertain if TSA was down-regulating other members of the complex, we examined the levels of BAF155, a member of the BRG-1 chromatin-remodeling core complex. Indeed, BAF155 displayed a similar expression profile as seen with BRG-1 and was down-regulated by 75% at 4 h (Fig. 6D). Similarly, BAF155 protein levels were decreased by 90% at 24 h post high TSA treatment. To extend our analysis, we examined the levels of a group of corepressor proteins over the same time course. Expression of the nuclear receptor corepressor NCoR increased up to 2-fold in response to the shorter term treatments of TSA (1-12 h) and then decreased by ϳ50% below basal levels after 24 h of treatment (Fig. 6D). Finally, the levels of both mSin3a and HDAC1 were unaffected by high TSA treatment at all time points, as would be predicted from the previous assay at 24 h (Fig. 6D). This series of experiments suggests that, despite the distinct effects of TSA on nuclear protein profiles observed, the Deacetylase Inhibition Does Not Significantly Change the Levels of Acetylated Histone H4 at the MMTV Promoter-To determine whether the increase in global histone acetylation in response to TSA treatment also occurred locally at the MMTV proximal promoter, we used ChIP assays with an antibody specific to the acetylated form of histone H4 (Fig. 7). R5020 treatment reduced the levels of acetylated histone H4 associated with the promoter (Fig. 7A, lanes 9 and 10), as has been observed previously by others (43). Surprisingly, treatment with high TSA did not significantly alter the levels of acetylated histone H4 (Fig. 7A, lanes 9 and 11) after 24 or 2 h of treatment (Fig. 7B, lanes 7-9). Immunoprecipitation with a nonspecific antibody to the IB kinase-␣ protein did not result in a significant background signal (Fig. 7, A and B, lanes 5-8  and 4 -6, respectively). Thus, although TSA treatment causes a global increase in histone H4 acetylation, specific locations of the genome such as the MMTV promoter may be relatively unaffected. DISCUSSION The assembly of gene regulatory sequences into defined chromatin structures provides an attractive and powerful means to control both constitutive and inducible gene expressions. Posttranslational modifications of histones represent a potentially important mechanism by which histone-DNA interactions can FIG. 7.
Inhibition of histone deacetylase activity does not significantly alter the levels of acetylated histone H4 at the MMTV promoter. A, cells were left untreated (lanes 1, 5, and 9) or were treated with R5020 (10 Ϫ8 M) for 1 h (lanes 2, 6, and 10), treated with high TSA for 25 h (lanes 3, 7, and 11), or pretreated with TSA for 24 h followed by R5020 for 1 h (lanes 4, 8, and 12). ChIP assays were performed using antibodies against IB kinase-␣ (IKK␣) or acetylated histone H4. B, cells were left untreated (lanes 1, 4, and 7) or were treated with high TSA for 2 h (lanes 2, 5, and 8) or 24 h (lanes 3, 6, and 9). ChIP assays were performed using antibodies against IB kinase-␣ or acetylated histone H4. be modulated to allow specific changes in gene activity. The acetylation and deacetylation of the core histones have been extensively investigated, and a general pattern has emerged, such that acetylation is associated with active chromatin, whereas deacetylation is linked to inactive chromatin (39,40). However, recent evidence suggests a more complex scenario with respect to individual genes and/or promoters (41).
We have used the MMTV promoter to assess the consequences of inhibiting histone deacetylase activity on transcriptional activation mediated by the progesterone receptor. Within human breast cancer cells, we observed the transcriptional repression of the MMTV promoter in response to TSAinduced histone hyperacetylation. These observations are distinct from a previous investigation that reported that moderate increases in histone acetylation induced by treatment with low concentrations of TSA activated MMTV promoter activity and chromatin remodeling (20). As indicated earlier, we did not see any clear changes in histone acetylation or promoter activity in response to low TSA. However, at higher concentrations of TSA (50 ng/ml), similar to those we employed (100 ng/ml), a similar decrease in hormone-induced transcription was observed (20), as reported here. In our studies, the cells exhibited a threshold effect for TSA such that exposure to low levels of TSA (5 ng/ml) failed to effectively hyperacetylate core histones and did not significantly affect the PR activation of the promoter (see Figs. 1-3 and 5). In contrast, exposure to high levels of TSA (100 ng/ml) resulted in a rapid and profound global hyperacetylation of histones and a significant loss of PR-induced activation. Mechanistically, this treatment alters promoter chromatin architecture such that the constitutively open or hypersensitive promoter reverts to a closed or hyposensitive structure. Moreover, hormone treatments were unable to reverse these structural changes. The consequence of this reversion in chromatin structure was the eviction of NF1 from the promoter (Fig. 3). Treatment with high TSA, which promotes these inhibitory events, had no effect on a naked DNA MMTV template (Fig. 4). Indeed, the transient MMTV promoter was susceptible to digestion with SstI and was bound by NF1 (Fig. 4), demonstrating that NF1 retains the ability to interact with DNA in the presence of TSA.
Given the somewhat unexpected relationship between the global levels of histone acetylation and the inhibition of MMTV transcription, we examined the local acetylation of histones at the MMTV promoter. Unexpectedly, ChIP analysis demonstrated that treatment with high TSA did not significantly alter the acetylation status of histone H4 at Nuc-B within the MMTV promoter. Therefore, although the general pools of histones within cells are hyperacetylated upon treatment with high TSA, histone acetylation within the MMTV promoter is unaffected, if not decreased. This raises the intriguing possibility that TSA may inhibit the activity of chromatin-remodeling and co-regulatory molecules that have been previously shown to participate in the activation of the promoter (31). Indeed, prolonged exposure to high levels of TSA resulted in the down-regulation of both PR isoforms, the chromatin-remodeling proteins BRG-1 and BAF155, coactivators, and histone acetylases NCoA1 and p300, as well as the nuclear corepressor NCoR. Interestingly, neither HDAC1 nor mSin3a was affected, nor were NF1 levels changed.
Treatment with high TSA rapidly induced global histone hyperacetylation, suggesting that it is an effective deacetylase inhibitor in these cells. Similarly, the closure of Nuc-B occurred rapidly, displaying progressive insensitivity to digestion indicative of a closed chromatin architecture at the promoter. Furthermore, loss of NF1 binding to the MMTV promoter was tightly associated with the closure of Nuc-B. In contrast, the levels of the PR or the corepressor NCoR were not decreased at early time points of high TSA treatment and did not correlate with the loss of hypersensitivity or transcription factor eviction. Rather, loss of PR-enhanced gene expression, global histone acetylation, and promoter closure appears to directly correlate with the loss of expression of the chromatin-remodeling proteins BRG-1 and BAF155.
This concept is not unique to the MMTV promoter because TSA treatment also reduces the levels of acetylated histone H4 associated with the active maternal H19 allele, and this correlates with a decrease in RNA levels (42). Moreover, in our studies, treatment with R5020, independent of TSA treatment, resulted in decreased histone H4 acetylation. This result is consistent with recently published data demonstrating decreased histone acetylation upon hormone stimulation of glucocorticoid-induced MMTV transcriptional activity (43). Consequently, our results obtained through ChIP analysis of the MMTV promoter do not offer a direct correlation between histone acetylation, chromatin structure, and transcriptional activity, suggesting a more complex mechanism of regulation. This is also consistent with previous studies in 3T3 fibroblasts stably transfected with an MMTV reporter demonstrating that global histone acetylation status and MMTV transcription may not be directly correlated (42). The lack of correlation between local histone H4 acetylation status and gene activation has also been investigated on a more global level (44). In that study, comparing the level of histone H4 acetylation within transcriptionally active chromatin, the authors found that H4 acetylation in coding and adjacent regions was not correlated with transcriptional activity.
We demonstrate that, for MMTV, exposure of cells to levels of TSA that result in global histone hyperacetylation inhibits PR-mediated transcription. These observations are consistent with previous studies in which inhibition of deacetylase activity was associated with gene inactivation (15)(16)(17)(18)21). In the case of the nuclear receptors, it has been proposed that histone acetylation may be viewed as a molecular switch between the inactive and active forms of the receptor, suggesting that action of both acetylases and deacetylases is important in the regulation of many genes (8). Our demonstration that TSA exposure results in the loss of chromatin-remodeling proteins with similar kinetics to the loss of PR activity represents an important advance in our understanding of complex interrelations between histone acetylation, chromatin remodeling, and cofactor regulation of gene expression. These observations suggest a novel mechanism by which the loss of expression of regulatory cofactors involved in chromatin remodeling results in the repression of PR-mediated transcriptional activity. As such, it contributes to the expanding body of evidence that places histone acetylation/deacetylation and chromatin structure as a central and important mechanism for regulating transcriptional activation.