Trichostatin A Induces Transforming Growth Factor β Type II Receptor Promoter Activity and Acetylation of Sp1 by Recruitment of PCAF/p300 to a Sp1·NF-Y Complex*

Transforming growth factor β type II receptor (TβRII) is a tumor suppressor gene that can be transcriptionally silenced by histone deacetylases (HDACs) in cancer cells. In this report, we demonstrated the mechanism by which trichostatin A (TSA), an inhibitor of HDAC, induces the expression of TβRII in human pancreatic cancer cell lines by modulating the transcriptional components that bind a specific DNA region of the TβRII promoter. This region of the TβRII promoter possesses Sp1 and NF-Y binding sites in close proximity (located at –102 and –83, respectively). Treatment of cells with TSA activates the TβRII promoter in a time-dependent manner through the recruitment of p300 and PCAF into a Sp1·NF-Y·HDAC complex that binds this DNA element. The recruitment of p300 and PCAF into the complex is associated with a concomitant acetylation of Sp1 and an overall decrease in the amount of HDAC associated with the complex. Transient overexpression of p300 or PCAF potentiated TSA-induced TβRII promoter activity. The effect of PCAF was dependent on its histone acetyltransferase activity, whereas that of p300 was independent. Stable transfection of PCAF caused an increase in TβRII mRNA expression, the association of PCAF with TβRII promoter, and the acetylation of Sp1. Taken together, these results showed that TSA treatment of pancreatic cancer cells leads to transcriptional activation of the TβRII promoter through modulation of the components of a Sp1·NF-Y·p300·PCAF·HDAC-1 multiprotein complex. Moreover, the interaction of NF-Y with the Sp1-associated complex may further explain why this specific Sp1 site mediates transcriptional responsiveness to TSA.

Transforming growth factor ␤ type II receptor (T␤RII) is a tumor suppressor gene that can be transcriptionally silenced by histone deacetylases (HDACs) in cancer cells. In this report, we demonstrated the mechanism by which trichostatin A (TSA), an inhibitor of HDAC, induces the expression of T␤RII in human pancreatic cancer cell lines by modulating the transcriptional components that bind a specific DNA region of the T␤RII promoter. This region of the T␤RII promoter possesses Sp1 and NF-Y binding sites in close proximity (located at ؊102 and ؊83, respectively). Treatment of cells with TSA activates the T␤RII promoter in a time-dependent manner through the recruitment of p300 and PCAF into a Sp1⅐NF-Y⅐HDAC complex that binds this DNA element. The recruitment of p300 and PCAF into the complex is associated with a concomitant acetylation of Sp1 and an overall decrease in the amount of HDAC associated with the complex. Transient overexpression of p300 or PCAF potentiated TSA-induced T␤RII promoter activity. The effect of PCAF was dependent on its histone acetyltransferase activity, whereas that of p300 was independent. Stable transfection of PCAF caused an increase in T␤RII mRNA expression, the association of PCAF with T␤RII promoter, and the acetylation of Sp1. Taken together, these results showed that TSA treatment of pancreatic cancer cells leads to transcriptional activation of the T␤RII promoter through modulation of the components of a Sp1⅐NF-Y⅐p300⅐PCAF⅐HDAC-1 multiprotein complex. Moreover, the interaction of NF-Y with the Sp1-associated complex may further explain why this specific Sp1 site mediates transcriptional responsiveness to TSA.
TGF-␤ 1 plays a significant role in the growth inhibition of most normal epithelial and some cancer cells (1). TGF-␤ medi-ates its biological effects through cell surface receptors known as TGF-␤ type I receptor (T␤RI) and TGF-␤ type II receptor (T␤RII). Its intracellular signaling is initiated upon the selective binding of the active cytokine to the T␤RII homodimer. T␤RII is a ubiquitously expressed and constitutively active serine/threonine kinase. Ligand binding to T␤RII induces the assembly of a heterotetrameric complex consisting of T␤RI and T␤RII. Once the receptor complex is formed, T␤RII phosphorylates and thereby activates the T␤RI serine/threonine kinase. Activation of T␤RI propagates downstream signaling via Smad family proteins. T␤RI directly interacts with and phosphorylates Smad2 and Smad3. These Smads bind Smad4 and then result in the translocation of this complex to the nucleus and modulate TGF-␤-responsive gene expression (2)(3)(4).
The TGF-␤ signaling pathway is inactivated in many tumors. Loss of negative growth regulation by TGF-␤ affords cells a selective growth advantage associated with decreased dependence of exogenous growth factor and increased tumorigenicity. Frequently, inhibition of TGF-␤ signaling occurs by either abolition of the function of a common mediator, Smad4, or interference with T␤RII function (5,6). Smad4 and T␤RII are tumor suppressor genes (5,6). It was reported that about 50% of pancreatic ductal adenocarcinomas express a low or undetectable level of T␤RII, although they have a normal T␤RII gene and downstream signaling intermediates (7). Additional studies have shown that the down-regulation of T␤RII in pancreatic ductal adenocarcinoma was most often caused by transcriptional repression but not by a mutation of the T␤RII gene (8). The decrease of expression of T␤RII in pancreatic ductal adenocarcinoma cells due to transcriptional repression contributes to a loss in growth control (9) and resistance to apoptosis (10). Restoration of expression of T␤RII in MIA PaCa-2 cells results in increased sensitivity to irradiation-induced apoptosis (10), consistent with T␤RII being a tumor suppressor in these cells.
The T␤RII promoter has been partially characterized. It lacks a distinct TATA box near the transcription initiation site. There are four major regulatory elements in T␤RII promoter: two positive regulatory elements (PRE-1 and PRE-2; located at Ϫ219 to Ϫ172 and ϩ1 to ϩ50), a negative regulatory element (NRE2; located at Ϫ100 to Ϫ67), and a core promoter (Ϫ47 to Ϫ1) (11). Moreover, T␤RII promoter contains many transcription factor binding sites, such as those of Sp1, NF-Y, AP1, CREB, and ERT (12,13). There are two consensus Sp1 binding sites, one in the core promoter (located at Ϫ25), and the other upstream of the NRE2 (located at Ϫ147). Recently, two novel regulatory sites at position Ϫ102 and Ϫ59 have been identified that bind the Sp family of proteins (14). The NF-Y transcription factor binds to an inverted CCAAT box in NRE2 of T␤RII promoter at position Ϫ83, which is in close proximity to the novel Sp1 site at position Ϫ102 (15).
Regulation of specific gene expression is achieved largely by controlling production of transcription factors responsible for the expression of target genes, alteration of the activity of transcription factor through post-translational modification, and localized remodeling of chromatin structures. Recently, histone acetylation and deacetylation have attracted tremendous attention. Acetylation of histone proteins, particularly H3 and H4, is thought to lead to relaxed chromatin structure and, therefore, to an enhanced rate of transcription. There are two classes of enzymes, histone deacetylases (HDACs) and histone acetyltransferase (HATs), involved in the regulation of the acetylation state of histones (16).
Several transcriptional co-activators with intrinsic acetyltransferase activity have been identified, including PCAF, p300/CREB-binding protein, GCN5, TAFII250, ACTR, and SRC-1. Among them, p300/CREB-binding protein and PCAF/ GCN5 are the best-studied co-activators (17). Current models suggest that these activators are brought to target gene promoters through interaction with sequence-specific DNA-binding proteins, where they may function 1) as a bridge to connect sequence-specific transcription factors to the transcription apparatus, 2) as a protein scaffold for the assembly of multicomponent complexes that confer transcriptional activation, and 3) as an acetylase targeting chromatin and/or as transcription factors to facilitate a transcriptional response (18,19).
In previous work, we showed that treatment of MIA PaCa-2 cells with a HDAC inhibitor, TSA, strongly induces T␤RII expression. A specific Sp1 site (Sp1C, located at Ϫ102 bp relative to transcription start site) and an adjacent NF-Y site (Ϫ83 bp) are required for TSA-mediated activation of the T␤RII gene (20,21). The purpose of the present investigation was to determine the mechanisms by which an HDAC inhibitor causes activation of the T␤RII promoter through modulation of components that bind to the Sp1C/NF-Y sites.
In this study, we demonstrate that TSA treatment activates the T␤RII promoter and leads to the acetylation of Sp1 protein in a time-dependent manner. Overexpression of p300 or PCAF potentiates TSA-induced T␤RII promoter activity. The effect of PCAF is dependent on its HAT activity, whereas that of p300 is independent. Moreover, PCAF can acetylate Sp1 in vivo. Taken together, these results show that the modulation of the Sp1⅐NF-Y⅐p300⅐PCAF⅐HDAC-1 multiprotein complex plays a pivotal role in the transcriptional activation of the T␤RII promoter through the Sp1C site by TSA.

EXPERIMENTAL PROCEDURES
Cells and Reagents-Human pancreatic cancer cell lines BxPC-3, PANC-1, CFPAC-1, and MIA PaCa-2 were procured from American Type Culture Collection. The cell line UK Pan-1 was established previously in our laboratory (22). The immortalized pancreatic epithelial cell line was kindly provided by Dr. Michel Ouellett (University of Nebraska). All the cell lines were cultured in a 37°C humidified atmosphere containing 5% CO 2 in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. TSA (Sigma) was dissolved in Me 2 SO and added to the culture medium at a final concentration of 100 ng/ml.
HDAC Activity Assay-The HDAC enzymatic activity was determined using a histone deacetylase assay kit (Upstate Biotechnology) per the manufacturer's instructions. Briefly, 10 g of nuclear extracts were incubated with [ 3 H]acetyl histone H4 peptide in the HDAC assay buffer for 24 h. The released [ 3 H]acetate was extracted with 600 l of ethyl acetate and separated by centrifugation, and then 200-l aliquots of the ethyl acetate phase were measured in duplicates for radioactivity by a scintillation counter.
Transfection and Luciferase Assay-Pancreatic cancer cell lines were seeded into 12-well plates at a density of 1 ϫ 10 5 cells/well the day before transfection. Transfection was performed by using FuGENE 6 (Roche Applied Science). The DNA mixture containing 0.5 g of RII promoter-luciferase construct; 0.5 g of p300-WT, p300 ⌬HAT, PCAF-WT, or PCAF ⌬HAT; 100 ng of Sp1 cDNA; and the relative empty vector were adjusted to 1.1 g of DNA per well. Twenty-four hours after transfection, the cells were treated with 100 ng/ml TSA or vehicle alone at the indicated time points. The T␤RII promoter activities were assayed by luminometer and dual luciferase assay kit (Promega). The luciferase activities were normalized against the protein concentration or cytomegalovirus-Renilla luciferase activity. All experiments were performed at least three times.
Co-immunoprecipitation-Co-immunoprecipitation was performed by incubating 100 g of nuclear extract from MIA PaCa-2 cells with 0.5 g of appropriate immunoprecipitation antibody in radioimmune precipitation assay buffer (20 mM Tris-HCl, pH 8.0, 100 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride, 1 g/ml aprotinin, and 1 g/ml leupeptin) overnight at 4°C with gentle mixing. Protein A-agarose beads (25 l) were used to precipitate the associated protein. Samples were then separated on an 8% SDS-PAGE and examined by Western blotting using indicated antibodies.
DNA Affinity Purification Assays (DAPAs)-DAPA was carried out as described by Walker et al. (24). Briefly, 1 g of biotin-end-labeled double-stranded oligonucleotides was incubated with 100 g of nuclear extracts from untreated or TSA-treated cells at the indicated time points in DAPA buffer (25 mM HEPES, pH 7.6, 60 mM KCl, 5 mM MgCl 2 , 7.5% glycerol, 0.1 mM dithiothreitol, and 0.25% Triton X-100). Oligonucleotide sequences used in DAPA are shown in Table I. The DNA⅐protein complexes were precipitated with 50 l of neutravidincoated agarose beads (Pierce), resolved by 6% SDS-PAGE, and detected by Western blot using antibodies to Sp1, HDAC-1 (Upstate Biotechnology), PCAF, p300, and NF-Y.
Chromatin Immunoprecipitation (ChIP) Assay-ChIP assay was performed as described by Boyd et al. (25). Cells (2 ϫ 10 7 ) were treated with or without 100 ng/ml TSA for 24 h. Formaldehyde was added to the culture medium to give a final concentration of 1%, and after incubation at room temperature for 15 min, cells were washed twice in phosphatebuffered saline, scraped, and lysed in lysis buffer (1% SDS, 10 mM Tris-HCl, pH 8.0, 1 mM phenylmethylsulfonyl fluoride, 1 g/ml pepstatin A, and 1 g/ml aprotinin) on ice for 10 min. Lysates were sonicated on ice to break the chromatin DNA to an average length of about 500 -600 bp and pre-cleaned with protein A-agarose. The chromatin⅐protein complexes were immunoprecipitated with anti-PCAF, -p300, and -Sp1 antibody (Upstate Biotechnology). Cross-linking of protein⅐DNA complexes was reversed at 65°C for 4 h. DNA was extracted three times with phenol/chloroform and precipitated with ethanol. Pellets were resuspended in TE buffer (10 mM Tris, pH 8.0, 1 mM EDTA) and subjected to PCR amplification using specific T␤RII promoter primers: forward primer, 5Ј-GTAAATACTTGGAGCGG-GAAC-3Ј (Ϫ182/Ϫ161); and reverse primer, 5Ј-ACTCACTCAACT-TCAACTCAGA-3Ј (ϩ54/ϩ33).
Stable Transfection-MIA PaCa-2 cells were transfected with a vector that expresses PCAF cDNA or PCAF ⌬HAT or with an empty vector without the PCAF insert by using commercially available FuGENE 6 transfection reagent (Roche Applied Science). Cells were cultured fur-

T␤RII Sp1C/NF-Y
Ϫ112 agc tct cgg agg ggc tgg tct agg aaa cat gat tgg cag cta cga ga Ϫ66 T␤RII Sp1C mut/NF-Y Ϫ112 agc tct cgg agg aat tcg tct agg aaa cat gat tgg cag cta cga ga Ϫ66 T␤RII Sp1C/NF-Y mut Ϫ112 agc tct cgg agg ggc tgg tct agg aaa cat ggt gta cag cta cga ga Ϫ66 ther and selected in medium containing 600 g/ml G418. Individual G418-resistant colonies were isolated by ring cloning after drug selection, established as individual clones, and used for additional analysis. mRNA Expression of T␤RII-RNA was extracted using TRIzol reagent (Invitrogen). Two g of total RNA were reverse transcribed into cDNA as described previously. PCR was performed using the following specific RII primers: forward primer, 5Ј-CAGAAATCCTGCATGAG-CAA-3Ј (ϩ212/ϩ231); and reverse primer, 5Ј-GCTGATGCCTGTCACT-TGAA-3Ј (ϩ516/ϩ535). Expression levels of ␤-actin were used as an internal loading control. The PCR products were run in 2% agarose gels, stained with ethidium bromide, and photographed.
In Vivo Sodium [ 3 H]Acetate Labeling and Immunoprecipitation Experiments-[ 3 H]Acetate labeling and immunoprecipitation were performed as described by Hung et al. (26), with a slight modification. MIA PaCa-2 cells stably transfected with expression vector for PCAF-WT, PCAF ⌬HAT defective mutant, or empty vector were plated on 10-cm plates. At 70% confluence, 1 mCi/ml sodium [ 3 H]acetate (4.1 Ci/mmol; ICN) and 100 ng/ml TSA were added to the medium and incubated for 2 h. Nuclear extracts were prepared and immunoprecipitated with Sp1 antibodies. The immunoprecipitated samples were captured with 30 l of 50% slurry of protein A-agarose and washed five times with radioimmune precipitation assay buffer. The beads were resuspended in scintillation fluid and counted for radioactivity.

Expression and Activity of HDAC in Pancreatic Epithelial
Cells and Pancreatic Cancer Cell Lines-HDACs are known to contribute to transcriptional repression of tumor suppressor genes and are reported to be overexpressed in cancer cells including pancreatic cancer (27). Consistent with this notion, we showed that an immortalized pancreatic epithelial cell line expressed lower levels of HDAC 1 protein compared with five pancreatic cancer cell lines (Fig. 1A). Expression levels of a second protein, c-Jun/AP1, were shown to indicate that pancreatic cancer cells do not show a global increase of expression of nuclear proteins when compared with normal pancreatic epithelial cells (Fig. 1A). As expected, the greater levels of HDAC expression were also related to an overall increase in the level of HDAC activity (Fig. 1B). We reported previously that the T␤RII gene is transcriptionally silenced in MIA PaCa-2 cells, at least in part, by a mechanism involving HDAC (20). We further found that treatment of pancreatic cancer cell lines with the HDAC inhibitor TSA leads to a transcriptional activation of the T␤RII promoter and an increased expression of the T␤RII receptor (20). Treatment of the pancreatic cancer cell lines with TSA caused an increase in T␤RII promoter activity of all five pancreatic cancer cell lines, indicating that the increase in promoter activity by TSA was not restricted to a single pancreatic cancer cell line (Fig. 2).
These results support our previous finding that HDACs can regulate the expression of the T␤RII gene (20). In that report, we showed that a specific Sp1 site located at Ϫ102 is required for TSA responsiveness. Moreover, we (20) and others (21) using a separate cell type showed that an adjacent NF-Y site located at Ϫ83 is critical to suppression of T␤RII promoter activity by HDACs. In this study, we evaluated the mechanism by which HDAC inhibitors modulated the components of the transcriptional complex that binds to this region of the T␤RII promoter. MIA PaCa-2 cells were primarily used as a model to examine this mechanism because they appear to show severe repression of T␤RII gene expression.
TSA Alters the Sp1⅐NF-Y Multiprotein Complexes-The Sp1 and NF-Y binding sites within the NRE2 region are reported to play key roles in the TSA activation of T␤RII promoter (20,21). We surmised that treatment with TSA would cause changes within the transcriptional complexes that bind to this region. Therefore, we determined whether the amount of Sp1, NF-Y, and other potential transcriptional regulators was changed in MIA PaCa-2 cells after TSA treatment. The expression levels of PCAF and p300, which have intrinsic HAT activity, and HDAC-1 were determined because these factors may be involved in the effect of increasing T␤RII promoter activity by TSA. Immunoblotting experiments with nuclear extracts from MIA PaCa-2 cells showed that the relative amounts of Sp1, NF-Y, p300, PCAF, and HDAC-1 were unchanged after TSA treatment (Fig. 3A, N/E). However, it is possible that TSA might induce posttranslational modifications or increase the recruitment of transcriptional components to the specific site, making the transcriptional complex more efficient in its transcriptional activation.
Therefore, we next investigated whether the TSA-induced T␤RII promoter activity is dependent on the modulation of multiprotein complexes. Co-immunoprecipitation was performed to confirm the association of proteins that might likely form complexes with Sp1 and NF-Y. Nuclear extracts isolated from MIA PaCa-2 cells with or without TSA treatment were immunoprecipitated with Sp1 and NF-Y antibodies, respectively. The results show that Sp1 and NF-Y were present in the same complex (Fig. 3A). We also determined whether HDAC or PCAF and p300, two known cofactors with intrinsic histone acetyltransferase activity, formed complexes with Sp1 and NF-Y. These studies demonstrate that both Sp1 and NF-Y were able to associate with p300, PCAF, and HDAC-1 (Fig. 3A). After treatment of cells with TSA, the association of PCAF with either Sp1 or NF-Y complex was increased. p300 was only faintly detected by these immunoprecipitation assays but also appeared to show some increase in association with Sp1 and NF-Y (Fig. 3A). In contrast to PCAF and p300, HDAC-1 appeared to show a slight decrease in association with Sp1 or NF-Y after TSA treatment (Fig. 3A). Thus, our results suggested that changes in the composition of multiprotein complexes occurred during the TSA-induced T␤RII promoter activation.
The immunoprecipitation analysis reflects only changes in association of these various transcriptional components within the total pool of nuclear proteins. Therefore, we next sought to determine whether the transcriptional components that bind to the Sp1C/NF-Y site of the T␤RII promoter were also changed after treatment with TSA. To assess this possibility, DAPAs were performed with biotin-end-labeled oligonucleotides corresponding to the sequences from Ϫ112 to Ϫ66 of the T␤RII promoter (containing the Sp1C site and the adjacent CCAAT box to which NF-Y binds). The results showed that recruitment of p300 and PCAF to the transcription complex was increased after treatment with TSA and reached a peak after 12 h of treatment (Fig. 3B). Conversely, TSA treatment of cells caused a decrease of HDAC-1 binding, which was apparent at 12 h and decreased further at 24 h (Fig. 3B). Therefore, p300 and PCAF were recruited to the T␤RII promoter in response to treatment with TSA in a time-dependent manner. In contrast, the level of HDAC associated with the complex decreased in concordance with the increase in promoter activity. Next we determined whether the recruitment of PCAF and p300 could also explain the increase in T␤RII promoter activity induced by TSA in other pancreatic cell lines. We examined two additional cell lines, UK Pan-1 and BxPC3, by DAPA analysis and found results similar to those seen for MIA PaCa-2. We show a representative result for DAPA analysis of UK Pan-1 in Fig.  3C. Thus, TSA induced similar changes in the transcriptional complex that activates T␤RII promoter in different pancreatic cancer cell lines.
Furthermore, we performed a ChIP assay to examine whether TSA, under in vivo binding conditions, modulated the complex formation of p300, PCAF, and Sp1 to the T␤RII promoter. In agreement with the results of DAPA and co-immunoprecipitation, p300 and PCAF binding to the T␤RII promoter was greatly enhanced by TSA, whereas, Sp1 binding was not changed (Fig. 3D).

Role of the Sp1C Site and NF-Y Site in TSA-induced Activation of the T␤RII Promoter-
The role of the Sp1C and NF-Y sites required for TSA-mediated activation of the T␤RII promoter in multiprotein complex formation was next investigated by DAPA. These assays used biotin-end-labeled wild-type oligonucleotides corresponding to the sequences from Ϫ112 to Ϫ65 of the T␤RII promoter containing the Sp1C and NF-Y binding sites (Sp1C/NF-Y) with the Sp1C site mutated (Sp1C mut/NF-Y) or with the NF-Y binding site mutated (Sp1C/NF-Ymut). The results showed that when the Sp1C site was mutated, the binding of p300, PCAF, or HDAC was greatly diminished compared with that with wild-type oligonucleotides (Fig.  4A). This result suggested that the Sp1C site is required for the formation of Sp1⅐HDAC-1⅐p300⅐PCAF multiprotein complex. Surprisingly, the binding of Sp1 to the Sp1C mut/NF-Y was not decreased but increased (Fig. 4A). To elucidate the influence of Sp1C site mutation on Sp1 binding, an electrophoretic mobility shift analysis was performed using the T␤RII promoter wildtype probe only with Sp1C site or its mutant. These analyses indicated that Sp1 could not bind to the Sp1C site mutant, which confirmed that the Sp1C site was essential for the binding of Sp1 to this site (data not shown). This suggests that the Sp1 that binds to the Sp1 mut/NF-Y oligonucleotide (Fig. 4A) may be associated with NF-Y or a NF-Y protein complex.
When the NF-Y site was mutated, the binding of p300, PCAF, HDAC, and Sp1 was almost the same as that with wild-type oligonucleotides, except that the binding of NF-Y was greatly diminished to a point that was almost undetectable (Fig. 4B).
Overexpression of PCAF or p300 Up-regulates T␤RII Expression and Sp1 Acetylation-We further examined the functional roles of histone acetyltransferases, PCAF, and p300 in regulating T␤RII promoter activity. MIA PaCa-2 cells were co-transfected with a T␤RII promoter-luciferase construct and with either Sp1, PCAF, p300 or HAT mutants of PCAF, and p300. The results showed that TSA induces T␤RII promoter activity as expected. Transient transfection of PCAF could further increase the level of activity, especially when Sp1 was co-transfected with PCAF. However, the PCAF HAT defective mutant could not stimulate T␤RII promoter activity, suggesting that PCAF potentiates TSA-induced T␤RII promoter activity dependent on the HAT activity of PCAF (Fig. 5A).
Transient transfection with a vector expressing p300 gave a result similar to that seen with PCAF. However, the p300 HAT defective mutant could still stimulate T␤RII promoter activity. This suggested that p300 potentiates TSA-induced T␤RII promoter activity independent of p300 HAT (Fig. 5B). Interestingly, in the absence of TSA, transfection of p300 or PCAF only slightly activated T␤RII promoter, suggesting that TSA-mediated inhibition of HDAC activity facilitates p300 and PCAF to interact with other factors to induce T␤RII promoter activity.
We hypothesized that inhibition of HDAC activity may lead to regional acetylation of transcriptional factors that bind to the Sp1C site. To address this possibility, we tested whether the acetylation status of Sp1 in MIA PaCa-2 cells was altered after treatment of cells with TSA. At different times after TSA treatment, MIA PaCa-2 nuclear extracts were immunoprecipitated with a Sp1 antibody and immunoblotted with pan-acetyl antibody. As shown in Fig. 5C, Sp1 could be acetylated in vivo. Moreover, addition of TSA enhanced the acetylation of Sp1 in a time-dependent manner. The greatest level of Sp1 acetylation was seen after 24 h of treatment with TSA, which is consistent with the time-dependent increase of TSA-induced T␤RII promoter activation.
Because transient expression of PCAF-WT cDNA led to an increase in RII promoter activity dependent on its HAT activity, we stably transfected MIA PaCa-2 cells with PCAF-WT cDNA or its HAT defective mutant to test the effect of PCAF on T␤RII expression and Sp1 acetylation. MIA PaCa-2 cells were stably transfected with an empty vector or with vectors with FLAG-tagged PCAF or a PCAF-HAT defective mutant. The transfectants were denoted as MIA PaCa-2 (Neo), MIA PaCa-2 (PCAF-WT), and MIA PaCa-2 (PCAF ⌬HAT) clone cells, respectively. Transfected cells were cloned and screened with FLAG tag expression (Fig. 6A), and their T␤RII mRNA expressions were analyzed by reverse transcription-PCR. A representative result is shown in Fig. 6B. As compared with MIA PaCa-2 (Neo) cells, the MIA PaCa-2 (PCAF-WT) clone showed an increase in FIG. 3. TSA treatment alters the Sp1⅐NF-Y protein complex. A, immunoprecipitation assays showing that Sp1⅐NF-Y complexes with p300, PCAF, and HDAC-1. Nuclear extracts (100 g) from untreated or TSA-treated MIA PaCa-2 cells were immunoprecipitated with antibodies to Sp1 or NF-Y. The protein complexes were resolved by SDS-PAGE and detected with the indicated antibodies. Nuclear extracts (N/E; 15 g) were used as a control. B and C, DNA affinity precipitation assays showing the modulation of protein components that bind to the T␤RII promoter in response to TSA in MIA PaCa-2 and UK Pan-1 cells. Biotin-end-labeled oligonucleotides corresponding to the sequences from Ϫ112 to Ϫ66 of the T␤RII promoter (containing the Sp1C site and the adjacent inverted CCAAT box) were incubated with nuclear extracts from MIA PaCa-2 cells (B) or UK Pan-1 cells (C) treated with or without TSA for the indicated time periods. Bound materials were analyzed by immunoblot assay using anti-Sp1, NF-Y, PCAF, p300, and HDAC-1 antibodies. D, TSA treatment induces the association of T␤RII promoter with p300 and PCAF in vivo. Untreated or TSA-treated MIA PaCa-2 cells were formaldehyde-cross-linked. Chromatin was sonicated to yield 500 -600-bp DNA fragments and immunoprecipitated with anti-PCAF, p300, or Sp1 antibody. The T␤RII promoter region was amplified by PCR using specific primers as described under "Experimental Procedures." T␤RII mRNA expression, whereas MIA PaCA-2 (PCAF ⌬HAT) cells did not show any increase but did show a slight decrease. This result is in agreement with the level of T␤RII promoter activity observed in MIA PaCa-2 cells transiently transfected with PCAF-WT, PCAF ⌬HAT, or empty vector (Fig. 5A). To evaluate the effect of overexpression of PCAF-WT and PCAF ⌬HAT on their association with T␤RII promoter in vivo, ChIP assays were performed. The results showed that in MIA PaCa-2 (PCAF-WT) cells, the association of PCAF with T␤RII promoter increased, whereas in MIA PaCa-2 (PCAF ⌬HAT) cells, the association did not show a significant difference from that of MIA PaCa-2 (Neo) cells.
Because PCAF possesses intrinsic histone acetyltransferase activity, we hypothesized that it was responsible for the Sp1 acetylation observed after TSA treatment (Fig. 5C). To determine whether Sp1 is a target of PCAF HAT activity, we pulselabeled MIA PaCa-2 (Neo), MIA PaCa-2 (PCAF-WT), and MIA PaCa-2 (PCAF ⌬HAT) cells with [ 3 H]acetate for 2 h and then subjected them to cell lysis and immunoprecipitation with Sp1 antibodies. The liquid scintillation results shown in Fig. 6D demonstrate that Sp1 antibodies, but not isotype-matched control antibodies, precipitate [ 3 H]acetate-labeled Sp1, indicating that acetylation of Sp1 occurs in MIA PaCa-2 cells. With TSA treatment, the [ 3 H]acetate incorporation of Sp1 increased, which is consistent with the results of immunoprecipitation experiments using anti-pan-acetyl antibodies (Fig. 5C). In MIA PaCa-2 (PCAF-WT) cells, the [ 3 H]acetate incorporation was much higher than that in MIA PaCa-2 (Neo) and MIA PaCa-2 (PCAF ⌬HAT) cells, indicating that PCAF could stimulate Sp1 acetylation in vivo. DISCUSSION T␤RII is a tumor suppressor gene that is found to be epigenetically silenced or down-regulated in several different cell types (7,23). We previously showed that HDACs play a role in down-regulating T␤RII expression in the human pancreatic cancer cell line MIA PaCa-2 (20). In this report, we demonstrated the mechanism by which TSA, an inhibitor of HDAC, induces T␤RII gene expression in human pancreatic cancer cells. Moreover, TSA activated T␤RII promoter activity in a panel of five pancreatic cancer cells, further suggesting that HDACs play a general role in regulating T␤RII expression in pancreatic cancer cells. This mechanism involves the induced acetylation of Sp1 and changes in a multiprotein complex containing p300, PCAF, Sp1, and NF-Y. Treatment of pancreatic FIG. 5. The roles of p300 and PCAF in TSA-induced T␤RII promoter activity. A, MIA PaCa-2 cells were transiently co-transfected with a T␤RII promoter-luciferase reporter construct and PCAF wild type (PCAF-WT) or PCAF HAT mutant (PCAF ⌬HAT) or together with a Sp1 construct or the empty vector and then treated with or without TSA. Luciferase activity was measured at 48 h after transfection and normalized against protein concentration. B, MIA PaCa-2 cells were co-transfected with a T␤RII promoter-luciferase reporter construct and p300 wild type (p300 WT) or p300 HAT mutant (p300 ⌬HAT), Sp1 construct, or the empty vector and then treated with or without TSA. Luciferase activity was measured at 48 h after transfection and normalized against protein concentration. C, nuclear extracts from MIA PaCa-2 cells treated with or without TSA for the indicated time were immunoprecipitated with anti-Sp1 antibodies. Subsequently, immunoblotting was performed with anti-pan-acetyl antibodies. Input levels of Sp1 were detected by immunoblotting with anti-Sp1 antibodies.
cancer cells with TSA strongly activates the T␤RII promoter. Separate studies previously showed that an Sp1 site (Sp1C) located at Ϫ102 and a NF-Y site located at Ϫ83 bp within the NRE2 (Ϫ100 to Ϫ67) region of the T␤RII promoter are required for TSA-mediated transcriptional activation of T␤RII (20,21). The present study was undertaken to determine whether TSA modulated the transcriptional components that bound to these two sites and whether this modulation was involved in TSAmediated tansactivation of the T␤RII promoter. Results from immunoprecipitation experiments indicated that addition of TSA enhanced the acetylation of Sp1 in a time-dependent manner. Moreover, p300 and PCAF, two well-known cofactors with histone acetyltransferase activity, as well as HDAC were present in the same complex with Sp1⅐NF-Y. An increased association of p300 and PCAF and a deceased association of HDAC with Sp1⅐NF-Y complex on T␤RII promoter were observed after TSA treatment using in vitro (DAPA) and in vivo (ChIP) assays. Furthermore, in the absence of TSA, transient transfection of p300 or PCAF only slightly activated the T␤RII promoter, and co-transfection with Sp1 did not further increase T␤RII promoter activity. However, in the presence of TSA, both p300 and PCAF could greatly potentiate T␤RII promoter activity, and co-transfection with Sp1 further increased promoter activation. This implies that TSA-mediated inhibition of HDAC activity facilitates the functional collaboration of p300 and PCAF with Sp1 in T␤RII promoter activation. These studies further support the notion that modulation of this multiprotein transcriptional complex is essential for the T␤RII promoter activation.
To elucidate the roles of Sp1C and NF-Y sites in TSA-mediated reversal of T␤RII gene silencing, DAPA assays were performed using wild-type, Sp1C site-mutated, or NF-Y site-mutated NRE2 probe. The results showed that the Sp1C site is required for the multiprotein complex formation. It is interesting to note that although Sp1 could not bind to mutated Sp1C site, as confirmed by electrophoretic mobility shift analysis, the binding of Sp1 to NRE2 probe with Sp1C site mutated was not decreased but increased. It is possible that Sp1 may bind with NF-Y, NF-Y protein complex, or to the NF-Y site. These possibilities are consistent with other studies (28 -30) showing the physical and functional interaction between NF-Y and Sp1. Moreover, Roder et al. (30) observed that an Sp1 interaction domain is located between amino acids 55 and 139 of NF-YA. In contrast to Sp1C site mutation, NF-Y site mutation did not greatly influence the binding of p300, PCAF, Sp1, and HDAC to the NRE2 probe but abolished the binding of NF-Y. These studies are in agreement with previous studies (20,21) indicating that a specific Sp1 site (Sp1C) and an inverted CCAAT element within the NRE2 region located at Ϫ102 and Ϫ83 bp are required for TSA-mediated T␤RII induction. The study further demonstrated for the first time a physical and functional cooperativity among Sp1, NF-Y, p300, and PCAF in TSA-mediated transcriptional activation of T␤RII.
Both p300 and PCAF are known to possess intrinsic HAT activity (14,15). Whereas the HAT activity of PCAF is required for TSA-induced T␤RII promoter activity, the HAT domain of p300 is dispensable for TSA-mediated activation. This suggests overlapping but distinct roles of p300 and PCAF in T␤RII transcription. In TSA-mediated T␤RII promoter activation, p300, as an adapter protein, may bridge Sp1 and NF-Y with the basal transcription machinery aiding in the formation of the pre-initiation complex to activate transcription. Stable transfection of MIA PaCa-2 cells with a vector expressing PCAF increased the association of PCAF with T␤RII promoter and caused an increase in T␤RII mRNA. However, the PCAF HAT defective mutant failed to induce transactivation, confirming FIG. 6. Stable expression of PCAF cDNA in MIA PaCa-2 cells causes an increase in expression of T␤RII mRNA and acetylation of Sp1. A, MIA PaCa-2 cells stably transfected with an empty vector or vectors with FLAG-tagged PCAF or a PCAF defective mutant were screened with Western blot by FLAG antibody and denoted as MIA PaCa-2 (Neo), MIA PaCa-2 (PCAF-WT), and MIA PaCa-2 (PCAF ⌬HAT) clone cells, respectively. A Ponceau S-stained band migrating at a molecular weight of ␤-actin was used to ensure protein transfer to nitrocellulose filters and to judge relative protein loading. B, stable transfection of MIA PaCa-2 cells with PCAF cDNA causes an increase in RII expression. Reverse transcription-PCR was performed with appropriate primers from total RNA extracted to determine the mRNA expression of the T␤RII gene in MIA PaCa-2 (Neo), MIA PaCa-2 (PCAF-WT), and MIA PaCa-2 (PCAF ⌬HAT) cells. ␤-Actin mRNA levels were measured as an internal control. C, stable transfection of MIA PaCa-2 cells with PCAF cDNA causes an increase in the association of PCAF with T␤RII promoter. ChIP assay was performed to test the association of PCAF with T␤RII promoter in MIA PaCa-2 (Neo), MIA PaCa-2 (PCAF-WT), and MIA PaCa-2 (PCAF ⌬HAT) cells with or without treatment with TSA (100 ng/ml). D, overexpression of PCAF causes the acetylation of Sp1. Stable MIA PaCa-2 transfectants containing expression constructs for FLAG-tagged PCAF-WT, PCAF HAT defective mutant, or empty vector were incubated with [ 3 H]acetate (1 mCi/ml) for 2 h in the absence or presence of TSA (100 ng/ml). Nuclear extracts were prepared and subjected to immunoprecipitation with anti-Sp1 antibody. The immunoprecipitated beads were resuspended in scintillation fluid and counted for radioactivity. that the HAT domain of PCAF is essential for T␤RII transcription. To determine the target of the PCAF HAT domain, [ 3 H]acetate labeling and immunoprecipitation assay were performed and showed that Sp1 is one of the targets of PCAF HAT activity. The requirement of the p300 and PCAF HAT domain for transcription activity may vary for different promoters. Previous reports indicate that PCAF activates MDR1-, MyoD-, and p73-mediated p21 expression through its HAT domain (31)(32)(33) but that PCAF activation of a CRE-LacZ promoter construct, HTLV-1 LTR, is independent of its HAT domain (32). Similarly, previous reports (23, 31, 34 -37, 39) indicate that the requirement for p300 HAT activity differs among promoters. p300 HAT activity is not required for transcription activation of MyoD and retinoic acid receptor promoters (26), but it is required for Tax transactivation of the HTLV-1 LTR, Sp1-mediated p21 transcription, EKLF-induced ␤-globin expression, and T␤RI and COX-2 transactivation (23, 34 -37, 39) .
Based on the results presented in the current study, potential mechanisms for the activation of T␤RII promoter by TSA may be proposed. One possibility is that TSA treatment inhibits HDAC activity and leads to a local disruption of nucleosome structure of the T␤RII promoter and the decondensation of chromatin, which permits PCAF and p300 to be recruited into Sp1⅐NF-Y complex. This may be followed by release of HDAC from the complex, resulting in the activation of the T␤RII promoter. A second possibility is based on acetylation of proteins modifying their biological activity, which may include altering protein-protein interactions, DNA recognition, and protein stability. The acetylation of Sp1 and KLF/Sp1 family members, such as EKLF and KLF13, has been shown to enhance transcriptional potency and affect protein-protein interactions (38,40,41). TSA-mediated Sp1 acetylation may alter Sp1 configuration and contribute to the recruitment of p300 and PCAF to Sp1⅐NF-Y complex (39). Increased association of PCAF may further acetylate Sp1, which forms a positive feedback for Sp1 acetylation and the induction of T␤RII promoter activity.
In summary, we demonstrate here for the first time that the modulation of the Sp1⅐p300⅐PCAF⅐NF-Y multiprotein complex plays a pivotal role in the transcriptional activation of T␤RII by TSA in pancreatic cancer cells. The Sp1C site in the T␤RII promoter plays a crucial role for formation of the multiprotein complex. Furthermore, the induction of T␤RII by TSA is dependent on the HAT activity of PCAF and may be involved in the acetylation of Sp1.