The androgen receptor acetylation site regulates cAMP and AKT but not ERK-induced activity.

The androgen receptor (AR) regulates ligand-dependent gene transcription upon binding specific DNA sequences. The AR conveys both trans-activation and trans-repression functions, which together contribute to prostate cellular growth, differentiation, and apoptosis. Like histone H3, the AR is post-translationally modified by both acetylation and phosphorylation. The histone acetyltransferase p300 transactivates the AR and directly acetylates the AR in vitro at a conserved motif. Point mutations of the AR acetylation motif that abrogate acetylation reduce trans-activation by p300 without affecting the trans-repression function of the AR. The current studies assessed the functional relationship between acetylation and phosphorylation of the AR. Herein trans-activation of the AR acetylation site mutants were enhanced by the p42/p44 MAPK pathway but were defective in regulation by protein kinase A (PKA) signaling. PKA inhibition augmented ARwt activity but not AR acetylation mutant gene reporter activity and association at an androgen response element in chromatin immunoprecipitation assays. Mutations of the lysine residues at the AR acetylation site reduced trichostatin A (TSA) responsiveness and ligand-induced phosphorylation of the AR. The AR acetylation site mutant formed ligand-induced phosphorylation-dependent isoforms with distinguishable characteristics from wild type AR as determined with two-dimensional electrophoresis. Conversely, point mutation of a subset of AR phosphorylation sites reduced trichostatin A responsiveness and trans-activation by histone acetyltransferases. Together these studies suggest that acetylation and phosphorylation of the AR are linked events and that the conserved AR lysine motif contributes to a select subset of pathways governing AR activity.

tion factor and the basal apparatus and contribute to the assembly of high molecular weight "enhanceosomes" (reviewed in Ref. 9). p300/CBP convey enzymatic activity toward histones, with the relative activity correlating under certain circumstances with their transcriptional coactivator function. Acetylation facilitates binding of transcription factors to specific target DNA sequences by destabilizing nucleosomes bound to the promoter region of a target gene. Direct acetylation of nonhistone proteins, including transcription factors, regulates AR activity (reviewed in Refs. 10 and 11).
Acetylation of the tumor suppressor p53 (12), the transcription factors Kruppel-like factor (13), and the erythroid cell differentiation factor GATA-1 (14) enhanced the trans-activation function. More recently, nuclear hormone receptors and nuclear receptor coactivators were shown to serve as targets of acetylation. Direct acetylation of activator of thyroid and retinoic acid receptors ACTR (15) or the ER␣ (16) contributed to ligand-dependent transcriptional attenuation, whereas direct acetylation of the AR contributed to full ligand-dependent activity (4). Glutamine or threonine substitutions of lysine residues within the AR acetylation site enhanced DHT-dependent activation of androgen-responsive reporter genes, increased physical association with p300, and reduced binding to the N-CoR corepressor (17,18). Alanine substitution mutants of the AR-acetylated lysine residues showed reduced p300 binding and increased N-CoR binding (17). Together these studies suggested the lysine motif of the AR may serve as an important regulator of coactivator/corepressor binding and acetylation of the AR may be involved in regulating ligand-dependent activity.
Activity of the steroid receptors is also regulated by direct phosphorylation through distinct kinases. The estrogen receptor (ER␣) is phosphorylated by the mitogen-activated protein kinase (MAPK) and protein kinase A (PKA) pathways (19,20) to enhance ligand-independent transcriptional activity. Phosphorylation of the glucocorticoid receptor is induced during S-phase transition (21), and the progesterone receptor phosphorylation is regulated by diverse signals, including the cellcycle epidermal growth factor and the cAMP pathway (22). In prostate cancer cell lines the phosphorylated AR forms are active, and dephosphorylation of the AR inhibits AR activity (23,24). Phosphorylation of the 110-kDa AR protein occurs rapidly resulting in the formation of a 110 -112-kDa doublet, with a third 114-kDa hyperphosphorylation isoform appearing upon addition of DHT (24,25). Activation of the cAMP pathway leads to a rapid dephosphorylation of the AR likely through induction of PKA-inducible phosphatases (26). Activity of the AR is enhanced by induction of the MAPK kinase pathway (27,28).
Post-translational modification by acetylation and phosphorylation under several circumstances may be integrated processes (29 -31). Evidence supporting a model that these two post-translational processes may be convergent includes studies of the immediate early (IE) genes such as ternary complex factor (TCF) and histone H3. Contemporaneous with IE gene induction, histone acetylation spanning several nucleosomes is observed. Almost invariably a second nucleosomal modification that occurs upon IE gene induction is the phosphorylation of histone H3 (32) and the nucleosome-binding high mobility group protein HMG-14 (33). Histone H3 is itself both acetylated and phosphorylated, and phosphorylated H3 is more sensitive to the histone deacetylase inhibitor trichostatin A (TSA) than nonphosphorylated H3 (34), providing strong evidence that these two events can be convergent. CREB phosphorylation is augmented indirectly upon activation of the PKA pathway through histone acetylase-regulated alterations in the local chromatin (35,36). The possibility that nuclear receptors themselves may be the direct target of interdependent acetylation/ phosphorylation events remained to be determined.
In previous studies, AR acetylation site alanine substitution mutants conveyed wild type trans-repression of SP-1 and NFB activity, bound ligand with wild type affinity, and were sumoylated like the wild type receptors. However, induction by the HDAC inhibitor TSA and activation by the p300 coactivator binding was reduced in cultured cells (4,17). In view of previous studies suggesting H3 phosphorylation and acetylation may be linked events, we investigated the functional significance of the AR lysine residue motif in signaling by the kinases MAPKK, AKT, and PKA. Mutation of the AR acetylation site did not affect MAPK signaling to the AR. However, the AR acetylation site mutants were defective in regulation by the HDAC inhibitor TSA and by cAMP and AKT signaling. Inhibition of cAMP signaling augmented recruitment of the ARwt to the ARE of the PSA promoter in chromatin immunoprecipitation assays but failed to induce recruitment of the AR acetylation site mutant. Point mutations of six distinct AR phosphorylation sites identified one phosphorylation site regulating HDAC responsiveness. In contrast to the three phosphorylationdependent isoforms of the wild type AR (110, 112, and 114 kDa), the acetylation mutations lacked the 114-kDa hyperphosphorylated form, suggesting acetylation and phosphorylation of the AR are functionally convergent events. These studies suggest that the conserved AR lysine residues that are acetylated in vitro and in cultured cells may play a role in coordinating a subset of kinase modules signaling to the AR.

MATERIALS AND METHODS
Reporter Genes and Expression Vectors-The expression vectors for the protein kinase A catalytic subunit (PKAc), the PKAcmut subunit (37), and AKT (38) were previously described. Gal4-CREB, and Gal4CREB S133 mutant (39) were previously described. For androgen receptor-regulated gene transcription, a 600-bp fragment of the prostate-specific antigen (PSA) promoter with an additional 2.4-kb enhancer sequence cloned upstream of luciferase (PSA-LUC) was used (4). In addition, the androgen-responsive synthetic reporter construction MMTV-LUC (4) was used. The wild type human AR was subcloned from pARO into pcDNA3, and the AR acetylation site mutants AR K(632/633)A and AR K630A were derived by PCR-directed amplification using sequence-specific primers and cloned into pcDNA3 (4). The integrity of all constructs was confirmed by sequence analysis. The activating mitogenactivated protein kinase kinase expression plasmids MEKwt, MEK R⌬F , and MEK K97M (40) and the wt c-fos-LUC, TCF, or SRE mutants (pm12 and pm18) (41) were previously described. The ARwt and point mutants of the AR phosphorylation site (AR S81A , AR S94A , AR S256A , AR S308A , AR S426A , and AR S650A ) were described elsewhere (23).
Cell Culture, DNA Transfection, and Luciferase Assays-Cell culture, DNA transfection, and luciferase assays were performed as previously described (4,17,18). The prostate cancer cell lines DU145 and LNCaP and the HEK293T cell line were cultured in DMEM supplemented with 10% fetal calf serum, 1% penicillin, and 1% streptomycin. Cells were plated at a density of 1 ϫ 10 5 cells in a 24-well plate on the day prior to transfection with LipofectAMINE Plus (Invitrogen). The DNA/Lipo-fectAMINE mix was added to the cells in Opti-Mem. Cells were incubated in media containing 10% charcoal stripped fetal bovine serum prior to experimentation using dihydrotestosterone (DHT) (4). The 8-bromo-cAMP, forskolin, N-2-(p-bromocinnamyl)amino)ethyl-5-isoquinolinesulfonamide (H-89), and PD98059 (Calbiochem-Novabiochem International) were reconstituted and stored as recommended by the manufacturer. At least two different plasmid preparations of each construct were used. In cotransfection experiments, a dose response was determined in each experiment with 150 and 300 ng of expression vector and the promoter reporter plasmids (1.2 g). Luciferase activity was normalized for transfection using ␤-galactosidase reporters as an internal control. Luciferase assays were performed at room temperature using an Autolumat LB 953 (EG&G Berthold). The -fold effect was determined for 150 -300 ng of expression vector with comparison made to the effect of the empty expression vector cassette and statistical analyses was performed using the Mann Whitney U test.
Nuclear Extract Preparation and Western Blots-Preparation of the cytoplasmic and nuclear extracts from the transfected cells was performed essentially as described previously (40). The antibodies used in Western blot analysis were to the AR (N-20, Upstate Biotechnology, Lake Placid, NY) and to the guanine nucleotide dissociation inhibitor (GDI) (a generous gift from Dr. Perry Bickel, Washington University, St. Louis, MO), which was used as an internal control for protein abundance. For detection of protein, the membrane was incubated with anti-AR (N-20, 1:2000) at room temperature for 2 h. The blots were then washed three times with 0.1% Tween 20 phosphate-buffered saline and incubated with the appropriate horseradish peroxidase-conjugated second antibody. Proteins were visualized by the enhanced chemiluminescence system (Amersham Biosciences). The abundance of immunoreactive protein was quantified by phosphorimaging using an ImageQuaNT version 1.11 computing densitometer (Amersham Biosciences). Two-dimensional Gradient Gel Electrophoresis-Two-dimensional PAGE was performed as described by O'Farrell (42). Briefly, HEK293T cells were transfected with pcDNA 3 ARwt and pcDNA3-AR K630A plasmid DNA by the calcium phosphate method. Five hours after transfection, the medium was changed to phenol-free DMEM supplemented with 10% charcoal-dextran-treated fetal bovine serum and cells were treated either with DHT (100 nM) or vehicle for 24 h. Cell lysate containing 0.5 mg of total protein was pre-cleared and subjected to immunoprecipitation using anti-AR antibody and protein A-agarose beads overnight at 4°C. The beads were then washed four times with cell lysis buffer (50 mM HEPES, pH 7.2, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 0.1 mM phenylmethylsulfonyl fluoride with the protease inhibitors mixture (Roche Applied Science, catalog #1836145)) and were washed twice with ddH 2 O. The immunoprecipitated AR protein was eluted from the bead with 280 l of rehydration buffer (8 M urea, 2% (W/V) CHAPS, 2% IPG buffer (carrier ampholyte mixture) (pH 4.7), 7 mg/2.5 ml DTT, and a trace amount of bromphenol blue), and 250 l was applied to the Immobiline DryStrip (Amersham Biosciences, pH 4 -7, 13 cm). The gel strip was then overlaid with 2.5 ml of IPG Cover Fluid, and rehydration was allowed to occur overnight at room temperature in the Immobiline DryStrip Reswelling Tray. The first dimension isoelectric focusing was run on a Multiphor II (Amersham Biosciences) flat bed electrophoresis tank equipped with the Immobiline DryStrip kit as described by the manufacturer. The gel strip was then equilibrated and secured with 0.5% agarose in SDS-PAGE sample buffer on top of the stacking gel of a 1.5-mm-thick, 7% SDS-polyacrylamide gel. The separated proteins were transferred to polyvinylidene difluoride membrane for immunoblotting.
Chromatin Immunoprecipitation Assay (ChIP Assay)-ChIP analysis was performed following a protocol provided by Upstate Biotechnology under modified conditions. HEK293T cells were grown in DMEM with 10% charcoal-dextran-stripped serum for 3 days prior to treatment. After stimulation, the cells were cross-linked by adding 1.0% formaldehyde buffer containing 100 mM sodium chloride, 1 mM EDTA-Na (pH 8.0), 0.5 mM EGTA-Na, Tris-HCl (pH 8.0) directly to culture medium for 10 min at 37°C. The medium was aspirated, the cells were washed twice using ice-cold PBS containing 10 mM DTT and protease inhibitors. The cells were then lysed with 1% SDS lysis buffer and incubated for 10 min on ice. The cell lysates were sonicated to shear DNA to lengths between 200 and 1000 bp, and the samples were diluted to 10-fold in ChIP dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris, pH 8.1, 167 mM NaCl). To reduce nonspecific background, cell pellet suspension was pre-cleared with 60 l of Salmon Sperm DNA/Protein-A-agarose-50% slurry (Upstate Biotechnology) for 2 h at 4°C with agitation. Chromatin solutions were precipitated overnight at 4°C using 4 g of anti-FLAG (M2 antibody, Sigma), (Upstate Biotechnology) with rotation. For a negative control, mouse IgG was immunoprecipitated by incubating the supernatant fraction for 1 h at 4°C with rotation. 60 l of Salmon Sperm DNA/Protein-A-agarose slurry was added for 2 h at 4°C with rotation to collect the antibody-histone complex and washed extensively following the manufacturer's protocol. Input and immunoprecipitated chromatin were incubated at 65°C overnight to reverse cross-linking. After proteinase K digestion for 1 h, DNA was extracted using a Qiagen spin column kit. Precipitated DNAs were analyzed by PCR of 30 cycles. The following primers were used for PCR to identify the androgen-responsive element in the human PSA promoter: 5Ј-ACAGACCTACTCTGGAGGAAC-3Ј and 5Ј-AAGACAGCA-ACACCTTTTT-3Ј.
Immunofluorescence Staining-Expression vectors for pcMV10FLAG-AR wt , FLAG-AR K630A , FLAG-AR K632/633A , or FLAG-AR K630Q were transfected into LNCaP cells using Genejuice transfection reagents (Novagen). The cells were then treated with vehicle or 100 nM DHT for 24 h, washed with PBS, and fixed in 4% paraformaldehyde for 30 min. Anti-FLAG (rabbit M2, Sigma, 1:100) or anti-PSA monoclonal antibody were applied for detection of FLAG-AR and PSA expression after brief permeabilization of the cells using 1% Triton in PBS for 10 min. Secondary antibodies (goat anti-rabbit antibody conjugated with Alexa Fluor 568 for FLAG (catalog #A21069, Molecular Probes, Inc.) and goat antimouse antibody conjugated with Alexa Fluor 488 for PSA (catalog #A11017, Molecular Probes, Inc.) were 1:500 diluted in PBS/3% bovine serum albumin and incubated in a humidified chamber for 45 min at 37°C. The cells were washed with PBS and mounted with Vectashield medium (Vector Laboratory Inc., H-1400).

PKA Inhibition of Liganded AR Function Involves the AR-
conserved Lysine Acetylation Motif-The AR is modified by both phosphorylation and acetylation (4,17,18,23). In recent studies, acetylation and phosphorylation of histones were shown to cooperate in transcriptional activation (29,43). The androgen-responsive synthetic reporter gene MMTV-LUC was shown to be induced in the presence of ARwt by the specific histone deacetylase inhibitor TSA (4). Using anti-acetyl lysine antibodies the AR was shown to be acetylated in cultured cells, and either p300 or P/CAF acetylated the AR in vitro (4). Edman degradation analysis of the acetylated AR products demonstrated that lysines 630, 632, and 633 were preferentially acetylated. The KLKK sequence constitutes an acetylation motif that is conserved between species (Fig. 1A). The androgenresponsive MMTV-LUC reporter was induced 3-fold at 100 nM DHT (Fig. 1B). We have previously shown that the liganded AR does not induce either the pA 3 LUC vector or several other luciferase reporter genes (RSV-LUC, cyclin E-LUC, c-fos-LUC), suggesting that the induction of the MMTV-LUC reporter is promoter-specific (44). We investigated the relationship between AR acetylation and phosphorylation function. Activity of the liganded AR is inhibited by PKA stimulators in some (26) but not all studies (45)(46)(47), with cell-type, species and reporterdependent differences described. The isoquinolinesulfonamide H-89 is a specific inhibitor of PKA (48) preferentially inhibiting the PKA (K i , 50 nM) compared with the PKC pathway (K i , 76 M). In our previous studies, H-89 selectively inhibited PKAinduced cAMP-response element (CRE) reporter activity and PKA-induced kinase activity using peptide as substrate, indicating H-89 inhibits PKA signaling both in vitro and in cultured cells (37). Consistent with previous studies (26), inhibition of endogenous PKA activity with H-89 enhanced DHT-induced human AR activity 3-fold in prostate cancer cells (Fig. 1B). In contrast, in the presence of DHT, the AR K(632/633)A was induced less than 50%, and the AR K630A mutant was not induced by H-89 ( Fig. 1, B and D). The basal activity of both the ARwt and AR mutants was not significantly affected by H-89 (Fig. 1C).
Because these studies suggested PKA-mediated repression of liganded AR activity involve the AR residues acetylated in vitro, we examined the specificity of cAMP-dependent repression of liganded AR activity. Compared with the inactive PKA catalytic subunit mutant or empty expression vector cassette, expression of the PKA catalytic subunit construction induced the activity of the cAMP response element from the glycoprotein ␣ subunit promoter linked to the luciferase reporter gene 15-fold as previously shown (49) (Fig. 2A). Forskolin, an inducer of intracellular cAMP, induced ␣CRE-LUC activity a further 2-fold compared with empty vector. As predicted, H-89 inhibited PKA-induced CRE activity. The effect on DHT-induced AR activity was assessed using the MMTV-LUC reporter (Fig. 2B). The addition of forskolin reduced DHT-induced AR activity by 40% (p Ͻ 0.01) (Fig. 2B, lane 6 versus 8). Forskolin did not affect activity of the AR K630A (Fig. 2B, lane 10 versus 12) but reduced residual activation of the AR K(632/633)A mutant. An expression vector encoding the catalytic subunit of PKA was transfected into DU145 cells, together with either the AR wt or AR K630A . The 3.5-fold induction of AR activity by DHT was inhibited 30 -40% by overexpression of the PKA catalytic subunit (Fig. 2C). In contrast, the liganded AR K630A was not repressed by either the expression of the PKA catalytic subunit or the addition of forskolin (Fig. 2, B and D). To examine further the specificity of cyclic AMP signaling in DU145 cells, a heter-ologous reporter system was used (Fig. 2E). The CREB transcription factor linked to the Gal4 binding domain was assessed using the Gal4 DNA binding sites linked to a luciferase reporter gene. Consistent with studies performed on other cell types, coexpression of the catalytic subunit of protein kinase A activated transcriptional activity of CREB ϳ8-fold (Fig. 2F). The mutant of protein kinase A catalytic subunit reduced basal CREB activity ϳ50%. Point mutation of the protein kinase A phosphorylation site of CREB at serine 133 abrogated protein kinase A catalytic subunit-dependent activation of CREB. This point mutation was also defective in repression by the PKA catalytic subunit mutant. These studies demonstrate that the PKA catalytic subunit induces protein kinase A signaling in a specific manner in DU145 cells and does not have an independent effect on luciferase reporter gene activity. Together these studies suggest PKA expression or forskolin treatment activates cAMP signaling in DU145 cells and inhibits DHT-induced activity of the AR, but they do not affect the activity of the liganded AR K630A or AR K(632/633)A mutants.
The AR Acetylation Site Regulates AR Activity Induced by JNK but Not Endogenous MAPK in Prostate Cancer Cells-Recent studies demonstrated the AR acetylation site was required for MEKK1-mediated apoptosis in DU145 cells (17). To determine whether the AR acetylation site affects AR signaling by other serine threonine kinase pathways, experiments were conducted with regulators of the ERK mitogen activated protein kinase (MAPK) pathways. The c-fos promoter linked to a luciferase reporter gene was induced by coexpression of either MAPKK or the constitutively active MAPKK mutant (R⌬F) 8to 9-fold. The catalytically dead mutant of MAPKK (K97M) failed to induce c-fos reporter activity as previously described (50). The addition of PD98059 reduced MAPKK-dependent induction of c-fos by 40% at 5 M (Fig. 3A), indicating that PD98059 was capable of inhibiting ERK-dependent reporter gene activity in DU145 cells. We next examined regulation of the MAPK-responsive AP-1-responsive reporter gene, p3TP-LUX. The constitutively active MAPKK induced AP-1 reporter activity 9-fold (Fig. 3B). The catalytically dead mutant of MAPKK (K97M) failed to induce AP-1 reporter activity. PD98059 inhibited MAPKK-induced AP-1 reporter activity by 70% indicating the MAPK pathway inhibitor PD98059 (5 M) was effective in DU145 cells (Fig. 3B). The R⌬F mutant induced MMTV-LUC activity 2-5 fold and further enhanced AR activity in the presence of DHT (Fig. 3C). The activity of AR K630A was induced by expression of R⌬F 2-fold in the absence and 3-fold in the presence of DHT. AR K(632/633)A was induced 2-fold in the absence and 4-fold in the presence of DHT. The -fold induction or activation by the MAPK pathway of the AR acetylation mutants was preserved. Because activation of MAPK signaling pathway in DU145 cells induced activity of the AR acetylation mutant, further analysis was conducted to determine the specificity of MAPK-dependent induction of gene transcription in DU145 cells. Analysis was conducted of reporter genes encoding a region of the c-fos promoter between Ϫ355 and Ϫ279 that includes the SIE (sis-inducible element), TCF, and SRE site of the c-fos promoter. Previous studies have demonstrated the induction of c-fos promoter activity by epidermal growth factor required the MAPK signaling pathway in the TCF site. The c-fos promoter construct was induced 7-fold by the activating MAPK kinase expression vector but was not induced by the point mutant MEK K97M (Fig. 3D). In contrast, mutation of the SRE site reduced MAPK kinase-dependent induction of c-fos reporter activity to ϳ5-fold and mutation of the TCF site abrogated MAPK kinase-dependent induction of c-fos activity in DU145 cells. Together these studies demonstrate DNA sequence-dependent activation of reporter gene activity in DU145 cells by MAPK kinase. AR activity is inhibited by activation of the AKT pathway and phosphorylation of the AR is required for AKT-repression (51). To examine the functional significance of the AR acetylation site in AKT-induced prostate cellular apoptosis, studies were performed in DU145 cells as previously described (52). Cells were cotransfected with wild type or mutant ARs, the AKT expression vector, or control empty vector. Cells were treated for 24 h with either DHT (10 Ϫ7 M) or control vehicle. Coexpression of AKT and the ARwt in the presence of ligand reduced AR activity ϳ50% as previously described (53). Coexpression of AKT with either AR K630A or AR K(632/633)A had no significant effect on the activity of the mutant AR compared with vector controls (Fig. 4A). Similar observations were made using the androgen-responsive PSA-LUC reporter (Fig. 4B). To examine further the specificity of AKT signaling in DU145 cells, the c-fos-responsive promoter was again employed. Overexpression of AKT induced-c-fos activity 2-fold. This activity was inhibited by the phosphatidylinositol 3-kinase inhibitor LY294002 ϳ50%. Coexpression of the dominant negative mutant of AKT K179M reduced basal c-fos promoter activity consistent with the presence of basal AKT activity in DU145 cells. Together these studies demonstrate an important function of the AR acetylation site in regulating AKT-dependent signaling in cultured human prostate cancer cells.
Experiments were conducted to determine the functional consequence of the AR acetylation site in cultured cells. Initial experiments were conducted comparing the effect of the AR acetylation mutant on the wild type androgen receptor. The wild type androgen receptor induced the PSA promoter ϳ5-fold in the presence of DHT. Coexpression of the AR acetylation mutant (AR K630A ) reduced AR-dependent induction of the PSA promoter ϳ40% in equal molar amounts (Fig. 5A). The AR LNCaP mutant induced PSA activity ϳ2-fold. Coexpression of the AR acetylation site mutant AR K630A abrogated DHT-induced activation of PSA by the LNCaP AR mutant (Fig. 5B). Comparison was next made of the AR acetylation site on abundance of the PSA protein in LNCaP cells (Fig. 5C). Comparison was made between the AR wt , the AR K630A , AR K632/633A , and an acetylation mimic mutant of the acetylation site AR K360Q . Immunohistochemical staining was conducted for either the FLAG epitope of the coexpressed androgen receptors or PSA. Coexpression of the acetylation mimic mutant increased PSA abundance within cells containing the activating mutant receptor. The addition of DHT further enhanced the induction of PSA in the presence of either the AR wt or AR K360Q . However, cells coexpressing the AR K630A or AR K632/633A demonstrated significantly reduced PSA immunostaining. Together these studies suggest the acetylation site mutant functions as a dominant negative inhibitor of androgen receptor signaling.

The AR Acetylation Site Mutation Affects the Isoelectric Focus of the AR in Both the Liganded and Unliganded State-The
ARwt and acetylation site mutants bound ligand with similar affinity by Scatchard analysis and gave similar gross structural analysis by trypsin digestion (17). The current studies suggested a role for the acetylation site in regulating ligand-induced covalent modifications (e.g. phosphorylation and acetylation) in the receptor. In previous studies, nondenaturing isoelectric focusing in ultra-thin polyacrylamide gels with subsequent two-dimensional electrophoresis was used to assess conformation of the liganded human AR (54). To assess the effect of mutation within the AR acetylation site on covalent of the ARwt (C) or AR mutants (D) was determined in DU145 cells using the MMTV-LUC reporter. Transfected cells were treated with either vehicle or forskolin (20 M) or PKA expression vector to activate cAMP signaling. The data are shown as relative luciferase activity (mean Ϯ S.E.) for at least six separate transfections. E, schematic representation of Gal4-CREB and Gal4-CREB serine 133 mutant. The heterologous reporter UAS 5 TATALUC containing five gal4 binding sites is also shown. F, Gal4-CREB or Gal4-CREB serine 133 mutant were transfected into DU145 cells with PKA wild type or PKA mutant expression plasmid. The induction of UAS 5 TATALUC activity is shown (mean Ϯ S.E.).

FIG. 3. Induction of AR activity by MAPK does not require the AR acetylation site.
A, the native c-fos promoter linked to a luciferase reporter gene was induced by coexpression of either MAPK kinase-activating mutant (⌴⌭⌲ R⌬F ) or MEKwt but not the kinase-dead mutant (K97M). B, PD98059 inhibition of ERK-dependent reporter activity was assessed using the AP-1-responsive reporter gene, p3TPLUX and co-expressed MAPK kinase-activating mutant (⌴⌭⌲ R⌬F ). C, the androgen-responsive reporter MMTV-LUC was used to assess AR-dependent MAPK signaling. Cells transfected with either ARwt or AR mutants with or without MAPK kinase-activating mutant (⌴⌭⌲ R⌬F ) were treated with DHT or vehicle for 24 h. D-F, the c-fos promoter luciferase reporter with point mutations in the sis-inducible element, TCF, or SRE site as indicated were assessed for induction by the MAPK kinase expression vectors. Data are mean Ϯ S.E. modification of the AR, the ARwt and AR K630A were expressed in HEK293 cells and treated with either vehicle or DHT for 24 h. The AR was immunoprecipitated with an hAR antibody and the immunoprecipitated AR was subjected to two-dimensional gel electrophoresis. In the absence of ligand, the ARwt migrated as a protein of ϳ110 kDa and focused as two groups of species, with three dominant protein spots of pH 4.5 and two major spots at pH 6 ( Fig. 6A). Additional minor spots spread down to around pH 5 (Fig. 6B) consistent with previous studies (54). In the presence of ligand, the ARwt focused at one major spot at pH 5.2 with spots of less protein distributed at pH 4.5, 4.8, 5.5, and 6 ( Fig. 6B). In the absence of ligand, the AR K630A mutant showed three spots around pH 4.5 with migration at about 110 kDa (Fig. 6C). Upon DHT treatment, the AR K630A mutant migrates as a major spot at pH 5.2 and a minor spot at pH 4.8. Spots within other pH range seen in the ARwt in the presence of ligand were not detectable (Fig. 6D) even with substantially longer exposures (data not shown). The experiments were conducted on multiple occasions with the same findings. These results suggest that mutation of the AR acetylation site affects its covalent modifications; leading to changes in the isoelectrophoretic profile of the AR. Substitution of alanine for lysine (AR K630A ) would be anticipated to reduce positive charge and P I (isoelectric point) as observed. The addition of ligand induced multiple ARwt species which were not observed with the AR K630A , suggesting several covalent modifications are dependent upon the lysine residue.
The AR Acetylation Site Regulates Formation of the 114-kDa Form of the AR-The AR is synthesized as a nonphosphorylated protein and migrates as a 110-kDa protein during SDS-PAGE. The AR becomes phosphorylated at serines within 15 min after synthesis and then migrates as a doublet of 110 -112 kDa. In response to ligand, a third (114 kDa) isoform is induced (25) and, in the presence of protein kinase A stimulators (cAMP or forskolin), the AR is rapidly dephosphorylated (26). To assess the electrophoretic properties of the ARwt and AR K630A mutant, HEK293T cells were transfected with the AR expres-sion vectors and cellular extracts assessed by SDS-PAGE. In DHT-treated cells, the three AR isoforms (110 -114 kDa) were observed (Fig. 7, A and B, a-*c). The incubation of extracts with calf intestinal alkaline phosphatase abolished the presence of 112-and 114-kDa bands (Fig. 7A, lane 2 versus lane 1). The ARwt and AR acetylation mutant were transfected into HEK293T cells, treated with DHT for 24 h, and subjected to Western blotting. In contrast with the ARwt, the AR K630A or AR K(632/633)A mutants did not display the 114-kDa isoform (Fig.  7B), suggesting ligand-induced phosphorylation requires the acetylation site. Multiple experiments of Western blotting of transfected cells demonstrated similar levels of expression of the ARwt and mutants receptors in the basal and the DHTtreated state (data not shown) as previously described (4,17). These studies suggested acetylation may be required for the AR phosphorylation and formation of the higher molecular weight forms.
The subcellular localization of the nuclear receptors is regulated by the ligand. The DNA-binding domain and the flanking hinge regions harbor a bi-or tripartite-type nuclear localization signal (NLS), consisting of two or three clusters of lysine and arginine residues (5,55). The AR contains a second NLS embedded in the ligand-binding domain, similar to ER␣ and progesterone receptor (55). The acetylation motif (KLKK) of the AR in the hinge region constitute the core of the first NLS (NL1) of the AR. To examine if the acetylation site mutant would affect the nuclear trafficking of the AR, the DNA constructs expressing either ARwt or AR K(632/633)A mutant were transfected into HEK293T cells. The cells were then treated with or without 100 nM DHT for 24 h. The cytoplasmic/nuclear cellular lysates of the cells were extracted and subjected to SDS-PAGE and Western blotting with anti-AR antibody. DHT treatment enhanced nuclear transfer of the ARwt. The AR K632/633A mutant was also detectable in the nuclear fractions after DHT treatment (Fig. 7C), consistent with previous observations that deletion of the AR acetylation motif (AR ⌬629 -633 ) results in delayed but none-the-less complete ligand-dependent The current studies suggest the AR acetylation site regulates both ligand-induced and H89-induced AR signaling. To examine further the evidence for cross-talk between acetylation and phosphorylation at the level of the androgen receptor, chroma-tin immunoprecipitation assays were conducted to assess whether PKA signaling may regulate androgen receptor occupancy at an ARE in the context of its local chromatin structure. The ARE of the PSA promoter was used for ChIP assays. HEK293T cells were transfected with either FLAG-tagged AR wt (Fig. 8A) or FLAG-tagged AR K630A (Fig. 8B). DHT induced recruitment of the androgen receptor to the ARE. Treatment of cells with TSA enhanced recruitment of the ARE to the PSA promoter, consistent with the induction of ARE signaling by the HDAC inhibitor TSA. Furthermore, the PKA inhibitor H89 induced recruitment of AR to the ARE. The AR acetylation site mutant demonstrated substantially reduced recruitment to the PSA promoter in the presence of DHT. The relative induction of the AR K630A mutant to the PSA promoter by TSA was also substantially reduced. H89 treatment of cells resulted in no detectable recruitment of the AR acetylation site mutant to the PSA promoter. Although the relative abundance of the AR wild type or AR acetylation site mutant in nuclear extracts in the presence of DHT was similar (Fig. 7C), the recruitment to an ARE in the context of local chromatin assessed using ChIP assays was substantially reduced. These studies suggested that protein kinase A-dependent regulation of AR recruitment to an ARE was abrogated by the acetylation site mutation.
The defective recruitment of the AR acetylation site mutant to an ARE in ChIP assays suggested evidence for cross-talk between phosphorylation and acetylation of the androgen receptor. To determine whether the AR phosphorylation in turn affected acetylation-dependent functions of the AR, a series of point mutation substitutions of the androgen receptor were assessed (23). Comparison was made among mutants of the androgen receptor at the phosphorylation sites where the phosphorylated serine residues were substituted with alanine (Fig.  9A). The androgen-responsive reporter gene PSA luciferase was examined, and a comparison was made between the androgen receptor wild type and the phosphorylation site mutants (Fig. 9B). Consistent with the previous publication (23), each of the phosphorylation site mutants induced androgen signaling ϳ5-fold. The S308A mutant enhanced DHT-dependent trans-activation significantly more than AR wild type as previously described in LNCaP cells (23).
To assess the role of HDAC-dependent signaling to the phosphorylation site mutants, the specific HDAC inhibitor TSA was FIG. 6. Isoelectric focusing and two-dimensional electrophoresis of liganded AR and the AR acetylation site mutants. A, isoelectric focusing and two-dimensional electrophoresis of cell extracts transfected with either AR wt (A and B) or AR K630A (C and D) and either treated with vehicle or DHT for 24 h as indicated. The extracts were immunoprecipitated with anti-AR antibody, and the immunoprecipitate was subjected to isoelectric focusing and two-dimensional electrophoresis. The AR immunoreactive band is circled. IgG(H), IgG heavy chain.
used. Regulation of AR signaling was selectively reduced by the S94A mutant. Induction by TSA was reduced ϳ50% (Fig. 9C). In previous studies of the p300-dependent trans-activation of AR signaling, the HAT domain was required, and the acetylation site mutants were defective in p300-dependent trans-activation (4). The phosphorylation site mutants were therefore examined for trans-activation by p300. The S94A mutant was significantly reduced in trans-activation by p300 (Fig. 9D). Western blot analysis was conducted of the AR phosphorylation site mutants, and normalization for protein transfer was conducted using the control GDI. The relative abundance of the phosphorylation site mutants was similar when normalized to the loading control, other than the S424A, which showed ϳ50% reduction in protein abundance in the presence of DHT. Together these studies demonstrate that the AR phosphorylation site residue Ser 94 plays an important role in activation by the p300 histone acetyltransferase and regulation by the histone deacetylase inhibitor TSA. DISCUSSION The current studies demonstrate the conserved AR acetylation site plays an important role in a subset of kinase signaling pathways previously shown to regulate AR activity. AR acetylation-defective mutants showed reduced regulation by AKT, PKA, and JNK, whereas MAPK and sumoylation of the AR were unaffected by point mutation within this acetylation site. Together with previous studies demonstrating this site does not affect AR trans-repression of NFB or SP-1 activity (11,18), the current findings suggest post-translational modification by acetylation may coordinate a distinct subset of AR functions.
Growing evidence suggests a model in which acetylation and phosphorylation of histones contribute to a local signaling module (31, 56 -58). Post-translational modification within the histone tails in turn contribute, through changed charge, to recruitment of other histone acetyltransferase or chromatin modeling proteins. In view of the recent studies that AR acetylation mutants substituting lysine for arginine or alanine are defective in binding the histone acetyltransferase p300 and show enhanced recruitment of nuclear N-CoR-Smad-HDAC-1 complexes (17), the current studies suggest transcription factor acetylation sites may also coordinate recruitment of HAT, and serve as signaling modules, including phosphorylation.
The AR point mutants substituting lysine for alanine within the acetylation motif were defective in activation by coactivators (p300, SRC-1a, Ubc-9, and TIP60) and bound relatively more N-CoR suggesting a critical role for direct receptor acetylation in corepressor and coactivator engagement (17). How might acetylation within the AR hinge region affect H-89-dependent activation of the AR? The AR acetylation site does not fall within the ligand-binding pocket deduced from the crystal structure (59), consistent with the wild-type ligand binding of the AR acetylation site mutants. Expression levels of the ARwt, the AR K630A , or AR K(632/633)A mutant proteins were also similar in cultured cells (4,17). DHT induces AR phosphorylation (24) and forskolin-induced dephosphorylation of the AR in prostate cancer cells may involve PKA-inducible phosphatases such as nuclear protein phosphatase-1 (26). Analysis of Ser 641 , one of the two serine residues dephosphorylated in response to cAMP (serines 641 and 653) (26), demonstrated the importance of Ser 641 to the overall phosphorylation of the AR. Ser 641 is located in the AR hinge region in close proximity to the KXKK motif. One interpretation of the finding that the AR acetylation site mutants failed herein to form the 114-kDa isoform and were resistant to PKA repression is that ligand-induced coactivator recruitment, dependent upon the AR lysine residues, precedes phosphorylation of the AR. Because DHT-induced accumulation of AR in the nucleus occurs rapidly, within minutes, and appears to precede receptor phosphorylation, it has been proposed that AR phosphorylation may play a role in late FIG. 9. Regulation of AR phosphorylation site mutants by the histone deacetylase inhibitor TSA and the histone acetyl transferase p300. A, schematic representation of the AR phosphorylation site. B-D, the expression vectors encoding the AR phosphorylation site mutants were expressed in HEK293T cells and analyzed for regulation by DHT, TSA, or p300 trans-activation. The data are mean Ϯ S.E. for n Ͼ 6 separate transfections. E and F, Western blot analysis of HEK293T cells transfected with the expression vector of the phosphorylation site mutants was conducted with antibodies to the AR. GDI is control for protein abundance and transfers. phases of transcriptional regulation or receptor recycling (23). The failure of the liganded AR K630A mutant to form several species observed with the ARwt by two-dimensional electrophoresis is also consistent with the role of the lysine residues in subsequent ligand-induced covalent modifications.
In the current studies, chromatin immunoprecipitation assays demonstrated ligand-induced recruitment of the AR wild type to the androgen response element of the PSA promoter in response to the ligand DHT, the histone deacetylase inhibitor TSA, or the protein kinase inhibitor H89. The recruitment of the AR acetylation site mutant was dramatically reduced despite similar levels of mutant receptor in nuclear extracts. The defective recruitment of the AR acetylation site mutant is consistent with recent studies of p53 acetylation (60). These findings are also consistent with previous studies in which DHT recruited AR to the PSA promoter. TSA treatment hyperacetylated histone H3 at lysine 9 (34) and hyperacetylation of H3 lysine 9 is thought to correlate with transcriptional activation of target genes by altering chromatin structure and enhancing transcription factor recruitment to target DNA sequences. The PKA inhibitor H89 enhanced AR recruitment to an ARE, an effect abrogated upon mutation of the AR acetylation site lysine motif. How might PKA affect AR recruitment? PKA is known to induce phosphorylation of histone H3 at lysine 10 (29), which may in turn alter local chromatin structure to enhance transcription factor access. Such a possibility is supported by the finding that phosphorylation of histone H3 increases the susceptibility of histone H3 to HATs or HDAC inhibitor to induce histone H3 acetylation (29,43). Alternatively, recent evidence for cross-talk between PKA signaling and histone hyperacetylation is a finding that PP1 physically associates with HDAC, contributing to regulation of CREB activity (61). The possibility that AR-bound HDAC could recruit PP1 to regulate AR signaling remains to be determined.
Additional evidence for interdependence of phosphorylation and acetylation was derived from analysis of AR phosphorylation site mutants. TSA induction of AR signaling was maintained for each of the AR phosphorylation site mutants except substitution of serine 94. Similarly, induction of AR signaling by the histone acetyltransferase p300 was selectively reduced by this mutation of the phosphorylation site. Further mechanistic insight into the relationship between AR acetylation and phosphorylation will require a better understanding of how AR phosphorylation regulates the diverse functions of the AR.
The AR acetylation mutants AR K630A or AR K(632/633)A did not form the 114-kDa phosphorylated isoform observed with the ARwt in the presence of ligand. Several recent observations suggest acetylation and phosphorylation of histones may function as integrated post-translational modifications and in some circumstances can cooperate to activate transcription (29,43,62). Histone H3 phosphorylation at Ser 10 enhances acetylation at Lys 14 (43). The functional cooperation between phosphorylation and acetylation may involve a phosphate-dependent stabilization of the enzyme substrate complex. Thus the histone acetyltransferase Gcn5 displays a 10-fold greater preference for phosphorylated H3 over nonphosphorylated H3 (43). Cooperation between phosphorylation-acetylation cascades of transcription factors had been shown for p53 function in response to DNA damage (31,63). Although in vitro studies have shown phosphorylation of Positive Coactivator 4 by caseine kinase II inhibits acetylation by p300 (64) and an inverse correlation between phosphorylation and acetylation of the forkhead transcription factor has been observed (65), further analysis of these proteins will require identification of specific residues and mutational analyses.
In the current studies, mutation of the AR lysine residues did not affect ERK signaling to the AR. These findings are consistent with analyses of histone H3 lysine acetylation, which is also independent of MAPK (34). Although several studies have implicated direct transcription factor acetylation in regulating reporter gene expression, only recently has the functional cellular phenotype governed by transcription factor acetylation been examined (66). The AR is overexpressed or promiscuously activated in human prostate cancer, and the effectiveness of androgen ablation therapy in reducing prostate cancer cellular growth suggests a key role for the liganded AR in aberrant prostate cellular growth (67). AR acetylation site lysine residue substitutions with glutamine to mimic charge changes induced by lysine acetylation promote contact-independent growth of prostate cancer cell lines in soft agar and in nude mice and resistance to flutamide. Furthermore, AR acetylation mimic mutants selectively enhance activity of the cell-cycle control genes cyclin D1 and cyclin E (18). Acetylation site mutations of p53 were defective in repression of Ras-induced transformation, suggesting an important functional role for p53 acetylation in vivo. (66). The identification of acetylation as a posttranslational modification required for regulation of the AR by a distinct subset of signaling pathways (AKT, PKA, and JNK) and not MAPK or sumoylation suggests these residues contribute to a new type of signaling specificity.
Post-translational modification of histones contributes to a signaling "code" and forms a platform for recruitment of distinct proteins. Lysine residues of histones, which can be either acetylated or methylated, result in a mutually exclusive modification. The in vitro binding of TAF II 250 to multiply-acetylated H-4 tails (68) suggests pairs of acetylated residues with particular spacing recruit specific bromo-domain containing proteins. In the same manner that combinatorial modifications, including phosphorylation and acetylation, form signaling platforms recruiting proteins to histones, these modifications appear to encode signaling platforms for nuclear proteins. The recruitment of candidate proteins to lysine residues of nuclear receptors may result in steric hindrance and thereby exclude the binding of other kinases that regulate AR signaling.