p53 Represses Androgen-induced Transactivation of Prostate-specific Antigen by Disrupting hAR Amino- to Carboxyl-terminal Interaction*

Prostate-specific antigen (PSA) is highly overexpressed in prostate cancer. One important regulator ofPSA expression is the androgen receptor (AR), the nuclear receptor that mediates the biological actions of androgens. AR is able to up-regulate PSA expression by directly binding and activating the promoter of this gene. We provide evidence here that that this AR activity is repressed by the tumor suppressor protein p53. p53 appears to exert its inhibition of human AR (hAR) by disrupting its amino- to carboxyl-terminal (N-to-C) interaction, which is thought to be responsible for the homodimerization of this receptor. Consistent with this, p53 is also able to block hAR DNA binding in vitro. Our previous data have shown that c-Jun can mediate hAR transactivation, and this appears to result from a positive effect on hAR N-to-C interaction and DNA binding. Interestingly, c-Jun is able to relieve the negative effects of p53 on hAR transactivation, N-to-C interaction, and DNA binding, demonstrating antagonistic activities of these two proteins. Importantly, a p53 mutation found in metastatic prostate cancer severely disrupts the p53 negative activity on hAR, suggesting that the inability of p53 mutants to down-regulate hAR is, in part, responsible for the metastatic phenotype.

Transcriptional regulation by nuclear receptors is critical to vertebrate development. This is well exemplified by the role of the androgen receptor (AR) 1 in male sexual development, which is disrupted in males with androgen insensitivity syndrome (for review, see Ref. 1). Such disruptions result in males who are unable to respond to androgens because of mutations in the AR and who therefore can develop, in the most severe case, the sexual phenotype of a female. In keeping with this, development of the prostate, a male-specific gland, is dependent upon the activity of AR and its cognate androgen ligands, testosterone and dihydroxytestosterone (DHT) (2)(3)(4). Testosterone and DHT act by specifically binding to the AR and activating it such that it can up-regulate the transcription of specific responsive genes. This paradigm of receptor activation is a defining characteristic of nuclear receptors (for reviews, see Refs. [5][6][7][8][9]. Nuclear receptors regulate transcription by binding, as either homodimers or heterodimers, to DNA in a sequencespecific manner. While the receptors for retinoids, thyroids, and vitamin D bind to DNA as heterodimers with the retinoid X receptors, the AR and receptors for glucocorticoids, progestins, and estrogens carry out the same activity as homodimers (for reviews, see Refs. 5 and 7). Interestingly AR and estrogen receptor ␣ have been suggested to form antiparallel homodimers based on the demonstration of a high affinity interaction between the amino and carboxyl termini of the receptor (10 -13). With respect to AR, the physiological significance of this N-to-C interaction is suggested by the recent finding that several mutations responsible for androgen insensitivity syndrome disrupt this interaction without affecting ligand binding (12). Moreover we have recently demonstrated by mutational analysis that disruption of the N-to-C interaction blocks hAR DNA binding in vitro. 2 Hence it may be concluded that hAR N-to-C interaction is critical for the ability of this receptor to up-regulate the transcription of androgenresponsive genes.
One clinically important androgen-responsive gene is prostate-specific antigen (PSA), which encodes a serine protease that is a diagnostic marker for prostate (for reviews, see Refs. 14 and 15). The androgen regulation of PSA gene expression is mediated by the presence of three androgen-responsive elements within the 5.8-kilobase pair PSA promoter, all of which have been shown to bind AR (16 -18). Two of these androgenresponsive elements are located within the proximal region of the promoter at Ϫ170 and Ϫ395 (17,18), and the third is far more distal at Ϫ4148 (16). Interestingly, of the two proximal elements, the more distally located element (Ϫ395) has the unique activity of functioning as an androgen-responsive element only in the presence of the more proximal (Ϫ170) element and thus is called an androgen-responsive region (17). In addition to these androgen-regulated elements, PSA promoter activity has been shown to be regulated by other types of elements, including GATA (19) and elements for PDEF, a novel ets transcription factor (20). Both the GATA-binding proteins (GATA-2 and GATA-3) and prostate-derived Ets factor have been reported to modulate androgen induction of the PSA pro-moter (19,20). Thus, it is clear that the complex nature of the PSA promoter renders AR activation of this promoter regulatable by other DNA-binding proteins.
p53 is a DNA-binding protein that has been shown to inhibit the production of PSA when it is expressed in tumors in nude mice (21)(22)(23). This negative effect on PSA production is consistent with the known negative activities of p53 on transcription that have been seen on many cellular and viral promoters (for review, see Ref. 24). These activities do not depend on sequence-specific DNA binding by p53 since they have been observed on promoters that do not contain p53-responsive elements (25)(26)(27)(28). However, this negative transcriptional activity by p53 does require its carboxyl-terminal 30 amino acids, which have been identified to be involved in the interaction of p53 with the TATA box-binding protein (TBP) (29). This finding suggests that the TBP interaction with p53 through the carboxyl terminus is inhibitory to the transcription process. In addition, p53 is known to repress the transcription process when bound to DNA demonstrated by fusing p53 to heterologous DNA-binding domains (DBDs) (30). This same approach has been used recently to identify an eight-amino acid sequence in the carboxyl terminus of p53 that is essential for transcriptional repression (31). In contrast to the negative transcriptional activities of p53 that can occur on or off DNA, the positive transcriptional activities of this protein have only been observed on DNA (for review, see Ref. 24). Because DNA binding is essential, the transactivation function of p53 relies on the central portion of the protein. This portion contains a zinc finger-like DBD that is responsible for sequence-specific DNA binding (32). In addition, the amino-terminal region (amino acids 1-42) of p53, which serves as the binding site for the p53 repressors E1B or MDM2 (33), harbors a transcriptional activation domain (29, 34 -36). Interestingly this amino-terminal region serves as a second site for interaction with TBP (and TBP-associated factors) (29,35,37), but this interaction results in transcriptional activation in contrast to the repression that results from the TBP interaction with the carboxyl terminus of p53.
Mutations of the p53 gene are associated with many cancers, including prostate cancer (for reviews, see Refs. 24 and 38). In view of the known p53 negative effects on transcription, the finding that expression of p53 in prostate tumors reduces production of PSA (21)(22)(23) opens the possibility that p53 regulates PSA promoter activity. Here we report that p53 is capable of negatively regulating AR-induced transactivation of the PSA promoter. This activity appears to be mediated by the ability of p53 to disrupt hAR N-to-C interaction and in vitro DNA binding.
The reporter plasmids have the gene for either luciferase (LUC) or chloramphenicol acetyltransferase (CAT) driven by different promoters. The AR-inducible reporter plasmid PSA-LUC was constructed by polymerase chain reaction amplification of the PSA promoter (43) using the upstream oligonucleotide 5Ј-GATCGGTACCCATTGTTTGCTGCA-C-3Ј and the downstream oligonucleotide 5Ј-GATCCCCGGGTCCGGG-TGCACCTCC-3Ј and PSA-630-CAT (a generous gift from Dr. Charles Young) as template. This resulted in the amplification of the PSA promoter from Ϫ624 to ϩ12. Polymerase chain reaction fragments were digested and ligated into the KpnI/SmaI site of pGL3 (Promega). The p53-inducible p50-2-CAT reporter (a generous gift from Dr. Thomas Shenk) was described previously (30). For the mammalian two-hybrid system, the previously described reporter 17 M-tk-CAT (44) was used. Transfection efficiency was standardized using either pRL-tk-LUC (Promega) for the luciferase assay or pCH110 for the CAT assay.
Cell Culture and Transfection-PC-3 cells were maintained in Dulbecco's modified Eagle's medium (Sigma) supplemented with 10% fetal bovine serum (Hyclone). Cells were plated at about 60% confluency in 60-mm dishes and transiently transfected using the calcium phosphate precipitation method (44). Cells were washed with phosphate-buffered saline 24 h after the introduction of DNA and subjected to fresh Dulbecco's modified Eagle's medium ϩ 10% fetal bovine serum containing 100 nM hormone or ethanol carrier. Cells were then harvested 24 -48 h postwash in passive lysis buffer (Promega) for the luciferase assay or ␤-galactosidase buffer (44) for the CAT assay.
Luciferase and CAT assays-Luciferase activity was assayed using the Dual Luciferase Assay kit from Promega. 100 l of cell extract were assayed for firefly luciferase activity using a TD 20/20 luminometer (Turner Designs). Transfection efficiency was normalized by Renilla luciferase activity according to the instructions of the manufacturer. CAT assays were standardized according to ␤-galactosidase activity and performed as described previously (44). All transfections were performed at least three times, and the results presented for luciferase and CAT assays are the mean values of three independent experiments. Standard deviations are indicated as error bars.
Gel Mobility Shift Assay-COS cells were grown in 100-mm dishes and transfected using LipofectAMINE 2000 (according to instructions from Invitrogen). Cells were treated with 100 nM DHT 24 h prior to harvesting. Cells were harvested in ice-cold phosphate-buffered saline and centrifuged at 5000 rpm for 5 min. 10% of the cells were used to do a ␤-galactosidase assay for quantification of transfection efficiency. The remainder of the cells were resuspended in Buffer I (10 mM Tris-HCl, pH 7.5; 10 mM NaCl; 5 mM MgCl 2 ) and incubated at 4°C for 5 min. 0.3 M sucrose was then added, and cells were lysed with a Dounce homogenizer. Nuclei were pelleted by centrifuging lysed cells at 2500 rpm (600 ϫ g) for 10 min. The nuclear pellet was washed once with Buffer II (Buffer I containing 0.3 M sucrose). Then the nuclear pellet was resuspended in Buffer III (50 mM Tris-HCl, pH 8; 150 mM NaCl; 5 mM EDTA; 0.1% Nonidet P-40) with protease inhibitors and incubated with shaking at 4°C for 30 min. The lysed nuclei were centrifuged at 15,000 rpm for 15 min, and the supernatant, constituting the nuclear extract, was saved. The amount of extract used was standardized according to ␤-galactosidase activity.
Gel mobility shift assays were performed with nuclear extracts containing the same amount of ␤-galactosidase activity. These reactions were performed in a final volume of 20 l in DNA binding buffer (10 mM Tris, pH 8; 0.1 mM EDTA; 4 mM dithiothreitol), which also containing 1 g of poly(dI-dC), 100 mM KCl, and 100,000 cpm 32 P-labeled probe (5Ј-GATCCAAAGTCAGAACACAGTGTTCTGATCAAAGA-3Ј), an androgen-responsive element. After the addition of an equal number of ␤-galactosidase units of nuclear extract, the reactions were gently vortexed and incubated for 15 min at 25°C. The samples were run on a 6% polyacrylamide gel for 1.5 h at room temperature after which the gel was dried and exposed to autoradiography. To do antibody supershifts, 1 l of either the anti-hAR antibody PA1-111A or the anti-birch profilin antibody 4A6 (a kind gift from Dr. Martin Rothkegel) was added prior to addition of probe.

SDS-Polyacrylamide Gel Electrophoresis and Western Blot
Analysis-Cell extracts were prepared in two ways for Western blot analysis. For PC-3 cells, the pellets from transfected cells, remaining after lysis and extraction for luciferase and CAT assays, were boiled for 5-10 min in SDS Sample Buffer (63 mM Tris, pH 6.8; 20% glycerol; 2% SDS; 5% ␤-mercaptoethanol). The amount of boiled pellet used was standardized according to internal control activity (Renilla luciferase activity for luciferase and ␤-galactosidase for CAT assays). For transfected COS cells, the nuclear extracts prepared for gel mobility shift assays were also used in Western blot analysis. Proteins were separated by SDSpolyacrylamide gel electrophoresis and were electrophoretically transferred onto nitrocellulose blots (Micron Separations Inc.). The nitrocellulose blots were blocked with nonfat dry milk and subsequently probed with either the anti-p53 antibody 13-4000 (Zymed Laboratories Inc. Laboratories) or the anti-hAR antibody PA1-111A (Affinity BioRe-agents) and developed using the enhanced chemiluminescence (ECL) kit from Amersham Pharmacia Biotech.

RESULTS
p53 Negatively Regulates hAR-induced Transactivation of the PSA Promoter-We have previously observed the c-Jun positive effect on hAR transactivation in COS, HeLa, NIH3T3, P19, and F9 cells (44 -48). Since this receptor is not known to have a physiological role in the tissues from which these cells were derived, we elected to use cells originating from the prostate, whose development is tightly regulated by androgens and AR. These are PC-3 cells, and they were derived from bone metastasis of prostate cancer (49). Although PC-3 cells have lost expression of AR (50), they have been used extensively in transfection studies to study AR-mediated transcription (for reviews, see Refs. 51 and 52). Therefore, we used these same cells in this study to determine whether p53 modulates hARinduced transactivation of the PSA promoter. This was done by transiently transfecting PC-3 cells with the androgen-responsive reporter PSA-624-LUC and expression plasmids for hAR and p53, both of which are not expressed in these prostate cells (50,53). Transfection of hAR in PC-3 cells resulted in DHT-dependent activation of the PSA promoter, which did not occur in the absence of transfected hAR (Fig. 1), consistent with the lack of endogenous hAR. Co-expression of p53 completely blocked this ligand-dependent hAR activity without significantly affecting the basal activity of the promoter in the absence of transfected hAR (Fig. 1). These results demonstrate that p53 is capable of repressing hAR transactivation of the PSA promoter.
Like other transcription factors, p53 consists of multiple functional domains required for the multitude of cellular functions that this protein has (for review, see Ref. 24). Previous data show that regions important for p53-mediated transcriptional repression include the carboxyl terminus (29) and, on some TATA box-containing promoters, the central region encoding the DBD (32). To determine whether either of these regions is important for hAR inhibition, two p53 mutants were studied for repression of hAR transactivation. Mutant p53W248, containing an Arg to Trp change at position 248, is deficient in specific DNA binding (41), and p53⌬C30 has a deletion of the last 30 amino acids that results in the loss of interaction with the TBP (40). While both mutant proteins are able to repress hAR transactivation of the PSA promoter, they are significantly weaker than wild-type p53 ( Fig. 2A). This difference in activity is not due to different levels of protein expression of the mutants since Western blot analysis demonstrated that the mutant p53 proteins are slightly more highly expressed than wild-type p53 (Fig. 2B). Moreover the expression of transfected hAR is not affected by any of the p53 proteins (Fig. 2C), clearly showing that inhibition of hAR transcriptional activity is not due to reduced hAR protein levels. These data show that the p53 DBD and carboxyl terminus are important for its repression of hAR transactivation.
p53 Blocks hAR Amino-to Carboxyl-terminal Interaction and DNA Binding-Previous work has demonstrated an in vivo interaction between the AB and E regions of AR (10,11) and that this interaction is stimulated by the coactivators cAMPresponse element-binding protein (CREB)-binding protein (54), steroid receptor coactivator-1 (SRC-1) (54), and transcription intermediary factor-2 (TIF2) (55). We have recently discovered that c-Jun can have the same enhancing effect on this hAR N-to-C interaction. 2 Moreover some mutations within the hAR ligand-binding domain associated with androgen insensitivity syndrome have been shown to disrupt this interaction (12). To test the possibility that p53 may be acting on this function of FIG. 1. p53 represses hAR transactivation of the PSA promoter. PC-3 cells were transfected with 5 g of PSA-LUC reporter plasmid, 2.5 g of hAR, and 0.5 or 1 g of p53 as indicated. 100 nM DHT was used as indicated. Note that luciferase activity is represented relative to activity of the third condition, which was transfection of hAR without DHT, and this was set to 1.
FIG. 2. Mutants of p53 are weakly able to repress hAR transactivation of the PSA promoter. A, PC-3 cells were transfected with 5 g of PSA-LUC reporter plasmid, 2.5 g of hAR, and 1 g of different p53 expression plasmids as indicated. 100 nM DHT was used in all cases. Note that luciferase activity is represented relative to activity of the first condition, which was transfection of only the reporter plasmid, and this was set to 1. B and C, expression of p53 proteins in transfected PC-3 cells. Cell pellets from the transfection described in A were subjected to Western blot analysis using either an anti-p53 antibody (B) or anti-hAR anti-body (C). Note that neither mutation in the p53 protein significantly alters its mobility in SDS-polyacrylamide gel electrophoresis. WT, wild type. the hAR, we established a mammalian two-hybrid system using the expression plasmids GAL-hAR(DE) and VP16-hAR(AB) (Fig. 3A) and the reporter plasmid 17 M-tk-CAT in transient transfection experiments in PC-3 cells. As shown by us (47) and others (56 -59), GAL-hAR(DE) exhibited little transcriptional activity either in the absence or presence of DHT (Fig. 3B). However, upon co-transfection of VP16-hAR(AB), GAL-hAR(DE) exhibited a DHT-dependent transactivation of the 17 M-tk-CAT (Fig. 3B). These results indicate that the hAR undergoes an N-to-C interaction in the presence of ligand as demonstrated previously (10,11). Interestingly p53 is able to repress the hAR N-to-C interaction in a dose-dependent manner (Fig. 3C). Note that p53 had no effect on the activity of GAL-VP16, 3 excluding a possible p53 interference of either GAL(DBD) or VP16 function. To analyze the specificity of this p53 effect, we utilized the two repression-deficient mutants of p53. Importantly neither p53⌬C30 nor p53W248 was able to significantly block hAR N-to-C interaction (Fig. 3C), consistent with their inability to affect hAR transactivation. Thus, these results suggest that p53 blocks hAR transcriptional activity by interfering with the N-to-C interaction of this receptor.
If this N-to-C interaction is required for hAR dimerization and subsequent DNA binding, then inhibition of this interaction should lead to reduced hAR DNA binding. To analyze this, we measured the in vitro DNA binding ability of hAR transiently expressed in COS cells in either the absence or presence of transfected p53 proteins (Fig. 4A). Extracts from cells transfected with hAR exhibited significant DNA binding (Fig. 4A,  compare lanes 1 and 2). The presence of hAR was confirmed by a supershift with an anti-hAR antibody (Fig. 4A, lane 3), but no supershift was seen with a heterologous antibody (Fig. 4A, lane  4). Importantly cell extract containing transfected hAR and p53 exhibited strongly reduced hAR DNA binding (Fig. 4A, lane 5). Expression of another unrelated protein, R-cadherin, had no effect on hAR DNA binding (Fig. 4A, lane 6), demonstrating that the negative effect on hAR DNA binding is p53-specific. Importantly the p53 inhibition of hAR DNA binding is dose-dependent since transfecting higher p53 levels led to a greater reduction in hAR DNA binding (Fig. 4A, compare lane 8 to  lanes 9 -11). To further examine the specificity of this p53 effect 3 C. J. Fisher and L. Shemshedini, unpublished data.

FIG. 3. Wild-type p53, but not the mutants, blocks the N-to-C interaction of hAR.
A, schematic diagram of hAR fusion proteins used in the mammalian two-hybrid system to measure the hAR N-to-C interaction. hAR(DE) is encoded by amino acids 624 -918, and hAR(AB) is encoded by amino acids 1-555. Note that the two black bars in the AB region and the black bar in the E region represent AF-1 and AF-2, respectively. B, the mammalian two-hybrid system can be used to measure the in vivo interaction of hAR AB and DE regions. PC-3 cells were transfected with 1 g of 17 M-tk-CAT reporter plasmid, 5 g of GAL-hAR(DE), and 3 g of VP16-hAR(AB) as indicated. 100 nM DHT was used as indicated. Note that CAT activity is represented relative to activity of the last condition, which was transfection of both GAL-hAR(DE) and VP16-hAR(AB) in the presence of DHT, and this was set to 1. C, p53 represses hAR N-to-C interaction. PC-3 cells were transfected as in B with 1, 3, and 5 g of p53 (wild type (WT) and mutants) as indicated and treated with 100 nM DHT in all cases. Note that CAT activity is represented relative to activity of the first condition, which was transfection of both GAL-hAR(DE) and VP16-hAR(AB) in the presence of DHT, and this was set to 1.

FIG. 4. Wild-type p53, but not mutants, blocks hAR DNA binding.
A, COS cells were transfected with 8 g of hAR and with or without 1, 5, or 8 g of wild-type (WT) p53 or 8 g of p53 mutants or R-cadherin (R-cad) and treated with 100 nM DHT. Nuclear extracts were tested for hAR DNA binding using a gel mobility shift assay. Antibody supershift analysis was done with either an anti-hAR or anti-birch antibody as a control. Note that S is specific binding, SS is supershift, NS is nonspecific binding, and FP is free probe. B, the same nuclear extracts were subjected to Western blot analysis using an anti-hAR antibody. on hAR DNA binding and correlate it to the effects on transactivation and N-to-C interaction, the two p53 mutant proteins were used. Both mutants p53⌬C30 and p53W248 were unable to block hAR DNA binding (Fig. 4A, compare lane 14 to lanes 16  and 17). Note that the level of hAR expression was not altered by the expression of p53 proteins (Fig. 4B, compare lane 2 to lanes 3 and 4, lane 6 to lanes 7-9, or lane 12 to lanes [13][14][15]. These results are consistent with the p53 activity on hAR N-to-C interaction and argue that p53 inhibits hAR transactivation by blocking receptor DNA binding through a negative effect on receptor homodimerization. c-Jun Relieves p53 Repression of hAR Activity-We have previously demonstrated that c-Jun can enhance hAR transactivation of MMTV (44,48) and hKLK2 (46) promoters, N-to-C interaction, 2 and DNA binding. 2 Since all of these c-Jun effects are opposite to the p53 effects reported here, we wanted to determine whether c-Jun can relieve these p53 negative effects on hAR activity. The first activity studied was hAR transactivation of PSA-LUC. As shown in Fig. 5A, c-Jun is able to both enhance hAR transactivation of the PSA promoter and relieve the p53 inhibition of this activity. In the same way, c-Jun can alleviate the p53 negative activity on hAR N-to-C interaction (Fig. 5B). Finally we examined the c-Jun effect on p53 repression of hAR DNA binding. As in experiments of Fig. 4, extracts from transfected COS cells were used in gel mobility shift studies measuring hAR DNA binding (Fig. 5C). Once again p53 interfered with hAR DNA binding (Fig. 5C, compare lanes 2  and 6), and importantly this interference was significantly relieved by co-transfected c-Jun (Fig. 5C, compare lanes 6 and  7). Note that c-Jun can weakly but significantly enhance hAR DNA binding in the absence of transfected p53 (Fig. 5C, com-pare lanes 2 and 5), consistent with what has been observed previously. 2 These results together strongly suggest that c-Jun and p53 have antagonistic effects on hAR and that both proteins target the N-to-C interaction of this receptor and therefore influence subsequent DNA binding and transcriptional activation.
The p53-hAR Interaction Is Not Mutually Inhibitory-Several nuclear receptors and AP-1 are known to have mutually inhibitory activity (for review, see Ref. 60). In addition, glucocorticoid receptor and p53 have recently been shown to block the transactivation functions of each other (61,62). To determine whether hAR and p53 engage in a similar reciprocal interaction, p53 transactivation was measured using the p53inducible reporter p50-2-CAT (30) in transient transfection experiments in PC-3 cells. While the reporter plasmid was strongly activated by p53, co-transfected hAR had no effect on this activity either in the absence or presence of DHT (Fig. 6). Thus, unlike the typical nuclear receptor-AP-1 interaction, which is mutually inhibitory, the hAR-p53 interaction acts in only one direction and results in down-regulation of hAR-dependent transcription. DISCUSSION The importance of hAR in prostate cancer is strongly suggested by several observations. First, prostate cancers are largely dependent on androgens for growth and survival (for reviews, see Refs. 51 and 52). Second, mutations have been identified within the hAR protein in late-stage disease that can affect the activity of the receptor (63). Third, amplification of the hAR gene has been observed in about 30% of prostate tumors recurring during androgen-ablation therapy (64). Finally, researchers have discovered an inverse correlation between the length of the polyglutamine repeat in the hAR amino terminus and the onset and severity of prostate cancer (65-67). Interestingly gene amplification and a shorter polyglutamine FIG. 5. c-Jun relieves the negative effects of p53 on hAR activity. A, c-Jun relieves p53 inhibition of hAR transactivation of PSA-CAT. PC-3 cells were transfected with 5 g of PSA-CAT reporter plasmid, 2.5 g of hAR, 3 g of p53, and 1 or 5 g of c-Jun as indicated. 100 nM DHT was used in all cases. Note that CAT activity is represented relative to activity of the first condition, which was transfection of hAR in the presence of DHT, and this was set to 1. B, c-Jun relieves p53 inhibition of hAR N-to-C interaction. PC-3 cells were transfected with 1 g of 17 M-tk-CAT reporter plasmid, 5 g of GAL-hAR(DE), 3 g of VP16-hAR(AB), 5 g of p53, and 1 or 5 g of c-Jun as indicated. 100 nM DHT was used in all cases. Note that CAT activity is represented relative to activity of the first condition, which was transfection of GAL-hAR(DE) and VP16-hAR(AB) in the presence of DHT, and this was set to 1. C, c-Jun relieves p53 inhibition of hAR DNA binding. COS cells were transfected with 5 g of hAR, 5 g of p53, and 5 g of c-Jun as indicated and treated with 100 nM DHT. Nuclear extracts were tested for hAR DNA binding using a gel mobility shift assay. Antibody supershift analysis was done with either an anti-hAR or anti-birch antibody as a control. Note that S is specific binding, SS is supershift, NS is nonspecific binding, and FP is free probe.
FIG. 6. hAR does not affect p53-dependent transcription. PC-3 cells were transfected with 5 g of p50-2-CAT reporter plasmid, 2.5 g of p53, and 1 or 3 g of hAR as indicated. 100 nM DHT was used as indicated. Luciferase activity is represented relative to activity of the first condition, which was transfection of only reporter plasmid, and this was set to 1.
p53 Inhibits AR Transactivation of PSA Promoter repeat length both lead to enhanced hAR transcriptional activity, suggesting that a higher transcriptional activity of this receptor correlates with a higher risk and higher grade of prostate cancer (for review, see Ref. 51). In view of these observations, it can be reasoned that accessory factors that affect hAR transcriptional activity may influence the onset of prostate cancer disease. For the AR as well as for other nuclear receptors, these accessory factors can be either positive or negative. Those that have been shown to act as positive factors on AR include the p160 family of coactivators (55,68), CREBbinding protein (69,70), p300 (71), ARA 70 (72), cyclin E (73), BRCA-1 (74), TRAM-1 (75), Tip60 (76), and c-Jun (44,45,47,48). In contrast, relatively few proteins have been identified and demonstrated to have negative activity on AR-mediated transcription. We have previously demonstrated that the protooncoprotein c-Fos is able to block hAR activity by interacting with c-Jun and presumably blocking its coactivation activity (46). More recently a yeast two-hybrid system was used to identify a novel AR-interacting protein, HBO1, which was shown to specifically repress AR-dependent transcription (77). In this article we provide evidence for a novel repressor of hAR activity. This repressor is the well characterized tumor suppressor protein p53, whose gene is known to be mutated in over half of human cancers and thus represents the most common genetic alteration found in clinical tumors (for review, see Ref. 24). In prostate cancer, p53 mutations are found in about 23% of metastatic tumors (for review, see Ref. 38).
p53 is known to have both positive and negative activities on transcription. The negative activities identified thus far require the carboxyl terminus of p53 and may or may not depend on specific p53 DNA binding (for review, see Ref. 24). Our data show that both the carboxyl-terminal region (last 30 amino acids) and DBD of p53 are important in the inhibition of hAR transactivation of the PSA promoter since either deletion of the former (found in mutant p53⌬C30) or disruption of the latter (found in mutant p53W248) severely compromises the negative activity of p53. The finding with the carboxyl-terminal region is not surprising in view of previous studies reporting that this region harbors a potent transcriptional repression domain (30,31). Since this domain has been shown to physically interact with the general transcription factor TBP (29), this interaction may play a role in p53-mediated repression of hAR. On the other hand, it is possible that deletion of the carboxyl-terminal region is such a dramatic change that it alters the conformation and, thus, the activity of the p53 protein. Evidence for this has been reported by several studies in which sequence-specific DNA binding by p53 is activated by deletion or disruption of the carboxyl-terminal region (78,79). Likewise it is possible that p53⌬C30 is defective in hAR repression because of a conformational change that influences another region of the protein that may be necessary for this negative activity. One possible region is an eight-amino acid sequence (residues 339 -346) close to the p53 carboxyl end that was recently reported to represent a minimal transcriptional repression domain (31). While the importance of these eight amino acids in p53 repression of hAR is not known, their activity may be influenced by the carboxylterminal deletion through a conformational effect. Unlike the gross change in p53⌬C30, the mutant p53W248 has a single amino acid change that likely results in a more subtle conformational change. Nevertheless this mutation is sufficient to disrupt both p53 sequence-specific DNA binding (41) and, as is shown here, repression of hAR. These results suggest that p53 DNA binding is necessary for its negative effect on hAR. To test this hypothesis, the DNA sequence of our PSA promoter was analyzed for p53-responsive elements, and several elements were identified at Ϫ275 to Ϫ221. Disruption of these elements reduced hAR transactivation of this promoter but had no effect on p53 repression of hAR, 4 suggesting that p53 binding to PSA promoter is not critical for its negative activity. This is supported by our finding that p53 has the same negative effect on another androgen-responsive promoter, the human kallikrein-2 (hKLK2). 4 Additionally our data suggest that p53 blocks hAR activity by interfering with its N-to-C interaction and subsequent DNA binding. Neither effect depends on p53 DNA binding; thus, this would be consistent with p53 having an effect on hAR transactivation that is not promoter-specific as our data thus far suggest. Thus, it is conceivable that the point mutation found in p53W248 disrupts not only DNA binding but also some other activity that is critical for the p53 repression of hAR.
Data from several laboratories (10 -12) suggest that the hAR N-to-C interaction is physiologically relevant and is likely responsible for the ability of this receptor to homodimerize in an antiparallel orientation and to subsequently bind to DNA. Physiological relevance has come from a study showing that two mutations in the hAR ligand-binding domain causing androgen insensitivity syndrome disrupt the N-to-C interaction without influencing ligand binding (12) and several other studies demonstrating that hAR coactivators, including SRC-1 (54,80), TIF2 (55), CREB-binding protein (54), and c-Jun, 2 mediate this interaction. The importance of the N-to-C interaction in DNA binding was recently demonstrated by single mutation in the hAR activation function-2 (AF-2) that abolishes both N-to-C interaction and DNA binding in vitro. 2 In view of these data, it might be expected that a repressor of hAR activity may disrupt the N-to-C interaction and DNA binding of this receptor. Indeed this is precisely what was observed with p53. Importantly the two repression-defective mutants p53⌬C30 and p53W248 were deficient in disrupting hAR N-to-C interaction and DNA binding, confirming the specificity of the p53 activity on hAR. These results also argue that disruption of N-to-C interaction and DNA binding may be the mechanism by which p53 exerts its negative effect on hAR transcriptional activity.
How could p53 disrupt the hAR N-to-C interaction? p53 may have a direct effect by physically associating with hAR and thereby interfering with its dimerization function. However, to date we have been unable to detect a protein-protein interaction between p53 and hAR using in vitro precipitation methods and mammalian two-hybrid assays. 3 A second possibility is that p53 has an indirect effect by blocking the activity of a protein mediating the hAR N-to-C interaction. Several proteins have been shown to have such a mediating activity, including CREB-binding protein (54), TIF2 (55), and SRC-1 (80). We provide evidence here that c-Jun, one protein recently demonstrated to mediate hAR N-to-C interaction, 2 is able to relieve the negative activity of p53 on hAR N-to-C interaction, DNA binding, and transactivation. Hence it is possible that p53 can interfere with the positive effect of c-Jun, or other protein, on hAR N-to-C interaction. However, this apparently does not happen as a result of a physical interaction between p53 and c-Jun since we have failed to measure such an interaction. 3 It is also possible that p53 and c-Jun may act on hAR via mechanistically distinct pathways. p53 may block hAR activity via a product of a p53-activated gene that can interfere with hAR N-to-C interaction. Of the few known p53 target genes, including p21 (81), MDM2 (82,83), cyclin G (84), IGF-BP3 (85), Bax (86), and GADD45 (87), none of these express proteins that have been implicated in hAR activity. It remains to be seen if there may be a currently unknown p53 target gene(s) that may express such a protein(s).
Regardless of how the p53 disruption of hAR N-to-C interaction is mediated, the finding that p53 blocks hAR transcriptional activity opens the possibility that this tumor suppressor may be an important modulator of androgen-regulated gene expression. If this is the case, then p53 mutations that disrupt the negative activity of this protein on hAR would lead to higher hAR transactivation. Since higher hAR transcriptional activity correlates with a higher risk and higher grade of prostate cancer (for review, see Ref. 51), it is possible that such p53 mutations may be associated with a more malignant prostate cancer. It now appears that prostate cells that harbor p53 mutations are selected for the establishment of prostate cancer metastases (for review, see Ref. 38). It is also known that many of these late-stage and metastatic prostate cancer cells have functional AR and AR-dependent transcription (for review, see Refs. 51 and 52). Interestingly the p53 mutation R248W, which has been shown in this study to disrupt the p53 negative activity on hAR, is found in lymph node metastasis of prostate cancer (88). Thus, it is conceivable that the aggressive nature of prostate cancer cells harboring p53 mutations may be due, in part, to the inability of the p53 mutants to down-regulate AR-dependent transcription. However, it is important to remember that mutations such as R248W interfere with not only the p53 activity on hAR but also with the ability of this tumor suppressor to activate direct target genes. Therefore, the metastatic phenotype of prostate cancer might very well depend on changes in gene expression that are regulated by p53 either via a direct transcriptional effect or through its negative effect on hAR.