Transcriptional Interferences between Normal or Mutant Androgen ...

0013-7227/99/$03.00/0 Endocrinology Copyright © 1999 by The Endocrine Society

Vol. 140, No. 1 Printed in U.S.A.

Transcriptional Interferences between Normal or Mutant Androgen Receptors and the Activator Protein 1—Dissection of the Androgen Receptor Functional Domains* ´ ATRICE TE ´ ROUANNE, JEAN-MARC LOBACCARO†, NICOLAS POUJOL, BE ´ VIRGINIE GEORGET, STEPHANE FABRE, SERGE LUMBROSO, AND CHARLES SULTAN Institut National de la Sante´ et de la Recherche Me´dicale (J.-M.L., N.P., B.T., V.G., S.L,. C.S)., INSERM U439, Pathologie Molećulaire des Rećepteurs Nucleáires, 34090 Montpellier, France; Unite´ Biochimie Endocrinienne du De´veloppement et de la Reproduction, Hoˆpital Lapeyronie (N.P., S.L., C.S.), 34295 Montpellier, France; Physiologie Compareé et Endocrinologie (S.F.), CNRS-ERS63, Les Ce´zeaux, 63177 Aubie`res, France; Unite´ Endocrinologie Pe´diatrique (C.S.), Hoˆpital A. de Villeneuve, 34295 Montpellier, France ABSTRACT We investigated the interferences of the normal or mutated androgen receptor with the activator protein-1 (AP-1) by assessing their effects on transcriptional activity in CV-1 cells. A luciferase reporter gene was constructed downstream from either a promoter for the mouse vas deferens protein, or a trimerized 12-O-tetradecanoyl phorbol-13-acetate-response element site whose transcriptions are activated by androgen and 12-O-tetradecanoyl phorbol-13-acetate, a potent AP-1 activator. The blockade of dephosphorylation by protein phosphatases identifies the protein phosphatases that modulate the AP-1/androgen receptor cross-talk. Using engineered or naturally occurring androgen receptor mutants that are responsible for complete or partial androgen insensitivity syndromes, we defined the subregions involved in the cross-talk of the androgen receptor with

the AP-1 factors. First, it appears that the 188 first amino acids of the N-terminal domain of the androgen receptor are necessary to obtain a full transrepression. Second, a functional and intact ligand binding domain is critical for the modulation of androgen/AP-1 pathway interactions. Third, normal DNA binding capacity of the androgen receptor is not required. Two mutants at positions 568 and 581 of the DNA binding domain demonstrate that the transactivation and transrepression functions of the androgen receptor can be dissociated. Collectively, these data indicate that several segments of the androgen receptor are involved in cross-talk with the AP-1 pathway. Mutations within the DNA binding domain of the androgen receptor highly impair these interferences. (Endocrinology 140: 350 –357, 1999)

T

HE ANDROGEN receptor (AR) belongs to a large family of ligand binding transcription factors that includes receptors for steroid/thyroid hormones, retinoic acids, vitamin D, and other proteins of unknown function (1). These nuclear receptors contain three functional domains. The amino-terminal region contains sequence information that specifies gene recognition for transcriptional activation. In the AR, the carboxy-terminal domain is responsible for ligand binding and for the interactions with heat shock pro-

teins (HSP), which prevent the receptor from binding to a DNA in the unliganded state (2). This domain also has a transactivation function, signals for nuclear localization, and is involved in the dimerization process (3 and 4). Finally, the central DNA binding domain recognizes specific DNA sequences in regulatory regions of hormone-responsive genes. On the basis of a three-dimensional model of the DNA binding domain/DNA complex (5), we have shown that the least conserved amino acids among the glucocorticoid (GR) and AR family are distributed on the surface of the protein, on the side opposite to the DNA binding region. These amino acids do not significantly affect the monomer-DNA or the monomer-monomer interactions. We therefore proposed that some of the 18 amino acids in the DNA binding domain that differ between AR and GR may be involved in interactions with accessory proteins that control target gene transcription. These differences also suggested that even if AR and GR are structurally very close, the conclusions drawn concerning one receptor do not necessarily pertain to the other. Although the AR structure has been studied with the use of wild-type and mutant ARs from patients with androgen resistance, the action of the androgen-AR complex in target cells is only partially understood. This is mainly because AR

Received April 29, 1998. Address all correspondence and requests for reprints to: Prof. Charles Sultan, INSERM Unite´ 439, 70 rue de Navacelles, 34090 Montpellier, France. E-mail: [email protected]. * Part of this work was presented at the 10th International Congress of Endocrinology, San Francisco, California, June 12–15, 1996 and was awarded the Henning Andersen Prize (best abstract) at the 35th Annual Scientific Meeting of the European Society for Pædiatric Endocrinology, Montpellier, France, September 15–18, 1996. This work was supported by the Institut National de la Sante´ et de la Recherche Me´dicale; the Universite´ Montpellier I; the Association pour la Recherche sur le Cancer, Grant 6711; and the Fe´de´ration Nationale des Centres de Lutte contre le Cancer, Grant 705193. † Present address: Howard Hughes Medical Institute (HHMI) and Department of Pharmacology, The University of Texas Southwestern Medical Center at Dallas, Dallas, Texas, 75235-9050.

350

MUTANT ANDROGEN RECEPTORS AND AP-1 CROSS-TALK

is believed to be part of a large, multicomponent transcription regulation complex, the precise components of which are incompletely defined. Moreover, it is well known that nuclear receptor action can also be modulated by transcription factors. One of these factors is the composite transcription factor activating protein 1 (AP-1), composed either of a c-Fos/c-Jun heterodimer or a c-Jun homodimer. Hence, AP-1 has been demonstrated to have a positive (6) or a negative (7) effect on the AR-mediated transcription activity. There are also controversial data regarding the mechanism involved in these transcriptional interferences based on gel retardation assays. Bubulaya et al. (6) suggested only a direct interaction between AR and AP-1. Conversely, Aarnisalo et al. (8) reported that the intracellular CREB-binding-protein (CBP) could mediate the interactions in part through competition, as has been demonstrated for other nuclear receptors (9). Although the transcriptional interactions between AR and AP-1 have been studied, few of the regions in the AR involved in these interactions have been defined. To this end, we studied several engineered and naturally occurring AR mutations. After transient transfections in CV-1 cells, we investigated the capacity of these mutants to interfere with AP-1 on an androgen (transactivation) or an AP-1 (transrepression) modulated gene. In addition, we used staurosporine, a microbial alkaloid that blocks phosphorylation by membrane-translocated protein kinase C (10) in presence of the AP-1 activator 12-O-tetradecanoyl phorbol-13-acetate (TPA). Calyculin-A and okadaic acid, specific inhibitors of protein phosphatases (11), were also used to study the reversible phosphorylation of proteins involved in AP-1/ androgen pathway interactions. Mutants of the three functional domains of the AR showed that the hormone-binding domain, part of the DNA binding domain, and the first part of the N-terminal domain are all necessary to establish the fully competent three-dimensional structure of the interface with the AP-1 pathway. Materials and Methods Recombinant plasmids Mutant expression vectors (schematically illustrated in Fig. 1) were constructed from pCMV-AR (12). Mutants Gly568Trp, Val581Phe, and Arg608Lys were previously described by our group in patients with complete or partial androgen insensitivity syndromes (13–15). Mutants Cys579Tyr and Arg585Lys were found in two patients with complete androgen insensitivity syndrome (manuscript in preparation), whereas Arg607Gln was detected in two unrelated families with partial androgen insensitivity syndrome (16). Two truncated ARs were constructed by cloning the AocI-AgeI fragment containing the ligand and DNA binding domains and part of the amino terminal domain of the AR complementary DNA (cDNA) into the KpnI/BamHI-cleaved pUC19. Deletions were then constructed using single restriction fragment sites and ligations: the KpnI-EspI fragment was recloned into the KpnI/EspI-cleaved pCMVAR. ARD661–919, deleted from amino acid 661, was obtained with the Tth-111I/SmaI and Tth-111I/StuI restriction enzymes. ARD1– 188, starting with a methionine at position 189, was obtained by excision of the BglII/AflII fragment of pCMV-AR. This mutant corresponds to the A form of the AR described by Wilson and McPhaul (17). All of the recombinant plasmids were verified by sequencing. The mouse c-Fos and c-Jun constructs were provided by Dr. D. Chalbos (INSERM U148, Montpellier, France). These cDNAs (18) were present in pCI vector: c-Fos was inserted into the SpeI/EcoRI site, and c-Jun was inserted into the SalI/BglII site for cell transfections. The

351

FIG. 1. Schematic location of the amino acid substitutions and truncations of the androgen receptor. NTD, N-terminal domain; DBD, DNA binding domain; LBD, ligand binding domain; CAIS, complete androgen insensitivity syndrome; PAIS, partial androgen insensitivity syndrome; MBC, male breast cancer. Numbering is according to Lubahn et al. (47). Inset, Immunoblot analysis of the wild-type AR and the different mutants expressed in COS-7 cells and immunostained with antibody SpO61. pCMV-Luc was provided by Dr. P. Balaguer (INSERM U439, Montpellier, France). The mouse vas deferens protein (MVDP)-Luc reporter gene was constructed from the promoterless basic plasmid pGL3 (Promega Corp., Lyon, France) expressing luciferase under the control of the androgenregulated promoter of the 0.8-kb fragment of the MVDP (19), which did not contain any functional TPA-response element (TRE) (20). The TPAregulated reporter gene TRE3-tk-Luc described by Astruc et al. (21) is under control of the AP-1 complex.

Transfection of COS-7 cells and immunoblot analysis COS-7 cells were plated in 100-mm dishes (1.2 3 10-6 cells), and 8 h later they were transiently transfected with 10 mg of expression plasmid of wild-type or mutants using the calcium phosphate precipitate, as previously described (22). The suspension was added to the culture media for 16 h. Cells were cultured in DMEM with 3% FCS and harvested 2 days later in 100 ml of 160 mm Tris, pH 6.8, 200 mm dithiothreitol, 20% glycerol, 4% SDS, and 0.004% bromophenol blue in the presence of protease inhibitors (1 mm PMSF, 0.05 mm leupeptin, 0.01 mm pepstatin). Samples were boiled for 5 min and centrifuged at 13,000 3 g for 10 min. Analysis of COS-7 cell extracts was done by immunoblotting with the polyclonal antibody SpO61 (23) directed against the Nterminal domain of the human AR, using the ECL detection system (Amersham, Cergy, France). For binding assays, COS-7 cells were transfected in 12-well tissue culture dishes with 0.1 mg of different expression vectors (mutants and wild-type AR) and 0.1 mg of pCMV-b galactosidase. Whole steroid binding assays were performed as described by Georget et al. (24).

Cotransfections of CV-1 cells and luciferase assays CV-1 cells were grown in DMEM supplemented with 10% FCS (Gibco BRL, Cergy, France) and cultivated in 6-well plates. The transfections were performed using the calcium-phosphate technique as previously described (22). Briefly, cells were cotransfected with various amounts

352


Endo • 1999 Vol 140 • No 1

(10 –100 ng) of wild-type or mutant AR, or c-Fos or c-Jun expression vectors, 500 ng of pCMV5-b-galactosidase as an internal control for transfection efficiency, and 1.5 mg of the reporter gene. Transactivation has been defined as the capacity of the AR to induce the expression of the androgen-regulated gene in the presence or absence of TPA, which stimulates PKC by mimicking the action of the diacylglycerol and ultimately activates AP-1. Conversely, transrepression has been defined as the capacity of the AR to repress the expression of the TPA-regulated gene in the presence or absence of androgen. For the transactivation assays, cells transfected with the androgen target gene MVDP-Luc were incubated for one day with various concentrations of the androgen agonist methyltrienolone (R1881, NEN Life Science Products, Paris, France) or ethanol, in the presence or absence of TPA. For the transrepression assays, cells transfected with the TPA target gene (TRE3-tk-Luc) were incubated in the presence or absence of R1881 for 30 h and in the presence of TPA. CV-1 cells were chosen for the luciferase assays because they give a higher induction factor with TRE3-tk-Luc than COS-7 cells. Cells were harvested after incubations and assayed for luciferase and b galactosidase activities and protein content. The fold increase in luciferase activity was determined relative to basal activity in the absence of androgen or TPA and was corrected for transfection efficiency using the b-galactosidase activity. In each experiment, the mock pCMV5 was used as control (22). For the study of phosphorylation processes, staurosporine (30 nm), calyculin-A (1 nm) or okadaic acid (1 nm) was added 10 min before TPA (50 nm). It should be noted that at more than 10 nm, both calyculin-A and okadaic acid presented significant toxicity, as shown by the cells that became rapidly (within 2 h) detached from the culture dish.

Results Expression and ligand binding properties of wild-type and mutant receptors

Protein expression was assessed by immunoblot analysis of extracts of COS-7 that had been transiently transfected with wild-type or mutant ARs. No change in expression level or proteolysis was observed between the wild-type receptor and the different mutants (Fig. 1). The discrepancy between the predicted molecular mass based on the AR sequence and the AR protein size determined by SDS-PAGE, except for the mutant ARD1–188, has already been reported by Jenster et al. (25). The N-terminal region harboring most of the phosphorylation sites is believed to be responsible for this mobility retardation. In addition, the different AR mutants have also been studied for their androgen-binding capacity by saturation curves and Scatchard analyzes. As already described, only AR mutant ARD661–919 (3) was unable to bind androgen. The other ARs had Kd values within the same range (0.6 6 0.2 nm) as the wild-type (data not shown). In CV-1 cells, TPA and androgen have a reciprocal antagonistic action on both MVDP- and TRE3-tk-Luc reporter genes

Transcriptional cross-modulation between AR and AP-1 factor was studied in CV-1 cells lacking endogenous AR. Transient transfections of MVDP-Luc reporter gene in these cells have already been proved to give a suitable system for studying the effects of androgen (19). Different amounts of expression vector (20, 50, 100 ng) and two R1881 concentrations (0.1 nm and 10 nm) were used to study the production of luciferase (Fig. 2A). Whatever the amount of expression vector, TPA (50 nm) clearly antagonized the effect of R1881 on the luciferase transactivation by the wild-type AR. We observed up to a 50% decrease in luciferase activity. When CV-1 cells were cotransfected with the empty pCMV vector,

FIG. 2. A, Transactivation of the MVDP-Luc target gene in absence or presence of TPA in CV-1 cells. Transactivation was tested in CV-1 cells by cotransfecting the reporter gene MVDP-Luc with indicated amount of expression vector. Twenty, 50, 100 ng of AR expression vector and 1.5 mg of MVDP-Luc were used. Sixteen hours after transfection, the cells were incubated for 30 h in fresh DMEM medium without serum, in presence of 0.1 or 10 nM R1881. Ten hours before harvesting, 50 nM TPA was added. The total amount by well was kept constant by adding empty pCMV as needed. Presented as a bar diagram are the averages of the relative light units (RLU) with standard deviations from three independent experiments in duplicate. B, Repression of the TRE3-tk-Luc target genes in CV-1 cells. One hundred nanograms of empty vector or AR expression vector and 1.5 mg of TRE3-tk-Luc were used. Cells were incubated with R1881 10 nM alone for 30 h, or with TPA 50 nM for 10 h. The results, expressed as induction factors, are the averages with standard deviations from at least three independent experiments in duplicate. In the absence of TPA, the value of induction factor was 1. The baseline was not significantly modified by incubation with androgen alone (less than 10%).

no increase or decrease in basal luciferase activity was observed after TPA incubation. A reciprocal antagonistic effect between TPA and androgens was observed on the TRE3-tkLuc gene, demonstrating a transrepression activity of AR (Fig. 2B): R1881 inhibited the transcription of the TPA-regulated luciferase by 43%, reducing the fold induction from 5.4 to 2.3. No decrease in reporter gene activity occurred if the empty pCMV vector was used in lieu of pCMV-AR. Effects of time or TPA concentration on MVDP-Luc and TRE3-tk-Luc reporter genes in CV-1 cells

Both MVDP-Luc and AP-1 transactivation functions were tested during various periods (6 –24 h) with different TPA concentrations (0.1–50 nm). As shown in Fig. 3A, the antagonistic action of the TPA on the androgen induced MVDPLuc activation was time dependent. The inhibition was approximately in the same range between 10 –12 h of TPA treatment. At 24 h, no luciferase activity was detected. The AP-1 regulated luciferase transactivation increased after 6 h


FIG. 3. A, MVDP- and TRE3-tk- controlled luciferase activity in presence of 50 nM of TPA for different times. To measure inhibition of transactivation in presence of TPA, transfected cells with 100 ng of AR expression vector and 1.5 mg of MVDP-Luc were incubated with R1881 10 nM for 30 h combined with TPA for 6, 8, 10, 12, and 24 h. For assays of AP-1 transactivation, cells were transfected with TRE3tk-Luc (1.5 mg) and 100 ng of mock pCMV5 vector and incubated with R1881 10 nM and TPA as described above. B, CV-1 cells were transfected with TRE3-tk-Luc or MVDP-Luc as described above and incubated with R1881 10 nM for 30 h and combined for 10 h with various amounts of TPA (0.1, 0.5, 1, 5, 10, and 50 nM). In both figures, the transactivation of the two reporter genes is expressed as fold-induction of luciferase activity. Each data point was tested in duplicate and induction of luciferase activity was calculated from three different experiments.

to 10 h of TPA incubation. Prolonged treatment with TPA caused a decrease in AP-1 activity. Based on these results, we chose to study the transrepression/transactivation mechanisms after 10 h of TPA treatment. Figure 3B shows that the TPA concentration had to be equal to 1 nm to induce an AP-1 transactivation with the TRE3-tk-Luc reporter gene in CV-1 cells, and higher than 1 nm to trigger the inhibition of MVDPLuc transactivation. c-Jun and c-Fos differentially alter the transcriptional activity of AR

To determine whether the TPA effect on the androgenregulated MVDP-Luc gene induction in CV-1 cells was due to an increase in activated forms of AP-1 components, we transfected c-Jun or c-Fos expression vectors together with the wild-type AR. Transfection of increasing amounts (50, 100, 200, 400 ng) of c-Jun inhibited the androgen-induced luciferase activity from 18 to 86%. Conversely, the transfection of 400 ng of c-Fos only induced a decrease of 50% of luciferase activity (Fig. 4A). To demonstrate that this inhibition was not due to a transcriptional squelching, we cotransfected different combina-

353

FIG. 4. A, Androgen-regulated luciferase activity in CV-1 cells transfected by c-Fos or c-Jun. Cells were transfected by 100 ng of AR, 1.5 mg of reporter gene MVDP-Luc, and either c-Fos or of c-Jun expression vectors (amount indicated as ng in the figure). The luciferase signal measured in the cells transfected by the empty vector pCI was arbitrarily set at 100%. B, Effects of different cotransfections on basal activity of constitutive gene reporter pCMV-Luc. CV-1 cells were cotransfected with 1.5 mg of pCMV-Luc and expression vectors as follows: empty expression vectors (pCMV: 100 ng and pCI : 400 ng) and expression vectors (AR: 100 ng, c-Fos and c-Jun: 400 ng). Cells were treated with R1881 for 30 h after removal of the calcium precipitate. Columns represent the average of three experiments in duplicate.

tions of empty vector (pCMV or pCI) or vectors expressing AR, c-Fos or c-Jun with the constitutively active reporter vector pCMV-Luc. No change in the luciferase activity was found (Fig. 4B). Effects of staurosporine, calyculin-A and okadaic acid on TPA/androgen interactions on MVDP-Luc

To identify the biological processes that control the reversible phosphorylation of proteins involved in the TPA/ androgen interactions, staurosporine was used as an inhibitor of the PKC. The inhibition of serine/threonine-specific protein phosphatase-1 and/or -2A (PP-1 and PP-2A) was obtained by the specific inhibitors, calyculin-A and okadaic acid, respectively, which induced a potentiation of the TPA effects. At a concentration of 1 nm, okadaic acid inhibits PP-2A but does not inhibit the protein tyrosine phosphatases, other phosphatases or the kinases that have been tested (26). Conversely, calyculin-A, 20- to 300-fold more potent than okadaic acid, has been used as a specific PP-1 class phosphatase inhibitor (11). In CV-1 cells transfected by the androgen-regulated reporter gene, staurosporine (30 nm) administered along with

354


FIG. 5. Effects of 30 nM staurosporine, 1 nM calyculin-A, and 1 nM okadaic acid on R1881-induced luciferase activity in CV-1 cells. Transfected cells with 1.5 mg of MVDP-Luc and 100 ng of AR were treated for 20 h with or without 10 nM R1881 and then incubated for another 10 h with the various compounds indicated in the figure. Results are expressed as the percentage of luciferase activity in CV-1 cells from at least three independent experiments in duplicate. The 100% and 0% values were obtained with 10 nM R1881 and ethanol, respectively. Effects of staurosporine, calyculin-A and okadaic acid were compared with that of TPA alone. Student’s t test was used for statistical analysis. ***, P 5 0.001; *, P 5 0.01.

TPA slightly blocked TPA action on the wild-type AR (66 6 2% vs. 54 6 2% for TPA alone, P 5 0.01) (Fig. 5). This partial effect may be due to the low dose of staurosporine used. Indeed, Young et al. (27) only obtained 70% of the initial signal after incubation of LNCaP cells with 10 nm TPA and 200 nm staurosporine. Our data further support the hypothesis that protein kinase C activation is involved in the TPAmediated suppression of androgen action in the transfected CV-1 cells. The experiments with protein phosphatase inhibitors showed that okadaic acid or calyculin-A in combination with TPA was more effective than TPA alone. However, no difference was observed between the two compounds (37 6 1% vs. 35 6 1%, respectively), suggesting that both PP-1 and PP-2A are involved in processes leading to the AP-1 activity in CV-1 cells (Fig. 5). Androgen receptor mutants present various interaction profiles in CV-1 cells

Based on the previous experiments, 100 ng of various AR expression vector, 10 nm of R1881 and 50 nm of TPA were used. Treatments with TPA were performed for 10 h. As previously shown (Fig. 2B), the mock vector did not inhibit or increase induction of luciferase. According to the results summarized in Fig. 6, A and B, mutants have been classified in three types: a) mutants that were unable to transactivate and to transrepress AP-1 activity, b) mutants that only had a transrepressive activity, and c) mutants that did not highly differ from the wild-type AR in either transactivation or transrepression function. The case of mutants Gly568Trp and Cys579Tyr will be reported in the following section.

Endo • 1999 Vol 140 • No 1

FIG. 6. Transactivation of MVDP-Luc (A) and transrepression of TRE3-tk-Luc (B) reporter genes, respectively, in CV-1 cells. One hundred nanograms of AR expression vectors and 1.5 mg of each reporter gene were used. Cells were incubated with R1881 10 nM for 30 h or TPA 50 nM for 10 h, alone or combined. Presented as a bar diagram are the averages of the induction factors with standard deviations from at least five independent experiments in duplicate. DBD, LBD, and NTD indicate the mutants from the DNA binding, the ligand binding, and the N-terminal domains, respectively; CAIS, complete androgen insensitivity syndrome; PAIS, partial androgen insensitivity syndrome; MBC, male breast cancer.

a) Mutant ARD661–919, which does not bind androgen but is constitutively active, did not repress AP-1 activity. An absence of AP-1 repression was also observed with the mutant Leu707Arg (data not shown), which lacks androgen binding and causes complete androgen insensitivity (28). This indicates that a functional and intact ligand binding domain is critical for the modulation of androgen/AP-1 pathway interactions. b) Mutants that cause complete androgen insensitivity syndrome and lack any transcriptional activity, such as Val581Phe or Arg585Lys, had a capacity to transrepress similar to that of the wild-type. So, mutations that dramatically affect the DNA binding capacity of the AR do not necessarily alter the interactions with the AP-1 pathway. This suggests that a different structural organization of the DBD of the AR is involved in these processes. c) The transcriptional activity of mutants Arg607Gln and Arg608Lys, described in men with partial androgen insensitivity syndrome (15, 16, 29), was inhibited by TPA as well as the wild-type receptor from 14.8 to 6.3 and from 14 to 5.6, respectively. However, the Arg608Lys mutation was more efficient in transrepression than the wild-type receptor (P , 0.02), while Arg607 slightly impaired the AP-1 activity.


The 1–188 mutant gave a slight transactivation with MVDP-Luc and 10 nm R1881. This transactivation was little affected by TPA and the AP-1 activity was weakly repressed. Amino acids in the DNA binding domain discriminate between transactivation and transrepression

Among the different mutants, Gly568Trp presents an interesting profile. Indeed, the MVDP-dependent luciferase activity induced by this mutant is less inhibited by TPA incubation (75 6 6% of the reference signal) than the wildtype AR or the other mutants (Fig. 6A), and the TPA-induced luciferase activity is not impaired by this mutant (110 6 2% of the initial signal vs. 44 6 7% for the wild-type AR, P , 0.001) (Fig. 6B). These results suggest that the glycine residue at position 568 is more important for transrepression than transactivation. Cys579Tyr shows an intermediary effect between the wild-type and the Gly568Trp, with a lower level of transrepression (72 6 8% of the initial signal). The striking activation of the TRE-regulated gene by Gly568Trp in absence of TPA cannot be explained at this time. For the mutant Gly568Trp, the action of transfected c-Fos and c-Jun resulted in an inhibition in the same range as that observed with the wild-type receptor, and it mimicked the effect of extended TPA incubations (data not shown). Discussion

Several diseases, such as the androgen insensitivity syndromes, X-linked spinal and bulbar muscular atrophy, and prostate cancer, illustrate the crucial role of the AR in the transmission of the androgen signal (30). The mechanisms underlying the actions of this receptor are thus of considerable interest. Although in vitro studies (31) and development of a three-dimensional model of the DNA binding domain (5, 19, 32) have proved useful for studying the structure-function relationship of the AR, the mechanisms modulating AR transcriptional activation and repression are still poorly understood (33). This is mainly due to the complexity of the target cells, the small number of known androgen-regulated genes, and the lack of knowledge about growth factors modifying the androgen response. One may hypothesize, furthermore, that a single modification of growth factor expression is able to modify the final phenotype, which would thus explain why subjects with identical AR gene mutations may have different phenotypes (31, 34). This hypothesis can be tested by evaluating the action of TPA, which mimics the actions of a number of different growth factors by stimulation of the PKC pathway (35). In this study, we provide evidence that TPA and androgen have a reciprocal antagonistic action in CV-1 cells: TPA decreases AR transcriptional activity measured with MVDPLuc, whereas AP-1 activity is inhibited by androgens. We also report a dose-dependent inhibition by exogenous c-Jun on AR transcriptional activity. When c-Fos was transfected into CV-1 cells, the degree of inhibition was lower. Such a mechanism might depend in part on the ratio of AR to c-Jun and c-Fos; however, one cannot exclude other co-regulator proteins. Bubulya et al. (6) have demonstrated that c-Jun

355

mediates an androgen-induced transactivation in COS-7 cells and that c-Fos is able to block AR activity in the presence or absence of transfected c-Jun. Sato et al. (7) have found that mutual repression of DNA binding activities is due to a direct interaction between AR and c-Jun in the LNCaP cell line. These authors found no difference in the inhibition of the luciferase activities in CV-1 cells transfected by c-Jun or cFos. These results, however, cannot be compared with our data because the authors only tested one concentration for the AP-1 components (500 ng/well) and rat AR cDNA (500 ng/well), i.e. higher than our conditions of transient transfection. The discrepancy of the agonistic and antagonistic effects between both the AR and AP-1 pathways could be relevant to the cell type and/or the availability of CBP concentration (36 –38). To provide insight into the mechanism of cross-modulation between the androgen and the AP-1 pathways, we used AR mutations to identify the subregions of the regions involved in these interactions. The results from the aminoterminal deletion (1–188) showed a transactivation equal to 30% of the AR and a slight inhibition of the AP-1 activity. Similar data have been reported by Kallio et al. (36): deletion of amino acid residues 40 –107 in the rat AR (40 –126 in the human AR) did not significantly modify the androgen/AP-1 cross-talk, while a more extensive deletion in the N-terminal domain, from amino acid 38 to 296 (38 –298 in the human AR), abolished the repressive activity of AR. It is interesting to note that this last deletion overlaps the N-terminal portion that we deleted (from amino acid 1 to 188). Similarly, Chamberlain et al. (39) have shown that amino-terminal deletion (40 –218) induced an activation and repression equal to the 50% of the wild-type. One of the key subregions may thus be located within amino acids 126 –218. Recently, an important helical region (189 –201) was described in the t1-core transactivation domain of the human GR (40). We propose that the homologous helical region in human AR starts at Ser 187. The deletion 1–188 could therefore prevent this a-helix formation by suppression of the two initial amino acid residues conserved in AR and GR (Ser187 and Thr188). Further studies are underway to delineate this region precisely, to determine its participation in the interferences with AP-1 and to analyze its coactivator recruitment. The data obtained from the mutants of the ligand binding domain led us to hypothesize that a liganded and active AR is necessary and sufficient for interaction with AP-1. Transrepression is impaired by mutant ARD661–919, which is constitutively active because it lacks an important part of the ligand binding domain encompassing the HSP-binding region. This suggests that the nuclear localization of the AR alone is not sufficient to repress AP-1 activity. Thus, in addition to an androgen binding, transrepression requires the integrity of the regions conformationally modified by this ligand binding. Concerning the mutants within the DNA binding domain, we showed that Val581Phe and Arg585Lys, which do not bind DNA and are thus totally inactive on an androgenregulated gene, have the same capacity of inhibition as the wild-type receptor, even though transrepression by the Arg585Lys mutant is lower. A functional DNA binding capacity, i.e. the ability to bind a target sequence, is thus not

356


required to repress the AP-1 activity, as described for the AR-Ets interaction (41) and in contrast to the AR-RelA interaction (42). It seems that the spatial position of the substituted amino acids in the DNA binding domain is more important than their contribution to the overall structure of the receptor. Our three-dimensional model suggests that valine 581 and arginine 585 are located in the protein-DNA interface and are unable to interact with other factors, whereas glycine at position 568 is located on the opposite surface, possibly interacting with proteins involved in the control of target gene transcription or part of the AR itself. Even though this model has been defined with the AR bound to DNA, we assume that in solution and liganded, the overall structure of the receptor is maintained. The substitution of this glycine, a small and polar amino acid, by a tryptophane, a more cumbersome and hydrophobic residue, would potentially modify or disrupt any protein-to-protein interactions, but it appears to be more important for transrepression than for transactivation. Similar dissociation of repression activity and transcriptional activity has been described for GR (43). Mutant Cys579Tyr presents similarly interesting results. Cysteine at position 579 is one of the four residues that coordinate the zinc ion to maintain the correct geometry of the first zinc finger. One might thus assume that substitution of this residue would definitively alter the functionality of the AR. However, mutant Cys579Tyr is half as active in transcription as the wildtype receptor. The functionality of this mutant could be explained by the fact that a tyrosine residue may also coordinate a zinc ion but with lower energy (44). Even though the first zinc finger is not totally destabilized, the structure of this domain is modified. This would explain why this mutant partially inhibits TPA-induced luciferase activity even though this difference is less clear cut than for mutant Gly568Trp. The importance of the spatial location of the substitution is also established by the study of mutants Arg607Gln and Arg608Lys. Arginine at position 608 interacts with residues of the first and second zinc fingers and is conserved among all members of the nuclear receptor superfamily. Substitution of this arginine with a lysine is predicted to produce a significant conformational change (19). Molecular modeling revealed that both arginines are partially surface-exposed, and may participate in a region of protein-protein interaction, in agreement with the hypothesis proposed for GR by Scheena et al. (45). Because of this modification in the second zinc finger structure, one may speculate that amino acid displacements are responsible for the increased transrepression capacity of the mutant Arg608Lys. In summary, the transrepression function appears to be mediated by a structure insensitive to some of the mutations that impair the major DNA recognition function of the receptor. Wise et al. (46) have similarly shown that a c-Jun mutant deficient in transactivation is able to interact with AR activity. These data demonstrate that transactivation and transrepression activity may also be separated for c-Jun. Conversely, mutations affecting amino acids located on the exposed surface of the DNA binding domain, including positions 568, 579, 608, and 562 (36), negatively or positively modulate interactions with AP-1 factor. In addition to defining the regions and/or amino acids

Endo • 1999 Vol 140 • No 1

involved in AR functioning, analysis of the interactions between natural AR mutants and the AP-1 pathway may help us to understand the mechanisms by which phenotypes are determined, or at least to raise hypotheses. Among the natural mutants, two present puzzling transactivation functions. Gly568Trp and Arg608Lys, which have been described in severe forms of partial androgen insensitivity syndrome, present normal transactivation capacity. These data are paradoxical, and one may speculate that a patient’s phenotype is due not only to a direct action of androgens on androgenregulated genes (transactivation function) but also to the cross-talk between androgen-mediated signaling systems and growth factor pathways, as suggested by Reinikainen et al. (35). In conclusion, our data demonstrate that the complex interaction between the androgen and AP-1 pathways involves several subregions of the AR, presumably jointly establishing the fully competent three-dimensional structure of the repression interface. These findings open up a new field of investigation into the identification and characterization of new AR subregions involved in the complex processes that modulate androgen action in target cells. Acknowledgments We are grateful to Dr. D. Chalbos for providing the c-Fos and c-Jun plasmids, Dr. T. R. Brown for the pCMV-AR plasmid, Dr. P. Balaguer for pCMV-Luc, and Dr. M. Pons for the TAT-tk-Luc and TRE3-tk-Luc plasmids. We also thank Dr. A. O. Brinkmann (Department of Endocrinology and Reproduction, Erasmus University, Rotterdam, The Netherlands) and Drs. J. L. Borgna, J. C. Nicolas, and M. Pons (INSERM U439) for critically reading this manuscript. We especially acknowledge Dr. J. D. Wilson (Department of Internal Medicine, The University of Texas Southwestern Medical Center at Dallas, Dallas, TX) for his advice and comments on the manuscript.

References 1. Mangelsdorf DJ, Thummel C, Beato M, Herrlich P, Schutz G, Umesono K, Blumberg B, Kastner P, Mark M, Chambon P 1995 The nuclear receptor superfamily: the second decade. Cell 83:835– 839 2. Marivoet S, Vandijck P, Verhoeven G, Heyns W 1992 Interaction of the 90-kDa heat shock protein with native and in vitro translated androgen receptor and receptor fragments. Mol Cell Endocrinol 88:165–174 3. Simental JA, Sar M, Lane MV, French FS, Wilson EM 1991 Transcriptional activation and nuclear targeting signals of the human androgen receptor. J Biol Chem 266:510 –518 4. Jenster G, van der Korput HAGM, van Vroonhoven C, van der Kwast TH, Brinkmann AO 1991 Domains of the human androgen receptor involved in steroid binding, transcriptional activation, and subcellular localization. Mol Endocrinol 5:1396 –1404 5. Lobaccaro JM, Poujol N, Chiche L, Lumbroso S, Brown TR, Sultan C 1996 Molecular modeling and in vitro investigations of the human androgen receptor DNA-binding domain: application for the study of two mutations. Mol Cell Endocrinol 116:137–147 6. Bubulya A, Wise SC, Shen XQ, Burmeister LA, Shemshedini L 1996 c-jun can mediate androgen receptor-induced transactivation. J Biol Chem 271:24583–24589 7. Sato N, Sadar MD, Bruchovsky N, Saatcioglu F, Rennie PS, Sato S, Lange PH, Gleave ME 1997 Androgenic induction of prostate-specific antigen gene is repressed by protein-protein interaction between the androgen receptor and AP-1/c- Jun in the human prostate cancer cell line LNCaP. J Biol Chem 272:17485–17494 8. Aarnisalo P, Palvimo JJ, Janne OA 1998 CREB-binding protein in androgen receptor-mediated signaling. Proc Natl Acad Sci USA 95:2122–2127 9. Kamei Y, Xu L, Heinzel T, Torchia J, Kurokawa R, Gloss B, Lin SC, Heyman RA, Rose DW, Glass CK, Rosenfeld MG 1996 A CBP integrator complex mediates transcriptional activation and AP-1 inhibition by nuclear receptors. Cell 85:403– 414 10. Tamaoki T, Nomoto H, Takahashi I, Kato Y, Morimoto M, Tomita F 1986 Staurosporine, a potent inhibitor of phospholipid/Ca11 dependent protein kinase. Biochem Biophys Res Commun 135:397– 402

MUTANT ANDROGEN RECEPTORS AND AP-1 CROSS-TALK 11. Ishihara H, Martin BL, Brautigan DL, Karaki H, Ozaki H, Kato Y, Fusetani N, Watabe S, Hashimoto K, Uemera D, Hartshorne DJ 1989 Calyculin A and okadaic acid: inhibitors of protein phosphatase activity. Biochem Biophys Res Commun 159:871– 877 12. Brown TR, Lubahn DB, Wilson EM, French FS, Migeon CJ, Corden JL 1990 Functional characterization of naturally occurring mutant androgen receptors from subjects with complete androgen insensitivity. Mol Endocrinol 4:1759 –1772 13. Lobaccaro JM, Belon C, Lumbroso S, Olewniczack G, Carre´-Pigeon F, Job JC, Chaussain JL, Toublanc JE, Sultan C 1994 Molecular prenatal diagnosis of partial androgen insensitivity syndrome based on the Hind III polymorphism of the androgen receptor gene. Clin Endocrinol (Oxf) 40:297–302 14. Lumbroso S, Lobaccaro JM, Belon C, Martin-Coignard D, Chaussain JL, Sultan C 1993 A new mutation within the DNA-binding domain of the androgen receptor gene in a family with complete androgen insensitivity syndrome and positive receptor-binding activity. Fertil Steril 60:814 – 819 15. Lobaccaro JM, Lumbroso S, Belon C, Galtier-Dereure F, Bringer J, Lesimple T, Heron JF, Pujol H, Sultan C 1993 A second instance of male breast cancer linked to a germline mutation of the androgen receptor gene. Nat Genet 5:109 –110 16. Sinisi A, Lobaccaro JM, Lumbroso S, Pasquali D, Poujol N, Sultan C, Bellastella A 1996 Partial androgen insensitivity due to androgen receptor gene DNA binding domain point mutation in two prepubertal brothers. Horm Res [Suppl] 46:116 17. Wilson CM, Mcphaul MJ 1994 A and B forms of the androgen receptor are present in human genital skin fibroblasts. Proc Natl Acad Sci USA 91:1234 –1238 18. Philips A, Chalbos D, Rochefort H 1993 Estradiol increases and anti-estrogens antagonize the growth factor-induced activator protein-1 activity in MCF7 breast cancer cells without affecting c-fos and c-jun synthesis. J Biol Chem 268:14103–14108 19. Poujol N, Lobaccaro JM, Chiche L, Lumbroso S, Sultan C 1997 Functional and structural analysis of R607Q and R608K androgen receptor substitutions associated with male breast cancer. Mol Cell Endocrinol 130:43– 47 20. Fabre S, Manin M, Pailhoux E, Veyssie`re G, Jean C 1994 Identification of a functional androgen response element in the promoter of the gene for the androgen-regulated aldose reductase-like protein specific to the mouse vas deferent. J Biol Chem 269:5857–5864 21. Astruc ME, Chabret C, Bali P, Gagne D, Pons M 1995 Prolonged treatment of breast cancer cells with antiestrogens increases the activating protein-1 mediated response: involvement of the estrogen receptor. Endocrinology 136:824 – 832 22. Lobaccaro JM, Lumbroso JM, Poujol N, Georget V, Brinkmann AO, Malpuech G, Sultan C 1995 Complete androgen insensitivity syndrome due to a new frameshift deletion in exon 4 of the androgen receptor gene: functional analysis of the mutant receptor. Mol Cell Endocrinol 111:21–28 23. van Laar JH, Voorhost-Ogink MM, Zegers ND, Boersma WJA, Claassen E, van der Korput JAGM, Ruizeveld de Winter JA, Van der Kwast TH, Mulder E, Trapman J, Brinkmann AO 1989 Characterization of polyclonal antibodies against the N-terminal domain of the human androgen receptor. Mol Cell Endocrinol 67:29 –38 24. Georget V, Lobaccaro J, Te´rouanne B, Mangeat P, Nicolas J, Sultan C 1997 Trafficking of the androgen receptor in living cells with fused green fluorescent protein-androgen receptor. Mol Cell Endocrinol 129:17–26 25. Jenster G, Deruiter PE, Vanderkorput HAGM, Kuiper GGJM, Trapman J, Brinkmann AO 1994 Changes in the abundance of androgen receptor isotypes: effects of ligand treatment, glutamine-stretch variation, and mutation of putative phosphorylation sites. Biochemistry 33:14064 –14072 26. Haystead TAJ, Sim ATR, Carling D, Honnor RC, Tsukitani Y, Cohen P, Hardie DG 1989 Effects of the tumor promotor okadaic acid on intracellular protein phosphorylation and metabolism. Nature 337:78 – 81 27. Young CY, Murtha PE, Zhang J 1994 Tumor-promoting phorbol ester-induced cell death and gene expression in a human prostate adenocarcinoma cell line. Oncol Res 6:203–210 28. Lumbroso S, Lobaccaro JM, Georget V, Leger J, Poujol N, Terouanne B,

29.

30. 31. 32.

33. 34.

35. 36. 37. 38. 39. 40. 41.

42. 43. 44. 45. 46. 47.

357

Evainbrion D, Czernichow P, Sultan C 1996 A novel substitution (Leu707Arg) in exon 4 of the androgen receptor gene causes complete androgen resistance. J Clin Endocrinol Metab 81:1984 –1988 Wooster R, Mangion J, Eeles R, Smith S, Dowsett M, Averill D, Barrett-Lee P, Easton DF, Ponder BAJ, Stratton MR 1992 A germline mutation in the androgen receptor gene in two brothers with breast cancer and Reifenstein syndrome. Nat Genet 2:132–134 Brinkmann A, Jenster G, Risstalpers C, Vanderkorput H, Bruggenwirth H, Boehmer A, Trapman J 1996 Molecular basis of androgen insensitivity. Steroids 61:172–175 Quigley CA, Debellis A, Marschke KB, Elawady MK, Wilson EM, French FS 1995 Androgen receptor defects: historical, clinical, and molecular perspectives. Endocr Rev 16:271–321 Bruggenwirth HT, Boehmer ALM, Lobaccaro JM, Chiche L, Sultan C, Trapman J, Brinkmann AO 1998 Substitution of Ala(564) in the first zinc cluster of the deoxyribonucleic acid (DNA)-binding domain of the androgen receptor by Asp, Asn, or Leu exerts differential effects on DNA binding. Endocrinology 139:103–110 Kallio PJ, Palvimo JJ, Janne OA 1996 Genetic regulation of androgen action. Prostate 45–51 Rodien P, Mebarki F, Mowszowicz I, Chaussain JL, Young J, Morel Y, Schaison G 1996 Different phenotypes in a family with androgen insensitivity caused by the same M780I point mutation in the androgen receptor gene. J Clin Endocrinol Metab 81:2994 –2998 Reinikainen P, Palvimo JJ, Janne OA 1996 Effects of mitogens on androgen receptor-mediated trans activation. Endocrinology 137:4351– 4357 Kallio PJ, Poukka H, Moilanen A, Janne OA, Palvimo JJ 1995 Androgen receptor-mediated transcriptional regulation in the absence of direct interaction with a specific DNA element. Mol Endocrinol 9:1017–1028 Deruiter PE, Teuwen R, Trapman J, Dijkema R, Brinkmann AO 1995 Synergism between androgens and protein kinase-c on androgen-regulated gene expression. Mol Cell Endocrinol 110:R1–R6 Fabre S, Darne C, Veyssiere G, Jean C 1996 Protein kinase C pathway potentiates androgen-mediated gene expression of the mouse vas deferens specific aldose reductase-like protein (MVDP). Mol Cell Endocrinol 124:79 – 86 Chamberlain NL, Whitacre DC, Miesfeld RL 1996 Delineation of two distinct type 1 activation functions in the androgen receptor amino-terminal domain. J Biol Chem 271:26772–26778 Almlof T, Gustafsson JA, Wright APH 1997 Role of hydrophobic amino acid clusters in the transactivation activity of the human glucocorticoid receptor. Mol Cell Biol 17:934 –945 Schneikert J, Peterziel H, Defossez PA, Klocker H, Delaunoit Y, Cato ACB 1996 Androgen receptor-ets protein interaction is a novel mechanism for steroid hormone-mediated down-modulation of matrix metalloproteinase expression. J Biol Chem 271:23907–23913 Palvimo JJ, Reinikainen P, Ikonen T, Kallio PJ, Moilanen A, Janne OA 1996 Mutual transcriptional interference between RelA and androgen receptor. J Biol Chem 271:24151–24156 Heck S, Kullmann M, Gast A, Ponta H, Rahmsdorf HJ, Herrlich P, Cato ACB 1994 A distinct modulating domain in glucorticoid receptor monomers in the repression of activity of the transcription factor AP-1. EMBO J 13:4087– 4095 Furey W, Robbins A, Clancy L, Winge D 1986 Crystal structure of Cd, Zn metallothionein. Science 231:704 –710 Scheena M, Freedman LP, Yamamoto KR 1989 Mutations in the glucocorticoid receptor zinc finger region that distinguish interdigitated DNA binding and transcriptional enhancement activities. Genes Dev 3:1590 –1601 Wise SC, Burmeister LA, Zhou XF, Bubulya A, Oberfield JL, Birrer MJ, Shemshedini L 1998 Identification of domains of c-Jun mediating androgen receptor transactivation. Oncogene 16:2001–2009 Lubahn DB, Brown TR, Simental JA, Higgs HN, Migeon CJ, Wilson EM, French FS 1989 Sequence of the intron/exon junctions of the coding region of the human androgen receptor gene and identification of a point mutation in a family with complete androgen insensitivity. Proc Natl Acad Sci USA 86: 9534 –9538