Mutant and Wild-Type Androgen Receptors Exhibit Cross-Talk on ...

Mutant and Wild-Type Androgen Receptors Exhibit Cross-Talk on Androgen-, Glucocorticoid-, and Progesterone-Mediated Transcription

Paul M. Yen, Ying Liu, Jorma J. Palvimo, Mark Trifiro, Jeannie Whang, Leonard Pinsky, Olli A. Ja¨nne, and William W. Chin Division of Genetics (P.M.Y., Y.L., J.W., W.W.C.) Department of Medicine Brigham and Women’s Hospital and Harvard Medical School Boston, Massachusetts 02115 Divisions of Genetics and Endocrinology (M.T., L.P.) Department of Medicine Sir Mortimer B. Davis-Jewish General Hospital Lady Davis Research Institute for Medical Research McGill University, Montreal, Quebec H3T 12E2 Canada Institute of Biomedicine (J.J.P, O.A.J.) Department of Physiology University of Helsinki FIN-00014, Helsinki, Finland

Androgen, glucocorticoid, and progesterone receptors (ARs, GRs, and PRs) often can regulate transcription via composite hormone response elements in target genes. We have used artificial and natural mutant ARs from patients with androgen resistance to study their effects on dominant negative activity on wild type AR, GR, and PR function on mouse mammary tumor virus (MMTV) and tyrosine aminotransferase (TAT) promoters. Artificial ARs that contained internal deletions within the amino-terminal region had minimal transcriptional activity but blocked ligand-mediated transcription by wild type AR. Mutants containing deletions of the DNA-binding and ligand-binding domains had minimal or weak dominant negative activity. We then tested the ability of wild type and mutant ARs to modulate GR- and PR-mediated transcriptional activity. The amino-terminal deletion mutants exerted dominant negative effects on GR- and PR-mediated activity, both in the absence and presence of testosterone. Surprisingly, wild type AR, which had approximately 20% of the maximal transcriptional activity of GR on the MMTV promoter, also had dominant negative activity on dexamethasone-regulated transcription mediated by

GR. This dominant negative activity likely involves DNA binding because a point mutation in the DNAbinding domain abrogated such activity of an aminoterminal deletion mutant. Additionally, natural human AR mutants from patients with androgen resistance, which do not bind either DNA or ligand, did not block dexamethasone-mediated transcription. In summary, these studies suggest that mutant and wild type ARs can display dominant negative activity on other steroid hormone receptors that bind to a composite hormone response element. This cross-regulation may be important in regulating maximal transcriptional activity in tissues where these receptors are coexpressed and may contribute to the phenotype of patients with steroid hormone resistance. (Molecular Endocrinology 11: 162–171, 1997)

INTRODUCTION Steroid hormone receptors can recognize common hormone response elements (HREs). In particular, androgen, glucocorticoid, and progesterone receptors (ARs, GRs, and PRs) can regulate transcription via the mouse mammary tumor virus promoter (MMTV) (1–4). This promoter region between ⫺187 to ⫺69 contains two palindromic HREs separated by 40 bp and two

0888-8809/97/$3.00/0 Molecular Endocrinology Copyright © 1997 by The Endocrine Society

162

Dominant Negative Activity by Androgen Receptors

single downstream half-sites arranged as direct repeats. Functional studies have shown the upstream palindrome is sufficient for transcriptional activation by GR, and the proximal palindromic element and two half-sites also can mediate transcriptional activation by GR (4, 5). Deletion analysis has shown that dexamethasone, progesterone, and testosterone can regulate transcription via nucleotide sequences between ⫺201 and ⫺69 of the promoter (4, 5). DNAase I footprinting analyses of the promoter region showed nearly identical patterns for GR and PR, suggesting that each of the receptors binds to common sites coinciding with the palindrome and half-site sequences (6). Methylation interference studies also demonstrated that GR and PR make similar contacts with nucleotide sequences within these composite HREs (4, 7). Similar observations were found comparing AR, GR, and PR binding to the HRE sequence of the tyrosine aminotransferase (TAT) gene (8, 9). Although there have been studies examining the transcriptional activity of AR, GR, and PR on the MMTV promoter, there have been limited studies examining cross-talk and dominant negative activity by steroid hormone receptors on this or other composite response elements. Recently, mutant ARs have been shown to exert dominant negative activity on wild type AR-mediated transcriptional activity using a reporter containing two consensus androgen response elements (AREs) (10). In particular, deletion of a subregion within the the amino-terminal domain abrogated ligand-mediated transcriptional activity by AR, suggesting that there is an activation factor (AF) domain within this region similar to the AF-1 region of PR and estrogen receptor, and the ␶1 region of GR (3, 10, 11). Additionally, these deletion mutants displayed dominant negative activity on wild type AR function. In contrast, deletions of the DNA-binding and ligand-binding domains had little effect on dominant negative activity. In this study, we investigated a battery of artificial truncation and deletion mutants as well as natural AR mutants for their dominant negative activity on wild type AR, GR, and PR transcription on the MMTV promoter. Our studies demonstrate that the AF-1 deletion mutant of AR had virtually no transcriptional activity but exhibited strong dominant negative activity on AR, GR, and PR-1 (longer PR isoform, also known as PRB) activity. Surprisingly, wild type AR also had dominant negative activity on GR and PR-1 transcription. Studies with natural and artificial AR mutants strongly suggest that DNA binding by wild type and mutant AR is important in mediating the dominant negative activity. These results also suggest that there is cross-talk among steroid hormone receptors that may be important for target gene regulation in tissues where these receptors are coexpressed. Additionally, they indicate that receptor mutations in patients with androgen resistance may affect the functions of other receptors (e.g. GR) and thereby contribute to the patient phenotype in tissues where AR and GR are expressed.

163

RESULTS We first examined a battery of AR deletion mutants (Fig. 1A) for their effects on the transcriptional activity of the MMTV promoter in CV-1 cells. Wild type AR (WT AR) mediated approximately a 125- to 175-fold increase in transcriptional activity over vector alone in the presence of testosterone (Figs. 1B and 2). This increase appeared to be maximal as transfection of 3-fold more AR expression plasmid gave a similar fold-induction (data not shown). Deletion of amino acids 38–296 (⌬38-296) and 46–408 (⌬46-408) in the amino-terminal region almost completely abrogated ligand-dependent activation, whereas deletion of amino acids 40–147 (⌬40-147) had minimal effect. Of note, these mutants have similar ligand- and DNAbinding affinities as WT AR (9, 10). Thus, similar to previous findings with a consensus ARE (10), these data suggest that there is a potent AF region between amino acids 147 and 296. Deletion of the DNA-binding domain (⌬557-610) and the most carboxy-terminal 115 amino acids (⌬788-902) also completely abrogated transactivation. As a control for appropriate intracellular localization, cells transfected with mutant receptors were immunostained with anti-AR antibody. Such studies showed that these receptors are localized mainly in the nucleus (U. Karvonen, P. J. Kallio, O. A. Ja¨nne, and J. J. Palvimo, manuscript submitted). Jenster et al. (12) also previously showed nuclear localization of comparable human AR mutants. We next examined the dominant negative activity on WT AR function by the transcriptionally inactive AR deletion mutants. Deletion mutants ⌬46-408 and ⌬38296 completely blocked WT AR transctivation at 1:1 and 3:1 ratios of cotransfected mutant AR/WT AR expression plasmids (Fig. 2, and data not shown). The DNA-binding domain deletion mutant had no effect, and the carboxy-terminal deletion mutant had a weak effect on WT AR function at a 3:1 expression plasmid ratio. Taken together, these data are in agreement with those on a consensus ARE (10) and suggest that although amino-terminal deletion mutants of AR lose their ability to transactivate in the presence of testosterone, they exhibit potent dominant negative activity on WT AR function. Since GR also is a potent transcriptional activator via the MMTV promoter, we investigated the dominant negative activity of these mutants on GR-mediated transactivation in the absence or presence of testosterone. GR had a maximal transcriptional activity of approximately 1000-fold in the presence of dexamethasone (Fig. 3). AR deletion mutants ⌬46-408 and ⌬38296 had potent dominant negative activity on GRmediated transcriptional activation, both in the presence or absence of testosterone. Surprisingly, wild type AR and ⌬40-147 also blocked GR-mediated transcriptional activation, almost to the level of AR alone with testosterone. For WT AR, ⌬46-408, and ⌬40-147, addition of testosterone further augmented

MOL ENDO · 1997 164

Vol 11 No. 2

Fig. 2. Dominant Negative Activity by AR Mutants on Wild Type AR-Mediated Transcription CV-1 cells were transfected with 0.1 ␮g expression vector encoding WT AR and 0.3 ␮g AR deletion mutants, 1.7 ␮g MMTV-luciferase reporter, and 1.0 ␮g RSV-␤ galactosidase control vector. In some samples, pSG vector was added so that each sample contained the same amount of DNA. Cells were treated with 10⫺6 M testosterone for 48 h and harvested, and cell lysates were prepared and analyzed for luciferase activity as in Fig. 1. Data are expressed as means ⫾ SD (n ⫽ 4).

Fig. 1. AR Mutants and Their Transcriptional Activities A, Rat AR mutants used in these studies. B, Transcriptional activity via the MMTV promoter by wild type AR and AR deletion mutants. CV-1 cells were transfected with 0.1 ␮g expression vector encoding WT AR or AR deletion mutants, 1.7 ␮g MMTV-luciferase reporter, and 1.0 ␮g RSV-␤ galactosidase control vector. Cells were treated with 10⫺6 M testosterone for 48 h and harvested, and cell lysates were prepared and analyzed for luciferase activity as described in Materials and Methods. Luciferase activity was normalized to fold basal luciferase activity with 1-fold basal activity defined as reporter activity with empty expression vector in the absence of hormone. Data are expressed as means ⫾ SD (n ⫽ 4). *, point mutant; AD, amino-terminal domain deletions; DBD, DNA-binding domain deletion; and LBD, ligand-binding domain (carboxy terminus) deletion.

dominant negative activity. In contrast to the other AR mutants and WT AR, ⌬557-610 and ⌬788-902 had weaker dominant negative activity on GR-mediated transactivation. The weak dominant negative activity by ⌬557-610 may be due to squelching of common cofactors important for GR-mediated transcription. We also examined the effects of these mutants on the transcriptional activity of the more potent PR isoform,

PR-1, and observed similar results with wild type AR and the AR mutants (data not shown). To compare further the dominant negative activity of WT AR with ⌬46-408, we performed dose-response studies in which the amount of GR expression plasmid was constant, and increasing amounts of AR expression plasmid were cotransfected (Figs. 4). At a 3:1 ratio of AR:GR expression plasmid, AR blocked GR-mediated transcription to about 20% maximal transcription whereas a 1:1 ratio of ⌬46-408:GR expression plasmid blocked transcription to about the same level. Increasing amounts of ⌬46-408 almost completely blocked GRmediated transcription. Interestingly, when low amounts of AR and ⌬46-408 expression plasmids were cotransfected (1:1), there was a greater testosterone-dependent effect on dominant negative activity. To examine whether the dominant negative activity of ⌬46-408 required DNA binding, we examined whether a double AR mutant containing the same internal deletion and a mutation of the fourth coordinating cysteine of the first zinc finger (⌬46-408 C562G) was able to mediate this effect. The latter mutation has been shown to abrogate DNA binding of several nuclear hormone receptors, including AR (Refs. 13–15 and J. J. Palvimo and O. A. Janne, unpublished results). In contrast to ⌬46-408, this mutation was unable to block GR-mediated transcription, suggesting that DNA binding is important in mediating this effect (Fig. 5A). We also examined AR/GR cross-talk by ex-


165

6B). We observed that WT AR maintained good dominant negative activity, whereas AR DBDmut and LBDmut did not. These findings further support the role of DNA binding in mediating dominant negative activity. In combination with the earlier findings (Figs. 3 and 4), they also suggest that AR may modulate GR-mediated transactivation in tissues where both receptors are coexpressed. In support of this possibility, we also examined the expression of rat AR and human GR in CV-1 cells and showed that they are similarly expressed in CV-1 cells using a whole cell ligandbinding assay (Table 1). Moreover, they suggest that natural mutant ARs may have impaired dominant negative activity on GR-mediated transactivation (Fig. 6B).

DISCUSSION Fig. 3. Dominant Negative Activity by WT AR and AR Mutants on GR-Mediated Transcription CV-1 cells were transfected with 0.1 ␮g expression vector encoding GR and 0.3 ␮g WT AR and AR deletion mutants, 1.7 ␮g MMTV-luciferase reporter, and 1.0 ␮g RSV-␤ galactosidase control vector. In some samples, pSG vector was added so that each sample contained the same amount of expression vector. Cells were treated with 10⫺6 M testosterone and dexamethasone as indicated for 48 h and harvested, and cell lysates were prepared and analyzed for luciferase activity as in Fig. 1. Data are expressed as means ⫾ SD (n ⫽ 4). dex, Dexamethasone.

amining this effect using a reporter containing the TAT promoter and observed similar effects with ⌬46-408 and ⌬46-408 C562G (Fig. 5B). To examine further the mechanism for dominant negative activity by AR, we used two natural mutant human ARs from patients with androgen resistance to study their effects on AR- and GR-mediated transcription (Fig. 6). One of these mutants (AR DBDmut) contains a point mutation in the second zinc finger of the DNA-binding domain and has lost the ability to bind DNA, and the other mutant (AR LBDmut) has a point mutation in the ligand-binding domain and binds testosterone poorly (16, 17). AR DBDmut was unable to mediate liganddependent transcriptional activity; however, when cotransfected with wild type human AR (WT hAR), it surprisingly enhanced ligand-mediated transcriptional activity to a greater extent than that of WT AR (Fig. 6A). These findings suggest that the AR DBDmut either titrated out a repressor of AR function or, more likely, WT hAR and DBDmut may form heterodimers that are potent transcriptional activators. Studies using two other DNAbinding natural human AR mutants (R614H and ⌬614) also showed similar enhancement of WT AR transcriptional activity (Ref. 18 and L. Pinsky, unpublished results). The AR LBDmut had no transcriptional activity or dominant negative activity. We also tested the dominant negative activity of these mutants on GR-mediated transactivation (Fig.

These studies describe several novel features of AR function. First, we have demonstrated the critical role of an amino-terminal subregion for ligand-dependent transactivation (amino acids 147–296). In contrast to other steroid hormone receptors (3, 4), this AF-1 domain appears to be a major contributor to ligandmediated transcription. Second, we have generated potent dominant negative inhibitors of AR, GR, and PR function when this subregion is deleted. Moreover, WT AR had dominant negative activity on GR function. Given the observation that the PR isoform, PR-2, has potent dominant negative activity on GR, AR, and PR-1 activity (Refs. 19–21 and P. M. Yen, unpublished results), these findings suggest that coexpression of these steroid hormone receptors not only may allow multiple ligands to regulate target gene expression but also may enable mutual modulation of target gene transcription by these receptors. In general, when two steroid hormone receptors are coexpressed in saturating amounts and both ligands are present, the resultant transcriptional activity will approach the transcriptional activity of the less potent receptor. Recently, putative coactivators for nuclear hormone receptors called SRC-1 and TIF2 have been identified (22, 23). Interestingly, SRC-1 augments transcriptional activation by PR-1, but had minimal effects on GR, suggesting that some coactivators for steroid hormone receptors may be receptor-specific. Differential expression of receptor-specific coactivators in tissues and cells could contribute to differences in transcriptional activation observed among different steroid hormone receptors in certain target genes. Our observation of different levels of maximal transcriptional activation by steroid hormone receptors on the MMTV promoter in CV-1 cells (GR⬎PR-1⬎AR⬎PR-2) would be consistent with this model. Several lines of evidence suggest that the mechanism of dominant negative activity by AR and AR mutants likely involves competitive DNA binding to the composite response element. First, our studies with AR truncation and deletion mutants show that deletion

MOL ENDO · 1997 166

Vol 11 No. 2

Fig. 4. Dose-Response of WT AR and ⌬46-408 Dominant Negative Activity on GR-Mediated Transcription CV-1 cells were transfected with 0.1 ␮g expression vector encoding GR and 0, 0.1, 0.3, or 0.5 ␮g WT AR or ⌬46-408, 1.7 ␮g MMTV-luciferase reporter, and 1.0 ␮g RSV-␤ galactosidase control vector. In some samples, pSG vector was added so that each sample contained the same amount of DNA. Cells were treated with 10⫺6 M testosterone and dexamethasone as indicated for 48 h and harvested, and cell lysates were prepared and analyzed for luciferase activity as in Fig. 1. Data are expressed as % maximal activity with GR transcriptional activity in the presence of dexamethasone ⫽ 100%. Data are expressed as means ⫾ SD (n ⫽ 3). Asterisks refer to significant differences from GR values (P ⬍ 0.01). A, Dose-response with WT AR; B, Dose-response with ⌬46-408.


167

Fig. 5. Dominant Negative Activity by DNA-Binding Mutant of ⌬46-408 (⌬46-408 C562G) on GR-Mediated Transcription CV-1 cells were transfected with 0.1 ␮g expression vector encoding GR and 0.3 ␮g WT AR, ⌬46-408, or ⌬46-408 C562G, 1.7 ␮g MMTV-luciferase or TAT-CAT reporter, and 1.0 ␮g RSV-␤ galactosidase control vector. In some samples, pSG vector was added so that each sample contained the same amount of DNA. Cells were treated with 10⫺6 M dexamethasone and/or testosterone as indicated for 48 h and harvested, and cell lysates were prepared and analyzed for luciferase activity as in Fig. 1 and CAT activity in Materials and Methods. Data are expressed as means ⫾ SD (n ⫽ 3). A, MMTV-luciferase reporter study. B, TAT-CAT reporter study.

MOL ENDO · 1997 168

Vol 11 No. 2

Fig. 6. Transcriptional Activity and Dominant Negative Activity of Natural AR Mutants A, Transcriptional activity of AR mutants. CV-1 cells were transfected with indicated amounts of hAR, DBDmut, and LBDmut in pRSV, with 1.7 ␮g MMTV-luciferase reporter, and 1.0 ␮g RSV-␤ galactosidase control vector. Cells were treated ⫺/⫹ 10⫺6 M testosterone for 48 h, harvested, and analyzed for luciferase activity as in Fig. 4. B, Dominant negative activity by AR mutants on GR-mediated transcriptional activity. CV-1 cells were transfected with 0.1 ␮g hGR and 0.5 ␮g hAR, DBDmut, and LBDmut in pRSV, with 1.7 ␮g MMTV-luciferase reporter, and 1.0 ␮g RSV-␤ galactosidase control vector and treated with 10⫺6 M dexamethasone and/or testosterone. pRSV vector was added to some samples so that each sample contained the same amount of DNA. Samples were harvested, assayed, and analyzed as in Fig. 4. For AR, 100% maximal activity ⫽ transcriptional activity with 0.1 ␮g hAR expression vector in the presence of 10⫺6 M testosterone. For GR, 100% maximal activity is defined the same as in Fig. 4. Data are expressed as means ⫾ SD (n ⫽ 3).

of the DNA-binding domain greatly weakens dominant negative activity on AR- and PR- mediated transactivation and weakens dominant negative activity on GRmediated transactivation. Second, a point mutation of the DBD of the potent dominant negative inhibitor,

⌬46-408, abrogates DNA binding and its effect on GR-mediated transcription. Third, our results using a natural AR mutation, which only contains a single amino acid deletion in the DNA-binding domain, further demonstrates the importance of the integrity of


Table 1. Whole Cell Ligand-Binding Studies of CV-1 Cells Cotransfected with AR, GR, and AR ⫹ GR Expression Plasmids Receptor

Ligand

Receptors/Cell

Concentration (fmol/mg)

AR alone AR ⫹ GR GR alone AR ⫹ GR

Mib Mib Dex Dex

8637 ⫾ 2293 8547 ⫾ 1364 14553 ⫾ 633 13002 ⫾ 2400

43 ⫾ 11 42 ⫾ 7 71 ⫾ 3 64 ⫾ 12

Expression plasmids (7.5 ␮g) (pSG-hAR, pSG-rAR, or pSGhAR ⫹ pSG-rAR) were cotransfected into confluent 10-cm plates containing CV-1 cells. Total amount of DNA per dish was adjusted to 15 ␮g by adding empty pSG5 vector when appropriate. [3H]Dexamethasone and [3H]mibolerone binding in whole cell lysates were prepared and measured as described previously (10). Triplicate samples were used for each experimental condition, and similar results were obtained in two other experiments. The data are expressed as mean values ⫾ SE (n ⫽ 3). Mib, Mibolerone; Dex, dexamethasone.

the DNA-binding domain on mediating dominant negative activity of GR-mediated transactivation. Fourth, we have observed only weak dominant negative activity by AR or AR ⌬46-408 on thyroid hormone receptor- or estrogen receptor-mediated transactivation (P. M. Yen and Y. Liu, unpublished results), suggesting that squelching of a common coactivator such as SRC-1 is unlikely. Last, WT AR, AR ⌬46-408, and GR can bind to the TAT HRE with similar affinities (9, 24–26), suggesting that they may have similar DNAbinding properties for the MMTV and TAT HREs, and thus could serve as competitors for DNA binding. Taken together, the foregoing data favor a model in which transcriptionally less active AR or inactive mutant ARs compete with GR and PR-1 for binding to HREs. Titration of a common cofactor that interacts with the DNA-binding domains of AR and GR is less likely based on these experiments but cannot be excluded. Additionally, it is possible that AR and AR mutants may form heterodimers with GR or PR-1 similar to AR/⌬46-408 dimers, mineralocorticoid receptor/GR dimers, and PR DBD/GR DBD dimers reported previously (9, 10, 27, 28). Interestingly, Bamberger et al. (29) recently showed that an alternative splice variant of GR, GR␤, which does not bind dexamethasone and has dominant negative activity on wild type GR transcriptional activity, may utilize a similar mechanism by competing with wild type GR for binding to the HRE. However, it should be noted that we have not observed AR/PR-1 heterodimer formation in solution or on DNA in coimmunoprecipitation experiments (J. Whang and P. M. Yen, unpublished results). It also is interesting that AR and AR mutants had dominant negative activity on GR- and PR-mediated transcription in the absence or presence of testosterone. Control experiments showed that 10⫺6 M dexamethasone and progesterone did not stimulate transcriptional activity by AR (Y. Liu and P. M. Yen, unpublished results), suggesting that these hormones

169

did not account for testosterone-independent dominant negative activity by AR. Immunostaining experiments have shown that unliganded rat AR and human AR are localized in the nuclei of transfected cells (Ref. 12 and U. Karvonen, P. J. Kallio, O. A. Ja¨nne, and J. J. Palvimo, manuscript submitted). Additonally, it has been shown that unliganded AR can bind to AREs in vitro with similar kinetics and affinity as liganded AR (9, 10). Ikonen et al. (30) also showed that unliganded AR expressed in COS cells also could bind to an ARE in vitro. Given these findings, it is possible that both unliganded and liganded AR compete for DNA binding to HREs in these cotransfection studies. Of note, we observed testosterone-dependent effects on AR and ⌬46-408 dominant negative activity when lower amounts of expression plasmid were used (Fig. 4), suggesting that at lower receptor concentrations, unliganded AR may be complexed with endogenous heat shock proteins and unable to bind DNA. We also examined AR- and GR-mediated transcription using a reporter plasmid containing the TAT promoter and upstream sequences. These studies showed that AR, ⌬46-408, and ⌬46-408 C562G had similar effects on GR-mediated transcription as observed with the MMTV-luciferase reporter. These findings suggest that cross-talk between AR and GR may occur on different target genes that utilize different promoters. Additionally, whole cell ligand-binding studies showed that AR and GR are similarly expressed in cotransfection studies of CV-1 cells. These findings then would suggest that AR modulation of GR activity can occur at relatively low amounts of receptors per cell and does not require excessive amounts of AR expression. The AR and GR expression level is similar to endogenous receptor levels in cells and tissues in which both receptors are coexpressed (31–35). Another potential consequence for this cross-talk between AR and GR is that patients with androgen resistance may be more sensitive to glucocorticoids in certain tissues than normal patients due to lack of dominant negative activity by mutant ARs. Studies of urinary free cortisol collections or graded dexamethasone suppression tests, in patients with androgen resistance, or target gene responsiveness to dexamethasone in Tfm mice may shed light on the physiological significance of this cross-talk (36). Our functional studies thus raise the possibility that mutations in one receptor may modulate the function(s) of another receptor when there is cross-talk between the receptors, particularly in tissues in which AR and GR are coexpressed (e.g. skin, bone, and prostate). The foregoing studies, as well as the development of estrogen receptor mutations that have dominant negative activity (37), raise the possibility of using mutant steroid hormone receptors in transgenic animals to study the effects of steroid hormone receptors in growth, development, metabolism, and oncogenesis. These receptors could even be targeted to specific tissues, as has been recently shown with dominant negative retinoic acid receptors (38, 39), and provide

MOL ENDO · 1997 170

an alternative to knockout mice for studying loss of steroid hormone receptor function. However, careful studies of potential cross-talk among steroid hormone receptors need to be performed before such models can be used. Finally, cross-talk among related receptors in regulating target gene expression via composite HREs may be yet another mechanism for fine control of gene expression as shown here as well as by that recently demonstrated in the regulation of the proliferin gene by GR and mineralocorticoid receptor (40). This cross-talk may modulate ligand-dependent transcription in tissues where AR, GR, and PR-1 are coexpressed. Understanding the interplay among these different receptors should prove to be an exciting and interesting area for future studies.

MATERIALS AND METHODS

Vol 11 No. 2

Acknowledgments The authors would like to thank Dr. Pierre Chambon (INSERM, Strasbourg, France) for the GR, PR-1, and PR-2 expression plasmids, Dr. Ronald Evans (Salk Institute, La Jolla, CA) for the pMTV-luc reporter plasmid, and Dr. S. Stoney Simons, Jr. (NIH, Bethesda, MD) for the TAT-CAT reporter plasmid, and Drs. Remco Spanjaard and Anath Shalev (Harvard Medical School, Boston, MA) for helpful discussions.

Received June 25, 1996. Revision received October 31, 1996. Accepted November 11, 1996. Address requests for reprints to: Paul M. Yen, G.W. Thorn Research Building, Room 907, Brigham and Women’s Hospital, 20 Shattuck Street, Boston, Massachusetts 02115. This work was supported by NIH Grant K080K02186 and a Clinical Research Grant from The March of Dimes Foundation (P.M.Y.) and The Medical Research Council of The Academy of Finland, The Finnish Foundation for Cancer, and The University of Helsinki (J.J.P. and O.A.J.), and grants from Medical Research Council, Canada, and FRSQ, Que´bec (M.I.T. and L.P.).

Preparation of Vectors Previously constructed pSG expression vectors encoding rat AR and various deletions mutants (⌬40-147, ⌬38-296, ⌬46408, ⌬557-610, ⌬788-902, and ⌬46-408 C562G) (10, 41) and pSV40 expression vectors encoding human AR and natural AR mutants (DBDmut Phe at amino acid position 582 was deleted, and LBDmut containing an amino acid substitution of Arg to Cys at amino acid position 773) (16, 17) and human GR, PR-1, and PR-2 in pSG (Dr. P. Chambon, INSERM, Strasbourg, France) were used (42). The pRSV-GR and MMTV-luciferase reporter plasmid containing the MMTV promoter region and luciferase cDNA (pMTV-luc) (Dr. R. Evans, Salk Institute, La Jolla, CA) (26) and the TAT-chloramphenicol acetyltransferase (CAT) reporter plasmid containing the minimal TAT promoter and 3.0 kb upstream sequences also were used in some experiments (Dr. S. Stoney Simons, Jr., NIH, Bethesda, MD) (43). Clones were isolated, sequenced, prepared, and purified by affinity chromatography (QIAGEN, Chatsworth, CA) before being used in transfections. Cotransfection Studies CV-1 cells were grown in DMEM/10% FCS. The serum was stripped of steroid hormones by incubating with charcoal for 12 h at 4 C and constant mixing with 5% (wt/vol) AG1-X8 resin (Bio-Rad, Richmond, CA) twice for 12 h at 4 C before ultrafiltration. The cells were transfected with expression (0.1 ␮g) and reporter (1.7 ␮g) plasmids as well as a RSV-␤galactosidase control plasmid (1 ␮g) (44) in 3.5-cm plates using the calcium-phosphate precipitation method (45). Cells were grown for 48 h in the absence or presence of 10⫺6 M dexamethasone, progesterone, or testosterone (Sigma), and harvested. Cell extracts were analyzed for both luciferase (46) and ␤-galactosidase (44) activities to correct for transfection efficiency. Except where indicated, the corrected luciferase activities of untreated samples were normalized to the luciferase activities of samples containing vector alone in the absence of ligand (1-fold basal). In experiments studying TAT-CAT reporter activity, chloramphenicol acetyltransferase enzyme was measured using a CAT ELISA kit (BoehringerMannheim, Mannheim, Germany). For the whole cell ligand-binding studies, 7.5 ␮g expression plasmid (pSG-hGR, pSG-rAR) were cotransfected into confluent 10-cm plates containing CV-1 cells. [3H]Dexamethasone and [3H]mibolerone binding in whole cells was prepared and measured as described previously (10).

REFERENCES 1. Majors J, Varmus HE 1983 A small region of mouse mammary tumor virus long terminal repeat confers glucocorticoid hormone regulation on a linked heterologous gene. Proc Natl Acad Sci USA 80:5866–5870 2. Chandler VL, Maler BA, Yamamoto KR 1983 DNA sequences bound specifically glucocorticoid receptor in vitro render a heterologous promoter hormone responsive in vivo. Cell 33:489–499 3. Gronemeyer H 1991 Transcription activation by estrogen and progesterone receptors. Annu Rev Genet 25:89–123 4. Truss M, Beato M 1993 Steroid hormone receptors: interaction with deoxyribonucleic acid and transcription factors. Endocr Rev 14:459–479 5. Ham J, Thomson A, Needham M, Webb P, Parker M 1988 Characterization of response elements for androgens, glucocorticoids, and progestins in mouse mammary tumour virus. Nucleic Acids Res 16:5263–5277 6. Ahe D, Janich S, Scheidereit C, Renkawitz R, Schutz G, Beato M 1985 Glucocorticoid and progesterone receptors bind to the same sites in two hormonally regulated promoters. Nature 313:706–709 7. Cairns C, Gustafsson J-A, Carlstedt-Duke J 1991 Identification of protein contact sites within the glucocorticoid/progestin response element. Mol Endocrinol 5:598–604 8. Truss M, Chalepakis G, Beato M 1990 Contacts between steroid hormone receptors and thymines in DNA: an interference method. Proc Natl Acad Sci USA 87:7180–7184 9. Kallio PJ, Palvimo JJ, Mehto M, Ja¨nne OA 1994 Analysis of androgen receptor-DNA interactions with receptor proteins produced in insect cells. J Biol Chem 269: 11514–11522 10. Palvimo JJ, Kallio PJ, Ikonen T, Mehto M, Ja¨nne O 1993 Dominant negative regulation of trans-activation by the rat androgen receptor: roles of the N-terminal domain and heterodimer formation. Mol Endocrinol 7:1399–1407 11. Jenster G, van der Korput HA, Trapman J, Brinkman AO 1995 Identification of two transcription activation units in the N-terminal domain of the human androgen receptor. J Biol Chem 270:7341–7346 12. Jenster G, van der Korput HA, van Vroonhoven C, van der Kwast TH, Trapman J, Brinkman AO 1991 Domains of the human androgen receptor involved in steroid bind-


13.

14.

15.

16.

17.

18.

19.

20.

21. 22. 23.

24.

25.

26.

27. 28.

29.

ing, transcriptional activation, and subcellular localization. Mol Endocrinol 5:1396–1404 Nagaya T, Madison LD, Jameson JL 1992 Thyroid hormone receptor mutants that cause resistance to thyroid hormone: evidence for receptor competition for DNA sequences in target genes. J Biol Chem 27:13014–13019 Yen PM, Wilcox EC, Hayashi Y, Refetoff S, Chin WW 1995 Studies on the repression of basal transcription (silencing) by artificial and natural human TR␤ mutants. Endocrinology 136:2845–2851 Yen PM, Liu Y, Sugawara A, Chin WW 1996 Vitamin D receptors repress basal transcription and exert dominant negative activity on triiodothyronine-mediated transcriptional activity. J Biol Chem 271:10910–10916 Beitel LK, Prior L, Vasilou DM, Gottlieb B, Kaufman M, Lambroso R, Alvarado C, McGillivray B, Trifiro MA, Pinsky L 1994 Complete androgen insensitivity due to mutations in the probable alpha-helical segments of the DNA-binding domain in the human androgen receptors. Hum Mol Genet 3:21–27 Prior L, Bordet S, Trifiro MA, Mhatre A, Kaufman M, Pinsky L, Wrogeman K, Belsham DD, Pereira F, Greenberg C 1992 Replacement of arginine 773 by cysteine or histidine in the human androgen receptor causes complete androgen insensitivity with different receptor phenotypes. Am J Hum Genet 51:143–55 Kazemi-Esfariani P, Beitel LK, Kaufman M, Gottlieb B, Alvarado C, Trifiro M, Pinsky L, Evidence for transcriptional superactivity of human steroid receptors by interaction with DNA-binding-deficient androgen receptors. Program of the 76th Annual Meeting of The Endocrine Society, Anaheim, CA, 1994 (Abstract 738) Vegeto E, Shahbaz MM, Wen DX, Goldman ME, O’Malley BW, McDonnell DP 1993 Human progesterone receptor A form is a cell- and promoter-specific repressor of human progesterone receptor B function. Mol Endocrinol 7:1244–1255 Tung L, Mohamed MK, Hoeffler JP, Takamoto GS, Horwitz KB 1993 Antagonist-occupied human progesterone B-receptors activation transcription without binding to progesterone response elements and are dominantly inhibited by A-receptors. Mol Endocrinol 7:1256–1259 McDonnell DP, Goldman ME 1994 RU486 exerts antiestrogenic activities through a novel progesterone receptor A form-mediated mechanism. J Biol Chem 269:11945–11949 Onate SA, Tsai SY, Tsai M-J, O’Malley BW 1995 Sequence and characterization of a coactivator for the steroid hormone receptor superfamily. Science 270:1354–1357 Voegel JJ, Heine MJS, Zechel C, Chambon P, Gronemeyer H 1996 TIF2, a 160 kD transcriptional mediator for the ligand-dependent actiation AF-2 of nuclear receptors. EMBO J 15:3667–3675 Rodriguez R, Carson MA, Weigel NL, O’Malley BW, Schrader WT 1989 Hormone-induced changes in the in vitro DNA-binding activity of the chicken progesterone receptor. Mol Endocrinol 3:356–362 Cairns C, Gustafssson J-A, Carlstedt-Duke J 1991 Identification of protein contact sites within the glucocorticoid/progestin response element. Mol Endocrinol 5:598–604 Dahlman-Wright K, Wright APH, Gustafsson J-A 1992 Determinants of high-affinity DNA binding by the glucocorticoid receptor: evaluation of receptor domains outside the DNA-binding domain. Biochemistry 31:9040–9044 Liu W, Wang J, Sauter NK, Pearce D Steroid receptor heterodimerization demonstrated in vitro and in vivo. Proc Natl Acad Sci USA 92:12480–12484 Trapp T, Rupprecht R, Castren M, Reul JMHM, Holsboer F 1994 Heterodimerization between mineralocorticoid and glucocoricoid receptor: a new principle of glucocorticoid acton in the CNS. Neuron 13:1457–1462 Bamberger CM, Bamberger AM, deCastro M, Chrousos GP 1995 Glucocorticoid receptor beta, a potential en-

171

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

dogenous inhibitor of glucocorticoid action in humans. J Clin Invest 95:2435–2541 Ikonen T, Palvimo JJ, Kallio PJ, Renkainen P, Janne OA 1994 Stimulation of androgen-regulated transactivation by modulators of protein phosphorylation. Endocrinology 135:1359–1366 Horwitz KB, Costlow ME, McGuire WL 1975 MCF-7: a human breast cancer cell line with estrogen, androgen, progesterone, and glucocorticoid receptors. Steroids 26:786–795 Norris JS, Kohler PO 1977 The coexistence of androgen and glucocorticoid receptors in the DDT1 cloned cell line. Endocrinology 100:613–618 Snochowski M, Dahlberg E, Gustafsson JA 1980 Characterization and quantification of the androgen and glucocorticoid receptors in cytosol from rat skeletal muscle. Eur J Biochem 111:603–616 Davies P, Rushmere NK 1990 Association of glucocorticoid receptors with prostate nuclear sites for androgen receptors and with androgen response elements. J Mol Endocrinol 5:117–127 Masuyama A, Ouchi Y, Sato F, Hosoi T, Nakamura T, Orimo H 1992 Characteristics of steroid hormone receptors in cultured MC3T3-El osteoblastic cells and effect of steroid hormones on cell proliferation. Calcif Tissue Int 51:376–381 Yarbrough WG, Quarmby VE, Simental JA, Joseph DR, Sar M, Lubahn DB, Olsen KL, French FS, Wilson EM 1990 A single base mutation in the androgen receptor gene cause androgen insensitivity in the testicular feminized rat. J Biol Chem 265:8893–8900 Ince BA, Schodin DJ, Shapiro DJ, Katzenellenbogen BS 1995 Repression of endogenous estrogen receptor activity in MCF-7 human breast cancer cells by dominant negative estrogen receptors. Endocrinology 136:3194–3199 Saitou M, Sugal S, Tanaka T, Shimouchi K, Fuchs E, Narumiya S, Kakizuka A 1995 Inhibition of skin development by targeted expression of a dominant-negative retinoic acid receptor. Nature 374:159–162 Imakado S, Bickenbach JR, Bundman DS, Rothnagel JA, Attar PS, Wang X-J, Walczak VR, Wisniewski S, Pote J, Gordon JS, Heyman RA, Evans RM, Roop DS 1995 Targeting expression of a dominant-negative retinoic acid receptor mutant in the epidermis of transgenic mice results in loss of barrier function. Genes Dev 9:317–329 Pearce D, Yamamoto KR 1993 Mineralocorticoid and glucocorticoid receptor activities distinguished by nonreceptor factors at a composite response element. Science 259:1161–1165 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 Kastner P, Bocquel M-T, Turcotte B, Garnier J-M, Horwitz KB, Chambon P, Gronemeyer H 1990 Transient expression of human and chicken progesterone receptors does not support alternative translational initiation from a single mRNA as the mechanism generating two receptor isoforms. J Biol Chem 265:12163–12167 Oshima H, Simons Jr SS 1992 Modulation of transcription factor activity by a distant steroid modulatory element. Mol Endocrinol 6:416–428 Edlund T, Walker MD, Barr PJ, Rutter WJ 1985 Cell specific expression of the rat insulin genes: evidence for role of two distinct 5⬘ flanking sequences. Science 230:912–916 Sambrook J, Fritsch EF, Maniatis T 1989 Molecular Cloning: A Laboratory Manual, ed. 2. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY DeWet JR, Wood KV, DeLuca M, Helsinki DR, Subramani S 1987 Firefly luciferase gene: structure and expression in mammalian cells. Mol Cell Biol 7:725–737