Androgens, progestins, and glucocorticoids induce follicle-stimulating hormone beta-subunit gene expression at the level of the gonadotrope.

FSH is produced by the pituitary gonadotrope to regulate gametogenesis. Steroid hormones, including androgens, progestins, and glucocorticoids, have all been shown to stimulate expression of the FSHbeta subunit in primary pituitary cells and rodent models. Understanding the molecular mechanisms of steroid induction of FSHbeta has been difficult due to the heterogeneity of the anterior pituitary. Immortalized LbetaT2 cells are a model of a mature gonadotrope cell and express the endogenous steroid receptor for each of the three hormones. Transient transfection of each receptor, along with ligand treatment, stimulates the mouse FSHbeta promoter, but induction is severely diminished using receptors that lack the ability to bind DNA, indicating that induction is likely through direct DNA binding. All three steroid hormones act within the first 500 bp of the FSHbeta promoter where six putative hormone response elements exist. The -381 site is critical for FSHbeta induction by all three steroid hormones, whereas the -197 and -139 sites contribute to maximal induction. Interestingly, the -273 and -230 sites are also necessary for androgen and progestin induction of FSHbeta, but not for glucocorticoid induction. Additionally, we find that all three receptors bind the endogenous FSHbeta promoter, in vivo, and specifically bind the -381 site in vitro, suggesting that the binding of the receptors to this element is critical for the induction of FSHbeta by these 3-keto steroid hormones. Our data indicate that androgens, glucocorticoids, and progestins act via their receptors to directly activate FSHbeta gene expression in the pituitary gonadotrope.

F SH AND LH are produced in the gonadotrope cells of the anterior pituitary. These hormones act on the gonads to regulate critical aspects of reproduction, including steroidogenesis, gametogenesis, and ovulation (1)(2)(3). FSH and LH are heterodimeric glycoproteins composed of a shared ␣-subunit and a unique ␤-subunit (4). The ␤-subunit confers the biological specificity of FSH and LH, and synthesis of this subunit appears to be the rate-limiting step for production of FSH and LH (5,6). Both peptide and steroid hormones regulate FSH and LH synthesis to produce the hormonal pattern necessary for the estrous cycle and normal reproductive function. The hypothalamic neuropeptide, GnRH, mediates the synthesis and secretion of both FSH and LH (7). In addition, the modulation of FSH␤ gene expression by the activin/inhibin/ follistatin system has been well characterized (8,9). Recent data show that activin can also regulate LH␤ gene expression (10)(11)(12)(13)(14).
A fundamental concept in the study of reproduction is the regulation of the hypothalamic-pituitary-gonadal axis by steroid hormone feedback. In addition to exerting effects at the level of the hypothalamus, steroids control the transcription of FSH␤ and LH␤ subunit genes at the level of the pituitary, although their mechanisms of action have not been as well characterized as GnRH and activin. More specifically, androgens, progestins, and glucocorticoids have all been shown to induce FSH␤ gene expression in the anterior pituitary [recently reviewed by Burger et al. (15)]. Although androgens and progestins are produced by the gonads, and glucocorticoids are produced by the adrenal glands, all of these 3-keto steroids possess a similar chemical structure and mechanism of action. Moreover, all three of their respective receptors, the androgen receptor (AR), progesterone receptor (PR), and glucocorticoid receptor (GR), are related members of the class I steroid receptor family (16,17) and bind DNA directly at response elements containing the halfsite TGTTCT organized as 15-bp inverted repeats with 3-bp spacers (18,19). In the classical mechanism of action, the receptors homodimerize upon ligand binding and bind their response elements causing trans-activation of specific genes (20). More recently, it has been recognized that the steroid receptors can also function in a nonclassical manner to modulate transcription through indirect DNA binding via interactions with other transcription factors such as activation protein 1 (21)(22)(23).
Several lines of evidence indicate that androgens regulate the levels of FSH␤ and LH␤ mRNAs in the anterior pituitary. Studies in castrated, GnRH antagonist-treated rats have shown a selective increase in FSH␤ mRNA upon treatment with testosterone (24)(25)(26)(27). Subsequent studies in primary rat pituitary cells confirmed that the activation of FSH␤ expression by testosterone occurs at the level of the pituitary (28)(29)(30). In contrast to FSH␤, androgens repressed LH␤-subunit gene expression in both castrated, GnRH antagonist-treated rats (26,27) and primary pituitary cell culture (29). Additionally, the effects of androgens on transcription of the LH␤ subunit have been studied in gonadotrope-derived immortalized cell lines. Jorgensen and Nilson (31) found that androgens suppress bovine LH␤ subunit gene expression through protein-protein interaction between ligand-bound AR and steroidogenic factor 1. Using a Ϫ617/ϩ44 rat LH␤ promoter, Curtin et al. (32) determined that androgen treatment suppressed GnRH-induced LH␤ subunit gene transcription as well. However, they found that the suppression occurred primarily as a consequence of direct protein-protein interaction between AR and Sp1.
Progesterone also appears to mediate FSH␤ gene expression at the level of the pituitary. FSH␤ mRNA levels were increased in rats treated with estrogen and progesterone (33). The presence of estrogen and progesterone also further augmented the increase in FSH␤ mRNA expression in response to GnRH stimulation (34). Furthermore, antiprogestins blocked FSH secretion and mRNA expression during the preovulatory FSH surge (35) as well as blocking the secondary FSH surge (36,37). Both the rat and ovine FSH␤ promoters have been examined for progesterone responsiveness. Reporter genes containing either the proximal rat or ovine FSH␤ promoter responded to progestin treatment in primary rat or ovine pituitary cultures, respectively (38,39). Within the ovine FSH␤ promoter, six progesterone response elements (PREs) were shown to bind PR (38), whereas three PREs bound PR in the rat promoter (39), although the functional role of the individual elements was not determined. Less is known about the effects of progestins on LH␤ mRNA expression, although no effect was reported on LH␤ mRNA levels in ovariectomized rats treated with estrogen and progestin (34) or on LH␤ mRNA levels in rat primary pituitary cells treated with both estrogen and progestin (40).
Similar to androgens and progestins, glucocorticoids affect FSH␤ expression at the level of the pituitary. Several studies have demonstrated a selective increase in FSH␤ gene expression in response to glucocorticoids in both intact animals (41,42) as well as in primary pituitary cells (30,43,44). The effect of the glucocorticoids on FSH␤ appears to be at the level of transcription as no effect on FSH␤ mRNA half-life was detected (43). Unlike FSH␤, a direct effect of glucocorticoids on LH␤ expression has not been observed (41)(42)(43), although Rosen et al. (45) did demonstrate that LH␤ induction by GnRH was inhibited by glucocorticoids.
Given the importance of androgens, progestins, and glucocorticoids in the regulation of FSH␤ gene expression, the purpose of this study was to investigate the molecular mechanisms of this regulation at the level of the gonadotrope. The identification of PREs in the proximal rat and ovine FSH␤ promoters almost a decade ago by O'Conner et al. (39) and Webster et al. (38) suggested that there could be a direct mechanism of action. Additionally, because PR contains a DNAbinding domain (DBD) that is highly conserved with other members of the steroid receptor superfamily, including AR and GR, these researchers hypothesized that all of these receptors could bind to the FSH␤ promoter and mediate its transcription. However, due to the fact that gonadotropes constitute only about 10% of the total secretory cells in the anterior pituitary (46) and that a majority of these cells express steroid receptors (47)(48)(49)(50), the mechanism of action awaited the availability of an appropriate gonadotrope cell model system.
The development of the L␤T2 gonadotrope-derived immortalized cell line provided a model system in which to study gonadotropin gene expression in a pure population of gonadotrope cells. The L␤T2 cell line endogenously expresses many markers of a mature gonadotrope including FSH␤, LH␤, GnRH receptor (GnRH-R), activin, activin receptor, follistatin, and inhibin (13,51,52). It has also been shown to endogenously express the estrogen receptor (ER) (53) and AR (54). These properties make the L␤T2 cell line an excellent model system for directly studying the regulation of gonadotropin gene expression by steroids.
Previously, L␤T2 cells have been used to study how androgens repress the induction of LH␤ gene transcription by GnRH (31,32). More recently, Spady et al. (55) found that an ovine FSH␤ reporter gene was activated in response to testosterone in the L␤T2 cell line. In the current study, we used the L␤T2 cell line to determine whether steroid hormones directly regulate the expression of the murine FSH␤ gene at the level of the gonadotrope and whether this regulation occurs through binding of the steroid receptor to the FSH␤ promoter. In particular, we used transient transfections with a FSH␤-luciferase reporter gene to examine the responsiveness of the proximal murine promoter to steroid regulation and to analyze the role of putative hormone response elements (HREs). We also used chromatin immunoprecipitation (ChIP) and gel-shift analysis to determine whether the steroid hormone receptors can bind to the proximal mouse FSH␤ promoter in vivo and in vitro. Overall, we show that steroid hormone regulation of the FSH␤ promoter can occur at the level of the gonadotrope through direct DNA binding of these steroid receptors to specific HREs.

Androgens, Progestins, and Glucocorticoids Mediate Transcription of the Murine FSH␤ Gene
Gonadotrope cells in the anterior pituitary of rodents have been shown to express the receptors for androgens, progestins, and glucocorticoids (56)(57)(58)(59), and the L␤T2 cell line has been shown previously to express AR (54). We tested whether the L␤T2 cells also express PR and GR to determine whether these cells constitute a relevant model system for studying the mechanisms of steroid responsiveness. Oligonucleotide primers encompassing 550 bp of AR, 406 bp of PR, and 298 bp of GR were used to amplify the respective receptors from cDNA generated by reverse transcription of mRNA from the L␤T2 cell line. All three receptors were amplified by PCR at the expected sizes, and no bands were seen in the control lanes that lacked reverse transcriptase (Fig. 1A). Next, ChIP with antibodies specific to AR, PR, and GR was used to determine whether the endogenous steroid receptors could bind the mouse FSH␤ promoter in vivo in L␤T2 cells. The results shown in the upper panel of Fig. 1B demonstrate that all three ligand-bound receptors bind to the endogenous FSH␤ promoter (Fig. 1B, upper panel; lanes 3-5). The strong signal for PR in the ChIP assay does not necessarily indicate that PR is more abundant than AR or GR in the L␤T2 cells because quantitative PCR was not performed. Instead, it may reflect the affinity of the PR antibody for the receptor. In contrast, there was no precipitation of FSH␤ promoter DNA with a nonspecific mouse IgG control (lane 2). The FSH␤ promoter was also amplified from the input chromatin (lane 1) as a positive control for genomic DNA preparation and PCR conditions. As a control for specificity, primers encompassing part of the downstream coding region of the FSH␤ gene were also used in PCR (Fig. 1B, lower panel). Although these primers amplified FSH␤ from the input chromatin as expected (lane 1), no bands were amplified from the precipitated DNA (lanes 2-5).
To monitor changes in promoter activity upon steroid treatment, L␤T2 cells were transiently transfected with the proximal 1000 bp of the mouse FSH␤ 5Јregulatory region linked to a luciferase reporter gene (Ϫ1000FSH␤luc). Because the L␤T2 cells are murine in origin, we investigated whether androgen regulation occurs in a similar manner on the murine promoter as it does on the ovine promoter (55). In previous studies, androgen induction of the ovine FSH␤ promoter in the presence of endogenous AR resulted in a modest stimulation of transcription (1.5-to 1.9-fold) (55). Given this relatively weak induction, we amplified the steroid hormone response on the FSH␤ promoter by transfecting the cells with 200 ng of rat AR.
Treatment of the cells with 100 nM testosterone for 24 h resulted in a 9-fold induction of the mouse FSH␤ promoter (Fig. 1C). Dihydrotestosterone (DHT) was also tested to determine whether it could induce FSH␤ because DHT cannot be aromatized to estrogen. Treatment with 100 nM DHT activated FSH␤ to a similar extent as testosterone, implying that estrogenic activity does not play a role (Fig. 1C). To determine whether other 3-keto steroid hormones such as progestins or glucocorticoids could also regulate murine FSH␤ gene expression, the cells were transfected with 200 ng of rat PR B or GR. Interestingly, progesterone and corticosterone also activated FSH␤ transcription. Treatment of 100 nM progesterone resulted in a 27fold induction, whereas 100 nM corticosterone activated FSH␤ 4-fold (Fig. 1C).
To further ascertain whether estrogen can modulate transcription of FSH␤-subunit gene expression, the L␤T2 cells were transfected with ER␣ or ER␤ and treated with 100 nM 17␤-estradiol for 24 h. Estrogen did not stimulate transcription of the Ϫ1000FSH␤luc reporter gene compared with the vehicle control in the presence of either ER␣ or ER␤ under these conditions (Fig. 1C). Moreover, estrogen did not positively regulate FSH␤ gene expression in the presence of both ERs (data not shown). As a control, a luciferase reporter gene driven by a consensus estrogen response element inserted upstream of a minimal Herpes virus thymidine kinase promoter was induced under these conditions (data not shown). Our results agree with previous experiments in rats in which estrogen did not alter the expression of FSH␤ mRNA in ovariectomized rats treated with a GnRH antagonist (60) or in female rat pituitary fragments (61).

Androgens, Progestins, and Glucocorticoids Stimulate FSH␤ Gene Expression in a Receptorand Dose-Dependent Manner
In addition to investigating hormone responsiveness, we examined whether the induction of FSH␤ by androgens, progestins, and glucocorticoids was dependent upon receptor concentration. L␤T2 cells were transfected with increasing concentrations of the receptor and then treated for 24 h with 100 nM of a synthetic analog of the relevant steroid hormones. There was a trend toward a positive induction with the endogenous receptors, and the addition of even 100 ng exogenous AR, PR, or GR all significantly induced FSH␤ gene expression (Fig. 2). The small induction with the endogenous receptors likely reflects titration of the receptors by the transfected reporter genes. For each receptor, the level of FSH␤ induction correlated with the amount of exogenous receptor, and the induction did not reach saturation even with the addition of 800 ng of receptor. These results indicate that the steroid receptors are necessary for transcriptional activation of FSH␤ by androgens, progestins, and glucocorticoids and that the L␤T2 cells are highly responsive to the presence of additional receptor. Furthermore, these data suggest that the steroid responsiveness is not likely due to an artificially high level of steroid receptors because the receptor is limiting in the cells.
To test whether the induction of FSH␤ occurs via a hormone-dependent mechanism, the L␤T2 cells were transfected with AR, PR, or GR and then treated with increasing concentrations of the respective hormone. For each of the three hormones, treatment of the cells with a synthetic steroid analog stimulated FSH␤ promoter activity in a dose-dependent manner (Fig. 3). Treatment with 100 pM of R1881, a synthetic androgen, activated FSH␤ 4-fold, with a maximal induction of 12-fold at the 100 nM dose (Fig. 3A). Treatment with the synthetic progestin, R5020, resulted in activation of the FSH␤ promoter at 100 pM with a maximal induction of 40-fold at 10 nM (Fig. 3B). Similarly, treatment with dexamethasone, a synthetic glucocorticoid, activated FSH␤ with a concentration of 1 nM (3-fold) with saturation occurring at 100 nM (14-fold) (Fig. 3C), demonstrating that the steroid hormone regulation of FSH␤ occurs in a saturable, dose-dependent manner and that physiological concentrations of hormone can activate the FSH␤ gene promoter.

Induction by Steroid Hormone Receptors Is Gene Specific
To determine whether the transcriptional activation by the steroid hormone receptors is specific to the FSH␤ gene, the effects of androgens, progestins, or glucocorticoids on the regulation of the LH␤-subunit gene were examined. In this set of experiments, the proximal 1.8 kb of the rat LH␤ 5Ј-regulatory region linked to a luciferase reporter gene (1.8LH␤luc) was transiently transfected into L␤T2 cells. The cells were also transfected with 200 ng of AR, PR, or GR. Treatment of the cells with 100 nM R1881 repressed the LH␤ promoter by 17% (Fig. 4A) whereas treatment with the progestin, R5020, caused a 48% decrease in LH␤ gene expression (Fig. 4B). Treatment with dexamethasone also repressed LH␤ by 32% (Fig. 4C). Thus, our results demonstrate that ligand-bound AR, PR, or GR can suppress LH␤ gene transcription, whereas they induce FSH␤ mRNA levels. This suggests that androgens, progestins, and glucocorticoids differentially regulate LH␤ and FSH␤ subunit gene expression at the level of the gonadotrope.

Fig. 1. Androgens, Progestins, and Glucocorticoids Induce FSH␤ Gene Expression in Immortalized Gonadotropes
A, RT-PCR was used to detect expression of the steroid receptors in L␤T2 cells. Amplified AR is visible at 550 bp, PR at 406 bp, and GR at 298 bp. In reverse transcription of total RNA isolated from L␤T2 cells, ϩ indicates the presence of reverse transcriptase, and Ϫ indicates no reverse transcriptase. B, ChIP was performed using cross-linked protein/chromatin from L␤T2 cells and antibodies directed against AR, PR, and GR or, as a negative control, against nonspecific IgG. Upper panel, PCR primers encompassing the proximal promoter of FSH␤ were used to detect precipitation of genomic DNA. Lower panel, PCR primers encompassing the downstream FSH␤ coding region were used as a control for specificity. PCR amplification was performed on 0.2% chromatin input (lane 1), and chromatin was precipitated with either mouse IgG (lane 2), AR (lane 3), PR (lane 4), or GR (lane 5) antibodies. C, The Ϫ1000FSH␤luc reporter gene was transiently transfected into L␤T2 cells along with 200 ng of the respective steroid receptor expression vectors. After overnight starvation in serum-free media, the cells were treated with 100 nM testosterone, DHT, progesterone, corticosterone, or 17␤-estradiol for 24 h. Luciferase activity was normalized to ␤-galactosidase activity and set relative to the empty reporter vector. The results represent the mean Ϯ SEM of at least three experiments performed in triplicate and are presented as fold induction of hormone treatment relative to the vehicle control (ethanol). For cells transfected with AR, and treated with testosterone or DHT, * indicates significantly different from the vehicle-treated control, using one-way ANOVA followed by Tukey's post hoc test. For L␤T2 cells transiently transfected with PR, GR, ER␣, or ER␤ and treated for 24 h with progesterone, corticosterone, or 17␤-estradiol, respectively, * indicates significantly different from the respective vehicle-treated control using Student's t test. RT, Reverse transcriptase.

DNA Binding by Steroid Hormone Receptors Is Necessary for Transcriptional Activation of the FSH␤ Gene Promoter
To determine whether direct DNA binding by AR, PR, or GR plays a critical role in the transactivation of FSH␤ by steroid hormones, we transfected L␤T2 cells with steroid receptor mutants deficient in DNA binding. These mutant receptors included an AR C562G mutation in the DNA binding domain (DBD) (62), a PR C577A mutation in the DBD (63), and a GRdim4 mutation containing the dim1 mutation, A458T, which has been previously described (22) as well as three other mutations (N454D, R460D, D462C) that further ensure loss of dimerization ability (21,64), and thereby prevent DNA binding by the mutant receptor. To ensure that the mutant receptors were expressed at similar levels as the wild-type receptors, we overexpressed the mutant receptors in either L␤T2 cells, in the case of mutant AR, or in Cos-1 cells, in the case of mutant PR and GR. For PR and GR, Cos-1 cells were used because overexpression of either wild-type or mutant receptors could not be detected by Western blot in L␤T2 cell extracts, perhaps due to a lack of stability of The Ϫ1000FSH␤luc reporter gene was transiently transfected into L␤T2 cells along with increasing receptor concentrations ranging from 0-800 ng/well, as indicated. After overnight starvation in serum-free media, the cells were treated for 24 h with the indicated hormone. The results represent the mean Ϯ SEM of at least three experiments performed in triplicate and are presented as fold induction relative to the vehicle control. A, L␤T2 cells were transiently transfected with increasing amounts of AR and then treated with 100 nM R1881. B, L␤T2 cells were transiently transfected with increasing amounts of PR and then treated with 100 nM R5020. C, L␤T2 cells were transiently transfected with increasing amounts of GR and then treated with 100 nM dexamethasone (Dex).

Fig. 3. Steroid Hormones Mediate Transcription of FSH␤ in a Dose-Dependent Manner
The Ϫ1000FSH␤luc reporter gene was transiently transfected into L␤T2 cells along with the respective receptor expression vector indicated on the graph. After overnight starvation in serum-free media, the cells were treated for 24 h with the indicated hormone concentrations. The results represent the mean Ϯ SEM of at least three experiments performed in triplicate and are presented as fold induction relative to the vehicle control. A, L␤T2 cells were transiently transfected with the AR expression vector and then treated with R1881 concentrations ranging from 10 pM to 100 nM. B, L␤T2 cells were transiently transfected with the PR expression vector and then treated for 24 h with 10 pM to 100 nM R5020. C, L␤T2 cells were transiently transfected with the GR expression vector and then treated with 100 pM to 1 M dexamethasone (Dex). the full-length receptor or lower transfection efficiency than Cos-1 cells. As demonstrated in Fig. 5, both wild-type and mutant receptors were expressed at similar levels (Fig. 5, A, C, and E, arrowhead). Equal amounts of protein were loaded as indicated by the similar level of expression of nonspecific bands in all three lanes of the gels. Although both wild-type and mutant receptors were expressed at similar levels, we observed no significant induction of the FSH␤ promoter with the AR C562G mutant compared with the induction with the wild-type receptor upon treatment with 100 nM R1881 (Fig. 5B). Activation of the FSH␤ promoter by the PR C577A mutant was also diminished after treatment with 100 nM R5020 (Fig. 5D). Additionally, no FSH␤ induction with the GRdim4 mutation was seen upon treatment with 100 nM dexamethasone (Fig. 5F). These data indicate that AR, PR, and AR all need to bind DNA directly to stimulate transcription of the FSH␤ gene in the presence of the appropriate steroid hormone.

HREs in the Proximal Promoter Are Critical for FSH␤ Induction by Androgens, Progestins, and Glucocorticoids
To map the promoter elements required for steroid hormone responsiveness in the mouse FSH␤ gene promoter, truncation/deletion analysis was used. L␤T2 cells were transiently transfected with truncations of the mouse FSH␤ gene ranging from 1526 to 95 bp upstream of the transcription start site, and the ability of steroids to induce FSH␤ was measured (Fig. 6). Truncation of the promoter to Ϫ95 resulted in a loss of responsiveness for each of the three hormones. Although some response to androgens was found in the upstream regions of the FSH␤ promoter, a considerable degree of responsiveness was conferred in the first 500 bp of the FSH␤ promoter (Fig. 6A). Interestingly, progestins and glucocorticoids did not appear to confer any additional response to FSH␤ after the first 500 bp of the promoter (Fig. 6, B and C). Although differences in steroid responsiveness may exist on the FSH␤ promoter, all three steroid receptors studied can act within the first 500 bp of the FSH␤ promoter, implying that all three regulate the FSH␤ promoter directly at the level of the gonadotrope.
Because the responses to androgens, progestins, and glucocorticoids mapped to a proximal region of the FSH␤ promoter and the responses also required a receptor capable of binding DNA, we examined the mouse promoter sequence in this region for putative HREs. This region contains putative HREs at Ϫ381/ Ϫ367, Ϫ230/Ϫ216, and Ϫ139/Ϫ125 of the mouse promoter (see Fig. 7). These elements contain three or four of the G/C residues that have been shown to be critical for high-affinity binding of steroid receptors to DNA. These elements are also analogous to elements in the proximal rat and ovine FSH␤ promoters that were identified as PR-binding sites, although the functional role of these sites in progesterone responsiveness was not previously determined (38,39). Three additional elements were originally identified in the rat and ovine promoters. These correspond to putative HREs at Ϫ273/Ϫ259, Ϫ197/Ϫ183, and Ϫ175/Ϫ161 of the mouse promoter but are less conserved than the previous three elements.
To investigate the importance of the individual elements, we created mutations in the six putative HREs by site-directed mutagenesis in the context of the Ϫ1000FSH␤luc reporter gene. To disrupt binding of The 1.8LH␤luc reporter gene was transiently transfected into L␤T2 cells along with the indicated receptor expression vector. After overnight starvation in serum-free media, the cells were treated for 24 h with hormone. The results represent the mean Ϯ SEM of at least three experiments performed in triplicate and are presented as fold induction relative to the vehicle control. *, Indicates that the hormone treatment was significantly different from the vehicle-treated control using Student's t test. A, L␤T2 cells were transiently transfected with the AR expression vector and then treated with 100 nM R1881. B, Cells were transiently transfected with the PR expression vector and then treated with 100 nM R5020. C, Cells were transiently transfected with the GR expression vector and then treated with 100 nM dexamethasone (Dex).
the steroid receptors to the putative HREs, mutations were made in the G/C residues critical for high-affinity binding (Fig. 7). Not surprisingly, none of the mutations affected basal expression of the FSH␤ gene (data not shown). Mutation of the Ϫ381 site completely prevented the response to androgen (100 nM R1881), whereas the androgen responsiveness with mutations in the Ϫ273, Ϫ230, Ϫ197, and Ϫ139 HREs was diminished to approximately 46%, 55%, 42%, and 65%, respectively, of the wild-type FSH␤ reporter gene (Fig.  8A). Interestingly, the Ϫ381 and Ϫ197 HRE mutations reduced the response to progesterone (100 nM R5020) to approximately 26% and 29%, respectively, whereas the Ϫ273, Ϫ230, and Ϫ139 HRE mutations decreased the response to progesterone to 55%, 83%, and 76%, respectively, of wild type (Fig. 8B). The Ϫ381 HRE mutation also completely abrogated responsiveness of the FSH␤ reporter gene to glucocorticoid (100 nM dexamethasone). Induction by dexamethasone was inhibited with the Ϫ197 and Ϫ139 HRE mutations to approximately 21% and 59%, respectively, but, in contrast to the other two hormones, no significant inhibition was seen with mutation of the Ϫ273 and Ϫ230 HREs (Fig. 8C). Mutation of the Ϫ175 HRE had no functional effect on the androgen, progestin, or glucocorticoid responsiveness of the FSH␤ Fig. 5. AR, PR, and GR Require DNA Binding to Facilitate FSH␤ Gene Expression A, C, and E, Wild-type and mutant steroid receptors were overexpressed, and protein levels were analyzed by Western blot to ensure that wild-type and mutant receptors were expressed at a similar level. The arrowhead denotes the band specific for the respective steroid receptors, and the nonspecific bands demonstrate equal protein loading. A, WT AR, mutant AR (ARC562G), or empty vector control (pSG5) were overexpressed in L␤T2 cells. B, WT PR, mutant PR (PRC577A) or empty vector control (pCMV5) were overexpressed in Cos-1 cells. C, WT GR, mutant GR (GRdim4) or empty vector control (pSG5) were overexpressed in Cos-1 cells. B, D, and F, Ϫ1000FSH␤luc reporter gene was transiently transfected into L␤T2 cells along with the wild-type or mutant receptor as indicated. After overnight starvation in serum-free media, the cells were treated for 24 h with hormone. The results represent the mean Ϯ SEM of at least three experiments performed in triplicate and are presented as fold induction relative to the vehicle control. *, Indicates that the hormone treatment was significantly different from the vehicle-treated control using Student's t test. B, L␤T2 cells were transiently transfected with the wild-type or mutant AR (ARC562G) expression vector and then treated with 100 nM R1881. D, Cells were transiently transfected with wild-type or mutant PR (PRC577A) expression vector and then treated with 100 nM R5020. F, Cells were transiently transfected with wild-type GR or GRdim4 mutant expression vector and then treated with 100 nM dexamethasone (Dex). WT, Wild type. promoter (Fig. 8). Clearly, the Ϫ381 and Ϫ197 HREs play a prominent role in the response to all three of the steroid hormones. The Ϫ139 element also appears to be necessary for induction by all three hormones and, whereas the Ϫ273 and Ϫ230 elements play a role in androgen and progesterone responsiveness, they do not appear to be necessary for glucocorticoid responsiveness.

AR, PR, and GR Bind to the ؊381 HRE in the Mouse FSH␤ Promoter
We used an EMSA to assess whether AR, GR, and PR can bind to the HREs in the proximal mouse FSH␤ promoter. Although we did not detect steroid receptor binding using L␤T2 nuclear extracts, we observed binding of AR, PR, and GR from baculovirus-infected insect whole-cell extracts to the wild-type Ϫ381 element (Fig. 9, lane 3), whereas no binding to the Ϫ381 HRE mutant was seen (lane 4). For each of the receptors, the resulting complex on the Ϫ381 oligonucleotide was supershifted by a receptor-specific antibody (lane 5) but not by IgG (lane 6). The complexes also The Ϫ1526FSH␤luc, Ϫ1000FSH␤luc, Ϫ500FSH␤luc, or Ϫ95FSH␤luc reporter genes were transiently transfected into L␤T2 cells along with the indicated receptor expression vector. After overnight starvation in serum-free media, the cells were treated for 24 h with hormone. The results represent the mean Ϯ SEM of at least three experiments performed in triplicate and are presented as fold induction relative to the vehicle control. Significantly different levels of hormone induction among truncations are indicated by differing letters, a, b, or c as determined by one-way ANOVA followed by Tukey's post hoc test. A, L␤T2 cells were transiently transfected with the reporter genes and the AR expression vector after which the cells were treated with 100 nM R1881. B, L␤T2 cells were transiently transfected with the reporter genes and the PR expression vector and then the cells were treated with 100 nM R5020. C, L␤T2 cells were transiently transfected with the reporter genes and the GR expression vector after which the cells were treated with 100 nM dexamethasone (Dex).  To determine whether AR, PR, or GR could bind the other functional HREs, the EMSA was repeated with probes encompassing the Ϫ273 and Ϫ197 sites as The wild-type Ϫ1000FSH␤luc reporter gene, or one of the six mutants, was transiently transfected into L␤T2 cells along with the indicated receptor expression vector. After overnight starvation in serum-free media, the cells were treated for 24 h with hormone. The results represent the mean Ϯ SEM of at least three experiments performed in triplicate and are presented as fold induction relative to the vehicle control. *, Indicates that the hormone treatment was significantly different from the vehicle-treated control using Student's t test. #, Indicates that the induction of the mutant reporter gene is significantly different from the induction of the wild-type reporter gene using one-way ANOVA followed by Tukey's post hoc test. A, L␤T2 cells were transiently transfected with the indicated reporter gene and the AR expression vector and then treated with 100 nM R1881. B, L␤T2 cells were transiently transfected with the indicated reporter gene and the PR expression vector and then treated with 100 nM R5020. C, L␤T2 cells were transiently transfected with the indicated reporter gene and the GR expression vector and then treated with 100 nM dexamethasone (Dex). WT, Wild type. well as the Ϫ230 and Ϫ139 sites. As seen in Fig. 10, an antibody specific for PR produced a supershift when the Ϫ197 probe was incubated with baculovirus-infected insect whole-cell extracts overexpressing PR (arrowhead), whereas no shift was seen with the negative nonspecific IgG control. This supershift was detectable after a much longer exposure than the EMSA in Fig. 9B. These results show that PR binds weakly to the Ϫ197 HRE compared with the Ϫ381 site and contrast with previous gel shifts that used HREs from the rat and ovine promoters (38,39). In addition, we saw a very weak supershift on the Ϫ230 and Ϫ139 sites that could indicate potential binding at these sites by PR as well. We did not observe binding to the other functional HREs by AR or GR under these EMSA conditions (data not shown) with the exception of the binding of AR to the Ϫ230 HRE as shown in Fig. 9. These results indicate that, although these HREs play a functional role in hormone responsiveness, they do not bind steroid receptors with high affinity on their own.
Because AR, PR, and GR did not bind to the other functional HREs in a convincing fashion (except PR binding to the Ϫ197 site), we tested whether other proteins from L␤T2 nuclear extracts bound the Ϫ273, Ϫ230, Ϫ197, or Ϫ139 sites. Although no difference was seen between vehicle-treated and hormonetreated L␤T2 nuclear extracts (data not shown), basal protein complexes bound all four probes (Fig. 11). Arrows denote specific complexes that were competed by unlabeled-self but not by unlabeled-mutant, competitor, using the Ϫ273 probe (Fig. 11A), the Ϫ230 probe (Fig. 11B), the Ϫ197 probe (Fig. 11C), and the Ϫ139 probe (Fig. 11D). Competition with mutant probes containing the same mutations used in the transient transfection experiments was used to determine whether any of the basal complexes were sensitive to those specific mutations. As can be seen, each of the gel shifts contained at least one complex that was sensitive to a mutation in the functional HRE, indicating that these basal proteins may also contribute to steroid binding or induction although, as mentioned previously, mutation of any of the HREs did not alter basal activity of the FSH␤ promoter (data not shown) nor did steroid treatment alter these EMSA complexes.

DISCUSSION
Gonadal steroid hormone feedback is a crucial component of the control of gonadotropin synthesis in the pituitary gonadotrope (15). In addition, although exposure to stress hormones such as corticosterone or dexamethasone can inhibit ovulation in primates and rodents (45,(65)(66)(67)(68)(69)(70), there is evidence that glucocorticoids play a physiological role in activating FSH␤ gene expression. For instance, both progestin and glucocorticoid levels peak during proestrus (71,72), and this coincides with increased FSH␤ mRNA expression and the secondary FSH␤ surge that occurs in the morning of estrus in rodents (73). Furthermore, glucocorticoids, as well as progestins, appear to be necessary for the secondary FSH␤ surge (37,74). These studies complement the numerous experiments in rats and mixed rat pituitary cells that found induction of FSH␤ mRNA after treatment with androgens, progestins, or glucocorticoids (24,26,28,29,33,34,(41)(42)(43).
In the current study, we characterized the specific molecular mechanisms of steroid hormone regulation of the FSH␤ gene in L␤T2 cells. Our results demonstrate that the proximal murine FSH␤ promoter is activated by androgens, progestins, and glucocorticoids (as well as their synthetic analogs) in the context of immortalized gonadotropes in culture (Figs. 1 and 3) in contrast to the lack of estrogen responsiveness on the murine FSH␤ promoter (Fig. 1). We also show, for the first time, that progestins and glucocorticoids, like androgens, can suppress transcription of the LH␤-subunit gene at the level of the gonadotrope (Fig. 4), although glucocorticoids have previously been shown to inhibit GnRH induction of LH␤ (45). However, because no obvious HREs exist in the LH␤ promoter, it is likely that the mechanism(s) of action are indirect, similar to the aforementioned studies concerning the suppression of LH␤ by androgens. The study of the mechanism(s) of LH␤ suppression by progestins and glucocorticoids is currently ongoing in our laboratory. Although FSH and LH are produced in the same cell type and both of their ␤-subunits are positively regulated by GnRH and activin (10,(75)(76)(77), the expression profiles of FSH␤ and LH␤ have been shown to differ over the menstrual cycle (73,78). It is intriguing to postulate that some of the observed differential regulation of the gonadotropin ␤-subunits may be due to these three 3-keto steroid hormones.
Androgens, progestins, and glucocorticoids could affect FSH␤ synthesis directly or by indirect means such as altering the balance of the autocrine activin/ inhibin/follistatin system (30,79). Our data indicate that induction of FSH␤ by androgens, progestins, and glucocorticoids is dependent on the presence of the appropriate hormone-bound steroid receptor because expression of the murine FSH␤luc reporter gene depends on the receptor levels (Fig. 2) and hormone concentration (Fig. 3). We also provide evidence that the murine FSH␤ promoter is directly bound and regulated by all three steroid hormone receptors. Using ChIP, we show that the endogenous AR, PR, and GR can bind the FSH␤ promoter in vivo in L␤T2 cells (Fig.  1B). Moreover, the use of mutant steroid receptors lacking the ability to bind DNA demonstrated that the steroid hormone receptors must bind DNA directly to modulate FSH␤ gene expression (Fig. 5). We also show that all three of the steroid hormone receptors can bind to specific sites within the FSH␤ promoter in vitro using gel-shift analysis (Figs. 9 and 10). Furthermore, mutation of putative HREs in the context of the Ϫ1000FSH␤luc reporter gene abolished the responsiveness of FSH␤ to androgens, progestins, and glu- cocorticoids (Fig. 8). Thus, this is the first demonstration that AR, PR, and GR can all directly induce murine FSH␤ gene expression.
It has been suggested previously that steroids (androgens in particular) can induce the expression of the GnRH-R (32,55) and that this regulation may be one of the contributing factors to induction of FSH␤ by steroids. However, our results demonstrating the suppression of the LH␤ gene by steroids would argue against a role of the GnRH-R. Furthermore, Bedecarrats and Kaiser (80) have shown that up-regulation of GnRH-R by overexpression or a higher-pulse frequency of GnRH inhibits FSH␤ expression. These results all point to direct regulation of the gonadotropin ␤-subunit genes by steroid hormones.
In addition to establishing that androgens, progestins, and glucocorticoids directly modulate FSH␤ transcription, we characterized the roles played by the HREs contained within the FSH␤ promoter. Because the responsiveness to steroid hormones mapped within the first 500 bp of the murine FSH␤ promoter (Fig. 6), we concentrated on the putative HREs in this region of the promoter. The six elements characterized had all been shown to bind PR in earlier studies using the rat and ovine promoters (38,39), although the relative importance of these sites had not been evaluated previously. In agreement with the data obtained using the rat promoter, the Ϫ381 HRE specifically bound PR as well as binding AR and GR (Fig. 9). Our gel-shift experiments suggest that the Ϫ381 HRE may be critical for direct regulation of the murine FSH␤ promoter. This hypothesis is further substantiated by the fact that mutation of this site greatly reduced or abolished FSH␤ responsiveness to androgens, progestins, and glucocorticoids (Fig. 8). Interestingly, the Ϫ381 HRE is an 11/12 match for a direct repeat element as opposed to a classical palindromic inverted repeat. This type of element has been reported to selectively confer specificity to AR on several enhancers and promoters [recently reviewed by Verrijdt et al. (81)]. However, in the context of the FSH␤ promoter, this sequence appears to be recognized by all three steroid receptors.
This raises the question of whether there is a specific response to these 3-keto steroid hormones on the murine FSH␤ promoter because androgens, progestins, and glucocorticoids have such different physiological effects. Much of the specificity may be provided by different levels of the steroid hormones. For instance, there is a circadian variation in glucocorticoid levels that can be affected by internal or external stress. Additionally, progesterone levels vary dramatically over the estrous cycle, and high levels are produced during pregnancy. Testosterone levels also increase in the male at puberty and, after peaking over the following decade, progressively decline with age. Factors such as temporal expression of the steroid receptors or subtle sequence preferences in the multiple HREs present in the FSH␤ promoter or in residues flanking the HREs, which result in different conforma-tional changes in the receptors and, thus, differential cofactor recruitment or protein-protein interactions with adjacent DNA-binding proteins, could also have an effect on specificity.
Because the magnitude of steroid hormone response imparted by each HRE is often weak, multiple HREs are often found in close proximity in the promoters of steroid-responsive genes (82). This situation also appears to apply to the murine FSH␤ promoter because mutation of the Ϫ273, Ϫ230, Ϫ197, and Ϫ139 HREs all reduce induction of FSH␤ by both androgens and progestins, suggesting that they are needed for the full induction. Glucocorticoid responsiveness seems to be less sensitive to mutation of the HREs although mutation of the Ϫ197 and Ϫ139 HREs did have a considerable effect. In addition to binding of AR, PR, and GR to the Ϫ381 HRE, we observed weak binding by PR to the Ϫ197 HRE and AR to the Ϫ230 site (Figs. 9 and 10). These data suggest several possibilities. First, these elements (with the exception of the Ϫ381 HRE) may have a low affinity for steroid receptors, and binding is difficult to detect using gelshift analysis although binding could be aided by highaffinity binding at the Ϫ381 site. Second, nuclear proteins present in the L␤T2 cells may be necessary for the steroid receptors to bind to these elements. Finally, these sites could be required for the binding of other transcription factors that would alter steroid hormone responsiveness. In support of the last two possibilities, we observed basal protein complex formation on the Ϫ297, Ϫ230, Ϫ197, and Ϫ139 elements using L␤T2 nuclear extracts that were sensitive to mutations in these HREs (Fig. 11) although mutations in the HREs did not affect basal activity of the FSH␤ promoter nor did steroid treatment of the cells alter the complexes (data not shown).
All of the functional HREs we characterized in the mouse promoter have a degree of conservation across multiple species, suggesting that they may play roles in steroid regulation of other mammalian species (Fig.  7). Not surprisingly, the putative HRE with the least conservation, the Ϫ175 element, had no functional effect in androgen, progestin, or glucocorticoid responsiveness of the murine FSH␤ promoter (Fig. 8). Although the HREs in the proximal FSH␤ promoter are well conserved in mammals, there appear to be species-specific differences in the regulation of the FSH␤ promoter by steroids. Our experiments with the murine FSH␤ promoter in L␤T2 cells and previous studies examining endogenous FSH␤ expression in rodents and in rat pituitary cells contrast with the regulation of FSH␤ by androgen and progestins observed in sheep and primates. In particular, androgens have been shown to repress FSH secretion at the level of the pituitary in rams and male rhesus monkeys (83,84) as well as in mixed pituitary cell cultures from transgenic mice containing 10 kb of the human FSH␤ promoter (85). Moreover, FSH␤ mRNA levels have been shown to decrease after progestin treatment in ovine mixed pituitary cell cultures (86) and in ovariectomized ewes pretreated with estrogen and treated with progesterone that received a hypothalamic-pituitary disconnection (87). Progesterone treatment also repressed the luciferase activity in pituitary cells derived from a transgenic mouse containing 4.7 kb of the ovine FSH␤ promoter linked to a luciferase reporter gene (88). Curiously, this same FSH␤luc reporter gene was activated by progestins when transiently transfected into ovine mixed pituitary culture (38) and induced by androgens when transfected into L␤T2 cells (55).
One possibility is that the results obtained in rodents, as compared with sheep and primates, differ due to species-specific steroid effects on the FSH␤ promoter. For instance, the Ϫ381 HRE that we found to be essential for hormone responsiveness on the murine FSH␤ promoter is conserved between mice and rats but not present in sheep. However, speciesspecific differences would not explain why the ovine FSH␤luc reporter gene was activated by progestins (38) whereas the endogenous gene was repressed (86). It is also plausible that a factor present in sheep and primate gonadotropes is necessary for the negative regulation; however, this is difficult to address because the available immortalized gonadotrope-derived cell lines and transgenic animal models are based on the mouse. Finally, it is credible that paracrine influences from other secretory cell types present in the anterior pituitary could be responsible for the negative regulation observed with the ovine and primate promoters. For example, androgens have been shown to regulate follistatin mRNA levels (27,30,89) and thus could indirectly mediate transcription of FSH␤ mRNA by impacting the autocrine/paracrine activity of activin in the gonadotrope. In addition, folliculostellate cells in the pituitary have been shown to secrete follistatin and overgrow mixed pituitary cell cultures (89). Clearly, further investigation is needed to discern whether steroids play distinct roles in the regulation of FSH␤ mRNA levels in different mammalian species.
In summary, we have demonstrated that androgens, progestins, and glucocorticoids via their receptors directly induce expression of the murine FSH␤-subunit gene at the level of the gonadotrope and that this regulation occurs through binding of the receptors to the Ϫ381 HRE in the proximal FSH␤ promoter. Additionally, this activation is specific for FSH␤, because LH␤ gene expression is repressed by androgens, progestins, and glucocorticoids. Although our studies do not exclude the possibility of steroidal regulation of GnRH synthesis as a central regulator of gonadotropin production, they do suggest that in endocrine disorders, such as polycystic ovary syndrome and Cushing's disease with excess androgens and glucocorticoids, respectively, fertility could be impacted by altered FSH␤ and LH␤ gene expression directly at the level of the gonadotrope, as well as the hypothalamus. Further examination of androgen, progesterone, and glucocorticoid regulation of the gonadotropin genes in the anterior pituitary is needed to better understand the physiological role these steroid hormones play in the maintenance of reproductive fitness. To begin to address their roles, selective ablation of the individual steroid hormone receptors in the gonadotrope cells needs to be performed using the Cre-LoxP system in mice. These conditional knockouts will determine whether AR, PR, and GR are necessary for gonadotropin regulation or whether there is a more dominant steroid regulation of the gonadotrope through afferent communication with the GnRH neuron. Evidence suggests that, individually, each of these steroids contributes to reproductive fitness through their cognate receptors. However, it is likely that complex interactions among them and other regulators of the hypothalamicpituitary-gonadal axis such as GnRH and activin exist and are also critical for reproduction.

Construction of Reporter Plasmids
Three luciferase reporter gene plasmids were generated by PCR amplification of the mouse FSH␤ promoter from a genomic clone kindly provided by Malcolm Low. Each PCR product was ligated into a pGL3 luciferase reporter plasmid (Promega Corp., Madison, WI) that had been digested with KpnI and HindIII. Initially, a luciferase reporter plasmid driven by 1526 bp of the mouse FSH␤ promoter (Ϫ1526FSH␤luc) was constructed. Two other reporter plasmids, Ϫ1000FSH␤luc and Ϫ500FSH␤luc were generated in a similar manner. Construction of the Ϫ95FSH␤luc plasmid was described previously (90). We confirmed the sequences of all promoter fragments with dideoxynucleotide sequencing by the DNA Sequencing Shared Resource, University of California San Diego Cancer Center.

Mutagenesis
We used the QuikChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA) to generate mutations in six putative HREs in the proximal FSH␤ promoter. The mutagenesis was performed using the Ϫ1000FSH␤luc plasmid and the appropriate oligonucleotides according to the manufacturer's protocol. The following oligonucleotides were used for mutagenesis: Ϫ139HREmut We also used the QuikChange Site-Directed Mutagenesis Kit to generate mutant PRs and GRs unable to bind to DNA. The PRC577A mutant is analogous to a C587A mutation in the human PRB that prevents DNA binding (63). The GRdim4 mutant contains four point mutations that prevent dimerization of the receptor and thus DNA binding: N454D, A458T, R460D, and D462C (64). The mutagenesis was performed using the pCMV5-rPR B or the pSG5-rGR plasmid, respectively, and the appropriate oligonucleotides according to the manufacturer's protocol. The following oligonucleotides were used for mutagenesis: PRC577Amut [5Ј-TCACTATGGTGT-GCTTACCTGTGGGAGCGCCAAGGTCTTCTTTAAGAGGG-3Ј] and GRdim4mut [5Ј-AAGGACAGCACGATTACCTTTG-TACTGGAGATAACTGTTGCATCAT-3Ј]. As with the reporter plasmids, all mutations were confirmed by dideoxynucleotide sequencing. Construction of the pSG5-rARC562G mutant was described previously (62).

Cell Culture and Transient Transfection
All of the transient transfection experiments were performed with the L␤T2 cell line (13). L␤T2 cells were maintained in 10-cm diameter dishes in DMEM (Cellgro, Mediatech, Herndon, VA) supplemented with 10% fetal bovine serum (Omega Scientific, Inc., Tarzana, CA) at 37 C with 5% CO 2 . One day before transfection, 3 ϫ 10 5 cells per well were plated into 12-well plates. Transient transfection was performed using Fugene 6 reagent (Roche Molecular Biochemicals, Indianapolis, IN) following the manufacturer's instructions. Each well was transfected with 0.4 g of reporter plasmid. Additionally, the cells were transfected with 0.2 g of the appropriate steroid receptor (unless otherwise noted). As an internal control for transfection efficiency, we transfected the L␤T2 cells with 0.2 g of a reporter plasmid containing ␤-galactosidase driven by the Herpes virus thymidine kinase promoter (tk-␤gal) or the Rous sarcoma virus promoter (RSV-␤-gal). The cells were switched to serum-free DMEM supplemented with 0.1% BSA, 5 mg/liter transferrin, and 50 nM sodium selenite 6 h after transfection. The following day, the cells were treated with ethanol (vehicle control) or hormone for 24 h. The cells were washed once with 1ϫ PBS and then lysed with 0.1 M K-phosphate buffer, pH 7.8, containing 0.2% Triton X-100. After lysing the cells, we assayed the luciferase activity using a buffer containing 100 mM Tris-HCl, pH 7.8, 15 mM MgSO 4 , 10 mM ATP, and 65 M luciferin. ␤-Galactosidase activity was measured using the Galacto-light assay (Tropix, Bedford, MA) according to the manufacturer's protocol. Both luciferase and ␤-galactosidase activities were measured using an EG&G Berthold Microplate Luminometer (PerkinElmer Corp., Norwalk, CT).

ChIP
L␤T2 cells were grown to confluency in 15-cm plates, treated with 100 nM hormone, and proteins were cross-linked to DNA by the direct addition of 1% formaldehyde to the cell medium. The nuclear fraction was obtained, and chromatin was sonicated to an average length of 1 kb in sonication buffer [50 mM HEPES, 140 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% Na-deoxycholate, and 0.1% sodium dodecyl sulfate (SDS)]. The lysate was diluted with ChIP dilution buffer (0.01% SDS, 1.1% Triton, 1.2 mM EDTA, 16.7 mM Tris pH 8, 167 mM NaCl) to a total of 3.5 ml and precleared with 100 l Protein A/G PLUS-Agarose beads (Santa-Cruz Biotechnology, Inc., Santa Cruz, CA). Protein-DNA complexes were incubated overnight with the C-19x AR polyclonal antibody (Santa Cruz Biotechnology, Inc.), the 1294 PR mouse monoclonal antibody (provided by Dean Edwards), the N499 GR rabbit polyclonal antibody (donated by Keith Yamamoto), or a nonspecific IgG control (Santa Cruz) and precipitated with Protein A/G beads (Santa Cruz). A fraction of the protein-DNA was not precipitated but set aside as the input. The agarose beads were washed in the following order: low-salt wash buffer [0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris (pH 8), 150 mM NaCl], high-salt wash buffer [0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris (pH 8), 500 mM NaCl], LiCl wash buffer [250 mM LiCl, 1% Nonidet P-40, 1% Nadeoxycholate, 1 mM EDTA, 10 mM Tris (pH 8)] and twice with Tris-EDTA buffer. The protein-DNA complexes were eluted with elution buffer (1% SDS, 0.1 M NaHCO 3 ), and the crosslinks were reversed with the addition of 200 mM NaCl and incubation at 65 C for 4 h. The DNA was phenol-chloroform extracted, precipitated, and then resuspended in 50 l of water. The primers used in PCR for the FSH␤ promoter were 5Ј-GGTGTGCTGCCATATCAGATTCGG-3Ј and 5Ј-GCAT-CAAGTGCTGCTACTCACCTGTG-3Ј and spanned the 280-bp sequence in the mouse FSH␤ gene from Ϫ223 to ϩ57. The primers for the FSH␤ coding region were 5Ј-GC-CGTTTCTGCATAAGC-3Ј and 5Ј-CAATCTTACGGTCTCG-TATACC-3Ј. The following PCR conditions were used: 4 min at 95 C, followed by 26 cycles consisting of 1 min at 95 C, 1 min at 60 C, and 1 min at 72 C, and an extension of 10 min at 72 C. The PCR product was labeled by including [␣-32 P]dATP in the nucleotide mix and run on a 5% acrylamide gel in 0.5ϫ Tris-borate-EDTA buffer. The gels were dried and subjected to autoradiography.

Preparation of Protein Extracts
Full-length, human AR containing a Flag epitope tag (92), human Flag-PR B (93), or Flag-GR (kindly provided by Steve Nordeen) was overexpressed in Sf9 insect cells via a baculovirus expression system by the University of Colorado Cancer Center Tissue Culture Core Facility. The Sf9 cells were inoculated with virus at a multiplicity of infection of 1.0 and grown for an additional 48 h at 27 C. Cells containing AR, PR, or GR were treated for the last 24 h before harvest with 1 M DHT, 200 nM R5020, or 500 nM triamcinolone acetonide (final concentration), respectively. The cells were harvested by centrifugation at 1500 rpm for 15 min, washed once in TG buffer (10 mM Tris-HCl, pH 8.0; and 10% glycerol) and frozen as a pellet at Ϫ80 C. We lysed the Sf9 cells in a homogenization buffer [20 mM Tris-HCl (pH 7.5), 350 mM NaCl, 1 mM dithiothreitol, 10% glycerol, 0.5 g/ml leupeptin, 10 g/ml bacitracin, 2 g/ml aprotinin, 1 g/ml pepstatin]. All procedures were done at 0Ϫ4 C. The cell lysate was centrifuged at 40,000 rpm for 30 min, and the supernatant was taken as a soluble whole-cell extract.
Nuclear extracts were prepared from L␤T2 and Cos-1 cells as previously described (94).

Western Blot Analysis
Wild-type and mutant steroid receptors, as well as their respective empty vector controls, were overexpressed in L␤T2 cells (AR, ARC562G, and pSG5) or Cos-1 cells (PR, PRC577A, pCMV5 and GR, GRdim4, pSG5) by transiently transfecting 10 g of DNA using Fugene 6 into 10-cm plates of cells and treating the cells with 100 nM R1881, R5020, or dexamethasone, respectively. Cells were harvested 24 h later by incubating cells in a lysis buffer [20 mM Tris-HCl (pH 7.4), 140 mM NaCl, 0.5% Nonidet P-40, 0.5 mM EDTA, 10 g/ml aprotinin, 10 g/ml pepstatin, 10 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride] for 30 min. The protein concentration was determined with Bradford reagent (Bio-Rad Laboratories, Inc., Hercules, CA), and an equal amount of protein per sample was loaded on a SDS-PAGE gel. After the proteins were resolved by electrophoresis and transferred to a polyvinylidene fluoride membrane, they were probed with specific antibodies for AR, PR, and GR. Specifically, the C-19x AR polyclonal antibody (Santa Cruz Biotechnology, Inc.), the 1294 PR mouse monoclonal antibody (provided by Dean Edwards), or the N499 GR rabbit polyclonal antibody (donated by Keith Yamamoto) were used. The bands were detected with secondary antibodies linked to horseradish peroxidase and enhanced chemiluminescence reagent (Amersham Biosciences, Piscataway, NJ).

Statistical Analyses
Transient transfections were performed in triplicate, and each experiment was repeated at least three times. The data were normalized for transfection efficiency by expressing luciferase activity relative to ␤-galactosidase activity and relative to the empty pGL3 plasmid to control for hormone effects on the vector DNA. The data were analyzed by Student's t test for independent samples or one-way ANOVA followed by post hoc comparisons with the Tukey-Kramer honestly significant difference (HSD) test using the statistical package JMP 5.0 (SAS Institute, Inc., Cary, NC). Significant differences were designated as P Ͻ 0.05.