Repression of DAX-1 and Induction of SF-1 Expression

Angiotensin II (Ang II) and adrenocorticotropic hormone stimulate aldosterone biosynthesis in the zona glomerulosa of the adrenal cortex through induction of the expression of the steroidogenic acute regulatory (StAR) protein, which promotes intramitochondrial cholesterol transfer. To understand the mechanism of this induction of the StAR protein, we have examined the effect of Ang II and forskolin, a mimicker of adrenocorticotropic hormone action, on two transcription factors known to modulate StARgene expression in opposite ways, DAX-1 and SF-1, in bovine adrenal glomerulosa cells in primary culture. Ang II markedly inhibited DAX-1 protein expression in a time- and concentration-dependent manner (to 38.7 ± 12.9% of controls at 3 nm after 6 h, p < 0.01), an effect that required de novo protein synthesis and ERK2/1 activation. This effect was associated with a concomitant decrease in DAX-1 mRNA and an increase in mitochondrial StAR protein levels. Similarly, forskolin dramatically repressed DAX-1 protein and mRNA expression (to 19.6 ± 1.8 and 50.3 ± 4.7% of controls, respectively,p < 0.01). Neither Ang II nor forskolin affected DAX-1 protein and mRNA stability. The aldosterone response to Ang II was markedly reduced (to 59 ± 4% of controls,p < 0.01) in transiently transfected cells overexpressing DAX-1. Whereas Ang II was without effect on SF-1 expression, forskolin significantly increased SF-1 protein and mRNA levels in a cycloheximide-sensitive manner (to 167.4 ± 16.6 and 173.1 ± 25.1% of controls after 6 h, respectively,p < 0.01). These results demonstrate that the balance between repressor and inducer function of DAX-1 and SF-1 are of critical importance in the regulation of adrenal aldosterone biosynthesis.

The biosynthesis of aldosterone, the main mineralocorticoid hormone, in the zona glomerulosa cells of the adrenal cortex is placed under the control of three principal physiological factors, the octapeptide hormone, angiotensin II (Ang II), 1 extra-cellular potassium (K ϩ ), and adrenocorticotropic hormone (ACTH) (1). Whereas Ang II and K ϩ exert their effect by recruiting the calcium messenger system in glomerulosa cells (2,3), ACTH acts mainly by activating adenylyl cyclase and generating cAMP as an intracellular messenger (4). In the cascade of subsequent events leading to the final production of aldosterone, the regulated, rate-limiting step is the transfer of cholesterol from the relatively cholesterol-rich outer mitochondrial membrane to the cholesterol-poor inner mitochondrial membrane (5). Cholesterol is then converted to pregnenolone by the cytochrome P450 side chain cleavage enzyme (P450 scc ).
In all steroidogenic tissues, intramitochondrial cholesterol transfer is facilitated by the steroidogenic acute regulatory (StAR) protein (6). In the adrenal zona glomerulosa cell, the expression of the StAR protein is rapidly increased by factors that activate mineralocorticoid biosynthesis. Indeed, Ang II and ACTH have been shown to stimulate StAR mRNA and StAR protein expression and to concomitantly increase aldosterone production in bovine zona glomerulosa cells (7,8). Moreover, in human H295R adrenocortical carcinoma cells, which bear only very few ACTH receptors, challenge with cAMP analogs, used to mimic adenylyl cyclase activation, Ang II, or K ϩ , leads to StAR mRNA and protein expression (9,10). Whereas the initial signal transduction mechanisms mediating the steroidogenic action of these activators of aldosterone biosynthesis are well characterized (3), the events occurring downstream of Ca 2ϩ or cAMP signal generation and leading to the induction of StAR protein expression are poorly understood.
No consensus cAMP response element has been found in the ϳ3.6-kilobase 5Ј-flanking region of the mouse, rat, human, porcine, and bovine StAR gene (11). In contrast, in all species, several putative binding sites for steroidogenic factor-1 (SF-1), also called Ad4BP, are present in the StAR gene promoter (10,(12)(13)(14)(15)(16). The number and localization of these binding sites vary from one species to the other. SF-1 is a nuclear transcription factor that was first identified in adrenal cortical cells (17). The orphan nuclear receptor SF-1 plays a critical role in adrenal and gonadal differentiation, development, and function (18). Furthermore, SF-1 has also been shown to regulate the expression of genes encoding cytochrome P450 hydroxylases and to efficiently transactivate the StAR gene in transient transfection assays in various cell types (13,15,16,19). Although the extent of SF-1 involvement in the regulation of StAR gene expression may present species-and cell type-dependent differences (20), it appears that activation of the cAMP-signaling pathway leads to increased phosphorylation (21) and/or expression (22) of SF-1 protein.
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF421373.
In addition to response elements for SF-1, the StAR gene promoter also bears a binding site for another orphan member of the nuclear receptor superfamily, DAX-1 (dosage-sensitive sex reversal adrenal hypoplasia congenita critical region on the X chromosome, gene 1) (23,24). DAX-1 has been shown to act as a powerful repressor of StAR gene expression. Indeed, overexpression of DAX-1 in Y-1 mouse adrenal tumor cells inhibits steroid synthesis, and DAX-1 represses both basal and cAMPinduced StAR promoter activity by binding to DNA hairpin secondary structures on the StAR gene promoter or to the SF-1 protein itself (25). Furthermore, overexpression of DAX-1 in Y-1 adrenocortical cells impairs basal and cAMP-stimulated steroid production (26). Conversely, cAMP down-regulates DAX-1 expression in cultured rat Sertoli cells (26). Finally, SF-1 and DAX-1 are co-localized in various endocrine and steroidogenic tissues, suggesting that these two nuclear proteins may be linked in function.
Because Ang II and ACTH stimulate StAR protein expression, the present study was undertaken to investigate whether the mechanism of this response involves a modulation of DAX-1 and SF-1 expression. We report here that both Ang II, a calcium mobilizing hormone, and forskolin, used as a mimicker of ACTH action to generate a pure cAMP signal, markedly inhibit DAX-1 expression at the mRNA and protein level. We also show that forskolin significantly increases SF-1 expression. This study provides evidence that the removal of the suppressor effect of DAX-1 on StAR expression is an important mechanism through which activators of aldosterone biosynthesis increase StAR expression.

EXPERIMENTAL PROCEDURES
Bovine Adrenal Zona Glomerulosa Cell Culture and Treatments-Bovine adrenal glands were obtained from a local slaughterhouse. Zona glomerulosa cells were prepared by enzymatic dispersion with dispase and purified on Percoll density gradients as previously reported (27). Primary cultures of purified glomerulosa cells were maintained in Dulbecco's modified Eagle's medium as described in detail elsewhere (27). The cells were grown on 6-well tissue culture plates (3 ϫ 10 6 cells/well) and kept in serum-free medium for 24 h before experiments, which were performed on the third day of culture. Cells were then washed and incubated at 37°C in serum-free medium containing various agents for varying periods of time as appropriate. At the end of the incubation period, the media were collected, and cells were processed for protein or total RNA extraction as described hereafter.
Determination of Aldosterone Production-Aldosterone content in incubation media was measured by direct radioimmunoassay using a commercially available kit (Diagnostic Systems Laboratories, Webster, TX). Aldosterone production was normalized and expressed per milligram of cellular protein.
Western Blot Analysis-For the determination of protein expression levels, bovine glomerulosa cells were washed twice in ice-cold phosphate saline buffer (PBS) and lysed in PBS containing 1% (v/v) Triton X-100, 0.1% (w/v) sodium dodecyl sulfate, 0.5% (w/v) deoxycholate, 1 mM phenylmethylsulfonyl fluoride, and 1 g/ml leupeptine. The cell debris were removed by centrifugation at 20,000 ϫ g at 4°C for 15 min, and supernatants were used as cell lysates. Proteins were quantified using a protein microassay (Bio-Rad). Equal amounts of protein (20 g) were resolved by 12% SDS-polyacrylamide gel electrophoresis and transferred onto polyvinylidene difluoride membranes. Western blot analysis of DAX-1 and SF-1 proteins was carried out with a mouse monoclonal antiserum directed against DAX-1 (kindly provided by Dr. Enzo Lalli, Strasbourg, France) or a rabbit polyclonal antiserum to Ad4BP/SF-1 (kindly supplied by Dr. Ken Morohashi, University of Tsukuba, Japan). Mitochondrial StAR protein levels were determined by Western blot as previously described (8). Immunoreactive proteins were visualized by the enhanced chemiluminescence method (ECL, Amersham Biosciences). Band intensity was quantified with a Molecular Dynamics Computing Densitometer.
Transient Transfection of Bovine Glomerulosa Cells with DAX-1 cDNA-For transfection experiments, 800,000 bovine glomerulosa cells were seeded in 60-mm Petri dishes. Cells were transfected using Effectene transfection reagent (Qiagen, Basel, Switzerland) according to the manufacturer's instructions. One g of pBKCMV-hDAX-1 encoding the human DAX-1 protein was introduced into cells at a DNA:Effectene ratio of 1:50. The same conditions were used with the empty vector pBKCMV for mock-transfected cells. Cells were incubated with the transfection complexes for 8 h then washed twice with phosphate saline buffer and incubated for 48 h in fresh medium supplemented with serum. In separate experiments with pBKCMV-GFP encoding the green fluorescent protein we observed that the transfection efficacy amounted to 35-40% under the above conditions (data not shown). Cells were then challenged with Ang II, and DAX-1 protein levels were determined as described.
Cloning of Bovine DAX-1 from Zona Glomerulosa Cells-Total cellular RNA was isolated from bovine glomerulosa cells using the SV Total RNA Isolation system (Promega, Zurich, Switzerland) according to the manufacturer's instructions. Total RNA (500 ng) was reverse-transcribed and amplified using the Access RT-PCR system (Promega). A fragment of bovine DAX-1 cDNA (214 base pairs) was amplified using human DAX-1 primers (5Ј-AGGGGACCGTGCTCTTTAAC-3Ј forward and 5Ј-ATGATGGGCCTGAAG AACAG-3Ј reverse). Primers corresponded to positions ϩ1145-1164 and 1339 -1358 of the human DAX-1 (NCBI/GenBank TM accession number NM_000457). The PCR product was purified from a 1% agarose gel, cloned into pGEM-T Easy vector (Promega), and amplified in JM 109 competent cells. The plasmid insert was sequenced by automatic sequencing using the DyEnamics Terminator sequencing kit (Amersham Biosciences) and Applied Biosystem 3100 sequencer.
Reverse Transcription-Polymerase Chain Reaction-RT-PCR was used to evaluate DAX-1 and SF-1 mRNA abundance in response to various treatments. RT-PCR measurements of mRNA were performed using Promega reagents and 500 ng of total RNA per reaction. Reverse transcription with avian myeloblastosis virus RT (5 units/reaction, 48°C for 30 min) primed with specific primers was followed by PCR using Tfl DNA polymerase (5 units/reaction) with the following cycling parameters: denaturation for 5 min at 95°C followed by 30 and 25 cycles for DAX-1 and SF-1, respectively, of 95°C for 1 min, 56°C for 1 min, and 68°C for 30 s and a final extension at 68°C for 10 min. The primers were bovine SF-1 (5Ј-GCAGAAGAAGGCACAGATTC-3Ј (forward) and 5Ј-TGGGTACTCAGACTTGATGG-3Ј (reverse). Primers corresponded to positions ϩ439 -459 and 693-713 of the bovine Ad4BP mRNA (NCBI/GenBank TM accession number D13569). DAX-1 primers were as described above. To correct for potential variations in RT-PCR efficiency, an internal control, a fragment of the glyceraldehyde-3phosphate dehydrogenase (GAPDH) gene was co-amplified in each sample using as the sense primer 5Ј-ATGGTGAAGGTCGGAGTG-3Ј and as the antisense primer 5Ј-TGCAGAGATGATGACCCTC-3Ј. Primers corresponded to positions ϩ82-99 and 426 -444 of the human GAPDH (NCBI/GenBank TM accession number NM_002046). The combination of GAPDH and SF-1 primers yielded two products corresponding to 362 and 274 bp for GAPDH and SF-1, respectively, and that of GAPDH and DAX-1 yielded two products corresponding to 362 and 214 bp for GAPDH and DAX-1, respectively. Ten microliters of the PCR product were separated by electrophoresis on ethidium bromide-stained 1.3% agarose gels. The intensity of each band was normalized to the intensity of the corresponding GAPDH band. Bands were quantified by image analysis using the NIH Image 1.23 software (rsb.info.nih.gov/nih-image/download.html). All amplified sequences were confirmed by automatic sequencing. Plots of product amount versus cycle number were linear, confirming that all reactions were run in the exponential part of the progress curve.
Data Deposition-The partial sequence of bovine DAX-1 has been deposited in the GenBank TM data base (Bethesda, MD) under accession number AF421373.
Data Bank Search-The nucleic acid and derived amino acid sequences of DAX-1 cDNA were compared with known sequences provided by the National Center for Biotechnology Information (Bethesda, MD) using Blast searches. Comparison of nucleotide and deduced amino acid sequences were performed using ClustalW.
Statistical Analysis-The data presented are the mean Ϯ S.E. The statistical significance of differences between treatments was determined by one-way analysis of variance and Fisher's protected least significant difference test. Values of p Ͻ 0.05 or less were considered statistically significant.

RESULTS
Angiotensin II Inhibits DAX-1 Protein Expression-DAX-1 protein levels were high in unstimulated bovine adrenal glomerulosa cells. When cells were treated with Ang II for 6 h, a significant, concentration-dependent inhibition of DAX-1 pro-tein expression was observed between 0.1 and 10 nM Ang II, as determined by Western blot analysis. Maximal inhibition (to 38.7 Ϯ 12.9% of controls, n ϭ 3-9, p Ͻ 0.01) was achieved with 3 nM Ang II (Fig. 1, A and B).
Kinetics of the Effect of Ang II on DAX-1 Protein Expression-To determine the kinetics of the inhibition of DAX-1 protein expression induced by Ang II, we stimulated glomerulosa cells with 10 nM Ang II for 0 -6 h. As shown in Fig. 2, A and B, the inhibitory effect of Ang II was time-dependent, with DAX-1 protein levels reaching 61.5 Ϯ 4.6% (n ϭ 5, p Ͻ 0.01) and 40 Ϯ 3.4% of controls (n ϭ 5, p Ͻ 0.01) after 4 and 6 h, respectively. No further decrease in DAX-1 protein expression was observed thereafter (data not shown). The progressive inhibition of DAX-1 protein expression was accompanied with a time-dependent increase in aldosterone production from Ang II-treated adrenal glomerulosa cells (Fig. 2C).
Ang II-induced Inhibition of DAX-1 Protein Expression Requires Protein Synthesis-The inhibition of DAX-1 protein expression by Ang II depended upon de novo protein synthesis. Indeed, although glomerulosa cells stimulated with 10 nM Ang II for 6 h showed a significant decrease in DAX-1 protein levels (to 40.8 Ϯ 2.6% of controls, n ϭ 3, p Ͻ 0.01), simultaneous treatment with cycloheximide, a protein translation inhibitor, abolished the Ang II-induced inhibition of DAX-1 protein expression (Fig. 3, A and B). Cycloheximide also completely abolished the aldosterone response to Ang II, as previously reported by others (28) (Fig. 3C).
Involvement of ERK2/1 Activation in Ang II-induced Repression of DAX-1-Ang II is known to activate ERK2/1 in bovine adrenal glomerulosa cells (29). To determine whether inhibition of DAX-1 protein expression by Ang II is mediated by the mitogen-activated protein kinase pathway, we stimulated cells with Ang II alone or in combination with U0126 (Biomol, Plymouth Meeting, PA), an inhibitor of MEK-1, the kinase that phosphorylates and activates ERK2/1. As shown in Fig. 4, A and B, the inhibitory effect of Ang II on DAX-1 expression was completely abolished in the presence of U0126. In contrast, the p38 mitogen-activated protein kinase inhibitor, SB203580, did not affect Ang II action (data not shown).
To determine whether Ang II exerts its inhibitory effect directly on DAX-1 mRNA expression, we stimulated glomerulosa cells for 6 h with 10 nM Ang II alone or in combination with 10 M U0126, and DAX-1 mRNA levels were then determined by semi-quantitative RT-PCR. The results shown in Fig. 4, C and D, indicate that Ang II treatment led to a significant decrease in bovine DAX-1 mRNA levels (to 68.8 Ϯ 4.3% of controls, n ϭ 4, p Ͻ 0.01). This down-regulation of DAX-1 mRNA by Ang II was completely prevented by the MEK-1 inhibitor U0126.
The involvement of ERK2/1 in the functional response to Ang II was also observed at the level of steroid production. Indeed, as shown in Fig. 4E, two structurally unrelated MEK-1 inhibitors, U0126 and PD98059 (Alexis Biochemicals, Lä ufelfingen, Switzerland), significantly reduced aldosterone production elicited by Ang II.
Forskolin Inhibits DAX-1 Protein and mRNA Expression-We next examined whether another inducer of StAR expression and aldosterone production, forskolin, used as a mimicker of ACTH, which mobilizes the cAMP messenger system, also affects DAX-1 expression. As shown in Fig. 5, A and   B, treatment of adrenal glomerulosa cells with 25 M forskolin for 6 h dramatically inhibited DAX-1 protein expression (to 19.6 Ϯ 1.82% of controls, n ϭ 3, p Ͻ 0.001). As for Ang II, inhibition of DAX-1 protein expression by forskolin was concentration-and time-dependent (data not shown). Furthermore, the inhibition of DAX-1 protein expression by forskolin was completely prevented by cycloheximide. In contrast, the MEK-1 inhibitor, U0126, had no effect on forskolin action. The inhibition of DAX-1 protein expression provoked by forskolin was associated with a 56-fold increase in aldosterone production (Fig. 5E).
The inhibition of DAX-1 expression by forskolin was also observed at the transcriptional level, as demonstrated in Fig. 5, C and D. Indeed, in glomerulosa cells stimulated with forskolin, DAX-1 mRNA levels were decreased to 50.3 Ϯ 4.7% of control untreated cells (n ϭ 3, p Ͻ 0.0.01). This effect was not reversed by the MEK-1 inhibitor, U0126. Moreover, actinomycin D did not affect the reduction of DAX-1 mRNA expression elicited by forskolin (data not shown).
Ang II and Forskolin Do Not Affect DAX-1 Protein and mRNA Degradation-To determine whether Ang II and forskolin decreased DAX-1 protein levels by accelerating its catabolism, we incubated bovine glomerulosa cells for 6 h in the absence or presence of either agonist and then followed DAX-1 protein levels under conditions of protein synthesis blockade. As shown in Fig. 6A, DAX-1 protein levels were quite stable over time under basal conditions and in the absence of cycloheximide. In contrast, when de novo protein synthesis was prevented, the levels of DAX-1 protein decreased markedly with time, with a half-life of ϳ6 h. Neither Ang II nor forskolin accelerated this process (Fig. 6A).
The inhibition of DAX-1 mRNA expression could result from changes in transcription rate and/or in mRNA turnover. We therefore examined whether Ang II affected DAX-1 mRNA stability. Glomerulosa cells were treated for 6 h in the absence or presence of Ang II (10 nM), and then actinomycin D was added. After 6 h, DAX-1 mRNA levels decayed to 45.2% of the zero time value in cells treated with actinomycin D alone and to a similar value (53.4 Ϯ 4.9%, n ϭ 3, p Ͻ 0.01) in cells that had been pretreated with Ang II (Fig. 6B). Similar results were obtained with forskolin (data not shown).
Effect of Overexpression of DAX-1 on the Aldosterone Response to Ang II-The role of DAX-1 on the aldosterone response to Ang II was confirmed in bovine glomerulosa cells that had been transiently transfected with DAX-1 cDNA. In cells transfected with the empty pBKCMV vector, Ang II induced a robust aldosterone production (Fig. 7, Mock). Although the absolute aldosterone values were somewhat lower than in nontransfected cells, possibly because of the transfection procedure itself, the fold increase over basal was not significantly different (see, for example, Figs. 1D, 3C, or 4C). This aldosterone response was associated with the expected repression of DAX-1. In contrast, in cells transiently transfected with pBKCMV-hDAX-1, which expressed 20 -30-fold higher levels of the DAX-1 protein, the aldosterone response to Ang II was significantly reduced to 59.4 Ϯ 4% of the response measured in mock-transfected cells (n ϭ 3, p Ͻ 0.01).
Effects of Ang II and Forskolin on SF-1 Protein and mRNA Expression-Because the transcription factor SF-1 is known to regulate StAR protein expression, we examined whether the induction of StAR expression induced by both Ang II and forskolin involves changes in SF-1 protein expression in adrenal glomerulosa cells. As shown in Fig. 8, A and B, a significant increase in SF-1 protein expression was observed in cells stimulated for 6 h with forskolin (167.4 Ϯ 16.6% of controls, n ϭ 7, p Ͻ 0.001), whereas Ang II had no effect. The forskolin-induced increase in We next examined whether forskolin and Ang II affected SF-1 mRNA expression. Bovine glomerulosa cells were stimulated for 6 h with 10 nM Ang II or 25 M forskolin, and SF-1 mRNA levels were then determined by semi-quantitative RT-PCR. The results shown in Fig. 8, C and D, indicate that forskolin treatment led to a significant increase in bovine SF-1 mRNA levels (to 173.1 Ϯ 25.1% of controls, n ϭ 5, p Ͻ 0.05). In contrast, Ang II had no significant effect on SF-1 mRNA expression. The forskolin-induced increase in SF-1 mRNA expression was almost completely abolished by actinomycin D (Fig. 8,  C and D), suggesting a direct action of forskolin on SF-1 mRNA at the transcriptional level.

DISCUSSION
The present study was undertaken in an attempt to investigate the mechanism of the known induction of StAR protein expression in adrenal glomerulosa cells in response to two physiological activators of aldosterone biosynthesis, the octapeptide hormone angiotensin II and adrenocorticotropic hormone, ACTH (10,30). We have focused our attention on two orphan members of the nuclear receptor family of transcription factors, DAX-1 and SF-1. Indeed, on one hand, DAX-1 has been shown to repress StAR gene expression by binding to a hairpin structure located in the StAR gene promoter and to block steroidogenesis (25). On the other hand, multiple binding elements for SF-1 have been reported in the 5Ј-flanking region of the StAR gene and are required for maximal promoter activity (4).
Two main conclusions can be drawn from the present work. 1) Ang II, which recruits the calcium messenger signaling system (31), and forskolin, used as a mimicker of ACTH action via cAMP generation, both markedly repressed DAX-1 expression, and 2) forskolin significantly increased the expression of SF-1. It is worth stressing that these observations were obtained in bovine glomerulosa cells in primary culture, expressing normal levels of only endogenous DAX-1 and SF-1, rather than in artificial overexpression systems.
The inhibition of DAX-1 exerted by Ang II and forskolin was manifest at both the protein and mRNA levels. In fact, although most studies have focused on the developmental and tissular distribution of DAX-1 expression in various species (32,33), on its mechanism of action (25,34,35), and on the incidence of mutations in the DAX-1 gene on adrenal and reproductive functions (23,24,36,37), information on the regulation of DAX-1 expression in steroidogenic tissues is practically non-existent. In our hands, the aldosterone response was correlated with the extent of the repression of DAX-I, the highest amounts of the mineralocorticoid being produced when DAX-1 protein levels were the lowest.
In principle, two main mechanisms can account for the inhibition of DAX-1 expression observed in response to Ang II and forskolin; they are 1) an inhibition of DAX-1 gene transcription and/or translation, leading to reduced mRNA and protein levels, and 2) a decrease in DAX-1 mRNA and/or protein stability. Our data indicate that the latter mechanism is much less likely to occur. Indeed, neither Ang II nor forskolin accelerated the decay rate of DAX-1 mRNA or protein under conditions of blockade of transcription with actinomycin or blockade of protein translation with cycloheximide; a half-life of ϳ6 h was measured for the DAX-1 protein whether the agonists were present or not. Therefore, a more likely mechanism should be the induction by Ang II and forskolin of a repressor protein, which blocks DAX-1 gene transcription. Consistent with this hypothesis, the repression exerted by Ang II and forskolin was abolished under conditions of protein synthesis inhibition.
A number of studies suggest that the mitogen-activated protein kinase cascade is involved in the regulation of steroidogenesis. For example, ERK2/1 activation mediates the stimulation of steroid production in porcine (35), murine (38), and human (39) ovarian granulosa cells. In bovine adrenal glomerulosa cells, Ang II stimulates ERK2/1 (p42/44) activity via protein kinase C and Ras/Raf-1 kinase (29). Whether this activation contributes only to the trophic effect of Ang II on adrenal cells or also to the steroidogenic response was not known. In the present study, the ERK2/1-mediated inhibition of DAX-1 expression induced by Ang II would suggest a major role for ERK2/1 in Ang II-induced aldosterone production. In contrast, the role of ERK2/1 in the activation of adrenal steroidogenesis by ACTH or forskolin is controversial. Thus, forskolin was found to increase StAR expression and steroidogenesis in Y1 mouse adrenocortical tumor cells and to activate ERK2/1 (21). In contrast, in bovine fasciculata and rat glomerulosa cells, ACTH treatment did not lead to p42/44 mitogen-activated protein kinase (ERK2/1) activation (40,41). In agreement with the latter reports, in our hands the massive inhibition of DAX-1 expression elicited by forskolin was insensitive to two unrelated inhibitors of p42/44 activation, a finding that rules out an involvement of ERK2/1 in the steroidogenic response to cAMP.
The role of DAX-1 in the steroidogenic response to Ang II was corroborated in cells overexpressing DAX-1. Indeed, Ang IIinduced aldosterone production was markedly reduced in cells expressing 20 -30-fold higher levels of DAX-1. This finding is in agreement with a previous report showing that steroid production is impaired in Y-1 adrenocortical cells expressing DAX-1 (34) and, to the contrary, with the demonstration that mice lacking DAX-1 show an increased adrenal responsiveness to ACTH (42). In Y-1 cells expressing DAX-1, the steroidogenic response to forskolin was completely suppressed (34), whereas in our hands, the aldosterone response to Ang II was highly significantly but incompletely decreased in bovine glomerulosa cells overexpressing DAX-1. Because the transfection efficacy amounted to only 35-40% in these cells, it is likely that, in fact, the extent of inhibition in the transfected cells was somewhat underestimated. Whether blockade of aldosterone synthesis in cells overexpressing DAX-1 is actually complete is presently unknown. In fact, previous work from our laboratory would tend to suggest that this is not the case. Indeed, we have shown that Ang II also promotes aldosterone biosynthesis through non-genomic mechanisms via the calcium messenger system, such as intramitochondrial cholesterol transfer (7, 31), a process that could explain the residual aldosterone response to the hormone in the presence of high levels of DAX-1.
The two activators of mineralocorticoid biosynthesis also markedly differed in their effect on SF-1 expression. Indeed, on one hand, forskolin significantly increased SF-1 mRNA and protein expression. This result is consistent with a previous study showing that forskolin induces an increase in SF-1 protein expression in granulosa-derived cell lines (43). On the other hand, Ang II was without effect.
The StAR promoter is known to be responsive to stimulation by cAMP despite the fact that it lacks a consensus cAMP response element (14,15). Multiple SF-1 binding sites are present in the StAR promoter (13) and SF-1, which is known to positively regulate StAR expression (12,13,15,16,19,44,45), mediates cAMP responsiveness (13,15).
In addition, SF-1-mediated StAR gene transcription has been shown to be enhanced by phosphorylation of a single serine residue (Ser-203) located in the hinge region of the protein (46,47). Ser-203 is a substrate for ERK2/1 in vitro and is critical for activation of SF-1 by multiple components of the mitogen-activated protein kinase pathway (46,47). In Y1 adrenocortical cells, forskolin stimulation leads to activation of ERK2/1 and phosphorylation of SF-1 (21), although it is not known whether this response is accompanied by an increase in SF-1 protein expression. Our results in bovine glomerulosa cells differ from those of the latter work in that we did not observe an involvement of ERK2/1 activation in the repressor effect of forskolin. In contrast, our data clearly demonstrate that forskolin markedly increases SF-1 protein expression, a result that is also compatible with an increase in the amount of phosphorylated SF-1.
In contrast to forskolin, Ang II had no effect on SF-1 expression. In human H295R adrenocortical carcinoma cells, deletion analysis of the StAR gene promoter has shown that Ang II and dibutyryl cAMP activation of reporter gene expression requires the presence of one critical SF-1 binding site located proximally to the transcriptional start site, within the 150 first bases (10,13). No effect of Ang II or forskolin on endogenous SF-1 protein expression or phosphorylation was reported. The proximal SF-1 binding element overlaps with a hairpin structure (nucleotides Ϫ61 to Ϫ27), which is a binding site for DAX-1 and mediates DAX-1-induced StAR gene repression (25). In fact, DAX-1 is likely to act as a suppressor of SF-1-induced transcription. Indeed, recent studies demonstrate a direct interaction between DAX-1 and SF-1, potentially recruiting co-repressors such as NcoR (nuclear receptor co-repressor) to the transcription factor complex (48,49). Our results demonstrating that Ang II and forskolin block DAX-1 expression are, therefore, compatible with a model in which binding of DAX-1 to the hairpin structure and/or heterodimer formation between DAX-1 and SF-1 is prevented, the latter thus becoming able to enhance StAR gene expression.
In summary, we have shown that two activators of aldosterone biosynthesis in adrenal zona glomerulosa cells mobilizing distinct signaling pathways markedly lower the expression levels of DAX-1, a known repressor of the StAR protein, which plays a crucial role in the activation of steroidogenesis. Our studies provide the first evidence that the balance between DAX-1 and SF-1 expression levels, i.e. between repressor and inducer functions, may play a key role in the fine tuning of the mineralocorticoid response. In view of the specific distribution of these two transcription factors, this mechanism may be of general relevance in all steroid-producing tissues.