SF-1 (Steroidogenic Factor-1) and C/EBPβ (CCAAT/Enhancer Binding Protein-β) Cooperate to Regulate the Murine StAR (Steroidogenic Acute Regulatory) Promoter

Abstract

The steroidogenic acute regulatory (StAR) protein mediates the rate-limiting step of steroidogenesis, which is the transfer of cholesterol to the inner mitochondrial membrane. In steroidogenic tissues, StAR expression is acutely regulated by trophic hormones through a cAMP second messenger pathway, leading to increased StAR mRNA levels within 30 min, reaching maximal levels after 4–6 h of stimulation. The molecular mechanisms underlying such regulation remain unknown. We have examined the StAR promoter for putative transcription factor-binding sites that may regulate transcription in a developmental and/or hormone-induced context. Through sequence analysis, deoxyribonuclease I (DNAse I) footprinting and electrophoretic mobility shift assays (EMSAs), we have identified two putative CCAAT/enhancer binding protein (C/EBP) DNA elements at −113 (C1) and −87 (C2) in the mouse StAR promoter. Characterization of these sites by EMSA indicated that C/EBPβ bound with high affinity to C1 and C2 was a low-affinity C/EBP site. Functional analysis of these sites in the murine StAR promoter showed that mutation of one or both of these binding sites decreases both basal and (Bu)₂cAMP-stimulated StAR promoter activity in MA-10 Leydig tumor cells, without affecting the fold activation[ (Bu)₂cAMP-stimulated/basal] of the promoter. Furthermore, we have demonstrated that these two C/EBP binding sites are required for steroidogenic factor-1 (SF-1)-dependent transactivation of the StAR promoter in a nonsteroidogenic cell line. These data indicate that in addition to SF-1, C/EBPβ is involved in the transcriptional regulation of the StAR gene and may play an important role in developmental and hormone-responsive regulation of steroidogenesis.

INTRODUCTION

The rate-limiting step of steroidogenesis is the delivery of cholesterol from cellular stores to the inner mitochondrial membrane, where it is converted to pregnenolone by the cytochrome P450 side-chain cleavage enzyme (P450_scc; Refs. 1, 2). The steroidogenic acute regulatory (StAR) protein mediates this transfer of cholesterol to the inner mitochondrial membrane and thus, is required for this regulatory step (reviewed in Refs. 3–6). Expression of the StAR protein is rapidly stimulated in steroidogenic tissues in response to trophic hormone through a cAMP second messenger pathway (7, 8). At the level of StAR gene transcription, it has been reported that an increase in mRNA levels is detectable within 30 min after stimulation of MA-10 cells with (Bu)₂cAMP (9). It has also been reported that the initial cAMP-stimulated induction of StAR mRNA does not require de novo protein synthesis (10). Therefore, the complement of transcription factors required to confer hormone-responsive activation of the StAR promoter must be present before stimulation. In addition, posttranslational modification of one or more transcription factors in response to cAMP stimulation may play a critical role in the activation of this promoter.

Recently it has been suggested that steroidogenic factor-1 (SF-1), an orphan nuclear receptor, may play important roles in the regulation of StAR transcription (10–14). SF-1 has been shown to play a role in the transcriptional regulation of many genes involved in steroidogenesis, including steroid hydroxylase genes (15–17), as well as LHβ (18), the ACTH receptor (19), and the GnRH receptor (20). SF-1 null mice have also revealed a role for SF-1 in the development of the gonads and adrenal glands (11, 21). Combined, these reports suggest an important role for SF-1 in the regulation of steroidogenesis at a number of levels. However, there are many lines of evidence suggesting that SF-1, although important for StAR transcription, may not be a key transcription factor in the acute regulation of the StAR gene in response to cAMP stimulation. For example, transfection studies in nonsteroidogenic cell lines have shown that SF-1 is capable of transactivating a StAR reporter (13, 14). Yet, when multiple SF-1-binding sites were mutated in the mouse StAR promoter and analyzed in MA-10 cells, the cAMP responsiveness (fold activation) from the promoter was not disrupted (10). These data indicate that SF-1 is required for proper activation of the StAR promoter but may not confer cAMP responsiveness in steroidogenic cells. These findings have led us to examine other transcription factors, whose activity is acutely regulated in response to trophic hormone in steroidogenic tissues, for their involvement in the transcriptional regulation of the StAR gene.

Recent studies in our laboratory have suggested that the CCAAT/enhancer binding protein (C/EBP) family of basic leucine zipper transcription factors may be involved in the regulation of steroidogenesis in Leydig cells. Thus far, six members of the C/EBP family have been identified: C/EBPα, C/EBPβ, C/EBPδ, C/EBPε, C/EBPγ (Ig/EBP), and C/EBPζ (CHOP; Ref. 22). C/EBPβ is the only member of the C/EBP family expressed in unstimulated primary Leydig cell cultures and MA-10 cells (A. J. Reinhart, D. Nalbant, S. C. Williams, and D. M. Stocco, unpublished observation) and C/EBPβ levels increase in MA-10 cells by 4.5-fold upon 4 h treatment with 1 mm (Bu)₂cAMP (23). It has also been shown that C/EBPβ activity can be altered, presumably by protein kinase A (PKA), upon treatment with cAMP analogs (24–27). Therefore, C/EBPβ was examined as a candidate transcription factor in the transcriptional regulation of the StAR gene.

In the present study, we examined the StAR promoter for potential binding sites for transcription factors that may be involved in the regulation of StAR transcription. Two putative C/EBP response elements were identified; one of these sites was shown to bind to C/EBPβ, and we determined that the other site was a low-affinity protein-binding site. Functional analysis revealed that mutation of these sites decreased basal and cAMP-stimulated activity from the StAR promoter in MA-10 cells. Furthermore, we report that SF-1-dependent transactivation of the StAR promoter in COS-1 cells required these putative C/EBP response elements, suggesting that C/EBPβ and SF-1 may interact to regulate StAR gene transcription.

RESULTS

Characterization of the C/EBP-Binding Sites within the Mouse StAR Promoter

Previous studies have established that the elements required to confer cAMP responsiveness to the mouse StAR gene are contained within the first 254 bp of the promoter (10). Therefore, we examined this promoter region from mouse (10), rat (28), human (14), ovine (J. L. Juengel and G. D. Niswender, personal communication) and porcine (29), to identify protein-binding sites that might mediate this response. The 5′-flanking regions of the StAR gene from these five species were aligned (Fig. 1A) and revealed that the greatest degree of sequence similarity was found within 120 bases of the transcription start site. Previously identified elements in this region included a TATA box at nucleotide −35 (all coordinates are relative to the transcriptional start site in mouse at +1) and three SF-1-binding sites corresponding to −42, −91, and −135 in the mouse sequence (10, 12, 14, 28). Several additional blocks of homology were revealed by these analyses. Two of these (named C1 and C2) displayed significant homology to the consensus binding site for members of the CCAAT/enhancer binding site family of transcription factors (ATTGCGCAAT; Ref. 22). Naturally occurring C/EBP sites generally exhibit divergence from this consensus sequence, but typically retain one well conserved half-site (22). The C1 site is centered at −113 and contains one perfect half-site (GCAAT) and an overall 7 of 10 match to the consensus sequence (Fig. 1B). The C1 site is almost completely conserved in all species (9 of 10 matches). The second potential C/EBP site (C2) also displays 7 of 10 matches to the consensus sequence in mouse (Fig. 1B) and contains an almost perfect half-site (ACAAT). However, this sequence is only partially conserved in the five species shown here (Fig. 1A).

Figure 1.

Identification of Two Putative C/EBP-Binding Sites in the StAR Promoter A, The sequences of the 5′-flanking regions of the mouse (M), rat (R), ovine (O), porcine (P), and human (H) StAR genes were aligned using the ClustalW (K. C. Worley, Human Genome Center, Baylor College of Medicine, Houston, TX) sequence alignment program, and identical bases in all five sequences are indicated with asterisks. Binding sites for SF-1 are shaded and labeled. Two putative C/EBP-binding sites were identified based on their similarity to the consensus C/EBP-binding site and are shaded and labeled. Other highly conserved sequences discussed in the text are boxed and labeled A and B. B, The two putative C/EBP-binding sites from the mouse promoter were compared with the consensus C/EBP-binding site determined by binding site selection (26 ). Sequence identities are indicated with vertical bars between the sequences. Both C1 and C2 share seven bases with the consensus site.

The sequence alignment in Fig. 1A revealed the presence of two additional well conserved sequence elements that have not yet been characterized (Fig. 1A, boxed regions A and B). The first is a 10-bp element centered at −63 of the mouse promoter, which resembles a binding site for members of the GATA family of transcription factors. Although GATA-4 and a testis-specific version of GATA-1 are expressed in testis and MA-10 Leydig cells, it is not clear whether they are involved in regulating the StAR gene (30, 31). The second putative element is a perfectly conserved 6-bp sequence (TGATGA) centered at −53 of the mouse promoter, which does not match the binding site of known transcription factors in the transfac matrix table (release 3.2) when searched using TFSEARCH (version 1.3) (© 1995 Yutaka Akiyama, Kyoto University, Kyoto, Japan). Further analyses are clearly required to address the possible roles of these elements in regulating StAR gene expression.

DNAse I footprint analysis of the mouse StAR promoter from− 66 to −254 was performed to identify possible transcription factor-binding sites in this region. A schematic diagram of the radioactive probe used in the footprint analysis is presented in Fig. 2A. Addition of 25 and 50 μg of nuclear extract prepared from (Bu)₂cAMP-stimulated MA-10 cells revealed a broad region of protection interspersed with two DNAse I-hypersensitive sites, indicated by arrows (Fig. 2B). The protected regions included the C1 site and the SF-1 element at −135 and are marked by vertical bars adjacent to the sequence (Fig. 2B).

Figure 2.

DNAse I Footprint Analysis of the StAR Promoter Revealed a Protected Region Encompassing the Putative C/EBP Site at −113 A double-stranded probe corresponding to the coding strand of the StAR promoter spanning −254 to −66 was radiolabeled and used in DNAse I footprinting assays to detect regions of the StAR promoter that were protected from DNAse I digestion by the presence of proteins bound to the promoter. A, Schematic representation of the StAR promoter from −254 to −66 indicating the SF-1, C1, and C2 elements. The line depicts the DNAse I-protected region that is flanked by the hypersensitive sites. B, Radiolabeled probe was incubated for 30 min in the absence (−) or presence of 25 μg or 50 μg of nuclear extract purified from (Bu)₂cAMP-treated MA-10 cells and then DNAse I treatment was for 15 or 30 sec as indicated. The arrows indicate the hypersensitive sites, and the vertical lines indicate the protected regions. The StAR promoter sequence is shown on the left with the vertical lines again indicating the protected regions. The GCAAT sequence is within the protected regions, and the C1 site used for EMSA spans −124 to −101. The SF-1 element at −135, CCACCTTGG, is shown in the sequence within a protected region. The C2 region begins at −95 and continues to the undigested probe. Four separate DNAse I footprint analyses have qualitatively shown the same results: the presence of two hypersensitive sites and protection of the C1 region. The lowercase c indicates a base change in the sequence due to the engineered BglII site.

We next performed electrophoretic mobility shift assays (EMSAs) to identify proteins in Leydig cells that bind the C1 and C2 sites. Radioactively labeled oligonucleotides containing the C1 or C2 binding sites were incubated with nuclear extracts prepared from resting and (Bu)₂cAMP-stimulated MA-10 cells. Protein-DNA complexes with similar mobilities were formed on both the C1 and C2 oligonucleotides, although complex formation occurred more efficiently on the C1 oligonucleotide (Fig. 3). Competition binding studies were carried out to determine whether the protein-DNA complexes represented specific interactions and whether the proteins binding to both sites were related (Fig. 3). Addition of 100-fold molar excess of unlabeled C1 oligonucleotide abolished formation of the protein-C1 complex (Fig. 3, lane 2). However, addition of a similar amount of a related oligonucleotide (C1m; Fig. 3, lane 3), in which residues within the predicted C/EBP recognition sequence had been mutated, failed to prevent complex formation, indicating that proteins were specifically binding to the potential C/EBP motif. Cross-competition assays were also performed using the C1 and C2 oligonucleotides. Unlabeled C1 blocked complex formation on the C2 oligonucleotide (Fig. 3, lane 7), whereas C1m did not (Fig. 3, lane 8), suggesting that proteins with similar DNA recognition specificities bound these two sites. The unlabeled C2 oligonucleotide was incapable of competing for complex formation with the C1 or the C2 oligonucleotide (Fig. 3, lanes 4 and 9), probably reflecting the apparent greater protein-binding affinity of the C1 oligonucleotide as compared with C2. Collectively, these data indicate that the C1 and C2 motifs appear to be binding sites for identical or related proteins in MA-10 nuclear extracts.

Figure 3.

The Putative C/EBP-Binding Sites Form Complexes with Proteins Expressed in MA-10 Nuclear Extracts Five micrograms of protein prepared from nuclear extracts of MA-10 cells stimulated for 6 h with (Bu)₂cAMP were incubated with ³²P-labeled probes representing the putative C/EBP sites at −113 (C1) or −87 (C2) in the presence or absence of 100-fold molar excess of unlabeled competitor oligonucleotides. The competitor oligonucleotides were either the unlabeled wild-type C/EBP binding sites (C1, C2) or mutant versions of these sites (C1m, C2m). DNA-protein complexes were subjected to electrophoresis through a 4% nondenaturing polyacrylamide gel, and then dried gels were visualized by phosphoimagery and autoradiography. The inset at the bottom of the gel is a phosphoimage of the region of the gel maked with an asterisk, which shows binding to the C2 oligonucleotide.

Having established that the C1 site binds with high affinity to proteins present in nuclear extracts from stimulated MA-10 cells, we next sought to determine whether C/EBPβ binds to C1, and whether this binding was regulated by (Bu)₂cAMP stimulation of the cells. Supershift experiments were performed using an antiserum raised against the amino terminus of C/EBPβ (32). Initially recombinant C/EBPβ was produced in rabbit reticulocyte lysates and incubated with radiolabeled C1 oligonucleotide in the absence and presence of the C/EBPβ antiserum. A specific DNA-protein complex was formed when the C1 probe was mixed with recombinant C/EBPβ that was absent in unprogrammed lysates (compare Fig. 4, lanes 1 and 2). This complex was completely supershifted by the C/EBPβ antiserum (lane 3). Nuclear extracts prepared from both (Bu)₂cAMP-stimulated and unstimulated MA-10 cells formed complexes with the C1 probe, and further incubation with the C/EBPβ antiserum resulted in a supershifted complex in both extracts (Fig. 4, lanes 5 and 7). The intensity of this complex was relatively unaffected by (Bu)₂cAMP stimulation (Fig. 4, compare lanes 4–5 with 6–7 and 8–9 with 10–11), suggesting that there are proteins other than C/EBPβ in these complexes, whose levels are not altered by (Bu)₂cAMP-stimulation, that may be rate-limiting in the formation of the complex. The major shifted complex in MA-10 nuclear extracts could be resolved into three distinct bands (Fig. 4, right panel, labeled a, b, and c), and the supershifted complex appeared to be derived primarily from the middle band. Multiple C/EBP-like binding activities have been observed in nuclear extracts from many cell types, which are unaffected by addition of currently available antisera directed against known C/EBP family members. These complexes may represent unrelated proteins with DNA-binding specificities similar to C/EBP, or heterodimeric or posttranslationally modified complexes that are not recognized by C/EBP antisera. The C/EBPβ-C1 complex (band b), which contributes to the supershifted band does not migrate with the complex formed with recombinant C/EBPβ, indicating that C/EBPβ may exist in a heterodimeric form in MA-10 cells. At present, the putative partner(s) for C/EBPβ are unknown.

Figure 4.

C/EBPβ Binds to the C1 Site In vitro transcribed and translated C/EBPβ or 5 μg of protein isolated from nuclear extracts prepared from unstimulated or (Bu)₂cAMP-stimulated MA-10 cells were incubated with a ³²P-labeled C1 oligonucleotide and then incubated in the presence or absence of 1 μl of a C/EBPβ-specific antiserum. DNA-protein complexes were subjected to electrophoresis through a 4% nondenaturing polyacrylamide gel for approximently 2 h at 200 V, and then dried gels were visualized by autoradiography. SS indicates the position of the supershifted band. The right panel is a example of a 4% nondenaturing polyacrylamide gel that was run for 1.5 h at 200 V and was exposed to x-ray film for less time than the example in (a) to achieve a higher resolution of the bands, which shows that three bands, labeled a, b and c, resolve in the major shifted complex.

Since C/EBPα and C/EBPβ share binding site preferences (22, 33), and have both been shown to be expressed in reproductive tissues (34, 35), we examined their expression in MA-10 cells. By Western analysis, we have determined that C/EBPβ, but not C/EBPα, could be detected in MA-10 cells, and that MA-10 nuclear extracts did not contain C/EBPα-binding activity as evidenced by EMSA using the C1 oligo (data not shown).

C/EBP DNA Elements Are Required for Activation of the StAR Promoter

To assess the role of the C1 and C2 sites in StAR promoter function, we compared the activity in MA-10 cells of the wild-type StAR promoter to mutants carrying changes in either or both of these sites. The mutations were tested in the context of a 966-bp fragment of the StAR gene that had previously been shown to support basal and cAMP-inducible expression in MA-10 cells (10). The same mutations in the C1 and C2 sites used in the EMSAs were introduced into the StAR promoter either alone or in combination. Each construct was transfected into MA-10 cells, and luciferase activities were measured in untreated cells and cells incubated in the presence of (Bu)₂cAMP for 6 h. The wild-type promoter construct (−966 StAR Luc) displayed low basal activity, which was stimulated 6.2-fold by (Bu)₂cAMP (Fig. 5). Mutation of either the C1 or C2 site alone (−966 StAR C1m and −966 StAR C2m) resulted in significantly lower basal activities, 20% and 15% of the wild-type value, respectively. Mutation of both the C1 and C2 sites (−966 StAR C1m, C2m) resulted in a further decrease in basal promoter activity to 10% of the wild-type activity; however, the cAMP responsiveness of the promoter was again relatively unchanged (Fig. 5). Although the absolute cAMP-induced activities of the mutated reporters was lower than the wild-type promoter, the fold activation of all four reporters was similar. These data indicate that the C1 and C2 sites are important for high-level basal expression from the StAR promoter.

Figure 5.

C/EBP DNA Elements Are Required for Activation of the StAR Promoter in MA-10 Cells MA-10 cells were transfected with 2 μg of either StAR −966, StAR −966C1m, StAR −966C2m, or StAR −966C1m,2 m, and then reporter activity was measured from (Bu)₂cAMP-stimulated or unstimulated cells. In all cases, 75 ng of pRL-SV40 were also transfected as a control for transfection efficiency. Data are represented as reporter activity divided by the activity of SV40-RNU. Fold activation represents the (Bu)₂cAMP-stimulated reporter activity divided by the unstimulated level. Data represent averages ± sem from three experiments including the StAR −966C1m,2 m reporter, and two additional experiments not including the StAR −966C1m,2 m reporter, all of which were normalized to the activity of the StAR −966 reporter from unstimulated cells in each experiment. Statistical analysis between control (StAR −966) and mutant reporters revealed some significant differences as determined by Student’s unpaired two-tailed t tests (#, P < 0.05 between unstimulated reporters; *, P < .05 between (Bu)₂cAMP-stimulated reporters).

SF-1 Requires C/EBP DNA Elements to Activate the StAR Promoter

To better understand the role of C/EBPβ in StAR gene transcription, we conducted transactivation experiments in COS-1 cells, a nonsteroidogenic cell line that does not express C/EBPβ (S. C. Williams, unpublished observation), or SF-1 (14). C/EBPβ and SF-1 expression vectors were cotransfected, either alone or in combination, with the wild-type and mutant promoter constructs mentioned above. StAR promoter activity was stimulated approximately 2-fold by C/EBPβ, and this effect was lost or diminished when either the C1 or C2 sites were mutated (Fig. 6; p-966 StAR C1m and C2m). A StAR promoter construct carrying mutations in both sites (p-966 StAR C1m,C2m) was completely unresponsive to C/EBPβ. SF-1 stimulated StAR promoter activity approximately 5-fold; however, mutating the C1 site, and especially the C2 site, diminished this activity (Fig. 6). In fact, the promoter construct bearing mutations in the C2 site was totally unresponsive to SF-1, despite the fact that the predicted SF-1 sites in this promoter remain intact. Coexpression of C/EBPβ and SF-1 did not result in additive or synergistic activation of the StAR promoter (Fig. 6). This observation may be due to the high levels of both C/EBPβ and SF-1 protein in the transiently transfected cells, resulting in squelching or titration of required cofactors, leading to a decrease in reporter activity. These data indicate that C/EBPβ can stimulate the activity of the StAR promoter and that efficient SF-1-dependent activation of this promoter requires intact C1 and C2 sites.

Figure 6.

C/EBP Sites Are Required for SF-1 to Transcativate the StAR Promoter COS-1 cells were transfected with 2 μg of either StAR −966, StAR −966C1m, StAR −966C2m, or StAR −966C1m,2 m, and cotransfected with 2 μg of either or both C/EBPβ or SF-1 expression plasmids, and reporter activity was measured. Data represent averages ± sem from three experiments that were normalized to the activity of StAR −966 reporter alone. All values were compared with the p-966 StAR Luc reporter cotransfected with the control vector by Student’s one-tailed unpaired t test (*, P < 0.05).

C/EBPβ and SF-1 Physically Interact in Vitro

The data presented above indicate a functional interaction between SF-1 and proteins binding to the C1 and C2 sites. Due to the close proximity of SF-1 and putative C/EBPβ binding sites, we next examined whether these proteins might physically associate. Bacterially expressed glutathione S-transferase (GST) or a chimeric GST-SF-1 protein was mixed with radiolabeled recombinant C/EBPβ in the presence of a portion of the StAR promoter (−5 to −158) and a reversible protein cross-linking agent. After purification on glutathione-agarose beads, the remaining proteins were resolved by SDS-PAGE (Fig. 7). Recombinant C/EBPβ was specifically retained by GST-SF-1, indicating that SF-1 and C/EBPβ are likely to physically associate, and that this association may be necessary for efficient activation of the StAR promoter.

Figure 7.

SF-1 and C/EBPβ Physically Interact in Vitro Bacterially expressed GST-SF-1 or GST was incubated with radiolabeled recombinant C/EBPβ in the presence of a fragment of the StAR promoter spanning −5 to −158. Proteins were reversibly cross-linked and purified using glutathione-agarose beads and then subjected to SDS-PAGE.

DISCUSSION

The promoter regions of eukaryotic genes are generally composed of multiple binding sites for transcriptional activators and repressors that act in combination to regulate expression of a linked gene (36–38). Comparison of the 5′-flanking sequence of the StAR gene revealed the existence of several blocks of conserved sequences within 120 bp of the transcription start site. Previous analyses had revealed the existence of binding sites for the orphan receptor, SF-1, and a variant TATA-like element located 25–30 bp from the start site. We report here the identification of two novel elements required for high-level expression from the StAR promoter in Leydig cells.

The upstream element, C1, is strongly bound by recombinant C/EBPβ and by C/EBPβ in MA-10 nuclear extracts and appears to be the primary site through which C/EBPβ activates the StAR promoter. We have considered two putative, nonexclusive functions for C/EBPβ in StAR gene regulation, namely, the activation of StAR gene expression during development and the rapid activation of StAR transcription in response to trophic hormone. The observation that mutation of the C1 site did not abolish cAMP induction (fold activation) of the StAR promoter suggests that either C/EBPβ does not mediate the cAMP-dependent regulation of StAR gene expression, at least during the acute phase of stimulation, or (Bu)₂cAMP stimulation of Leydig cells does not completely mimic all of the effects of hormone stimulation. However, mutating the C1 site decreased basal level activity from the StAR promoter to 20% of the wild-type level. This finding, combined with our previous observation that both C/EBPβ and StAR protein levels increase during Leydig cell development (23), indicates a role for C/EBPβ in developmental regulation of StAR gene expression. In support of a broad role of C/EBPβ in the developmental regulation of steroidogenesis, analysis of the promoter regions of genes encoding steroidogenic enzymes in Leydig cells, such as 3β-hydroxysteroid dehydrogenase (3βHSD), cytochrome P450 side-chain cleavage (P450scc), and 17,α-hydroxylase (CYP17), has revealed the presence of putative C/EBP sites (our unpublished observation), although functional studies of these sites are lacking at present. Therefore, C/EBPβ may participate in the regulation of multiple Leydig cell genes during development.

Our mutational analysis also identified a second element (C2) that is required for high-basal level expression of the StAR promoter. Although the C2 site is not as highly conserved as the C1 site, it appears to be at least equally important for StAR gene transcription as its mutation decreased promoter activity to 15% of wild-type levels. We initially considered the C2 site to be a binding site for C/EBPβ based on the presence of an almost perfect half-site in the mouse, porcine, and human genes, and a 7 of 10 match to the consensus C/EBP binding site in the mouse gene. However, the C2 site complexed weakly with nuclear extract from MA-10 cells and was unable to compete for C/EBP binding to the C1 element. In addition, mutation of the C2 site in the StAR promoter had only a slight negative effect on transactivation by C/EBPβ in COS-1 cells. These data could be interpreted in two ways. First, the C2 site may not be a bona fide C/EBPβ binding site in vivo, instead serving as the binding site for another, as yet unidentified, protein. Some candidate proteins might be members of the CCAAT box-binding protein families, such as the constitutively expressed nuclear factor-Y [NF-Y; a heterotrimer of NF-YA, NF-YB, and NF-YC (39, 40)]. Second, the C2 site may be a weak binding site for C/EBPβ, and efficient usage of this site by C/EBPβ may require the presence of cooperating factors such as SF-1. In support of this hypothesis, a cryptic C/EBP-binding site is present in the promoter of the liver-specific cytochrome P450 2D5 (2D5) gene as part of a bi-partite binding site for C/EBPβ and Sp1 (41). C/EBPβ is unable to bind to, or to activate transcription through, this element in the absence of Sp1, and the selective interactions with Sp1 explain the difference in the ability of C/EBPβ and C/EBPα to activate the 2D5 promoter (41). Additional studies are required to identify the proteins that bind the C2 site in vivo and whether C/EBPβ must interact with other proteins to bind to C2.

We (in this report) and others (14) have demonstrated that exogenously expressed SF-1 is capable of transactivating the StAR promoter in COS-1 cells. However, SF-1-dependent activation was diminished or lost when either one or both the C1 or C2 sites were mutated, despite the fact that these mutations would not be expected to significantly affect SF-1 binding to its cognate sites. Furthermore, we have shown that SF-1 and C/EBPβ associate in vitro. We interpret these results to indicate that SF-1 physically interacts with C/EBPβ and possibly other proteins bound to these sites and that these interactions are a prerequisite for SF-1 action. These interactions could be direct protein-protein interactions or may involve the recruitment of accessory factors such as coactivators to the promoter. In regard to the former possibility, C/EBPβ has been shown to functionally and/or physically interact with numerous members of the steroid hormone superfamily, including the estrogen receptor (42), glucocorticoid receptor (43), and hepatocyte nuclear factor-4[ HNF4 (44)]. SF-1 has also been shown to interact with a number of proteins that may be involved in the transcriptional regulation of the StAR gene, including the steroid receptor coactivator-1[ SRC-1/NCoA-1 (45, 46)] and DAX-1 [for dosage-sensitive sex reversal-adrenal hypoplasia congenita critical region on the X chromosome, gene 1 (47)]. The requirement of intact C1 and C2 sites for efficient SF-1-mediated transactivation of the StAR promoter and the physical interaction observed between SF-1 and C/EBPβ distinguishes the StAR promoter as an important model with which to investigate how transcription factors may cooperate to regulate transcription.

A central unanswered question surrounding the StAR gene is the mechanism by which the promoter responds acutely to trophic hormone stimulation and, more specifically, how the interactions between transcription factors bound to the StAR promoter affect cAMP-dependent regulation of the StAR gene. We have shown in this report that the complex of proteins bound to the C1 site is relatively unaffected by (Bu)₂cAMP stimulation. We have also reported that disruption of the C1 and C2 sites interfered with basal (unstimulated) transcription of the StAR gene, and that these sites were required for SF-1-dependent transcription from the StAR promoter. Additionally, it has been shown that StAR transcriptional activation does not require de novo protein synthesis (10). Collectively, these findings indicate that C/EBPβ (and/or other proteins) bound to the C1 and C2 sites interacts with SF-1 regardless of cAMP stimulation, and that cAMP stimulation regulates a step distal to the formation of this complex, which, in turn, may activate transcription from the StAR promoter. There are several nonexclusive mechanisms by which this may transpire. For example, upon stimulation with trophic hormone, increased cAMP levels cause the release of the catalytic subunit of PKA, which can enter the nucleus (48). Posttranslational modifications have been shown to activate C/EBPβ independently from its ability to bind DNA (25), and a specific target residue for PKA has been identified in the C/EBPβ basic region (27). This posttranslational modification of C/EBPβ may activate transcription either directly through C/EBPβ or through recruitment of coactivators to the promoter. An alternative model involves the orphan nuclear receptor DAX-1. DAX-1 has been shown to be a powerful repressor of StAR promoter activity, through binding to a hairpin loop structure located proximal to the C2 site (49). Our data demonstrate that similar complexes, which include C/EBPβ, are formed on the C1 site regardless of cAMP stimulation. SF-1 and other factors may also bind to the promoter in the absence of cAMP stimulation, which should result in a high level of basal transcription from the promoter. Since DAX-1 has been shown to repress StAR, its presence on this promoter could effectively inhibit transcription, even in the presence of positive factors such as SF-1, Sp1, and C/EBPβ. Upon cAMP stimulation, DAX-1 may be displaced from the promoter, or disassociated from corepressors, allowing high levels of transcription from the StAR promoter. DAX-1 null mice have recently been described, and steroidogenesis appears to be relatively normal (50). The observed phenocopy appears to be less severe than expected, given that mutations in the DAX-1 gene in humans results in X-linked, adrenal hypoplasia congenita (AHC). The precise role of DAX-1 in the regulation of the StAR gene remains a most interesting question, which clearly requires additional study and a more detailed analysis of the DAX-1 null mice. Work is currently underway to study in more detail the direct interaction between C/EBPβ and SF-1 and to study whether transcriptional coactivators and repressors are involved in SF-1- or C/EBPβ-mediated regulation of the StAR gene.

MATERIALS AND METHODS

Cell Culture

The MA-10 mouse Leydig tumor cell line was a generous gift from Dr. M. Ascoli (Department of Pharmacology, University of Iowa College of Medicine, Iowa City, Iowa). The cells were grown in Waymouth’s MB/752 medium containing 15% horse serum and 40 μg gentamycin sulfate/ml (referred to as WAY+). COS-1 African green monkey kidney cells were obtained from the American Type Culture Collection (Manassas, VA) and were maintained in DMEM supplemented with 10% FBS and 100 U of penicillin/ml and 10 U of streptomycin sulfate/ml. All cells were grown at 37 C in a humid atmosphere of 5% CO₂. Media, additives, and serum were purchased from Gibco BRL (Gaithersburg, MD).

Plasmids and Construction of StAR Promoter Mutants

pMEX C/EBPβ has been described previously (51), and the SF-1 expression plasmid was a generous gift of Dr. Keith Parker (University of Texas, Southwestern Medical School, Dallas, TX).

Site-directed mutagenesis was performed to mutate both of the C/EBP-binding sites in the StAR promoter. The Gene Editor kit (Promega Corp., Madison, WI) was used to introduce mutations into p-966 Luc (described in Ref. 10); referred to in this study as Star −966) to eliminate the C/EBP elements such that the mutated sequences would contain a novel SalI restriction site. The oligonucleotides used to introduce the mutations were as follows (mutations are underlined):

C1m: CACTGCAGGATGGTCGACTCATTCCATCCT

C2m: CTTGACCCTCTGGTCGACGACTGATGACTT

C1,2m: GCACTGCAGGATGGTCGACTCATTCCATCCTT - GACCCTCTGGTCGACGACGATGAC.

Resulting plasmids were partially sequenced to confirm that the C/EBP-binding sites had been mutated as expected.

Transfections

MA-10 and COS-1 cells were transfected by electroporation. Briefly, 350 μl of a suspension of cells (12.5 × 10⁶ cells/ml) were mixed with various amounts of effector and reporter plasmids along with sheared salmon sperm DNA (Sigma Chemical Co., St. Louis, MO.) as carrier DNA to equalize the total amount of DNA electroporated to 70 μg in each electroporation. For the transfection studies in MA-10 cells, 75 ng of pRL-SV40 vector (a plasmid that constitutively expresses Renilla luciferase under the control of the SV40 promoter; Promega Corp.) was also transfected in all cases as a transfection control. Cells were electroporated in cuvettes (Invitrogen, Carlsbad, CA) with a gap width of 4 mm, using the electro cell manipulator 600 (BTX Inc., San Diego, CA) using the following parameters: capacitance = 960 μFarads; voltage = 250 V; resistance = 129 Ω, yielding an electroporation time of 20–30 msec. After electroporation, 1 ml normal growth medium was added to the cuvette, and the cells were incubated for 15 min at room temperature (RT). The cells were then brought to 12 ml in normal growth medium, and 2 ml were placed in each well of a six-well (35-mm) dish. Twenty-four hours after electroporation, the medium was replaced with fresh medium. Twenty-four hours later, three wells of each six-well plate were treated for 6 h with 1 mm (Bu)₂cAMP (Sigma Chemical Co., St. Louis MO), in 1 ml of WAY+, while the control wells received only 1 ml of WAY+. After (Bu)₂cAMP stimulation the cells were harvested for luciferase assays as described below.

Luciferase Assays

Extracts for luciferase assays were prepared using luciferase assay system reporter lysis buffer (Promega Corp.). At the time of harvesting, medium was removed, and the cells were rinsed three times with ice-cold PBS. Reporter lysis buffer (250 μl) was added to the cells, and the cells were scraped into 1.5-ml centrifuge tubes. The cellular debris was then pelleted by centrifugation at 13,800 × g at 4 C, and the supernatant fluid was placed in a 1.5-ml centrifuge tube and was either used immediately or stored at −80 C.

Luciferase assays were performed using luciferase or dual luciferase assay kits (Promega Corp.) exactly as described in the protocol provided with the kit. Relative light units were measured using a Monolight 2010 luminometer (Analytical Luminescence Laboratory, San Diego, CA). Student’s unpaired one-tailed or two-tailed t tests were performed using Statview SE+ graphics software (Abacus Concepts, Berkeley, CA).

EMSA

Nuclear extracts were prepared from confluent cell cultures as described (52). Briefly, cell monolayers were rinsed three times with ice-cold PBS and scraped in 1 ml of PBS into 1.5-ml centrifuge tubes. The cells were pelleted by centrifugation at 1500 × g for 3 min. The pellets were resuspended in 400 μl of buffer A (10 mm HEPES, pH 7.9; 10 mm KCl; 0.1 mm EDTA; 0.1 mm EGTA; 1 mm dithiothreitol; 1 mm phenylmethylsulfonyl fluoride). The cells were swelled for 15 min at 4 C, and then 25 μl of 10% NP-40 were added and the tubes were vortexed. The homogenates were centrifuged 30 sec at 13,800 × g in a microfuge to pellet the nuclei; then 50 μl of buffer C (20 mm HEPES, pH 7.9; 0.4 m NaCl; 1 mm EDTA; 1 mm EGTA; 1 mm dithiothreitol; 1 mm phenylmethylsulfonyl fluoride) was added, and the samples were vigorously rocked for 15 min at 4 C. The nuclear lysate was then centrifuged for 5 min at 13,800 × g in a microfuge at 4 C, and the supernatant fluid was placed into a fresh microfuge tube and stored at −80 C or used immediately. The pSVSportC/EBPβ plasmid (provided by Dr. Elmus Beale, Texas Tech University, Health Sciences Center, Lubbock, TX) was transcribed using SP6 polymerase and was translated using the TNT kit (Promega Corp.). The double-stranded DNA probes used were C1 (C/EBP binding site at −113) and C2 (C/EBP binding site at −87), and mutants of C1 and C2 in which underlined bases have been mutated to disrupt the specific binding of C/EBP proteins:

C1: GGCTGCAGGATGAGGCAATCATTCCA

C1m: GGCTGCAGGATGGTCGACTCATTCCA

C2: GGGACCCTCTGCACAATGACTGATG

C2m: GGGACCCTCTGGTCGACGACTGATG

To generate radioactive probes, the sense and antisense oligonucleotides were heated to 75 C for 5 min and then slowly cooled over 2 h to room temperature in annealing buffer [10 mm Tris-HCl (pH 7.5), 100 mm NaCl, 1 mm EDTA]. 5′-GGG overhangs present in the double-stranded oligonucleotides were filled in using α [³²P] dCTP 3000 Ci/mmol (DuPont NEN, Boston, MA) and Klenow (Promega Corp.) at 37 C for 30 min. The ³²P-labeled probes were purified using Probe Quant spin columns (Pharmacia Biotech, Piscataway, NJ). Binding reactions were performed by mixing 5 μg of nuclear extract with a binding cocktail containing 4% Ficoll, 10 mm HEPES (pH 7.9), 1 mm EDTA (pH 8.0), and 1 μg poly (dI:dC) and the labeled probe at a final concentration of 5 nm in 15 μl. Where noted, the protein was first incubated 20 min at room temperature, in binding cocktail with 100-fold molar excess of the unlabeled competitor DNA before addition of the labeled DNA. For the supershift experiments, the binding reaction was performed as described above for 20 min, after which 1 μl of the C/EBPβ antiserum was added and the reaction was incubated an additional 20 min at room temperature. After the binding reaction, the entire reaction was subjected to electrophoresis through a 4% nondenaturing polyacrylamide gel. The gel was then dried and autoradiography and phosphorimagery (Molecular Dynamics, Inc., Sunnyvale, CA) were performed.

DNAse I Footprint

To generate a radiolabeled DNA probe, the region spanning −254 to −35 of the StAR promoter was amplified using the following oligonucleotide primers; 5′-primer is a 20 mer that spans bases −254 to −235 and has additional bases at the 5′-end to generate an MluI restriction endonuclease site, and the 3′-primer is a 44 mer that spans −101 to −35 and; contains point mutations to generate a BglII resriction endonuclease site centered at base −63 and a XhoI resriction endonuclease site centered at −95. The amplification product was cloned into the MluI-SmaI sites of the pSport vector (Life Technologies, Gaithersburg, MD), and the sequence was verified by the dideoxynucleotide sequencing method of Sanger using the T7 Sequenase Kit Version 2 (Amersham Pharmacia Biotech, Arlington Heights, IL). The StAR promoter fragment was excised from pSport by MluI and KpnI digestion and gel purified using QIAquick gel extraction kit (Qiagen, Chatsworth, CA) and treated with calf intestinal phosphatase (CIP, Promega Corp.). CIP was inactivated by phenol-cholorform extraction, and the DNA was precipitated with ethanol. Two picomoles of probe were radiolabed usingγ -[³²P]ATP (DuPont NEN, Boston MA) and T4 polynucleotide kinase (Promega Corp.) followed by inactivation of the kinase and digestion with BglII. The resultant probe is labeled on the coding strand and spans −254 to −66 of the StAR promoter. The probe was purified by phenol-chloroform extraction and ethanol precipitation, after which DNAse I footprint analysis was performed using the Core Footprinting System (Promega Corp.) with minor modifications. In brief, 20–40 fmol of probe (50K-100K cpm) were added to the DNA protein-binding reaction (10 mm Tris/Cl, pH 8.0, 150 mm KCl, 2.5 μg poly dI:dC, 4 μg/ml calf thymus DNA, 10% glycerol, and 25–50 μg MA-10 nuclear extract. The reaction was incubated on ice for 30 min and then transferred to 25 C, and CaCl₂ and MgCl₂ were added to a final concentration of 2.5 mm and 5 mm, respectively. One unit of DNAse I (Promega Corp.) was added to the reaction and was incubated for 15 sec in the presence of nuclear extract or 30 sec in the presence or absence of nuclear extract. The reactions were stopped by the addition of an equal volume of stop buffer containing 200 mm NaCl, 30 mm EDTA, 1% SDS, and 100 μg/ml yeast RNA, and the DNA was recovered by phenol-chloroform extraction and ethanol precipitation. The reactions were resuspended in formamide loading buffer, and the DNA was resolved on a 6% polyacrylamide sequencing gel. The gel was dried and exposed to x-ray film. Maxam and Gilbert (53) chemical sequencing reactions were performed using 40 fmol of probe following standard protocols (54).

GST Pull Down Assay

A GST-SF-1 fusion protein was prepared and isolated according to standard protocols (52). The plasmid containing the SF-1 coding sequence in the pGEX-1λT vector (Pharmacia Biotech, Piscataway, NJ), was described previously (55), and the control GST plasmid was pGEX4T-3 (Pharmacia Biotech). Both GST and GST-SF-1 were subjected to PAGE and stained with Coomassie blue. The appearance of a band at the correct molecular weight confirmed that intact proteins were produced. In vitro transcribed and translated C/EBPβ was prepared as described above, except that the STP3 kit (Novagen, Madison, WI) was used in the presence of ³⁵[S]methionine according to protocols supplied by the manufacturer. A binding reaction was prepared containing 10 μl of in vitro transcribed and translated ³⁵S-labeled C/EBPβ, 20 μg of the GST-SF-1 or GST protein bound to 70 μl of glutathione linked to beaded agarose (Sigma Chemical Co.), 1 μg poly (dI:dC) in the binding cocktail described in the EMSA methods above, with 50 ng of a PCR product spanning −5 to −158 of the mouse StAR promoter (primer sequences avaliable upon request). The binding reaction was incubated 20 min at 4 C, DTSSP (Pierce Chemical Co., Rockford, IL), a reversible cross-linking reagent, was added to a final concentration of 5 mm and incubation continued for an additional 20 min at 4 C. The binding reactions were then centrifuged at 13,000 × g, the supernatant was removed, and the pellets were washed four times in the binding cocktail. Pellets were resuspended in denaturing sample buffer and incubated at 100 C for 10 min and subjected to SDS-PAGE. Autoradiography and phosphorimagery were performed on dried gels.

Acknowledgments

The authors would like to thank Dr. Steven King for many helpful discussions and Dr. Joseph Orly for sharing data before publication and for helpful discussions. We acknowledge the technical assistance of Deborah Alberts, Matthew Dyson, Rebecca Combs, Darrell Eubank, and Demet Nalbant. We also thank Drs. Mark McLean, Holly LaVoie, and Jennifer Juengel for providing us with StAR promoter sequences before publication and Dr. Keith Parker for providing us with plasmids.

This research was supported by NIH Grants HD-17481 (D.M.S.) and DK-51656 (B.J.C.) and a Scientist Development Grant from the American Heart Association (S.C.W.).