The Id2 transcriptional repressor is induced by follicle-stimulating hormone and cAMP.

Id (inhibitor of DNA binding/differentiation) proteins repress differentiation and promote cell division by dimerizing with and inhibiting the action of basic helix-loop-helix transcription factors including those that bind to E-box motifs. Of the four characterized Id proteins, only Id2 is found in the nucleus of Sertoli cells that support the development of spermatozoa in the testis. Differential display analysis of rat primary Sertoli cell mRNA identified Id2 as being inducible by forskolin, a stimulator of cAMP production. Northern blot analysis confirmed that Id2 mRNA expression peaked in Sertoli cells 6-12 h after stimulation with forskolin or follicle-stimulating hormone (FSH), the major physiological stimulator of cAMP in Sertoli cells. Similarly, Id2 promoter activity in Sertoli cells was induced after forskolin or FSH stimulation as well as by overexpression of protein kinase A. Forskolin induction of the Id2 promoter required sequences located between positions -122 and -82. Protein(s) of 40-45 kDa were found to bind two activated transcription factor/cAMP-response element-like sites and a GATA motif within the regulatory region. The induction of the Id2 gene by FSH corresponded with a decrease in protein binding to an E-box consensus motif and decreased E-box-mediated transcription. Together, these findings raise the possibility that FSH-mediated induction of Id2 and resultant inhibition of basic helix-loop-helix transcription factor-regulated genes in Sertoli cells may contribute to the regulation of spermatogenesis.

The development of spermatozoa within the testis requires specialized hormonal inputs and precise temporal control of gene expression. Follicle-stimulating hormone (FSH) 1 and testosterone are the major hormonal regulators of spermatogenesis (reviewed in Ref. 1). Both hormones act on Sertoli cells to modulate cell processes and gene expression patterns that are required to support developing germ cells. FSH and testosterone actions are transduced through specific receptors present on the cell surface (FSH receptor) or within the cell (androgen receptor). Therefore, the proper expression of receptors for FSH and testosterone in Sertoli cells is essential for the process of spermatogenesis.
Recently, transcription of the FSH receptor and androgen receptor genes was shown to be regulated by members of the basic helix-loop-helix (bHLH) family of transcription factors (2)(3)(4)(5). bHLH proteins are part of a larger group of HLH proteins that share an HLH structural domain (two amphipathic helices separated by a loop) that is required for dimerization with other HLH proteins (6). Adjacent to the HLH domain, bHLH proteins contain a basic region of amino acids that is required to bind DNA. bHLH proteins bind as dimers to DNA sequences in gene promoters sharing the core hexanucleotide sequence CANNTG called an E-box (6). Two broad functional groups of bHLH proteins have been defined by their pattern of expression. Class A bHLH proteins are expressed in most tissues, whereas the expression of class B members is tissuespecific. In general, cell type-specific bHLH proteins form dimers with ubiquitously expressed bHLH proteins.
One major consequence of bHLH protein expression is the activation of genes required for the differentiation of tissues during development as well as the maintenance of normal physiological differentiation (7). For example, the differentiation processes of neurogenesis, myogenesis, and hematopoiesis as well as pancreatic development require bHLH proteins (reviewed in Ref. 7). Furthermore, bHLH proteins are thought to be required for testicular development and Sertoli cell differentiation (8,9).
A subset of HLH proteins called inhibitors of DNA binding/ differentiation or Id proteins represses the actions of bHLH transcription factors. Id proteins contain the helix-loop-helix region required for dimerization but lack the basic DNA binding domain. Therefore, Id proteins act as dominant negative repressors of transcription by dimerizing with and sequestering the ubiquitously expressed class A E-box HLH proteins (10), and in some cases, class B (tissue-specific) HLH proteins (11). Four subtypes of Id proteins have been characterized in mammalian cells, Id1, Id2, Id3, and Id4 (11,12,14,15). In an immunohistochemical analysis of adult mouse testis, Id1, Id2, and Id4 were observed to be expressed in germ cells at various specific stages of development (16). In this same study, Id3 was identified in the cytoplasm of Sertoli cells, whereas Id2 was the only Id protein detected in the nuclei of Sertoli cells. A more recent study detected all four Id mRNAs and proteins in primary Sertoli cells isolated from 20-day-old rats (17). Although the cellular localization of the proteins was not investigated, Id2 and Id3 expression levels were found to be higher than Id1 and Id4. Based on these available data, Id2 is likely the major Id protein regulating transcription in Sertoli cells. Overexpression of Id proteins in Sertoli cells has been shown to activate the E-box-regulated transferrin gene promoter (17) and inhibit the c-Fos and FSH receptor promoters through E-box motifs (2,9). The importance of Id2 for Sertoli cell function and germ cell development is highlighted by the finding that Id2 null mice display defective spermatogenesis (18) in addition to retarded growth and neonatal morbidity (19).
In this study, we present evidence demonstrating that Id2 mRNA expression is transiently induced by the cAMP-elevating agents forskolin and FSH in Sertoli cells. We also show that the Id2 promoter is induced by forskolin, cAMP, or protein kinase A through a motif located 120 to 82 bp upstream of the transcription start site. This region is shown to bind proteins of 41 and 45 kDa. Furthermore, we demonstrate that inhibition of nuclear protein binding to E-box motifs corresponds to the induction of Id2 by FSH and that FSH induction of E-boxmediated transcription is delayed until Id2 levels decrease. These studies suggest that FSH-mediated induction of Id2 expression may be important for modulating the expression of Sertoli cell genes that are required to support spermatogenesis.

EXPERIMENTAL PROCEDURES
Isolation of Primary Sertoli Cells and Cell Culture-Sertoli cells were isolated from 16-day-old Sprague-Dawley rats as described previously (20). Decapsulated testes were digested with collagenase (0.5 mg/ml, 33°C, 12 min) in enriched Krebs-Ringer bicarbonate medium followed by three washes in enriched Krebs-Ringer bicarbonate medium (1 ϫ g, 3 min) to isolate seminiferous tubules. Tubules were digested with trypsin (0.5 mg/ml, 33°C, 12 min), and cell aggregates were passed repeatedly through a drawn-out Pasteur pipette. An equal volume of Dulbecco's modified Eagle's medium containing 10% fetal calf serum was added to the Sertoli cells, which were then pelleted (100 ϫ g, 5 min) and resuspended in serum-free medium containing 50% Dulbecco's modified Eagle's medium, 50% Ham's F12, 5 g/ml insulin, 5 g/ml transferrin, 10 Ϫ6 M retinoic acid, 10 ng/ml epidermal growth factor, 3 g/ml cytosine ␤-D-arabinofuranoside, 2 mM glutamine, 1 mM sodium pyruvate, 100 units/ml penicillin, and 100 g/ml streptomycin. Sertoli cells were cultured on matrigel (Collaborative Research, Bedford, MA)coated dishes (32°C, 5% CO 2 ). Sertoli cells were routinely Ͼ95% pure as determined by phase microscopy and alkaline phosphatase staining. The mouse-derived Sertoli cell line MSC-1, which lacks FSH receptors (21), was cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM glutamine. Animals used in these studies were maintained and sacrificed according to the principles and procedures described in the National Institutes of Health Guide for the Care and Use of Laboratory Animals. These studies were approved by the University of Pittsburgh IACUC committee.
Differential Display-Sertoli cells were either untreated or treated with forskolin (10 M) and 3-isobutyl-1methylxanthine (IBMX) (0.5 mM) for 2 or 12 h, and extracted total RNA was subjected to reverse transcription and PCR amplification using the RNA image kit (GenHunter, Brookline, MA). Briefly, total RNA was incubated with reverse transcriptase and T 11 primer having a G, A, or C 1-bp anchor at the 3Ј end to produce cDNAs. The cDNAs were amplified in the presence of [␣-33 P]dATP using a random 5Ј primer and a 3Ј-oligonucleotide primer "clamped" to the 5Ј end of the cDNA poly(A) tail. Amplified cDNAs were fractionated on 6% sequencing gels, and cDNAs that were reproducibly modulated by forskolin treatment were isolated from the gel and amplified again. Using three separate 3Ј primers and one 5Ј primer, ϳ20 regulated cDNAs were identified. The cDNAs that were induced by forskolin were isolated and sequenced.
Northern Analysis-Primary Sertoli cells from 16-day-old rats were stimulated with either ovine FSH (100 ng/ml) plus IBMX (0.5 mM) or forskolin (10 m) plus IBMX for 2, 6, 12, or 24 h. RNA was isolated by the method of Chomczynski and Sacchi (22). For Northern analysis, RNA (20 g) was fractionated on formaldehyde gels, transferred to polyvinylidene difluoride membranes (Micron Separations, Westboro, MA), and probed with a cDNA fragment including the entire murine Id2 coding region (kindly provided by Dr. Edward Prochownik, Children's Hospital, Pittsburgh, PA). Membranes were stripped and reprobed with ribosome S16 cDNA. Autoradiograph data were digitized using a flatbed scanner, and relative hybridization levels were determined using NIH image version 1.6 software.
Reporter Plasmid Constructs, Transient Transfections, and CAT Assays-The Id2CAT reporter plasmids obtained from Dr. Larry Kedes (University of Southern California, Los Angeles, CA) contain regions of the human Id2 promoter extending upstream from position ϩ30 as indicated (23). BTATACAT contains the two tandem B enhancer elements from the human immunodeficiency virus (HIV) long terminal repeat linked to the albumen gene TATA element (24). The MSC-1 Sertoli cell line and primary Sertoli cell cultures in 35-mm 2 dishes were transfected using FuGENE reagent (Roche Applied Science) according to the manufacturer's instructions. Primary Sertoli cells and MSC-1 cells were transfected with 0.5 g of reporter plasmid in the presence or absence of 0.5 g of expression vector. The Ϫ834Id2CAT or BTATA-CAT reporter plasmids were cotransfected with empty expression vector or pRSVCAT-␤ that expresses the C ␤ catalytic subunit of protein kinase A (25). Transfected cells were exposed to vehicle, ovine FSH (100 ng/ml), or forskolin (10 M) plus IBMX (0.5 mM) for 24 h. Studies of the Ϫ834, Ϫ180, Ϫ152, Ϫ132, Ϫ122, and Ϫ82 Id2CAT promoter constructs were performed in the presence of forskolin plus IBMX or vehicle. The E-box luciferase reporter contains two consensus E-box motifs upstream of a minimal thymidine kinase promoter driving the luciferase gene. Assays for chloramphenicol acetyltransferase (CAT) and luciferase activity were performed 72 h after transfection. For CAT assays, cell lysates were incubated with fluorescent BODIPY chloramphenicol (Molecular Probes, Inc., Eugene, OR) according to the manufacturer's instructions. The resulting products were separated by thin layer chromatography, and the levels of acetylated products were quantified using a Fluorimager and ImageQuant software (Amersham Biosciences). Luciferase assays were performed as described (26). CAT and luciferase activities were normalized for total protein as determined by Bradford assay.
DNA-Protein Binding Studies-32 P-radiolabeled DNA probes were generated by annealing nucleotide templates containing Id2 promoter sequences to complementary 10-base primers. The overhangs were filled in with Klenow in the presence of [␣-32 P]dATP and 5 mM dCTP, dGTP, and dTTP. The probes used and their coding strand template and consensus CRE (5Ј-GATCCGGCTGACGTCATGAAGCTAGATC-3Ј). Mutated nucleotides are underlined. In addition, 13 Id2 Ϫ120 to Ϫ78 probes were synthesized having sequential 2-bp transversion mutations incorporated from positions Ϫ108 to Ϫ83. Electrophoretic mobility shift assays (EMSAs) were performed as described (27). Briefly, 32 P-labeled probes were incubated with 2-5 g of nuclear extracts prepared from untreated 16-day rat primary Sertoli cell cultures as described previously (20). In some cases, EMSA assays employed NF-B p50 or CREB proteins prepared by coupled in vitro transcription-translation (28). Binding reactions were incubated for 15 min at room temperature in the presence of 1 g of poly(dI-dC), 5 g of bovine serum albumin, 5 mM dithiothreitol, 50 -100 mM NaCl or KCl, 20 mM HEPES, and 1 mM EDTA. For competition EMSAs, a 10 -50-fold molar excess of unlabeled double-stranded competitor oligonucleotides including the Id2 Ϫ120 to Ϫ78 region, the consensus CRE motif, the E-box probe, the Ϫ144 to Ϫ119 region of the CREB promoter (5Ј-CTGGGAG-AAGCGGAGTGTTGGTGAGT-3Ј), or the Ϫ152 to Ϫ120 region of the Id2 promoter (5Ј-GGAAGAACCAAGCCCACGCCCCGCGCCCGCGCC-CACC-3Ј) were co-incubated with labeled Id2 Ϫ120 to Ϫ78 probe and nuclear extracts. DNA-protein complexes were resolved via 5% PAGE under non-denaturing conditions in a Tris borate/EDTA buffer. For comparing protein binding affinities of various probes, labeled probes with similar specific activities (Ϯ20%) were used in simultaneous reactions. DNA-protein cross-linking studies were performed with radiolabeled probes prepared as described above except that labeling reactions contained a 1:1 mixture of bromodeoxyuridine and dTTP, and binding reactions were scaled up 10-fold. The resolved DNA-protein complexes were exposed to UV light (302 nm) for 15 min in situ and then excised from PAGE gels. The gel pieces were inserted into the wells of SDS-PAGE gels and fractionated. DNA-protein complex formation was visualized by autoradiography and quantified using NIH Image 1.6 software analysis of digitized images.

Identification of Forskolin-stimulated mRNAs in Sertoli
Cells-To identify candidate genes that are induced by FSH and cAMP in the testis, primary rat Sertoli cells were stimulated with forskolin, an inducer of adenylate cyclase, in conjunction with the phosphodiesterase inhibitor IBMX. A limited differential display analysis of gene expression was performed on RNA collected after 12 h of stimulation. From three exper-iments, a total of 20 mRNAs were found to be induced, and 6 were sequenced. Two of the partial transcripts did not match sequences present in the GenBank TM , European Molecular Biology Laboratory (EMBL), or Swiss-Protein data bases. The unique cDNAs are the subject of another study. Four other transcripts were identified by DNA sequence analysis. These included 1) ATP synthase lipid-binding protein, 2) S-100b (an EF-hand calcium-binding protein), 3) non-muscle myosin alkali light chain, and 4) Id2. Of the identified forskolin-induced cDNAs, the HLH protein Id2 was chosen for further study because of the possibility that Id2 may be a regulator of FSHmediated gene transcription in Sertoli cells.
Forskolin or FSH Enhance Id2 Gene Expression-To confirm that Id2 gene expression was stimulated by forskolin, Northern blotting experiments were performed using mRNA isolated from primary rat Sertoli cells stimulated with forskolin and IBMX. In agreement with earlier studies demonstrating that Id2 probes do not cross-hybridize with other Id mRNAs (17,29), only a single Id2 transcript was detected. In primary Sertoli cells, Id2 levels rose 2 h after forskolin stimulation, peaked at 6 -12 h, and fell to near basal levels by 24 h (Fig. 1). To determine whether the major physiological inducer of cAMP in Sertoli cells, FSH, could also induce the Id2 gene, FSH and IBMX were added to primary Sertoli cells for 2, 6, 12, and 24 h. Although the peak of Id2 mRNA induction was delayed, the pattern of Id2 induction by FSH was similar to that seen for forskolin.
The Id2 Promoter Is Induced by Forskolin-Transient transfection studies were performed to determine whether the forskolin-mediated induction of Id2 mRNA was due to increased transcription initiation from the Id2 promoter. Primary Sertoli cells were transfected with the -834Id2CAT (30) reporter plasmid containing 834 bp of the Id2 promoter linked to the CAT reporter gene. The -834Id2 promoter was induced 3.3-and 2.8-fold by forskolin and FSH, respectively ( Fig. 2A). Cotransfection of a vector expressing the catalytic subunit of PKA also stimulated the Id2 promoter (4.4-fold). In studies employing the MSC-1 mouse Sertoli cell line that does not express FSH receptors (21), overexpression of the catalytic subunit of PKA also induced the -834Id2 promoter (19.0-fold) (Fig. 2B). Transcription from a control reporter plasmid (BTATACAT) having the CAT gene driven by a TATA box and two B enhancer motifs (24) was not altered by forskolin, FSH, or PKA. Together, these data suggest that the induction of Id2 mRNA by forskolin and FSH is due to stimulation of the Id2 promoter via PKA.
The Ϫ122 to Ϫ82 Region of the Id2 Promoter Contains a cAMP-inducible Element-To localize the regions of the Id2 promoter that are required for basal and forskolin-inducible activity, primary Sertoli cells and the MSC-1 Sertoli cell line were transfected with CAT reporter plasmids containing vari- FIG. 1. Id2 mRNA is induced by forskolin or FSH. Total RNA (20 g) isolated from primary Sertoli cells stimulated with forskolin or FSH and IBMX for 0, 2, 6, 12, or 24 h was analyzed by Northern blot using an Id2 cDNA probe. Blots were stripped and reprobed with a S16 ribosomal protein cDNA probe. A representative Northern blot is shown on the left, and the mean (ϮS.E.) of observations from three experiments is displayed on the right. were transfected with CAT reporter plasmids containing either the Id2 promoter (-834Id2CAT) or two consensus B enhancer motifs linked to a TATA box (BTATACAT). The cells were co-transfected with either empty expression vector or an expression vector encoding the catalytic subunit of PKA (C ␤ ). After 48 h, cells were stimulated with forskolin (10 m) plus IBMX (0.5 mM), ovine FSH (100 ng/ml), or vehicle. Cells were recovered 24 h later, and CAT activity was adjusted for protein concentrations and expressed relative to either -834Id2CAT plus vehicle or BTATA plus vehicle. The basal activity of BTATACAT was routinely 3-4-fold higher than -834Id2CAT (data not shown). Data shown represent the mean (ϮS.E.) of three independent experiments performed in duplicate.
ous Id2 promoter deletion mutants (30). Compared with the full-length -834Id2 promoter, removal of sequences upstream of Ϫ180 and Ϫ152 increased promoter basal activity (1.5-and 1.8-fold, respectively, for primary Sertoli cells (Fig. 3A) and 2.7and 4.4-fold, respectively, for MSC-1 cells (Fig. 3B)). Together, these findings suggest that multiple sequences between positions Ϫ152 and Ϫ82 contribute to basal Id2 promoter activity, but the Ϫ122 to Ϫ82 region is required for forskolin induction of the Id2 promoter. These data are consistent with a previous study in glioma cells that identified a serum-and forskolin-inducible region in the Id2 promoter downstream of position Ϫ177 (31).
The Ϫ120 to Ϫ78 Region of the Id2 Promoter Binds Proteins of 41-45 kDa-We next investigated potential DNA-binding proteins that might be responsible for forskolin induction of the Id2 promoter through the Ϫ122 to Ϫ82 region. EMSA analysis of Sertoli cell nuclear extracts identified a triplet of DNAprotein complexes that were formed with a probe extending from positions Ϫ120 to Ϫ78 (Fig. 4A, lane 1). Competition studies revealed that a 50-fold excess of unlabeled Ϫ120 to Ϫ78 probe or a probe containing a consensus CRE competed for protein binding to the labeled probe (Fig. 4A, lanes 1-3). In contrast, competing probes containing an E-box motif, a region of the CREB promoter, or the Ϫ152 to Ϫ122 region of the Id2 promoter did not inhibit formation of the DNA-protein complexes (Fig. 4A, lanes 4 -6). Additional competition studies using a range of 10 -50-fold excess of unlabeled Ϫ120 to Ϫ78 or consensus CRE probes confirmed that these two sequences compete for binding with the labeled Ϫ120 to Ϫ78 probe but that the CRE motif was a less effective competitor (Fig. 4B).
Previous studies of the Id2 promoter described two ATF/ CRE-like motifs within the Ϫ120 to Ϫ78 region that contribute to promoter activity in C2 myoblasts but not HeLa cells (30). EMSA studies were performed to compare Sertoli nuclear extract binding to a probe containing wild-type Ϫ120 to Ϫ78 sequences and a probe in which the ATF/CRE-like sequences were mutated. These studies demonstrated that the ATF/CRElike sites are required to bind Sertoli nuclear proteins (Fig. 5A). The competition observed using the consensus CRE probe and the dependence on an ATF/CRE-like motif for protein interactions raised the possibility that the binding complex contained a member of the CREB-ATF family of proteins. However, in-FIG. 3. The Id2 promoter contains a cAMP responsive element between ؊120 and ؊82. Primary Sertoli cells (A) and MSC-1 Sertoli cells (B) were transfected with CAT reporter plasmids containing either the Ϫ834 Id2 promoter or Id2 promoter 5Ј deletion mutants as indicated. Cells were stimulated 48 h after transfection with vehicle (stippled bars) or with forskolin (forsk, 10 m) plus IBMX (0.5 mm) (black bars) and recovered 24 h later. CAT activity was adjusted for protein concentrations and expressed relative to -834Id2CAT plus vehicle (ϩ S.E.). The mean -fold induction relative to vehicle treated -834Id2CAT is provided to the right of the bar graphs for each promoter construct. Data shown represent at least three independent experiments performed in duplicate.
FIG. 4. The ؊120 to ؊78 region specifically binds nuclear proteins in Sertoli cells. A, EMSA studies were performed using a 32 Plabeled oligonucleotide containing the Ϫ120 to Ϫ78 region of the Id2 promoter and nuclear extracts from primary Sertoli cells. Binding reactions were performed in the absence and presence of a 50-fold molar excess of competitor oligonucleotides containing the same sequences as the probe (Id2 Ϫ120 to Ϫ78), a CRE, an E-box, the Ϫ144 to Ϫ119 region of the CREB promoter, or the Ϫ152 to Ϫ120 region of the Id2 promoter as indicated. B, competition EMSA assays were performed with the Id2 Ϫ120 to Ϫ78 probe and Sertoli cell nuclear extracts in the absence or presence of a 10-, 25-, or 50-fold excess of either unlabeled Id2 Ϫ120 to Ϫ78 probe or a probe containing a CRE motif. Equal levels of protein (ranging from 2 to 5 g) were used for each assay. DNA-protein complexes were visualized by autoradiography. The unbound radiolabeled probe is not shown. The data displayed are representative of at least three independent experiments. cubation of the Ϫ120 to Ϫ78 probe with an excess of CREB protein produced by in vitro translation did not result in a DNA-protein complex, whereas the CREB protein did form a complex with a consensus CRE probe (Fig. 5B). As expected, in vitro translated NF-B p50 protein did not bind to the CRE or the Ϫ120 to Ϫ78 probe. These results suggest that CREB is not present in the complex binding to the Id2 ATF/CRE-like site.
To determine the relative molecular weight of the proteins binding to the Ϫ120 to Ϫ78 region, UV cross-linking analysis was performed using a photoreactive probe (Fig. 5C). Two DNA-protein complexes of 46 and 50 kDa were resolved. Formation of the DNA-protein adducts was increased after stimulation with forskolin for 24 h, suggesting that elevated cAMP levels may induce the expression of the proteins binding to the Ϫ120 to Ϫ78 region of the Id2 promoter or increase their affinity for the regulatory region. Because the probe contributes ϳ5 kDa to the complex (32), the proteins binding to the Ϫ120 to Ϫ78 region of the Id2 promoter are estimated to be 41 and 45 kDa.
To localize the binding sites for Sertoli cell nuclear proteins binding to the Ϫ120 to Ϫ78 region of the Id2 promoter, a series of probes were synthesized having 2 bp mutations in the ATF/ CRE-like and flanking regions. This strategy determined that the fastest migrating DNA-protein complex required the sequence 5Ј-CTCGTC-3Ј (Ϫ106 to Ϫ101) for binding (Fig. 5D), which corresponds to the upstream ATF/CRE-like sequence identified previously (23). The two slower migrating complexes were formed due to proteins binding to the sequence 5Ј-GATA-GACGTG-3Ј (Ϫ98 to Ϫ89), which contains a consensus sequence for GATA protein binding and the overlapping downstream ATF-CRE-like sequence. Mutation of positions Ϫ98 and Ϫ97 as well as Ϫ94 and Ϫ93, within and flanking the GATA motif, abolished all DNA-protein interactions. Unfortunately, mutation of the TA to CC at positions Ϫ96 to Ϫ95 resulted in the formation of new DNA-protein complexes. A second mutation of this site from TA to AT also resulted in novel DNAprotein binding patterns (data not shown). Therefore, information regarding protein binding dependent on positions Ϫ96 to Ϫ95 is lacking. Nevertheless, our results identify two to three binding sites for Sertoli cell nuclear proteins within the Ϫ106 to Ϫ89 region that account for all the DNA-protein complexes that are formed.
FSH Stimulation Transiently Represses E-box Binding Activity-The stimulation of Id2 expression by inducers of cAMP would be expected to enhance Id2 dimerization with E-boxbinding proteins and inhibit E-box protein interactions with target DNA sequences. To test this idea, EMSA studies were performed with a probe containing a consensus E-box motif and nuclear extracts from Sertoli cells treated with FSH. E-box binding activity was high in extracts from untreated and 6-h stimulated Sertoli cells but decreased 85% after 12 h and then recovered to 45% of basal levels after 24 h of FSH stimulation (Fig. 6, A and C). In contrast, CREB binding to a consensus CRE motif increased after 12-24 h of FSH stimulation, perhaps reflecting elevated CREB protein levels as a result of previously described FSH induction of the CREB gene (33) (Fig. 6, B and C). Together, these data suggest that the transient induction of Id2 detected 6 -12 h after FSH stimulation corresponds with the down-regulation of E-box binding activity in Sertoli cells.
FSH regulation of E-box-mediated transcription in primary Sertoli cells was then investigated in transient transfection assays using a luciferase reporter gene driven by two E-box motifs upstream of a minimal thymidine kinase promoter. These studies revealed that luciferase activity increased 81% within 6 h of FSH stimulation but then decreased to 38% above basal levels by 12 h (Fig. 6D). Luciferase activity was then allowed to increase more significantly to 347% over basal levels by 24 h. This pattern of expression suggests that FSH can initially stimulate transcription of the reporter but that stimulation is transiently interrupted during the period when Id2 levels are elevated and heterodimerization of Id2 and E-box proteins is favored. The later induction of E-box-mediated transcription coincides with decreased Id2 expression and increased E-box binding activity. DISCUSSION Initial differential display analyses suggested that the cAMP-elevating agent forskolin increased Id2 mRNA levels within 12 h of stimulating primary Sertoli cell cultures. Subsequently, we confirmed that forskolin-or FSH-induced Id2 mRNA expression peaks after 6 -12 h and falls to near basal levels after 24 h. Stimulation of the Id2 gene is apparently independent of the effects of serum, a known inducer of Id2 (34, FIG. 5. The ؊120 to ؊78 region specifically binds 41-and 45-kDa proteins in Sertoli cells. A, EMSA analysis was performed with Sertoli nuclear extracts incubated with either the Id2 Ϫ120 to Ϫ78 probe (wt) or a similar probe having the ATF sites mutated (mt). B, Sertoli nuclear extracts as well as NF-B p50 and CREB produced by in vitro translation were incubated with the Id2 Ϫ120 to Ϫ78 or consensus CRE probes, and EMSA analysis was performed as in A. C, a photoreactive Id2 Ϫ120 to Ϫ78 oligonucleotide probe was incubated with nuclear extracts from untreated or forskolin stimulated Sertoli cells, and the mixtures were exposed to UV light. The cross-linked DNA-protein adducts resolved by SDS-PAGE are shown. The relative mobility of molecular size markers are indicated to the left. D, the wild-type Id2 Ϫ120 to Ϫ78 probe (lane 1) and mutant probes containing 2 bp changes as indicated (lanes [2][3][4][5][6][7][8][9][10][11][12][13][14] were incubated with Sertoli nuclear extracts, and EMSA analysis was performed. Lane 8 was inserted from a separate, identical experiment. The unbound probe was run off of the gel for all studies. The sequence of the Ϫ108 to Ϫ83 region is displayed above, and the 2 bp mutations corresponding to each lane are shown immediately below the wild-type sequence. DNA-protein complexes were visualized by autoradiography. Data shown are representative of at least three independent experiments. 35), because Id2 is induced in primary Sertoli cells cultured in serum-free medium as well as in MSC-1 cells maintained in 10% serum. The transient Id2 mRNA induction explains the results of a previous study in which no Id2 stimulation was observed 72 h after treatment with FSH or a cAMP analogue (17). It is likely that the induction of Id2 mRNA is terminated due to a negative feedback mechanism mediated by Id2 interfering with the binding of bHLH proteins to E-box motifs within the Id2 promoter. This negative feedback hypothesis is supported by earlier studies showing that the Id2 promoter is inducible through E-box motifs after the overexpression of bHLH proteins but that the co-expression of Id2 blocks the stimulation of the Id2 promoter (31,35,36).
Examination of Id2 promoter deletion mutants in MSC-1 and primary Sertoli cells revealed differences in the regions responsible for maintaining basal promoter activity. Although the Ϫ122 to Ϫ82 region of the promoter was required for basal activity in primary Sertoli cells, the removal of the Ϫ152 to Ϫ132 region resulted in a marked decline in basal Id2 promoter activity in MSC-1 cells. This difference in the control of basal activity is reminiscent of earlier studies by Kedes' group (23) that found the Ϫ152 to Ϫ132 region was required for Id2 promoter activity in C2 myoblasts but was less important in HeLa cells. It is possible that the variable contributions to promoter activity exhibited by the Ϫ152 to Ϫ132 region in primary Sertoli cells and the MSC-1 Sertoli cell line may be reflective of the differentiation state or proliferation capabilities of the cells (primary Sertoli cells isolated from 16-day-old rats are nearly fully differentiated and at the end of their proliferation period, whereas MSC-1 cells proliferate continuously).
In contrast to the differential basal regulation of the Id2 promoter in primary Sertoli cells versus the MSC-1 cell line, the induction of mRNA accumulation and promoter activity is similar for both types of Sertoli cells. Transient transfection studies of primary Sertoli cells demonstrated that the 834-bp Id2 promoter can be induced by forskolin, FSH, or the catalytic subunit of PKA, suggesting that FSH-induced increases in Id2 mRNA are due to increased transcription efficiency that is mediated via a cAMP/PKA-mediated pathway. In MSC-1 and primary Sertoli cells, promoter deletion analyses revealed that forskolin-inducible Id2 promoter activity required sequences located between Ϫ122 and Ϫ82. Previously, two binding motifs were identified between Ϫ110 and Ϫ90 of the Id2 promoter that are potential binding sites for the ATF and CREB transcription factors. These motifs were shown to confer doxorubicin-induced expression from the promoter and specifically bind nuclear proteins; however, neither of the major cAMP-responsive proteins, ATF-1 and CREB, was found to bind to the region (30).
In our studies of Sertoli cell nuclear protein binding to the Ϫ120 to Ϫ78 region, we detected a triplet of sequence-specific DNA-protein complexes. The competition for protein binding by a consensus CRE probe and the lack of DNA-protein complexes formed using a probe having mutated ATF/CRE-like motifs raised the possibility that CREB or ATF proteins might bind to the Ϫ120 to Ϫ78 region. However, binding studies employing recombinant CREB protein suggest that CREB is not the major protein binding to this region of the Id2 promoter and is not responsible for induction via the PKA pathway. Consistent with this idea is our inability to supershift the DNA-protein complexes with CREB antisera (data not shown). Furthermore, it is not likely that an ATF protein binds to the ATF/CRE-like sequences of the Id2 promoter sequence because of the known ATF proteins, ATF-1 is the most similar in size (35 kDa) to the proteins that could be cross-linked to the probe, but ATF-1 is at least 5 kDa smaller than the protein(s) that we identified.
The Ϫ106 to Ϫ89 region was found to be required for the formation of all three DNA-protein complexes observed with the Id2 Ϫ120 to Ϫ78 probe. The fastest migrating complex was disrupted by mutation of any 2 bp in the Ϫ106 to Ϫ101 region corresponding to the core of the upstream ATF/CRE-like sequence. With one exception, probes having mutations within the Ϫ98 to Ϫ89 region eliminated the formation of the two slower migrating complexes. We noted that the 5Ј end of the Ϫ98 to Ϫ89 region contains a consensus GATA-protein binding site. Of the six well characterized GATA proteins, the major forms expressed in Sertoli cells are GATA-1 and GATA-4 (37). Supershift EMSA assays and immunoprecipitation of UV cross-linked DNA-protein complexes using Sertoli cell nuclear extracts did not identify GATA-1 or GATA-4 bound to the Id2 Ϫ120 to Ϫ78 probe (data not shown). Although information could not be obtained for positions Ϫ96 to Ϫ95, flanking mutations of the Ϫ98 to Ϫ93 region resulted in the inhibition of all DNA-protein complex formation. One possible explanation for this result would be that a protein binding to the GATA motif is required to allow binding of proteins to the adjacent sequences. With the exception of the GATA binding site, a computer-assisted search of the region for binding motifs using TFSEARCH software (38) did not identify any highly probable binding sites for vertebrate transcription factors. Therefore, FIG. 6. FSH stimulation transiently represses E-box binding activity. EMSA studies were performed with nuclear extracts (4 g) from Sertoli cells stimulated with FSH for 0, 6, 12, or 24 h and probes containing a consensus E-box motif (A) or a consensus CRE (B). The migration of E-box-protein and CRE-protein complexes are indicated. DNA-protein complexes were visualized by autoradiography. The unbound radiolabeled probe is not shown. C, the mean binding activities (Ϯ S.E.) relative to control are shown for E-box (stippled rectangles) and consensus CRE probes (filled rectangles). The data displayed are representative of three (E-box probe) or four (CRE probe) independent experiments. D, Sertoli cells transfected with an E-box luciferase reporter plasmid (E-box luciferase) were stimulated with FSH for 0, 2, 6, 12, or 24 h. Luciferase activities (Ϯ S.E.) are expressed relative to untreated cells (0 h FSH ϭ 100%) for three experiments performed in duplicate. the factors responsible for FSH, forskolin, and PKA induction of the Id2 promoter remain to be determined.
Our transfection studies demonstrated that after an initial FSH-mediated induction of E-box-regulated transcription, further transcription was repressed until 12-24 h after FSH administration. This pattern of FSH-regulated transcription corresponded to the increased Id2 expression and decreased E-box binding activity that we observed in response to FSH. In previous studies of Sertoli cells, overexpression of Id2 was found to repress E-box-mediated transcription from the promoters of the FSH receptor (2,8). Together, these observations suggest that FSH-induced sequestration of E-box factors by Id2 contributes to the decreases in endogenous FSH receptor RNA observed within 6 -8 h of FSH stimulation prior to the later restoration of FSH receptor mRNA to normal levels that has been reported previously (39,40). In vivo, the levels of FSH receptor on the Sertoli cell surface rise and fall in a cyclical manner, causing related changes in cAMP levels (13,41). It would be expected that Id2 expression would respond to the cyclical changes in cAMP concentration. The induction of Id2 by FSH and subsequent transient negative feedback of Id2 in cultured Sertoli cells may contribute to the cyclical expression of FSH receptors on Sertoli cells in the testis. Together, the temporal induction of Id2 by FSH and Id2 down-regulation of E-box-mediated transcription implicate Id2 as an important transducer of hormonal signals that are required to support sperm development.