Identification and Characterization of the Gonadotropin-releasing Hormone Response Elements in the Mouse Gonadotropin-releasing Hormone Receptor Gene*

The response of the pituitary gonadotrope to gonadotropin-releasing hormone (GnRH) correlates directly with the concentration of GnRH receptors (GnRHR) on the cell surface, which is mediated in part at the level of GnRHR gene expression. Several hormones have been implicated in this regulation, most notably GnRH itself. Despite these observations and the central role that GnRH is known to play in reproductive development and function, the molecular mechanism(s) by which GnRH regulates transcription of the GnRHR gene has not been well elucidated. Previous studies in this laboratory have identified and partially characterized the promoter region of the mouse GnRHR gene and demonstrated that the regulatory elements for tissue-specific expression as well as for GnRH regulation are present within the 1.2-kilobase 5′-flanking sequence. By using deletion and mutational analysis as well as functional transfection studies in the murine gonadotrope-derived αT3-1 cell line, we have localized GnRH responsiveness of the mouse GnRHR gene to two DNA sequences at −276/−269 (designated Sequence UnderlyingResponsiveness to GnRH-2 (SURG-2), which contains the consensus sequence for the activating protein-1-binding site) and −292/−285 (a novel element designated SURG-1), and demonstrated that this response is mediated via protein kinase C. By using the electrophoretic mobility shift assay, we further demonstrate that a member(s) of the Fos/Jun heterodimer superfamily is responsible in part for the DNA-protein complexes formed on SURG-2, using αT3-1 nuclear extracts. These data define a bipartite GnRH response element in the mouse GnRHR 5′-flanking sequence and suggest that the activating protein-1 complex plays a central role in conferring GnRH responsiveness to the murine GnRHR gene.

reproductive development and function. Pituitary gonadotropes, which make up 8 -15% of all cells in the anterior pituitary gland (1), express cell surface, G protein-coupled receptors specific for GnRH (2,3). Activation of this receptor by GnRH stimulates intracellular signal transduction pathways to increase the synthesis and release of the pituitary gonadotropins, luteinizing hormone (lutropin; LH) and follicle-stimulating hormone (follitropin; FSH) (4,5). These hormones then enter the systemic circulation to regulate gonadal function, including steroid hormone synthesis and gametogenesis.
The biosynthesis and secretion of LH and FSH by pituitary gonadotropes are tightly regulated as evidenced by predictable and reproducible changes in circulating levels during the menstrual cycle. This regulation is dependent primarily on GnRH pulse amplitude and frequency, which varies with physiological state, with the rat estrous and human menstrual cycle, and with puberty, and the menopause. The response of pituitary gonadotropes to GnRH correlates directly with the concentration of GnRH receptors (GnRHR) on the cell surface, which are, in turn, regulated by a number of hormonal factors, most notably GnRH itself (6 -9). The highest concentration of GnRHR in the pituitary gland is associated with a GnRH pulse frequency of 30 min and results in the optimum synthesis and release of LH. Lower concentrations of GnRHR are seen with GnRH pulse frequencies of 2 h and correlate with optimum synthesis and release of FSH (8 -10). Continuous exposure to high concentrations of GnRH results in down-regulation of GnRHR mRNA (11). The difference in the concentration of GnRHR between high and low frequency GnRH pulses is 2-3fold (8,12). The difference in the concentration of GnRHR appears to be mediated at least in part at the level of GnRHR gene expression (6). GnRH regulation of GnRHR mRNA is well documented in rat pituitary cells (11). However, the cellular mechanism(s) by which GnRH regulates transcription of this and other genes has not been intensively investigated. This study was designed to identify and characterize the critical cis-DNA element(s) and cognate trans-factors that mediate the regulation of mouse GnRHR (mGnRHR) gene expression by GnRH.
Previous studies in this laboratory have identified and partially characterized the promoter region of the mGnRHR gene and demonstrated that the regulatory elements for tissue-specific expression as well as for GnRH regulation are present within a 1.2-kilobase (kb) 5Ј-flanking region of the mGnRHR gene (designated Ϫ1164/ϩ62 relative to the major transcriptional start site (TSS)) (13). By using deletion and mutational analysis as well as functional transfection studies in the murine gonadotrope-derived ␣T3-1 cell line, we have localized GnRH responsiveness of the mGnRHR gene to two distinct DNA elements that appear to be both necessary and sufficient to mediate a full GnRH response. The first and critical element (5Ј-TATGAGTC-3Ј), designated the Sequence Underlying Responsiveness to GnRH-2 (SURG-2), lies at position Ϫ276/Ϫ269 and contains the consensus sequence for the canonical 12-Otetradecanoylphorbol-13-acetate response element, also known as the activating protein-1 (AP-1)-binding site. The second element (5Ј-GCTAATTG-3Ј), designated SURG-1, lies at position Ϫ292/Ϫ285 and appears to be a novel enhancer element. The importance of these two elements in mediating GnRH responsiveness of the mGnRHR gene was confirmed by demonstrating that both SURG-1 and SURG-2 are capable independently of conferring activity on a heterologous minimal promoter but that both elements are required for a full response. Our data further suggest that the response of the mGnRHR gene promoter to GnRH is mediated via the protein kinase C (PKC), and not protein kinase A (PKA), signal transduction pathway. By using the electrophoretic mobility shift assay, we further demonstrate that a member(s) of the Fos/Jun heterodimer superfamily is responsible in part for the DNAprotein complexes formed on SURG-2, using ␣T3-1 nuclear extracts, and that such proteins are rapidly induced by GnRH stimulation. We propose therefore that GnRH-stimulated activity of the mGnRHR gene is regulated by two distinct elements within the GnRHR gene promoter and that the key component of this mechanism involves the AP-1 protein complex that activates transcription in a cell-specific fashion.
Reporter Plasmids and Expression Vectors-A fusion construct was prepared by ligation of the 1.2-kb 5Ј-flanking region of the mGnRHR gene (designated Ϫ1164/ϩ62) into the luciferase reporter plasmid, pXP2, as described previously (13). The nucleotide sequence of the mGnRHR gene promoter used in these studies is based on previous work in this laboratory (13), with Ϫ1 assigned to the nucleotide immediately 5Ј of the major TSS. 5Ј-Deletions of the 1.2-kb GnRHR gene promoter were synthesized by polymerase chain reaction (PCR) using selected sense/antisense primers with the full-length construct as a template and incorporating HindIII/XhoI restriction enzyme sites at the ends. All primers included sufficient 5Ј-and 3Ј-flanking sequence (ϳ18 base pairs (bp)) to ensure specific annealing. The resultant PCR products encoding the desired 5Ј-deletion constructs (Ϫ765/ϩ62, Ϫ387/ ϩ62, Ϫ365/ϩ62, Ϫ341/ϩ62, Ϫ300/ϩ62, Ϫ232/ϩ62, Ϫ117/ϩ62, and Ϫ38/ ϩ62) were then each digested with HindIII and XhoI restriction enzymes and inserted into the HindIII/XhoI polylinker restriction sites upstream of the luciferase reporter in pXP2. An expression vector expressing ␤-galactosidase driven by the Rous sarcoma virus promoter (RSV-␤-galactosidase) was used as an internal standard and control.
The GH 50 -pXP2 construct was prepared by subcloning the rat growth hormone gene minimal promoter (GH (Ϫ50/ϩ1) , designated GH 50 ) into the BglII restriction site in pXP2. Dideoxy sequencing was used to confirm the orientation and sequence of the GH 50 insert. PCR-generated fragments of the region Ϫ387/Ϫ220 of the mGnRHR gene promoter were synthesized using selected sense/antisense primers with the full-length construct as a template and incorporating HindIII/XhoI restriction enzyme sites at the ends. The constructs were designated GH 50 /Ϫ387/ Ϫ220 (wild type), GH 50 /Ϫ308/Ϫ220, GH 50 /Ϫ267/Ϫ220, GH 50 /Ϫ387/ Ϫ264, GH 50 /Ϫ341/Ϫ264, GH 50 /Ϫ308/Ϫ264, and GH 50 /Ϫ387/Ϫ308. A series of scanner-linker mutations of the region Ϫ308/Ϫ264 of the mGnRHR gene incorporating HindIII/XhoI restriction enzyme sites at the ends were generated by synthesizing sense and antisense oligonucleotides with an overlap of ϳ15 bp, self-annealing the oligonucleotides, and reconstituting the DNA double strands using Sequenase 2.0 (United States Biochemical Corp.). By serially replacing 8-bp segments of the wild type sequence with the NotI restriction enzyme site starting at the 5Ј end, five separate mutants of the region Ϫ308/Ϫ264 of the mGnRHR gene promoter were created and designated regions A-E (Fig. 5A). The NotI restriction enzyme site (5Ј-GCGG2CCGC-3Ј) is novel to the mGn-RHR gene, and previous transfection experiments have shown no effect of this sequence on GnRH stimulation (14). Digestion with the restriction enzyme, NotI, and subsequent gel electrophoresis confirmed the presence of the NotI-containing insert. A point mutation of the AP-1binding site at position Ϫ269 (C269T) in the wild type Ϫ308/Ϫ264 construct was similarly generated. These constructs were then digested with HindIII and XhoI restriction enzymes and subcloned into HindIII/ XhoI polylinker restriction sites upstream of the rat growth hormone gene minimal promoter in GH 50 -pXP2 to generate wild type expression vector (designated GH 50 /Ϫ308/Ϫ264), a single point mutation (designated GH 50 /region E/mut), and five NotI mutants (designated GH 50 / region A, GH 50 /region B, GH 50 /region C, GH 50 /region D, and GH 50 / region E, according to the location of the NotI restriction enzyme site (Fig. 5A)).
Cell Culture and Transient Transfection-␣T3-1 (mouse gonadotrope) and CV-1 (African green monkey kidney fibroblast) cells were maintained in monolayer culture in high and low glucose Dulbecco's modified Eagle's medium (DMEM, Life Technologies, Inc.), respectively, and supplemented with 10% (v/v) fetal bovine serum (Life Technologies, Inc.), 100 units/ml penicillin, and 100 g/ml streptomycin sulfate (Life Technologies, Inc.) at 37°C in humidified 5% CO 2 /95% air. For transient transfection studies, cells were divided into six-well tissue culture plates and cultured overnight in DMEM in the absence of serum or antibiotics. Under these conditions, cells were 40 -50% confluent. Cells were then transfected by calcium phosphate co-precipitation, as described previously (15). Briefly, cells were incubated with the calcium phosphate-DNA precipitates for 4 h in media containing 10% (v/v) fetal bovine serum. In each experiment, test vector was standardized at 4 g of DNA per well. An expression vector expressing RSV-␤-galactosidase (1 g/well) was co-transfected in all experiments and used as an internal standard. Following a 4-h transfection, cells were washed once at room temperature with phosphate-buffered saline (pH 7.4). Thereafter, cells were treated with 100 nM GnRHAg or vehicle in serum-containing DMEM (2 ml/well) for 4 h immediately prior to harvest. These conditions were selected after optimization analysis to give maximal levels of expression and GnRHAg stimulation. Following the final incubation, the medium was aspirated, and cells were washed once with ice-cold phosphate-buffered saline. Cells were lysed in the wells by addition of 200 l of lysis buffer (125 mM Tris (pH 7.6), 0.5% (v/v) Triton X-100). Cellular debris was removed from lysate by microcentrifugation at 14,000 ϫ g for 10 min at 4°C. Supernatants were assayed immediately for luciferase and ␤-galactosidase activity by standard protocols. Briefly, luciferase activity was determined by adding 120 l of cell lysate to 200 l of luciferin substrate (Promega) and measuring luminescence with a LB-953 Autolumat (Berthold, Nashua, NH) luminometer set for a 30-s integration with no delay. ␤-Galactosidase activity was determined by adding 50 l of cell lysate to 297 l of substrate (0.1 M Na 2 HPO 4 buffer (pH 7.3), 0.013 M 2-nitrophenyl-␤-D-galactopyranoside, 0.1% (v/v) 1.0 M MgCl 2 , 0.35% (v/v) ␤-mercaptoethanol), incubating overnight at 37°C, and measuring colorimetrically at 410 nm in a Beckman DU640 spectrophotometer (Beckman, Fullerton, CA) after the addition of 100 l of 1.0 M sodium carbonate. Luciferase activity was normalized to expression of RSV-␤-galactosidase.
Northern Blot Analysis-To investigate the effect of GnRHAg stimulation on GnRHR mRNA, ␣T3-1 cells were treated with 100 nM Gn-RHAg or vehicle for varying time intervals (1, 2, 4, or 8 h), and total RNA was extracted from cells using the Qiagen "RNEasy" RNA extraction kit (Qiagen, Santa Clarita, CA). Total RNA (10 g/lane) was separated by electrophoresis on a denaturing 1.2% agarose gel containing 6.7% formaldehyde prior to capillary transfer onto sheets of Nytran (Schleicher & Schuell) immobilization membrane. Northern blot analysis was performed under high stringency conditions using a [ 32 P]UTPlabeled antisense riboprobe (5 ϫ 10 5 cpm/ml) prepared from the coding region (ϩ173/ϩ1153) of mGnRHR cDNA using T7 RNA polymerase (New England Biolabs, Inc., Beverly, MA). Washed blots were exposed to Kodak X-Omat/AR film at Ϫ70°C for 8 -38 h. Rat cyclophilin antisense riboprobe was used as an internal standard. The intensity of the individual RNA bands was quantified in a PhosphorImager (Molecular Dynamics, Sunnyvale, CA) according to the protocol outlined by the manufacturers. Measurements were standardized for cyclophilin mRNA.
Preparation of Nuclear Extracts-␣T3-1 cells were grown to 20, 40, 60, and 80% confluence and treated with 100 nM GnRHAg or vehicle for varying time intervals (1 or 4 h). Thereafter, cells were harvested, and nuclear extracts were prepared by the method of Andrews and Faller (16).
Electrophoretic Mobility Shift Assay (EMSA)-Probe was prepared for EMSA by digestion of the pXP2 plasmid containing the Ϫ308/Ϫ220 fragment of the mGnRHR gene promoter with HindIII and XhoI, followed by 5Ј-end-labeling with [␥-32 P]ATP by T4 polynucleotide kinase (New England Biolabs). Constructs were then purified using a Qiagen nucleotide removal kit. The binding reaction for EMSA was performed by incubating 50,000 cpm of DNA probe with 10 g of nuclear extract and 1 g of salmon sperm DNA in reaction buffer (20 mM HEPES (pH 7.9), 60 mM KCl, 5 mM MgCl 2 , 10 mM phenylmethylsulfonyl fluoride, 10 mM dithiothreitol, 1 mg/ml bovine serum albumin, and 5% (v/v) glycerol) for 30 min at 4°C. For competition studies, excess unlabeled DNA was added 5 min prior to the addition of probe. Protein-DNA complexes were resolved on 4% low ionic strength non-denaturing polyacrylamide gel electrophoresis in 0.5ϫ Tris borate/EDTA buffer (45 mM Tris-HCl (pH 8.0), 45 mM boric acid, 1 mM EDTA). Gels were then dried for 1 h and subjected to autoradiography for 24 -48 h. Antibody supershift experiments were performed using an anti-Fos antibody (Santa Cruz Biotechnology) which recognizes all members of the Fos oncoprotein family. Similar experiments were carried out using an anti-Jun blocking antibody (Santa Cruz Biotechnology) raised against the common DNA binding domain of all members of the Jun family. Antibody, either anti-Fos (1 l), anti-Jun (1, 2, 4, or 8 l), or both, was added to the EMSA reaction samples after 30 min and incubated at 4°C for an additional 2 h prior to gel electrophoresis. When indicated, the intensity of the individual protein bands was quantified in a PhosphorImager (Molecular Dynamics). Measurements were standardized to background intensity.
Characterization of the Signal Transduction Pathway Involved in the GnRHAg Stimulation of the mGnRHR Gene-To identify the second messenger pathway(s) involved in the GnRHAg stimulation of the mGnRHR gene, ␣T3-1 cells were transfected with GH 50 -pXP2, GH 50 / Ϫ308/Ϫ264 (wild type), GH 50 /region C (SURG-1 mutant), or GH 50 / region E (SURG-2 mutant) as described, and response to GnRHAg stimulation was measured in the presence or absence of selective agonists/antagonists. To investigate the role of the PKC signal transduction pathway, transfected cells were stimulated for 4 h with PMA (100 ng/ml), GnRHAg (100 nM), or both. Final concentrations of GnRHAg ( Fig. 2A) and PMA (data not shown) were chosen to give maximal stimulation at 4 h. Similar experiments were carried out in the presence or absence of 1 M GF-109203X (Sigma), a simplified derivative of staurosporine that acts as a competitive inhibitor for the ATP-binding site of PKC. This agent is selective for PKC isoforms ␣, ␤ 1 , ␤ 2 , ␥, ␦, and ⑀ as compared with PKA, phosphorylase kinase, various tyrosine kinases, and PKC isoform . In the latter experiments, cells were incubated with antagonist or vehicle for 30 min immediately prior to as well as during the 4-h period of stimulation. Reagents were initially dissolved in dimethyl sulfoxide (Me 2 SO) and subsequently diluted in culture medium to give a final concentration of 0.01% (v/v) Me 2 SO in each experiment. To investigate the role of the PKA signal transduction pathway, transfected cells were stimulated for 4 h with forskolin (25 M), GnRHAg (100 nM), or both. Identical studies were carried out using a second PKA agonist, 8-bromo-cAMP (1 mM). The effect of 1 M SQ 22536 (Sigma), a selective inhibitor of PKA activity, was investigated in selected experiments.
Statistical Analysis-Transfections were performed in triplicate and repeated multiple times. Data in each experiment were normalized to the basal level of activity of either pXP2 or GH 50 -pXP2 (designated 1-fold). Data were then combined across experiments. Results were expressed as mean Ϯ S.E. for basal and GnRHAg-stimulated activities for each construct, and fold stimulation in response to GnRHAg was calculated. One-way analysis of variance (ANOVA) followed by post hoc comparisons with Fisher's protected least significant difference test was used to assess whether changes in GnRH responsiveness among different GnRHR promoter-luciferase reporter constructs were significant. Significant differences were designated as p Ͻ 0.05. When appropriate, data were analyzed by the Student's t test for independent samples.

RESULTS
Northern Blot Analysis-A selected Northern blot of RNA extracted from ␣T3-1 cells stimulated with GnRHAg (100 nM) or vehicle for varying time intervals (1, 2, 4, or 8 h) and hybridized with mGnRHR antisense riboprobe is shown in Fig.  1A. The size of the major mGnRHR mRNA was 4.5 kb as previously reported (17) and that of cyclophilin mRNA was 0.8 kb. The intensities of individual bands were quantified in a PhosphorImager and corrected for total RNA content by using cyclophilin mRNA levels. Results (expressed as percent of time 0) demonstrate a significant increase in GnRHR mRNA in response to GnRHAg stimulation that was maximal at 4 h (1.84 Ϯ 0.2-fold stimulation; p Ͻ 0.01) but decreased thereafter (Fig. 1B). These data are consistent with previous reports indicating that GnRH is capable of regulating GnRHR gene expression, both in the short term (up-regulation) and in the long term (down-regulation) (6,11). The following study was then designed to identify and characterize the critical cis-DNA element(s) and cognate trans-factors responsible for this regulation in ␣T3-1 cells.
Optimization of the GnRHAg Response-To optimize conditions for GnRHAg responsiveness, ␣T3-1 cells were transiently transfected with the full-length 1.2-kb 5Ј-flanking region of the mGnRHR gene (Ϫ1164/ϩ62) in pXP2 luciferase vector for 4 h and then subjected to GnRHAg stimulation for varying times (2, 4, or 20 h) and with varying concentrations of GnRHAg (0, 10, 100, or 500 nM). On review of the GnRHAg dose-( Fig. 2A) and time-response curves (Fig. 2B) for ␣T3-1 cells, GnRHAg stimulation was optimum at a concentration of 100 nM for 4 h. A longer transfection time resulted in increased basal expres-sion of luciferase activity and a significantly diminished response to GnRHAg ( Fig. 2A). Optimum response to GnRHAg stimulation was also seen in ␣T3-1 cells grown to 40 -50% confluence and cultured overnight in high glucose DMEM in the absence of serum or antibiotics (data not shown). Cells grown to Ն80% confluence or cells cultured overnight in medium containing serum and antibiotics showed a marked reduction in GnRHAg responsiveness (data not shown). CV-1 cells transiently transfected with Ϫ1164/ϩ62 also demonstrated a modest increase in basal luciferase activity with time but showed no response to GnRHAg (Fig. 2C), suggesting that the GnRH response is cell-specific.
Identification of Two GnRH Response Elements (GnRH-REs) in the mGnRHR Gene Promoter-Transfection of ␣T3-1 cells with Ϫ1164/ϩ62 resulted in a 9.1 Ϯ 1.3-fold increase in luciferase activity in response to GnRHAg stimulation as compared with vector alone (p Ͻ 0.0001; ANOVA) (Fig. 3). Transfection with serial 5Ј-deletion mutants of the full-length construct demonstrated a significant decrease in GnRHAg-stimulated luciferase activity between the constructs Ϫ765/ϩ62, Ϫ387/ ϩ62, Ϫ365/ϩ62, Ϫ341/ϩ62, Ϫ300/ϩ62 (in which fold stimulation was not different from each other nor from Ϫ1162/ϩ62 but were all significantly different from pXP2 vector alone (p Ͻ 0.01; ANOVA)) and the remaining constructs Ϫ232/ϩ62, Ϫ117/ ϩ62, and Ϫ38/ϩ62 (in which fold stimulation was not significantly different from each other nor from pXP2 vector alone) (Fig. 3). Despite a 2.4 Ϯ 0.4-fold and 2.3 Ϯ 0.4-fold stimulation in the Ϫ232/ϩ62 and Ϫ117/ϩ62 constructs, respectively, these measurements were not significantly different from pXP2 alone (p ϭ 0.56 and p ϭ 0.57, respectively; ANOVA; n ϭ 13 separate experiments). Basal luciferase activity was not significantly different between the various constructs (data not shown). These data suggest the presence of an element(s) within the region Ϫ300/Ϫ232 of the mGnRHR gene promoter which is necessary for GnRH responsiveness.
To determine whether the putative element(s) in this region is not only necessary but also sufficient to mediate a GnRHAg response, PCR-generated fragments of the region Ϫ387/Ϫ220 were placed in control of a heterologous promoter, and response to GnRHAg stimulation was measured. A number of heterologous minimal promoters were tested, including the human ␣-subunit gene promoters (␣ (Ϫ99/ϩ1) or ␣ (Ϫ77/ϩ1) ), the rat growth hormone promoter (GH 50 ), and the herpes simplex virus thymidine kinase promoters (PT (Ϫ109/ϩ1) or PT (Ϫ81/ϩ1) ). The GH 50 minimal promoter was found to be most appropriate for our system (data not shown) and was therefore used in all subsequent experiments. GH 50 -pXP2 alone did not increase luciferase activity in response to GnRHAg stimulation, but incorporation of selected mGnRHR constructs upstream of GH 50 allows for a 10 -12-fold response to GnRHAg (Fig. 4).
To identify and characterize further the GnRH-RE(s) within the region Ϫ308/Ϫ264 of the mGnRHR gene promoter, five scanner-linker mutants of this region were synthesized as detailed above (Fig. 5A), placed upstream of the GH 50 minimal promoter in GH 50 -pXP2, and transfected into ␣T3-1 cells. Wild type expression vector (GH 50 /Ϫ308/Ϫ264) resulted in a 7.8 Ϯ 1.2-fold increase in luciferase activity in response to GnRHAg stimulation, which was similar to that measured in the scanner-linker mutant constructs for regions A, B, and D (Fig. 5B). However, transfection with the mutant construct of region E (GH 50 /region E, SURG-2mut), in which the putative AP-1-binding site had been mutated, completely abrogated the GnRHAg response (p Ͻ 0.0001 compared with wild type, but not significant compared with GH 50 -pXP2; ANOVA). Similarly, a single point mutation of the AP-1-binding site at position Ϫ269 (GH 50 / region E/mut) completely eliminated the GnRHAg response, suggesting that region E at position Ϫ276/Ϫ269 of the mGn-RHR gene promoter is critical for GnRH responsiveness. Transfection with the mutant construct of region C (GH 50 /region C, SURG-1mut) resulted in a 3.6 Ϯ 0.4-fold increase in luciferase activity, which was significantly different from both GH 50 -pXP2 alone and from the wild type mutant, GH 50 /Ϫ308/Ϫ264 (p ϭ 0.041 and p ϭ 0.0003, respectively; ANOVA) (Fig. 5B). Once again, basal luciferase activity was not significantly different between the various constructs (data not shown). These data suggest that region E (SURG-2), which contains the consensus sequence for the AP-1-binding site, is critical for GnRH responsiveness of the mGnRHR gene and that region C (SURG-1) contains an enhancer element which may be necessary for the full GnRH response.
To investigate further the importance of SURG-2 and SURG-1 in the GnRHAg-stimulated response of the mGnRHR gene, single or multiple copies of these regions were synthesized by PCR, either alone or in combination, placed upstream of the GH 50 minimal promoter in GH 50 -pXP2, and transfected into ␣T3-1 cells as described. Both GH 50 /SURG-2 and GH 50 / SURG-1 demonstrated a significant increase in luciferase activity in response to GnRHAg stimulation as compared with GH 50 (8.8 Ϯ 2.7-fold and 2.9 Ϯ 0.6-fold, respectively; p Ͻ 0.001),
Identification and Characterization of trans-Factors by EMSA-Using nuclear extracts from ␣T3-1 cells and 32 P-endlabeled Ϫ308/Ϫ220 of the mGnRHR gene promoter as probe, two distinct protein-DNA bands could be identified on EMSA that were not present with probe alone (Fig. 7). Nuclear extracts from cells grown to approximately 40% confluence appeared to give optimum binding as compared with nuclear extracts derived from cells grown to approximately 20, 60, or 80% confluence (Fig. 7). These results are in keeping with data from transfection studies suggesting that cells grown to Ͼ40 -50% confluence showed a marked reduction in GnRHAg responsiveness (data not shown). Further EMSA experiments were therefore standardized to nuclear extracts from ␣T3-1 cells grown to 40 -50% confluence. GnRHAg stimulation (100 nM for 4 h) of ␣T3-1 cells prior to preparation of nuclear extract increased the intensity of the lower band by 1.9 Ϯ 0.4-fold (p Ͻ 0.05; Student's t test), suggesting the presence of a specific, GnRH-responsive protein within the complex (Fig. 7). A shorter GnRHAg stimulus (100 nM for 1 h) appeared to give similar results (1.8 Ϯ 0.5-fold (data not shown)). The intensity of the upper band did not change significantly with GnRHAg stimulation.
Since SURG-2 containing the AP-1-binding site had been shown to be critical for GnRH responsiveness in the mGnRHR gene (above), anti-Fos and anti-Jun antibodies (Santa Cruz Biotechnology) were used in antibody-supershift EMSA experiments to identify and characterize further the trans-factors present within the protein-DNA complex. Supershift of the lower band with an anti-Fos antibody suggests the presence of a Fos protein within the complex (Fig. 8A). Similarly, incubation with an anti-Jun blocking antibody resulted in diminution in binding of the lower band, suggesting the presence of a Jun protein within the complex (Fig. 8B). However, this effect was moderate at best and required large amounts of anti-Jun antibody (8 l/lane as compared with only 2 l/lane for positive control). Positive control for the anti-Jun blocking antibody FIG. 6. Confirming the importance of regions SURG-1 and SURG-2 on GnRHAg-stimulated luciferase activity in ␣T3-1 cells. A, GH 50 -linked constructs containing single or multiple copies of SURG-2 (region E, AP-1-binding site) and/or SURG-1 (region C) were synthesized as detailed above and transfected into ␣T3-1 cells. Measurements are expressed as fold stimulation of luciferase activity by GnRHAg (100 nM for 4 h). Results are mean Ϯ S.E. from multiple experiments. *, p Ͻ 0.0001 compared with pXP2 (positive control). **, p Ͻ 0.001 compared with all other reactions. #, p Ͻ 0.0001 compared with all other reactions. q, p Ͻ 0.001 compared with GH 50 , GH 50 / SURG-1, and GH 50 /SURG-2/Not-1/ SURG-2. B, similar experiments were carried out using the full-length (1.2 kb) mGnRHR gene promoter containing mutations of SURG-2 (Ϫ1164/ϩ64/SURG-2mut) or SURG-1 (Ϫ1164/ϩ64/SURG-1mut). *, p Ͻ 0.0001 compared with pXP2 (positive control). #, p Ͻ 0.02 compared with Ϫ1164/ϩ62 and Ϫ1164/ϩ62/SURG-1mut but not significant compared with pXP2. q, p Ͻ 0.01 compared with all other reactions.
included EMSA with purified human c-JUN protein (Promega) using the consensus AP-1-binding site (Santa Cruz Biotechnology) as probe (Fig. 8B). These data suggest that the lower band on EMSA is a complex with a member(s) of the Jun/Fos heterodimer superfamily. This is consistent with our transfection data suggesting that SURG-2 at position Ϫ276/Ϫ269 of the mGnRHR gene promoter, which contains the AP-1-binding site, is critical for GnRH responsiveness. Exactly which of the Jun and Fos family members make up this heterodimer complex has yet to be determined. We are also currently investigating the identity of the protein(s) responsible for the upper band seen on EMSA.
GnRHAg Stimulation of the mGnRHR Gene Promoter Is Mediated via PKC-A dose-response curve for PMA was used to standardize all subsequent experiments at a final concentration of 100 ng/ml PMA for 4 h (data not shown). ␣T3-1 cells were transfected with GH 50 -pXP2, GH 50 /Ϫ308/Ϫ264 (wild type), GH 50 /region C (SURG-1mut), or GH 50 /region E (SURG-2mut) and stimulated for 4 h with PMA (100 ng/ml), GnRHAg (100 nM), or both. In the wild type sequence, PMA stimulation resulted in a 15.2 Ϯ 3.3-fold increase in luciferase activity, which was not significantly different from the 9.6 Ϯ 2.1-fold increase seen with GnRHAg stimulation (p ϭ 0.4; Student's t test). However, simultaneous stimulation with both PMA and GnRHAg gave an additive effect resulting in a 25.2 Ϯ 1.1-fold increase in luciferase activity (p ϭ 0.02 and p ϭ 0.001 compared with PMA alone and GnRHAg alone, respectively; ANOVA) (Fig. 9A). The addition of the selective PKC antagonist, GF-109203X (1 M; Sigma), resulted in complete abrogation of the response to PMA, to GnRHAg, and to both PMA and GnRHAg to a level similar to that seen in GH 50 -pXP2 vector alone (Fig. 9A). Similar results were seen with SURG-1mut, although the magnitude of the response to both GnRHAg and PMA was significantly diminished. SURG-2mut, on the other hand, failed to demonstrate an increase in luciferase activity in response to either PMA or GnRHAg (Fig. 9A). Taken together, these data suggest that both PMA and GnRHAg stimulation of the mGnRHR gene promoter are mediated through SURG-2.
Similar experiments were carried out to investigate the role of the PKA signal transduction pathway in the response of mGnRHR gene to GnRH. Exposure to forskolin alone (25 M) did not stimulate luciferase activity, and the addition of forskolin did not influence GnRHAg-stimulated luciferase activity in either GH 50 /Ϫ308/Ϫ264 (wild type), SURG-1mut, or SURG-2mut (Fig. 9B). Once again, the magnitude of the response seen with SURG-1mut was significantly lower than that seen with the wild type construct. Similarly, the absence of GnRHAgstimulated luciferase activity with SURG-2mut was consistent with previous experiments. Furthermore, incubation with the selective PKA antagonist, SQ 22536 (1 M; Sigma), had no effect on GnRHAg-stimulated luciferase activity (Fig. 9B). Identical results were observed using 8-bromo-cAMP (1 mM; data not shown).

DISCUSSION
The maintenance of normal reproductive function in all vertebrate species is dependent on the regulation of LH and FSH synthesis and release by pituitary gonadotropes. Although the synthesis and intermittent release of the pituitary gonadotropins are affected by a number of endocrine, paracrine, and autocrine factors, the most important influence appears to be that of GnRH (6 -9). In this study, we have defined the dimeric GnRH-RE within the 1.2-kb 5Ј-flanking sequence of the mGn-RHR gene, and we have demonstrated that the AP-1 complex plays a central role in conferring GnRH responsiveness to the mGnRHR gene.
We have used ␣T3-1 cells, a well characterized mouse pituitary gonadotrope cell line, as a model for the analysis of FIG. 7. Identification of a GnRH-responsive trans-factor by EMSA. ␣T3-1 cells were grown to 20, 40, 60, and 80% confluence and treated with GnRHAg (100 nM) or vehicle for 4 h prior to preparation of nuclear extract. Using ␣T3-1 nuclear extracts and the Ϫ308/Ϫ220 PCRgenerated fragment of the mGnRHR gene promoter as probe, EMSA identified two distinct protein-DNA complex bands that were not present in probe alone (designated by arrows). Nuclear extracts from cells grown to approximately 40% confluence appeared to give optimum binding as compared with nuclear extract derived from cells grown to 20, 60, or 80% confluence. Stimulation of ␣T3-1 cells by Gn-RHAg prior to preparation of nuclear extract increased the intensity of the lower band by 1.9 Ϯ 0.4-fold as measured by a PhosphorImager.
cis-regulatory elements in the mGnRHR gene. This cell line, obtained by targeted tumorigenesis in the mouse pituitary with the SV40 large T antigen driven by the human glycoprotein hormone ␣-subunit promoter (18), has been used to study many aspects of gonadotrope physiology. A number of studies have shown that ␣T3-1 cells constitutively express GnRHR and are capable of binding and responding to exogenous GnRH (13,18). Characterization of this cell model has demonstrated many similarities in the GnRH response compared with that in mouse primary pituitary cells, including the specific intracellular signal transduction pathways activated, the degree of stimulation of the gonadotropin subunit promoter activities, and the presence of differential regulation of GnRHR and ␣-subunit gene promoter activities by GnRH (4,14,19). ␣T3-1 cells thus appear to be a useful model for the study of the regulation of expression of the GnRHR gene by GnRH.
The responsiveness of pituitary gonadotropes to GnRH correlates directly with changes in GnRHR concentrations. It has been suggested that the concentration of GnRHR on the cell surface is mediated in turn, at least in part, at the level of gene expression (6,12). Data from Northern blot analyses presented above (Fig. 1), which demonstrate a significant increase in GnRHR mRNA in response to GnRHAg stimulation which was maximal at 4 h, would support this conclusion. These findings are consistent with previous reports in primary monolayer cultures of rat pituitary cells in which GnRHR mRNA levels were significantly increased by pulses of GnRH (10 nM, 5 min/ pulse) at all pulse frequencies tested over a 24-h period (12). In contrast, Alarid and Mellon (20) found no change in GnRHR mRNA levels in ␣T3-1 cells in response to continuous exposure to GnRHAg for 1-24 h, and Mason et al. (21) demonstrated a time-and dose-dependent decrease in the level of GnRHR mRNA in ␣T3-1 cells in response to GnRH or GnRHAg. The disparity among these results may be related to cell culture conditions and the timing of GnRH stimulation. In our hands, optimal response of mGnRHR promoter-transfected ␣T3-1 cells   FIG. 8. Identification and charac-terization of trans-factors binding to critical cis-DNA elements in the mGnRHR gene promoter by EMSA. A, using nuclear extracts prepared from ␣T3-1 cells and the Ϫ308/Ϫ220 fragment of the mGnRHR gene promoter as probe, the two protein-DNA complex bands could again be identified by EMSA. Control was probe alone. An increase in intensity of the lower band was again seen with GnRHAg stimulation. Supershift of the lower band with anti-Fos antibody (Santa Cruz Biotechnology) suggests the presence of a Fos protein within this DNA-protein complex (see arrows). B, similar experiments were carried out using an anti-Jun blocking antibody (Santa Cruz Biotechnology). Results demonstrate a moderate but significant diminution in binding of the lower band suggesting the presence of Jun protein within the complex (see small arrow). To determine the specificity of the anti-Jun antibody, EMSA experiments were carried out using purified c-JUN protein (Promega) and the consensus AP-1-binding site (5Ј-TGAGTCA-3Ј) as probe. Control was probe alone. Addition of excess anti-Jun antibody resulted in significant diminishment in the intensity of the lower band (see large arrow). Addition of excess anti-Fos antibody had no effect on binding (data not shown).
FIG . 9 to GnRHAg stimulation was seen after 4 h transfection and 4 h GnRHAg stimulation. A longer transfection time resulted in increased basal expression of luciferase activity but a significantly diminished response to GnRHAg (Fig. 2B).
The mGnRHR gene has been isolated, and its major TSS has been identified (17,22,23). A 1.2-kb 5Ј-flanking region of the mGnRHR gene has been characterized and shown to be active in transfection studies (13). This region has also been used in transgenic mice to show that it is sufficient to mediate gonadotrope-specific expression in vivo (24). Preliminary studies on the 5Ј-flanking putative promoter region of the mouse, human, and sheep GnRHR genes reveal complex organization with multiple TSS that are occasionally associated with TATA boxes (13,25). In the mGnRHR gene, the major TSS was shown to be located 62 nucleotides upstream of the translational start site by primer extension and ribonuclease protection analysis of ␣T3-1 gonadotrope mRNA (13). Functional analysis by transient transfection of ␣T3-1 cells with the 1.2-kb 5Ј-flanking region of the mGnRHR gene (Ϫ1164/ϩ62 [Figs. 2 and 3]) confirmed previous observations (13,25) that this region contains an element(s) that is necessary for tissue-specific basal expression as well as for GnRH responsiveness. By using deletion and mutational analysis in ␣T3-1 cells, Duval et al. (26) recently identified a tripartite enhancer that appears to be responsible for regulating cell-specific basal expression of the GnRHR gene. Individual elements of this putative enhancer include binding sites for steroidogenic factor-1 (SF-1), AP-1, and a novel element designated GnRH receptor activating sequence. Although a number of hormones, including GnRH (7-9), estradiol (9,27), and activin A (28), either alone or in combination, are known to affect transcriptional activation of the mGnRHR gene, neither the cis-regulatory element(s) nor their cognate binding protein(s) has previously been identified. In this study, transfection of ␣T3-1 cells with serial 5Ј-deletion mutants of the 1.2-kb 5Ј-flanking region of the mGnRHR gene (Ϫ1164/ ϩ62) demonstrated the existence of an element(s) within the region Ϫ300/Ϫ232 of the mGnRHR gene promoter that is necessary for GnRH responsiveness (Fig. 3). Further studies show that the 44-bp region Ϫ308/Ϫ264 is also sufficient to mediate GnRH responsiveness in the mGnRHR gene (Fig. 4). Transfection studies using a series of scanner-linker mutations of this region further localized GnRH responsiveness to two distinct GnRH-REs, designated SURG-1 and SURG-2 (Fig. 5B). SURG-2 (5Ј-TATGAGTC-3Ј) lies at position Ϫ276/Ϫ269 and contains the consensus sequence for the AP-1-binding site. The AP-1-binding site is a ubiquitous DNA consensus motif (5Ј-TGAGTCA-3Ј) that binds a family of dimeric transcription factors, known collectively as AP-1, which are composed of Jun, Fos, and/or ATF (activating transcription factor) subunits. SURG-1 (5Ј-GCTAATTG-3Ј) lies at position Ϫ292/Ϫ285 and appears to be a novel element. The relative importance of SURG-1 and SURG-2 in mediating GnRH responsiveness of the mGnRHR gene was confirmed by demonstrating that these two elements, either alone or in concert, are capable of conferring activity on an heterologous minimal promoter, GH 50 , in ␣T3-1 cells (Fig. 6A). In addition, a point mutation of SURG-2 (C269T) in the full-length 1.2-kb 5Ј-flanking region of the mGn-RHR gene completely abrogated the GnRHAg-stimulated response, whereas mutation of SURG-1 diminished but did not completely abolish this stimulation (Fig. 6B). These data suggest that SURG-2 (AP-1-binding site) is critical to GnRH responsiveness in the mGnRHR gene, whereas SURG-1 acts as an enhancer element to facilitate full GnRH response. The putative repressor element in the region Ϫ264/Ϫ220 of the mGnRHR gene has not been further characterized.
Earlier studies identified a putative gonadotrope-specific element (5Ј-TGTCCTTG-3Ј) at position ϩ48/ϩ55 of the mGnRHR gene and suggested that this element may be important in conferring GnRH responsiveness (13). This sequence was first described in the human ␣-subunit gene as an element that binds the nuclear orphan receptor, SF-1 (29), and appears to be important for gonadotrope-specific expression of the ␣-(30) and LH␤-subunit genes (31). It has also been shown to be important in regulating cell-specific basal expression of the mGnRHR gene (26). Our studies, however, suggest that the putative gonadotrope-specific element is not involved in GnRHAg responsiveness of the mGnRHR gene. Functional transfection studies using serial 5Ј-deletion constructs of the 1.2-kb putative mGnRHR gene promoter (Ϫ1164/ϩ62) showed that downstream constructs (Ϫ232/ϩ62, Ϫ117/ϩ62, and Ϫ38/ϩ62) did not significantly stimulate luciferase activity in response to GnRHAg (Fig. 3). Results from mutational studies in the fulllength Ϫ1164/ϩ62 construct (Fig. 6B) further confirm these observations. These data are in keeping with observations made in SF-1 knock-out mice. Targeted disruption of the murine ftz-F1 gene encoding SF-1 results in adrenal and gonadal hypoplasia (32). Such SF-1 knock-out mice exhibit malformations of the ventromedial hypothalamus as well as selective deficiency of GnRHR, LH␤, and FSH␤ mRNA in the pituitary (32). However, treatment with GnRH results in partial restoration of gonadotropin subunit gene expression as well as detectable levels of mGnRHR mRNA (32). Taken together, these results along with our data suggest that SF-1 is not a critical element for mGnRHR gene expression.
Using nuclear extracts from ␣T3-1 cells, with or without GnRHAg stimulation, two distinct protein-DNA bands were identified on EMSA. Nuclear extracts from cells grown to approximately 40% confluence appeared to give optimum binding. These results are in keeping with data from transfection studies suggesting that cells grown to 40 -50% confluence showed optimal response to GnRHAg stimulation. The lower proteincomplex band on EMSA, but not the upper band, appeared to be GnRH-responsive. Using antibody blocking and supershift EMSA experiments, we have demonstrated that the lower band represents a complex containing a member(s) of the Jun/Fos heterodimer superfamily, also known as the AP-1 protein complex (Fig. 8). These data are consistent with the functional transfection studies demonstrating that the AP-1-binding site (SURG-2) is critical for GnRH responsiveness of the mGnRHR gene. Anti-Fos supershift EMSA experiments resulted in a supershift of the entire lower band. Anti-Jun blocking EMSA, on the other hand, resulted in only a modest inhibition in binding. It is possible that Jun may not itself bind to cisregulatory elements and that AP-1 stimulation of mGnRHR gene transcription may be mediated by Jun interaction with another protein(s), perhaps Fos, that directly binds to the proximal mGnRHR promoter. A similar observation was made by Bruder et al. (33) investigating the role of AP-1 in the repression of GnRH gene transcription in GT1-7 neuronal cells. Exactly which of the Jun and Fos family members make up this heterodimer complex has yet to be determined. The identity of the DNA-protein complex responsible for the upper band seen on EMSA has yet to be characterized but may represent a trans-factor binding to the SURG-1 cis-regulatory element.
Serum is a highly effective stimulus of primary response genes, including fos and jun. Endogenous steady-state levels of Fos and Jun in ␣T3-1 cells cultured in the presence of 10% fetal calf serum are relatively high (34). However, levels of c-fos, c-jun, as well as junB mRNA have been shown to decrease progressively over a 6-h period in ␣T3-1 cells cultured in serum-free medium. The final steady-state mRNA level of all three of these trans-factors after 6 h was around 3-5% of that observed in cells cultured in the presence of 10% fetal calf serum (34). In transcription studies detailed above, optimal response to GnRHAg stimulation was observed in ␣T3-1 cells cultured overnight in serum-free medium (data not shown). By having subsequently identified a member of the Jun/Fos heterodimer superfamily as the critical element in GnRH responsiveness of the mGnRHR gene, we hypothesize that an overnight incubation in serum-free medium might decrease basal (endogenous) levels of Jun and Fos, thereby allowing for an enhanced response to GnRHAg.
The expression of the GnRHR, ␣-, FSH␤-, and LH␤-subunit genes are all strictly regulated in pituitary gonadotropes by GnRH and other hormones. Separate and independent manipulation of each of these related genes in a single cell type requires a complex regulatory system. The individual components of this regulatory system are poorly defined but probably include activation of distinct signal transduction pathways, the presence of specific cis-regulatory elements in the promoter regions of each of these genes, and the incorporation of different trans-factors and/or coactivators/corepressors which differentially regulate gene transcription. The GnRH-RE(s) in each of the gonadotropin subunit genes have been partially characterized. In the LH␤ gene, two putative Sp1-binding sites in the proximal promoter region appear to play an important role in conferring GnRH responsiveness (35), and two transcription factors, SF-1 and early growth response-1 (Egr-1), are known to be involved in tissue-specific expression of this gene (31,36). There is no consensus Sp1-binding site in the mGnRHR gene. Analysis of the ␣-subunit promoter in ␣T3-1 cells and in transgenic mice (14,37) have led to the identification of multiple cis-regulatory elements that appear to be important for tissuespecific basal expression of the ␣-subunit gene. Such elements include a binding site for a LIM homeodomain protein (38), several canonical E boxes (39), the ␣ACT element that binds members of the GATA binding factor family (40), and the gonadotrope-specific element that binds the SF-1 transcription factor (29). A binding site for Ets factor (a family of transcription factors that have been implicated in mediating transcriptional responses to mitogen-activated protein kinase activation) has also been identified in the ␣-subunit gene promoter (41). The precise GnRH-RE(s) in the common ␣-subunit gene, however, has not been characterized. This may be due in part to the observation that the common ␣-subunit gene is less well regulated by GnRH as compared with the GnRHR, LH␤-, and FSH␤-subunit genes (10). Recent studies in the ovine FSH␤ gene have identified two AP-1 enhancers in the proximal promoter that appear important for tissue-specific basal FSH␤ expression in vivo (42). These same AP-1 elements also appear to mediate GnRH-stimulated transcription of the ovine FSH␤ gene in primary cultures of ovine pituitary cells (43). There are no confirmed AP-1 consensus sequences in either common ␣or LH␤-subunit gene promoters.
The mechanism(s) by which a common and ubiquitous cisregulatory element, such as the AP-1-binding site, is able to regulate differentially both GnRHR and FSH␤-subunit genes within the pituitary gonadotrope cell remains unclear. A number of potential mechanisms exist. Specific homo-and heterodimer members of the AP-1 family, for example, may differentially regulate target genes through a common cis-regulatory element. Alternatively, the same AP-1 trans-factor may interact with different protein kinases and/or transcriptional coactivators/corepressors to affect distinct biological functions. Our data suggest another possibility, namely the incorporation of one or more secondary cis-elements, such as SURG-1 for the GnRHR gene (above). Exactly which of the AP-1 family members binds to the GnRHR and FSH␤-subunit promoter regions are not known. There is, however, evidence to suggest that different signal transduction pathways may be involved in the regulation of these two genes. Although GnRHAg stimulation of both FSH␤-subunit (43) and GnRHR genes (above) appears to be mediated through PKC, inhibition of PKC activity completely blocked GnRHAg-mediated stimulation of the GnRHR gene (above) but only partially blocked that of the FSH␤-subunit gene (43).
The single copy GnRHR gene is well conserved between the species, as is its putative promoter sequence. Indeed, there appears to be 69 -71% homology of the entire 1.2-kb 5Ј-flanking region among the mouse (13), rat (44), human (45,46), and sheep genes (47). The sequence homology of SURG-1 and SURG-2 between various species (Fig. 10) shows a relatively high concordance between sequences in the mouse, rat, and human. The sheep GnRHR gene promoter is poorly characterized but does not appear to contain a consensus AP-1-binding site (47). It is likely that different mechanisms are involved in the GnRH-mediated activation of the GnRHR gene in different species. For example, the 5Ј-flanking region of the human Gn-RHR gene is far more complex than that of the other species (45,46,48). It is larger (ϳ2.3 kb), contains multiple TSS, and numerous putative cis-regulatory sequences have been identified by sequence homology, including thyroid hormone-RE, glucocorticoid/progesterone-RE, cAMP-RE, PEA-3, AP-1, AP-2, and Pit-1 sites (46,48).
The intracellular signal transduction pathways within pituitary gonadotropes, which are involved in regulating gonadotropin subunit and GnRHR gene transcription, are still not clearly described but likely include phosphoinositides, calcium, and cAMP as second messengers and/or mitogen-activated protein kinase cascades (see Ref. 2 for review). GnRH induction and basal regulation of the ␣-subunit gene seems to occur through the PKC/mitogen-activated protein kinase pathway, whereas induction of the LH␤ gene is dependent on calcium influx (49). In this study, PMA stimulation of luciferase activity in ␣T3-1 cells transfected with Ϫ308/Ϫ264 of the mGnRHR gene was similar to that seen with GnRHAg (Fig. 9A). The addition of GF-109203X (Sigma), a inhibitor selective for the PKC isoforms ␣, ␤ 1 , ␤ 2 , ␥, ␦, and ⑀, resulted in complete abrogation of the response to either PMA or GnRHAg. Simultaneous stimulation with optimal doses of both PMA and GnRHAg resulted in an additive effect as compared with each agonist alone, which was similarly completely blocked by GF-109203X. These data suggest that both PMA and GnRHAg stimulation of the mGnRHR gene are mediated via PKC. The additive effect of PMA on GnRHAg stimulation implies either that these two agents act through different PKC isoforms or that they act synergistically through the same PKC pathway. Whatever the mechanism, it is clear from the transfection data presented above that both agents act at least in part through the SURG-2 consensus sequence. Similar experiments were carried out using forskolin and 8-bromo-cAMP to investigate the role of the PKA signal transduction pathway in the GnRHAg response. Neither agonist was able to stimulate luciferase activity in ␣T3-1 cells transfected with Ϫ308/Ϫ264 nor were they able to influence GnRHAg-stimulated luciferase activity in such cells.
Similarly, the addition of SQ 22351 (Sigma), a competitive inhibitor of adenylate cyclase, had no effect on GnRHAg stimulation (Fig. 9B). These findings were not unexpected given that no cAMP-response element-like sequence has been identified in the mGnRHR gene (13,25), although there may be cAMP-response element-like elements in the rat and human GnRHR genes (44,45). Taken together, these data suggest that the response of the mGnRHR gene to GnRHAg stimulation is mediated via PKC and not PKA. These observations are consistent with a number of previous reports suggesting that PKC and its activators increase GnRH binding activity in pituitary gonadotropes (50,51) but in contrast with other studies in which phorbol esters did not affect levels of GnRHR mRNA in ␣T3-1 cells, whereas forskolin decreased GnRHR mRNA (20). Whether this discrepancy can be explained on the basis of post-transcriptional modification has yet to be determined. Although GnRH is known to induce levels of cAMP in gonadotropes both in vitro and in vivo (52,53), GnRHAg stimulation of gonadotropin secretion appears to be independent of changes in cAMP (54). These data do not exclude the possibility that second messengers such as calcium and/or mitogen-activated protein kinase cascades may be involved in this response downstream of PKC. Indeed, the observations that both GnRH and PMA induce rapid increases in mRNA levels for primary response genes (including jun and fos) with a peak response at around 30 min (34), whereas maximal response of the mGn-RHR gene to GnRHAg stimulation is achieved at around 4 h (above), suggest that a more complex intracellular signal transduction pathway may be involved.
While this paper was being completed, Lin and Conn (55) reported on transcriptional activation of the mGnRHR gene by GnRH and cAMP in GGH 3 cells (GH 3 cells stably expressing GnRHR). By using the same full-length mGnRHR promoter construct as that detailed above (13), the authors localized the major putative GnRH-RE(s) to the region Ϫ331/Ϫ255 (relative to the major TSS) of the mGnRHR gene promoter, which is in keeping with our data. In our studies, response to GnRHAg stimulation (100 nM for 4 h) ranged from 10-to 12-fold ( Fig.  2-3) as compared with a 2-fold response to GnRHAg stimula-tion (Buserelin; 100 nM for 6 h) reported by Lin and Conn (55). This discrepancy may be accounted for by the use of different cell lines (␣T3-1 cells and GGH 3 cells, respectively) but is more likely due to differences in cell culture conditions. As demonstrated above (Fig. 2B), optimal response to GnRHAg stimulation was seen after 4 h transfection. Longer transfection times were associated with higher basal luciferase activity but a marked reduction in GnRHAg responsiveness. In the study by Lin and Conn (55), cells were transfected for 24 h. By using serial 5Ј-deletion constructs of the full-length (12 kb) mGnRHR gene promoter in transfection studies, the authors demonstrated a 1.5-2-fold response to both GnRHAg and dibutyryl-cAMP in the Ϫ255/ϩ62 construct, which they suggested was statistically significant. Comparable studies detailed above (Fig. 3) demonstrated a similar 2.3-2.4-fold response to Gn-RHAg stimulation in the Ϫ232/ϩ62 and Ϫ117/ϩ62 constructs, but these measurements were not statistically different from pXP2 alone.
In summary, we have used deletion and mutational analysis as well as functional transfection studies in the murine gonadotrope-derived ␣T3-1 cell line to localized GnRH responsiveness of the mGnRHR gene to two DNA sequences at Ϫ276/ Ϫ269 (SURG-2, the AP-1 consensus binding site) and Ϫ292/ Ϫ285 (a novel element designated SURG-1), and we demonstrated that this response is mediated via PKC. By using EMSA, we further demonstrate that a member(s) of the Fos/ Jun heterodimer superfamily is responsible for the DNA-protein complexes formed using ␣T3-1 nuclear extracts. These data define the dimeric GnRH-RE in the mGnRHR gene promoter and suggest that the AP-1 complex plays a central role in conferring GnRH responsiveness to the mGnRHR gene.