Gonadotropin-releasing Hormone Receptor Initiates Multiple Signaling Pathways by Exclusively Coupling to G q/11 Proteins*

The agonist-bound gonadotropin-releasing hormone (GnRH) receptor engages several distinct signaling cas-cades, and it has recently been proposed that coupling of a single type of receptor to multiple G proteins (G q , G s , and G i ) is responsible for this behavior. GnRH-depend- ent signaling was studied in gonadotropic a T3–1 cells endogenously expressing the murine receptor and in CHO-K1 (CHO#3) and COS-7 cells transfected with the human GnRH receptor cDNA. In all cell systems studied, GnRH-induced phospholipase C activation and Ca 2 1 mobilization was pertussis toxin-insensitive, as was GnRH-mediated extracellular signal-regulated kinase activation. Whereas the G i -coupled m2 muscarinic re- ceptor interacted with a chimeric G s protein (G s i5) containing the C-terminal five amino acids of G a i2 , the hu- man GnRH receptor was unable to activate the G protein chimera. GnRH challenge of a T3–1, CHO#3 and of GnRH receptor-expressing COS-7 cells did not result in agonist-dependent cAMP formation. GnRH challenge of CHO#3 cells expressing a cAMP-responsive element-driven firefly luciferase did not result in increased reporter gene expression. However, coexpression of the human GnRH receptor and adenylyl cyclase I in COS-7 cells led to clearly discernible GnRH-dependent cAMP formation subsequent to

The decapeptide GnRH 1 plays a central role in the neuroendocrine control of reproduction. GnRH is synthesized in the hypothalamus and released into the hypophyseal portal circulation in a pulsatile fashion to interact with a membranous receptor on gonadotropes in the anterior pituitary gland (1). In these cells, the synthesis and the secretion of gonadotropins is subject to differential regulation by GnRH (2), and the neuropeptide is additionally involved in the long term maintenance of the gonadotrope phenotype (3). Cloning of cDNAs coding for GnRH receptors of various species demonstrated that GnRH interacts with a membrane receptor belonging to the large superfamily of heptahelical G protein-coupled receptors (2,4).
Apart from the effects of GnRH on the hypothalamic-pituitary-gonadal axis, extrapituitary actions in the central and peripheral nervous system, as well as in several extraneural and also in neoplastic tissues have been described (5). Current evidence is consistent with the notion of an autocrine/paracrine regulatory GnRH system exerting a growth regulatory effect on various cell types (6). In a number of human malignancies, such as breast, ovary, endometrium, and prostate cancers, the expression of GnRH and its receptor, as well as a direct antiproliferative effect of GnRH analogues, could be demonstrated (7)(8)(9). Most notably, the nucleotide sequence of the GnRH receptor in breast, ovarian, and prostate cancer cells is identical to that expressed in the pituitary gland (10). In vivo studies with nude mice strongly suggest that the antitumor activity of a GnRH receptor antagonist is not only exerted through suppression of the pituitary-gonadal axis but also through a direct effect of the GnRH analogue on tumor cells (11).
Considering the plethora of cellular effects elicited by GnRH, one may ask where whithin the neuropeptide-induced signal transduction cascade divergent pathways are engaged. One potential answer would be to postulate an inherent ability of the GnRH receptor to couple to multiple G proteins, which would then enable the receptor to activate multiple signal transduction pathways (2,12). This view, however, is not uncontested, and despite the deluge of circumstantial evidence, direct proof of the activation of multiple G proteins by the agonist-bound GnRH receptor is still missing.
To better understand the cellular actions of GnRH and its analogues, we set out to study the G protein coupling profile of the murine GnRH receptor, endogenously expressed in ␣T3-1 cells, and of the human receptor expressed in CHO and COS-7 cells. We provide direct evidence that the GnRH receptor in its native environment exclusively couples to G q/11 proteins. Therefore, divergent GnRH-induced signal transduction is determined downstream of the receptor/G protein interface.

Materials
All chemicals were obtained from Sigma with the following exceptions: buserelin was purchased from Welding, and the ECL system and myo-[ 3 H]inositol (18.6 Ci/mmol) were from Amersham Pharmacia Biotech. Pertussis toxin (PTX) and cholera toxin (CTX) were from List Biological Laboratories, and human chorionic gonadotropin was from Calbiochem. LipofectAMINE was obtained from Life Technologies, fura-2/AM was from Molecular Probes, Dowex 1-X8 100 mesh resin was * This work was supported by the Deutsche Forschungsgemeinschaft and Fonds der Chemischen Industrie. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
from Bio-Rad, and rabbit polyclonal anti-ERK2 antibody was from Santa Cruz Biotechnology. Cetrorelix was a gift from Schering AG. The adenylyl cyclase type I (AC-I) cDNA was a gift from Dr. A. G. Gilman. Human V2 vasopressin receptor, m2 muscarinic receptor, and G s i5 cDNAs were provided by Dr. J. Wess, and the plasmid pADneo2-C6-BGL was provided by Dr. A. F. Czernilofsky (13). Cell culture medium and supplies were from Life Technologies, Inc.

Methods
Cell Culture and Transfection-COS-7 and ␣T3-1 cells were cultured in DMEM containing 10% heat-inactivated fetal calf serum, penicillin (50 units/ml), and streptomycin (50 g/ml) under 7% CO 2 at 37 C°. For transfections, 1 ϫ 10 6 cells were seeded into 60-mm dishes. Twenty-four hours later, cells were transfected with various cDNA constructs (3 g of plasmid DNA per dish) by lipofection. Cells permanently expressing the human GnRH receptor (CHO#3) and the human V2 vasopressin receptor (CHO V2-R) were generated as described before by double selection cotransfecting the puromycin and hygromycin resistance plasmids pBSpac⌬p and pSK/HMR272, respectively (14,15). Single cell clones were obtained following a limited dilution procedure. CHO#3 and CHO V2-R cells were cultured in Ham's F-12 medium as described above.
Radioactive Labeling of Buserelin-Buserelin was labeled with 125 I by the chloramine T method at the Department of Isotope Chemistry of Schering AG. The tracer was purified by reverse-phase high pressure liquid chromatography on a Spherisorb ODS II column by eluting with 50% acetonitrile, 0.15% trifluoroacetic acid at a flow rate of 0.5 ml/min. The retention time of mono-125 I-buserelin was approximately 23 min. The specific activity of the tracer was 2000 Ci/mmol.
Membrane Preparation-Plasma membranes were prepared from cells grown to confluence. Cells were harvested; pelleted (1000 ϫ g for 10 min); resuspended in a buffer consisting of 0.25 M sucrose, 0.01 M triethanolamine, pH 7.4; and subsequently disrupted by nitrogen cavitation. Nuclei were pelleted at 750 ϫ g for 5 min, and the supernatant was centrifuged at 30,000 ϫ g for 30 min at 4°C. Membrane pellets were homogenized in a glass-Teflon homogenizer in assay buffer (0.25 M sucrose, 0.01 M triethanolamine, pH 7.4, ovalbumin 1 mg/ml) and stored as 200-l aliquots at Ϫ70°C. The protein content of samples was determined by the method of Bradford (16).
Radioligand Binding Assays-Saturation binding studies were performed with 10 g of membrane protein for ␣T3-1 and 40 g for CHO#3 cells, 1500 -200,000 cpm 125 I-buserelin, and assay buffer in an incubation volume of 100 l per sample. Incubations were carried out for 90 min at room temperature. Nonspecific binding was determined in the presence of excess unlabeled buserelin (10 Ϫ6 M). Bound and free ligand were separated by filtration on Whatman GF/C filters using an Amicon 10-fold sampling device. Filters were presoaked with 0.3% polyethyleneimine for 30 min in order to reduce nonspecific binding. The filters were washed twice with 5 ml 0.02 M Tris/HCl, pH 7.4. Bound radioactivity was determined with a ␥-counter. Binding data were analyzed with the help of the computer program LIGAND (17).
Measurement of Intracellular Inositol Phosphate and cAMP Accumulation-For cAMP and inositol phosphate (IP) measurements, cells were seeded into 12-well plates (3 ϫ 10 5 cells/well for ␣T3-1 and CHO#3 cells; 2 ϫ 10 5 cells/well for COS-7 cells) 3 days prior to functional assays. For cAMP determinations, cells were washed once in serum-free DMEM, followed by a 20-min preincubation with the same medium containing 1 mM 3-isobutyl-1-methylxanthine (Sigma) for 20 min at 37°C in a humidified 7% CO 2 incubator. Subsequently, cells were stimulated with appropriate concentrations of agonist or antagonist for 1 h. Reactions were terminated by aspiration of the medium and addition of 1 ml 5% trichloroacetic acid. The cAMP content of the cell extracts was determined as described previously (15). For IP determinations, cells were incubated with 2 Ci/ml of myo-[ 3 H]inositol for 18 h. Thereafter, cells were washed once with serum-free DMEM containing 10 mM LiCl. Agonist-induced increases in intracellular IP levels were determined by anion exchange chromatography.
Measurement of Intracellular Ca 2ϩ Mobilization-Determination of [Ca 2ϩ ] i was performed after loading of cells with 5 M fura-2/AM as described previously (18). Briefly, semiconfluent grown cells were detached and incubated (2 ϫ 10 7 cells/ml) in incubation buffer (138 mM NaCl, 6 mM KCl, 1 mM MgSO 4 , 1 mM CaCl 2 , 1 mM Na 2 HPO 4 , 5 mM NaHCO 3 , 5, 5 mM glucose, 10 mM HEPES, 0.1% bovine serum albumin, pH 7.4) containing 5 M fura-2/AM for 30 min at 37°C. The cells were then pelleted and resuspended at a density of 1 ϫ 10 6 cells/ml in incubation buffer in 2-ml aliquots at room temperature in the dark until used. To obtain R max and R min , Triton X-100 (reduced form, Sigma; final concentration, 0.1%) and EGTA (pH 8.5; final concentration, 20 mM) were added sequentially to the cell suspensions. Autofluorescence was recorded in the presence of 20 mM MnCl 2 . Concentration of cytosolic Ca 2ϩ was calculated according to Grynkiewicz et al. (19).
cAMP Response Element-dependent Luciferase Assays-For luciferase assays CHO cells permanently expressing the human GnRH receptor (CHO#3) or the human V2 receptor (CHO-V2-R) were transiently transfected with the plasmid pADneo2-C6-BGL containing six heterologous cAMP-responsive element sequences. Two days later, cells were stimulated for 5 h with 20 M forskolin, 1 M GnRH, or 100 nM argininevasopressin (AVP). After removal of the incubation medium, cells were washed with phosphate-buffered saline and subsequently lysed in assay buffer containing ATP and 1% Triton X-100. After luciferin addition the luciferase activity was measured in a Lumat LB9501 luminometer (Berthold).
ERK Activity Assays-For ERK mobility shift assays, ␣T3-1 cells (5 ϫ 10 5 cells/well) were grown to confluence in six-well dishes, washed once with phosphate-buffered saline, and incubated in serum-free DMEM for 16 h in the absence or presence of PTX (0.1 g/ml). Thereafter, cells were washed with phosphate-buffered saline again and treated for 10 min with various compounds (see figure legend) dissolved in DMEM. After an additional washing step with ice-cold phosphatebuffered saline, the cells were lysed on ice in SDS sample buffer (pH 6.8). Lysates were sonicated and boiled for 3 min. Forty l of these lysates were used for SDS-polyacrylamide gel electrophoresis on 10% polyacrylamid gels run at 10 mA overnight. ERK2 was detected with anti-ERK2 antibodies by immunoblotting.
For immune complex kinase assays, endogenously expressed ERK was immunoprecipitated with ERK2 antibodies and protein A-Sepharose from ␣T3-1 cells grown to semiconfluence in 100-mm dishes. Immune complexes were washed three times with kinase buffer (40 mM HEPES, pH 7.5, 5 mM magnesium acetate, 2 mM dithiothreitol, 1 mM EGTA and 200 M Na 3 VO 4 ), and reactions were performed with 250 g/ml myelin basic protein (Sigma). The kinase assay was started in the presence of 50 M ATP solution containing 2 Ci of [␥-32 P]ATP (NEN Life Science Products). Incubations were carried out at room temperature for 20 min and were terminated by the addition of 88% formic acid. Reaction mixtures were then spotted onto Whatman p81 chromatography paper and washed four times in 150 mM phosphoric acid. Incorporated radioactivity was determined by liquid scintillation spectrometry. All assays were performed in duplicate. Basal cpm values in independent experiments ranged between 1500 and 7000 cpm, whereas probes without ERK2 antibody were below 150 cpm.
Photolabeling of Receptor-activated G Proteins-Photolabeling of membrane G proteins and immunoprecipitation were performed as described previously (20). The following antisera were used: AS 348 (␣ s ), AS 370 (␣ q/11 ), AS 233 (␣ 12 ), AS 343 (␣ 13 ), AS 266 (␣ i common ), and AS 6 (␣ o common ). Antisera were raised against peptides corresponding to specific regions of G protein ␣ subunits and have been described before (20,21). Immunoprecipitated G protein ␣ subunits were separated on 13% polyacrylamide gels and visualized by autoradiography of dried gels with Kodak X-Omat AR-5 films (Eastman Kodak Co.) or with a phosphoimaging screen.

Membrane Expression of the Mouse and Human GnRH
Receptor in ␣T3-1 and CHO#3 Cells-To functionally characterize the human GnRH receptor (22), CHO-K1 cells were generated that permanently express the human receptor. The CHO#3 cells used throughout this study represent one clonal cell line isolated by limiting dilution. To compare ligand binding properties of the mouse and human GnRH receptors, membranes prepared from ␣T3-1 and CHO#3 cells were incubated with the agonist 125 I-buserelin. The radioligand bound to both membrane preparations in a saturable and specific manner. Scatchard transformation of the binding data (Fig. 1) revealed a single high affinity binding site for mouse and human GnRH receptors, with K d values of 0.29 nM and 0.23 nM for ␣T3-1 and CHO#3 cells, respectively. B max values of 1.1 (␣T3-1 cells) and 0.3 pmol/mg of protein (CHO#3 cells) reflected high levels of membrane expression in both cell systems (Fig. 1).
Insensitivity of GnRH-mediated Signaling to Pertussis Toxin in ␣T3-1 and CHO#3 Cells-It has been suggested that PTXsensitive G proteins contribute to GnRH-induced IP production in primary pituitary cells as well as activation of ERK-mitogenactivated protein kinases in ␣T3-1 cells (23)(24)(25). To address the issue of GnRH receptor coupling to G proteins of the G i/o family in ␣T3-1 and CHO#3 cells, agonist-stimulated IP accumulation was determined with or without PTX pretreatment (Fig. 2). As illustrated in Fig. 2, A and D, ␣T3-1 as well as CHO#3 cells responded to GnRH challenge with a concentration-dependent increase in IP accumulation with EC 50 values of 2.8 and 1.4 nM for ␣T3-1 and CHO#3 cells, respectively. PTX pretreatment neither affected basal nor GnRH (1 M)-stimulated IP production in ␣T3-1 (Fig. 2B) and CHO#3 (Fig. 2E) cells. To check whether PTX was active in these experiments, IP accumulation due to activated endogenously expressed LPA receptors was measured as recently demonstrated for human fibroblasts (26). Fig. 2, C and F, illustrates that LPA-induced IP production in ␣T3-1 as well as in CHO#3 cells were completely blocked with PTX pretreatment.
In accord with our observations, PTX did also not affect GnRH-induced calcium transients in the two cellular systems studied. When peak intracellular Ca 2ϩ -concentrations in response to increasing agonist concentrations were recorded in cells loaded with the fluorescent Ca 2ϩ indicator fura-2/AM, EC 50 values of 1.2 and 0.1 nM were determined for ␣T3-1 (Fig.  3A) and CHO#3 cells (Fig. 3B), respectively. Compared with the half-maximal GnRH concentrations needed for IP accumulation (see Fig. 2), the concentration response curves depicted in Fig. 3 are left-shifted toward lower agonist concentrations. No effect of PTX on the agonist-induced calcium transients in either cell line was detected at any hormone concentration tested (see Fig. 3). Viability of PTX was confirmed using single cell fluorometrics as described previously (27); LPA-induced Ca 2ϩ -ocillations in CHO#3 cells were completely abolished after pretreatment with 100 ng/ml PTX (data not shown).
To assess the participation of G i/o proteins to GnRH-induced ERK activation in aT3-1 cells, we performed ERK mobility shift and in vitro kinase assays on whole cell lysates obtained from for ␣T3-1 cells either stimulated with 1 M GnRH, 100 ng/ml of the phorbol ester PMA or incubated with 0.1% bovine serum albumin, serving as a control. GnRH-induced ERK mobility shifts that mirror increased phosphorylation of ERKs were observed for the 42-kDa isoform of ERK2 10 min after addition of GnRH or PMA (Fig. 4). Pretreatment of cells with 100 ng/ml PMA for 24 h to down regulate protein kinase C isoforms completely suppressed ERK activation by GnRH or short-term PMA challenge. On the contrary, preincubation of cells with PTX (100 ng/ml) did not interfere with agonistinduced ERK activation (see Fig. 4). Comparable results were obtained with the immune complex kinase assay in which GnRH elicited a 2.7-fold ERK activation that was not inhibited with PTX pretreatment but was blocked by the addition of the GnRH-specific antagonist cetrorelix (10 M). As a control for PTX activity, ERK activation in response to 10 M LPA was blocked after pretreatment of ␣T3-1 cells with PTX (100 ng/ ml). Collectively, these data strongly argue against a participation of G i/o proteins in GnRH-evoked signaling pathways in ␣T3-1 and CHO#3 cells.
Inability of the GnRH Receptor to Couple to the Chimeric G s i5 Protein-Several mutagenesis studies have shown that the C terminus of the G protein ␣ subunits plays a pivotal role in defining the specificity of receptor/G protein interaction (28). In particular, the C-terminal five amino acids are sufficient to switch the receptor specificity of G␣ q /G␣ i2 chimeric proteins. Therefore, we took advantage of the chimeric G protein ␣ subunit G s i5 characterized by the replacement of the C-terminal five amino acids of G s by the corresponding amino acids of the G i2 , thus enabling a G i -activating receptor to stimulate adenylyl cyclase. Transient expression in COS-7 cells of the human GnRH receptor in conjunction with or without G s i5 did not confer the ability upon the agonist-bound receptor to initiate intracellular cAMP production (Fig. 5). When expressed alone, the activated m2 muscarinic receptor also failed to raise intracellular cAMP levels, whereas carbachol stimulation of the m2 receptor coexpressed with G s i5 led to clearly discernible cAMP accumulation (see Fig. 5). These data show that coexpression of the G s i5 chimera with G i coupling receptors results in cAMP production and further entertain the notion that the human GnRH receptor is devoid of any G i coupling potential.
Lack of GnRH-induced cAMP Production in ␣T3-1 and CHO#3 Cells-Besides being able to couple to G q/11 proteins, the GnRH receptor has repeatedly been portrayed as a G sactivating receptor as well (29). Therefore, we tested the ability of ␣T3-1 and CHO#3 cells to respond to GnRH stimulation with increased cAMP accumulation (Fig. 6). Upon GnRH (1 M) challenge, the low intracellular cAMP levels remained indistinguishable from baseline values, whereas CTX and forskolin induced profound cAMP formation in ␣T3-1 (Fig. 6A) and CHO#3 (Fig. 6B) cells.
In an additional effort not to miss even small protracted rises in intracellular cAMP levels evoked by GnRH stimulation, CHO#3 cells were transiently transfected with a reporter plasmid (pADneo2-C6-BGL) containing the firefly luciferase gene under the transcriptional control of multiple cAMP response elements (13). CHO-V2-R cells permanently expressing the human V2 vasopressin receptor served as a positive control. Both cell lines were transfected with equal amounts of reporter plasmid, and luciferase activity in whole cell lysates was determined after a 5-h incubation with saturating concentrations of GnRH and AVP or after addition of forskolin (Fig. 7). In CHO#3 cells, neither GnRH nor AVP challenge resulted in increased luciferase activity, whereas forskolin was effective in inducing reporter gene expression (Fig. 7A). The V2 vasopressin receptor-expressing CHO cells responded to AVP stimulation with a substantial enhancement of luciferase activity com-parable to the one observed after forskolin addition (Fig. 7B).
GnRH-mediated cAMP Production in COS-7 Cells via Ca 2ϩ / Calmodulin-sensitive Adenylyl Cyclase I-In contrast to various reports on G s -dependent cAMP formation initiated by the activated GnRH receptor (12,30), we were unable to detect GnRH-stimulated cAMP production in ␣T3-1 and CHO#3 cells.
In an attempt to reconcile these discrepant results, we embarked on the hypothesis that GnRH-dependent Ca 2ϩ transients may activate Ca 2ϩ /calmodulin-sensitive adenylyl cyclase isoforms (31). One candidate enzyme, AC-I, is highly expressed in the pituitary gland (32). Thus, COS-7 cells were cotransfected with GnRH receptor and AC-I cDNAs, and cAMP accumulation subsequent to hormonal treatment was measured (Fig. 8). The G s coupling luteinizing hormone receptor served as a positive control. In AC-I-expressing COS-7 cells, GnRH treatment resulted in a substantial cAMP accumulation that was suppressed by the presence of the potent GnRH receptor antagonist [D-pGlu 1 , D-Phe 2 , D-Trp 3,6 ]-LH-RH (see Fig. 8). Increased baseline cAMP-levels in AC-I expressing cells were interpreted to reflect the basal activity of the enzyme.
Exclusive Coupling of the GnRH Receptor to G Proteins of the G q/11 Family in ␣T3-1 and CHO#3 Cells-To study GnRH receptor/G protein interaction under in situ conditions, G proteins in membranes prepared from ␣T3-1 (Fig. 9A) and CHO#3

FIG. 3. PTX-insensitive Ca 2؉ mobilization via the GnRH receptor. ␣T3-1 (A) and CHO#3 (B) cells were detached and loaded with fura-2/AM in incubation buffer (see under "Experimental Procedures").
Measurements were carried out with 1 ϫ 10 6 cells/ml in incubation buffer within seconds after the addition of agonist. Fluorescence was monitored at 37°C at various GnRH concentrations for untreated (E) or PTX-pretreated (q) cells with an LS 50 B dual wavelength fluorescence spectrophotometer (Perkin-Elmer).

FIG. 4. GnRH-mediated ERK2 activation in ␣T3-1 cells. A, cells
were incubated with (؉) or without (-) 100 ng/ml PMA or 0.1 g/ml PTX for 18 h. Confluent, serum-starved cells were then challenged with 1% bovine serum albumin (control), 1 M GnRH, or 100 ng/ml PMA for 5 min. An ERK mobility shift assay (see under "Experimental Procedures") was used to determine ERK activation. Similar results were obtained in three independent experiments. B, serum-starved cells in 100-mm dishes were stimulated for 5 min with 200 nM GnRH or 10 M LPA as indicated with (؉) or without (-) PTX pretreatment or in the presence of 10 M cetrorelix (antag.), which was added simultaneously with GnRH. ERK activity was determined with an in vitro kinase assay using myelin basic protein as a substrate (see under "Experimental Procedures"). Data represent means Ϯ S.E. of two independent experiments, each performed in duplicate. (Fig. 9B) cells were photolabeled with the nonhydrolyzable GTP analog [␣-32 P]GTP azidoanilide in the absence or presence of saturating GnRH concentrations. Stimulation with 1 M GnRH led to an increased incorporation of radioactivity into G q/11 proteins (Fig. 9, A and B), whereas the other PTX-insensitive G protein families, G 12/13 and G s , were not activated (see Fig. 9, A and B). PTX-sensitive G i and G o proteins were also found not to productively interact with the agonist-bound GnRH receptor (see Fig. 9, A and B). To show that the GnRH receptor effect on G q/11 proteins was specific in ␣T3-1 and CHO#3 cells, an antagonist control experiment using 10 M cetrorelix was performed (see Fig. 9C). G protein expression in specific in ␣T3-1 and CHO#3 cells was confirmed by immunoblotting (see Fig. 9D).

DISCUSSION
In the gonadotrope, GnRH elicits several distinct physiological responses via activation of a single class of membrane receptors (2, 12). One possibility to explain the engagement of parallel signal transduction pathways at the molecular level is the activation of several G proteins by a single agonist-occupied receptor (33). Along these lines, the proposal has recently been put forward that apart from interacting with G q/11 proteins (34), the GnRH receptor would also activate G i and G s proteins (12). Moreover, the propensity of the GnRH receptor to couple to multiple G proteins was inferred not to be gonadotropespecific but rather to represent an inherent ability of the GnRH receptor itself (12). To test this frequently reiterated paradigm of the multiple G protein coupling ability of the GnRH receptor (2,12,23,35,36), we set out to study the G protein coupling profile of the murine GnRH receptor endogenously expressed in ␣T3-1 cells as well as of the human receptor (22) in CHO-K1 and COS-7 cells.
Functional studies on the GnRH receptor have been greatly facilitated by the development of the ␣T3-1 cell line isolated from anterior pituitary tumors of transgenic mice (37). Consistent with their derivation from the gonadotrope lineage, glycoprotein hormone ␣-subunit is synthesized and secreted by these cells. In our studies, the total number of specific binding sites for GnRH analogues amounted to 1.1 pmol/mg of protein, slightly higher than those for normal mouse (0.33 pmol/mg) and rat (0.31 pmol/mg) pituitary (38), whereas the B max values for CHO#3 cell membranes (0.3 pmol/mg) heterologously expressing the the human GnRH receptor cDNA were identical with those obtained for pituitary membranes. K d values in the subnanomolar to low nanomolar range reflected high affinity binding of the GnRH superagonist 125 I-buserelin to two cell models examined by us. Whereas these data conform to those reported by McArdle et al. (39), they contrast with a rather low binding affinity of 125 I-buserelin (K d ϭ 41 nM) to the recombinant rat GnRH receptor stably expressed in GH 3 somatotropes (30). Considering the similarity of binding characteristics between our experimental systems and gonadotropic cells, it is unlikely that receptor/G protein interaction was substantially affected by receptor overexpression.
When GnRH-induced phospholipase C-␤ activation was studied in ␣T3-1 and CHO#3 cells, we were unable to detect any inhibitory effect of PTX pretreatment on IP formation and Ca 2ϩ mobilization. However, in our hands, a similar regimen substantially reduced IP formation via G i -coupled receptors (18). Stimulation of phospholipase C activity by G i coupling receptors is thought to occur via G␤␥ subunits released from activated G i proteins (40). The observation that transient expression of a C-terminal ␤-adrenergic receptor kinase (GRK2) peptide (␤ARK-ct) serving as a G␤␥ sequestrant suppresses GnRH-dependent IP formation in GH 3 cells stably transfected with the rat GnRH receptor, appears to be compatible with the latter notion (41). However, as other presumably G i -independent signaling pathways, e.g. cAMP release, are similarly suppressed in the latter study (41), the functional sequelae of ␤ARK-ct expression are not necessarily indicative of GnRHinduced G i activation. A general impairment of G protein mediated signaling processes due to a quantitative sequestration of G␤␥ subunits cannot be ruled out under the above-mentioned experimental conditions.
GnRH-mediated ERK activation in ␣T3-1 cells allegedly depends on a dual mechanism involving protein kinase C on the one hand and PTX-sensitive G proteins on the other hand (25). These observations are at odds with our results, which do not at all entertain the notion of an involvement of G i/o proteins in GnRH-dependent ERK activity but corroborate the finding that protein kinase C activity plays a pivotal role in this signaling pathway (42,43).
To further assess a potential G i coupling ability of the human GnRH receptor, coexression studies with the chimeric G protein G s i5, in which the C-terminal five amino acids of G s were replaced by the corresponding amino acids of G i2 , were per- FIG. 9. Activated G proteins in membrane preparations from ␣T3-1 (A and C) and CHO#3 (B and C) cells stimulated with 1 M GnRH. Membranes (100 g/tube for ␣T3-1 or 75 g/tube for CHO#3 cells) were photolabeled with [␣-32 P]GTP azidoanilide in the absence (-) or presence (ϩ) of GnRH and immunoprecipitated with the ␣ s , ␣ q/11 , ␣ 12 , ␣ 13 or ␣ q/11, ␣ i common (A and B), and ␣ o common (A) antisera or with ␣ q/11 antibodies alone (C) as described. In C, the GnRH antagonist cetrorelix (10 M) was added together with GnRH (lanes 3 and 6) to membranes from ␣T3-1 (lanes 1-3) and CHO#3 (lanes 3-6). Precipitated proteins were resolved in SDS-polyacrylamide gels and visualized by autoradiography. Molecular mass markers are indicated. Similar results were obtained in four (A) or two (B and C) independent experiments. D, membrane proteins of ␣T3-1 and CHO#3 cells (50 g per lane) were resolved on 10% SDS-polyacrylamide gels and subsequently blotted onto nitrocellulose membrane. Membranes were cut into strips and incubated with antisera against G␣ subunits as indicated. Molecular mass markers are shown on the right. formed in COS-7 cells. Several previous studies have shown that the C-terminal five amino acids play a key role in dictating the specificity of receptor/G protein coupling (28). Whereas coexpression of G s i5 with the m2 muscarinic receptor resulted in carbachol-induced cAMP accumulation, no functional interaction between the GnRH receptor and the chimeric G protein was observed. The notion of GnRH receptor/G i interaction has largely been derived from circumstantial evidence, for instance agonist-dependent palmitoylation of G protein ␣-subunits (12). In dispersed pituitary cell cultures, however, a challenge with phorbol myristic acid, a protein kinase C activator, was able to partially mimic the GnRH effect (44), thus casting doubts on the specificity of the experimental approach chosen. Moreover, when down-regulation of activated G proteins was used as a read-out system, stimulation of ␣T3-1 cells with a GnRH agonist resulted in enhanced agonist-dependent proteolysis of G q/11 proteins, whereas G i2 remained unaffected (45). In keeping with the latter findings, our data strongly suggest that human and mouse GnRH receptors are unable to couple to G i proteins.
Treatment of rat pituitary cell cultures with CTX increases luteinizing hormone release in response to GnRH (46), and it was subsequently inferred that the GnRH receptor might be positively linked to adenylyl cyclase. Whereas in gonadotropes GnRH-stimulated cAMP formation has never been detected (12,47), somatotropic GH 3 pituitary cells stably transfected with the rat GnRH receptor cDNA respond to prolonged incubation (Ͼ24 h) with GnRH agonists in the presence of a phosphodiesterase inhibitor with a modest increase in cAMP secretion into the culture medium (30). Moreover, in permeabilized gonadotropes, cAMP synergizes with Ca 2ϩ and phorbol esters in stimulation of exocytotic gonadotropin release (48), suggesting that the adenylyl cyclase/cAMP and phospholipase C/Ca 2ϩ / diacylglycerol pathways may converge at a later stage of the GnRH signal transduction pathway.
Therefore, we set out to examine whether GnRH stimulation of ␣T3-1 and CHO#3 cells would lead to increased cAMP formation. Dual coupling of a single receptor to G s and G q/11 proteins has been demonstrated for several G protein-coupled receptors (33), for instance the H 2 histamine (49) and the cholecystokinin type A receptor (50). In the two cell lines tested, GnRH was ineffective in raising intracellular cAMP levels, whereas the inhibition of the intrinsic GTPase activity of G␣ s and the direct activation of adenylyl cyclase by CTX and forskolin, respectively, resulted in marked cAMP accumulation. To determine whether the activated GnRH receptor could provoke a subtle and prolonged increase in the basal rate of cAMP formation, CHO#3 cells were transfected with a reporter plasmid containing the firefly luciferase under the transcriptional control of several cAMP-responsive elements (13). However, no significant cAMP responses to GnRH were detectable.
Although our findings are consistent with other studies on gonadotropes (51,52), they differ from results with the recombinant rat GnRH receptor heterologously expressed in Sf9 insect cells (53) or in COS-7 cells (54). It is noteworthy, however, that in transfected somatotropic GH 3 (12,29) as well as in COS-7 cells (54) GnRH-initiated cAMP formation increases over at least 4 orders of magnitude, whereas in the latter cell model, agonist-induced IP formation is accomplished within 2 log orders of GnRH concentrations. Thus, it is imaginable that in GH 3 and in COS-7 cells GnRH-dependent cAMP formation is caused by mechanisms other than adenylyl cyclase activation through G s proteins, for instance mediated by elevated cytosolic Ca 2ϩ concentrations (31). Certain adenylyl cyclase isoforms (AC-I and AC-III) can be activated independently from G s by Ca 2ϩ /calmodulin, and the AC-I isoform was found to be ex-pressed in the anterior piuitary gland (32). Along these lines, GnRH-dependent cAMP formation could be reconstituted in COS-7 cells by coexpression of the human GnRH receptor together with AC-I, thus delineating an alternative signaling pathway that may help to reconcile some of the conflicting results obtained when studying GnRH-stimulated cAMP formation in different cell systems. In addition, we cannot exclude the possibility that species differences between the mouse and human receptors studied by us and the rat receptor (30,54) may contribute to the discrepant functional data. Mammalian GnRH receptors lack intracellular C termini that have been implicated in receptor expression and receptor/G protein interaction and desensitization phenomena (55,56). Interestingly, the recently cloned catfish GnRH receptor posesses a C-terminal tail, and GnRH-dependent cAMP production in transfected 293T cells was detected (56,57). However, as yet, it is unclear whether this property relates to the structural peculiarity of the receptor and whether activation of G s proteins underlies the latter signaling pathway.
By definition, receptor-catalyzed GDP/GTP exchange is the sole direct measure to monitor G protein activation, whereas other approaches, such as the recording of changes in the palmitoylation pattern of G protein ␣ subunits (29), yield only indirect results. Therefore, the experimental approach that was chosen to monitor GnRH receptor/G protein interaction consisted of a combination of receptor-dependent photolabeling of G protein ␣ subunits with subsequent immunoprecipitation. Photolabeling experiments employing the nonhydrolyzable GTP analog [␣-32 P]GTP azidoanilide showed that in membranes of gonadotropic ␣T3-1 cells or of CHO#3 cells the GnRH receptor couples neither to G i nor to G s or G 12/13 proteins but exclusively activates G q/11 proteins. The experimental approach has previously been adopted for the signaling analysis of various G protein-coupled receptors (18,49). Thus, we show in the present study that the GnRH receptor physiologically expressed in gonadotropic cells elicits multiple cellular responses via selective coupling to G q/11 proteins.
The observation that many hormones and neurotransmitters evoke a plethora of physiological responses led to the concept of G protein-mediated signal transduction as a complex signaling network with divergent and convergent pathways at each transduction level (33). Thus, very rarely does signal transduction occur in a strictly linear fashion, e.g. one receptor coupling to one G protein subsequently activating one effector. In fact, most G protein-coupled receptors have the propensity to interact with more than one G protein family (58). In the present study, we provide evidence that the murine GnRH receptor in gonadotropic cells and the human GnRH receptor heterologously expressed in CHO-K1 and COS-7 cells exclusively couple to G proteins of the G q/11 family. In the case of the GnRH receptor, signaling diversity occurs at a level downstream of the receptor/G protein interface and does not reflect multiple coupling potential of a single type of receptor. Therefore, the GnRH receptor represents a valuable tool to study in depth the cellular events in response to a single class of activated G proteins.