Critical Involvement of the Hinge Region of the Follicle-stimulating Hormone Receptor in the Activation of the Receptor*

The follicle-stimulating hormone receptor (FSHR) is a G-protein-coupled receptor with a large hormone-specific extracellular domain (amino acids (aa) 1–366) and a characteristic seven-transmembrane domain (TMD; aa 367–695). The extracellular domain is composed of leucine-rich repeats (LRRs; aa 18–259) connected to TMD by a hinge region (HinR; aa 260–366), whose role in the hormone action is not clearly understood. We generated a novel polyclonal HinR antibody that specifically stimulates cAMP production by HEK 293 cells expressing FSHR in a hormone-independent manner. The monovalent antibody retained the stimulatory potential. The segment of aa 296–331 in HinR was identified as the binding site for the stimulatory antibody. Deletions of this entire segment or any 10 amino acids within this segment from FSHR led to complete loss of antibody response and, surprisingly, response to the hormone as well despite all mutants exhibiting cell surface receptor density and affinity comparable with the wild type receptor. Interestingly, these mutants exhibited higher basal cAMP production. The mutant lacking LRRs with the intact HinR (Δ18–259) showed suppressed basal activity, which increased significantly with deletions extending to aa 331. These data suggest that the segment 296–331 acts as a tethered inverse agonist of the TMD and plays a very critical role in the hormonal activation of FSHR. The mutants lacking LRRs failed to bind FSH, whereas deletions in HinR had no effect on hormone binding, indicating that the LRRs constitute the primary high affinity binding site, whereas HinR may not play a significant role in FSH binding.

The human glycoprotein hormones, follicle-stimulating hormone (FSH), 3 thyroid-stimulating hormone (TSH), luteinizing hormone (LH), and the chorionic gonadotropin, are heterodimeric proteins with an identical ␣ subunit annealed noncovalently with the hormone specific ␤ subunit (1). The receptors for these hormones belong to the family of G-protein-coupled receptors and have a hormone-specific large extracellular domain (ECD) believed to be involved in specific interactions with the hormone and a characteristic seven-transmembrane domain (TMD) involved in transducing the signal. The ECD of these receptors can be further divided into two distinct regions, the N-terminal leucine-rich repeat domain (LRRs), with 9 -10 such LRRs being present in each receptor and the hinge region (HinR) that connects the LRRs to the TMD (2,3). The roles of these domains in hormone binding and signal transduction have been intensively investigated in the past few years (4 -6).
Several mechanisms of receptor activation have been proposed (2,6). In a widely discussed model, the ECD, particularly of the thyroid-stimulating hormone receptor (TSHR), serves as an inverse agonist, stabilizing the TMD in an inactive conformation, and the hormone binding or the activating mutations disengage the inhibitory ECD-TMD interactions (7)(8)(9)(10)(11)(12)(13)(14)(15). This hypothesis is supported by the observation that the mutants lacking the ECD exhibited high basal activities (9,11). A recent report by Mizutori et al. (7) narrowed this inverse agonistic activity of the TSHR-ECD to a 14-amino acid region within the HinR and showed that the LRRs are not involved in this process.
In addition, they also demonstrated a significant contribution of the HinR to high affinity hormone binding and signal transduction. In contrast, the Luteinizing hormone receptor (LHR) ECD had no negative influence on the basal activity, since deletion of the entire ECD did not cause any significant increase in the basal cAMP production. In fact, this deletion tended to reduce the LHR basal activity. Moreover, the mutationally induced constitutive receptor activation was diminished in the absence of ECD (16,17). These observations indicate specific and distinct roles for the ECDs in TSHR and LHR. However, a direct role for FSHR-ECD in such processes has not been reported.
The crystal structure of human FSH-human FSHR (hFSH-hFSHR) indicated that the hormone binds to the ␤ strands on the LRRs toward the concave portion of the receptor (18), whereas its interactions with the HinR were not delineated at all. However, such interactions cannot be completely ruled out (19 -21). The structure of the HinR of FSHR or TSHR is not known and is difficult to model because of low homology with other proteins in the data base.
In the present study, we have explored the role of the HinR of hFSHR, both as a silencer and an activator of the TMD. We have employed antibodies to different rat FSHR ECD fragments to identify the regions of the receptor involved in hormone binding and signaling and report a novel rat FSHR-HinR antibody that can stimulate the receptor (both rat and human) in the absence of the hormone. After precisely locating the stimulatory antibody binding site in the HinR of the FSHR-ECD, we further demonstrate that this site plays a dual role in 1) silencing the basal activity in the absence and 2) receptor activation in presence of the hormone while not being involved in the primary high affinity hormone binding.

EXPERIMENTAL PROCEDURES
Hormones and Antibodies-The iodination grade-purified hormones and cAMP antiserum were obtained from the National Hormone and Peptide Program, NIDDK, National Institutes of Health. The recombinant hFSH used in the study was expressed using the Pichia pastoris expression system, purified from the fermentation medium using a procedure similar to that used for purification of the recombinant chorionic gonadotropin (22) and shown to be fully biologically active in vitro. 4 Molecular Biology-The restriction and modifying enzymes were purchased from MBI Fermentas and New England Biolabs. Phusion TM hot start high fidelity DNA polymerase was obtained from New England Biolabs. The transfections were carried out using the Lipofectamine transfection reagent purchased from Invitrogen. All other reagents were obtained from Sigma. The propagation and construction of different expression vectors were according to standard procedures (47).
FSHR Fragments-The constructs expressing rat FSHR fragments were generously provided by Dr. James Catterall (Population Council, New York) (23). The receptor fragments corresponding to aa 18 -142 (RF1), aa 218 -336 (RF2), and aa 23-336 (RF3) were expressed and purified under urea denaturing conditions using Ni 2ϩ -nitrilotriacetic acid chromatography, and the denaturant was removed by stepwise dialysis against distilled water, concentrated, and kept frozen until used in the experiments. The receptor fragment corresponding to amino acids 18 -366 (RF5) was expressed as a glutathione S-transferase fusion protein and purified from the bacterial lysate using the glutathione S-transferase affinity matrix (48).
Receptor Antibodies-Polyclonal antibodies were generated against the purified receptor fragments using the procedure described earlier (24 -26). The ability of these antibodies to recognize the respective receptor fragments as well as the fulllength receptor was demonstrated by carrying out an enzymelinked immunosorbent assay with different receptor fragments as described earlier (22). The IgGs were purified using Protein G-Sepharose (GE Healthcare) chromatography. The Fab fragments were generated from the IgG by papain digestion and purified by removing the Fc and the undigested IgGs by Protein A-Sepharose chromatography. The RF2 region-specific antibodies were purified by affinity chromatography using the RF2 affinity matrix prepared by coupling the protein to NHS-Sepharose (GE Healthcare) and eluting the bound antibodies with 0.1 M glycine-HCl, pH 2.8, followed by immediate neutralization with 1 M Tris-HCl, pH 8.0.
Constructions of hFSHR Mutants in pcDNA3.1-Myc-His Vector-A synthetic 57-nucleotide double-stranded fragment corresponding to the cognate signal sequence of hFSHR was first cloned into pcDNA3.1-Myc-His (Invitrogen) at the Hin-dIII and EcoRI site, thus creating a base vector with hFSHR signal peptide. The constructs expressing the N terminally truncated hFSHR fragments were generated by inserting the appropriate PCR fragments in frame into this base vector downstream of the signal sequence between the EcoRI-XhoI sites without the termination codon, thus fusing each receptor fragment to the Myc and His tags at the C terminus of the protein. Deletion constructs were generated by amplifying the fragments upstream and downstream of the deletion sites and cloning them sequentially into the pcDNA3.1-Myc-His between KpnI-EcoRI and EcoRI-XhoI sites, respectively.
Site-directed Mutagenesis-Artificial mutation P272G/S273I or the mutations reported elsewhere in the primary amenorrhea patients (A419T and L610V) (27)(28)(29) were introduced into the hFSHR by a two-stage PCR protocol for the site-directed mutagenesis (30). The mutations were confirmed by DNA sequencing. A189V and D567G hFSHR were kind gifts from Dr. Ilpo Huhtaniemi (Imperial College, London) and Dr. Jorg Gromoll (Institute of Reproductive Medicine, Munster, Germany), respectively.
Construction of Stable Cell Line Expressing hFSHR-The HEK 293 cells were transfected with the full-length hFSHR cDNA cloned into pcDNA3.1 obtained from the University of Missouri (Rolla, MO), and the transformants were selected on a medium containing G-418 (500 g/ml). The clones expressing hFSHR were identified by 125 I-hFSH binding (31). The protocol for preparation of particulate membranes was adapted from the method described by Kene et al. (31).
Transfection Experiments-HEK 293 cells were seeded into 6-well (ϳ1 ϫ 10 6 /well/2 ml) or 24-well plate (ϳ3 ϫ 10 5 cells/ well/500 l) or 48-well (ϳ1 ϫ 10 5 cells/well/250 l) tissue culture plates in DMEM supplemented with 10% fetal bovine serum and grown to obtain nearly 80 -90% confluence and transiently transfected with different recombinant hFSHR plasmids (3200 ng of the plasmid DNA/ml of the plating medium) using Lipofectamine TM reagent as per the manufacturer's protocol (Invitrogen). After 4 h of incubation with the transfection mixture, the cells were replenished with fresh medium and used for transgene expression studies 48 h later. In each experiment, parallel plates were transfected simultaneously to determine the ligand binding to the intact cells and membrane preparations, cAMP production, and Western blot analysis.
Receptor Binding-Radioiodination of hFSH and RF2 Fab was carried out using the IODO-GEN method (32). The specific binding of 125 I-hFSH/RF2 Fab to hFSHR and its mutants was demonstrated as described earlier (33). The binding data obtained with the displacement analysis were converted into the Scatchard plots. The ability of various antibodies (IgGs) to inhibit 125 I-hFSH binding was investigated by preincubating the receptor preparations with antibodies followed by the addition of 125 I-hFSH and determining its binding (33). The data were analyzed with the Prism version 5.0 (GraphPad Software, Inc.).
Binding Assays with Intact Cells-The HEK 293 cells transiently transfected with various expression constructs were washed twice with 1ϫ Dulbecco's phosphate-buffered saline (Invitrogen) containing 1 mg/ml bovine serum albumin and incubated with 2 ϫ 10 5 cpm of 125 I-hFSH (ϳ2 ng/ml) for 3 h at 37°C in a CO 2 incubator in a total volume of 500 l. At the end of the incubation, the cells were washed three times with 1ϫ Dulbecco's phosphate-buffered saline with 1 mg/ml bovine serum albumin and solubilized in 1 N NaOH (500 l), and an aliquot of the lysate was counted in a ␥-counter. The nonspecific binding was determined by incubating the labeled hormone with the cells in presence of excess of the unlabeled hFSH (1 g/ml).
In Vitro Bioassay and cAMP Measurement-Approximately 1 ϫ 10 5 cells/well (stable cell lines or transiently transfected) were plated in a 48-well plate and 24 h later were incubated with fresh medium containing 1 mM phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine for 30 min at 37°C in a CO 2 incubator (100 l) and then exposed to varying concentrations of hFSH or IgGs for 15 min at 37°C (100 l). The cells were lysed with the addition of 200 l of 0.2 N HCl, and the total cAMP produced was determined by cAMP RIA as described elsewhere (31,34).
Gel Electrophoresis and Immunoblotting-Equal amounts of purified membrane preparations of the HEK 293 cells expressing various hFSHR constructs were resolved on 12.5% SDS-PAGE under reducing conditions and either stained with Coomassie Brilliant Blue or transferred to polyvinyl difluoride membranes and subjected to Western blot analysis using monoclonal anti His tag antibody or FSH receptor antibodies.

RESULTS
Characterization of Human FSHR Cell Line-The cell line expressing human FSHR (hFSHR cells) generated in the laboratory exhibited high affinity and capacity for hormone binding (K D ϭ 1.37 nM, B max ϭ 4.5 pmol/mg membrane protein) and responded to hFSH with a dose-dependent increase in total cellular cAMP with EC 50 of 0.56 nM (Fig. 1). Characterization of the rat FSHR cell line was described earlier (31).
Characterization of FSHR-specific Antibodies-Antibodies raised to different rat FSHR fragments were characterized for their binding to different rat FSHR fragments (supplemental Fig. 1, a (Western blot) and b (enzyme-linked immunosorbent assay)). Although RF2 antibodies could bind to RF2, RF3, RF5, and the peptide 296 -331, but not to the peptide 26 -48, confirming that this antibody recognizes the HinR, RF5 antibodies recognized all of the fragments of rat FSHR.
Effect of RF2-IgG on hFSH Binding and Response-The ability of different rat FSHR fragment-specific antibodies to influence the functionality of hFSHR was investigated by determining their effects on hormone binding and the basal as well as the hormone-stimulated cAMP production.
The cells expressing hFSHR were preincubated with 50 g/ml normal rabbit or RF2/RF3/RF5 antibodies for 1 h at 37°C, followed by the addition of 125 I-hFSH for an additional 1 h, and labeled hormone bound to the cells was determined. As shown in Fig. 2B, RF2-IgG inhibited hormone binding by nearly 30%, whereas RF5-IgG inhibited hFSH binding by nearly 80%.
To investigate the effect of RF2-IgG on hFSH-stimulated response, the cells were first incubated with the antibody, followed by the addition of the hormone and determination of cAMP production at the end of 15 min of incubation. As shown in Fig. 2C, although RF2-IgG had only a marginal effect (11% inhibition) on response to the hormone, RF5 antibody inhibited the response significantly (40% inhibition).
Interestingly, RF2-IgG itself stimulated cAMP production by the hFSHR-expressing cells. As shown in Fig. 2D, RF2-IgG (250 g/ml) stimulated a 34-fold increase in cAMP production over the basal level, whereas RF3 and RF5 antibodies did not stimulate cAMP production at all. The Fig. 2E shows the dose-response curve for the total RF2-IgG, which shifted to the left with the immunoaffinity-purified RF2-IgG (RF2-IgG-IP), suggesting ϳ35-fold purification of the stimulatory antibodies. RF2 immunoaffinity chromatography of RF5-IgG (which inhibited hFSH binding and response; Fig. 2, B and C) yielded antibodies (RF5-IgG-IP) that showed stimulation of cAMP production similar to that shown by the RF2-IgG, whereas the unbound antibodies retained the inhibitory activity (Fig. 2F). The stimulatory activities of both immunoaffinity-purified RF2 and RF5 IgGs could be blocked by preincubating the antibodies with RF2 protein prior to the addition to the hFSHR cells (Fig.  3C).
The specificity of the stimulatory activities of RF2-IgG-IP and RF5-IgG-IP was further confirmed by incubating the antibodies with HEK 293 cell lines expressing human LHR and TSHR that were generated in the laboratory. These cell lines produce high affinity receptors and respond to the respective hormones in terms of cAMP production (Fig. 3B). Both RF2-IgG-IP and RF5-IgG-IP failed to stimulate cAMP production by these cells, thus confirming the specificity of the stimulatory effect. Moreover, RF2-IgG-IP and RF5-IgG-IP failed to recognize human LHR and human TSHR in immunoblotting analysis with the membrane preparations derived from the HEK 293 cells expressing these receptors (Fig. 3A), confirming the binding specificity of these antibodies.

FIGURE 2. Effect of receptor antibodies on FSH-FSHR interactions.
A, schematic representation of various fragments of rat FSHR used in the study. B, hFSHR membrane preparation was incubated with IgGs of different rat FSHR antibodies (50 g/ml) for 1 h at 30°C, followed by the addition of 125 I-hFSH and continuation of the incubation for 1 h at 30°C, at the end of which the hormone-receptor complex was precipitated by the addition of 2.5% polyethylene glycol and centrifugation at 5000 ϫ g. The radioactivity in the pellet was counted. The nonspecific binding was determined by incubating the labeled hormone with an excess (1 g/ml) of the unlabeled hFSH. The numbers above the bars denote the percentage of inhibition of 125 I-hFSH binding in the presence of antibodies. C, the hFSHR cells were incubated with IgG of FSHR antibodies (50 g/ml) in the presence of hFSH (1 nM) for 15 min at 37°C, and the cAMP produced was determined by RIA. The numbers above the bars denote the percentage of inhibition of cAMP production in the presence of antibodies. The basal cAMP production was determined without the addition of either hFSH or antibodies in all of the experiments. D, the hFSHR cells were incubated with IgGs of FSHR antibodies (250 g/ml) in the absence of hFSH for 15 min at 37°C, and cAMP produced was determined by RIA. The numbers above the bars denote -fold increase in cAMP production in the presence of antibodies over the basal level. E, the hFSHR cells were incubated with increasing concentrations of NR-IgG, total RF2, and immunoaffinity-purified RF2-IgG-IP for 15 min at 37°C, and cAMP produced was determined by RIA. F, the hFSHR cells were incubated with the NR-IgG, total RF5-IgG, RF5-IgG-IP, RF3-IgG, and RF3-IgG-IP for 15 min at 37°C, and cAMP produced was determined by RIA. In the same experiment, the effect of RF5-IgG in the unbound fraction (UF) of RF2 immunoaffinity purification (50 g/ml) on response to hFSH was determined as described in C. All of the experiments presented above are representative of two independent experiments.
To demonstrate that the stimulatory effect could be obtained with the monovalent antibody, Fab fragments were prepared from the RF2-IgG-IP. Direct binding of the Fab to the receptor was demonstrated by radioiodinating the Fab and determining its specific binding to the hFSHR cells. As shown in Fig. 4A, 125 I-labeled Fab could bind to the hFSHR cells but not to the HEK 293 cells. 125 I-Fab binding to hFSHR could be inhibited by the unlabeled Fab as well as the RF2 protein, further confirming the specificity of the antibody binding. As shown in Fig. 4B, RF2 Fab, but not the normal Fab, stimulated cAMP production by the hFSHR cells, clearly indicating that the monovalent RF2 antibody was capable of eliciting the response.
The ability of either the hormone or RF2-IgG-IP to influence each other's binding was investigated by determining the affinities of each ligand for the receptor in the presence of the other. Binding of 125 I-hFSH to the receptor was determined in the presence of increasing concentrations of the unlabeled hFSH in the absence or presence of RF2-IgG-IP/NR-IgG, and the binding data were converted into a Scatchard plot. The dissociation constants for hFSH (K D ) in the presence and absence of RF2-IgG-IP were 0.63 and 0.67 nM, respectively, and in the presence of NR-IgG, it was 0.60 nM, indicating that the affinity of the receptor for hFSH was not altered by RF2-IgG-IP. Similarly, RF2 Fab binding to the hFSHR was not affected by the hormone (K D of 125 I-Fab in the presence and absence of hFSH was 1.753 and 1.69 nM, respectively) (Fig. 5B). The simultaneous additions of RF2-IgG-IP (25 g/ml) and a subsaturating concentration of hFSH (0.1 nM) resulted in an additive response (Fig. 5C). Further, the purified, functional ECD of hFSHR, expressed using the P. pastoris expression system, captured on a solid phase using RF2-IgG-IP was able to bind 125 I-hFSH (data not shown).
These data together suggest that RF2 antibody and hFSH do not affect each other's binding, and since they bind to the receptor simultaneously, they have independent binding sites on the receptor.
Identification of the Stimulatory Antibody Binding Site-To identify the exact site recognized by the stimulatory antibodies, the ability of different hFSHR fragments to inhibit the stimulatory effect of RF2/RF5-IgG-IP was next investigated by first incubating the IgGs with different overlapping hFSHR peptides corresponding to aa 218 -336, followed by the addition of the complex to the hFSHR cells and determination of cAMP produced. Earlier, it was demonstrated that RF2 protein inhibits the effect of stimulatory antibodies (Fig. 3C). Different fragments of the RF2 region inhibited the stimulatory effect to varying degrees, the peptide 296 -331 being the most effective (Table 1, ϳ85% inhibition), suggesting that this segment  . RF2 Fab retains stimulatory potential. A, the hFSHR cells were incubated with 125 I-RF2 Fab (2 ϫ 10 5 cpm) for 3 h at 37°C. The cells were washed twice with Dulbecco's phosphate-buffered saline and lysed with 1 N NaOH, and an aliquot was counted for radioactivity. The nonspecific binding was determined by incubating the cells with the labeled probe in presence of excess of RF2 Fab (10 g/ml) and RF2 protein (25 g/ml). B, the hFSHR cells were incubated with increasing concentrations of RF2-IgG-IP or RF2 Fab fragments for 15 min at 37°C, and the cAMP produced was quantified by RIA as described in the legend to Fig. 2D. All of the experiments presented above are representative of two independent experiments. encompasses the antibody binding site and thus could be directly involved in antibody-stimulated activation of hFSHR. The peptide 26 -48, did not neutralize the stimulatory effect of the antibodies, confirming the specificity of the assay.
Effect of Deletions in hFSHR on Response to the Hormone and the Antibody-Direct involvement of aa 296 -331 in the receptor activation was next investigated. Different deletion mutants of hFSHR were generated with a hexahistidine tag (His tag) at the C-terminal end of the receptors (Fig. 6A), and their responses to hFSH and RF2-IgG-IP were determined in transient transfection experiments.
The mutant receptor expression on the cell surface was demonstrated by different methods. Direct binding of 125 I-hFSH to the intact cells was determined by incubation of 125 I-hFSH with the cells and determining the specifically bound hormone. The mutant receptors were characterized for their affinity by carrying out 125 I-hFSH binding to the membrane preparations obtained from the cells after transient transfection and converting the binding data into Scatchard plots. Receptor expression was also confirmed by Western blot analysis of the membrane preparations with His tag antibody. Finally, the cell surface expression of the  receptors was demonstrated by immunocytochemistry using RF2 and RF5 antibodies.
As shown in Tables 2-5, all mutant receptors except for those lacking LRRs showed cell surface binding of the hormone with affinities similar to that of the full-length receptor. The mutants lacking LRRs did not bind the hormone, but their presence in the membranes was confirmed by immunoblotting with His tag monoclonal antibody. Finally, immunocytochemistry with RF5 antibody showed that ⌬18 -259 and ⌬18 -331 hFSHR were translocated to the cell surface (supplemental Fig. 3). Interestingly, A189V hFSHR, which has been shown to be incapable of translocation to cell surface (35), did not show any hormone binding to the intact cells (Table 5), further supporting the cell surface expression analysis.
Consequent to the loss of hormone binding (Table 2), the deletion mutants ⌬18 -259 hFSHR, ⌬18 -331 hFSHR, and ⌬18 -366 hFSHR completely lost the hormone response (Fig.  6C) despite adequate cell surface expression (Fig. 6B and supplemental Fig. 3). The mutant with the intact HinR (⌬18 -259 hFSHR) exhibited low basal cAMP production and nearly wild type response to the antibody. With further deletions in the ECD (⌬18 -331 hFSHR and ⌬18 -366 hFSHR), loss of antibody stimulation was observed primarily due to the absence of the stimulatory antibody binding site. However, these two mutants showed a significant increase in the basal cAMP production (Fig. 6C). The mutant ⌬296 -331 hFSHR (lacking the stimulatory antibody binding site) exhibited adequate cell surface receptor density with wild type affinity but failed to respond either to the hormone or the antibody ( Table 2, Fig. 6C, and supplemental Fig. 3). However, this mutant exhibited a relatively higher basal cAMP production. These data suggest that amino acids within 296 -331 in the HinR are probably involved in attenuation of the basal activity as well as playing a critical role in receptor activation by both ligands.
To identify the precise region involved in the antibody stimulation, mutant receptors with progressive sequential deletions of ten amino acids starting with amino acids 260 -329 were characterized for binding and response to both ligands. All of these deletion mutants showed near wild type cell surface expression of the receptors with affinities equivalent to the wild type receptor (Table 3). However, the effect of these deletions on the basal as well as ligand-stimulated cAMP production varied with individual deletions. Whereas the mutant ⌬260 -269 hFSHR exhibited wild type-like basal cAMP levels, ⌬270 -279 hFSHR exhibited very high basal cAMP production, which was brought down to the relatively lower levels in the ⌬280 -289 hFSHR mutant. All of the three mutants responded to both the hormone and the antibody. The maximal levels of cAMP produced in response to the antibody were nearly the same as in the wild type. However, the response to the hormone was very low compared with the wild type (Fig. 7). Further progressive deletions until amino acid 329 (⌬290 -299, ⌬300 -309, ⌬310 -319, and ⌬320 -329) resulted in complete absence of response to both hormone and the antibody with a concomitant increase in the basal cAMP production. These results were the same as with the parent deletion ⌬296 -331 hFSHR analyzed before (Fig. 6C).
Effect of Activating Mutations of the HinR on hFSHR Deletion Mutants-Point mutations at amino acids 272 and 273 (P272G and S273I) have been shown to constitutively activate the receptor, resulting in higher basal cAMP production (12). Both of these mutations were introduced in combination in hFSHR wild type, ⌬18 -259 hFSHR, and ⌬296 -331 hFSHR mutants. As expected, introduction of these double mutations in the wild type receptor resulted in high basal activity (ϳ20-fold), which remained responsive to the hormone and exhibited a small, but significant, increase in cAMP production in response to RF2-IgG-IP (Fig. 8). Introduction of these mutations in ⌬18 -259 hFSHR resulted in higher basal activity, which could not be stimulated further with the antibody. Similarly, the same mutations in ⌬296 -331 hFSHR also neither increased the basal cAMP production (which was already higher) nor exhibited any response to the antibody (Fig. 8).
Next, the effect of stimulatory antibodies inactivating and activating mutations of hFSHR was investigated. The RF2-IgG-IP was unable to rescue the inactivating mutants A419T and L601V (27,29), whereas it could stimulate further a constitutively activated mutant D567G (28) ( Table 5 and Fig. 9).

DISCUSSION
Antibodies to glycoprotein hormones and their receptors have often provided insights into the mechanism of hormonereceptor interactions and signal transduction. Although the TSH receptor antibodies and their effects on the overall physiology have been well documented (36,37), reports of such antibodies against FSHR or LHR and their possible effects on the reproductive functions are not available. In this study, we show that an antibody specific for the HinR of FSHR can bypass the hormone and stimulate the receptor directly, thus eventually leading to identification of the region involved in hormonal activation of the receptor.
The three different domains of these receptors carry out distinct functions. The N-terminal LRR portions bind the hormones (18), whereas the role of TMD in signal transduction is well documented (2,3,6). Although the role of the HinR in hormone binding and signaling has not been conclusively demonstrated, there is evidence for its involvement in these events. Earlier studies predicted that the HinR is in contact with the exoloops (19,20), and in the case of TSHR, it acts as an inverse agonist of the TMD, with aa 371-384 being responsible for this inverse agonism (7). The present study clearly demonstrates the role of HinR of hFSHR in signaling. The novel feature of this study is that an antibody specific for the HinR stimulated the receptor bypassing the hormone but did not influence the hormone binding (Fig. 2, D and E). Further, it was demonstrated that this antibody recognizes amino acids 296 -331 within the HinR of hFSHR ( Table 1). The stimulatory nature of this antibody was rather unique, since it may be difficult to produce a similar antibody. However, another antibody raised against the entire ECD of rat FSHR (RF5) also yielded a small population of stimulatory antibodies upon immunoaffinity chromatography, pointing to the possibility of generating more such antibodies (Fig. 2F). However, it was recently reported that a TSHR stimulatory human monoclonal antibody (M22) produced from the lymphocytes of a patient with Graves' disease clearly recognized TSHR-LRRs and inhibited hormone binding to the receptor (38,39). The RF2 antibody recognizes different region of the receptor and does not interfere in the hormone-receptor interactions but still has the ability to activate the receptor (Fig. 5, A  and B). The antibody can also bind to hFSHR-ECD complexed with the hormone (data not shown). 5 Similarly, a monoclonal antibody (106.105) that recognizes hFSHR aa 318 -333 also did not interfere in the hormone-receptor interactions when it was used to capture soluble hFSHR-ECD, suggesting that the C-terminal portion of the ECD is not directly involved in hormone binding (40). These data strongly suggest that the HinR of FSHR may not play a significant role in high affinity hormone binding.
Further insights into the role of the HinR in hormone binding and signal transduction were obtained by creating several deletions and point mutations. An adequate expression of mutants on the cell surface is often a major concern in such experiments. Therefore, cell surface expression of these mutants was ensured by various different strategies, and most of the mutants with small deletions were present on the surface as indicated by hormone binding to the intact cells, and their affinities for the hormone were nearly the same as that of the wild type receptor. Expression of mutants lacking LRRs was confirmed indirectly by Western blot and immunocytochemistry.
Deletion of the entire LRRs in mutants ⌬18 -259 hFSHR, ⌬18 -331 hFSHR, and ⌬18 -366 hFSHR resulted in complete 5 G. Agrawal, J. A. Dias, and R. R. Dighe, unpublished observations. FIGURE 6. Response of various deletion mutants of FSHR to hFSH and RF2-IgG-IP. A, schematic representation of the deletion mutants of hFSHR (not drawn to scale). The various mutants generated were as follows: wild type (hFSHR WT; aa 1-695), ⌬18 -259 hFSHR, ⌬18 -331 hFSHR, ⌬18 -366 hFSHR, and ⌬296 -331 hFSHR. B, the HEK 293 cells were transiently transfected with various deletion mutants, and 48 h later, the membranes were prepared, electrophoresed, and probed with anti-His tag antibody. Each immunoblot is a representative of three independent experiments. C, the HEK 293 cells were transiently transfected with hFSHR and its mutant constructs, as described under "Experimental Procedures." The transfected cells were incubated with hFSH (10 nM), RF2-IgG-IP (25 g/ml), and NR-IgG (25 g/ml) for 15 min at 37°C, and the cAMP produced was determined by cAMP RIA. The results presented are representative of two independent experiments. The cell surface expression of the wild type or mutant hFSHR was determined in a parallel experiment.
loss of hormone binding and consequently response, confirming the involvement of LRRs in hormone recognition (18). Whereas the mutant ⌬18 -259 hFSHR retained response to the RF2-IgG-IP, further deletion of 72 amino acids (⌬18 -331 hFSHR mutant) abolished the antibody stimulation. Second, the mutant ⌬296 -331 hFSHR retained hormone binding; however, the antibody response was completely abolished, confirming the identity of the antibody binding site between amino acids 296 and 331.
These mutants revealed the important role played by the HinR in signaling. The basal activity of ⌬18 -259 hFSHR was low, but the deletion of next 72 amino acids (⌬18 -331 hFSHR) led to significant increase in the basal cAMP production, indicating that aa 259 -331 of the HinR has an attenuating effect on the basal activity of the receptor (Fig. 6C). The mutant ⌬296 -331 hFSHR, which also exhibited high basal activity, allowed us to further narrow down the attenuating function to this segment of the HinR. Thus, these data support the concept that the ECD of hFSHR acts as a tethered inverse agonist of the TMD. Residues within the HinR, particularly between positions 296 and 331, appear to be involved in inverse agonistic activity of the ECD.
Most interestingly, the mutant ⌬296 -331 hFSHR also showed no further response to the hormone, indicating that this part of the receptor is also critical for hormonal activation, perhaps playing a dual role in the attenuation of the basal activity and a direct involvement in the hormonal activation of the receptor (Fig. 6C). Progressive sequential deletions of 10 amino acids from 290 to 329 yielded similar results (high basal cAMP production with concomitant loss of hormone and antibody response), clearly demonstrating that the integrity of this region is absolutely essential for hormonal activation (Fig. 7). FIGURE 7. HinR is indispensable for signal transduction but dispensable for hormone binding. The HEK 293 cells were transiently transfected with hFSHR and its mutant constructs, and response to hFSH and RF2-IgG-IP was determined as described in the legend to Fig. 6C. The results presented are representative of two independent experiments.

TABLE 2 Effect of various deletions in hFSHR on receptor expression
The intact cells expressing the mutant receptors were incubated with 125 I-hFSH (2 ϫ 10 5 cpm, approximately 2 ng/ml) for 3 h at 37°C. The cells were washed twice with Dulbecco's phosphate-buffered saline and lysed with 1 N NaOH, and the aliquot was counted for radioactivity. The nonspecific binding was determined by incubating the cells with the labeled probe in the presence of excess of hFSH (1 g/ml) and was found to be similar to that seen with the HEK 293 cells transfected with pcDNA3.1-Myc-His. The membrane preparation obtained from different mutants was incubated with 125 I-hFSH in the presence of increasing concentrations of hFSH, and binding data were converted into a Scatchard plot to obtain K D and B max values. ND, not determined.

TABLE 3 Effect of various deletions in the HinR on receptor expression
The whole cell binding and affinities with membrane preparations were determined as described in Table 2. ND, not determined.

TABLE 4 Effect of activating mutations (PS272-273GI) in the HinR on receptor expression
The whole cell binding and affinities with membrane preparations were determined as described in Table 2. ND, not determined.

Effect of various mutations in hFSHR on receptor expression
The whole cell binding and affinities with membrane preparations were determined as described in Table 2. ND, not determined. Progressive 10-amino acid deletions from the aa 260 -289 resulted in high basal activity only with the ⌬270 -279 hFSHR, whereas it was almost equivalent to the wild type in the mutants ⌬260 -269 and ⌬280 -289. As discussed below, the mutations in the HinR at the conserved proline and serine residues at the 272-and 273-positions were proposed to disengage the suppressive HinR-TMD interactions (12). Thus, a high basal activity displayed by ⌬270 -279 hFSHR was not surprising, since it lacked these critical residues. cAMP levels in all three mutants in response to RF2-IgG-IP were similar to the response exhibited by the wild type receptor, but hormonal response was significantly lower despite having adequate receptor densities and normal affinity (Table 3 and Fig. 7). These data suggest that aa 260 -289 do play a key role in hormone signaling but not in hormone binding.

Construct
Since the deletion of aa 296 -331 or any smaller deletion within this segment increases basal cAMP production, it was possible that these two parts function together to suppress the basal cAMP production. This hypothesis was tested by mutat-ing P272G/S273I in the mutants ⌬18 -259 hFSHR and ⌬296 -331 hFSHR. The ⌬18 -259 mutant, which showed low basal cAMP production and responded to the RF2 antibody, showed a higher basal activity and did not respond further to the antibody with the P272G/S273I mutations. The mutant P272G/ S273I ⌬296 -331 hFSHR was similar to the base mutant (⌬296 -331 hFSHR) with high basal level and lack of responsiveness to both antibody and the hormone (Fig. 8 and Table 4). However, a synergistic or additive effect on the basal cAMP production was not observed. These results suggest that both sites contribute together to the inverse agonistic activity of the HinR, and disrupting either site results in a receptor activation that could not be suppressed by the other. Hence, as proposed earlier (7,9), these mutations in the HinR in fact appear to disengage the suppressive HinR-TMD interactions. Thus, the stimulatory antibodies, by binding to 296 -331 segment, perhaps bring about the disengagement. However, although the function of 272-273 site appears to be restricted to suppression of the basal activity, the region 296 -331 has a more profound and multifaceted involvement in the receptor function, first being critical for hormonal activation and second being critical in suppression of the basal constitutive activity of the receptor.
A large array of reproductive abnormalities is associated with malfunctioning of FSHR. To explore the possibility of using the stimulatory antibodies for therapeutic purposes, three inactivating mutations of hFSHR were analyzed. In corroboration with the earlier reports (27,29), the mutants A419T and L601V are incapable of transducing the signal, despite having adequate cell surface expression and wild type affinities for the hormone ( Fig. 9 and Table 5), mainly because of defective TMD. The RF2 antibody failed to elicit any response from these mutants, suggesting that its ability to activate the receptor depends on the status of the TMD. Interestingly, the activating mutant D576G, which showed very high basal cAMP production, could be stimulated by both antibody and the hormone to nearly wild type levels ( Fig. 9), suggesting that in this mutant, the interactions between the HinR and TMD are similar to that of wild type, and higher basal cAMP production could be due to different interactions of the TMD with the G-proteins.
The increase in cAMP production seen in the wild type receptor in response to the hormone was much higher than that observed with the RF2 antibody. The same was the case with different mutants. This is probably due to the multiple contact points between the hormone and the receptor leading to full stimulation of the receptor that is almost certainly not achieved by the antibody or the activating mutations. A similar observation has been reported for TSHR mutants lacking the ECD or harboring activating mutations (9). This excludes the model in which normal activation of the receptor would simply result from the release of a silencing effect exerted by the ECD on the TMD via interaction with the extracellular loops (9).
Activation of glycoprotein hormone receptors by the ligands is most likely to be a two-stage process, with LRRs being involved in the high affinity initial binding and the secondary binding sites present in the HinR. Previously, we have demonstrated that the complimentarity determining regions of a ScFv that recognized an epitope formed jointly by the ␤L2 and ␣L3 loops of chorionic gonadotropin exhibited homology to the  amino acids within the HinR of the LHR (aa 252-259 and aa 324 -332), suggesting that this portion of the HinR is in contact with the hormone (41). Perhaps the interaction between the ␤L2 ϩ ␣L3 region of the hormone and the specific HinR residues, which is probably a secondary event following high affinity interaction with LRRs, results in a conformational change leading to the maximal receptor activation (Fig. 10). It is not clear whether the initial high affinity binding to the LRRs results in such a conformational change in the HinR leading to generation of secondary binding sites. Although the HinR specific antibody and the HinR deletion mutants used in this study had no effect on hormone binding, there are other reports suggesting involvement of HinR in the process (7,42). A very recent report identified Asp-330 and Tyr-331 in the HinR of the LHR as the key residues in the hormone-induced receptor activation (43), supporting the previous studies where tyrosine sulfation at these key residues was demonstrated to be crucial for receptor activation (44,45). All of these studies further highlight the importance of the HinR in signal transduction. Whether this entire region is interacting with any of the exoloops or helices of the TMD and thus brings about attenuation of signaling needs to be investigated. This possibility has been suggested in a previous report by Nishi et al. (10), where the fusion of the ectodomains of LH and FSH receptors with the TMD of fly LGR2 receptor resulted in a high constitutive basal cAMP production, which could be attenuated by replacing the key residues within the exoloop-2 of the TMD with the human LHR sequences, suggesting a direct interaction of HinR with the exoloops of the TMD. However, later it was demonstrated that deletion of the entire LHR-ECD caused no significant increase in the basal cAMP production (16,17,46). Nevertheless, the present study with hFSHR and previous studies with the TSHR (7,9,11,14,15) support an activation model in which the ECD of the receptor functions as a tethered inverse agonist constraining the receptor in an inactive state. Thus, ECDs of glycoprotein hormone receptors differ in their abilities to regulate basal cAMP production, and hFSHR appears to be more similar to TSHR rather that the LHR with respect to the activation mechanism.
The observations that the deletions in the HinR do not affect hormone binding clearly support the notion that LRRs constitute the primary hormone binding sites, and the HinR may or may contribute to the binding event. This is further supported by the observations that the mutants lacking LRRs are unable to bind the hormone. In conclusion, our data suggest that the HinR of hFSHR contributes more to the signaling with LRRs, being clearly involved in high affinity hormone binding. Moreover, the inverse agonistic activity that attenuates the basal constitutive activity of the TMD is restricted to the HinR, whereas LRRs have no role in this process. FIGURE 10. Schematic representation of glycoprotein hormone receptor activation and the possible role of the HinR. Based on the observed constitutive activation of hFSHR mutated in the HinR, it is proposed that the receptors are present in a constrained conformation in the unliganded state (a). Upon agonist binding to the LRRs in the ectodomain (c) or following point mutations of the key residues (Pro-272 and Ser-273) in the HinR or/and deletion of amino acids between 296 and 331 or upon binding of the stimulatory antibodies, the receptor assumes a relaxed conformation (R*) (b) capable of activating the G s protein. Since some of the hFSHR HinR mutants could be further stimulated by its ligand, an additional role for the interaction of the common ␣ subunit with the HinR and TMD region can be envisaged, resulting in the stabilization of the active conformation of the receptor (HR***), leading to its full activation.