The ectodomain of the luteinizing hormone receptor interacts with exoloop 2 to constrain the transmembrane region: studies using chimeric human and fly receptors.

Lutropin (LH) and follitropin (FSH) receptors belong to a group of leucine-rich repeat-containing, G protein-coupled receptors (LGRs) found in vertebrates and flies. We fused the ectodomain of human LH or FSH receptors to the transmembrane region of fly LGR2. The chimeric human/fly receptors, unlike their wild type counterparts, exhibited ligand-independent constitutive activity. Because ectodomains likely interact with exoloops to constrain the receptors, individual exoloops of the chimeric receptor containing the ectodomain of the LH receptor and transmembrane region of fly LGR2 was replaced with LH receptor sequences. Chimeric receptors with the ectodomain and exoloop 2, but not exoloop 1 or 3, from LH receptors showed decreases in constitutive activity, but ligand treatment stimulated cAMP production. Furthermore, substitution of key resides in the hinge region of fly LGR2 with LH receptor sequences led to constitutive receptor activation; however, concomitant substitution of the homologous exoloop 2 of the LH receptor decreased G(s) coupling. These results suggest that the hinge region of the LH receptor interacts with exoloop 2 to constrain the receptor in an inactive conformation whereas ligand binding relieves this constraint, leading to G(s) activation.

The receptors for lutropin (LH), 1 follitropin (FSH), and thyrotropin (TSH) belong to the large G protein-coupled receptor (GPCR) family with seven-transmembrane (TM) helices, but are unique in having a large N-terminal extracellular (ecto-) domain containing leucine-rich repeats important for interaction with the glycoprotein ligands (1,2). Recent studies indicate the evolution of a large family of the leucine-rich repeat-containing, G protein-coupled receptors (LGRs) with at least seven members in mammals, including the well studied glycoprotein hormone receptors (3,4) and LGR4 -7 (5,6). In addition, homologous LGRs were found in fly (LGR1 and LGR2) (7,8), nematode (nLGR) (9), sea anemone (10), and snail (11). These genes can be divided into three subgroups, each with unique structural characteristics (6). Understanding of the evolution-ary relationship of these receptors and the availability of recombinant proteins provided the opportunity to elucidate their ligand-signaling mechanisms.
The allosteric ternary complex model proposes the isomerization of GPCRs from an inactive to an active state capable of coupling to G proteins (12). This isomerization involves conformational changes that may be induced or can occur spontaneously, thus allowing the constrained receptor to relax into an active conformation. The active conformation could be achieved following ligand activation of GPCRs or by gain-of-function mutations discovered in constitutively activated GPCRs. Although most of the gain-of-function mutations were found in the TM region of different GPCRs (12)(13)(14), studies using thyrotropin and gonadotropin receptors indicate that point mutations in the ectodomain could also confer conformational changes in their TM regions. These point mutations were found in the hinge region of TSH and gonadotropin receptors (15)(16)(17)(18) and suggest the importance of the ectodomains of these receptors in constraining their TM region. In addition, deletion of the ectodomain of the TSH receptor also led to the constitutive activity of the TM region (19), whereas recent photoaffinity labeling and cross-linking experiments suggest a direct interaction between the hinge region and exoloop 2 of the LH receptor (20).
Chimeric receptor analysis has been useful in revealing the importance of specific domains of the glycoprotein hormone receptors in ligand signaling (21,22). Because the ectodomain and TM region of the three glycoprotein hormone receptors are compatible, chimeric receptors with the ectodomain of one glycoprotein hormone receptor, and the TM region and C-terminal tail of a different receptor, showed normal ligand signaling based on the specificity of the ligand-binding ectodomain. We recently found that fly LGR1, but not fly LGR2, are constitutively activated when overexpressed in mammalian cells (7). Here, we show that a chimeric receptor with the ectodomain from the human LH receptor, and the TM region and C-tail from fly LGR2, showed constitutive activity. Taking advantage of the apparent incompatibility of these human and fly receptors, we modified the structure of this chimeric receptor and demonstrated that the interactions between the ectodomain and exoloop 2 of the LH receptor are important in receptor constraint.

EXPERIMENTAL PROCEDURES
Hormones and Reagents-Purified human CG (hCG, CR-129) was supplied by the National Hormone and Pituitary Program (NIDDK, National Institutes of Health, Bethesda, MD); human recombinant FSH (Org32489) was from Organon (Oss, The Netherlands). Anti-FLAG M1 monoclonal antibody and FLAG peptide were purchased from Sigma Chemical Co. 125 I-Na was from Amersham Biosciences, Inc.
Construction of Mutant Receptor cDNAs-PCR-based mutagenesis was performed to generate mutant chimeric receptor cDNAs as previ-* This study was supported by National Institutes of Health Grant HD23273. 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.
ously described (22) using cDNA encoding human LH receptors (23), human FSH receptors (22), or fly LGR2 (8). PCR was performed with Vent DNA polymerase (New England Biolabs, Inc., Beverly, MA) in accordance with the manufacturer's instructions. To facilitate cell surface expression of the fly receptor in mammalian cells, its signal peptide was replaced with the prolactin signal peptide for secretion and tagging with the FLAG M1 epitope as previously described (7). The junctional sequences for LDR and FDR are -EPDAFNPCED (LH receptor)/ (LGR2)LFDWWTLRCG-and -KPDAFNPCED (FSH receptor)/LFD-WWTLRCG (LGR2)-, respectively. To introduce the amino acid in the hinge region of human LH receptor to fly LGR2, PCR-based mutagenesis was performed using overlapping primers containing mutated sequences as described previously (22). All cDNAs were subcloned into the expression vector pcDNA3.1 Zeo (Invitrogen Corp., Carlsbad, CA), and plasmids were purified using a Maxi plasmid preparation kit (Qiagen, Inc., Valencia, CA). Fidelity of the PCR was confirmed by sequencing on both strands of the final constructs before use in expression studies.
Transfection of Cells and Analysis of Signal Transduction-Human 293T cells derived from embryonic kidney fibroblasts were maintained in Dulbecco's modified Eagle's medium/Ham's F-12 (DMEM/ F12) supplemented with 10% fetal bovine serum, 100 g/ml penicillin, 100 g/ml streptomycin, and 2 mM L-glutamine. Before transfection, cells (2 ϫ 10 6 /culture) were seeded in 10-cm dishes (Nalge Nunc International, Naperville, IL). When cells were 70 -80% confluent, transient transfection was performed with 10 g of plasmid using the calcium phosphate precipitation method (24) following replacement of culture media. After 18 -24 h of incubation with the calcium phosphate-DNA precipitates, media were replaced with DMEM/F12 containing 10% fetal bovine serum. Forty-eight hours after transfection, cells were washed twice with Dulbecco's phosphate-buffered saline (D-PBS), harvested from culture dishes, and centrifuged at 400 ϫ g for 5 min. Cell pellets were then resuspended in DMEM/F12 supplemented with 1 mg/ml of bovine serum albumin. Cells (2 ϫ 10 5 /ml) were placed on 24-well tissue culture plates (Corning, Inc., Corning, NY) and preincubated at 37°C for 30 min in the presence of 0.25 mM 3-isobutyl-1-methyl xanthine (Sigma Chemical Co.) before treatment with or without hormones for 16 h.
Total cAMP in each well was measured in triplicate by a specific radioimmunoassay as previously described (25). All experiments were repeated at least four times using cells from independent transfections.
Ligand Binding Analysis-Human CG and human FSH were iodinated by the lactoperoxidase method (26) and characterized by radioligand receptor assays using recombinant human LH or FSH receptors expressed in 293T cells. Specific activity and maximal binding of the labeled hCG were 150,000 cpm/ng and 30%, respectively. These values for labeled FSH were 220,000 cpm/ng and 15%, respectively. To estimate ligand binding to the cell surface, transfected cells were washed twice with D-PBS and collected in D-PBS before centrifugation at 400 x g for 5 min. Pellets were resuspended in D-PBS containing 1 mg/ml bovine serum albumin (binding assay buffer). Resuspended cells (2 ϫ 10 5 /tube) were incubated with increasing doses of 125 I-hCG or 125 I-FSH at room temperature for 22 h in the presence or absence of unlabeled hCG (Pregnyl, Organon, 100 IU/tube) or FSH (Pergonal, Organon, 50 IU/tube), respectively. At the end of the incubation, cells were centrifuged and washed twice. Radioactivity in the pellets was determined in a ␥-spectrometer. Data from saturation binding studies were used to derive equilibrium dissociation constant (K d ) values based on Scatchard plot analysis.
Determination of Epitope-tagged Receptors on the Cell Surface-Transfected 293T cells were resuspended and incubated with FLAG M1 antibodies (50 g/ml) in Tris-buffered saline (pH 7.4) containing 5 mg/ml bovine serum albumin and 2 mM CaCl 2 (assay buffer) for 4 h at room temperature in siliconized centrifuge tubes. Cells were then washed twice with 1 ml of assay buffer following centrifugation at 14,000 x g for 15 s. The I 125 -labeled second antibody (anti-mouse IgG from sheep: ϳ400,000 cpm) was added to the resuspended cell pellet and incubated for 1 h at room temperature. Cells were then washed twice with 1 ml of assay buffer by repeated centrifugation before determination of radioactivity in the pellets using a ␥-spectrometer. Background binding was determined by adding an excess amount of the synthetic FLAG peptide at a concentration of 100 g/ml.

Chimeric Receptors with the Ectodomain from Human LH or FSH Receptors and the TM Region from Fly
LGR2 Showed Constitutive Activity-Although both wild type human LH receptor and fly LGR2 showed minimal basal cAMP production when overexpressed in human 293T cells, the chimeric receptor LDR, with the ectodomain from the LH receptor and the TM region from the Drosophila melanogaster LGR2, showed major increases in basal cAMP production ( Fig. 1). Due to variations in receptor expression levels, all cAMP data for this and subsequent experiments were normalized based on cell surface M1 antibody binding of the tagged FLAG sequence (Table I). For comparison, cAMP data without normalization were also presented (Table II). Of interest, the chimeric LDR showed high affinity (K d : 34 pM, Table I) to labeled hCG and responded to hCG treatment with further increases in cAMP production (Fig. 1). These findings are consistent with the hypothesis that the ectodomain of wild type fly LGR2 constrains its own TM region and substitution with the incompatible ectodomain of the human LH receptor leads to constitutive activation of the chimeric receptor. In addition, the results indicate that fly LGR2 is capable of interacting with the Gs protein from mammalian cells.

FIG. 1. The chimeric receptor LDR with the ectodomain from the human LH receptor and the TM region and C-tail from Drosophila melanogaster
LGR2 showed constitutive activity: further stimulation following hCG treatment. Human 293T cells were transfected with expression plasmids encoding the wild type human LH receptor (LHR WT), the fly LGR2, or the chimeric human/fly LDR. Basal and hCG-stimulated cAMP production of the chimeric receptor LDR was compared with that mediated by the wild type human LHR or the fly LGR2. Due to differences in receptor expression levels, all data were normalized based on cell surface M1 antibody binding as shown in Table I. Lower panel, diagrammatic representation of different receptors.
We further tested whether a chimeric receptor FDR with the ectodomain from the human FSH receptor and the TM region from the fly LGR2 could also confer constitutive activity. As shown in Fig. 2A, the chimeric FDR also showed major increases in basal cAMP production. In addition, treatment with FSH further increased cAMP production. This chimeric receptor exhibited high affinity to labeled FSH, comparable with that of the wild type FSH receptor based on Scatchard plot analysis (Fig. 2B). The cell surface expression of FDR, estimated using M1 antibody binding, was 59 Ϯ 1% of that of the wild type FSH receptor.
Substitution of Exoloop 2 from the LH Receptor in the Chimeric LDR Reduced Basal Activity but Not Agonist Activation-Based on the hypothesis that the ectodomain of the LH receptor could contact one or more of the exoloops in the TM region to constrain the receptor in an inactive conformation, we further substituted individual exoloops of the chimeric LDR with the LH receptor sequence. As shown in Fig. 3, the length of the three exoloops is conserved between the human LH receptor and fly LGR2. Among these residues, 36 (9/25), 33 (7/21), and 33 (4/12)% are conserved in exoloops 1, 2, and 3, respectively. Although all three mutants with individual ex-oloop substitution were expressed on the cell surface (Table I  and Fig. 4A), two mutants, LDR(EL-1) and LDR(EL-2), showed high affinity binding to labeled hCG based on Scatchard plot analysis (Table I and Fig. 4C), but the LDR(EL-3) mutant showed negligible ligand binding, likely the result of an abnormal conformation.
Of interest, the constitutively active LDR could be silenced when exoloop 2, but not exoloop 1, was replaced with the LH receptor sequence (Fig. 4B). In addition, treatment with hCG restored cAMP production mediated by LDR(EL-2) to levels comparable with those of LDR-expressing cells. These data suggest that an attraction between exoloop 2 of the LH receptor and its homologous ectodomain in the chimeric LDR(EL-2) mutant constrains the receptor in an inactive conformation. Ligand binding to the ectodomain disrupts this interaction, resulting in receptor activation. In addition, treatment of LDR(EL-2) with increasing doses of hCG indicated a dose-dependent stimulation of this chimeric receptor with an ED 50 of 1.7 ng/ml (Fig. 4D), a value higher than that for the wild type LH receptor (ED 50 : 0.25 ng/ml). Although LDR(EL-3) also showed low basal cAMP production, its inability to respond to hCG stimulation is consistent with the observed defective ligand binding of this chimeric receptor (Table I).
To further confirm the importance of exoloop 2, but not exoloops 1 and 3, in mediating the interactions between the ectodomain and TM region of the chimeric LDR receptor, we substituted two individual exoloops of the chimeric receptor (Fig. 4A). As shown in Table I and Fig. 4C, LDR(EL-1,2) and LDR(EL-1,3) were expressed on the cell surface and exhibited high affinity ligand binding. In contrast, LDR(EL-2,3) showed negligible ligand binding despite adequate cell surface expression. Furthermore, LDR(EL-1,2) showed lower basal cAMP production as compared with the chimeric LGR, whereas LDR(EL-1,3) still exhibited constitutive activity (Fig. 4B), confirming the important role of exoloop 2 in receptor constraint. Of interest, treatment with hCG stimulated cAMP production by LDR(EL-1,2) to levels comparable with those induced in the wild type LH receptor following ligand stimulation. Although LDR(EL-1,3) remained responsive to hCG treatment with a slight increase in cAMP production, the ligand binding of this receptor was partially defective. As shown in Table I, LDR(EL-1,3) exhibited lower maximal binding ability despite high cell surface expression. In contrast, the ligand signaling mechanism was defective for LDR(EL-2,3), suggesting the expression of a dysfunctional protein (Fig. 4B).

Studies Using Mutant Fly LGR2 Revealed Interactions between the Exoloop 2 and the Hinge Region of the LH Receptor-
Earlier studies suggested the important role of several residues in the hinge region (Fig. 5A, asterisks) of the human LH receptor for the conformational constraint of this protein (18). Single point mutations of serine 277 and proline 276 of the LH receptor were associated with ligand-independent activation of the receptor. In addition, photoaffinity labeling and affinity crosslinking of peptide fragments of the LH receptor suggested a potential interaction between the hinge region and the exoloop 2 of this protein (20). Taking advantage of similar structural arrangements of fly LGR2 and the human LH receptor, we replaced key residues in the hinge region of fly LGR2 with those of the LH receptor (from SYAY to TYPS, Fig. 5A) and expressed the mutant receptor in 293T cells. After correcting for cell surface expression levels based on M1 antibody binding of tagged sequences (Fig. 5B, parentheses), basal cAMP production was determined for wild type and mutant receptors. As shown in Fig. 5B, the mutant fly LGR2 (TYPS) showed constitutive activity suggesting a loss of receptor constraint. We hypothesized that the observed receptor constraint could be   the gain-or loss-of-function of different mutants. The present results are consistent with cell-free studies in which synthetic peptides corresponding to exoloop 2 were found to compete for the binding between labeled hCG and a stretch of amino acids corresponding to the hinge region (20).
There are at least three different steps involved in the ligand signaling of the LH and related glycoprotein hormone receptors, each probably requiring unique but overlapping domains of the receptor. First, the heterodimeric ligands (LH and hCG) interact with the leucine-rich repeats in the ectodomain of the LH receptor. Based on structural modeling, the seven to nine leucine-rich repeats of the LH receptor are believed to form a 1/3 donut structure important for interaction with its large ligands. Recent studies indicated that leucine-rich repeats 2 and 4 are crucial for hormone binding (27,28). Second, ligand binding leads to the disruption of the constraint on the TM region exerted by the interactions between the ectodomain (likely the hinge region) and exoloop 2. Indeed, point mutations of key residues in the hinge region of all three glycoprotein hormone receptors led to constitutive activation of these receptors (15)(16)(17)(18). Furthermore, an earlier affinity labeling study (20) and the present findings provide evidence for direct interactions between the hinge region and exoloop 2. Third, the relaxed TM region, as the result of ligand binding, interacts with the G s protein to activate the adenyl cyclase enzyme.
Analysis of chimeric human/fly receptors provided a unique opportunity to separate the first and second steps in LH receptor ligand signaling. Although the chimeric LDR showed optimal ligand binding and constitutive activity, this receptor was constrained following introduction of exoloop 2, but not exoloop 1, of the LH receptor. Of interest, the constrained LDR(EL-2) could be activated following ligand stimulation, indicating that the ligand signaling mechanism remains intact in this mutant receptor containing minimal LH receptor sequence in the TM region. A recent study further indicated that the N-terminal region of leucine-rich repeat 4 in the ectodomain is responsible for interaction with hCG, whereas the C-terminal region of this repeat is important for signal generation by binding the ex-FIG. 4. Replacement of one or two exoloops of the chimeric human/fly LDR with the LH receptor sequence indicated that exoloop 2 of the human LH receptor could mediate the ability of the ectodomain to constrain the TM region. A, diagrammatic drawing of different mutant receptor constructs. B, human 293T cells were transfected with expression plasmids encoding the wild type human LH receptor (LHR WT), LDR, or LDR with further substitution of one or two exoloops of the LH receptor. Basal and hCG-stimulated cAMP production mediated by different receptors was shown. All data were normalized based on cell surface M1 antibody binding indicated in Table I. C, Scatchard plot analyses of the binding of labeled hCG to wild type and different mutant receptors. D, dose-dependent stimulation of cAMP production induced by hCG in cells expressing wild type or the LDR(EL-2) mutant receptor. All data were normalized based on cell surface M1 antibody binding. oloop 2 (29). The exact relationship among exoloop 2, leucinerich repeat 4, and the hinge region awaits structural analysis following future crystallization of the receptor.
Although alanine-scanning mutagenesis of exoloop 2 in the LH receptor indicated that ligand binding affinity was enhanced by some alanine substitution (30), similar mutagenesis of individual residues in this region did not lead to constitutive activation of the receptor (31). Based on the screening of a comprehensive series of overlapping synthetic peptides, several synthetic peptides corresponding to sequences surrounding the hinge region competed for the binding of radiolabeled hCG to rat ovarian LH receptors (32). However, loop 3, but not loop 2, showed activity in this assay, probably due to the requirement of extremely high concentrations of these peptides in the radioligand receptor assay. Likewise, our attempts to activate the wild type LH receptors following incubation with synthetic peptides corresponding to either exoloop 2 or the hinge region of the LH receptor were unsuccessful (data not shown).
Although the chimeric receptor approach provides support for the role of the exoloop 2 in ligand signaling mediated by the LH receptor, the role of exoloop 3 cannot be ruled out. Earlier mutagenesis studies (33,34) have indeed suggested an important role of this region in LH receptor activation. In the present study, the mutant LDR(EL-3) and LDR(EL-2,3) showed low basal cAMP production and normal or adequate cell surface expression. However, ligand binding and activation of these receptors were defective, suggesting mis-folding of these mutants. Because a mutant LH receptor with only transmembrane loop 1 is capable of ligand binding (35), the observed decreases in hCG binding in LGR(EL-3) and LDR(EL-2,3) suggested the involvement of exoloop 3 in the ligand binding step.
Although our earlier data suggested that fly LGR1, but not fly LGR2, exhibited constitutive activity when overexpressed in mammalian cells (7), the present results indicate that fly LGR2 is also likely to signal through the Gs protein, and the G protein-signaling mechanism could be established between the two fly receptors and G s proteins expressed in human cells used for transfection. The observed ligand signaling of different chimeric receptor mutants consisting of human LH receptor and fly LGR2 also suggested that fly LGR2 likely belongs to the same subgroup of LGRs together with glycoprotein hormone receptors and fly LGR1 (8). Despite the lack of knowledge on the endogenous ligand(s) for the fly LGR2, understanding of the signaling pathway for this orphan receptor could facilitate future identification of its ligands using bioinformatic and biochemical approaches (36).
In conclusion, the present chimeric receptor analyses, together with earlier mutagenesis and cross-linking studies, provide a unified model of ligand signaling for the LH receptor. In this model, the hinge region and exoloop 2 of the receptor interact to constrain the unliganded transmembrane protein in an inactive conformation. Following ligand (hCG) occupancy of the leucine-rich repeats (2 and 4) in the ectodomain of the receptor (28), the ␣ subunit of hCG displaces the hinge region from exoloop 2 (20), leading to the relaxation of the TM region. With or without the additional stabilizing effects of the agonist on the TM region, ligand-induced conformation changes in this region allows subsequent activation of the G s protein. Thus, the FIG. 5. Replacement of key residues in the hinge region of fly LGR2 by the human LH receptor sequence led to constitutive receptor activation: decreases in basal cAMP production following replacement of both the hinge region and exoloop 2 of the fly LGR2 with the human LH receptor sequences. A, sequence comparison between the hinge region of fly LGR2 and the human LH receptor. The asterisks indicate two key residues previously found to be essential for constraining the human LH receptor. Conserved residues are shaded. B, human 293T cells were transfected with expression plasmids encoding the wild type human LH receptor (LHR WT), or fly LGR2 with or without substitution of the hinge region by the LH receptor sequence (SYAY to TYPS). The double fly mutant, fly LGR2 (TYPS)(EL-2), has both hinge and exoloop 2 from the LH receptor. Basal and hCG-stimulated cAMP production mediated by different receptors is shown. All data were normalized based on cell surface M1 antibody binding as indicated in parentheses with the LHR WT arbitrarily set as 100%. Lower panel, diagrammatic representation of different receptors.
␤ subunits of different glycoprotein hormones provide specificity in their binding to the receptor-specific leucine-rich repeats whereas the common ␣ subunit is important in the displacement of the hinge-exoloop 2 interactions. These findings provide the basis for future structural approaches to understand the detailed mechanism of ligand signaling mediated by glycoprotein hormone receptors and other GPCRs with a large ectodomain.