Asp330 and Tyr331 in the C-terminal Cysteine-rich Region of the Luteinizing Hormone Receptor Are Key Residues in Hormone-induced Receptor Activation*

The luteinizing hormone (LH) receptor plays an essential role in male and female gonadal function. Together with the follicle-stimulating hormone (FSH) and thyroid stimulating hormone (TSH) receptors, the LH receptor forms the family of glycoprotein hormone receptors. All glycoprotein hormone receptors share a common modular topography, with an N-terminal extracellular ligand binding domain and a C-terminal seven-transmembrane transduction domain. The ligand binding domain consists of 9 leucine-rich repeats, flanked by N- and C-terminal cysteine-rich regions. Recently, crystal structures have been published of the extracellular domains of the FSH and TSH receptors. However, the C-terminal cysteine-rich region (CCR), also referred to as the “hinge region,” was not included in these structures. Both structure and function of the CCR therefore remain unknown. In this study we set out to characterize important domains within the CCR of the LH receptor. First, we mutated all cysteines and combinations of cysteines in the CCR to identify the most probable disulfide bridges. Second, we exchanged large parts of the LH receptor CCR by its FSH receptor counterparts, and characterized the mutant receptors in transiently transfected HEK 293 cells. We zoomed in on important regions by focused exchange and deletion mutagenesis followed by alanine scanning. Mutations in the CCR specifically decreased the potencies of LH and hCG, because the potency of the low molecular weight agonist Org 41841 was unaffected. Using this unbiased approach, we identified Asp330 and Tyr331 as key amino acids in LH/hCG mediated signaling.

the production of gonadal androgens, which in women are almost completely converted to estrogens. Furthermore, in women the mid-cycle LH peak triggers ovulation of the mature oocyte ready to be fertilized, whereas the closely related pregnancy hormone, chorionic gonadotropin (hCG), supports the corpus luteum of pregnancy. Both LH and hCG act by binding and activating the LH receptor. The LH receptor has a modular architecture consisting of an ectodomain or extracellular hormone binding domain (ECD), linked to a seven-transmembrane signal transduction domain (7TMD). Together with the receptors for thyroid stimulating hormone (TSH) and follicle stimulating hormone (FSH), the LH receptor belongs to the glycoprotein hormone receptor family. The extracellular ligand binding domain of this family consists of a stretch of nine leucine-rich repeats (LRRs), flanked by N-terminal and C-terminal cysteine-rich clusters (NCR and CCR, respectively). The crystal structures that have been determined for major parts of FSHR and TSHR ECDs show that the LRRs are organized as ␤ sheets, and give rise to a curved helical tube of which the concave surface forms the hormone-binding surface (1,2). The NCR provides an additional ␤ strand to the binding surface. Because the CCR was not included in the protein expression procedures to obtain the crystals, its structure remains elusive.
LRRs are considered versatile binding motifs often involving protein-protein interactions, and are observed in a large variety of proteins (3,4). Also the non-glycoprotein hormone receptor LRRs contain an NCR and CCR and in those LRR proteins of which crystal structures have been resolved, the NCR and CCR form ␤ strands that combine seamlessly with the ␤ sheets formed by the LRRs (5,6). Thus, the CCR of the glycoprotein hormone receptors may also be expected to form a prolongation of the LRR ␤ sheets.
The function of the CCR has been addressed in more detail by receptor mutagenesis studies. Ligand binding studies with mutant rat LH receptors indicate that LRR1-6 contribute most to high-affinity ligand binding, whereas a combined deletion of LRR9 and the CCR did not alter hCG affinity (7). Truncation of the rat LH receptor demonstrates that neither the 7TMD nor the CCR are essential to high affinity binding (8). A similar approach in the human LH receptor, systematically truncating the receptor at exon boundaries showed a stepwise decrease in affinity loss with severity of truncation (9). Here, deleting both 7TMD, and the C-terminal part of the CCR, did not alter hCG affinity. Notably, the above mentioned deletion and truncation mutants failed to be expressed at the plasma membrane.
In most LRR proteins the CCR has four cysteines, of which the first two are either directly adjacent or separated by one amino acid resulting in two disulfide bonds (10). Also in the CCR of the LH receptor the first two (of a total of six) cysteines are located adjacently. Because the CCR is a conserved feature, yet is not required for high-affinity ligand binding, its role may be to structurally preserve the ECD structure, and/or to selectively transduce signals from the ECD to the 7TMD. Crystal structures of the glycoprotein Ib␣ (GPIb␣) suggest a role for the CCR in responding to ligand binding: a 16-residue ␤-hairpin (␤-switch) is formed in the CCR of GPIb␣ upon binding of the von Willebrand factor (6). Noteworthy, all disease-causing gain-of-function mutations are present in this ␤-switch (11,12). Also for the TSH receptor such a concept has been proposed. Based on a series of constitutively active TSH receptors, Mueller and co-workers (13) proposed an activation mechanism in which the CCR is key. The authors propose that the CCR acts as a lever linking the ECD LRRs to the 7TMD, in which the conserved cysteine tandem of the CCR serves as a fulcrum. Modification on either site of the lever, by mutation, ligand interaction, or tryptic clipping, changes the conformation of the CCR, and results in 7TMD activation (13). Also in the LH receptor, mutation of Ser 277 close to the CCR leads to constitutive receptor activation (14), supporting a similar mechanism in this receptor as well.
Besides via the extracellular domain, the LH receptor can also be activated directly via the 7TMD. Illustrating for this phenomenon is that most of activating LH receptor mutations observed in patients are single amino acid changes within the transmembrane helices (for reviews see Refs. 15 and 16). Recently, some thienopyr(im)idine derivatives (i.e. Org 41841 and Org 42599) were characterized as a novel class of low molecular weight LH receptor agonists (17,18). These ligands can activate the receptor by interacting directly with the 7TMD, rather than with the receptor ectodomain (19,20).
In the present study we unravel the constellation of disulfide bridges in the CCR, and further set out to characterize the requirements of the CCR for hormone-induced signaling. Hereto we exchanged and deleted (parts of) the LH receptor CCR, followed by detailed alanine scanning. This led to the identification of two amino acid residues, Asp 330 and Tyr 331 , as key CCR residues, involved in LH/hCG mediated signaling. In addition we introduce the low molecular weight agonists Org 41841 and Org 42599 as tools to evaluate intrinsic receptor functionality.

EXPERIMENTAL PROCEDURES
Materials-Unless stated otherwise, cell culture reagents were purchased from Invitrogen (Paisley, UK): restriction enzymes and polymerases from Roche; fine chemicals from Sigma; oligonucleotides were from Biolegio (Nijmegen, The Netherlands). hCG, LH, Org 41841, and Org 42599 are kind gifts of NV Organon, Oss, The Netherlands.

Construction of WT and Mutant LH Receptor Expression
Plasmids-Construction of the expression plasmid pSG5-hLHR WT is described elsewhere (21). An N-terminal HA tag was inserted immediately C-terminal to the signal peptide by inserting the primer-dimer (5ЈGTACCCATACGATGTTCCAGAT-TACGC and complement) into the Eco47III site at position 84 in the LH receptor cDNA. In this background, the deletion or domain-exchange LH receptor mutants were created by a series of fusion PCRs (see supplemental materials). In short, two or three separate PCR were done to amplify different domains of the LH receptor or, if required, other receptors (pSG5-FSHR (22), pSVL-TSHR (23), pCDNA3-GP1b␣ (6), and pOTB7-Nogo66-receptor (5)). By using extended oligos in these PCRs, overlap was created between bordering PCR fragments and in a second series of PCR, this overlap was used to fuse individual fragments and amplify the fusion product. The resulting product was subcloned into the pSG5-5ЈHA-hLHR WT construct. Alanine scanning mutagenesis was done using the QuikChange Site-directed Mutagenesis kit (Stratagene, La Jolla, CA) according to the manufacturer's protocol. Oligonucleotides used for mutagenesis are depicted in supplemental materials. All constructs were verified by dideoxynucleotide sequencing.
Cell Culture and Transfection-HEK 293 cells were maintained in culture medium, Dulbecco's modified Eagle's medium/F-12 ϩ Glutamax supplemented with 1 ϫ 10 5 IU/liter penicillin, 0.1 mg/liter streptomycin, and 10% fetal bovine serum. Two days prior to transfection, cells are plated at a density of 20% in 25-cm 2 tissue culture flasks (Nalge Nunc International, Rochester, NY). For transfection, 7 g of LH WT/mutant receptor cDNA (in pSG5), 1.4 g of pCRE 6 lux, and 1.4 g of pRL-SV40 was added to 350 l of 150 mM NaCl and mixed with 350 l, 0.03 mg/liter polyethyleneimine (linear M r ϳ 25,000; Polysciences Inc., Warrington, PA) in 150 mM NaCl. Medium was replaced by serum-free medium and supplemented with a cDNA/polyethyleneimine mixture. After 4 h, medium was supplemented with fetal bovine serum to a final concentration of 10%.
Reporter Gene Analysis-To analyze LH receptor activity, WT and mutant expression constructs were transfected to HEK 293 cells (see above). 24 h post-transfection, cells were detached using trypsin/EDTA, and transferred to white Clear TC-treated plates (Greiner Bio-one Alphen a/d Rijn, The Netherlands), 48 h post-transfection, medium was replaced by serum-free culture medium supplemented with 0.1% bovine serum albumin and 25 mM HEPES and a 2-fold dilution series of LH, hCG, Org 41841, or Org 42599. For mass to mole conversion for LH and hCG, molecular masses of 29.5 and 36.7 kDa were assumed, respectively (24,25). After 6 h stimulation, wells were aspired and cells were lysed with 25 mM Tris phosphate, pH 7.8, 8 mM MgCl, 1 mM dithiothreitol, 15% glycerol, 1% Triton X-100 and analyzed for luciferase (cAMP-responsive element driven) and Renilla luciferase (transfection control) activities using the Dual-Luciferase Reporter Assay System according to the manufacturer's instructions (Promega Corporation, Madison, WI). Data were analyzed using Prism3 (Graphpad Software Inc., San Diego, CA). Only Ն3-fold changes in EC 50 (compared with WT) are considered relevant, and are tested for statistical significance by Student's t test. p values Ͻ0.05 were considered to indicate a significant difference (*, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001).
ELISA-WT and mutant expression constructs were transfected to HEK 293 cells (see above). 24 h post-transfection, cells were detached using trypsin/EDTA, and transferred to 96-well clear plates (Corning Inc., Corning, NY) that were polylysine coated for ELISA. 48 h post-transfection, cells were washed with PBS and fixed with freshly made 4% paraformaldehyde in PBS (30 min, 22°C) and blocked with 3% bovine serum albumin in PBS (1 h, 37°C). Wells were washed 4 times with PBS and incubated (1 h, 37°C) with 1:1000 dilution of Anti-HA-peroxidase, (3F10, Roche, Mannheim, Germany) in PBS, 3% bovine serum albumin. Wells were washed 6 times for 10 min with PBS and peroxidase activity was assayed using 1-Step Ultra TMB-ELISA (Pierce) according to the manufacturer's instruction.

RESULTS
Constellation of Disulfide Bridges in the CCR-Despite the apparent structural resemblance of glycoprotein hormone receptors, the overall sequence identity in the ECD is low, approximately 40% (26). However, several sequence motifs are conserved: 4 cysteines in the N-terminal part of the ECD, 9 leucine-rich repeat motifs (XXLXLXX), and 6 cysteines at the C-terminal side of the extracellular domain. These cysteines are always present in even numbers, suggesting that they are conserved as disulfide pairs. To study the importance of the 6 cysteines in the CCR we individually mutated them to alanine and assayed the mutant receptors for functionality in a CRE-driven reporter gene assay. Mutation of Cys 304 or Cys 336 to alanine does not affect the maximum response to LH (Fig. 1A), and only results in a small (3-5-fold) decrease in LH potency (supple-mental materials Fig. S1). In contrast, mutation of Cys 279 , Cys 280 , Cys 343 , or Cys 353 completely abolished LH-mediated signaling (Fig.  1A), and mutant receptors could not be detected at the plasma membrane (Fig. 1B). Remarkably, CRE signaling remained possible after stimulation of these mutant receptors with Org 42599 (Fig. 1A). Org 42599 is a low molecular weight LH receptor agonist, based on lead optimization of Org 41841 (18) and, as Org 41841 binds to and activates the receptor directly at the 7TMD (17,19,20), and does not require the ECD. It should be noted that, compared with WT, not only the maximal effect, but also the potency of Org 42599 was reduced (supplemental materials Fig. S1). We observed similar shifts in potencies and maximal effects of both Org 42599 and LH when we transfected decreasing amounts of WT LH receptor plasmid (supplemental Fig. S2).
We utilized the residual Org 42599 responsiveness of individual Cys to Ala mutant LH receptors to probe disulfide bridges in the CCR. We argued that mutation of both cysteines of the same disulfide bridge would have the same effect as mutation of only one cysteine, because in both cases only one bridge is broken. In contrast, mutating two cysteines that are not interconnected would break two disulfide bridges and would aggravate the effect of the individual mutations. To test this hypothesis we substituted the following pairs of cysteines by alanines: Cys 279 ϩ Cys 343 , Cys 279 ϩ Cys 353 , Cys 280 ϩ Cys 343 , and Cys 280 ϩ Cys 353 . As is clearly illustrated by Fig. 1C, only the combinations C279A ϩ C343A and C280A ϩ C353A responded to Org 42599, suggesting that disulfide bonds are present between cysteines 279 and 343, and between 280 and 353. A third disulfide bridge is possible between Cys 304 and Cys 336 . Both single cysteine to alanine mutants as well as the double alanine substitution C304A,C336A only result in a small (maximum 5-fold) decrease in potency of LH. From these data we cannot conclude whether a disulfide bridge is present or there is no disulfide bridge at all. Clear, however, is that the contribution of Cys 304 and Cys 336 to the overall receptor structure is limited. Assuming the third disulfide bridge between Cys 304 and Cys 336 , an overall structure results as depicted in Fig. 2.
LH/hCG-mediated Signaling Requires CCR of Glycoprotein Hormone Receptor-Because sequence homology between the glycoprotein hormone receptors is relatively low, we defined the CCR as the region between LRRs and 7TMD, ranging from the first to the last cysteine (in the LH receptor: Cys 279 to Cys 353 ). To probe the specific relevance of the LH receptor CCR for the function of the LH receptor, we deleted it, or replaced it by CCRs of other related (FSH receptor and TSH receptor) or unrelated proteins (Nogo66 receptor (NgR) and GPIb␣). The latter two are not 7TMD receptors, but do contain LRRs with flanking CCR and NCR of which crystal structures are available (5, 6). Fig. 3 shows that the introduction of non-related CCRs (NgR and GPIb␣), as well as the complete deletion of the CCR, rendered the mutant receptors completely unresponsive to LH. Analysis of cell surface expression (by ELISA) showed that these mutant receptors were not expressed at the cell surface (data not shown). In contrast, the CCR of the FSH and TSH receptors are much better tolerated as replacements of the original CCR. These exchange mutants are well expressed (not shown). Furthermore, the mutant LH receptor in which the complete CCR is replaced by that of the FSH receptor (hereafter referred to F281-352) responds to LH and hCG with unaltered potency (Fig. 3 and Table 1). Using the TSH receptor as CCR donor, the mutant receptor is also functional, but the potency of LH and the maximal effect are reduced ( Fig. 3 and Table 1).

Fragment Exchange within CCR Shows That Amino Acids 305-335
Are of Key Importance in LH/hCGmediated Activation-The six cysteines in the CCR of the receptor define four intercysteine stretches consisting of amino acids 281-303, 305-335, 337-342, and 344 -352 (Fig. 2). To unravel the function of the CCR we replaced several of these stretches of the LH receptor with the corresponding sequences of the FSH receptor and tested the mutant receptors in our CRE-driven reporter gene assay. Replacing the first intercysteine domain or a combination of domains 1, 2, and 3 by their FSH receptor counterparts (F281-303 and F281-342) did not alter the potency of LH, whereas replacement of intercysteine domains 2, 3, and 4 (F305-352) resulted in an 8-fold decrease of LH potency (Table 1). Zooming in on this region, we individually replaced intercysteine stretches 2, 3, and 4 (F305-335, F337-342, and F344 -352, respectively). Replacement of intercysteine stretch 2 (F305-335) resulted in a nearly 30-fold decrease in LH potency, whereas the other exchanges were relatively ineffective (Table  1). Using the homologous hormone hCG a similar trend is observed, although effects are generally more moderate ( Table  1). None of the intercysteine replacement mutants displayed altered receptor expression (supplemental Fig. S3).
Focused Deletion within the CCR Leads to Complete Nonresponsiveness to LH/hCG-As reported earlier (27), deletion of the exon 10-encoded part of the LH receptor (⌬290 -316), resulted in a mutant receptor for which potencies of especially LH, and to a lesser extent hCG, were reduced (15-and 4-fold, respectively, Table 2). Because replacement of the second intercysteine stretch with its FSHR counterpart (F305-335) also effectively reduced the potencies of LH and hCG (28-and 13-fold, respectively), we expected the overlap with exon 10 to be involved. However, deleting this overlap, i.e. the C-terminal 12 amino acids of exon 10 (⌬305-316), left the LH and hCG potencies unaffected (Table 2). Because the patient phenotype could not be mimicked by deleting amino acids 305-316, we also set out to delete the other amino acids encoded by exon 10 (⌬290 -303). Remarkably, also this deletion did not result in large changes of LH/hCG potencies (Table 2).
Apparently, the exon 10-encoded part of the second intercysteine stretch does not mediate the decreased potency observed for FSH receptor exchange mutant F305-335. In sharp contrast, deleting the exon 11-encoded part of this second intercysteine stretch (⌬317-335) rendered the LH receptor completely unresponsive to both hCG and LH ( Fig. 4A and Table 2) without altering its plasma membrane expression (Fig.  4C). Because the mutation may cause intrinsic incapability of the receptor to signal, we applied the low molecular weight agonist Org 41841. In contrast to the glycoprotein hormones,  Org 41841 binds directly to the transmembrane region and may therefore activate the receptor irrespective of mutations in the CCR. Indeed, despite the complete abolishment of the response to LH and hCG, the ⌬317-335 mutant receptor responded with unaltered potency to Org 41841 (Fig. 4B). Deletion of the N-terminal amino acids of intercysteine domain 2, ⌬317-324 did not result in changes in LH/hCG potencies, indicating the importance of amino acids 325-335 in LH/hCG mediated signaling.
Alanine Scanning Mutagenesis Shows That Asp 330 and Tyr 331 Are Crucial Residues in CCR-Because deletion of amino acids 317-335, but not 317-324, was deleterious to LH/hCG induced signaling, we applied alanine scanning on amino acids 325-335 to identify important residues in the 324 -335 region. Hereto, amino acids Glu 325 to Phe 335 were individually mutated to alanines. Major changes in potency for LH, but not for Org 41841 were observed for D330A and Y331A mutant LH receptors (22-and 75-fold, respectively; Fig. 5 and Table 3). For mutant receptors W329A and F335A, decreases in LH potency were smaller (3-and 4-fold, respectively). None of the other mutations resulted in a major change in LH potency. Again, for hCG similar, but milder, changes were observed, whereas potencies of Org 41841 remained unaltered for all mutant receptors (Table 3).

DISCUSSION
Recently, structures of the extracellular domain of both FSH (2) and TSH receptors (1) have been resolved. Unfortunately, the CCR was not included in these crystallization attempts and the structure of the CCRs thus remains elusive. To the best of our knowledge, attempts to provide structural clues on the structure of the CCR have proven unsuccessful. Because disulfide bridges may be critical in upholding the structure of the CCR, we set out to unravel the constellation of the six cysteines in this domain. Mutation of individual cysteines to alanines confirmed the finding from Zhang and co-workers (28), that Cys 279 , Cys 280 , Cys 343 , and Cys 353 are required for LH receptor function, whereas Cys 304 and Cys 336 are not. Receptors in which the crucial cysteines were exchanged no longer responded to LH, but were still responsive to the low molecular weight agonist Org 42599, albeit with reduced maximum effect and potency ( Fig. 1 and supplemental Fig. S1). Reductions in both potency and maximal effect were also observed after transfecting decreasing amounts of WT LH receptor expression plasmid. This suggests that mutation of any of the 4 cysteines compromises receptor structure to such an extent that expression levels are strongly reduced. The remaining receptors still appear responsive to Org 42599 but not to LH, indicative for an intact 7TMD with a disrupted ECD.
We used the residual potency of Org 42599 to probe disulfide bridge formation within the CCR and the findings are consist-

TABLE 1 LH or hCG activation of WT or CCR exchange mutants of the LH receptor
EC 50 values were determined using a CRE-driven reporter gene assay after transfection in HEK293 cells. Numbers in brackets indicate which amino acids are replaced by FSH receptor (F) of TSH receptor (T) counterparts. Data are calculated as the mean Ϯ S.E. (*, p Ͼ 0.05; **, p Ͼ 0.01; ***, p Ͼ 0.001 versus WT).

TABLE 2 LH or hCG activation of WT or partial CCR deletion mutants of the LH receptor
Ligand induced activation in HEK 293 cells transiently transfected with WT or mutant LH receptors in which part of the CCR is deleted. EC 50 values were determined using a CRE-driven reporter gene assay after transfection in HEK293 cells. Numbers in brackets indicate which amino acids of the LH receptor are deleted (⌬). Data are calculated as the mean Ϯ S.E. (*, p Ͼ 0.05; **, p Ͼ 0.01; ***, p Ͼ 0.001 versus WT). ent with disulfide bonding of Cys 279 with Cys 343 and Cys 280 with Cys 353 . Because single mutations C304A and C336A as well as the combination of them resulted only in minor changes in LH potency, we cannot discriminate between the absence of a third disulfide bridge, or the presence of a third bond that is of little structural importance. The putative third disulfide bridge would result in a cysteine constellation as schematically represented in Fig. 2. Notwithstanding this putative third disulfide bridge, the main disulfide bonding of Cys 279 with Cys 343 and Cys 280 with Cys 353 results in a knot-like structure for the CCR in which amino acid residues 280 -343 of the CCR branch off from this cluster of cysteines, facilitating close proximity between the ECD and 7TMD.
The CCR has often been described as a mere capping structure, stabilizing the LRR structure of the ECD although in some cases the CCR has a role in signaling interaction with protein ligands (6). We decided to address the potential role of the CCR as a "capping structure" to stabilize the LRR structure, by deleting the complete CCR of the LH receptor. Indeed, this mutant receptor was not expressed at the plasma membrane, indicating the importance for correct folding. We argued that if the sole function of a CCR is "capping" the LRR, a non-related CCR   instead of the native CCR might also be capable to perform this function. We substituted the CCR of the LH receptor with that of the Gp1b␣ and the NgR. Both proteins are membrane linked (via a single transmembrane domain or a GPI anchor, respectively) and their crystal structures have been resolved (5,6), revealing the classical LRR architecture, in which both NCR and CCR are a continuation of the LRR ␤-strands. If replacement with these unrelated CCRs would generate a functional LH receptor, homology modeling of the LH receptor CCR with these proteins would be feasible. Unfortunately, replacement with any of these structures did not result in rescue of the LH/hCG functionality. In contrast, replacement of the CCR with its homologous structure derived from the related FSH or TSH receptors was able to reinstate LH/hCG responsiveness indicating that the overall fold of the CCR is conserved among the glycoprotein hormone receptors.
Indeed, replacement of the complete LH receptor CCR by the CCR of the FSH receptor was accepted without a change in receptor phenotype. Interestingly, partial replacement resulted in a full spectrum of effects: LH potencies were left unchanged (F281-303; F337-342), decreased (F305-352; F305-335), or even (marginally) increased (F281-342; F344 -352). Apparently, there is cooperativity within the CCR that is dependent on the presence of interacting motifs that probably have coevolved. Because F305-352 and especially F305-335 resulted in strong decreases in LH potencies, we assumed region 305-335 harbors important LH receptor residues.
Part of this 305-335 region is encoded by exon 10. Gromoll and co-workers (29) have described a patient in which exon 10 of the LH receptor was genomically deleted, resulting in a 28-amino acid deletion within the CCR of the LH receptor (⌬290 -317). This patient was virilized, as a result of responsiveness to placental hCG, but was blocked in pubertal development, because his LH receptor did not respond to LH. In vitro characterization of the ⌬290 -317 mutant receptor verified that the patient's symptoms were represented at the receptor level: a strongly decreased LH potency, and more marginally affected hCG potency (Table 2 and Ref. 27). However, deletion of the amino acids common to exon 10 and 305-335, nor the other part of exon 10 had a strong effect on LH or hCG potencies. Apparently, not the specific amino acids encoded by exon 10, but other characteristics, such as the overall length, of this fragment are important. Possibly, the exon 10-encoded part of the receptor is required to create the topology in which critical residues in the CCR can optimally interact with other elements of the receptor or with hormones LH and hCG themselves.
In contrast to exon 10, deletion of the exon 11-encoded amino acids of 305-335 and ⌬317-335 did decrease LH/hCG potency, even to a level at which no LH/hCG-mediated activation could be detected. By means of alanine scanning Asp 330 and Tyr 331 were identified as key residues within the CCR. Interestingly, using a different approach, Costagliaola and colleagues (30) identified Tyr 385 of the TSHR, analogous to the LH receptor Tyr 331 as a key residue. These authors demonstrated that Tyr 385 of the TSHR was sulfated and that the introduction of this negative charge was essential for TSH signal transduction (30). Sulfation is also suggested for the analogous Tyr res-idues in the LH and FSH receptor (30,31), the mechanistic role of sulfation in signal transduction remains unknown.
Throughout this study, we observed that changes in potency of LH coincided with changes in potency of hCG, although the latter were always milder. The hormones share a common ␣-subunit, and also LH␤ and hCG␤ are highly similar. hCG␤ is mainly distinguished by the presence of a highly glycosylated 30-amino acid extension called the C-terminal peptide (Ref. 32). This C-terminal peptide therefore likely mediates the difference in CCR mutation sensitivity. In the absence of a crystal structure of the LH receptor encompassing both CCR and hCG, the mode of interaction between CCR and C-terminal peptide remains speculative. It is, however, clear that when studying the effects of mutations of the LH receptor, LH should be included as a ligand of choice, and not just the easier available hCG.
Our data suggest that the CCR plays a pivotal role in LH/hCG-mediated signaling. Because high-affinity LH/hCG binding is thought to reside within the LRRs of the ECD, the CCR must act at a later stage of receptor activation. Generally two phenomena could account for this later stage in activation: either the CCR interacts with LH/hCG via low affinity interaction and "guides" the hormone to interact with the 7TMD, as recently suggested by Rapoport and co-workers (33), or the CCR itself acts as the true agonist. The latter option implies that the LRR changes conformation upon LH/hCG binding; this conformational change in turn causes the CCR, and likely Asp 330 and Tyr 331 , to establish contacts with the 7TMD and that this finally leads to activation. Such a phenomenon is well accepted in protease-activated G protein-coupled receptors. In these receptors a cryptic ligand, present in the receptor, is unmasked upon protease cleavage of the receptor (34). Also for class B GPCRs, an "agonist within" has been proposed that is unmasked upon ligand binding to the extracellular domain (35). The CCR of the TSH receptor is proposed to act as an inverse agonist until ligand binding, upon which it switches to an agonist (26,36). For the LH receptor, our results exclude that the CCR keeps the LH receptor inactive: mutation of the key amino acids, Asp 330 and Tyr 331 , would remove such an inhibition, yet no increase in constitutive receptor activation is observed.
Recently a new class of thienopyrimidine derivatives have been synthesized that act as selective LH receptor agonists (17). The prototype of this class is Org 41841, and acts as a submicromolar potency agonist of the LH receptor (Table 2 and Ref. 17). Based on this prototype, also the more potent Org 42599 was developed (18). In contrast to glycoprotein hormones hCG and LH, Org 41841 does not bind to the ECD of the receptor, but directly to the 7TMD (17,19,20). In this study we exploited this feature to study the functional intactness of the mutant LH receptor proteins, i.e. do the expressed proteins survive the folding quality control system and are they able to signal to the intracellular transduction systems? Thus, the ⌬317-335 mutant LH receptor is expressed at the plasma membrane (as determined by cell surface ELISA), yet does not respond to LH and hCG. Org 41841 provided a valuable tool to demonstrate that the absence of effect of LH or hCG did not result from receptor misfolding, because Org 41841 activates this mutant LH receptor with the same potency and effect as the WT recep-tor. Therefore it can be concluded that the ⌬317-335 mutant LH receptor is intrinsically capable of signaling, and thus correctly folded.
In summary, we unraveled the requirements of the CCR for hormone-induced signaling by means of rational mutagenesis. By exchanging part of the CCR for its FSH receptor counterpart, we characterized amino acids 305-335 as an important region within the CCR. Subsequent deletion mutagenesis within this domain allowed us to narrow this region down to amino acids 317-335, because deletion of these amino acids completely eliminated LH/hCG mediated signaling. Noteworthy, signaling induced by the low molecular weight agonist Org 41841 was unaffected, establishing Org 41841 as a valuable tool to evaluate intrinsic receptor functionality. Alanine scanning finally demonstrated that within the CCR, Asp 330 and Tyr 331 are key residues for LH/hCG-mediated signaling.