Modeling and mutagenesis of the binding site of Calhex 231, a novel negative allosteric modulator of the extracellular Ca(2+)-sensing receptor.

A model of the Ca2+-sensing receptor (CaSR) seven transmembrane domains was constructed based on the crystal structure of bovine rhodopsin. This model was used for docking (1S,2S,1'R)-N1-(4-chlorobenzoyl)-N2-[1-(1-naphthyl)ethyl]-1,2-diaminocyclohexane (Calhex 231), a novel potent negative allosteric modulator that blocks (IC50 = 0.39 microm) increases in [3H]inositol phosphates elicited by activating the human wild-type CaSR transiently expressed in HEK293 cells. In this model, Glu-8377.39 plays a pivotal role in anchoring the two nitrogen atoms of Calhex 231 and locating the aromatic moieties in two adjacent hydrophobic pockets delineated by transmembrane domains 3, 5, and 6 and transmembrane domains 1, 2, 3, and 7, respectively. To demonstrate its validity, we have mutated selected residues and analyzed the biochemical and pharmacological properties of the mutant receptors transfected in HEK293 cells. Two receptor mutations, F684A3.32 and E837A7.39, caused a loss of the ability of Calhex 231 to inhibit Ca2+-induced accumulation of [3H]inositol phosphates. Three other mutations, F688A3.36, W818A6.48, and I841A7.43, produced a marked increase in the IC50 of Calhex 231 for the Ca2+ response, whereas L776A5.42 and F821A6.51 led to a decrease in the IC50. Our data validate the proposed model for the allosteric interaction of Calhex 231 with the seven transmembrane domains of the CaSR. Interestingly, the residues at the same positions have been shown to delimit the antagonist-binding cavity of many diverse G-protein-coupled receptors. This study furthermore suggests that the crystal structure of bovine rhodopsin exhibits sufficient mimicry to the ground state of a very divergent class 3 receptor to predict the interaction of antagonists with the heptahelical bundle of diverse G-protein-coupled receptors.

The extracellular Ca 2ϩ -sensing receptor (CaSR) 1 plays an essential role in the regulation of Ca 2ϩ homeostasis. Located at the cell surface of the parathyroid cell, the CaSR is stimulated by serum Ca 2ϩ and controls parathyroid hormone release (1). Initially cloned from bovine parathyroid (2), the CaSR has been isolated from various species and tissues (3)(4)(5)(6). CaSR activation results in calcitonin secretion in the thyroid and Ca 2ϩ reabsorption in the kidney. The CaSR on nerve terminals may regulate neurotransmitter release (3,7), and its presence on oligodendrocyte cells suggests that it participates in the complex processes of myelination (8,9). Its physiological importance is further illustrated in several disorders linked to Ca 2ϩ homeostasis resulting in gain-or loss-of-function mutations (10).
The CaSR belongs to G-protein-coupled receptor (GPCR) class 3, which comprises eight metabotropic glutamate receptors and ␥-aminobutyric acid type B, vomeronasal, pheromone, and taste receptors. These GPCRs possess an unusual long bilobed amino-terminal extracellular domain resembling bacterial periplasmic binding protein implicated in nutrient transport and postulated to contain the ligand-binding sites of these receptors (11,12). The CaSR is activated by Ca 2ϩ and Mg 2ϩ present in the extracellular fluids and by charged molecules, including spermine, spermidine, ␤-amyloid peptides, and several antibiotics (2,4,8,(13)(14)(15). Recently, low molecular mass synthetic molecules activating the CaSR have been identified, and their pharmacological properties with respect to cloned CaSR have been reported (16 -19). It has been proposed that these molecules, named calcimimetics, interact allosterically within the seven transmembrane domains to potentiate the effect of Ca 2ϩ (20 -22). On the other hand, compounds that inhibit the effect of Ca 2ϩ on the CaSR are called calcilytics (23). Controlling transient parathyroid hormone release by blocking the parathyroid CaSR with such molecules has been hypothesized to produce anabolic effects in bone and represents a major therapeutic interest in the treatment of osteoporosis (23). Moreover, such compounds might be useful for studying the roles played by the CaSR in tissues under physiological and pathological states. NPS 2143 was the first negative allosteric modulator acting on the CaSR whose properties have been investigated both in vitro and in vivo (24,25). We have recently synthesized and evaluated the in vitro pharmacological properties of a novel structurally different series of calcilytics acting on the cloned rat CaSR (26). We now report the calcilytic properties of (1S,2S,1ЈR)-N 1 -(4-chlorobenzoyl)-N 2 -[1-(1-naphthyl)ethyl]-1,2-diaminocyclohexane (Calhex 231) (Fig. 1), * 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. which belongs to this family of molecules, and the characterization of its potency toward the human CaSR transiently expressed in HEK293 cells. Calhex 231 shows in vitro potency comparable to NPS 2143 in inhibiting Ca 2ϩ -induced activation of the human CaSR (25). We have developed a three-dimensional model of the transmembrane domains of the human CaSR based on the crystal structure of bovine rhodopsin (27). This model has allowed us to dock Calhex 231 into a hydrophobic cavity centered on Glu-837 (position 7.39 according to the numbering of Ballesteros et al. (28)) with two adjacent hydrophobic pockets. We used site-directed mutagenesis of amino acid residues likely involved in the recognition of Calhex 231 to demonstrate the validity of this model and to propose a possible binding mode of this negative allosteric CaSR modulator within the seven transmembrane domains.
Site-directed Mutagenesis-To mutate amino acids possibly involved in the binding site of Calhex 231, the coding region of the human wild-type (WT) CaSR, kindly provided by Prof. M. Freichel (6), was first cloned in the HindIII-XbaI sites in a modified pUC18 plasmid in which SacI and SmaI restriction sites were removed (pUCmCaSR). A SacI-BamHI insert encompassing the coding region corresponding to TM2-7 was then cloned into the pBluescript SK ϩ plasmid (Stratagene), and site-directed mutagenesis was performed. A SacI-SmaI fragment containing the mutation was then replaced in the pUCmCaSR plasmid to obtain the final mutant CaSR. Finally, WT and mutant CaSR coding regions were subcloned into HindIII-XbaI sites of the pcDNA3 expression vector (Invitrogen). All point mutants were constructed using the QuikChange site-directed mutagenesis kit (Stratagene) with specific oligonucleotides (Eurobio, Les Ulis, France) to convert residues to alanine. 2 Sequencing was performed on both strands in pBluescript SK ϩ containing the mutated fragment and in the final pcDNA3 vector (Eurogentec, Ivoz-Ramet, Belgium).
Cell Culture and Transfection-HEK293 cells (Eurobio) were cultured in Dulbecco's modified Eagle's medium supplemented with 10% dialyzed fetal calf serum (Invitrogen) and were transiently transfected using a Gene-Pulser apparatus (Bio-Rad) by electroporation (270 V and 975 microfarads). Briefly, 4 g of pcDNA3 plasmid containing human WT or mutant CaSR DNA were supplemented with 6 g of pRK5 plasmid and used to transfect 10 6 cells in a total volume of 300 l of electroporation buffer ( (6) was amplified by PCR, inserted into a glutathione S-transferase (GST) fusion pGEX-4T-1 vector (Amersham Biosciences), and sequenced (Eurogentec). The production of a GST fusion protein in Escherichia coli BL21 was then carried out (29), and 100 g were injected into a rabbit to generate 141Ab antiserum.
Alignment of Amino Acid Sequences-The amino acid sequences of four human GPCRs were retrieved from the Swiss Protein Database (Swiss-Prot accession numbers P41180 (CaSR), Q13255 (metabotropic glutamate receptor-1), Q9UBS5 (␥-aminobutyric acid receptor subunit 1), and P07550 (␤ 2 -adrenoreceptor)). Their transmembrane domains were extracted and aligned with those of bovine rhodopsin (accession number P02699) using the in-house developed GPCRmod program. 3 A ClustalW multiple alignment (32) was then performed on extra-and intracellular loops while maintaining the relative alignment of the seven transmembrane domains fixed. A slow pairwise alignment using the BLOSUM matrix series (33) and a gap opening penalty of 15.0 were chosen for aligning the amino acid sequences with the sequence of bovine rhodopsin.
Preparation of Starting Protein Coordinates-The three-dimensional model of the human CaSR was constructed using a previously described procedure (34). Briefly, starting from the x-ray structure of bovine rhodopsin (Protein Data Bank code 1f88), a first model of the seven transmembrane domains was obtained by mutating the side chains of the amino acids in rhodopsin. Standard geometries for the mutated side chains were given by the BIOPOLYMER module of SYBYL (Tripos Associates, Inc, St. Louis, MO). Whenever possible, the side chain torsional angles were kept to the values occurring in bovine rhodopsin. Otherwise, a short scanning of side chain angles was performed to remove steric clashes between the mutated side chain and the other amino acids. The third intracellular loop between TM5 and TM6, which shows a high degree of variability, was not included in any of the models. This loop is believed for most GPCRs to be far away from the TM-binding cavity (35). We therefore assume that omitting this loop should not influence our docking results. The observed insertions/deletions in the loops of the CaSR were achieved through a simple knowledge-based loop search procedure as previously reported (34). Special caution had to be given to extracellular loop 2 (ECL2), which has been described in bovine rhodopsin to fold back over the heptahelical bundle (27) and therefore limits the size of the active site. Two models of this loop were proposed. The first one assumes a rhodopsin-dependent folding and was obtained by direct threading to the rhodopsin template. In this model, one residue was inserted eight positions before Cys-765, and nine amino acids were deleted after His-766. The second model assumes a rhodopsin-independent fold and was obtained by searching for two loops of 12 and 11 residues linking Ser-750 to Thr-764 and His-766 to Ser-769, respectively. After the heavy atoms were modeled, all hydrogen atoms were added, and the protein coordinates were then minimized with AMBER6 using the AMBER95 force field as previously described (34).
Modeling the Calhex 231-bound CaSR-To obtain a ligand-bound model of the CaSR, the above-described coordinates were refined as previously described (34) to enlarge the binding cavity. Briefly, Calhex 231 was first manually docked into the TM cavity by anchoring its protonable nitrogen atom to the only negatively charged residue of the binding site (Glu-837) and fitting the shape of the two aromatic moieties into the two proximal hydrophobic pockets. The bulky naphthalene group was docked into the bigger of the two pockets (pocket A: Pro-682, Phe-688, Val-689, Tyr-744, Pro-748, Leu-776, Trp-818, Phe-821, and Ala-840), whereas the substituted phenyl moiety was located into the smaller of the two pockets (pocket B: Phe-612, Ala-615, Leu-616, Ser-665, Phe-668, Ile-669, Phe-684, Val-838, and Ile-841). After parameterization of Calhex 231 for the AMBER6 force field using a previously reported protocol (36), the resulting protein-ligand complex was then refined by minimization using the above-described AMBER parameters. Removing the ligand atoms from the minimized complex finally yielded one set of coordinates for the Calhex 231-bound receptor model.
Automated Docking of Calhex 231-To verify that the above-described coordinates were not biased by the manual docking procedure, the Surflex docking program (37) was used to automatically dock Calhex 231. An idealized active-site ligand or protomol (38) was first generated from 33 consensus positions (34) supposed to map the TM cavity of most GPCRs. This protomol consists of the preferred locations of various molecular probes (CH 4 , CϭO, N-H), which are then used by the docking engine to search for the best three-dimensional morphological similarity between the protomol and the ligand to dock. A proto_ thresh value of 0.5 and a proto_bloat value of 0 were used to generate a compact protomol. A Tripos mol2file of Calhex 231, obtained from a two-dimensional sketch as previously reported (34), was docked into the TM cavity using standard parameters of Surflex used in the "whole" docking approach (37). The best 30 solutions were finally stored in mol2 format.

RESULTS
Potency of Calhex 231-In a recent preliminary report, we described the synthesis and characterization of a novel series of molecules displaying calcilytic properties toward the rat CaSR (26). We have now synthesized Calhex 231, which belongs to this family of molecules, and we have investigated its potency toward the human CaSR by measuring its effects on Ca 2ϩinduced accumulation of [ 3 H]IP, a well characterized response linked to CaSR activation (2,13,39). Increasing the concentration of extracellular Ca 2ϩ from 0.3 to 10 mM caused a 10-fold increase in [ 3 H]IP accumulation in HEK293 cells transiently transfected with a plasmid containing the human WT CaSR, whereas we did not detect a significant increase in [ 3 H]IP accumulation in cells transiently transfected with an empty control plasmid ( Fig. 2A) (data not shown). Analysis of the dose-response curve led to an EC 50 for Ca 2ϩ of 3.4 Ϯ 0.1 mM (mean Ϯ S.E., n ϭ 10). These data fit well with the affinity for Ca 2ϩ -mediated increases in IP accumulation previously determined for the human CaSR (39). Preincubation of HEK293 cells expressing the human WT CaSR with Calhex 231 caused a concentration-dependent inhibition of the IP response to 10 mM Ca 2ϩ (Fig. 2B). Analysis of the dose-response curve led to an IC 50 for Calhex 231 of 0.39 Ϯ 0.08 M (mean Ϯ S.E., n ϭ 7). These data indicate that Calhex 231 is a potent calcilytic of the human CaSR transiently expressed in HEK293 cells.
Molecular Modeling of the Seven Transmembrane Domains of the Human CaSR-To elucidate the binding mode of Calhex 231, we postulated that the ligand-binding pocket could be localized within the seven transmembrane domains of the CaSR. First, we developed a model of the human CaSR based on the x-ray structure of bovine rhodopsin (27), which was used as a template to model the seven transmembrane domains of the CaSR. Using the in-house developed GPCRmod program, 3 we could unambiguously assign the positions of the seven helices using GPCR-dependent TM-specific amino acid fingerprints (Fig. 3A) and thread CaSR three-dimensional coordinates onto those of bovine rhodopsin. Two models were generated differing only in the fold of ECL2 (Fig. 3, B and C). The first one assumes a conserved folding of ECL2 over the seven-TM bundle as in bovine rhodopsin. The rationale for this first choice was the presence of two conserved cysteine residues at position 677 in TM3 and position 765 in ECL2 of the CaSR, which are also present in bovine rhodopsin, to form a disulfide bridge. The second model of ECL2 was derived independently from the rhodopsin structure, as it is questionable whether the particular fold of ECL2 in rhodopsin is a common feature of most GPCRs (Fig. 3C).
As previously reported (34), the TM cavity was enlarged to accommodate a ligand for automated docking. This procedure requires the manual positioning of Calhex 231. As there is only a single accessible negatively charged residue in the TM cavity (Glu-837 7.39 ) available to neutralize the positively charged secondary amine of the ligand, anchoring Calhex 231 into the TM cavity was straightforward (see "Experimental Procedures"). After energy refinement of the receptor-ligand complex and subsequent expansion of the binding cavity, the CaSR forms a well defined hydrophobic cavity centered on Glu-837 located in the seven transmembrane domains with the two adjacent hydrophobic pockets A and B. The bigger pocket A is delineated by hydrophobic side chains between TM3, TM5, and TM6, whereas the smaller pocket B is located between TM1, TM2, TM3, and TM7.
Hypothesized Binding Mode of Calhex 231-Automated docking of Calhex 231 with the recently described Surflex docking program (37) disclosed a preferred binding mode (Fig. 4, A  and B) in which both nitrogen atoms are H-bonded to Glu-837. The close proximity of the protonated secondary amine to the negatively charged Glu-837 side chain indicates a likely ionic interaction between both moieties. The naphthalene moiety is embedded in pocket A and interacts with neighboring hydrophobic side chains (Pro-682, Phe-688, Val-689, Tyr-744, Pro-748, Leu-776, Trp-818, and Phe-821). The p-chlorophenyl group is buried in the additional pocket B (Phe-612, Ala-615, Leu-616, Phe-668, Ile-669, Phe-684, Val-838, and Ile-841). Interestingly, the cyclohexyl scaffold is proposed to be located in a small hydrophobic niche delimited by Pro-682, Phe-684, and Gly-685. The important methyl group at the chiral carbon atom directly faces the Phe-821 aromatic ring. The proposed interaction model suggests a tight binding of Calhex 231, as 85% of it overall surface (573 of 674 Å 2 ) is buried upon binding to the TM cavity.

Generation of Point Mutations and Characterization of the Mutant CaSR-We have mutated into alanine Thr-764 and
His-766 located in ECL2 as well as seven other amino acid residues located in TM3 and TM5-7 by site-directed mutagenesis to investigate their possible interactions with Calhex 231. These mutants and the WT receptor were transiently transfected into HEK293 cells. We then analyzed their ability to respond to Ca 2ϩ by measuring [ 3 H]IP accumulation and their expression by Western blotting using a specific rabbit antiserum (141Ab) developed against the carboxyl-terminal tail of the human CaSR.
This antiserum was generated against a 330-amino acid polypeptide starting from amino acid 747 of the CaSR and fused to GST. The 141Ab antiserum was first evaluated by Western blot analysis against the human WT receptor. Under reducing conditions, two polypeptides migrating with a mobility corresponding to relative molecular masses of 150 and 130 kDa were identified in membrane preparations from cells transfected with the WT receptor, whereas these signals were absent in mock cell preparations, thereby indicating their specificity (Fig. 5). Our results are in agreement with previous carbohydrate analysis of the human receptor expressed in HEK293 cells showing that the polypeptides of higher molecular mass correspond to N-linked glycosylated receptors expressed at the cell surface, whereas the polypeptides of lower molecular mass represent intracellular mannose-modified receptors (40). A polypeptide complex migrating above 200 kDa was also identified and might correspond to intermolecular linked dimers as previously observed (40). Importantly, the expression pattern of the different mutant receptors was comparable to that of the WT receptor as accessed by immunoblotting (Fig. 5), indicating that the alanine substitution at the various positions analyzed did not abolish the expression of the CaSR.
All tested mutants responded to Ca 2ϩ (Fig. 6, A and B; and Table I). The mutants harboring the F684A or F688A mutation showed a rightward shift in sensitivity, with EC 50 ϭ 5.9 Ϯ 0.4 and 5.9 Ϯ 0.2 mM (mean Ϯ S.E., n ϭ 3; p Ͻ 0.001), respectively, whereas three mutants with the L776A, F821A, or I841A mutation showed a moderate but significant increase in Ca 2ϩ sensitivity, with EC 50 Ͻ 3 mM (n ϭ 3; p Ͻ 0.05). The other mutants studied did not differ in Ca 2ϩ sensitivity compared with the WT receptor. The maximal response to Ca 2ϩ was reduced by 2-fold for mutants harboring the F684A or F688A mutation compared with the WT receptor, and we observed only a moderate decrease in the maximal response (30% reduction) for mutants with the W818A or E837A mutation (Table I). These data demonstrate that the mutant and WT receptors are functional after transfection in HEK293 cells.

Functional Analysis of ECL2 Mutants for Calhex 231 Inhibition of Ca 2ϩ -promoted Increases in the IP Response-Because
all the mutants examined displayed a maximal IP response at 10 mM Ca 2ϩ (Fig. 6, A and B; and Table I), we chose to analyze the effects of Calhex 231 in inhibiting this maximal response for each mutant. We first transfected into HEK293 cells the mutants harboring the T764A or H766A mutation located in ECL2, and we constructed the dose-response curves for Calhex 231 inhibition of Ca 2ϩ -induced accumulation of IP. Calhex 231 dose-dependently inhibited the IP response induced by 10 mM Ca 2ϩ with a potency in the T764A (IC 50 ϭ 0.28 Ϯ 0.05 M) and H766A (IC 50 ϭ 0.64 Ϯ 0.03 M) (mean Ϯ S.E., n ϭ 3) mutant receptors similar to that in the WT receptor. These data indicate that these amino acid residues do not participate significantly in the binding of Calhex 231 to the receptor.
Functional Analysis of Calhex 231 Inhibition of Mutants Located in Transmembrane Domains-We then analyzed the effect of mutations located in TM3 and TM5-7 on Calhex 231 inhibition of Ca 2ϩ -induced increases in the IP response ( Fig. 7 and Table I). The dose-response curves of the ligand for the two mutants with the F684A or E837A mutation located in TM3 and TM7, respectively, were profoundly affected and are shown in Fig. 7 (A and B, respectively). Calhex 231 lost its ability to block the Ca 2ϩ -induced IP response in the CaSR with the point mutation F684A or E837A (Ͻ30 and Ͻ20% inhibition by 10 M Calhex 231, respectively). These data indicate that these two amino acid residues are crucial for Calhex 231 recognition. Its dose-response curve for each of three other mutants with the F688A, W818A, or I841A mutation was right-shifted, indicating a marked increase (ϳ10-fold) in the IC 50 value. It is worth noting that Phe-688 and Ile-841 are located near the two crucial residues Phe-684 and Glu-837, respectively, confirming the importance of these two regions for Calhex 231 recognition. The last two mutations studied, L776A and F821A, led to significant decreases in the IC 50 of Calhex 231 in inhibiting the Ca 2ϩ -induced IP response for the mutants (IC 50 ϭ 0.07 Ϯ 0.03 and 0.06 Ϯ 0.01 M, respectively) compared with the WT receptor (IC 50 ϭ 0.39 Ϯ 0.08 M; p Ͻ 0.01) DISCUSSION In this study, we have reported the characterization of Calhex 231, a novel negative allosteric modulator of the human CaSR. We used molecular modeling approaches, mutagenesis, and functional activity (phospholipase C) to identify for the first time residues involved in the binding pocket of a negative modulator of the CaSR. The amino-terminal domain of the CaSR is thought to contain the Ca 2ϩ -binding sites and has been submitted to extensive mutation and deletion studies, which have given insight on the mechanism of CaSR activation (41)(42)(43). However, little is known about the binding sites of positive or negative allosteric modulators of the CaSR. An amino acid residue (Glu-837) located in TM7 has been reported to interact A. Logean, submitted for publication). Residues in boldface are typical fingerprints (31) from either class 1 (␤ 2 -adrenergic receptor and bovine rhodopsin) or class 3 (CaSR, metabotropic glutamate receptor-1, and ␥-aminobutyric receptor-1) GPCRs. Boxes correspond to mutation effects described herein. The residue numbering of Ballesteros et al. (28) is indicated above the proposed sequence alignment. B, close-up of the sequence alignment of ECL2 of the human CaSR with bovine rhodopsin. Residues neighboring the conserved cysteine residue involved in a disulfide bridge with the third transmembrane domain are boxed. C, two putative models of ECL2 of the CaSR obtained either by direct threading to the rhodopsin x-ray structure (in green) or by a loop search procedure (in cyan). The SCG residues of bovine rhodopsin (in white) facing retinal (27) are displayed as sticks with the following color coding: carbon atom, white; oxygen atom, red; nitrogen atom, blue; and sulfur atom, yellow. The retinal structure and location of TM4 and TM5 are shown. Arrows indicate the path of the main chain.
with the calcimimetic NPS R-568 (21), whose pharmacological properties with respect to cloned CaSR have been previously reported (16,18). At the present time, the sites of interaction with the CaSR of NPS 2143, the first and sole calcilytic whose pharmacokinetic properties have been reported in vitro and in vivo (24,25), have not yet been described. We recently identi-fied a novel class of molecules inhibiting the effect of Ca 2ϩ on the cloned rat CaSR expressed in Chinese ovary cells (26). We have now synthesized Calhex 231, which belongs to this family of molecules, and have shown that it behaves as a potent and high affinity negative allosteric modulator of the human CaSR. Although the CaSR and rhodopsin display little amino acid identity, we have generated a three-dimensional model of the seven transmembrane domains of the CaSR that has allowed us to identify putative residues implicated in the recognition of Calhex 231. We submitted nine of these residues to mutations and found that seven of them affect the binding affinity of Calhex 231 as measured by inhibition of Ca 2ϩ -induced IP accumulation, a well characterized functional response linked to CaSR activation in these cells (39), thus confirming that Calhex 231 is a negative allosteric modulator of the CaSR.
A significant decrease in the effect of Calhex 231 was observed after mutation of Ile-841 7.43 . In our model, Ile-841 7.43 interacts with the p-chlorophenyl moiety of Calhex 231. This position has already been shown to directly contact adenosine as well as peptide receptor antagonists (Table II). Surprisingly, the mutation of two positions (Leu-776 5.42 and Phe-821 6.51 ) led to receptor mutants with significantly enhanced Calhex 231 antagonist activities. In many monoamine receptors, position 5.42 is a serine that has been shown to directly interact with the catechol group of many GPCR ligands (49). Phe-821 6.51 is involved in the conformational switch triggering of GPCR activation (48). The consequence of these particular mutations is difficult to explain at the molecular level. Both contribute to enlarge pocket A and perhaps induce slight conformational changes allowing better accommodation of the bulky naphthalene group of Calhex 231.
Altogether, the present data suggest that the bovine rhodopsin x-ray structure exhibits a significant molecular mimicry to the ground state of not only class 1 monoamine receptors (28), but also the transmembrane domains of the very divergent class 3 GPCRs, as very recently proposed for the metabotropic type 1 receptor (50). Interestingly, the central positions 6.48 (Trp-818) and 7.39 (Glu-837) also play a key role in recognizing a noncompetitive metabotropic glutamate receptor-1 antagonist (50), suggesting that the TM-binding cavities of class 3 GPCRs largely overlap.
However, some discrepancies with the bovine rhodopsin structure remain. In opposition to retinal, the two residues (Thr-764 and His-766) adjacent to a conserved cysteine (Cys-765) involved in disulfide bridging between ECL2 and TM3 (Cys-677) do not contribute to the binding site of Calhex 231. Thus, it is likely that the overall folding of ECL2 differs significantly from that of bovine rhodopsin bound to retinal.
In summary, we have, to our knowledge, identified for the first time the amino acids involved in defining the ligandbinding pocket of a negative allosteric modulator of the CaSR. Because calcilytics are proposed to represent a novel therapeutic approach for treating osteoporosis (23), it is of major interest to delineate the residues involved in their recognition. Our study should facilitate the understanding of how calcilytics interact with the CaSR and the development of molecules with increased affinity and selectivity. Moreover, Calhex 231 represents a novel calcilytic that should be highly useful for studying the role of the CaSR in tissues such as bone, kidney, and brain under physiological and pathological conditions.