The biologically crucial C terminus of cholecystokinin and the non-peptide agonist SR-146,131 share a common binding site in the human CCK1 receptor. Evidence for a crucial role of Met-121 in the activation process.

The cholecystokinin (CCK) receptor-1 (CCK1R) is a G protein-coupled receptor, which mediates important central and peripheral cholecystokinin actions. Our aim was to progress in mapping of the CCK1R binding site by identifying residues that interact with the methionine and phenylalanine residues of the C-terminal moiety of CCK because these are crucial for its binding and biological activity, and to determine whether CCK and the selective non-peptide agonist, SR-146,131, share a common binding site. Identification of putative amino acids of the CCK1R binding site was achieved by dynamics-based docking of the ligand CCK in a refined three-dimensional model of the CCK1R using, as constraints, previous results that identified contact points between residues of CCK and CCK1R (Kennedy, K., Gigoux, V., Escrieut, C., Maigret, B., Martinez, J., Moroder, L., Frehel, D., Gully, D., Vaysse, N., and Fourmy, D. (1997) J. Biol. Chem. 272, 2920-2926 and Gigoux, V., Escrieut, C., Fehrentz, J. A., Poirot, S., Maigret, B., Moroder, L., Gully, D., Martinez, J., Vaysse, N., and Fourmy, D. (1999) J. Biol. Chem. 274, 20457-20464). By this approach, a series of residues forming connected hydrophobic clusters were identified. Pharmacological and functional analysis of mutated receptors indicated that a network of hydrophobic residues including Cys-94, Met-121, Val-125, Phe-218, Ile-329, Phe-330, Trp-326, Ile-352, Leu-356, and Tyr-360, is involved in the binding site for CCK and in the activation process of the CCK1R. Within this hydrophobic network, the physico-chemical nature of residue 121 seems to be essential for CCK1R functioning. Finally, the biological properties of mutants together with dynamic docking of SR-146,131 in the CCK1R binding site demonstrated that SR-146,131 occupies a region of CCK1R binding site which interacts with the C-terminal amidated tripeptide of CCK, i.e. Met-Asp-Phe-NH(2). These new and important insights will serve to better understand the activation process of CCK1R and to design or optimize ligands.

Cholecystokinin (CCK) 1 is a neuropeptide that has a wide spectrum of biological actions. CCK is composed of several molecular variants, the octapeptide (CCK-8: Asp-Tyr(SO 3 H)-Met-Gly-Trp-Met-Asp-Phe-NH 2 ) being the major fully active one (1,2). Two CCK receptors have been characterized pharmacologically, biologically and subsequently cloned, the CCK1 receptor (abbreviated CCK1R, previously named CCKA receptor) and the CCK2 receptor (abbreviated CCK2R, previously named CCK-B/gastrin receptor), which both belong to the superfamily of G protein-coupled receptors (3,4). The CCK1R and CCK2R can exist in several conformational states, which bind CCK with high, low, and very low affinities, respectively, and share the functional coupling to phospholipase C, via binding to heterotrimeric GTP-binding protein(s) (5)(6)(7). CCK1R-mediated effects include control of gallbladder contraction, pancreatic exocrine secretion, gastric emptying and gut motility, and satiety (8,9). The wide spectrum of biological functions regulated by the CCK1R makes it a candidate target for a therapeutic approach in a number of diseases related to nutrient assimilation. This led a number of academic and pharmaceutical research groups to design specific and highly potent agonists and antagonists for this receptor (8,9).
Pharmacological studies have shown that chemically distinct molecules such as peptides, peptoids, and non-peptides can bind to the CCK1R with very close affinities (8,9). On the other hand, within each chemical family of CCK1R ligands, compounds having close structures are agonists, partial agonists, or antagonists, indicating that appropriate modifications within the pharmacophore switches an agonist to an antagonist and vice versa. This can be illustrated with both peptide and non-peptide ligands of CCK1R. For instance, JMV 179, a CCK heptapeptide analogue in which the C-terminal amidated phenylalanine and the L-tryptophan have been replaced by a phenylethyl ester moiety and a D-tryptophan, respectively, is a full CCK1R antagonist (10). This antagonist has been converted to JMV 180, a peptide exhibiting dual agonistic/antagonistic activity, by exchanging the D-tryptophan for an L-tryptophan (7,11). Another interesting example came from the discovery of the non-peptide CCK1R agonist, SR-146,131, by chemical modification of the CCK1R antagonist, SR-27,897 (12,13) (Fig. 1). These examples, which could probably be extended to multiple G protein-coupled receptors, raise the important questions of whether closely related ligands having opposite biological activities share the same binding site and of which intrinsic mechanism(s) at the binding site level govern(s) G protein-coupled receptor functioning.
One of our recent objectives has been to define the agonist binding site on the CCK1R and to identify interactions between critical residues of that binding site and chemical functions of the pharmacophoric domain of CCK ( Figs. 1 and 3). Amino acids within three regions of the CCK1R were identified as belonging to the binding site for CCK. Trp-39 and Gln-40, located at the top of transmembrane helix I, were shown to interact directly with the N-terminal portion of CCK (14). , located in the second extracellular loop, were then shown to interact with the sulfated tyrosine (15,16). More recently, Arg-336 and Asn-333, at the top of helix VI, were demonstrated to pair with the Asp carboxylate and the Cterminal amide of CCK, respectively (17). The first two identified amino acids of the CCK1R (Trp-39 and Gln-40) contribute weakly to CCK1R affinity for and response to CCK, whereas all others play a more critical role because of their interaction with residues of CCK, which are essential for both binding and biological activity of CCK. However, contact points within the CCK1R binding site for other key residues of CCK such as the Trp, Met, and Phe residues have not yet been identified.
To progress in the mapping of the CCK1R binding site(s), the three-dimensional model of the CCK1R⅐CCK complex was optimized, leading to the identification of putative amino acids involved in the interaction with the Met and Phe residues of the ligand CCK. Mutation of candidate residues and extensive characterization of the resulting mutants allowed us to position the C-terminal biological part of CCK in hydrophobic pockets formed by aromatic and nonaromatic amino acids located in the upper half of transmembrane helices III, V, VI, and VII. Our data, therefore, refute the model of the CCK1R⅐CCK complex proposed by other investigators in which the C terminus of CCK interacts with an amino acid residue (Trp-39) of helix I (18). Furthermore, binding site for the non-peptide agonist SR-146,131 was identified and experimentally validated as overlapping with that of the C-terminal tripeptide of CCK. Finally, the role of Met-121 located on helix III in the activation of the CCK1R by agonists was demonstrated.
Computer Modeling of the CCK1R and CCK1R⅐CCK Complex-The model of empty CCK1R was built using the transmembrane (TM) helical arrangement found in the bacteriorhodopsin crystal structure as starting point (21). It was then modified according to the rhodopsin crystal structure (22,23) and to the mutant data base "input/output" information scheme defined in the Viseur program (24). Extracellular and intracellular loops connecting the transmembrane helices were then added to the preliminary seven-helix bundle, and the structural model was optimized by the use of simulated annealing procedures. The entire system was finally relaxed and submitted to a 1-ns molecular dynamics with possible translational and rotational movements of individual TM helices taken into account. The final model respects most transmembrane arrangements found in the recent x-ray structure of rhodopsin (23). For docking of CCK ligand into the CCK1R binding site, experimental data that identified contact points between residues Trp-39 and Gln-40 and the N-terminal moiety of CCK served as a first constraint to place CCK within the CCK1R grove (14). In a first step, manual docking was achieved by taking into account molecular electrostatic potentials at the top of the receptor grove. The resulting complex was submitted to annealing molecular dynamics calculations. The resulting theoretical positioning of CCK into the CCK1R binding site was experimentally validated by two-dimensional site-directed mutagenesis. By doing so, Met-195-Arg-197, Arg-336, and Asn-333 were shown to belong the CCK1R binding site and to interact with the sulfated tyrosine of CCK, the Asp-8 carboxylate, and the C-terminal amide, respectively (15)(16)(17). The program package (Insight II, Discover, Homology, Biopolymer) from Molecular Simulations Inc. (San Diego, CA) was used for all the calculations.
Site-directed Mutagenesis and Transfection of COS-7 Cells-Mutant receptor cDNAs were constructed by oligonucleotide-directed mutagenesis (QuikChange site-directed mutagenesis kit, Stratagene, France) using the human CCK1R cDNAs cloned into pRFENeo vector as template (25). Oligonucleotides were designed to include a silent restriction site to facilitate analysis of mutant constructs by restriction endonuclease digestion. The presence of the desired and the absence of undesired mutations were confirmed by automated sequencing of both cDNA strands (Applied Biosystems). COS-7 cells (1.5 ϫ 10 6 ) were plated onto 10-cm culture dishes and grown in Dulbecco's modified Eagle's medium containing 5% fetal calf serum (complete medium) in a 5% CO 2 atmosphere at 37°C. After overnight incubation, cells were transfected with 2.5 g/plate of pRFENeo vectors containing the cDNA for the wild-type or mutated CCK1 receptors, using a modified DEAE-dextran method. Cells were transferred to 24-well plates at a density of 80,000 -150,000 cells/well 24 h after transfection. Receptor Binding Assay-Approximately 24 h after the transfer of transfected cells to 24-well plates, the cells were washed with phosphate-buffered saline, pH 6.95, 0.1% BSA and then incubated for 60 min at 37°C in 0.5 ml of Dulbecco's modified Eagle's medium, 0.1% BSA with either 71 pM 125 I-BH-(Thr,Nle)-CCK-9 or 1.83 nM [ 3 H]SR-27,897 in the presence or the absence of competing agonists or antagonists. The cells were washed twice with cold phosphate-buffered saline, pH 6.95, containing 2% BSA, and cell-associated radioligand was collected with 0.1 N NaOH added to each well. The radioactivity was directly counted in a ␥ counter (Auto-Gamma, Packard, Downers Grove, IL) or added to scintillant and counted for the tritiated radioligand.
Inositol Phosphate Assay-Approximately 24 h after the transfer to 24-well plates and following overnight incubation in complete medium containing 2 Ci/ml myo-2-[ 3 H]inositol, the transfected cells were washed with Dulbecco's modified Eagle's medium and then incubated for 30 min in 1 ml/well Dulbecco's modified Eagle's medium containing 20 mM LiCl at 37°C. The cells were washed with PI buffer at pH 7.45: phosphate-buffered saline containing 135 mM NaCl, 20 mM HEPES, 2 mM CaCl 2 , 1.2 mM MgSO 4 , 1 mM EGTA, 10 mM LiCl, 11.1 mM glucose, and 0.5% BSA. The cells were then incubated for 60 min at 37°C in 0.3 ml of PI buffer with or without ligands at various concentrations. The reaction was stopped by adding 1 ml of methanol/chlorhydric acid to each well, and the content was transferred to a column (Dowex AG 1-X8, formate form, Bio-Rad) for the determination of inositol phosphates. The columns were washed twice with 3 ml of distilled water and twice more with 2 ml of 5 mM sodium tetraborate, 60 mM sodium formate. The content of each column was eluted by addition of 2.5 ml of 1 M ammonium formate, 100 mM formic acid. 0.5 ml of the eluted fraction was added to scintillant, and ␤ radioactivity was counted.
Membrane Preparation-Approximately 65 h after transfection, the cells were washed three times with phosphate-buffered saline, pH 6.95, scraped from the plate in 10 mM Hepes buffer, pH 7.0, containing 0.01% soybean trypsin inhibitor, 0.1% bacitracin, 0.1 mM phenylmethylsulfonyl fluoride and frozen in liquid N 2 . After thawing at 37°C, the cells were subjected to another cycle of freeze/thawing and then centrifuged at 25,000 ϫ g for 20 min. The membrane pellet was resuspended in 50 mM Hepes buffer, pH 7.0, containing 115 mM NaCl, 5 mM MgCl 2 , 0.01% soybean trypsin inhibitor, 0.1% bacitracin, 1 mM EGTA, 0.1 mM phenylmethylsulfonyl fluoride (binding buffer); aliquoted; and stored at Ϫ80°C until use. Membrane protein concentrations were determined using the Bio-Rad protein assay kit. To assess the effect of GTP␥S on CCK binding, membranes from transfected COS-7 cells (0.4 -4 g of proteins) were incubated with 71 pM 125 I-BH-(Thr,Nle)-CCK-9 in the absence or in the presence of increasing concentrations GTP␥S in binding buffer for 120 min at 25°C. Nonspecific binding was measured in the presence of 1 M CCK.

Importance of Methionine 7 and Phenylalanine 9 of CCK for
Binding and Activation of the CCK1R-Previous structureactivity studies using synthetic replicates of CCK and pancreatic acini from rodents, a biological model naturally expressing CCK1R, have clearly shown the importance of both Met and Phe residues for binding and activity of CCK; however, no such study has been reported for human CCK1R in any expression system (26). Therefore, we first determined to what extent the Met and Phe side chains contribute to the affinity of CCK for human CCK1R expressed in COS-7 cells and to its capacity to induce production of total inositol phosphates. As shown in Fig.  2, replacement of Met-7 in CCK by an Ala residue caused 4000and 390-fold decrease in affinity and potency, respectively. In contrast, substitution of Met-7 by Nle did not affect affinity and potency of CCK. Furthermore, exchange of Phe for Ala was found to induce a 4900-and 2700-fold decrease in affinity and potency of the analogues, respectively. The efficacies of (Nle-7)-CCK, (Met-7)-CCK, and (Ala-7)-CCK were comparable, whereas that of (Ala-9)-CCK reached only 40% referred to the parent analogue (Nle-7)-CCK. Hence, both Met-7 and Phe-9 side chains contribute significantly to receptor binding and activation potency of CCK, confirming data obtained previously on receptors from rodents and guinea pig (27,28). These results also confirm that replacement of Met-7 with Nle does not affect affinity and activity of CCK.

Identification of Candidate Amino Acid Residues of the CCK1R for Interaction with Nle/Met-7 and Phe-9 of CCK Using
Molecular Modeling-Considering the major anchoring points of CCK inside the receptor discovered in our previous studies, it appears that the Nle/Met-7 residue is located in the vicinity of a hydrophobic pocket constituted by residues Leu-50, Ile-51, Leu-53, Pro-101, Val-125, Met-121, Ileu-352, and Leu-356. In the model of the (WT)-CCK1R, both Nle-7 or Met-7 side chains are positioned in the same way. The Phe-9 aromatic side chain of CCK is also positioned into a well defined cavity delineated by Phe-330 at the bottom, and Pro-177, Val-125, Leu-214, Ile-329 around Phe-330, which itself belongs to a large aromatic cluster constituted by the side chains of Cys-94, Phe-130, Trp-166, Phe-170, Phe-218, Phe-322, Phe-323, Trp-326, and Tyr-360. These two pockets are connected via the Val-125 side chain so that helix movements changing the structure of one of them may have consequences on the other. Examination of the organization of the two clusters in the three-dimensional model of the CCK1R⅐CCK complex suggests that not only a single, but several hydrophobic residues are contributing to the binding energy between Nle/Met-7 or Phe side chains of CCK and the CCK1R. As a consequence and unlike the charged residues of the binding pocket that were characterized previously, a mutation of only one of these amino acid residues of the CCK1R was not expected to induce changes in affinity and biological potency of the CCK1R to an extent comparable with effects caused by replacements of Nle/Met or Phe in CCK.
Effects of CCK1R Mutations on CCK1R Expression and Affinity for the Non-peptide Antagonist SR-27,897-Among all residues forming hydrophobic clusters around the Nle/Met-7 and Phe-9 side chains, in a first instance those in closest contact were chosen for mutagenesis experiments (Fig. 3). These were exchanged for amino acids lacking the chemical functions thought to be responsible for the interactions. In addition, Met-121 and Ile-329, which appear to be important for the equilibrium within the hydrophobic clusters surrounding Nle/Met-7 and Phe-9 of CCK, were each exchanged for more bulky and hydrophobic residues, namely Val and Phe, respectively. In a first series of experiments, COS-7 cells expressing mutated receptors were assayed for binding of the non-peptide antagonist [ 3 H]SR-27,897. Indeed, binding of this antagonist, unlike that of an agonist, offers the advantage of allowing detection of CCK1R independent of the coupling state to G protein(s), thus yielding accurate expression levels of all mutants (15,17). Ligand binding data are summarized in Table I. No binding of [ 3 H]SR-27,897 was found with the (I329F)-CCK1R mutant, a result that could be interpreted at this stage of the study either by an absence of expression of the mutant or by a direct or indirect effect of the mutation on receptor affinity for the ligand. Results with all other CCK1R mutants confirmed their expression at COS-7 cell surface at levels varying from 0.6 to 10.0 pmol/10 6 cells, which permitted further characterization. In addition, the binding data clearly revealed that all mutants (except (I329F)-CCK1R) bind the non-peptide antagonist with an affinity very close to that of (WT)-CCK1R. This finding indicates that the mutations did not dramatically disturb the conformational state of the CCK1R and that the mutated residues are not directly involved in the binding pocket of the antagonist SR-27,897.
Effects of Mutations on the CCK1R Affinity for CCK-To verify if the amino acid residues identified by molecular modeling are effectively involved in the CCK1R binding pocket for CCK, COS-7 cells expressing the mutated receptors were tested for binding of the agonist radioligand ( 125 I-BH-(Thr,Nle)-CCK-9). Binding parameters calculated from Scatchard analysis of binding values are summarized in Table I. According to these binding data, the CCK1R mutants can be classified in four categories. A first series of mutants, i.e. 6-, and 32-fold decrease in their potency to stimulate inositol phosphate production if compared with the (WT)-CCK1R. The biological efficacy of (C94L)-and (F330A)-CCK1R was reduced to 45 and 57% of that of the (WT)-CCK1R. It is worthy to note that (I329F)-CCK1R, which failed to bind both the CCK and SR-27,897 radioligands, was capable of inducing production of inositol phosphates after (Nle-7)-CCK stimulation, although its potency was 507-fold decreased compared with the (WT)-CCK1R. This result suggests that the absence of binding was likely because of a drastically reduced affinity of the mutant for CCK and SR-27,897. Indeed, this explanation was indirectly confirmed with experiments, which showed that the potency of SR-27,897 to inhibit CCK-induced production of inositol phosphates by COS-7 cells expressing (I329F)-CCK1R was 625-fold decreased if compared with the (WT)-CCK1R (1500 nM versus 24 nM, n ϭ 2, not shown). Finally, and very surprisingly, the mutant (M121V)-CCK1R failed to induce production of inositol phosphate upon stimulation with (Nle-7)-CCK, whereas (M121A)-CCK1R retained the ability to couple to phospholipase C, even though with a moderate efficacy (48% of that of (WT)-CCK1R; see Table II and Fig. 4). These data suggest that Met-121 plays a crucial role in the coupling of CCK1R to G proteins(s), although this residue is located in the upper part of helix III. Indeed, the importance of the side chain of residue 121 in the CCK1R for G protein(s) coupling was further evaluated by experiments, which showed that the nonhydrolyzable analogue of GTP, GTP␥S, did not dissociate binding of 125 I-BH-(Thr,Nle)-CCK-9 to (M121V)-CCK1R, whereas it dissociated binding of CCK to (M121A)-and (WT)-CCK1R to extents that agree with the biological efficacies of these receptors (Fig. 4).
Effets of Nle/Met Exchange in CCK and Met-121 Mutations in CCK1R-In all our studies regarding mapping of the CCK1R binding sites including the current one, we used a CCK analogue in which Met-7 was replaced by a Nle residue because of its high stability compared with Met in terms of possible oxidation on handling (19). The use of this analogue was supported by previous findings that confirmed its full biological potency. This was also confirmed in the present study, as shown in Fig. 2, where replacement of Met-7 with Nle was found to be without any effect on affinity and activity of the ligand. According to the three-dimensional model of the CCK1R⅐CCK complex, Nle (or Met) of CCK is inserted into a hydrophobic pocket including residues Leu-50, Ile-51, Leu-53, Cys-94, Met-121, Val-125, Ile-352, and Leu-356. To analyze whether the biological efficacy of mutated receptors was de-  pendent on the presence of Met-7 in CCK, a series of experiments was performed. Both competition binding and inositol phosphate production assays clearly demonstrated that (Met-7)-CCK bound to and stimulated all the mutants in a manner identical to that for the (Nle-7)-CCK analogue (data not shown), except for (M121V)-CCK1R and (M121A)-CCK1R. Indeed, (M121V)-CCK1R, which was unable to induce production of inositol phosphates upon (Nle-7)-CCK stimulation, responded to a (Met-7)-CCK stimulation with a maximum that reached 33% of that obtained with the (WT)-CCK1R, and with a potency that was only 12.5-fold decreased (EC 50 : 12.4 Ϯ 1.1 nM versus 1.0 Ϯ 0.1 nM, n ϭ 3) (Fig. 5). Interestingly, in a competition binding assay with 125 I-BH-(Thr,Nle)-CCK-9, the affinity of (Met-7)-CCK for (M121V)-CCK1R was very similar to that of (Nle-7)-CCK (IC 50 : 25.4 Ϯ 0.2 nM versus 19.0 Ϯ 1.3). Moreover, (M121A)-CCK1R was found to couple to production of inositol phosphates upon (Met-7)-CCK stimulation with a higher efficacy than it did upon (Nle-7)-CCK stimulation (70% versus 48% relative to (WT)-CCK1R, not illustrated). Therefore, double exchange of Met tor Nle in CCK and Met-121 to Val in CCK1R yielded an inactive CCK1R⅐CCK complex. On the other hand, the presence of Met-7 in CCK was required for functional coupling of (M121V)-CCK1R and enhanced that of the (M121A)-CCK1R mutant.
The intent of the mutation of Met-121 to Val was to disturb the hydrophobicity pattern within the pocket surrounding Nle of CCK. This mutation, which was expected to decrease the binding affinity of the CCK1R for CCK, effectively caused a 16-fold drop in the affinity of the receptor for the ligand and, surprisingly, made activation of that receptor dependent on the presence of Met-7 in CCK. To further explore the role of residue 121 of the CCK1R and of Met-7 of CCK, the (M121Q)-CCK1R mutant was expressed and a (Gln-7)-CCK analogue was synthesized. Competition binding to (WT)-CCK1R revealed that substitution of Met-7 residue by Gln in the ligand affects only weakly the affinity of the peptide (K d(1) : 12.5 Ϯ 2.4 nM, K d (2) : 332 Ϯ 78 nM, n ϭ 3, not shown). In agreement with this high binding affinity, (Gln-7)-CCK was found to stimulate inositol phosphates with high potency (D 50 : 13.8 Ϯ 2.2 nM) (Fig. 5). Furthermore, the mutant (M121Q)-CCK1R binds 125 I-BH-(Thr,Nle)-CCK-9 at two classes of binding sites with binding parameters similar to those of the (WT)-CCK1R (K d (1) : 0.8 Ϯ 0.6 nM, K d (2) : 120 Ϯ 50 nM, not illustrated) and this binding is coupled to the production of inositol phosphates at a potency similar to the (WT)-CCK1R upon (Nle-7)-CCK stimulation (EC 50 : 1.8 Ϯ 0.6 nM, E max : 95% of (WT)-CCK1R). Finally, (Gln-7)-CCK stimulates the (M121V)-CCK1R mutant as did (Met-7)-CCK (Fig. 5). Altogether, this set of results excludes a direct involvement of the sulfur atoms of Met-7 of CCK and Met-121 of the CCK1R in the mechanism of activation of the receptor. However, the presence of a polar atom within the hydrophobic cluster surrounding these two residues appears to be required for CCK1R activation.
Using a molecular dynamic simulated annealing procedure, three-dimensional models of the different CCK1R⅐CCK complexes was generated. As illustrated in Fig. 6, the location of the C-terminal part of CCK was identical in the (WT)-CCK1R⅐(Met-7)-CCK and (WT)-CCK1R⅐(Nle-7)-CCK complexes (Fig. 6a), which is consistent with the identical binding and functional properties of these complexes (Fig. 2). Conversely, in the (M121V)-CCK1R⅐(Nle)-CCK complex, the position of the ligand differs as shown in Fig. 6b. Indeed, in this complex, the Val-121 and Nle-7 residues are in tight contact leading to a deeper insertion of the peptide into the receptor grove. As a consequence, interactions between Asn-333 and the C-terminal amide of the ligand are precluded and the aromatic ring of the CCK Phe residue is moved into the direction of the Val-121 side chain. Reintroduction of Met-7 in CCK reverses these altered docking modes in the (M121V)-CCK1R⅐(Nle)-CCK complex, and the resulting structural models of the (M121V)-CCK1R⅐(Met)-CCK and (WT)-CCK1R⅐(Nle)-CCK complexes become very similar (Fig. 6, c and d). Hence, the positioning of the C-terminal part of CCK apparently differs in the (M121V)-CCK1R⅐(Nle)-CCK complex as the only inactive complex from that in the active complexes.
Effects of CCK1R Mutations on SR-146,131-induced Production of Inositol Phosphates-SR-146,131 is an agonist having high affinity and specificity for the CCK1R. Although of nonpeptidic nature, this compound exhibits some structural similarities with the C-terminal part of CCK docked into the threedimensional model of the CCK1R (Fig. 1). This observation raised the interesting question of whether these structural similarities imply an overlapping of the binding sites for the two ligands, SR-146,131 and CCK. A way to address this question would be to analyze the effects on binding of SR-146,131 by the mutation of those residues in the receptor which were found to be involved in CCK binding. However, as no labeled SR-146,131 was available, these effects were determined by measuring SR-146,131-stimulated production of inositol phosphates. The results (Table III) revealed that residues Met-195 and Arg-197, which were shown previously to interact with the sulfated tyrosine of CCK, did not affect recognition of SR-146,131. Conversely, residues Asn-333 and Arg-336, which were demonstrated previously to pair with the C-terminal amide and the carboxylate side chain of Asp-8 of CCK, respectively, are likely involved in recognition of SR-146,131 because their mutation caused a 120-and 126-fold decrease in potency of inositol phosphate responses, respectively. Among the residues of the receptor that were shown to interact with the Nle/Met-7 and Phe residues, several seem to be involved in SR-146,131-induced inositol phosphate production. In fact, the mutants (  (Table III).
Docking of SR-146,131 into the CCK1R Structural Model-On the basis of above experimental data, dynamic docking of SR-146,131 into the three-dimensional model of the CCK1R was carried out using a molecular dynamic simulated annealing procedure described elsewhere. Compared with the docking of CCK the positioning of SR-146,131 exhibits some similarities. Indeed, the SR-146,131 carboxylate group interacts with the guanidino function of Arg-336, and its chloro-dimethoxy aromatic ring is positioned inside the same pocket as the Phe-9 side chain of CCK (Fig. 7). Asn-333 of the CCK1R is also involved in interactions with the non-peptide agonist. This interaction occurs through a possible hydrogen bond between the Asn-333 carboxamido group and the chlorine of the ligand, which could be activated by the two O-methyl groups. The cyclohexane moiety of SR-146,131 is positioned in proximity to the Cys-94, Val-125, Ile-352, and Leu-356 side chains, which are involved in the binding site of CCK (Fig. 7). It is noteworthy that the sulfur atom and the carbonyl of SR-146,131 are positioned almost at the same place as the sulfur atom of the CCK Met-7 in the CCK1R⅐CCK complex, with the indole group of SR-146,131 occupying the position of CCK Trp-6 -Met-7 backbone.

DISCUSSION
The aim of the present study was to advance in the knowledge of the binding site of CCK1 receptor for CCK, by identifying amino acid residues that interact with two crucial residues of the CCK, namely Met and Phe for which no receptor partner was yet identified. For this purpose, approaches of  For modeling experiments, a structural model of the CCK1R⅐CCK complex was generated in which amino acid residues of the binding site for CCK were shown previously to be involved by pharmacological and functional analysis of CCK1R mutants using several peptidic and non-peptidic ligands (14 -17). In this structural model, the C-terminal portion of CCK was strongly constrained in the receptor grove because of interactions between Arg-336 and the carboxylate of the penultimate Asp residue of CCK, and between Asn-333 and the C-terminal CCK amide. These interactions restrict considerably possible movements of the C-terminal part of CCK within the bottom of the receptor grove. In fact, after dynamic docking, two closely linked hydrophobic clusters appeared as likely candidates for interactions with the Nle/Met 7 and Phe side chains of the ligand, which are so critical for binding affinity and biological potency of the neuropeptide. This CCK1R⅐CCK structural model was well supported by a large set of experimental data and their physicochemical interpretation. The observation that mutation of residues such as Leu-50, Ile-51, Leu-53, and Val-125 affect only slightly recognition of CCK by the receptor is in full agreement with their expected low contribution to the stability of CCK1R⅐CCK complex. The moderate decrease in affinity and potency (10-and 3-fold) caused by mutation of Phe-330 was consistent with a T-shaped interaction between the aromatic rings of CCK Phe-9 and receptor Phe-330. The positioning of the C terminus of CCK was furthermore in agreement with the dramatic effect of the Ile-329 3 Phe mutation on CCK-induced CCK1R activation, which likely results from a reduced ability of the mutant to bind CCK. Indeed, in the three-dimensional model of (I329F)-CCK1R⅐CCK complex, exchange of Ile-329 for a Phe residue causes rotation of the amide of CCK away from Asn-333, its partner in (WT)-CCK1R. Consequently, the interaction between the CCK amide and the carboxamide of Asn-333 is lost (data not shown). The similar properties of (I329F)-CCK1R and the previously characterized (N333A)-CCK1R support this explanation (17). Indeed, the (N333A)-CCK1R mutant mediates CCK-induced inositol phosphate production with a 1350-fold lower potency than (WT)-CCK1R and with an efficacy 60% of that of (WT)-CCK1R (17). Two sets of experimental data obtained with the receptor mutated at residues Ile-352 and Met-121 further validate the location of the C-terminal portion of CCK. Exchange of Ile-352 for an Ala caused a 232-fold shift in the potency of CCK1R to induce inositol phosphates, a result that was likely the result of an important decrease in the binding affinity of (I352A)-CCK1R for CCK. This result agrees with observations from the three-dimensional receptor model, which suggest that, in the empty receptor, Ile-352 of TM7 interacts with Ile-329 of TM6 and that, upon docking of CCK, the Nle/Met-7 side chains disrupt the interaction between these two residues.
To confirm the existence of the hydrophobic cluster surrounding Nle/Met of CCK, we predicted that mutation of Met-121 to Val will decrease binding affinity of the CCK1R for CCK because of presence of bulky isopropyl side chain of Val residue, whereas mutation of Met-121 to Ala will not significantly affect binding. Experimental data confirmed the modeling predictions because mutants (M121V)-and (M121A)-CCK1R bound CCK with affinities that were 16-and 1.8-fold lower than that of the (WT)-CCK1R, respectively. Proximity between Met-121 and Met/Nle of CCK was further documented by the set of data showing that Met-7 of CCK and Met-121 of the CCK1R were interchangeable to yield an active CCK1R⅐CCK complex (see below).
The present study also provided important new data concerning the comparison between the binding site for the nonpeptide agonist SR-146,131 and the binding site for CCK. Both the experimental data obtained with this compound and the CCK1R mutants and its dynamic docking to the three-dimensional receptor model agree with a positioning of the nonpeptide agonist into the part of the CCK1R binding site that interacts with the C-terminal tripeptide of CCK, Met-Asp-Phe-NH 2 . Residues from transmembrane segments VI (Arg-336, Asn-333, Phe-330, Ile-329) and VII (Ile-352, Leu-356) clearly appear to be involved in recognition of SR-146,131. Our findings are consistent with a recent report showing involvement of several of these residues, namely Asn-333, Arg-336, Ile-329, and Leu-356 in CCK1R binding site for SR-146,131, albeit the proposed model for the CCK1R.SR-146,131 complex differs slightly from ours (29,30). However, in the two models, the location of SR-146,131 in respect to Met-121, Leu-356, and Cys-94 is very similar. Additional evidence for a similar location of CCK and SR-146,131 relative to the Met-121 residue is derived from the capability of the non-peptide agonist to stimulate coupling of (M121V)-CCCK1R to phospholipase-C as does (Met-7)-CCK. The fact that the CCK1R binding site for SR-146,131 is overlapping part of the binding site for the biologically essential region of CCK strongly validates docking of the C terminus of CCK into a cavity formed by the transmembrane helices V, VI, and VII. This docking mode of the C terminus of CCK into CCK1R is further supported by an NMR study of the interactions between CCK and a fragment of CCK1R comprising the top portion of helix VI and the third extracellular loop (31). On the other hand, it differs from that derived from photoaffinity labeling studies (18,32,33). In the photoaffinity labeling experiments with a CCK photoprobe in which the C-terminal Phe residue was replaced by benzophenylalanine, Trp-39 at the top of helix I was identified, supporting the hypothesis that the C terminus of CCK was in close proximity of the first helix (32). Based on these findings, a model was proposed of the peptide-occupied CCK1R in which CCK resembles an hairpin lying at the receptor extracellular surface with the C-terminal Phe of CCK being in contact with Trp-39 of the receptor (18). It is very likely that the large amount of energy required to generate the covalent bond from a p-nitrophenylalanine (or benzophenone) (30 min of UV irradiation at a wavelength of 300 nm) significantly affected the conformation of both the CCK1R and CCK, leading to movement of CCK within its binding site. On the other hand, by definition, photoaffinity labeling using structurally modified CCK could not identify amino acids of the receptor in interaction with native CCK.
Finally, with the current study, we succeeded in providing the first data related to the process of CCK1R activation following agonist binding. Indeed, we found that an exchange of Met-121 for a Val residue leads to a CCK1R mutant that is unable to induce inositol phosphate accumulation upon (Nle-7)-CCK stimulation. With this mutant, 30% of biological activity is recovered upon stimulation by (Met-7)-and (Gln-7)-CCK and 84% upon stimulation by SR-146,131. From a general point of view, results with mutants at position 121 show that ascribing a functional role to a residue must take into account the fact that impact of mutations on pharmacological and functional properties of a receptor can depend on the nature of the amino acid by which a critical residue is substituted as well as on the ligand used to analyze the mutants. Direct involvement of the sulfur atom of Met-7 of CCK or Met-121 of CCK1R in the mechanism of receptor activation can be ruled out by results demonstrating that the Gln side chain can mimic the Met side chain. However, the presence of a polar atom within the hydrophobic cluster surrounding the Met-121/Met-7 residues seems to be required for CCK1R activation. Indeed, in terms of production of inositol phosphates, the relative efficacies of the mutated complexes were (M121V)-CCK1R⅐(Nle-7)-CCK ϭ 0, Ͻ (M121V)-CCK1R⅐(Met-7)-CCK Ͻ (M121A)-CCK1R⅐(Nle-7)-CCK Ͻ (M121A)-CCK1R⅐(Met-7)-CCK. The more the side chain of residues in the vicinity of position 121 is hydrophobic, the less efficient is phospholipase C activation. In agreement with this view, efficacy of mutant M121A represents ϳ50% of that of (WT)-CCK1R, a value in accordance with moderate "hydrophobic weight" of the methyl of Ala relative to isopropyl of Val. The much higher efficacy of SR-146,131 compared with (Met-7)-CCK in stimulating functional coupling of (M121V)-CCK1R to phospholipase C can tentatively be ascribed to the presence of two polar elements, i.e. a sulfur atom and a carbonyl moiety in the vicinity of position 121 of the receptor. Another major support for the peculiar role of the residues in position 121 of the receptor and 7 of the ligand was derived from molecular dynamic modeling of the different complexes, which show the importance of the amino acid side chains in these positions for the correct positioning of the C-terminal part of CCK, particularly toward Phe-330 of the CCK1R. An interaction between the aromatic ring of CCK Phe-9 and that of Phe-330 (T-shape) seems to be important for CCK1R full activation. This view is in line with the 60% decrease in efficacy to stimulate inositol phosphate production caused by an exchange of Phe-9 with Ala in CCK (Fig. 2). Furthermore, stimulation of inositol phosphate production was strongly affected by mutation of residue Phe-330, as it was affected by mutations of Cys-94 and Phe-218. In addition to their role in CCK-induced production of inositol phosphates, binding results also argue in favor of a role of the residues Met-121, Cys-94, Phe-218, and Phe-330 in the conformational stability of CCK1R. Indeed, the data showed that exchange of Met-121 for a Val or an Ala converted the whole CCK1R population into a single, relatively high, affinity state as did mutation of residues Cys-94, Phe-218, and Phe-330. All these data can be interpreted by considering the prevailing model for G protein-coupled receptor activation, which is the allosteric ternary complex formed between the receptor R, the agonist L, and G protein(s). According to this model, in absence of any agonist stimulation, R is believed to undergo spontaneous conformational changes, however, with the inactive conformation being energetically the most stable. Binding of an agonist would either induce or stabilize active receptor species (R*), or both (34 -36). Accordingly, Met-121, Phe-330, Cys-94, and Phe-218 would represent key residues allowing the receptor either to be stabilized in an inactive conformation in absence of ligand and to undergo proper conformational changes for G protein(s) coupling and phospholipase C activation in presence of the agonist.
Further investigations are obviously required to determine precisely how certain residues such as those pointed out in this study regulate affinity state and/or activation of the CCK1R. In light of recent data with rhodopsin and ␤ 2 -adrenergic receptor, we hypothesize that these residues play a pivotal role for helix movements upon CCK binding thereby enabling G protein(s) to efficiently couple with previously buried region of that receptor (37,38). Several conserved regions at the bottom of helix III (such as (D/E)-R-Y motif), helix VII (such as N-P-X-X-Y), and helix VI near the third intracellular loop boundary have been shown to be directly involved in activation process of G proteincoupled receptors (39 -42).
To summarize, the current study has provided much information regarding positioning of the C-terminal part of CCK, which triggers the biological activity of the peptide, in the CCK1R. SR-146,131, a non-peptide agonist, was found to occupy this critical region of CCK1R binding site. Data showing the importance of several amino acid residues involved in this region of CCK1R binding site will be used to investigate CCK1R functioning.