Mutation of Asn-391 within the Conserved NPXXY Motif of the Cholecystokinin B Receptor Abolishes Gq Protein Activation without Affecting Its Association with the Receptor*

Among the most conserved regions in the G-protein-coupled receptors is the (N/D)PX2–3Y motif of the seventh transmembrane domain (X represents any amino acid). The mutation of the Asn/Asp residue of this motif in different G-protein-coupled receptors was shown to affect the activation of either adenylyl cyclase or phospholipase C. We have mutated the Asn residue (Asn-391) of the NPXXY motif in the CCKBR to Ala and determined the effects of the mutation on binding, signaling, and G-proteins coupling after expression of the mutated receptor in COS cells. The mutated receptor displayed similar expression levels and high affinity CCK binding compared with the wild type CCKBR. However, unlike the wild type CCKBR, the mutated receptor was completely unable to mediate activation of either phospholipase C and protein kinase C-dependent and -independent mitogen-activated protein kinase pathways, indicating an essential role of Asn-391 in CCKBR signaling. Coimmunoprecipitation experiments allowed us to show that the inactive mutant retains an intact capacity to form stable complexes with Gqα subunits in response to CCK. These results indicate that the formation of high affinity CCK-receptor-Gq protein complexes is not sufficient to activate Gq and suggest that Asn-391 is specifically involved in Gq proteins activation.

Cholecystokinin receptors belong to the superfamily of Gprotein-coupled receptors that are characterized by seven transmembrane ␣-helical domains (TM) 1 connected by extracellular and intracellular loops. Cholecystokinin binds with high affinity and exerts multiple physiological functions through two pharmacological subtypes, CCKAR and CCKBR (1)(2)(3). It is well described that CCKBR stimulates phospho-lipase C ␤ (PLC) and mitogen-activated protein kinase (MAPK) pathways (3). The insensitivity to pertussis toxin of CCKBR-PLC activation suggests coupling through the G q family of G-proteins (3,4). Occupancy of the CCKBR that leads to MAPK activation involves two different pathways, one involving protein kinase C (PKC) and the other one independent of PKC (5,6). The molecular mechanisms within the G-protein-coupled receptors that lead to the activation of these different intracellular effectors still remain largely unknown for most of the receptors and in particular for the CCKBR. Highly conserved residues found in G-protein-coupled receptors have been shown to be critical for maintaining normal receptor conformation and for other fundamental properties of the receptor such as high affinity binding, G-protein coupling, signaling, and regulation.
Among the most conserved regions is the (N/D)PX 2-3 Y motif that is present at the end of the seventh transmembrane domain of all G-protein-coupled receptors (where X represents any amino acid). The mutation of the Asn residue (that may be exceptional an Asp as in the thrombin and gonadotropin-releasing hormone receptors) in different G-protein-coupled receptors was shown to affect the activation of either adenylyl cyclase (7) or phospholipase C (7)(8)(9)(10)(11)(12)(13). Recently, this Asn was also reported to be important for the direct association between certain G-protein-coupled receptors and the small G-proteins, ARF and RhoA, to mediate activation of phospholipase D (14). Consistent with the importance of the Asn/Asp residue of TM7 in the activation of different pathways, a computational modeling study has reported a special structural property of the conserved N/DP motif different from a regular Pro-kink that would deviate the TM7 from an ideal ␣-helix and allow a local flexibility in the TM7 of G-protein-coupled receptors (15). The authors proposed that the flexibility introduced by this specific structural perturbation may play a role in G-protein-coupled receptors activation by functioning as a sensitive conformational switch.
To determine the role of the conserved Asn residue (Asn-391) of the NPXXY motif present in the CCKBR, we have mutated this Asn to Ala and determined the effects of the mutation on binding, signaling, and G-protein coupling. We found that the mutation of Asn-391 to Ala switches the CCKBR to a high affinity inactive conformation for the transmission of signal to different effectors. However, the mutated receptor retains an intact capacity to form stable complexes with G q ␣ subunits in response to CCK demonstrated by coimmunoprecipitation of the mutated receptor-G q ␣ complexes. Our results indicate that the formation of high affinity CCK-receptor-G q protein complexes is not sufficient to activate G q and suggest that Asn-391 is specifically involved in G q protein activation. To date, this is the first identification of a single residue in a G-protein-coupled receptor that appears to act as a molecular switch for receptormediating G q protein activation.
Construction of Mutant Receptor cDNAs-Mutant receptor cDNAs were constructed by oligonucleotide-directed mutagenesis (Quick-Change TM Site-directed Mutagenesis Kit, Stratagene, France), using the rat CCKBR cDNA subcloned in pCDL-SR␣ (pCDL-SR␣-WT-CCKBR) as template. Mutations were confirmed by DNA sequencing using an automated sequencer (Applied Biosystems).
Transfection of Wild Type and Mutant Receptor cDNAs into Mammalian Cells-COS-7 cells were grown in Dulbecco's modified Eagle's medium supplemented with 5% fetal calf serum. Usually 2 g (6 g for immunoprecipitation experiments) of plasmids coding for wild type or mutant CCKBR were transiently transfected into COS-7 cells using the DEAE/dextran method as described previously (19). For binding on whole cells or measurement of inositol phosphate accumulation experiments, 24 h after transfection, the transfected cells were transferred to 24-well culture plates and seeded at a density of approximately 10 5 cells/well.
Binding of 125 I-BH-CCK-9 to Transfected COS-7 Cells-Forty eight hours after transfection, binding experiments on whole cells were performed as we described previously (20,21) by incubating cells for 60 min at 37°C with 50 pM 125 I-BH-CCK-9 with or without increasing concentrations of unlabeled peptides. Binding studies in the presence of aluminum fluoride were performed as described previously (22) by using 50 pM 125 I-BH-CCK-9. For binding studies in presence of GTP␥S, plasma membranes from transfected COS cells were prepared as described (23) and incubated (50 -100 g) in binding buffer (50 mM HEPES (pH 7.5), 5 mM MgCl 2 , 115 mM NaCl, 0.01% soybean trypsin inhibitor, 0.1% bacitracin, 1 mM EGTA, 0.1 mM phenylmethylsulfonyl fluoride, and 0.2% bovine serum albumin) with 50 pM 125 I-BH-CCK for 60 min at 25°C in the presence of increasing GTP␥S concentrations. After incubation, bound and free radioligand were separated by centrifugation (10 min, 12,000 ϫ g), and the pellets were counted in a gamma counter. Binding assays were performed in duplicate in at least three separate experiments. Nonspecific binding (determined in the presence of 1 M CCK-9) was always less than 10% of total binding. Binding data were determined using the nonlinear, least squares, curve-fitting computer program, Ligand (24), and GraphPad Prism Program (Software). K i values were calculated as K i ϭ IC 50 /(1 ϩ (labeled ligand)/K d of labeled ligand).
Measurement of Total Inositol Phosphate Accumulation-Twenty four hours after transfection, the transfected cells were incubated overnight in Dulbecco's modified Eagle's medium with 2 Ci/well of myo-2-[ 3 H]inositol (18.6 Ci/mmol) (NEN Life Science Products). The next day, total inositol phosphate accumulation was measured after 1 h stimulation at 37°C by the indicated concentrations of CCK as described previously (20,21). EC 50 was calculated using GraphPad Prism Program Software.
Luciferase Reporter Assay-Elk reporting system (PathDetect Elk-1 Trans-Reporting System, Stratagene, Netherlands) was used to measure MAPK activation. COS-7 cells were transfected with 2 g of GAL-4-Elk1 fusion plasmid (pFA2-Elk-1), 4 g of GAL-4-luciferase reporter plasmid (pFR-Luc), 1 g of pCMV-␤-galactosidase plasmid, and either 2 g of pCDL-SR␣-WT-CCKBR or pCDL-SR␣-N391A-CCKBR plasmids. One day after transfection, the cells were transferred to 6-well plates at a density of 5⅐10 5 cells/well, serum-starved overnight, and then incubated for 7 h with the test substances. At the end of the incubation, cells were washed twice with phosphate-buffered saline and lysed with 300 l of cell lysis buffer (Luciferase Cell Culture Lysis Reagent, 5ϫ, Promega) for 15 min at room temperature. Cells were then centrifuged to eliminate cell debris. 40 l of cell lysate was mixed with 100 l of luciferase assay reagent (Promega), and luminescent intensity was measured using a ECG Bertold luminometer. Luciferase activity was expressed as relative light units and normalized for ␤-galactosidase activity. ␤-Galactosidase activity was measured by the luminescent light derived from 90 l of each sample incubated in 200 l of 1 mg/ml O-nitrophenyl-␤-D-galactopyranoside and used to correct transfection efficiency among the different treatment groups.
Western Blotting Analysis-Proteins (whole cell lysates or immunoprecipitates), prepared as described above, were separated by SDSpolyacrylamide gel electrophoresis on 5/15% linear gradient and transferred to Immobilon polyvinylidene difluoride membranes (Millipore). Then, membranes were saturated with saline buffer (10 mM Tris, 140 mM NaCl (pH 7.4) containing 5% (w/v) non-fat dried milk) and incubated overnight at 4°C with the indicated antibodies (anti-HA (3F10 clone) or anti-G q ␣/G 11 ␣ antibodies). Membranes were washed three times with saline buffer containing 0.5% non-fat dried milk and 0.5% Nonidet P-40 and incubated with 125 I-protein A (500,000 cpm/ml) for 1 h at room temperature. Proteins fixed on membranes were visualized by autoradiography.

The N391A-CCKBR Mutant Bound CCKBR Agonists and
Antagonists with High Affinity-To determine if the conserved Asn residue of the NPXXY motif (Asn-391) plays a role in binding and activation of the CCKB receptor, we have mutated this Asn to Ala. The mutant N391A-CCKBR and wild type (WT) CCKB receptors were transiently expressed in COS-7 cells. We first determined the affinity of N391A-CCKBR for CCK and several selective CCKBR agonists and antagonists by performing competition binding experiments using 125 I-BH-CCK-9 as radioligand. Scatchard analysis of CCK-9 competition binding to the WT-CCKBR and N391A-CCKBR mutant revealed that the N391A-CCKBR mutant has an affinity and a maximal binding capacity close to the WT-CCKBR (K d 2.21 Ϯ 0.41 nM and B max 1.31 Ϯ 1.18 pmol/10 6 cells versus K d 1.40 Ϯ 0.34 nM and B max 1.28 Ϯ 0.53 pmol/10 6 cells, respectively). Moreover, the N391A-CCKBR mutant bound CCK-4 and gastrin 17-I agonists with an affinity similar to the WT-CCKBR, whereas it bound the selective CCKB antagonists, PD135-158 and L365-260, with slightly increased affinities (Table I). These  (4,20,21). To assess the ability of the wild type and mutant receptors to activate the PLC-signaling cascade, we measured the accumulation of total inositol phosphates (IP). As illustrated in Fig. 1, CCK stimulated total IP production to the WT-CCKBR with an efficacy of 0.94 Ϯ 0.03 nM and a maximal potency of 11-fold over basal. Under similar experimental conditions, no detectable IP accumulation was measured with the N391A-CCKBR mutant even when exposed to 10 Ϫ4 M CCK. This result indicates that Asn-391 is essential for activation of the PLC pathway.
We then investigated the ability of WT-CCKBR and N391A-CCKBR mutant to activate the MAPK pathway. Both protein kinase C-dependent and -independent pathways were described to mediate CCKBR-MAPK activation in different cell lines (5,6). Since no such studies have been performed in COS cells, we first assessed whether the WT-CCKBR expressed in COS-7 cells was able to activate MAPK in response to CCK. We used Elk-1 and luciferase gene reporting systems to measure CCK-induced MAPK activation. Activation of MAPK is known to target transcription factors such as Elk-1 that regulates the activity of the promoter of the early response gene c-fos that was reported to be induced by the CCKBR (26,27).
As shown in Fig. 2, CCK induced luciferase activity in a concentration-dependent manner with an efficacy of 11.3 Ϯ 0.8 nM. A maximal increase of 18.5-fold in luciferase activity over basal was reached when WT-CCKBR-transfected cells were stimulated with 10 Ϫ6 M CCK. This effect was inhibited in a concentration-dependent manner by the highly specific MAPKK1-2 inhibitor PD98059 (28) (Fig. 3). These data indicate that CCK activates in COS cells a cascade of phosphorylation reactions that targets p42 and p44 MAPK. To determine if a PKC-independent pathway might be involved in CCKinduced MAPK activation, we used the PKC-specific inhibitor bisindolylmaleimide GF109203X (29). As shown in Fig. 4A, treatment of the cells with 5 M GF109203X completely abolished the PMA-stimulated MAPK activity. At a similar concentration, GF109203X inhibited 64% of CCK-induced MAPK activity (Fig. 4B), indicating that CCK-induced MAPK activation is both dependent and independent on PKC. We then determined if the N391A-CCKBR mutant was able to activate MAPK. Similar experiments were performed with COS-7 cells expressing the N391A-CCKBR mutant. No luciferase activity was detectable even at a concentration of 10 Ϫ4 M CCK (Fig. 5).
These results indicate that the mutation of Asn-391 to Ala completely inhibits both PKC-dependent and -independent activation of p42 and p44 MAPK.

CCK Binding on N391A-CCKBR Mutant Is Not Affected by GTP␥S and AlF 4
Ϫ -To determine if the impairment of mutant receptor activity could be associated to a default in G-protein activation, we measured the effects of the hydrolysis-resistant GTP analogue GTP␥S on 125 I-BH-CCK-9 binding to membrane preparations from COS-7 cells expressing the wild type and mutant receptors. As shown in Fig. 6, treatment with increasing concentrations of GTP␥S decreased the high affinity binding of 125 I-BH-CCK-9 to the WT-CCKBR by about 75%. In contrast, the high affinity binding of 125 I-BH-CCK-9 to the N391A-CCKBR mutant was not affected by concentrations of GTP␥S as high as 10 Ϫ3 M. We also determined the effect of AlF 4 Ϫ , which binds to the G␣GDP and mimics the gamma phosphate of GTP, on cells expressing the wild type and mutant receptors. In the presence of 30 mM NaF plus 10 M AlCl 3 , concentrations that we previously showed to affect high affinity binding of CCK to the WT-CCKBR (22), the N391A-CCKBR mutant exhibited unchanged high affinity CCK binding, whereas the WT-CCKBR showed a 90% decrease in CCK binding (Fig. 7). It is generally assumed that the high affinity state of agonists binding reflects the formation of a ligand-receptor-G-protein complex, and addition of guanine nucleotides or aluminum fluoride should reduce or eliminate such high affinity binding. The fact that the high affinity CCK binding to the N391A-CCKBR mutant is not affected in presence of GTP␥S or AlF 4 Ϫ might suggest that the mutant retained the capacity to interact with G-proteins but not to activate them. To determine whether the N391A-CCKBR mutant was able to couple to G-proteins when stimulated by CCK, we performed immunoprecipitation of the mutated receptor and investigated whether G-proteins were coassociated with the mutant.
The WT-CCKBR Stimulates Total Inositol Phosphate Production via G q Proteins-To date no study has reported direct association of the WT-CCKBR with a G-protein. A previous study on COS-7 cells expressing the WT-CCKBR showed that CCK-stimulated PLC activation was pertussis toxin-insensitive suggesting that the WT-CCKBR is coupled to a G-protein of the G q family (4). In order to verify the coupling of the WT-CCKBR to G q proteins, COS-7 cells were cotransfected with pCDL-SR␣-WT-CCKBR and increasing concentrations of pCIS-G q ␣ and assayed for total IP production after stimulation with 10 Ϫ6 M CCK. Coexpression of G q ␣ with the WT-CCKBR increased IP production in a concentration-dependent manner that reached a maximum when cells were transfected with 2 g of pCIS-G q ␣ (Fig. 8). In similar conditions, coexpression of G q ␣ with the N391A-CCKBR mutant did not induce any IP production indicating that the mutant does not stimulate IP production even when G q ␣ subunits are overexpressed (not shown). These results confirm that the WT-CCKBR couples to G q proteins, and thus in the subsequent experiments, we looked for the presence of G q proteins in the immunoprecipitates.
Association of the Inactive N391A-CCKBR Mutant with G q ␣ Subunits in Response to CCK-Immunoprecipitation of the wild type and mutant receptors was performed using an antibody directed against the hemagglutinin epitope (HA) fused to the amino terminus of the WT-CCKBR and N391A-CCKBR mutant transiently expressed in COS-7 cells. We verified that epitope-tagged wild type (HA-WT-CCKBR) and mutant (HA-N391A-CCKBR) receptors expressed in COS-7 cells displayed similar CCK affinities, maximal binding capacities, and IP production compared with untagged wild type and mutated receptors (not shown). In addition, to confirm that the transfection and expression efficiencies of both tagged receptors were similar, we performed immunoblotting of COS cells lysates expressing HA-tagged wild type and mutated receptors with anti-HA antibodies. The CCKBR was expected to migrate at a mass of 68 -97 kDa as determined previously with a photoreactive CCK probe in COS cells expressing the CCKBR (30). As shown in Fig. 9, equal amounts of HA-WT-CCKBR and HA-N391A-CCKBR migrating at a mass of 85 kDa were revealed (lanes 1 and 2, respectively), and in non-transfected cells no corresponding band was detected (lane 3).
The G q ␣ subunits associated with immunoprecipitated receptors were visualized by Western blotting with a well characterized, specific, and high affinity-purified polyclonal antibody raised against the carboxyl-terminal decapeptide of G q ␣/ G 11 ␣ subunits (31). We determined whether the G q ␣/G 11 ␣ antibodies were able to detect endogenous G q ␣/G 11 ␣ subunits by performing Western blot analysis on non-transfected COS-7 cells lysates. No endogenous G q ␣/G 11 ␣ subunits were detected with the G q ␣/G 11 ␣ antibodies (Fig. 10, lane 1). These results are consistent with a previous study reporting that COS-7 cells express low amounts of endogenous G q ␣/G 11 ␣ subunits (32). In contrast, when COS-7 cells were transfected with 2 g of pCIS-G q ␣, G q ␣ subunits were detected by G q ␣/G 11 ␣ antibodies migrating at a mass of 42 kDa (Fig. 10, lane 2). Thus, to permit visualization of receptor-G q ␣ complexes, COS-7 cells were cotransfected with plasmids coding for G q ␣ subunits and either HA-WT-CCKBR or HA-N391A-CCKBR mutant.
We first examined the association of HA-WT-CCKBR with G q ␣ subunits. COS-7 cells coexpressing HA-WT-CCKBR and G q ␣ subunits were stimulated with 10 nM CCK, a concentration that induces maximal inositol phosphate response (Fig. 1). The cells were then solubilized and subjected to immunoprecipitation with anti-HA antibodies. As shown in Fig. 11, upper panel, when cells coexpressing G q ␣ subunits and HA-WT-CCKBR were treated with CCK, G q ␣ subunits were coimmunoprecipitated with the HA-WT-CCKBR and identified after reaction with the anti-G q ␣/G 11 ␣ antibodies as a band that migrated at 42 kDa. In contrast, in CCK-untreated cells, a low basal association between G q ␣ subunits and HA-WT-CCKBR was observed. The effect of CCK on G q ␣-HA-WT-CCKBR complex formation was 2.8-fold Ϯ 0.8 of the basal level (mean Ϯ S.E. of five experiments). As control, non-transfected cell lysates were immunoprecipitated with anti-HA antibodies and immunoblotted with anti-G q ␣/G 11 ␣ antibodies. Under these conditions, no band corresponding to the G q ␣ subunits was detected (Fig.  11, upper panel). These results revealed that the wild type receptor coupled specifically to G q ␣ subunits when stimulated by CCK. FIG. 4. Effect of GF109203X on Elk-1 transcriptional activity in COS cells transiently expressing the WT-CCKB receptor. As described under "Experimental Procedures", COS-7 cells were cotransfected with pCDL-SR␣-WT-CCKBR, pFA2-Elk-1, pFR-Luciferase, and the pCMV-␤-galactosidase plasmids. A, transfected cells were treated with increasing concentrations of GF109203X for 30 min, exposed to 100 nM PMA for 7 h, and were then assayed for luciferase activity. B, as described above, transfected cells were treated with 5 M GF109203X, a concentration that completely abolished PMA-stimulated luciferase activity, and then stimulated by 10 Ϫ6 M CCK. Luciferase activity, normalized to ␤-galactosidase, is expressed as fold induction over basal (untreated cells). Data represent the mean Ϯ S.E. of three to four separate experiments each performed in duplicate.
Next, we performed similar experiments with COS-7 cells coexpressing the HA-N391A-CCKBR mutant and G q ␣ subunits. When cells were stimulated with 10 nM CCK and subjected to immunoprecipitation with anti-HA antibodies, G q ␣ subunits coimmunoprecipitating with the HA-N391A-CCKBR mutant were revealed with anti-G q ␣/G 11 ␣ antibodies (Fig. 11,  upper panel). A low basal association between G q ␣ subunits and HA-N391A-CCKBR was observed in CCK-untreated cells as with the wild type receptor. The effect of CCK on G q ␣-HA-N391A-CCKBR complex formation was 4.3-fold Ϯ 1.1 of the basal level (mean Ϯ S.E. of five experiments). These results suggest that the mutated receptor forms stable complexes with G q ␣ subunits when stimulated by CCK. In parallel, to verify the amounts of immunoprecipitated wild type and mutated receptors, the top of the membrane was cut after protein transfer and separately immunoblotted with anti-HA antibodies. As shown in Fig. 11, lower panel, equal amounts of epitope-tagged wild type and mutated receptors migrating at a mass of 85 kDa were immunoprecipitated in all assays except in non-transfected cells where no corresponding band was detected.
Together these data suggest an essential and specific role of Asn-391 in the activation process of G q proteins since the N391A-CCKBR mutant retains the capacity to interact with G q ␣ proteins in response to CCK. DISCUSSION According to the importance of the Asn/Asp residue of the highly conserved (N/D)PX 2-3 Y motif in the function of several G-protein-coupled receptors (7, 9 -13), we have investigated whether a similar role could be attributed to the Asn-391 of the NPXXY motif of the CCKBR. Mutation of this residue to Ala had no effect on agonists binding since the N391A-CCKBR mutant conserved a high affinity for CCK-9, CCK-4, and gastrin 17-I indicating that Asn-391 is not involved in agonistbinding sites. Moreover, this mutation produced a gain of affinity of 3-and 10-fold for the selective CCKBR antagonists, L365-260 and PD135-158. Because Asn-391 is located deep inside the transmembrane domain seven, it is likely that this residue is not directly involved in antagonist-binding sites. However, despite the high affinity binding displayed by CCK and selective agonists and antagonists for the mutant, this mutation completely abolished phosphoinositide hydrolysis and MAPK activation. This was not due to a defect of expression of the mutant at the cell surface since similar expression levels were measured for the mutant and wild type receptors.
We have shown in this study using the PKC-specific inhibitor bisindolylmaleimide GF109203X that 65% of CCK-induced p42 and p44 MAPK activation is PKC-dependent. Thus it appears that the remaining MAPK activation is independent of PKC and, accordingly, of PLC. These results are consistent with previous studies performed in other cell lines showing that both PKC-dependent and -independent pathways are involved in MAPK activation mediated by the CCKBR (5,6). The fact that the mutation of Asn-391 inhibits at least two different signaling cascades without affecting the agonist-binding site suggests that this residue has a critical role in the CCKBR activation processes. These results are in agreement with the absence of effects of GTP␥S and AlF 4 Ϫ on CCK binding to the N391A-CCKBR mutant, whereas similar treatment reduced CCK binding to the WT-CCKBR. The reduction in CCK binding observed for the WT-CCKBR in the presence of GTP␥S (or AlF 4 Ϫ ) is attributable to the dissociation of the receptor from ␣GTP␥S (or ␣GDP-AlF 4 Ϫ ) and ␤␥ subunits and thus reflects the presence of G-proteins and their persistent activation. Absence of effects of GTP␥S and AlF 4 Ϫ on CCK binding to the mutant receptor, in addition to its inactivation, could reflect an absence of G-protein association. Alternatively, the mutant could be associated to G-proteins, but the processes of guanine nucleotide exchanges and/or ␣ and ␤␥ subunit dissociation from the mutant receptor could be defective. Such a mechanism is supported by a previous study on the rhodopsin receptor showing that the mutation of two segments in the second and third intracellular loops produced mutants that normally bound G t in response to light but failed to release the bound G t in the presence of guanosine triphosphate (33). Moreover, these mutants were shown to bind the G t ␣␤␥-GDP form of G t that was defective in catalyzing GDP release (34). In order to determine whether the N391A-CCKBR mutant retains the capacity to interact with G-proteins, we performed immunoprecipitation of the mutant receptor and investigated whether G-proteins were coassociated with the mutant by Western blotting. To date, such approach has not been used to evaluate if inactive mutants were coupled to G-proteins in other G-protein-coupled receptors. We looked for the presence of G q proteins in the immunoprecipitates since we showed an increase in CCK-stimulated inositol phosphates response by coexpressing the WT-CCKBR with increasing amounts of G q ␣ subunits. Coupling of the WT-CCKBR to G q proteins was further demonstrated by coimmunoprecipitating the WT-CCKBR with G q ␣ when cells were stimulated by CCK. In similar conditions, we found significantly more G q ␣ subunits associated with the mutated receptor than with the wild type CCKBR (4.3-fold Ϯ 1.1 versus 2.8-fold Ϯ 0.8 of the basal level), indicating that the mutated receptor forms stable complexes with G q ␣ subunits. These re-sults suggest that the high affinity agonist state displayed by the mutated receptor is due to G-proteins association. In addition, the stability of the mutated receptor-G q complexes supports that the inactivity of the mutant likely reflects a defect in the dissociation of ␣ and ␤␥ subunits and would be consistent with the absence of effect of GTP␥S on CCK binding. However, the present study does not allow one to conclude about the state of the G q proteins associated with the mutated receptor that could contained GDP or GTP bound to the ␣ subunit or no nucleotides (empty state). Nevertheless, the fact that this mutation allows stable receptor-G-protein complexes to form but prevents G q activation argues in favor of a specific role of Asn-391 in G q protein activation. In agreement with the important role of Asn-391 in receptor activation, a computational modeling study has reported that the (N/D)P motif introduced a local flexibility in the seventh transmembrane domain of G-protein-coupled receptors that may play a role in receptor activation by functioning as a sensitive conformational switch (15). Consistent with this study, the activation of rhodopsin receptor has been correlated in a recent study with conformational changes at the cytoplasmic end of the seventh transmembrane helix by using an antibody directed against a part of the highly conserved NPXXY motif. This antibody recognized specifically light-induced exposure of the epitope, whereas no binding was detected in the dark. The accessibility of the epitope to the antibody was correlated with the formation of the metarhodopsin II photointermediate, an active state that binds G t and catalyzes guanine nucleotides exchanges. Based on these studies, it can be hypothesized that Asn-391 participates in CCKBR conformational changes that lead to G q activation; however, a direct role of Asn-391 in G q activation cannot be excluded.
In several G-protein-coupled receptors, this conserved Asn (Asp for the gonadotropin-releasing hormone receptor) was reported to interact with another highly conserved Asp (Asn for the gonadotropin-releasing hormone receptor) located in transmembrane domain 2 by performing reciprocal exchange of these two residues to restore activity of the mutants (11,12,35,36). This was not the case in the CCKBR, since mutation of Asn-391 to Asp neither affected CCK binding nor PLC activation, whereas mutation of Asp-100 to Asn decreased by 50% the production of IP, leaving open the possibility of other interactions (4).
In this study, a PKC-independent pathway was found to couple the CCKBR to MAPK activation. However, the molecular entities involved in such a coupling are not well known. Occupancy of the CCKBR expressed in Rat-1 cells by CCK has been shown to elicit MAPK activation through a pathway that was independent of pertussis toxin-sensitive G-proteins and PKC but was dependent on p74 Raf-1 (6). A ␤␥ dimer of a pertussis toxin-insensitive G-protein acting on a Ras-dependent/ PKC-independent pathway was described to be involved in MAPK activation mediated by the muscarinic-1 receptor (29,37). An implication of such ␤␥ dimer in addition to ␣ q subunit would be consistent with the complete loss of PKC-dependent and -independent MAPK activation displayed by the N391A-CCKBR mutant.
In conclusion, in the present study we show that the N391A-CCKBR mutant, which lacks the ability to activate PLC and MAPK but binds agonists with high affinity, is able to form stable complexes with G q in response to CCK, suggesting an essential and specific role of Asn-391 in the activation process of G q proteins. More generally, these results show that the formation of high affinity agonist-receptor-G q complexes is not sufficient to activate G q and thus indicate that specific determinants of the receptor are involved in G q activation, whereas others are involved in G q binding. Since this Asn residue is highly conserved in G-protein-coupled receptors, it would be of interest to determine if a similar role could be extended to others receptors of this family.