Identification of cytoplasmic domains of hVPAC1 receptor required for activation of adenylyl cyclase. Crucial role of two charged amino acids strictly conserved in class II G protein-coupled receptors.

The VPAC1 receptor mediates the action of two neuropeptides, vasoactive intestinal peptide (VIP) and pituitary adenylate cyclase-activating peptide. It is a class II G protein-coupled receptor-activating adenylyl cyclase (AC). The role of the N-terminal extracellular domain of hVPAC1 receptor for VIP binding is now established (Laburthe, M., Couvineau, A. and Marie, J. C. (2002) Recept. Channels 8, 137-153), but nothing is known regarding the cytoplasmic domains responsible for AC activation. Here, we constructed a large series of mutants by substituting amino acids with alanine in the intracellular loops (IL) 1, 2, and 3 and proximal C-terminal tail of the receptor. The mutation of 40 amino acids followed by expression of mutants in chinese hamster ovary cells showed the following. (i) Mutations IL1 result in the absence of expression of mutants, suggesting a role of this loop in receptor folding. (ii) All residues of IL2 can be mutated without alteration of receptor expression and AC response to VIP. (iii) Mutation of residues IL3 points to the specific role of lysine 322 in the efficacy of the stimulation of AC activity by VIP. This efficacy is reduced by 50% in the K322A mutant. (iv) The proximal C-terminal tail is equipped with another important amino acid since mutation of glutamic acid 394 reduces AC response by 50%. The double mutant K322A/E394A exhibits a drastic reduction of >85% in the efficacy of VIP in stimulating AC activity in membranes and cAMP response in intact cells without alteration of receptor expression or affinity for VIP. These data highlight the role of charged residues in IL3 and the proximal C-terminal tail of hVPAC1 receptor for agonist-induced AC activation. Because these charged residues are absolutely conserved in class II receptors for peptides, which are all mediating AC activation, they may play a general role in coupling of class II receptors with the Gs protein.

The VPAC1 receptor is a common receptor for vasoactive intestinal peptide (VIP) 1 and pituitary adenylate cyclase-acti-vating peptide or PACAP (1). It mediates the actions of the two neuropeptides in a large variety of tissues through coupling to G s proteins (2) and subsequent activation of adenylyl cyclase (3). Together with the VPAC2 receptor subtype, they mediate a large array of VIP or PACAP actions on exocrine secretions, release of hormones, relaxation of muscles, metabolism, immune functions, growth control of fetuses and tumor cells, and embryonic brain development (1)(2)(3)(4)(5). The VPAC1 receptor belongs to the class II subfamily of G protein-coupled receptors (GPCR), which includes receptors for peptides structurally related to VIP, receptors for other peptides such as the PTH, calcitonin, or CRF (6,7), and the so-called EGF-TM7 (8) or LNB-TM7 (9) receptors bearing unusually large and complex N-terminal extracellular domains.
The VPAC1 receptor, a prototypical class II receptor, has been extensively studied by site-directed mutagenesis and molecular chimerism with respect to determination of VIP binding domains (1)(2)(3) and agonist selectivity (10,11). These studies showed that the N-terminal extracellular domain of hVPAC1 receptor plays a crucial, although not exclusive, role in VIP binding (12). Moreover, microdomains consisting of small clusters of amino acids located in the N-terminal ectodomain (11) or at the junction of extraloop 1 and TM3 (10) were characterized as selectivity filters restricting access of VIP-related peptides to the hVPAC1 receptor. A three-dimensional model of the N-terminal ectodomain of hVPAC1 receptor has been recently developed (12). In sharp contrast, nothing is known regarding the regions of the intracellular domain of the hVPAC1 receptor involved in signal transduction e.g. adenylyl cyclase activation. More generally, the current knowledge of the coupling mechanism of class II GPCRs to signal transduction is limited as compared with that of class I GPCRs including the prototypical rhodopsin and ␤-adrenergic receptors (13). Since class II peptide receptors have low overall sequence identity with class I GPCRs and even lack the rigorously conserved residues found in the core of class I receptors including the (D/E)RY sequence, the determination of the class II receptor intracellular domains involved in coupling to signal transduction cannot be simply extrapolated from our current knowledge regarding class I GPCRs.
In this context, the present study explores the structurefunction relationship of the hVPAC1 receptor for VIP-induced adenylyl cyclase activation. For this purpose, we constructed a large series of mutants by substituting groups of amino acids or individual amino acids with alanine in the intracellular domains of the receptor including IL1, IL2, IL3, and the proximal C-terminal tail. The mutation of as many as 40 amino acids provides evidence that two charged residues in IL3 and the C-terminal tail are mandatory for coupling to adenylyl cyclase activation.

EXPERIMENTAL PROCEDURES
Materials-Restriction enzymes and culture medium were obtained from Invitrogen (Cergy-Pontoise, France). Eukaryotic expression vector was from Clontech (Palo Alto, CA). DNA Sequenase kit and radioactive reagents were from Amersham Biosciences. The site-directed mutagenesis kit was from Promega. Synthetic VIP was from Neosystem (Strasbourg, France). 125 I-VIP and 125 I-cAMP were prepared and purified in our laboratory as previously described (14). Synthetic oligonucleotides were from Invitrogen. The hVPAC1 receptor cDNA was cloned in our laboratory (15). A receptor construct containing an inserted FLAG sequence between Ala 30 and Ala 31 and fused in the C-terminal position with the green fluorescent protein (16) was used in all studies. It has the same phenotype as the wild-type native receptor with respect to VIP binding and activation of adenylyl cyclase (16). Other highly purified chemicals used were from Sigma. Modeler v6.0 (17) software was downloaded from salilab.org/modeler/. Graphical representation of the threedimensional model of the hVPAC1 receptor was performed with VMD software (www.ks.uiuc.edu/Research/vmd/).
Site-directed Mutagenesis Experiments-Oligonucleotide-directed mutagenesis was performed as previously described (10). Identification of the desired mutations was obtained by direct sequencing of the mutated region (10). Correct expression of each construction was determined par direct fluorescence of GFP on living transfected CHO cells as described (17).
Cell Transfection-Wild-type and mutated hVPAC1 receptors were stably transfected into CHO cells as previously described using 3 l of FuGENE 6 (Roche Applied Science) and 2 g of DNA construct. After 48 h, GFP fluorescence was used to check receptor expression, and transfected CHO cells were selected in the presence of Geneticin (G418) at a final concentration of 0.8 g/liter for 5 days, then grown in absence of G418 for 1 week, and used for cAMP measurement or membrane preparation as previously described (14).
Ligand Binding, Adenylyl Cyclase, and Intracellular cAMP Assays-The binding properties of wild-type and mutated hVPAC1 receptor were determined by competitive inhibition of 125 I-VIP binding to transfected cell membranes by unlabeled VIP as in Ref. 10. Briefly, membranes (200 g of protein) prepared from transfected cells were incubated for 60 min at 30°C in 20 mM HEPES buffer, pH 7.4, containing 2% (w/v) bovine serum albumin and 0.05 nM 125 I-VIP in the presence of increasing concentrations of unlabeled VIP. Nonspecific binding was determined in the presence of an excess of unlabeled VIP (1 M). The specific binding was calculated as the difference between 125 I-peptide totally bound and the nonspecific binding. Binding parameters (K d and B max ) were analyzed using Prism software (GraphPad Software, San Diego, CA). Adenylyl cyclase activity in membranes was measured in the presence of increasing concentrations of native VIP as described (14). Dose response curves were fitted, and concentrations of VIP giving half-maximal responses (EC 50 ) were calculated using the Prism Software. The efficacy of VIP in stimulating adenylyl cyclase for mutants is defined as (VIP (10 Ϫ6 M)-stimulated adenylyl cyclase activity for mutant/VIP (10 Ϫ6 M)-stimulated adenylyl cyclase activity for wild-type receptor) ϫ 100. Intracellular cAMP was assayed as described (11). Briefly, stably transfected CHO cells were grown in 12-well trays, the culture medium was discarded, and adherent cells were gently rinsed with phosphate-buffered saline. They were then incubated without or with VIP under continuous agitation in 0.5 ml of phosphate-buffered saline containing 2% (w/v) bovine serum albumin, 0.1% (w/v) bacitracin, and 1 mM 3-isobutyl-1-methylxanthine as described (11). At the end of incubation (30 min at 25°C) the medium was removed and cells lysed with 1 M perchloric acid. The cAMP present in the lysate was measured by radioimmunoassay as described (11). Cell number was determined in parallel wells, and data were calculated as pmol of cAMP/10 6 cells.
Internalization of Wild-type Receptor and K322A/E394A Mutant in CHO Cells-Subconfluent CHO cells expressing wild-type receptor or K322A/E394A mutant were grown in 24-well plates. Cells were treated with 5 nM VIP or not (control) for 2 h at 37°C. After treatment, cells were washed twice with 1 ml of phosphate-buffer saline and subse-quently used for binding or immunofluorescence studies. Binding parameters (K d and B max ) were determined by Scatchard analysis. Briefly, intact cells were incubated for 60 min at 25°C in 20 mM HEPES buffer, pH 7.4, containing 2% (w/v) bovine serum albumin, and 0.05 nM 125 I-VIP in the presence of increasing concentrations of unlabeled VIP, washed with 1 ml of phosphate-buffer saline, solubilized with 300 l of 0.25 M NaOH for 10 min at 25°C, and the amount of radioactivity was determined in a ␥-counter. Specific binding was calculated as the difference between the amount of bound 125 I-VIP in absence and in presence of 1 M unlabeled VIP. Cell number was determined in parallel wells, and data were calculated as number of sites/cell. The GFP fluorescence in living cells and immunofluorescence studies on stably transfected CHO cells using anti-FLAG antibodies were determined as described (12,16). Briefly, the cells were incubated for 30 min with the primary FLAG mouse monoclonal antibody and then with the secondary antibody (TRITC-sheep antimouse IgG). Selected fields were examined on a LEICA inverted DM IRB microscope using a (ϫ40) fluor oil immersion objective and the two filters N2.1 (515-590-nm bandpass) and L5 (480 -527-nm bandpass) for GFP and TRITC fluorescence, respectively. In the conditions of our studies with nonpermeabilized living cells, the anti-FLAG antibodies, which recognize a FLAG inserted in the N-terminal extracellular domain of the receptor, only bind to receptor constructs expressed at the plasma membrane (12,16).
Molecular Modeling-A three-dimensional model of the hVPAC1 receptor (136 -411) consisting of the seven transmembrane domains, three extra-and intracellular loops, and the first 18 amino acids of the C-terminal tail was built by comparative modeling using the recently determined 2.8-Å x-ray structure of rhodopsin as template (18). Briefly, preliminary alignment between the hVPAC1 and rhodopsin primary sequences was performed using a Clustal algorithm (19). In a first step, 53 sequences corresponding to class II receptors including VPAC1, VPAC2, PAC1, secretin, GRF, PTH, CRF, GLP-1, GLP-2, glucagon, GIP, and calcitonin receptors from several species (www.gpcr.org/7tm/) were aligned. In parallel, 83 sequences corresponding to rhodopsin from different species were aligned. These two alignments were used to select five members of class II receptors and opsins sharing the highest conserved residues within each family. The next step consisted of aligning bovine OPSD (PDB 1F88) and human OPSB, OPSG, OPSR, OPSD (NCB accession P03999, P04001, P04000, P08100, respectively) and human VPAC1, VPAC2, secretin, PAC1, GRF receptors (NCB accession P32241, P41587, P47872, P41586, Q02643, respectively). In order to refine this alignment, the HCA (hydrophobic cluster analysis) plots of these sequences were compared using VisualFasta software (Patrick Durand, Paris, France). Briefly, the hydrophobic clusters visualized by a two-dimensional helical plot of the sequences detecting the hydrophobic domains of these proteins were aligned in order to make the seven transmembrane domains coincident. The quality of final alignment between the hVPAC1 receptor and bovine rhodopsin (PDB 1F88) is evaluated by identity score (14%), homology score (61%), and HCA score (64%). The resulting alignment was used as input for Modeler v6.0 (16). Two hundred models were generated and ranked according to the Modeler objective function, and only the first model was used.

RESULTS
In order to determine the molecular motif(s) responsible for hVPAC1 receptor coupling to adenylyl cyclase as many as 40 amino acids were mutated into alanine in the intracellular domains of the receptor as previously defined by the Kyte-Doolittle hydropathicity plots (20). Mutated domains include IL1, IL2, IL3, and the proximal C-terminal tail of the receptor (Fig. 1). Depending on the context, mutations have been carried out per groups of amino acids or individually. Regarding IL1, which is a short loop, two blocks of amino acids were mutated resulting in mutants IL11 and IL12 (Table I). After transfection of CHO cells with the corresponding cDNAs, no binding of 125 I-VIP could be detected. Nor was there any ability of VIP to stimulate adenylyl cyclase activity whereas CHO cells expressing the wild-type receptor responded nicely to VIP (Table I).
Since IL1 is very short in the hVPAC1 receptor, we reasoned that mutations may have grossly altered receptor structure resulting in problems in expression of mutants. Indeed, no GFP fluorescence could be detected after transfection of IL11 and IL12 mutants (not shown) supporting that the mutated proteins are not expressed in transfected cells. Next, substitution hVPAC1 Receptor Cytoplasmic Domains of all residues of the IL2 has been performed by constructing three mutants IL21, IL22, and IL23 ( Fig. 1 and Table I). All three mutants expressed in CHO cells bound 125 I-VIP and Scatchard analysis of binding data revealed affinities similar to that of the wild-type hVPAC1 receptor. It should be also noted that the B max of mutants expressed in CHO cells is in the range between 461 and 580 fmol/mg of protein comparing well with the B max of the wild-type receptor ( Table I). The three mutants were further investigated for their ability to mediate stimulation of adenylyl cyclase by VIP in membrane preparations. The responses were identical to that mediated by the wild-type receptor both in terms of EC 50 and efficacy (Table I). These data are consistent with the fact that IL2 is not involved in the functional coupling of hVPAC1 receptor to adenylyl cyclase.
The role of residues present in the intracellular loop 3 was investigated next. This loop is large with 26 amino acids (see Fig. 1). In order to limit the number of mutants, we searched for conserved residues and aligned the amino acid sequences of IL3 from all human G s -coupled receptors having high (VPAC1 and VPAC2 receptors) or very low (PAC1, secretin, and GRF receptors) affinity for VIP (1). As shown in Table II, the sequence alignment reveals the presence of eight strictly conserved residues and an arginine versus lysine exchange (Arg 341 in the hVPAC1 receptor). These nine amino acids were individually mutated into alanine and the phenotypes of mutants analyzed after transfection in CHO cells (Table II). All mutants are functionally expressed in CHO cells as assessed by their B max determined in VIP binding experiments. All mutants but one have a K d in the nanomolar range like the wild-type receptor (Table II). The notable exception is the K322A mutant, which has a low affinity for VIP (K d ϭ 23.4 Ϯ 4.0 nM). The adenylyl cyclase assay provided data in agreement with the above described binding data. Indeed, the only mutant that differed from the wild-type receptor was again the K322A mutant, which exhibited an increased EC 50 and a significant 50% decrease of the VIP efficacy in stimulating cyclase activity (Table II). For some mutants, a slight decrease in binding affinity was observed as compared with the wild-type receptor but full efficacy and potency was maintained for VIP-stimulated adenylyl cyclase activity (Table II). This may be tentatively ascribed to differences in the binding and adenylyl cyclase assays.
Then, we investigated the role of the proximal C-terminal tail of the hVPAC1 receptor in receptor coupling to adenylyl cyclase. We substituted alanine in the first 20 residues in this domain. First, four mutants were constructed in which blocks of 3 or 4 amino acids were mutated to alanine residues (Table  III). All constructs exhibited binding parameters (K d and B max ) that are very similar to those of the wild-type receptor. This is in line with the similar EC 50 of VIP in stimulating adenylyl cyclase activity in CHO cells expressing the four mutants and wild-type receptors (Table III). However a significant alteration of efficacy was observed for the mutant CT1 in which alanine was substituted for the first three amino acids in the C-terminal tail. The efficacy of VIP in stimulating adenylyl cyclase activity was significantly decreased in CHO cells expressing this mutant as compared with cells expressing wildtype receptor (Table III). Since three amino acids were mutated in the CT1 construct, further mutants were constructed in order to determine which amino acid residue is reponsible for the low efficacy of the adenylyl cyclase response to VIP (Table  III). Mutation of glutamate 394 alone (mutant CT11) or of both glutamate 394 and valine 395 (mutant CT12) also resulted in a 50% decrease in the efficacy of VIP highlighting the crucial role of glutamate 394. It is again important to notice that these two latter mutants have binding parameters identical to those of the wild-type receptor, including B max (Table III). As a control, we also mutated glutamine 396 alone. The Q396A mutant has clearly the same phenotype as the wild-type receptor with respect to VIP binding and VIP-stimulated adenylyl cyclase activity (Table III). Other mutants CT2, CT3, and CT4 exploring other parts of the proximal C-terminal tail did not show any alteration of VIP efficacy (Table III).
The mutants K322A in IL3 and E394A in the proximal C-terminal tail exhibited similar phenotype with respect to VIP efficacy in stimulating adenylyl cyclase efficacy, e.g. 50% decrease. In this context, we investigated the combination of the two mutations in the new mutant K322A/E394A hVPAC1 receptor. After expression of this mutant in CHO cells, the efficacy of VIP in stimulating adenylyl cyclase was dramatically decreased by Ͼ80% as compared with control experiments with the wild-type receptor (Fig. 2). Binding isotherms for the K322A, E394A, K322A/E394A mutants versus wild-type receptor are shown in Fig. 2. Surprisingly, the double substitution does not affect the affinity of this mutant to bind VIP (K d ϭ 0.97 Ϯ 0.32 nM) as compared with the wild-type hVPAC1 receptor (K d ϭ 0.36 Ϯ 0.02 nM) while the single mutation of Lys 322 into alanine results in a significant decrease of the affinity for VIP (see above and Table II). Again, it was verified that receptor concentration in transfected cell membranes was similar for the double mutant (B max ϭ 483 Ϯ 54 fmol/mg of protein) and the wild-type receptor (B max ϭ 561 Ϯ 95 fmol/mg of protein). Therefore the dramatic reduction of VIP efficacy in stimulating adenylyl cyclase is clearly not related to a low expression of the double mutant in transfected cells. Further, we tested the ability of VIP to stimulate cAMP production in living CHO cells after transfection of the double mutant. It appeared clearly that under such more physiological conditions, the efficacy of VIP is considerably reduced by 85% as compared with that observed for the wild-type receptor (Fig. 3). The dramatic uncoupling of the K322A/E394A hVPAC1 receptor mutant to adenylyl cyclase is therefore not related to the artificial conditions of the incubation medium in the adenylyl cyclase assay since it can be observed in living cells. We further characterized the phenotype of this mutant by exploring its internalization upon VIP treatment of transfected cells. After a 2-h incubation period with 5 nM VIP (see "Experimental Procedures") the K322A/E394A hVPAC1 receptor mutant and the wild-type receptor underwent similar internalization as assessed by Scatchard analysis of binding data and immunofluorescence experiments carried out on viable nonpermeabilized cells using anti-FLAG antibodies (Fig. 4). It can be therefore concluded that the double mutation K322A/E394A specifically affects the coupling of the hVPAC1 receptor to adenylyl cyclase.

hVPAC1 Receptor Cytoplasmic Domains
Our mutagenesis approach was based on the secondary structure of hVPAC1 receptor as defined by hydropathicity plot of the primary sequence. We tried to further analyze the location of the two crucial residues Glu 394 and Lys 322 by constructing a homology three-dimensional model of hVPAC1 receptor core including loops based on the x-ray structure of bovine rhodopsin (18). The large extracellular N-terminal domain of the hVPAC1 receptor, which is crucial for VIP binding (1) and has no counterpart in rhodopsin, was of course excluded from the model. Despite the observation that rhodopsin is a class I receptor and hVPAC1 a class II receptor, an initial sequence alignment obtained with the ClustalW algorithm can be refined by the HCA of the two proteins (see "Experimental Procedures"). The HCA plot revealed a good alignment of transmembrane domains between hVPAC1 and rhodopsin with an overall 61% homology of amino acid sequences (see Fig. 5, top). A three-dimensional model of hVPAC1 (sequence 136 -411) receptor core was built using Modeler 6.0 (16) and refined by energy minimization. The local root mean square distance between the hVPAC1-(136 -411) three-dimensional model ob-tained and bovine rhodopsin is 1.73 Å, indicating similar geometrical parameters for the two proteins. The hVPAC1 receptor model (see Fig. 5, bottom left) shows the location of Glu 394 and Lys 322 , which are crucial for coupling the receptor to adenylyl cyclase activation. A distance of 17 Å between C␣ of the two residues can be calculated supporting the idea that the side chains of Glu 394 and Lys 322 are not interacting with each other, but probably with the G s protein. DISCUSSION In this work we demonstrate that two charged amino acids, Lys 322 in the third intracellular loop and Glu 394 in the proximal C-terminal tail of the hVPAC1 receptor are crucial for VIPinduced adenylyl cyclase activation. Indeed, their joint mutation into alanine drastically reduces hVPAC1-mediated adenylyl cyclase response without altering receptor expression and VIP binding. Since these amino acids are rigorously conserved in all class II GPCR for peptides (see Fig. 5, bottom right), their charge and location may prove to be important for coupling of this class of receptors to the G s protein.

hVPAC1 Receptor Cytoplasmic Domains
We have investigated all intracellular domains of the hVPAC1 receptor by substituting blocks of alanine or individual alanine to amino acids present in the native sequence. Our results are initially discussed below in the context of class I and class II GPCRs, while considering each domain.
(i) With regard to IL1, which is predicted to be a very short loop from the Kyte-Doolittle plot of the hVPAC1 sequence and even shorter in the rhodopsin structure, our data indicated that mutations in this loop result in the absence of expression of mutated proteins in transfected CHO cells. This suggests that residues in IL1 play a role in receptor protein folding in consonance with data reported with another class II receptor, the glucagon receptor (21) and also with the high sequence conservation of this loop in class II receptors (2). Although the absence of expression of IL1 mutants of hVPAC1 receptor does not allow us to conclude the role of IL1 in coupling to adenylyl cyclase activation, studies of glucagon receptor suggested that IL1 is important for protein folding but not for direct receptor-G protein coupling (21). Neither has IL1 a direct role in G protein coupling in the class I GPCRs (22). In this context, the fact that insertion of Ͼ10 amino acid peptides in IL1 in alternative spliced variants of some class II receptors, including the calcitonin receptor (23) or CRF receptor (24), affects G protein coupling may be due to global tertiary structure alteration of those receptors. We previously showed that the mutation of His 178 into arginine at the junction of IL1 and TM2 results in the constitutive activation of the hVPAC1 receptor (17). This histidine is highly conserved in class II receptors but its mutation into arginine results in constitutive G s activation in PTH1 receptor (25) but not in other class II GPCRs (26). As previously suggested (21), this histidine at the junction of IL1 and TM2 may be important for some structural component of receptor-G protein coupling but is probably not involved in direct receptor-G protein interaction.
(ii) All the amino acid residues of IL2 have been mutated into alanine in the hVPAC1 receptor without any alteration in receptor phenotype with respect to expression, affinity for VIP, and potency and efficacy of VIP in stimulating adenylyl cyclase activity. This supports the idea that IL2 in the hVPAC1 receptor is not important for receptor folding and function including coupling to G s . This contrasts markedly with the important role of the IL2 in class I receptors, which appear to function as a  hVPAC1 Receptor Cytoplasmic Domains switch enabling G protein activation (27,28). In particular the (D/E)RY sequence located close to the junction of TM3 with IL2 is intimately involved in G protein activation of class I GPCRs (29). Of particular significance in this context is the absence of (D/E)RY sequence in hVPAC1 receptor as well as all class II GPCRs (2,21). With regard to class II GPCRs few data are available but this loop appears to play a role in glucagon receptor signaling as measured by both cAMP accumulation and calcium flux assays (21). Mutations in IL2 of the rat PTH1 receptor result in selective loss of PTH-stimulated phospholipase C activity without affecting PTH-stimulated cAMP response (30). Mutation of a conserved glutamic acid in IL2 of the PAC1 receptor for PACAP has been shown to confer constitutive receptor activation with respect to cAMP production (31). Mutation into alanine of the corresponding glutamic acid in IL2 of hVPAC1 receptor does not confer constitutive activation. 2 The role of IL2 in class II GPCRs coupling to G proteins is therefore not consistent.
(iii) IL3 is the largest intracellular loop in the hVPAC1 receptor comprising 26 amino acids (see Fig. 1) as determined from Kyte-Doolittle hydropathicity plots. In order to limit the number of mutants, we aligned the amino acid sequences (see Table III) of IL3 from all human G s -coupled receptors having high (VPAC1 and VPAC2 receptors) or very low (PAC1, secretin, and GRF receptors) affinity for VIP (1). The mutation of the nine strictly conserved residues in IL3 in this receptor group showed only one hit with the K322A mutant. This mutant exhibits a decrease of affinity for VIP and a 50% reduction of the efficacy of VIP in stimulating adenylyl cyclase activity (see Table II). This is consistent with many observations indicating that the IL3 of class I receptors interacts with G␣ and G␤ subunits of G proteins and is important for signal transduction (Ref. 29 for review; Ref. 32). Regarding class II peptide GPCRs, IL3 appears to be important in glucagon receptor (21), GLP-1 receptor (33), secretin receptor (34), and PTH1 receptor (35) signaling. Moreover, splice variants of the specific PACAP receptor PAC1 have been described with insertion of 28 -56amino acid cassettes within IL3 (36). Not unexpectedly, important modifications of signal transduction were reported for the ability of PACAP to stimulate cAMP production and phospholipase C response (36). No similar splice variants have been reported yet for the hVPAC1 receptor (1). It is important to point out that Lys 322 in IL3 of the hVPAC1 receptor is absolutely conserved in all class II peptide receptors (see Fig. 5, bottom right).
(iv) The proximal C-terminal tail of hVPAC1 receptor is clearly equipped with another charged amino acid involved in receptor signaling since mutation of Glu 394 into alanine reduces cAMP response by 50% (see Table III). This glutamic acid residue is strictly conserved in all class II peptide GPCRs (see Fig. 5, bottom right). Compared with IL2 or IL3, the role of the C-terminal tail in G protein receptor coupling is less consistent in the superfamily of GPCRs even for rhodopsin-like receptors belonging to the class I subfamily. Indeed, some class I receptors use the proximal C-terminal tail for G protein coupling (37) whereas others activate G protein even though they are not equipped with a C-terminal tail (38). With regard to class II GPCRs, there is evidence for a role of the C-terminal tail in signal transduction for the porcine calcitonin receptor (39) and the human PAC1 receptor for PACAP (40). The C-terminal tail of the rat glucagon receptor was claimed not to be necessary for adenylyl cyclase coupling (41). However, this statement was based on a truncation approach, which leaves a short proximal 10 amino acid sequence of the C-terminal tail (41). This short tail includes the highly conserved glutamic acid residue demonstrated to be very important in the present study. It is therefore possible that the proximal C-terminal tail of the rat glucagon receptor does play an important for coupling to G s 2 M. Laburthe, unpublished observations. After exposure of cells to VIP, an important internalization of both wild-type and mutant receptors occurred since they could be barely TRITC-labeled as compared with control cells. Further, in the same field are shown GFP-labeled cells (panel A) due to the presence of GFP at the C terminus of wild or mutated VPAC 1 receptors. This GFP labeling is representative of total cellular expression of these receptors. VIP induced a similar receptor internalization for wild-type and mutated receptors as seen by the extinction of TRITC-labeled versus GFP-labeled cells.
protein. Alternative splicing that deletes exon 13, resulting in the absence of 14 amino acids in the predicted seventh transmembrane domain of the rabbit calcitonin receptor has been reported (42). This CTR ⌬e13 isoform lacks the highly conserved glutamic acid residue, which is important in hVPAC1 receptor. Interestingly, it still activates adenylyl cyclase but does not activate phospholipase C as compared with the C1a isoform with no deletion of exon 13 (42). This suggests that the junction between TM7 and C-terminal tail may also participate in the specificity of coupling to G proteins.
The present study highlights the crucial role of Lys 322 in IL3 and Glu 394 in the proximal C-terminal tail of the hVPAC1 receptor. In the absence of structural data regarding class II GPCRs, this observation cannot be further discussed in the context of this receptor subfamily apart from the fact that both amino acids are rigorously conserved in all members of this . ଙ is Pro, ᭜ is Gly, Ⅺ is Thr, and is Ser. Hydrophobic residues (Leu, lle, Val, Trp, Tyr, and Phe) are circled to delimitate hydrophobic clusters in the protein (gray areas). Vertical boxes represent the location of transmembrane domains (TM 1-7) of bovine rhodopsin and hVPAC1 receptor. Bottom left, ribbon representation of the three-dimensional model of the hVPAC1 receptor (136 -411). Gray color shows extracellular loops, intracellular loops, and C-terminal domain. Dark blue is TM1; light blue is TM2; green is TM3; yellow is TM4; orange is TM5; red is TM6; purple is TM7; and light brown is an eight helix along the membrane surface. Dark pink spacefills represent strictly conserved Lys 322 and Glu 394 residues, which are crucial for coupling the hVPAC1 receptor to adenylyl cyclase activation. Bottom right, alignment of lL3 region (318 -326) and proximal C-terminal (CT) domain (390 -398) of hVPAC1 receptor with corresponding domains in various class II GPCRs. The crucial Lys and Glu residues are highlighted in red. hVPAC1 Receptor Cytoplasmic Domains receptor family (see Fig. 5, bottom right). This suggests that they play a general role in the transduction signaling of class II receptors, at least regarding G s coupling, which is a common property of these receptors (1). Much more data are available for rhodopsin-like receptors (see Ref. 29; Ref. 43 for reviews). The crucial role of Lys 322 in IL3 of hVPAC1 receptor is in accordance with the well recognized importance of basic residues in IL3 of the m1 (44) or m3 (45) muscarinic receptors. It also fits well with recent covalent cross-linking experiments demonstrating physical interaction between IL3 of rhodopsin and the C-terminal sequence of the transducin ␣ subunit (46). The role of Glu 394 in the proximal C-terminal tail of hVPAC1 receptor is in accordance with the ability of the corresponding region of rhodopsin to constitute a portion of the binding domain for the ␤/␥ complex of transducin subunits (47). In this context, it is interesting to notice that in our model of hVPAC1 receptor the 21 amino acid residues after TM7 form an ␣-helix along the surface of the cell membrane (see Fig. 5, bottom left) similar to the crystal structure of rhodopsin (18). This is not only the trivial result of homology modeling since ab initio analysis of this region in the hVPAC1 receptor also provides clear evidence for the formation of an eighth ␣-helix. 2 However, a major difference between hVPAC1 and class II receptors with rhodopsin-like receptors is the absence of cysteine residues (1), the palmitoylation of which was shown to anchor this helix to the cell membrane (see Ref. 13 for review). Another important difference is the role of IL2. While this loop appears to provide contact points of class I receptors with the C terminus of ␣-subunits of G proteins (43), there is no evidence for a role for the residues present in IL2 of hVPAC1 receptor for coupling to the G s protein.
In conclusion, the hVPAC1 receptor, as a member of the class II GPCRs, is coupled to the adenylyl cyclase transduction pathway through charged amino acids in IL3 and the proximal C-terminal tail but probably does not need specific residues of IL2 for signal transduction. Therefore, it appears that the hVPAC1 receptor follows a paradigm reminiscent of, but not identical to, that of the superfamily of class I GPCRs. Differences may be tentatively related to the absence of (D/E)RY sequence in IL2 and of palmitoylation site in the C-terminal tail of class II receptors including hVPAC1.