Contribution of the Carboxyl Terminus of the VPAC1 Receptor to Agonist-induced Receptor Phosphorylation, Internalization, and Recycling*

When exposed to vasoactive intestinal peptide (VIP), the human wild type VPAC1 receptor expressed in Chinese hamster ovary (CHO) cells is rapidly phosphorylated, desensitized, and internalized in the endosomal compartment and is not re-expressed at the cell membrane within 2 h after agonist removal. The aims of the present work were first to correlate receptor phosphorylation level to internalization and recycling, measured by flow cytometry and in some cases by confocal microscopy using a monoclonal antibody that did not interfere with ligand binding, and second to identify the phosphorylated Ser/Thr residues. Combining receptor mutations and truncations allowed identification of Ser250 (in the second intracellular loop), Thr429, Ser435, Ser448 or Ser449, and Ser455 (all in the distal part of the C terminus) as candidates for VIP-stimulated phosphorylation. The effects of single mutations were not additive, suggesting alternative phosphorylation sites in mutated receptors. Replacement of all of the Ser/Thr residues in the carboxyl-terminal tail and truncation of the domain containing these residues completely inhibited VIP-stimulated phosphorylation and receptor internalization. There was, however, no direct correlation between receptor phosphorylation and internalization; in some truncated and mutated receptors, a 70% reduction in phosphorylation had little effect on internalization. In contrast to results obtained on the wild type and all of the mutated or truncated receptors that still underwent phosphorylation, internalization of the severely truncated receptor was reversed within 2 h of incubation in the absence of the agonist. Receptor recovery was blocked by monensin, an endosome inhibitor.

The neuropeptide vasoactive intestinal polypeptide (VIP) 1 exerts its multiple regulatory functions through interaction with two high affinity receptors named VPAC 1 and VPAC 2 . These are members of a family of G protein-coupled receptors (GPCRs), designated as Class II or B. This class also includes, among others, receptors for peptides of at least 20 amino acid residues like secretin, glucagon, glucagon-like peptides, growth hormone-releasing peptide, parathormone, and pituitary adenylate cyclase-activating peptide (1). VPAC 1 and VPAC 2 receptors are preferentially coupled to the G␣ s protein (1) responsible for increasing cyclic AMP concentrations but may also, with a lower efficiency, couple to G␣ i and G␣ q proteins (2) responsible for a [Ca 2ϩ ] i and inositol 1,4,5-trisphosphate increase. As with most, if not all, of the GPCRs, both VIP receptors are desensitized, sequestered, and down-regulated after exposure to agonist (3)(4)(5). This was observed in cells expressing native receptors as well as in transfected Chinese hamster ovary (CHO) cells and HEK 293 cells. It was recently demonstrated that VPAC 1 receptor phosphorylation and desensitization was enhanced by co-transfection with the G protein receptor kinases, GRK2, -3, -5, and -6 (5). Although the overexpression of arrestin or of a dominant negative mutant did not modify receptor internalization, the inhibitory effect of a dominant negative mutant of dynamin suggested the following sequence of events for receptor regulation: agonist stimulation, G protein kinase-mediated phosphorylation, ␤-arrestin translocation, and dynamin-dependant receptor internalization (5).
In the present work, we detailed the contribution of the carboxyl-terminal intracellular tail to receptor internalization by studying truncated and mutated human VPAC 1 receptors expressed in CHO cells. We developed a monoclonal antibody against the amino-terminal extracellular part of that receptor that permitted the quantification, by flow cytometry, of the receptors expressed at the cell membrane. Immunoprecipitation of the receptor after metabolic labeling with [ 32 P]orthophosphate, followed by SDS-PAGE and autoradiography, allowed for phosphorylation quantification. We therefore evaluated the link between receptor phosphorylation and receptor internalization. We also evaluated the recycling of the receptors to the membrane.
We found that VIP induced in the wild type (WT) receptor a rapid phosphorylation and internalization, which was not reversible within 2 h. Mutation into Ala or truncation of all of the Ser/Thr residues in the C-terminal tail and mutation of one Ser in the second intracellular loop abolished receptor phosphoryl-ation and internalization. A larger truncation of a domain located between the seven transmembrane helix and the Ser/Thr-containing region led to a receptor that was no longer phosphorylated but remained internalized. However, receptors were re-expressed at the membrane within 120 min. Single and combined mutations of the Ser/Thr residues indicated the possibility of alternative phosphorylation sites in mutant receptors and also indicated that phosphorylation of all of the identified sites was not necessary for receptor internalization.

Construction of Truncated and Mutated Receptors
The cell line expressing the VPAC 1 receptor has been detailed in a previous publication (6). Generation of the truncated receptors was achieved by introduction of a stop codon using the QuikChange sitedirected mutagenesis kit (Stratagene, La Jolla, CA) essentially according to the manufacturer's instructions as described (2). The expected mutation was confirmed by DNA sequencing on an ABI automated sequencing apparatus, using the BigDye Terminator Sequencing Prism Kit from ABI (PerkinElmer Life Sciences). The complete nucleotide sequence of each construction was verified by DNA sequencing. 20 g of the receptor-coding region were transfected by electroporation in the CHO cell line expressing aequorin and G␣ 16 (kindly provided by Vincent Dupriez, Euroscreen SA, Belgium) as described (2). Selection was carried out in culture medium (50% Ham/F-12, 50% Dulbecco's modified Eagle's medium, 10% fetal calf serum, 1% penicillin (10 milliunits/ml), 1% streptomycin (10 g/ml), 1% L-glutamine (200 mM)), supplemented with 600 g of Geneticin (G418)/ml of culture medium. After 10 -15 days of selection, isolated colonies were transferred to 24-well plates and grown until confluence, trypsinized, and further expanded in 6-well plates, from which cells were scraped and membranes were prepared for identification of receptor-expressing clones by an adenylate cyclase activity assay in the presence of 1 M VIP. The selected clones were expanded in the same medium as that used for the selection but in the absence of Geneticin.

Membrane Preparations
Membranes were prepared from scraped cells lysed in 1 mM NaHCO 3 by immediate freezing in liquid nitrogen. After thawing, the lysate was first centrifuged at 4°C for 5 min at 400 ϫ g, and the supernatant was further centrifuged at 20,000 ϫ g for 15 min. The pellet was resuspended in 1 mM NaHCO 3 and used immediately.

Binding Studies
Binding studies, using 125 I-labeled VIP, were performed for 30 min at 23°C in a total volume of 120 l containing 20 mM Tris-maleate, 2 mM MgCl 2 , 0.1 mg/ml bacitracin, 1% bovine serum albumin (pH 7.4), and 3-30 g of protein/assay. The assays were performed in such conditions that specific binding was strictly proportional to the amount of protein.
Bound and free radioactivities were separated by filtration through glass fiber GF/C filters presoaked for 24 h in 0.01% polyethyleneimine. The filters were rinsed three times with 20 mM sodium phosphate buffer (pH 7.4) containing 0.5% bovine serum albumin. Binding site density was evaluated as follows: density ϭ bound/free ϫ IC 50 /mg of protein.

Adenylate Cyclase Activity
Adenylate cyclase activity was determined by the Salomon (7) procedure as previously described (8). Membrane proteins (3-15 g) were incubated in a total volume of 60 l containing 0.5 mM [␣-32 P]ATP, 10 M GTP, 5 mM MgCl 2 , 0.5 mM EGTA, 1 mM cAMP, 1 mM theophylline, 10 mM phospho(enol)pyruvate, 30 g/ml pyruvate kinase, and 30 mM Tris-HCl at a final pH of 7.8. The reaction was initiated by membrane addition and was terminated after a 15-min incubation at 37°C by adding 0.5 ml of a 0.5% SDS solution containing 0.5 mM ATP, 0.5 mM cAMP, and 20,000 cpm of [ 3 H]cAMP. cAMP was separated from ATP by two successive chromatographies on Dowex 50Wx8 and neutral alumina.

Preparation of a Monoclonal Antibody for VPAC 1
Genetic Immunization and Generation of Monoclonal Antibodies-Genetic immunization and generation of monoclonal antibodies were performed according to Costagliola et al. (9). The protocol was approved by the local Ethical Committee for Animal Experimentation. Six-week-old Balb/c female mice were anesthetized by injections of 6 -10 mg/kg Ketamin HCl® associated with 0.1 ml/kg Rompum®. The anterior tibialis muscle of each leg was injected at day 0 with 100 l of 10 mM cardiotoxin (Latoxan, Rosans, France). Five days later, 50 g of the plasmid construct was injected in the same region in a final volume of 100 l of 0.09% NaCl. Injections were repeated 3 and 6 weeks thereafter. Blood samples were obtained from retro-ocular puncture 7 weeks after the initial immunization, and serum was tested for the presence of antibodies against the VPAC 1 receptor. The mouse selected (by FACS and Western blotting) for monoclonal antibody (mAb) production, was boosted by an IV injection of 100 l of a saline solution containing 10 6 CHO cells expressing the human VPAC 1 receptor. Three days later, splenocytes were fused with SP2O, a nonsecreting myeloma cell line, at a 3:1 ratio in the presence of polyethylene glycol. Fused cells were then delivered into 10 96-well plates and selected by 100 M hypoxanthine, 400 M aminopterine, and 16 M thymidine. Irradiated macrophages from mouse peritoneum were added to the well for supply of cytokines and growth factors. After 10 days, culture supernatants were screened by FACS (see below), and the cells producing antibodies were cloned by dilution. The monoclonal antibody selected (mAb-VPAC 1 ) was purified using ImmunoPure IgG purification kit (Pierce) and was of the IgG 2a subtype, based on the mouse mAb isotyping kit (Isotrip; Roche Applied Science).
Properties of the Selected Antibody-CHO cells expressing the recombinant human VPAC 1 receptor were detached from the plates using a 5 mM EDTA, 5 mM EGTA phosphate-buffered saline (PBS) solution, harvested by centrifugation (500 ϫ g, 4°C, 4 min), washed once with PBS solution, and resuspended to 3 ϫ 10 5 cells/tube in 100 l of PBS, 0.1% bovine serum albumin, containing 0.1 g of purified mAb VPAC 1 . After a 30-min incubation at 4°C, the cells were washed in the same buffer and centrifuged in the same conditions. They were then incubated for 30 min, on ice in the dark, with secondary antibody, an fluorescein isothiocyanate-conjugated ␥-chain-specific goat anti-mouse IgG (Sigma). The cells were again washed and resuspended in 250 l of PBS, 0.1% bovine serum albumin. The level of fluorescence was analyzed using a FACScalibur (BD Biosciences), and the data were processed using Cell Quest software. Basal fluorescence was determined from a sample of nontransfected CHO cells. The use of propidium iodide (10 g/ml) allowed exclusion of debris and dead cells from the analysis. The same procedure was used to evaluate the selectivity of the antibody; the level of fluorescence observed with CHO cells expressing the human VPAC 2 and the rat VPAC 1 and VPAC 2 receptors was not different from that of cells that did not express the human VPAC 1 receptor. Chimeric receptors made of different parts of the human VPAC 1 and VPAC 2 receptors and expressed in CHO cells (8) were only detected when the amino-terminal domain of the VPAC 1 receptor was conserved (Fig. 1). Furthermore, preincubation of membranes prepared from cells expressing the human VPAC 1 receptor with the monoclonal antibody at increasing concentrations did not modify the binding of 125 I-labeled VIP or the VIP-stimulated adenylate cyclase activity.

Receptor Internalization and Trafficking
Receptor internalization was defined as the percentage of cell surface receptors that were no longer accessible to the monoclonal antibody after agonist exposure. Cells expressing the VPAC 1 receptor were incubated with agonist at 37°C and, after washing three times with ice-cold phosphate-buffered saline, were processed for FACS analysis as described above. Details on specific protocols for evaluation of receptor recovery are given in the figure legends.
Confocal microscopy was also used to confirm receptor sequestration. Cells were cultured on 22-mm glass slides for 72 h. After a 30-min VIP treatment, cells were washed in PBS and fixed with Ϫ20°C absolute methanol for 10 min. Nonspecific protein binding was prevented by a 15-min incubation with 5% normal sheep serum. The cells were then incubated overnight at 4°C with the monoclonal anti-VPAC 1 antibody (1:250). The primary antibody was diluted in PBS, 1% normal sheep serum, 1‰ azide. The optimal working dilution had been previously determined empirically by serial dilutions for the antibody used. After three washes and a second 15-min incubation at room temperature with normal sheep serum, the cells were incubated for 30 min at room temperature with fluorescein isothiocyanate-conjugated ␥-chain-specific goat anti-mouse IgG (Sigma) (1:100) and diluted in the same solution as the primary antibody. Omission of the primary or secondary antibody resulted in the absence of labeling. Cells were finally incubated for 5 min at room temperature with Hoechst 33258 (Molecular Probes, Inc., Eugene, OR). After three rinses in PBS, coverslips were mounted with "Slow Fade Light" anti-fade mounting medium (Molecular Probes) in 50% glycerol (or the Calbiochem mounting medium) before viewing under an LSM510 NLO confocal microscope fitted on an Axiovert M200 inverted microscope equipped with a C-Apochromat ϫ63/1.2 numeric aperture water immersion objective (Zeiss). A ϫ2 electronic zoom was used across regions of interest. The 488-nm excitation wavelength of the Argon/2 laser, a main dichroic HFT 488, and a band pass emission filter (BP500 -550 nm) were used for selective detection of the green (fluorescein-5-isothiocyanate) fluorochrome. The nuclear stain Hoechst was excited in multiphotonic mode at 760 nm with a Mai Tai tunable broad band laser (Spectra-Physics, Darmstadt, Germany) and detected using a main dichroic HFT KP650 and a band pass emission filter (BP435-485 nm). Optical sections, 2.5 m thick, were collected for each fluorochrome sequentially. The images generated (512 ϫ 512 pixels, pixel size 0.14 m) were merged and displayed with the Zeiss LSM510 software and exported in jpg image format. All figures show a single optical section across the regions of interest. Scale bars represent 10 m.

Immunoprecipitation and Determination of Receptor Phosphorylation and Receptor Density
Cells were first cultured in phosphate-free Dulbecco's modified Eagle's medium for 16 h and then incubated for 2 h at 37°C in the presence of 0.1 mCi/ml acid-free [ 32 P]orthophosphate. At the end of this labeling period, agonist was added for 5 min. Phosphorylation inhibitors were added 30 min prior to agonist addition. Cells were then washed three times with ice-cold buffer consisting of 10 mM HEPES, 4.2 mM NaHCO 3 , 11.7 mM glucose, 1.2 mM MgSO 4 , 4.7 mM KCl, 118 mM NaCl, and 1.3 mM CaCl 2 , pH 7.4, and then lysed in 1.2 ml of a buffer consisting of 20 mM Tris, 100 mM (NH 4 ) 2 SO 4 , and 10% glycerol, pH 7.5. The cell lysate was centrifuged at 600 ϫ g at 4°C for 10 min, and the supernatant was centrifuged at 19,000 ϫ g for 30 min. The resulting pellet was resuspended in the same buffer containing 1% dodecylmaltoside (Roche Applied Science) and solubilized for 45 min at 4°C. The remaining insoluble material was eliminated by a further centrifugation. The supernatant (200 l) was added to 50 l of a 10% protein A-Sepharose suspension coated during 2 h with 2 g of purified mAb VPAC 1 . After a 150-min incubation under rotating agitation at 4°C, the Sepharose beads were separated by centrifugation and washed successively with the concentrated lysis buffer and then with a 2-fold diluted buffer and finally with water. The final bead pellet was resuspended in a buffer consisting of 125 mM Tris, 10% ␤-mercaptoethanol, 4% SDS, 20% glycerol, 0.02% bromphenol blue, pH 6.8. After heating at 60°C for 10 min, the samples were resolved by SDS-PAGE using a 10% gel. The gel was fixed and dried, and the phosphorylated bands were detected and quantified by phosphorimaging (Vilber Lourmat, Kaiser).
The contribution of GRK to agonist-dependent phosphorylation was assessed in in vitro assays. CHO cells expressing the VPAC 1 or the CCR5 receptor (kindly provided by Dr. C. Blanpain, IRIBHN, Brussels) were grown as previously described and resuspended in a buffer consisting of 20 mM Tris HCl, 10 mM MgCl 2 , 10 mM NaCl, and 1 mM EGTA added to a protease inhibitor mixture (Complete; Roche Applied Science). Fifty l of membrane (2 g/l of protein) were incubated with 50 M [␥-32 P]ATP for 5 min at 37°C in the presence of the tested agents. The final volume was 100 l. Reactions were started by the addition of ATP and stopped by a 30-s centrifugation at 13,000 ϫ g, removal of the supernatant, and solubilization of the membranes with 250 l of a buffer consisting of 20 mM Tris, 100 mM (NH 4 ) 2 SO 4 , 10% glycerol, a protease inhibitor mixture (Complete), and 1% dodecylmaltoside, pH 7.5. Human VPAC 1 solubilized receptor was then immunoprecipitated as previously described, and protein was resolved on 10% SDS-PAGE. CCR5-solubilized receptor was immunoprecipitated in 80 l of a 50% protein G-Sepharose suspension and 3 g of purified monoclonal antihuman CCR5 antibody (2D7; PharMingen). After five successive washes, the sample was resuspended in the buffer, heated at 50°C for 20 min, and processed as for the VPAC 1 receptor.
Receptor density was evaluated in all cases by binding studies using 125 I-labeled VIP as ligand as previously described (10) and confirmed in some cases by Western blotting. Because Western blotting could not be performed with the monoclonal antibody used for the FACS and immunoprecipitation studies, we used a polyclonal antibody generated by Prof. Schulz (Otto-von-Guericke-University, Magdeburg, Germany) directed against the 438 -457 sequence of the carboxyl terminus of the receptor (11). This antibody does not recognize any of the truncated receptors under study (data not shown), but Western blots performed on WT and selected mutated receptors validated binding data (see "Results"). Membranes were prepared as described above. The protein were resolved by 10% SDS-PAGE, transferred on a nitrocellulose membrane, and incubated with 1 g/ml primary antibody overnight at 4°C. We used as secondary antibody anti-rabbit peroxidase-conjugated antibody. Proteins were visualized using SuperSignal® West Pico reagent chemiluminescent substrate (Pierce and Perbio Science). 1 Receptor Phosphorylation and Internalization-VIP induced a rapid, dose-dependent stimulation of 32 P incorporation into a protein with an apparent molecular size of 75 kDa that immunoprecipitated with the mAb-VPAC 1 (Fig. 2). No signal was observed in nontransfected cells.

Agonist-induced VPAC
Receptor phosphorylation was agonist-dependent. Forskolin, phorbol esters, and the selective VPAC 1 antagonist were inac- tive per se; forskolin and phorbol esters did not modify VIPinduced phosphorylation. VIP-stimulated phosphorylation was not affected by 100 nM H-89, 6 M staurosporine, 300 nM K252a, and 100 M genistein but was inhibited by 30 M H-89, 100 M A3, and 200 -500 M CKI-7 (Fig. 3). The possible contribution of the G protein receptor kinases to receptor phosphorylation was evaluated on membranes incubated in the presence of radioactive ATP and 1 M VIP with 0.1 mM Zn 2ϩ and 1 g/ml heparin as inhibitors (12)(13)(14). A positive control consisted in CHO cell membranes expressing the chemokine receptor CCR5 stimulated by RANTES (15) (Fig. 3). VPAC 1 receptor internalization was estimated by flow cytometry by the decrease in fluorescence associated with the binding of the mAb-VPAC 1 . A typical experiment is shown in Fig. 4. Exposure to VIP induced a rapid and sustained decrease in the receptor number expressed at the cell surface that was completely blocked by preincubation in the presence of 0.5 M sucrose, suggesting that the receptor was internalized in endosomes. The selective VPAC 1 receptor antagonist was inefficient. Results observed at 5 and 30 min were presented in Fig. 4. Seventy-five percent of the receptors disappeared within 30 min.
Receptor internalization was further visualized by confocal microscopy; as shown in Fig. 5 (left panels), fluorescence signal was detected in unstimulated cells exclusively at CHO cell membranes, whereas the fluorescence was scattered in the cytosol after a 30-min stimulation with 1 M VIP. As negative control (right panel), we used the poly(A) mutant that was considered as noninternalized by the FACS technique (see below) and remained indeed at the cell membrane by confocal inspection (Fig. 5).
The reversibility was tested after 30-min exposure to agonist; after three washes, cells were incubated for 20 -120 min, and the receptors accessible to the antibody were again evaluated. There was no reappearance of the receptors (Fig. 6). We used as positive control the VPAC 2 receptor expressed in the same CHO cell line, which was similarly internalized but was re-expressed to the membrane within 120 min (Fig. 6); the results of the VPAC 2 receptor were detailed in Ref. 16.
Properties of Carboxyl-terminally Truncated Receptors-The truncated receptors studied were schematized in Fig. 7 and listed in Table I. They were all stably expressed in CHO cells. For each construction, at least four clones were generated and studied. To compare the different constructions, we detailed clones that expressed (when possible) a similar receptor density. The binding receptor properties and the capability of VIP to stimulate adenylate cyclase activity were detailed elsewhere for some of the truncated receptors (17) and are summarized in Table I. The IC 50 values of binding, the EC 50 values, and the maximal stimulatory effect of VIP on adenylate cyclase were comparable for all of the truncated receptors. We already reported that two receptors had an elevated basal adenylate cyclase activity that was decreased by the selective VPAC 1 receptor antagonist (17), suggesting a constitutive activity but one that did not modify VIP stimulation.
As compared with the 1-457 wild type receptor, the 1-444, 1-441, and 1-436 truncated receptors had a comparable 30% reduction of phosphorylation measured by densitometry on gels loaded with the same amount of receptors. The 1-433 and 1-429 truncated receptors had a 70% reduction in receptor phosphorylation (Fig. 8). The 1-421 truncated receptor retained only 10% of the VIP-stimulated phosphorylation. The shortest receptor tested, 1-398, exhibited a still detectable VIP-stimulated phosphorylation, but too low to be valuably quantified.
Receptor internalization was rapid (65-80% of the receptors were no more accessible to the antibody after 5 min of incubation with 1 M VIP, and this value remained stable for the next 25 min). Internalization was slow down for the 1-429 VPAC 1 receptor and for the shorter fragments 1-421, 1-417, and 1-402. For this last construction, 10 and 20% only of the receptors were not accessible after 5 and 30 min of incubation, respectively. Surprisingly, the shortest fragments 1-401, 1-400, 1-399, and 1-398 were internalized as rapidly and as efficiently as the wild type receptor (Table I).
As mentioned above, VIP-stimulated wild type receptor internalization was not reversible after repeated washings of the treated cells and further incubation for 120 min in absence of agonist. The same behavior was observed for the 1-444 to 1-429 truncated receptors. For the 1-421 to 1-402 truncated receptors, the results were difficult to analyze due to the low level of internalization, but no reappearance of the receptors was suspected. Surprisingly, reappearance of the internalized 1-401, 1-400, 1-399, and 1-398 truncated receptors was obvious (Fig. 6). Receptor reappearance was in all cases inhibited by 25 M monensine but not by 10 g/ml cycloheximide (data not shown).
Properties of Carboxyl-terminally Mutated Receptors-Serine and threonine residues were mutated in alanine separately or collectively to precise the phosphorylatable residues and the functional consequences on ligand recognition, adenylate cyclase activation, coupling to G protein, internalization, and reappearance of the receptors at the cell membrane. To validate the receptor quantification by binding assay, we performed a Western blot of the wild type and some point-mutated VPAC 1 receptors using the polyclonal anti-VPAC 1 antibody (Fig. 9). A same amount of receptor, evaluated by binding assay, was introduced in each lane. The results obtained after The results are summarized in Tables II and III, and the location of the mutations can be found in Fig. 7. The S455A mutant was not different from the wild type receptor except for a 40% reduced phosphorylation. The triple mutant S447A/ S448A/S449A (that eliminates a protein kinase A consensus sequence) also had properties undistinguishable from those of the wild type receptor except for a 30% reduction in receptor phosphorylation. Internalization was comparable with that of the wild type receptor, and re-expression of the receptor at the cell surface was not observed within 120 min. The S447A, S441A, T438A, T432A, S431A, S425A, and S422A mutants, on all of the parameters tested, were not different from the wild type receptor. The S435A and the T429A had as sole difference with the wild type receptor a 66% reduction in VIP-stimulated receptor phosphorylation. Combining the three mutations that decreased from at least 50% receptor phosphorylation led to the S455A/S435A/T429A triple mutant; surprisingly, the individual effects on receptor phosphorylation were not additive, the phosphorylation level reaching as for the single mutant 40% of that of the wild type receptor. However, at variance with the single mutants, internalization of the receptor was significantly slowed down (Tables II and III).
Mutations of Ser Residues in IC2-As mentioned above, the deletion of all of the Ser and Thr residues of the carboxyl terminus markedly reduced but did not abolish the VIP-stimulated receptor phosphorylation. We hypothesized therefore the possibility of a phosphorylatable Ser/Thr residue in the intracellular loops connecting the transmembrane domains. As we have previously shown (18), the importance of the distal part of the IC3 for receptor coupling to the G proteins and as a poorly coupled receptor could not be helpful, we first mutated the Ser 247 and Ser 250 residues in IC2. The single mutated S247A receptor was undistinguishable from the wild type receptor, but the S250A had a 50% reduction of VIP-induced phosphorylation (Fig. 10) without change in ligand recognition, basal and VIP stimulated adenylate cyclase values, and receptor internalization and trafficking (Table II). Combining truncation of the receptor C terminus containing all of the Ser/Thr residues and S250A mutation (S250A 1-421) completely abolished VIP-stimulated receptor phosphorylation and markedly slowed down receptor internalization (Fig. 10).
Surprisingly, combination of the Ser/Thr mutations in the C terminus that reduced receptor phosphorylation with the S250A mutation did not further reduce phosphorylation or did not further slow down receptor internalization; the S250A/ S435A/S455A/T429A receptor was not different from the S435A/S455A/T429A receptor. The activity, including receptor phosphorylation, internalization, and trafficking, of S250A/ S435A/S455A, S250A/S455A/T429A, and the S250A/S435A/ T429A were not different from any corresponding single mutant (Table III).
Mutation in Ala of all the Ser and Thr residues of the Cterminal tail and of Ser 250 led to a receptor with binding properties and adenylate cyclase activity not different from that of the wild type receptor but that was neither phosphorylated nor internalized (by the fluorescence-activated cell sorting technique and confocal microscopy; see carboxyl poly(A) in Table III and Fig. 5).

DISCUSSION
The aims of the present work were to identify the amino acid residues of the human VPAC 1 receptors that are phosphorylated during agonist stimulation and to correlate the phosphorylation level with receptor internalization and eventually reexpression to the membrane.
To identify the phosphorylated residues, we first searched for consensus sequences (19,20) and identified a protein kinase A (Ser 447 -Ser 448 -Ser 449 in the C terminus), a protein kinase C (Phe 249 -Ser 250 -Glu 251 -Arg 252 in IC2), and casein kinase (Ser 247 -Phe 248 -Phe 249 -Ser 250 in IC2; Ser 331 -Asp 332 -Ser 333 -Ser 334 and Ser 334 -Pro 335 -Tyr 336 -Ser 337 in IC3; Ser 422 -Gly 423 -Gly 424 -Ser 425 in the C terminus) consensus sites (underlined amino acids correspond to the phosphorylated residues). Because the receptor was not phosphorylated by forskolin or by phorbol esters, and because the VIP stimulated phosphorylation was not inhibited by low concentrations of H-89 or by staurosporine or K252a, we hypothesized that VIP-induced phosphorylation was not mediated by protein kinase A and protein kinase C. The partial inhibitory effect of CKI-7 (21), a selective inhibitor of casein kinase 1-␣, did not exclude involvement of that enzyme. However, phosphorylation by casein kinases implies the presence of an acidic function in the consensus (22), preferentially a phosphorylated Ser/Thr, to anchor the enzyme and trigger a phosphorylation cascade. There was no evidence that these potential initiators (Ser 247 , Ser 331 , and Ser 334 in the intracellular loops and Ser 422 in the first part of the C terminus) were indeed phosphorylated. In the distal part of the C terminus tail, however, a phosphorylation cascade starting with Thr 429 could involve Thr 432 , Ser 435 , Thr 438 , and Ser 441 . However, by single mutation, only Thr 429 and Ser 435 were identified as candidates for phosphorylation. GRK remained by exclusion the main kinase(s) identified. This was tested on membranes, since there is no known selective cell-permeable inhibitor; Zn 2ϩ and heparin are reported to antagonize GRK activity (12)(13)(14). In our experimental conditions, Zn 2ϩ was only partially efficient. However, in the positive control used (15), similar results were obtained. Finally, only Ser and/or Thr residues were phosphorylated; the tyrosine kinase inhibitor genistein was inactive, and the nonselective Ser/Thr kinase inhibitor A3 (23) completely blocked VIP-stimulated phosphorylation. Recent data (24) suggest that Ser 447 in the protein kinase A consensus in the carboxyl terminus could be phosphorylated, but it was not demonstrated that this was performed through kinase A activation; replacement of Ser 447 with Ala increased basal unstimulated phosphorylation and blunted the VIP-induced phosphorylation, a finding that was not observed in the present work.
The strategy used to identify the phosphorylated residues consisted of the progressive truncation of the carboxyl terminus and individual mutations in Ala of the suspected residues, followed by combined mutations.
For the interpretation of the results on the truncated receptors, we made the assumption that a decreased phosphoryla-   tion was due to the suppression of one phosphorylatable residue and that an unchanged phosphorylation level meant that the residues that were suppressed were not phosphorylated. In other words, we did not consider that the truncated receptors may be phosphorylated on residues other than those used in the wild type receptor. We also made the assumption that the phosphorylation level was directly linked to the number of phosphorylated residues, and we did not consider possible kinetic changes in the kinase and phosphatase activities. We also considered that the evaluation by binding studies of the number of receptors was appropriate and that phosphorylation was in any case proportional to the receptor density.
Considering these points, the similar 30% reduction in receptor phosphorylation of the three truncated forms 1-444, 1-441, and 1-436 suggested that at least one residue of Ser 447 , Ser 448 , Ser 449 , and Ser 455 was phosphorylated. Since 447-448-449 was a protein kinase A consensus sequence and since protein kinase A has been excluded, we first considered the residue Ser 455 as a good candidate. Mutation into Ala reduced receptor phosphorylation by 40%. However, the simultaneous mutation to Ala of the three adjacent Ser residues reduced the phosphorylation by 30%. The single mutation of the Ser 447 residue did not significantly modify VIP-induced phosphorylation under our conditions as already discussed. The marked decrease in receptor phosphorylation when comparing the 1-436 and the 1-433 mutants focuses on the Ser 435 residue; indeed, its replacement by Ala decreased by 70% the phosphorylation level. Since phosphorylation of the 1-429 truncated receptor was comparable with that of the 1-433, it was unlikely that Ser 431 and Thr 432 were phosphorylated. Mutation of these residues in Ala confirmed this hypothesis. The VIP-stimulated phosphorylation of the 1-421 receptor was extremely low but detectable. Removal of either Thr 429 , Ser 425 , or Ser 422 could be responsible for that decrease. Individual mutations of these Thr and Ser residues into Ala indicated that Thr 429 was the only residue to be phosphorylated. From the results on the truncated and the single mutation receptors as well as the effect of forskolin, 12-O-tetradecanoylphorbol-13-acetate, and inhibitors, we considered that the following residues were likely candidates for VIP-stimulated VPAC 1 receptor phosphorylation: Ser 455 , Ser 448 or Ser 449 , Ser 435 , and Thr 429 in the C terminus and also Ser 250 in the IC2 loop. A recent study on the 5-HT 2A receptor also implicated a serine located in the IC2 and a second in the C terminus in the agonist-mediated receptor desensitization (25).
However, the results obtained when combining mutations of these identified residues indicated that the effects on phosphorylation were not additive. This contrasts with results published, for instance, for the CCR5 receptor where four phosphorylatable serine residues were identified and each contributed equally to the total phosphorylation level (15).
In our model, combining double (data not shown), triple, and  quadruple mutations of the target residues identified by point mutation maintained a phosphorylation level of about 30% of that observed in the wild type receptor, a value reached with some single mutations. However, mutation of all of the phosphorylatable residues of the carboxyl terminus abolished receptor phosphorylation. This suggests that phosphorylation can operate on other residues when the preferred ones are missing. This alternative phosphorylation has been described for rhodopsin; rhodopsin kinase can efficiently phosphorylate other serine and threonine residues in the absence of the three sites preferentially phosphorylated (26,27). A second point to be considered is the fact that single mutation of Ser 435 and Thr 429 to Ala induced a more pronounced decrease in receptor phosphorylation than mutation of Ser 250 , Ser 455 , and the sequence Ser 447 -Ser 448 -Ser 449 . This suggested a hierarchy in the phosphorylation of the VPAC 1 receptor that could be explained by the fact that some residues are better substrates, constitute a kinase binding site, or trigger a phosphorylation cascade. This last point was already discussed. Hierarchical phosphorylation has already been reported for the ␦-opioid receptor (28), the N-formyl peptide receptor (29) and the CCK receptor (30,31). Whatever the explanation, we concluded that there is variability as to which residues are phosphorylated in mutated and probably also in truncated receptors.
Due to this variability, it is difficult to correlate phosphorylation data and receptor internalization. A quantitative aspect can be discussed; if we consider the mutant and the truncated receptors longer than residues 1-402, the receptor phosphorylation level must be reduced to at least 30% of the wild type receptor level to decrease receptor internalization. Thus, the phosphorylation rate is in excess for internalization. Such a low phosphorylation requirement has already been described for other receptors; internalization of the CCR5 receptor only requires the presence of two phosphorylated serines in the C terminus (32), even if, in vivo, four distinct C-terminal residues are phosphorylated (15). A stoichiometry of 2 mol of phosphate/ mol for the ␤ 2 -adrenergic (33) and m2 muscarinic receptor (34) is sufficient for their internalization, whereas additional phosphorylation of up to 10 -11 mol of phosphate/mol of receptor does not amplify the phenomenon. For these two receptors, the position of the phosphorylated sites was not critical. Receptor internalization was directly correlated to arrestin binding, and complete phosphorylation of the receptor was not necessary for arrestin-receptor complex stability.
Several mechanisms are possible for receptor internalization: first, an arrestin-, clathrin-, dynamin-dependent process; second, an arrestin and clathrin-independent but dynamin-dependent process through caveolae; third, an arrestin and clathrin-independent but dynamin-dependent process that does not require caveolae; and fourth, an arrestin-, clathrin-, and dynamin-independent process. Concerning the VPAC 1 receptor, the established facts are as follows: (a) a dynamin-dependent mechanism; (b) a VIP-dependent arrestin recruitment to the membrane without any effect of dominant negative mutant (5); (c) internalization in endocytic vesicles, which could be blocked by sucrose (present work); (d) a relative dependence on receptor phosphorylation (present work). Considering other class 2 GPCRs, the following appears to be true. (a) The secretin receptor is phosphorylated after agonist exposure, but phosphorylation is not required for internalization (35). Arrestin is recruited to the membrane, but there is no effect of dominant negative construction. Dominant negative dynamin was also without effect (36). (b) The parathyroid hormone receptor is internalized by an arrestin-dependent mechanisms but requires the presence of two highly conserved residues located in the core of the receptor: Asn 289 and Lys 382 . These residues could regulate a conformational modification necessary to translocation toward the endocytic endosomes (37). The use of a pathway that differs from the classical clathrin-coated pit pathway is not limited to class 2 GPCR; internalization of the class 1 GPCR 5-HT 2A receptor involves also atypic mechanisms (38).
In the present work, we showed that VPAC 1 receptor internalization occurs by two different mechanisms: a phosphorylation-dependent nonreversible pathway and a phosphorylationindependent pathway that allows rapid recycling of the receptor to the plasma membrane. This is the case of the truncated 1-421 to 1-402 receptors. A possible explanation for this is that the multiple positive charges in the 402-421 domain (Arg 403 , Arg 404 , His 406 , Lys 417 , and His 420 ) may prevent interactions of negatively charged residues located in the intracellular domains (Glu 394 and Glu 398 in the C-terminal tail but also Asp 327 and Asp 332 in IC3 and Glu 251 in EC2) with an unidentified intracellular partner. It must be noticed that Glu 394 was identified as necessary for coupling of VPAC 1 receptor to G␣ s (39).
In conclusion, the present data do not allow an unambiguous identification of the Ser/Thr residues of the VPAC 1 receptor that are phosphorylated in response to VIP. This is probably due to the possibility of alternative phosphorylation when key residues are mutated or eliminated by truncation. They clearly demonstrate that when all of the potential phosphorylation sites located in the C terminus and one Ser residue in the IC2 loop are mutated into Ala, the VIP-stimulated phosphorylation is abolished, and the receptor is no longer internalized, although it is still fully active. Truncation of the distal part of the C terminus containing all of the Ser/Thr residues also abolishes receptor phosphorylation and internalization. However, a receptor more proximally truncated, although still fully active and not phosphorylated is internalized rapidly, supporting the notion of recruitment of arrestin-insensitive/GRK-insensitive pathways. This internalization differs from that of the wild type receptor by its reversibility within 2 h, suggesting new interactions with the receptor trafficking machinery. (Department of Pharmacology and Toxicology, Otto-von-Guericke-University, Magdeburg, Germany). We are indebted to Perrine Hague and Huy Nguyen-Tran, from the "Laboratoire de Neurophysiologie," for skillful technical assistance in confocal microscopy.