Phosphorylation of the V2 vasopressin receptor.

The V2 vasopressin receptor undergoes ligand-induced sequestration and desensitization (Birnbaumer, M., Antaramian, A., Themmen, A. P. N., and Gilbert, S. (1992) J. Biol. Chem. 267, 11783-11788). The V2 receptor expressed in transfected cells labeled with [32P] orthophosphate was phosphorylated following the addition of 100 nM arginine vasopressin (AVP). Phosphorylation was complete 5 min after addition of AVP, and was not stimulated by increased levels of Ca2+ or cAMP. The half-maximal dose of AVP that stimulated phosphorylation was 2.4 ± 0.4 nM, similar to the receptor KD of 4.5 ± 0.4 nM. The role of phosphorylation on receptor desensitization was investigated by studying two vasopressin receptors 14 and 27 amino acids shorter than the wild type receptor. The missing segments were not needed for normal ligand binding or coupling to Gs, but the last 14 amino acids were required for phosphorylation. The truncated receptors exposed to 100 nM AVP were sequestered and desensitized. The R137H V2R mutant receptor that binds vasopressin with wild type-like affinity and does not couple to Gs (Rosenthal, W., Antaramian, A., Gilbert, S., and Birnbaumer, M. (1993) J. Biol. Chem. 268, 13030-13033) was phosphorylated and subjected to ligand-induced sequestration. These results established that phosphorylation is not essential for sequestration and desensitization of the V2 vasopressin receptor. Furthermore, they revealed that the conformation acquired after ligand occupancy is necessary for receptor phosphorylation and sequestration, while coupling to Gs is not.

The V2 vasopressin receptor undergoes ligand-induced sequestration and desensitization (Birnbaumer, M., Antaramian, A., Themmen, A. P. N., and Gilbert, S. (1992) J. Biol. Chem. 267, 11783-11788). The V2 receptor expressed in transfected cells labeled with [ 32 P] orthophosphate was phosphorylated following the addition of 100 nM arginine vasopressin (AVP). Phosphorylation was complete 5 min after addition of AVP, and was not stimulated by increased levels of Ca 2؉ or cAMP. The half-maximal dose of AVP that stimulated phosphorylation was 2.4 ؎ 0.4 nM, similar to the receptor K D of 4.5 ؎ 0.4 nM. The role of phosphorylation on receptor desensitization was investigated by studying two vasopressin receptors 14 and 27 amino acids shorter than the wild type receptor. The missing segments were not needed for normal ligand binding or coupling to G s , but the last 14 amino acids were required for phosphorylation. The truncated receptors exposed to 100 nM AVP were sequestered and desensitized. The R137H V2R mutant receptor that binds vasopressin with wild type-like affinity and does not couple to G s (Rosenthal, W., Antaramian, A., Gilbert, S., and Birnbaumer, M. (1993) J. Biol. Chem. 268, 13030 -13033) was phosphorylated and subjected to ligand-induced sequestration. These results established that phosphorylation is not essential for sequestration and desensitization of the V2 vasopressin receptor. Furthermore, they revealed that the conformation acquired after ligand occupancy is necessary for receptor phosphorylation and sequestration, while coupling to G s is not.
Exposure of receptors to ligands triggers not only their activation but also a chain of events, termed desensitization, that result in a reduction of cellular response to the agonist. Two major components have been identified for G protein-coupled receptors desensitization: sequestration/internalization of the occupied receptor (probably into vesicles physically separated from the plasma membrane), and phosphorylation of the intracellular domains of the receptor (1,2). The first mechanism reduces the number of receptors available for signal transduction, while the second may reduce the activity of the receptor, either directly or by promoting binding of the receptor to inhibitory proteins that block signaling through G proteins (arrestin-like) (see Bennett and Sitaramayya (3) and Attramadal et al. (4)). Receptor phosphorylation can be catalyzed by cytosolic kinases activated by the increase in second messenger concentrations (such as protein kinase A or protein kinase C), or by any of the receptor kinases that can phosphorylate G protein-coupled receptors following their activation and interaction with G proteins (GRKs) 1 (5)(6)(7)(8). Some receptors, such as the ␤ 2 -adrenergic and the angiotensin receptors, are phosphorylated by both mechanisms. The specific consequences of these phenomena on receptor activity have not yet been completely clarified, and it is still difficult to attribute the reduction in receptor function to one mechanism as opposed to the other (9).
The human V2 vasopressin receptor (V2R), is a member of the superfamily of G protein-coupled receptors that mediates the antidiuretic effects of vasopressin in the kidney. The V2R activated by arginine vasopressin (AVP) stimulates adenylyl cyclase activity, and the now elevated levels of cAMP initiate a phosphorylation cascade that increases the water permeability of the collecting duct, thus facilitating water reabsortion (10). Our previous studies, utilizing stably transfected L cells expressing the human V2R encoded by its gene, characterized the agonist-induced desensitization and sequestration of this receptor (11). The data demonstrated that the human V2R was not desensitized by protein kinase A, a result later confirmed by our observation that the cDNA encoding the protein does not predict the existence of a protein kinase A consensus site for phosphorylation (12). Since it was still unknown whether the human V2R was subjected to ligand-induced phosphorylation, this possibility was examined when it became possible to isolate the V2R from transfected cells by immunoprecipitation. To carry out these studies, we expressed in COS and HEK 293 cells both the wild type V2R and a mutated vasopressin receptor cDNA in which asparagine 22 has been replaced by glutamine, thus destroying the only glycosylation acceptor site present in the protein. The levels of expression, ligand binding affinity and coupling to G s for the nonglycosylated receptor (V2R Q ) are identical to these parameters for the wild type glycosylated protein (13). While the wild type receptor migrates in SDS-PAGE as a broad band, the nonglycosylated receptor migrates as a discrete band, thus facilitating the quantification of radioactivity incorporated into the protein. Rabbit polyclonal antibodies, raised against peptides corresponding to a segment of the third intracellular loop and the carboxyl-terminal segment of the V2R, were used to immunoprecipitate the receptor from cell extracts.
Here we report that the human V2R is subjected to ligandinduced phosphorylation and describe the characteristics of this reaction. Two receptor truncations were also analyzed for their ability to undergo phosphorylation. The V2R Q 358t receptor retains the DE motif, a possible consensus site for GRK phosphorylation (14), and lacks the last 14 amino acids that include a cluster of serines and threonines adjacent to the carboxyl terminus of the protein, another possible target for phosphorylation by receptor kinases (Fig. 1). The V2R Q 345t receptor lacks the DE motif and two other possible phosphorylation sites between the putative palmitoylation sites and amino acid 358. We describe the structural requirements for observing phosphate incorporation into the V2 receptor protein and the role of phosphorylation in the sequestration of the V2R. Additionally, we provide evidence that the uncoupled R137H mutant V2R occupied by AVP undergoes a conformational change that leads to receptor recognition by GRK and the sequestration apparatus, but it is not sufficient to activate G s .
Construction of Mutant V2 Receptors-The N22Q mutation and the premature stop codons were introduced into the V2R using a polymerase chain reaction-based approach (13). The resulting constructs were sequenced fully by the dideoxy chain termination method of Sanger et al. (15). For expression in eucaryotic cells, the cDNAs were cloned into the expression vector pcDNA3 (Invitrogen, Boston, MA).
Transient Expression Experiments-COS.M6 or HEK 293 cells were plated at a density of either 1.0 or 2.8 ϫ 10 6 cells/100-mm dish, respectively, and transfected the following day by a modification of the method of Luthman and Magnusson (16). Briefly, after rinsing with HBSS, each plate with COS.M6 cells received 800 l of HBSS, pH 7.05, containing 3 g of plasmid DNA mixed with 0.5 mg/ml DEAE-dextran. After 20 min at room temperature, 100 M chloroquine in DMEM containing 12.5 mM HEPES, pH 7.4, and 2% fetal bovine serum was added. After 3 h at 37°C, the cells were exposed to 10% dimethyl sulfoxide in HBSS for 2 min, rinsed twice with DMEM without additives, and returned to growth medium at 37°C.
HEK 293 cells were transfected the following day by replacing the growth medium with 6.7 ml of a mixture of 100 M chloroquine and 0.25 mg/ml DEAE-dextran in DMEM with 10% fetal bovine serum containing 3 g of DNA. After 2 h at 37°C, the solution was removed, and the cells were treated for 1 min at room temperature with 10% dimethyl sulfoxide in PBS. The cells were rinsed twice with PBS and returned to growth medium at 37°C.
Stable Expression in HEK 293 Cells-Cells were transfected by the calcium phosphate precipitation technique of Graham and Van der Eb (17). The day before transfection, 1-2 ϫ 10 6 cells were plated into each of two 100-mm plates. The DNA-calcium phosphate co-precipitate, containing 10 g of V2R cDNA cloned into pcDNA3, was prepared in a sterile hood immediately before use, all reagents were at 37°C. The reagents were mixed in a 15-ml sterile polystyrene tube in this order: 10 -100 l of plasmid DNA in 1 mM EDTA, 10 mM Tris-HCl, pH 7.5; sterile H 2 O to bring the volume to 900 l; 1 ml of 250 mM CaCl 2 and 100 l of 15 mM Na 2 HPO 4 , 50 mM HEPES, 150 mM NaCl, and 5 mM KCl, adjusted to pH 7.05 with NaOH. All reagents were added slowly dropwise, with gentle mixing after each addition. After 10 min at room temperature, half the whitish suspension was added dropwise to each plate containing cells and mixed by gentle swirling. After 18 h in the incubator, the medium was removed and cells were treated with 2 ml of 25% glycerol in HBSS at 37°C. After 1 min, the glycerol/HBSS mixture was diluted with 10 ml of HBSS, added slowly with continuous mixing. The solution was aspirated, and the cells were rinsed once with HBSS. Fresh medium was then added, and the plates were returned to the incubator. The next day, the cells were trypsinized and diluted with the selection medium containing 400 g/ml G418. Cells were then distributed into the wells of two 96-well microtitration plates (2,000 -4,000 cells/well), using a Costar transplate device. G418-resistant clones were picked after 16 -18 days, and expanded in 6-well plates to assay for stimulation of adenylyl cyclase activity, as described previously.
Phosphorylation of the V2R in Intact Cells-Transiently or stably transfected cells were plated in 6-well plates (6 -8 ϫ 10 5 cells/well for HEK, 1.5 ϫ 10 5 cells/well for COS). After 18 h, the cells were washed and incubated for 30 min with phosphate-free minimal essential medium, followed by the addition of 100 Ci/well [ 32 P]H 3 PO 4 . After 2 h at 37°C, the cells were exposed to vasopressin or other agents for the times and concentrations described in the text. Following treatment, the plates were chilled on ice and washed twice with phosphate-buffered saline, and the cells were solubilized for 1 h at 4°C in 300 l of RIPA (150 mM NaCl, 50 mM Tris-HCl, pH 8, 5 mM EDTA, 1% Nonidet P40, 0.1% sodium deoxycholate, 0.1% SDS) with protease inhibitors (0.1 mM phenylmethylsulfonyl fluoride, 1 g/ml soybean trypsin inhibitor, 1 g/ml leupeptin) and phosphatase inhibitors (10 mM sodium pyrophosphate, 10 mM NaF, 300 nM okadaic acid). Solubilization was helped by drawing the cells through needles of decreasing gauge (20 G, 25 G) fitted to a 1-ml plastic syringe. Cell extracts were then clarified by addition of 50 l of a 50% slurry of prewashed protein A-Sepharose in RIPA and centrifugation. Prewashed protein A-Sepharose was prepared by treating for 1 h with 25 mg/ml bovine serum albumin in RIPA, followed by two washes with RIPA alone. Two rabbit polyclonal antibodies purified by affinity column were used for immunoprecipitation. The antibodies were raised against peptides corresponding to the third intracellular loop (peptide VPGPSERPGGRRRGR, antibody no. 2), or the carboxyl terminus (peptide ARGRTPPSLGPNDES, antibody no. 3) of the human V2R. The anti-peptide polyclonal antibodies and the 12CA5 monoclonal (anti-HA epitope) antibody immunoprecipitated the same metabolically labeled protein from transfected cells (13,18). The clarified extracts were incubated overnight at 4°C with 9 g/ml peptide purified antibody, and the formed antigen/antibody complexes separated by incubating with prewashed protein A-Sepharose for 2 h at 4°C. The beads were centrifuged and washed five times with 600 l of RIPA (4 min), recovering them each time by centrifugation. The proteins were eluted for 20 min at room temperature with 80 l of 2 ϫ Laemmli buffer containing 10% ␤-mercaptoethanol. The samples were electrophoresed in 10% polyacrylamide gels and visualized by exposing the dried gels to Kodak X-Omat film at Ϫ70°C. Quantification of the 32 P incorporated into proteins was performed using a Molecular Dynamics Phosphor-Imager (Sunnyvale, CA). Identical rectangles containing the bands of interest were drawn to circumscribe the areas to be integrated. The values obtained were normalized for each sample, using a 36-kDa nonspecific band present in precipitated extracts from naive HEK cells when antibody no. 2 was used. Variation in the intensity of the 36-kDa band was due to differences in the manipulation of the Sepharose beads, not to the extent of phosphorylation of the receptor. Background measured at the level of migration of the receptor in lanes containing samples of nontransfected cells was subtracted from these values. The results are expressed as percent of the phosphorylation detected at 30 min ( Fig. 4A), or as percent of the phosphorylation detected with 100 nM AVP (Fig. 4B). The data in the graph are the means of three independent experiments Ϯ S.E.
Hormone Binding to Intact Cells-HEK 293 cells were plated in 24-well plates previously treated for 1 h with polylysine at a density of 1.5-2.5 ϫ 10 5 cells/well. Binding assays were performed the following day. Cells were washed twice with ice-cold D-PBS, after which each well received 0.5 ml of ice-cold D-PBS with 1 mM tyrosine, 1 mM phenylalanine, 0.2% glucose, 2% bovine serum albumin, and the appropriate dilution of [ 3 H]AVP. Plates were incubated for 2 h on top of crushed ice in the cold room before removal of the binding mixture by aspiration. After quickly rinsing twice with ice-cold D-PBS, 0.5 ml of 0.1 N NaOH was added to each well to extract bound radioactivity. After 30 min at Hormonal Treatment of Intact Cells-Desensitization treatments with 100 nM AVP were carried out in DMEM-high glucose at 37°C and terminated by rinsing the cells sequentially with 2 ml of ice-cold D-PBS, followed by two 40-s washes with 2 ml of an ice-cold solution of 5 mM acetic acid in Tris-buffered saline, and three rinses with normal D-PBS to ensure restoration of neutral pH. Cells were harvested to prepare homogenates, or subjected to saturation binding assays for receptor sequestration experiments.
Homogenate Preparation-Cells rinsed with D-PBS plus Ca 2ϩ and Mg 2ϩ were scraped off the dishes with a rubber policeman and were centrifuged. The supernatant was discarded. The cell pellet was resuspended in approximately 10 volumes of ice-cold 27% (w/v) sucrose, 1 mM EDTA, and 20 mM Na-HEPES, pH 7.8 (homogenization buffer) and homogenized in a Dounce homogenizer with 10 strokes of a tight-fitting pestle. Aliquots of the homogenate were used within 30 min to test for adenylyl cyclase activities.
Adenylyl Cyclase Activity in Cell Homogenates-Adenylyl cyclase activity was assayed as described previously (11). The incubation medium contained, in a final volume of 50 l, 0.1 mM [␣-32 P]ATP (1-5 ϫ 10 6 cpm), 4.0 mM MgCl 2 , 10 M GTP, 1 mM EDTA, 1 mM [ 3 H]cAMP (ϳ10,000 cpm), 2 mM isobutylmethylxanthine, a nucleoside triphosphate regenerating system composed of 20 mM creatine phosphate, 0.2 mg/ml (2,000 units/mg) creatine phosphokinase, 0.02 mg/ml myokinase (448 units/mg), and 25 mM Tris-HCl, pH 7.4. Incubations were at 32°C for 20 min. Hormones (diluted in 1% bovine serum albumin) were present at concentrations indicated in the figures. Reactions were stopped by the addition of 100 l of a solution containing 40 mM ATP, 10 mM cAMP, and 1% sodium dodecyl sulfate. The cAMP formed was isolated by a modification of the standard double chromatography over Dowex-50 and alumina columns (19,20). Under these assay conditions, cAMP accumulation was linear with time of incubation, for up to 40 min, and proportional to the amount of homogenate. The activities were expressed as picomoles of cAMP formed per min per mg of homogenate protein. Protein concentration of the samples was determined by the method of Lowry (21) using bovine serum albumin as standard.
Statistical Analysis of the Data-The values obtained with the Phos-phorImager, the AVP binding data, and the adenylyl cyclase activity determinations, shown in Figs. 4 -6, were fitted by the least square method using logistic functions. The minimization of the square differences was calculated with the function Solver in Microsoft Excel 5.0 C. The values reported in these figures were obtained from the fitting curves. A paired sample t test was applied when mentioned; differences were considered statistically significant if p Ͻ 0.005.

RESULTS
Ligand-dependent V2R Phosphorylation-To assess whether the V2 receptor undergoes ligand-induced phosphorylation, two different recombinant cDNAs were transfected into COS cells, one encoding the wild type V2R, the other encoding a mutant V2R cDNA in which asparagine in the glycosylation consensus sequence had been substituted by glutamine (V2R Q ). Cells previously incubated with 32 P were exposed to 100 nM AVP for 15 min at 37°C. Extraction and immunoprecipitation of the V2 receptor was performed as described under "Experimental Procedures" using rabbit polyclonal antibody no. 2. As illustrated in Fig. 2, the treatment with AVP promoted phosphorylation of the receptor. Cells expressing the wild type V2R yielded a phosphorylated broad band at 45-55 kDa that coincides with the migration observed for the glycosylated receptor (13). Cells expressing the nonglycosylated receptor treated under similar conditions yielded a radioactive band of approximately 40 kDa that had been previously identified as the nonglycosylated form of the V2R. In both cases, the phosphorylation of the receptor protein was AVP-dependent. Since this result proved that the absence of sugar did not interfere with the incorporation of phosphate into the receptor protein, and we had previously demonstrated the nonglycosylated receptor to be fully active (13), subsequent experiments to characterize V2R phosphorylation were performed with cells transiently or stably transfected with a plasmid encoding the nonglycosylated receptor (V2R Q ) to improve the quantitation of the radioactive band.
To examine the specificity of the reaction, and whether the kinase(s) responsible for the phosphorylation of the V2R could be activated by second messengers, HEK 293 clonal cells expressing the V2R Q were treated with 100 M ATP or 100 nM VIP to stimulate receptors intrinsic to these cells. In addition, cells were treated with 100 M forskolin to increase intracellular levels of cAMP, or with 400 nM phorbol 12-myristate 13acetate to test for the possible involvement of protein kinase C in receptor phosphorylation. As shown in Fig. 3, all these treatments failed to stimulate phosphorylation of the V2R, although they induced incorporation of radioactivity in several of the proteins that coprecipitated with the receptor, thus modifying the radioactive background in all lanes. Treatment of transiently transfected COS cells expressing the V2R with radioactive phosphate and 10 M isoproterenol, or with 100 M ATP to stimulate receptors intrinsic to the COS cells also failed to promote phosphorylation of the V2R (data not shown).
Exposure of the cells to 100 nM AVP at 37°C induced a fast rate of receptor phosphorylation. As illustrated in Fig. 4A, the incorporation of radioactive phosphate into the receptor protein was already detectable 30 s after AVP addition, the shortest time analyzed. Receptor phosphorylation reached apparent saturation in less than 10 min, and the level of radioactivity in the receptor band remained unchanged for at least 30 min in the presence of AVP. Experiments were performed to correlate phosphorylation and the extent of receptor occupancy by AVP. Thus, the experiments established that, as described for other receptors, the activity of the kinases involved depends on occupancy of the binding site by the agonist, strongly suggesting that a form of GRK was responsible for V2R phosphorylation.
The R137H mutant V2R, previously characterized by Rosenthal et al. (23) was examined next. This mutation, found in patients suffering the X-linked form of nephrogenic diabetes insipidus, encodes a full-length receptor that binds AVP with the same K D as the wild type receptor but fails to stimulate adenylyl cyclase activity (23). Experiments to test whether this mutant receptor could be phosphorylated by receptor kinases were performed. As previously reported, the R137H V2R is expressed at a lower level than the wild type receptor, 0.7 compared to 6.0 ϫ 10 6 sites/cell in our latest experiments. Immunoprecipitation of the receptor with antibody no. 3 helped to reduce the radioactive background in the vicinity of the receptor band and thus increased the sensitivity of the assay. Quantification of ligand-dependent phosphorylation with the PhosphorImager detected 7.5 Ϯ 0.4% phosphorylation in the R137H mutant compared to 100% for the wild type V2R. In the experiment shown in Fig. 5, the usual volume of extract containing the mutant receptor was loaded along with 0.1 volume of the extract containing wild type V2R to compensate for the reduced level of expression. As illustrated, it was evident that that the R137H V2R was phosphorylated in an AVP-dependent manner. A phosphorylated protein was detected at the same location prior to vasopressin treatment. COS.M6 cells not expressing the V2R were labeled with 32 P, exposed for 15 min at 37°C to 10 M isoproterenol to stimulate their cAMP production, and then subjected to extraction and immunoprecipitation with antibody no. 3. No radioactive bands were detected at the position where the receptor migrates, suggesting that the band seen in transfected cells was the mutant protein. Unable to identify a transfection-independent phosphorylated protein at this location, it was concluded that the mutant V2R has a detectable level of basal phosphorylation.
Role of the Carboxyl Terminus in V2R Phosphorylation-As in many other G protein-coupled receptors, the carboxyl-terminal portion of the V2R is rich in serines and threonines. To explore whether this segment played a role in phosphorylation, we mutagenized the V2R cDNA, introducing two premature stop codons at Cys-358 and Gly-345 to direct the synthesis of truncated receptor proteins. As shown in a representative experiment illustrated in Fig. 6, the truncations did not alter the binding affinity of the mutant receptors for AVP. The average K D values calculated from three independent experiments were: 4.8 Ϯ 0.7, 6.1 Ϯ 1.3, and 4.7 Ϯ 0.6 nM for the V2R Q , the V2R Q 358t, and the V2R Q 345t receptors, respectively. The level of expression of the truncated receptors, assessed by performing saturation binding assays in transiently transfected cells, was compared to that of the wild type receptor. Whereas the B max value obtained with the V2R Q 358t protein was similar to that obtained with the full-length V2R Q (136 Ϯ 6%), the B max value for the V2R Q 345t receptor was 45 Ϯ 8% lower than that of the V2R Q . The mutant receptors were next tested for their ability to mediate AVP stimulation of adenylyl cyclase activity by assaying homogenates prepared from transiently transfected cells. As illustrated in Fig. 7, no significant differences were detected either in the extent of stimulation of adenylyl cyclase activity, nor in the EC 50 values of these stimulations among the V2R Q 358t, V2R Q 345t, and full-length receptor. Thus, reduced expression of the V2R Q 345t receptor did not result in a detectable difference in its ability to stimulate the adenylyl cyclase system.
Although, according to the saturation binding data, both truncated proteins were present at levels comparable to those of the full-length receptor when tested in phosphorylation experiments, it was not possible to detect AVP-dependent incorporation of phosphate into either truncated receptor protein, revealing that the last segment of the carboxyl terminus is required for phosphorylation of the V2R. As illustrated in Fig.  8, no ligand-induced phosphorylation was detected for the V2R Q 358t immunoprecipitated with antibody no. 3 and analyzed along the R137H and the wild type V2R. The susceptibility of the truncated receptors to AVP-induced desensitization was examined next. Transiently transfected cells expressing the full-length and the truncated V2Rs were exposed to 100 nM AVP, and the ability of the receptors to mediate AVP stimulation of adenylyl cyclase activity was assessed in homogenates of those cells. The maximal extent of desensitization was evaluated for each construct in several independent experiments. As shown in Fig. 9, the truncated forms of the receptor were desensitized by exposure to 100 nM AVP. The V2R Q 358t showed only 15 Ϯ 5% desensitization, while the V2R Q 345t showed 26 Ϯ 3% desensitization compared to 35 Ϯ 3% desensitization of the V2R Q . Impaired desensitization of the truncated forms might be due to the absence of phosphorylation; the difference observed between the truncated receptors revealed the presence of segments in the carboxyl terminus capable of modulating desensitization.
Sequestration of Nonphosphorylated Receptors-To examine the role of phosphorylation in the sequestration/internalization of the V2R, the extent of sequestration that could be achieved with the fully active truncated receptors was compared to that of the full-length receptor. As shown in Fig. 10, when exposed to 100 nM AVP the truncated receptors disappeared from the cell surface, although at a lower rate than the full-length V2R Q . Thus, the last 25 amino acids of the V2R were not required for sequestration to occur. Deleting the last 14 amino acids from the V2R impaired sequestration to a greater extent than deleting the last 27 amino acids of the carboxyl-terminal segment. A similar finding has been reported for the calcitonin receptor (22). These data indicated that, for the V2R protein, phosphorylation was not required for receptor sequestration. Neither was sequestration blocked by the absence of coupling activity. As shown in Fig. 11, the R137H V2R mutant expressed in stably transfected L cells can be sequestered in 10-min exposure to 100 nM AVP to a similar extent as the wild type receptor. Furthermore, the wild type and the mutant receptor binding sites returned to the cell surface with the same time course following ligand-induced sequestration. DISCUSSION Similar to what has been described for other G proteincoupled receptors, binding of agonist to the V2 vasopressin receptor caused phosphorylation of the protein. Unlike what has been reported for the ␤ 2 -adrenergic or parathyroid hormone receptors (6, 7), there was minimal phosphorylation of the wild type V2R in the absence of ligand. While characterizing the pattern of expression of the V2R, Sadeghi et al. (18) demonstrated that only an 35 S-labeled broad band migrating between 45 and 55 kDa represented properly processed glycosylated vasopressin receptor, the other receptor bands present consisting mostly of immature receptor, as revealed by the composition of the sugar moiety. This immature protein was also apparent when the nonglycosylated receptor was expressed and represented 50% or more of the V2R molecules isolated from transiently transfected cells. Those experiments provided no information as to the location of the immature protein in the cell, nor did they addressed the issue of whether this protein was functional to any extent. The observation that the immature receptor protein was not phosphorylated in response to AVP suggested that this protein may not reach the plasma membrane, or if present on the cell surface, it cannot be a substrate for phosphorylation. The data indicate that only the mature receptor present on the plasma membrane (migrating at 40 kDa when the V2R is not glycosylated), can interact with the ligand and promote its own phosphorylation.
Because binding of AVP was a requirement for V2R phosphorylation and the extent of radioactive phosphate incorporation was proportional to the degree of occupancy of the receptor by agonist, the involvement of one of the kinases able to recognize G protein-coupled receptors as substrate was suggested (5). Incorporation of phosphate into the receptor came to completion within a few minutes, and it was sustained for as long as 30 min after addition of 100 nM AVP to the cells. As Fig. 3 illustrates, there was no detectable decay in the amount of phosphorylated receptor that could be isolated during this period. Probably the majority of the receptors had been removed from the cell surface in 30 min, thus these results suggest that the phosphorylated protein does not suffer significant degradation after sequestration/internalization. It is possible that the phosphorylated receptor, once sequestered, might recycle quickly back to the cell surface after losing its phosphate, to be phosphorylated again once occupied by AVP; or that after phos- FIG. 7. Stimulation of adenylyl cyclase activity by V2R Q , V2R Q 358t and V2R Q 345t receptors. Cell homogenates were used to measure the ability of the truncated receptors to stimulate adenylyl cyclase activity, as described under "Experimental Procedures." Data are expressed as picomoles/min/mg. A representative experiment is shown. The EC 50 values calculated from three independent experiments were 0.68 Ϯ 0.06, 0.45 Ϯ 0.07, and 0.60 Ϯ 0.08 for E, V2R Q ; Ⅺ, V2R Q 358t; and Ç, V2R Q 345t, respectively. The EC 50 values of the truncated receptors and the wild type were not significantly different (p Ͼ 0.05 for both).
FIG. 6. Binding of [ 3 H]AVP to the V2R Q , V2R Q 358t, and V2R Q 345t receptors. Two 100-mm dishes of COS cells were transfected with each cDNA construct and pooled together 48 h after transfection for plating. Saturation binding assays were performed as described under "Experimental Procedures." A representative experiment is shown. Scatchard analysis of three independent experiments gave the following values: K D , 4.8 Ϯ 0.7, 6.1 Ϯ 1.3, and 4.7 Ϯ 0.6 nM for the V2R Q , V2R Q 358t, and V2R Q 345t receptors, respectively. B max , 1.12 Ϯ 0.15, 1.56 Ϯ 0.20, and 0.64 Ϯ 0.13 ϫ 10 6 sites/cell for the V2R Q , V2R Q 358t, and V2R Q 345t receptors, respectively. phorylation the protein remains for all this time in an early endosomal compartment. The experimental evidence provides no explanation for the observed persistence of the phosphorylated protein or for its location.
The observation that the truncated receptors were able to fully stimulate adenylyl cyclase activity, revealed that the carboxyl terminus, although required for phosphorylation, did not play a significant role in coupling the receptor to G s . These data are consistent with the activity shown for the V1a/V2 chimeras reported by Liu and Wess (24), in particular the chimera where the whole segment of the V2R after the seventh transmembrane domain has been replaced by the equivalent segment of the V1a receptor.
When cells expressing the truncated receptors were exposed for 20 min to 100 nM AVP, the extent of sequestration observed was less pronounced for the V2R Q 358t than for the V2R Q 345t protein. Deletion of the last 14 amino acids resulted in a significant reduction in receptor sequestration, while deletion of the next 13 amino acids restored the ability of the receptor to be removed from the cell surface. These data suggested not a requirement but instead a modulatory role for this distal segment in the sequestration of the V2R, locating the sequestration signal itself elsewhere in the protein. The modulatory role of the terminal segment could depend on its interaction with arrestin-like molecules, known to participate in this process (25,26). Because the missing protein fragments are rich in serines and threonines, a correlation between absence of phosphorylation and the reduction in sequestration of the truncated receptors is likely.
As originally described for rhodopsin, activated G proteincoupled receptors are phosphorylated by GRK, and as a consequence are able to bind ␤-arrestin with increased affinity. Recently, Ferguson et al. (26) reported that deletion of the COOH terminus at Cys-341, or substitution of the phosphorylation sites of the ␤ 2 -adrenergic receptor resulted in reduction but not loss of internalization. As a matter of fact, these mutant ␤ 2 -adrenergic receptors were sequestered 50% as well as the wild type, a result similar to what has been observed with the V2 receptor. In the same report, these authors describe a role FIG. 10. Time course of AVP-induced sequestration of V2R Q , V2R Q 358t, and V2R Q 345t receptors in transiently transfected HEK 293 cells. HEK 293 cells transfected with the three different constructs were exposed to vehicle or to 100 nM AVP at 37°C for the times indicated to stimulate hormone-dependent sequestration. Cells were then chilled, washed with PBS, subjected to an acid wash to remove bound ligand, and neutralized as described under "Experimental Procedures." Saturation binding assays were performed to determine the number of high affinity binding sites remaining on the cell surface.
FIG. 11. Agonist effect on V2R Q and V2R Q -R137H receptor sequestration. Stably transfected L cells expressing the wild type V2R (wt) or the V2R-R137H (Q2-3) were treated for 10 min at 37°C with vehicle or 100 nM AVP. Cells were chilled, washed with PBS, subjected to acid wash to remove bound ligand, and neutralized, as described under "Experimental Procedures." Growth medium was added, and the cells were returned to the incubator for the times indicated. Binding assays at 20 nM [ 3 H]AVP in the presence and absence of 10 M cold AVP were performed in replicate at the indicated times to determine the number of binding sites present on the cell surface.  9. Desensitization of V2R Q , V2R Q 358t, and V2R Q 345t receptors. HEK 293 cells transfected with the three different constructs were exposed to vehicle or to 100 nM AVP for 15 min at 37°C. After an acid wash to remove surface ligand, as described under "Experimental Procedures," cell homogenates were prepared, and the ability of the receptors to mediate vasopressin stimulation of adenylyl cyclase activity was assessed in the presence of 100 nM AVP. Percent desensitization was calculated by dividing the difference in maximal activity between the vehicle and AVP treated samples by the maximal activity measured in the vehicle-treated sample. The results are presented as means Ϯ S.E. For V2R Q and V2R Q 345t, n ϭ 7, for V2R Q 358t n ϭ 4. The average of the adenylyl cyclase activities at 100 nM VIP expressed as picomoles of cAMP/min/mg in these experiments were: 59.0 and 54.0 for V2R Q ; 59.0 and 56.0 for V2R Q 345t; and 50.0 and 49.0 for V2R Q 358t, for the vehicle and the 100 nM AVP-treated cells, respectively.
for ␤-arrestin in promoting agonist-induced sequestration of nonphosphorylated ␤ 2 -adrenergic receptor. If arrestin binding promotes sequestration, the existence of residual sequestration for the nonphosphorylated ␤ 2 -adrenergic and the V2 receptors suggests that the intervention of the GRKs enhances but does not cause removal of receptors from the cell surface.
The structure of the occupied receptor, not the activation of a G protein, seems crucial for activating the receptor kinases. The observation that the uncoupled R137H mutant V2R is phosphorylated in response to agonist, indicates that it is possible to dissociate G protein activation from GRK activation as different functions of an activated receptor (27). Despite its functional flaws, the R137H V2R is recognized by the sequestration apparatus with only a small reduction in efficiency, and it can recycle to the cell surface after removal of the ligand in a manner similar to the wild type receptor. The phosphorylation in the absence of ligand detected for the R137H V2R is probably a unique property of the mutant receptor. Since the wild type receptor is present at 10-fold higher levels, it should be easy to detect a similar basal level of phosphorylation if it was present.
Although it is widely accepted that phosphorylation plays an important role in the desensitization of G protein-coupled receptors, it has not yet been clarified how or to what extent the presence of phosphate groups reduces coupling to G proteins. For other G protein-coupled receptors, the effect of second messenger dependent kinases can complicate the analysis of the data, while in the case of the V2R, there is no such interference. Nevertheless, a ligand-induced reduction in truncated receptors activity was observed in the absence of detectable phosphorylation. These data have revealed that, for the V2R, eliminating GRK phosphorylation reduces the extent of desensitization but does not block it.
In conclusion, we have demonstrated that the V2R is phosphorylated when occupied by AVP and that the last 14 amino acids of the protein are required for this effect. We also have demonstrated that abolishing phosphorylation by GRK reduced the extent of desensitization but did not block it. Thus, as observed for receptor sequestration, we concluded that GRKcatalyzed phosphorylation plays an enhancing, not a permissive, role in these phenomena.