Association of β-Arrestin with G Protein-coupled Receptors during Clathrin-mediated Endocytosis Dictates the Profile of Receptor Resensitization*

Resensitization of G protein-coupled receptors (GPCRs) following agonist-mediated desensitization is a necessary step for maintaining physiological responsiveness. However, the molecular mechanisms governing the nature of GPCR resensitization are poorly understood. Here, we examine the role of β-arrestin in the resensitization of the β2 adrenergic receptor (β2AR), known to recycle and resensitize rapidly, and the vasopressin V2 receptor (V2R), known to recycle and resensitize slowly. Upon agonist activation, both receptors recruit β-arrestin to the plasma membrane and internalize in a β-arrestin- and clathrin-dependent manner. However, whereas β-arrestin dissociates from the β2AR at the plasma membrane, it internalizes with the V2R into endosomes. The differential trafficking of β-arrestin and the ability of these two receptors to dephosphorylate, recycle, and resensitize is completely reversed when the carboxyl-terminal tails of these two receptors are switched. Moreover, the ability of β-arrestin to remain associated with desensitized GPCRs during clathrin-mediated endocytosis is mediated by a specific cluster of phosphorylated serine residues in the receptor carboxyl-terminal tail. These results demonstrate that the interaction of β-arrestin with a specific motif in the GPCR carboxyl-terminal tail dictates the rate of receptor dephosphorylation, recycling, and resensitization, and thus provide direct evidence for a novel mechanism by which β-arrestins regulate the reestablishment of GPCR responsiveness.

G protein-coupled receptor (GPCR) 1 signaling plays a pivotal role in regulating various physiological functions including vision, olfaction, neurotransmission, cardiac output, and fluid and electrolyte balance. The magnitudes of these physiological responses are linked intimately to the delicate balance between GPCR signal generation and signal termination. The termination of GPCR signaling is tightly regulated by a family of intracellular proteins termed ␤-arrestins (1)(2)(3)(4)(5). ␤-Arrestins bind with high affinity to agonist-activated GPCRs that have been phosphorylated by G protein-coupled receptor kinases (GRKs). The interaction of ␤-arrestins with a phosphorylated receptor uncouples the receptor from heterotrimeric G proteins, producing a nonsignaling, desensitized receptor. For many GPCRs, like the ␤ 2 -adrenergic receptor (␤ 2 AR), ␤-arrestins target the desensitized receptor to clathrin-coated pits for endocytosis (4, 6 -10). In this process, ␤-arrestins function as docking proteins that link receptors to components of the endocytic machinery such as AP-2 and clathrin (11,12). Intracellular trafficking of GPCRs following ␤-arrestin-mediated desensitization is necessary for receptor resensitization (5,13,14). Therefore, ␤-arrestins are involved not only in terminating receptor G protein coupling but also in initiating processes that regulate re-establishment of receptor responsiveness.
While it is evident that responsiveness to most GPCR-activating ligands can be regained, the biochemical and kinetic properties of the cellular processes mediating resensitization may differ considerably among receptors. Some internalized GPCRs recycle rapidly back to the plasma membrane fully resensitized, while others are retained inside the cell and recycle slowly or not at all (5, 14 -22). The molecular mechanisms governing the rate at which receptors recycle and re-establish agonist responsiveness are poorly understood; however, the dephosphorylation of GRK-phosphorylated receptors in early endosomes appears to be a critical step in the resensitization pathway (14,17,(23)(24)(25). An event presumably necessary for the dephosphorylation of GPCRs is their dissociation from ␤-arrestin, as suggested by in vitro evidence that the binding of visual arrestin to GRK-phosphorylated rhodopsin prevents rhodopsin dephosphorylation (26). A common assumption is that ␤-arrestins do not dissociate from desensitized receptors at the plasma membrane but traffic with them into early endosomes (27,28). However, recent observations have demonstrated that the fate of the GPCR-␤-arrestin complex can differ among receptors (29). For some receptors, the receptor-␤-arrestin complex dissociates at or near the plasma membrane shortly after the formation of clathrin-coated pits, and ␤-arrestin is excluded from receptor-containing endocytic vesicles. For other receptors, the receptor-␤-arrestin complex remains intact and is internalized into endosomes. The ability of ␤-arrestin to remain associated with some receptors but not others suggests that ␤-arrestin may regulate the cellular trafficking and dephosphorylation of receptors and ultimately their kinetics of resensitization.
In order to investigate the role of ␤-arrestin in the regulation of GPCR resensitization, two GPCRs, the ␤ 2 AR and the vasopressin V2 receptor (V2R), which share many of the same signaling properties but differ markedly in their ability to recycle and resensitize (5,14,20), were chosen. We demonstrate using mutagenesis and chimeric receptors that the ability of ␤-arrestin to remain associated with desensitized GPCRs and internalize with them into endosomes dictates the properties of GPCR resensitization. These observations, therefore, provide direct evidence for an important mechanism by which ␤-arrestins can regulate the physiological responsiveness of GPCRs.
Receptor Binding-Membrane binding assays were performed on transfected HEK-293 cells as described previously (33). Briefly, membrane proteins (2 g) from cells expressing the ␤ 2 AR and ␤ 2 AR-V2R chimera were incubated in phosphate-buffered saline (PBS) containing 0.1% bovine serum albumin (BSA) at room temperature in the presence of 15 pM [ 125 I]cyanopindolol and increasing concentrations of isoproterenol (10 pM to 30 M). Membrane proteins (10 g) from cells expressing the V2R and V2R-␤ 2 AR chimera were incubated in PBS containing 2% BSA at room temperature with increasing concentrations of [ 3 H]AVP (0.5 nM to 16.0 nM). Binding was terminated by rapid filtration and consecutive washes with ice-cold wash buffer (120 mM NaCl, 50 mM Tris-HCl, pH ϭ 7.2). Wild-type and chimeric receptor expression levels were measured on whole cells as described previously (32). Transfected HEK-293 cells expressing the V2R and V2R-␤ 2 AR chimera were incubated 2 h on ice in PBS containing 2% BSA with a saturating concentration of [ 3 H]AVP, and bound radioactivity was extracted with 0.1 M NaOH. Nonspecific binding was determined under each respective condition in the presence of 10 M propranolol or 10 M unlabeled AVP. Receptor expression levels varied between 2000 and 4000 fmol/mg of whole cell protein for experiments with ␤arr2-GFP and between 500 and 1500 fmol/mg of whole cell protein for all other experiments.
Receptor Sequestration and Recycling-Receptor sequestration was assessed by flow cytometry as described previously (30). Sequestration was defined as the fraction of total cell surface receptors that, after exposure to agonist, were removed from the plasma membrane and thus not accessible to antibodies from outside the cell. For recycling experiments, isoproterenol was removed by three rapid washes with serum-free medium at 37°C and AVP was removed by successive washes with ice-cold PBS, acid (150 mM NaCl/5 mM acetic acid), PBS, and serum-free medium.
Confocal Microscopy-␤arr2-GFP trafficking was visualized in transfected HEK-293 cells on a heated (37°C) microscope stage as described previously (31). Images were collected sequentially using single line excitation (488 nm) with a Zeiss laser scanning confocal microscope (LSM-510). For experiments assessing ␤arr2-GFP trafficking after agonist removal, cells were washed as described above to remove agonist and returned to a 37°C incubator for 60 min. Colocalization of ␤arr2-GFP with rhodamine-labeled receptors was performed on transfected cells pre-incubated in serum-free medium containing a rhodamineconjugated anti-HA 12CA5 mouse monoclonal antibody (1:100) for 45 min at 37°C. Cells were then washed three times with serum-free medium, treated with the appropriate agonist at 37°C for 30 min, and imaged by confocal microscopy. ␤arr2-GFP and rhodamine-labeled receptor fluorescence were performed using dual excitation (488, 568 nm) and emission (515-540 nm, GFP; 590 -610 nm, rhodamine) filter sets.
Adenylyl Cyclase Assays-Whole cell cyclase assays were performed on transfected HEK-293 cells using varying concentrations of isoproterenol (1 ϫ 10 Ϫ12 M to 1 ϫ 10 Ϫ5 M) or AVP (1 ϫ 10 Ϫ12 M to 1 ϫ 10 Ϫ5 M) as described previously (5). For membrane adenylyl cyclase assays, transfected HEK-293 cells were harvested by scraping in ice-cold lysis buffer (10 mM Tris-HCl, 5 mM EDTA, pH ϭ 7.4) and membranes were prepared by disruption with a Polytron homogenizer for 20 s at 20,000 rpm followed by centrifugation at 40,000 ϫ g. The cell membrane was resuspended in lysis buffer by Polytron homogenization for 15 s at 20,000 rpm, centrifuged, and resuspended in ice-cold assay buffer (75 mM Tris-HCl, 2 mM EDTA, 15 mM MgCl 2 , pH ϭ 7.4) to a final concentration of 1-2 g/l membrane protein. Equivalent amounts of membrane protein in 20-l aliquots were assayed for agonist-stimulated adenylyl cyclase activity in a final volume of 50 l as described previously (5).
Whole Cell Phosphorylation-Receptor phosphorylation was performed as described previously (5). In brief, transfected HEK-293 cells were labeled for 1 h at 37°C with [ 32 P]orthophosphate (100 Ci/ml) in phosphate-free medium. Cells were stimulated with agonist for 10 min at 37°C and then washed three times on ice with ice-cold PBS. For resensitization experiments, cells were washed to remove agonist as TABLE I Binding, expression, and adenylyl cyclase activation parameters for the wild-type and chimeric receptors The ligand binding properties, expression, and half-maximal effective concentration (EC 50 ) for the activation of adenylyl cyclase were determined for the wild-type and chimeric receptors in transfected HEK-293 cells. Isoproterenol competition binding assays were performed on membranes to determine the equilibrium dissociation binding constants (K d ) for the ␤ 2 AR and ␤ 2 AR-V2R chimera. K h and K l indicate the K d values for the high and low affinity states, respectively. [ 3 H]AVP saturation binding assays were performed on membranes to determine the affinity of AVP for the V2R and V2R-␤ 2 AR chimera. Saturation binding assays were performed on whole cells using [ 125 I]cyanopindolol or [ 3 H]AVP to determine the B max of the wild-type and chimeric receptors. The EC 50 for activation of adenylyl cyclase was determined on whole cells using varying concentrations of isoproterenol or AVP. All values are expressed as the mean Ϯ S.E. (n ϭ 3).
described above and either maintained on ice or allowed to recover at 37°C. All cells were scraped in radioimmune precipitation buffer (150 mM NaCl, 50 mM Tris, 5 mM EDTA, 10 mM NaF, 10 mM disodium pyrophosphate, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS) containing protease inhibitors and solubilized for 1 h at 4°C. After centrifugation, supernatants were collected and assayed for protein concentration (Bio-Rad DC protein assay kit). HA-tagged receptors were immunoprecipitated at 4°C using the anti-HA 12CA5 mouse monoclonal antibody. Equivalent amounts of receptor, as determined by receptor expression and the amount of solubilized protein in each sample, were subjected to SDS-polyacrylamide gel electrophoresis and processed for autoradiography. Receptor phosphorylation was quantitated using a Molecular Dynamics PhosphorImager and ImageQuant software. Data Analysis-The mean and standard error of the mean are expressed for values obtained from the number of independent experiments indicated. Statistical significance was determined using a twotailed t test. Binding and dose-response data were analyzed using GraphPad Prism software.

RESULTS
Construction and Characterization of the ␤ 2 AR-V2R and V2R-␤ 2 AR Chimeras-The ␤ 2 AR and V2R share many of the same signaling properties: both agonist-activated receptors couple to G s ␣ and stimulate adenylyl cyclase, desensitize in a GRK-dependent fashion, and internalize into early endosomes (5, 15, 34 -37). However, whereas the ␤ 2 AR recycles rapidly back to the plasma membrane fully resensitized, the V2R is retained inside the cell (5,14,20). ␤-Arrestin binding to many GPCRs is mediated by GRK-phosphorylated serine and threonine residues located in the receptor carboxyl-terminal tails. In addition, residues in the carboxyl terminus of the V2R have been shown to play a critical, but undefined, role in the intracellular retention of this sequestered receptor (20). Therefore, to gain insight into the mechanisms regulating the ability of these two receptors to recycle and resensitize, chimeric receptors were constructed in which the carboxyl-terminal tails of the ␤ 2 AR and V2R were exchanged, one for the other, after the putative sites of palmitoylation. The ␤ 2 AR-V2R chimera contains the first 341 amino acids of the ␤ 2 AR (Met-1 to Cys-341) fused to the last 29 amino acids of the V2R (Ala-343 to Ser-371). The V2R-␤ 2 AR chimera contains the first 342 amino acids of the V2R (Met-1 to Cys-342) fused to the last 72 amino acids of the ␤ 2 AR (Leu-342 to Leu-413). The chimeric receptors were essentially indistinguishable from their wild-type counterparts with respect to their affinity for agonist, level of expression, and half-maximal effective concentration (EC 50 ) for adenylyl cyclase activation (Table I).
Sequestration of the ␤ 2 AR, V2R, and ␤ 2 AR-V2R and V2R-FIG. 1. Sequestration of the ␤ 2 AR, V2R, and ␤ 2 AR-V2R and V2R-␤ 2 AR chimeras. A, wild-type and chimeric receptors were transiently expressed in HEK-293 cells together with empty vector (CON), the ␤-arrestin1 dominant negative mutant V53D, or the dynaminI dominant negative mutant K44A. After treating the cells for 30 min at 37°C with or without 10 M isoproterenol (␤ 2 AR and ␤ 2 AR-V2R chimera) or 100 nM AVP (V2R and V2R-␤ 2 AR chimera), cell surface receptors were assessed by flow cytometry. Receptor sequestration is expressed as a loss of cell surface immunofluorescence. B, wild-type and chimeric receptors were transiently expressed in COS-7 cells together with empty vector (CON), ␤-arrestin1, or ␤-arrestin2. Cells were treated and processed as described above. The data represent the mean Ϯ S.E. of three independent experiments (*, p Ͻ 0.05; **, p Ͻ 0.01; and ***, p Ͻ 0.001 compared with control values; t test).
FIG. 2. Cellular trafficking of ␤arr2-GFP with the ␤ 2 AR, V2R, and ␤ 2 AR-V2R and V2R-␤ 2 AR chimeras. HEK-293 cells were cotransfected with plasmids containing the cDNA for ␤arr2-GFP and the cDNA for the ␤ 2 AR (A), ␤ 2 AR-V2R chimera (B), V2R (C), or V2R-␤ 2 AR chimera (D). The distribution of ␤arr2-GFP fluorescence was visualized in cells before and after treatment with 10 M isoproterenol (␤ 2 AR and ␤ 2 AR-V2R chimera) or 100 nM AVP (V2R and V2R-␤ 2 AR chimera). Shown are confocal microscopic images of ␤arr2-GFP fluorescence in the same HEK-293 cells treated with agonist for 0, 2, and 15 min at 37°C. ␤ 2 AR Chimeras-Differences in the ability of the ␤ 2 AR and V2R to recycle and resensitize might be due to differences in their sequestration pathways. The ␤ 2 AR internalizes in a ␤-arrestin-and clathrin-dependent manner (34), but little is known about the sequestration pathway of the V2R other than it is blocked nonspecifically by sucrose (36). Therefore, we evaluated the ␤-arrestin and clathrin dependence of the V2R sequestration pathway. HEK-293 cells were transfected with each receptor alone or each receptor together with the ␤-arrestin1 dominant negative mutant V53D (V53D), which blocks ␤ 2 AR sequestration (34), or the dynaminI dominant negative mutant K44A (K44A), which blocks clathrin-mediated endocytosis (34). Agonist-induced sequestration of the ␤ 2 AR, V2R, and their chimeras was blocked by overexpression of V53D or K44A (Fig.  1A). V53D was less effective inhibiting sequestration of the V2R (30 Ϯ 4% reduction) compared with the ␤ 2 AR (62 Ϯ 1% reduction). The differential sensitivity to V53D was reversed in the chimeric receptors (31 Ϯ 4% and 58 Ϯ 1% reduction for the ␤ 2 AR-V2R and V2R-␤ 2 AR chimeras, respectively) and presumably reflects a tighter interaction of the endogenous ␤-arrestins with the V2R tail. COS-7 cells express lower levels of endogenous ␤-arrestins than many other cell types and therefore provide an alternative system for assessing the role ␤-arrestin plays in ␤ 2 AR and V2R sequestration (38). COS-7 cells were transfected with each receptor alone or each receptor together with ␤-arrestin1 or ␤-arrestin2 (Fig. 1B). In the absence of overexpressed ␤-arrestin, only a small amount of sequestration (compared with HEK-293 cells) was observed for the ␤ 2 AR, V2R, and their chimeras following a 30-min exposure to agonist. In contrast, overexpression of ␤-arrestin1 or ␤-arrestin2 enhanced the ability of each receptor to sequester to levels similar to that observed in HEK-293 cells. These sequestration studies demonstrate that the ␤ 2 AR and V2R both internalize in a ␤-arrestinand clathrin-mediated pathway.
Cellular Trafficking of ␤-Arrestin with the ␤ 2 AR, V2R, and ␤ 2 AR-V2R and V2R-␤ 2 AR Chimeras-We employed a functional ␤arr2-GFP to visualize whether differences exist in ␤-arrestin trafficking with the ␤ 2 AR and V2R in transfected HEK-293 cells (31). In the absence of agonist, ␤arr2-GFP was evenly distributed throughout the cytoplasm of cells expressing either the ␤ 2 AR or the V2R as indicated by the homogenous ␤arr2-GFP fluorescence (Fig. 2, A and C, 0 min). Addition of isoproterenol promoted the rapid redistribution of ␤arr2-GFP from the cytosol to the ␤ 2 AR at the plasma membrane ( Fig. 2A, 2   FIG. 3. Colocalization of ␤arr2-GFP with the internalized ␤ 2 AR, V2R, and ␤ 2 AR-V2R and V2R-␤ 2 AR chimeras. HEK-293 cells were co-transfected with plasmids containing the cDNA for ␤arr2-GFP and the cDNA for the ␤ 2 AR (A), ␤ 2 AR-V2R chimera (B), V2R (C), or V2R-␤ 2 AR chimera (D). Cells pre-labeled with rhodamine-conjugated anti-HA mouse monoclonal antibody were treated for 15 min at 37°C with 10 M isoproterenol (␤ 2 AR and ␤ 2 AR-V2R chimera) or 100 nM AVP (V2R and V2R-␤ 2 AR chimera). Shown are the confocal visualizations of the receptor immunofluorescence (red) and the ␤arr2-GFP fluorescence (green). Colocalization (yellow) of the receptor with ␤arr2-GFP is indicated in the overlay. min) (28,30). The punctate pattern of ␤arr2-GFP fluorescence at the plasma membrane reflects its localization with the receptor in clathrin-coated pits (11,29,31). Activation of the V2R with AVP also promoted the rapid redistribution of ␤arr2-GFP from the cytoplasm to the receptor at the plasma membrane in the same time frame and with the same punctate pattern as that observed for the ␤ 2 AR (Fig. 2C, 2 min). A more prolonged exposure to agonist produced a striking difference in the trafficking of ␤-arrestin. ␤arr2-GFP remained at the plasma membrane in cells expressing the ␤ 2 AR even after 1 h of agonist treatment ( Fig. 2A, 15 min and data not shown). In contrast, ␤arr2-GFP redistributed within 2 to 3 min of agonist stimulation to endocytic vesicles in cells expressing the V2R and remained in these vesicles even after 1 h of agonist treatment (Fig. 2C, 15 min and data not shown). The differential trafficking of ␤arr2-GFP with the ␤ 2 AR and V2R was completely reversed when the carboxyl-terminal tails of these two receptors were switched. ␤arr2-GFP redistributed to endocytic vesicles in cells expressing the ␤ 2 AR-V2R chimera but remained at the plasma membrane in cells expressing the V2R-␤ 2 AR chimera (Fig. 2, B and D, compare 15 min images). Similar results were found using a functional ␤arr1-GFP (data not shown).
To determine whether ␤-arrestin colocalized with the V2R in endocytic vesicles, we examined the agonist-induced redistribution of the receptor and ␤arr2-GFP in the same living HEK-293 cell. Cell surface receptors were pre-labeled with fluorescent antibodies prior to agonist stimulation. Under these conditions, V2R immunofluorescence was localized at the plasma membrane and ␤arr2-GFP fluorescence was evenly distributed in the cytoplasm (data not shown). Following a 15-min stimulation with AVP, an extensive colocalization (yellow) of the V2R immunofluorescence (red) and the ␤arr2-GFP fluorescence (green) was observed in endocytic vesicles (Fig.  3C). In contrast, ␤arr2-GFP fluorescence (green) did not colocalize with ␤ 2 AR immunofluorescence (red) emanating from endocytic vesicles (Fig. 3A). Switching the carboxyl-terminal tails of these two receptors completely reversed these phenotypes. ␤arr2-GFP colocalized with the ␤ 2 AR-V2R chimera in endocytic vesicles but did not colocalize with vesicles containing the V2R-␤ 2 AR chimera (Fig. 3, B and D). These results demonstrate that the endocytic pathways of the ␤ 2 AR and V2R share a common recruitment of ␤-arrestin to the receptor at the plasma membrane during the initial stages of clathrin-mediated endocytosis but then diverge. The ␤ 2 AR-␤-arrestin complex dissociates at or close to the plasma membrane, and ␤-arrestin is excluded from receptor-bearing endocytic vesicles. In contrast, the V2R-␤-arrestin complex remains intact and is internalized into endocytic vesicles. Moreover, these results demonstrate that the differential trafficking of ␤-arrestin to endosomes is mediated by the carboxyl-terminal tails of these two receptors.
Recycling and Resensitization of the ␤ 2 AR, V2R, and ␤ 2 AR-V2R and V2R-␤ 2 AR Chimeras-The ability of internalized ␤-arrestin-free and ␤-arrestin-complexed receptors to recycle back to the plasma membrane was examined by flow cytometry in transfected HEK-293 cells. Cells were treated with agonist for a period of 30 min to promote receptor sequestration. Agonist was then removed, and the return of sequestered receptors to the cell surface was monitored over time. Sixty minutes after agonist removal, 65 Ϯ 6% of the internalized ␤ 2 AR but only 11 Ϯ 4% of the internalized ␤ 2 AR-V2R chimera recycled back to the plasma membrane (Fig. 4A). Similarly, only 23 Ϯ 5% of the internalized V2R but 89 Ϯ 7% of the internalized V2R-␤ 2 AR chimera recycled back to the plasma membrane 60 min after HEK-293 cells were transfected with plasmids containing the cDNA for the ␤ 2 AR, ␤ 2 AR-V2R chimera, V2R, or V2R-␤ 2 AR chimera. Cells were treated for 30 min at 37°C with or without 10 M isoproterenol (␤ 2 AR and ␤ 2 AR-V2R chimera, A) or 100 nM AVP (V2R and V2R-␤ 2 AR chimera, B). The cells were then washed to remove agonist and maintained on ice or incubated at 37°C for 0, 15, 30, or 60 min. Cell surface receptors were assessed by flow cytometry and are plotted as percentage of total cell surface immunofluorescence measured in control cells not treated with agonist. Data represent the mean Ϯ S.E. of 4 -5 independent experiments. C and D, visualization of ␤arr2-GFP fluorescence after agonist removal. HEK-293 cells were co-transfected with plasmids containing the cDNA for ␤arr2-GFP and the cDNA for the ␤ 2 AR, ␤ 2 AR-V2R chimera, V2R, or V2R-␤ 2 AR chimera. Cells were treated with agonist as described above, washed to remove agonist, and incubated at 37°C for 60 min. The distribution of ␤arr2-GFP fluorescence was visualized in cells immediately before agonist removal (0 min) and immediately after the 60-min recovery period (60 min). Shown are confocal microscopic images of ␤arr2-GFP fluorescence in representative cells.
agonist removal (Fig. 4B). Thus, by switching the carboxylterminal tails of the recycling ␤ 2 AR and nonrecycling V2R and altering their ability to recruit ␤-arrestin into endocytic vesicles, the ability of these two receptors to recycle was completely reversed.
The fate of ␤arr2-GFP following agonist removal was also examined in this "recycling" paradigm for each of the wild-type and chimeric receptors. After a 30-min treatment with agonist, ␤arr2-GFP fluorescence was observed in a punctate pattern at the plasma membrane of cells expressing the recycling ␤ 2 AR and V2R-␤ 2 AR chimera and localized to endocytic vesicles in cells expressing the nonrecycling V2R and ␤ 2 AR-V2R chimera (Fig. 4, C and D, 0 min). The cells were then washed to remove agonist, and ␤arr2-GFP fluorescence was re-evaluated after a 60-min recovery period. In cells expressing the recycling ␤ 2 AR and V2R-␤ 2 AR chimera, ␤arr2-GFP redistributed from the plasma membrane back to the cytoplasm as reflected by the homogeneous ␤arr2-GFP fluorescence (Fig. 4, C and D, 60  min). In contrast, in cells expressing the nonrecycling V2R and ␤ 2 AR-V2R chimera, ␤arr2-GFP remained localized with the receptor in endocytic vesicles (Fig. 4, C and D, 60 min). These results demonstrate that the interaction of ␤-arrestin with the nonrecycling V2R and ␤ 2 AR-V2R chimera is a stable interaction, as it is preserved in endocytic vesicles even 1 h after agonist removal.
We next tested whether the differences in the ability of the wild-type and chimeric receptors to recycle lead to corresponding differences in the ability of these receptors to resensitize. Receptor-expressing HEK-293 cells were treated with vehicle for 15 min (Naive), with agonist for 15 min (Desensitized), or with agonist for 15 min and allowed to recover for 60 min in agonist-free medium (Resensitized) (Fig. 5). Cell membranes were then prepared and agonist-mediated adenylyl cyclase activity was measured for each condition. For both the ␤ 2 AR and ␤ 2 AR-V2R chimera, desensitization was characterized by a decrease in the maximal velocity (V max ) of adenylyl cyclase activity (Fig. 5, A and B). One hour after agonist removal, the recycling ␤ 2 AR had fully resensitized as indicated by the complete recovery in V max (100 Ϯ 3% of V max measured under Naive conditions) (Fig. 5A). In contrast, recovery of the V max for the nonrecycling ␤ 2 AR-V2R chimera was impaired by 66 Ϯ 3% (Fig. 5B). Similar results were obtained for the V2R and V2R-␤ 2 AR chimera. Desensitization was characterized for both receptors by a decrease in V max and a rightward shift in the EC 50 (Fig. 5, C and D). One hour after agonist removal, the recycling V2R-␤ 2 AR chimera fully resensitized as indicated by the complete recovery in V max (102 Ϯ 2% of V max measured under Naive conditions) (Fig. 5D). In contrast, recovery of the V max for the nonrecycling V2R was impaired by 54 Ϯ 1% (Fig. 5C). Thus, differences in the ability of the wild-type and chimeric receptors to recycle lead to corresponding differences in the ability of these receptors to resensitize and re-establish agonist responsiveness.
Phosphorylation and Dephosphorylation of the ␤ 2 AR, V2R, and ␤ 2 AR-V2R and V2R-␤ 2 AR Chimeras-Dissociation of ␤-arrestin from desensitized receptors may be necessary for their dephosphorylation and resensitization (26). As shown in Fig. 6, each of the wild-type and chimeric receptors expressed in HEK-293 cells are phosphorylated after 10 min of agonist treatment. To assess the rate of receptor dephosphorylation, receptor- FIG. 5. Resensitization of the ␤ 2 AR, V2R, and ␤ 2 AR-V2R and V2R-␤ 2 AR chimeras following agonist removal. HEK-293 cells were transfected with plasmids containing the cDNA for the ␤ 2 AR (A), ␤ 2 AR-V2R chimera (B), V2R (C), or V2R-␤ 2 AR chimera (D). Cells were incubated for 15 min at 37°C in the absence (Naive) or presence (Desensitized, Resensitized) of 10 M isoproterenol (␤ 2 AR and ␤ 2 AR-V2R chimera) or 100 nM AVP (V2R and V2R-␤ 2 AR chimera). The cells were then washed to remove agonist and maintained on ice (Naive, Desensitized) or incubated for 1 h at 37°C (Resensitized). Cell membranes were prepared and adenylyl cyclase activity measured under basal conditions and in the presence of increasing concentrations of agonist and 10 M forskolin. Adenylyl cyclase activity was normalized to the forskolin-stimulated cyclase response and is expressed as the percentage of the maximal response in Naive cells. The EC 50 values for ␤ 2 AR-and ␤ 2 AR-V2R-mediated activation of adenylyl cyclase were 91 and 61 nM, respectively. The EC 50 values for V2R-and V2R-␤ 2 AR-mediated activation of adenylyl cyclase were 4.7 and 4.0 nM, respectively. The data represent the mean Ϯ S.E. of four independent experiments performed in duplicate. expressing cells were treated for 10 min with agonist, washed to remove agonist, and either maintained on ice (Desensitized) or returned to a 37°C incubator for 30 or 60 min (Resensitized) (Fig. 7, A and B). A 48 Ϯ 5% reduction in the phosphorylation of the recycling ␤ 2 AR and a 67 Ϯ 7% reduction in the phosphorylation of the recycling V2R-␤ 2 AR chimera were observed 60 min after agonist removal. In contrast, very little dephosphorylation was observed for the nonrecycling V2R (3 Ϯ 6% decrease) and no dephosphorylation was observed for the nonrecycling ␤ 2 AR-V2R chimera (12 Ϯ 13% increase) after the 60min recovery period. These data suggest that the stability of the ␤-arrestin interaction with the carboxyl-terminal tail of GPCRs dictates the rate of receptor dephosphorylation.

Identification of Residues within the V2R Carboxyl Terminus That Allow ␤-Arrestin to Remain Associated with Receptors in
Endocytic Vesicles-To identify residues in the V2R tail that stabilize the receptor's interaction with ␤-arrestin, mutations were made in the putative phosphate acceptor sites (Fig. 8A). These mutant receptors expressed in HEK-293 cells sequestered to levels similar to that observed for the wild-type V2R and induced translocation of ␤arr2-GFP to the plasma mem-brane upon agonist activation (data not shown). Removal of the two clusters of serine/threonine residues contained within the last 10 amino acids of the V2R tail, either by truncation or alanine substitution, produced mutant receptors (V2R-S362X and V2R-SSSTSS/AAAAAA) that did not recruit ␤-arrestin into endocytic vesicles (Fig. 8A, images 2 and 3, respectively). Removal of the cluster of serine/threonine residues at the end of the V2R tail by alanine substitution produced a mutant receptor (V2R-TSS/AAA) that still recruited ␤-arrestin into endocytic vesicles (Fig. 8A, image 4). However, removal of the more proximal cluster of serine residues (Ser-362, Ser-363, and Ser-364) by alanine substitution produced a mutant receptor (V2R-SSS/AAA) that failed to recruit ␤-arrestin into endocytic vesicles (Fig. 8A, image 5). Similar results were found when this serine cluster was mutated to alanines in the ␤ 2 AR-V2R chimera (␤ 2 AR-V2R-SSS/AAA) (Fig. 8A, image 6). The importance of this serine cluster for recruiting ␤-arrestin into endocytic vesicles was further tested by adding only the last 10 amino acids of the V2R to the end of the full-length ␤ 2 AR. ␤arr2-GFP translocated to this mutant receptor (␤ 2 AR413-V2R10) at the plasma membrane upon agonist activation but did not inter- FIG. 7. Dephosphorylation of the ␤ 2 AR, V2R, and ␤ 2 AR-V2R and V2R-␤ 2 AR chimeras following agonist removal. HEK-293 cells were transfected with plasmids containing the cDNA for the ␤ 2 AR, ␤ 2 AR-V2R chimera, V2R, or V2R-␤ 2 AR chimera. Cells were treated for 10 min at 37°C with or without 10 M isoproterenol (␤ 2 AR and ␤ 2 AR-V2R chimera, A) or 100 nM AVP (V2R and V2R-␤ 2 AR chimera, B). The cells were then washed to remove agonist and either maintained on ice (Desensitized, D) or incubated at 37°C for 30 min (Resensitized, R30) or 60 min (Resensitized, R60). Receptors were then immunoprecipitated and assayed for phosphorylation. Upper panel shows a representative autoradiograph, and lower panel shows the mean Ϯ S.E. of three independent experiments quantified by PhosphorImager analysis. Values are expressed as the percentage of phosphorylation measured for the desensitized receptor.
nalize with the receptor into endocytic vesicles (Fig. 8A, image  8). Similar results were found when the last 29 amino acids of the V2R were added to the end of the full-length ␤ 2 AR (data not shown). However, when the last 10 amino acids of the V2R were positioned closer to the putative palmitoylated cysteine of the ␤ 2 AR, the mutant receptor (␤ 2 AR360-V2R10) gained the ability to recruit ␤-arrestin into endocytic vesicles (Fig. 8A,  image 9). These findings identify a cluster of three serine residues located in the V2R carboxyl terminus that mediate the trafficking of ␤-arrestin with the V2R into endocytic vesicles. Moreover, they suggest that the position of the serine cluster within the receptor carboxyl-terminal tail is critical for the formation of a stable receptor-␤-arrestin complex that internalizes into endocytic vesicles.
Whole cell phosphorylation assays were performed on HEK-293 cells expressing the wild-type or mutant V2Rs in order to assess whether the proximal cluster of three serine residues is actually phosphorylated. As shown in Fig. 9, agonist-induced phosphorylation of the V2R-SSSTSS/AAAAAA mutant, in which both the proximal and distal clusters of serine/threonine residues are mutated, was reduced 86.3 Ϯ 1.4% compared with the wild-type V2R. Agonist-induced phosphorylation of the V2R-TSS/AAA mutant, in which only the distal cluster of serine/threonine residues is mutated, was reduced 4.7 Ϯ 6.9%. However, agonist-induced phosphorylation of the V2R-SSS/ AAA mutant, in which only the proximal cluster of three serine residues is mutated, was reduced 84.2 Ϯ 0.6%. Therefore, the proximal cluster of three serine residues, which mediates the formation of stable V2R-␤-arrestin complexes that internalize into endocytic vesicles, is the principal site of V2R phosphorylation.

DISCUSSION
In the present study, we evaluated the role of ␤-arrestin in the resensitization of the ␤ 2 AR and V2R. Both these receptors, upon agonist activation, recruit ␤-arrestin to the plasma membrane and internalize in a ␤-arrestinand clathrin-dependent fashion. The ␤ 2 AR-␤-arrestin complex dissociates at or near the plasma membrane, presumably allowing the receptor to associate rapidly with receptor phosphatases and be dephosphorylated in the endosomal compartment. The dephosphorylated ␤ 2 AR is then recycled rapidly back to the plasma membrane, and normal receptor responsiveness is re-established. The more stable V2R-␤-arrestin complex does not dissociate at the plasma membrane and is internalized into endosomes. Preservation of this complex presumably prevents the association of this receptor with receptor phosphatases. Consequently, the V2R is not dephosphorylated efficiently in endosomes and re- FIG. 8. Cellular trafficking of ␤arr2-GFP with the V2R into endocytic vesicles is mediated by a cluster of three serine residues at the V2R carboxyl terminus. A, upper panel shows the amino acid composition of the carboxyl-terminal tails of the V2R, ␤ 2 AR, and various mutant receptors beginning with the putative sites of palmitoylation in bold (Cys-342 for the V2R constructs 1-5, and Cys-341 for the ␤ 2 AR constructs 6 -9). Underlined are the mutations made by alanine substitution and the last 10 amino acids of the V2R tail when added to the ␤ 2 AR. Lower panel shows the distribution of ␤arr2-GFP fluorescence in HEK-293 cells expressing wild-type or mutant receptors following a 30-min treatment at 37°C with 100 nM AVP (images 1-5) or 10 M isoproterenol (images 6 -9). The number on each image corresponds to the number of the transfected receptor construct shown in the upper panel. B, amino acid composition of the carboxyl-terminal tails of the V2R, NTR1, and AT1AR beginning with the conserved NPXXY motif in bold. Clusters of serine/threonine residues are underlined, and putative sites of palmitoylation are indicated with an asterisk (*). cycles only slowly back to the plasma membrane. The molecular determinants that stabilize the receptor-␤-arrestin complex appear to reside in the receptor carboxyl terminus since switching the tails of the ␤ 2 AR and V2R completely reverses their dephosphorylation, recycling, and resensitization kinetics. For the V2R, a cluster of three serine residues located in the carboxyl terminus and serving as the principal site for GRKmediated phosphorylation determines whether a stable V2R-␤-arrestin complex forms. Moreover, this same serine cluster plays a critical role in preventing V2R recycling (20). Therefore, the interaction of ␤-arrestin with phosphorylated serine/threonine clusters in the carboxyl-terminal tail of GPCRs regulates the rate of GPCR dephosphorylation, recycling, and resensitization by presumably impeding the interaction between GPCRs and receptor phosphatases.
␤-Arrestins bind with high affinity to agonist-activated, GRK-phosphorylated GPCRs in what is thought to involve the simultaneous engagement of two regions of the ␤-arrestin molecule with two corresponding regions of the receptor (28,39,40). A large region within the amino-terminal half of ␤-arrestin (termed the activation-recognition domain) recognizes the agonist-activated state of GPCRs. This domain of ␤-arrestin appears to bind the third cytoplasmic loop of a variety of GPCRs, including the m2 and m3 muscarinic acetylcholine receptors and the ␣2A adrenergic receptor (41). The amino-terminal half of ␤-arrestin also contains a positively charged, smaller domain of approximately 20 amino acids (termed the phosphorylationrecognition domain; net charge ϩ7), which appears to interact with the GRK-phosphorylated carboxyl termini of GPCRs (42). GRK-mediated phosphorylation of the serine cluster in the tail of the V2R would produce a localized concentration of negative charges. Therefore, the strong ionic interaction between the cluster of negative charges in the V2R tail and the cluster of positive charges in the phosphorylation-recognition domain of ␤-arrestin might be responsible for promoting the formation of a stable receptor-␤-arrestin complex that internalizes into endocytic vesicles. Consistent with this hypothesis, mutation of the serine cluster in the V2R-SSS/AAA and V2R-SSSTSS/ AAAAAA mutants results in a marked reduction (approximately 85%) in the level of agonist-induced phosphorylation of these receptors as well as the inability of these receptors to recruit ␤-arrestin into endocytic vesicles. Nevertheless, these mutant receptors still exhibit a small amount of agonist-induced phosphorylation, a robust recruitment of ␤-arrestin to the plasma membrane, and normal levels of sequestration. This suggests that other serine/threonine residues in the V2R may also serve as substrates for GRK-mediated phosphorylation, thus enabling ␤-arrestin to interact, albeit more weakly, with other intracellular regions of the agonist-occupied V2R.
We have recently demonstrated that ␤-arrestins dissociate from the dopamine D1A receptor (D1AR) and the endothelin type A receptor (ETAR) at the plasma membrane but traffic with the neurotensin receptor 1 (NTR1) and the angiotensin II type 1A receptor (AT1AR) into endocytic vesicles (29). A comparison of the carboxyl-terminal tails of these receptors with the ␤ 2 AR and V2R reveals striking similarities with respect to the absence or presence of clusters of putative phosphate acceptor sites (defined as serine/threonine residues occupying three consecutive positions or three out of four positions). Clusters of putative phosphorylation sites are noticeably absent from the carboxyl-terminal tails of the ␤ 2 AR, D1AR, and ETAR even though potential phosphate acceptor sites are abundant. In contrast, clusters of putative phosphorylation sites are not only present in the V2R, NTR1, and AT1AR carboxyl-terminal tails but also occupy similar positions with respect to the conserved NPXXY motif (marking the end of the seventh trans-membrane domain) and the end of the receptor (Fig. 8B). Evidence that the location of the cluster of putative phosphate acceptor sites is critical for the formation of a stable receptor-␤-arrestin complex comes from our finding that the ␤ 2 AR gains the ability to recruit ␤-arrestin into endocytic vesicles when the V2R serine cluster is added to the ␤ 2 AR truncated at position 360 but not when added to the full-length ␤ 2 AR. Clearly, an important goal of future studies will be to test whether the clusters of putative phosphate acceptor sites found in the tails of other GPCRs are critical to the recruitment of ␤-arrestin with these receptors into endocytic vesicles and to test whether these complexes can interact with other accessory proteins or signaling molecules.
In summary, ␤-arrestins are multifunctional proteins that play a pleiotropic role in regulation of GPCR responsiveness. They regulate GPCR desensitization by uncoupling receptors from their cognate G proteins and they regulate GPCR sequestration by targeting desensitized receptors to clathrin-coated pits for endocytosis. We now show that ␤-arrestins, by interacting with a specific motif in the carboxyl-terminal tails of GPCRs, dictate the rate of receptor dephosphorylation, recycling, and resensitization. Thus, the stability with which ␤-arrestins associate with GPCRs controls the intracellular trafficking of receptors and consequently the re-establishment of physiological responsiveness following desensitization.