Agonist-induced Formation of Opioid Receptor-G Protein-coupled Receptor Kinase (GRK)-Gβγ Complex on Membrane Is Required for GRK2 Function in Vivo

G protein-coupled receptor kinases (GRKs) catalyze agonist-induced receptor phosphorylation on the membrane and initiate receptor desensitization. Previous in vitro studies have shown that the binding of GRK to membrane-associated Gβγ subunits plays an important role in translocation of GRK2 from the cytoplasm to the plasma membrane. The current study investigated the role of the interaction of GRK2 with the activated δ-opioid receptor (DOR) and Gβγ subunits in the membrane translocation and function of GRK2 using intact human embryonic kidney 293 cells. Our results showed that agonist treatment induced GRK2 binding to DOR, GRK2 translocation to the plasma membrane, and DOR phosphorylation in cells expressing the wild-type DOR but not the mutant DOR lacking the carboxyl terminus, which contains all three GRK2 phosphorylation sites. DORs with the GRK2 phosphorylation sites modified (M3) or with the acidic residues flanking phosphorylation sites mutated (E355Q/D364N) failed to be phosphorylated in response to agonist stimulation. Agonist-induced GRK2 membrane translocation and GRK-receptor association were observed in cells expressing M3 but not E355Q/D364N. Moreover, over-expression of Gβγ subunits promoted GRK2 binding to DOR, whereas over-expression of transducin α or the carboxyl terminus of GRK2 blocked binding. Further study demonstrated that agonist stimulation induced the formation of a complex containing DOR, GRK2, and Gβγ subunits in the cell and that agonist-stimulated formation of this complex is essential for the stable localization of GRK2 on the membrane and for its catalytic activity in vivo.

G protein-coupled receptor kinases (GRKs) catalyze agonist-induced receptor phosphorylation on the membrane and initiate receptor desensitization. Previous in vitro studies have shown that the binding of GRK to membrane-associated G␤␥ subunits plays an important role in translocation of GRK2 from the cytoplasm to the plasma membrane. The current study investigated the role of the interaction of GRK2 with the activated ␦-opioid receptor (DOR) and G␤␥ subunits in the membrane translocation and function of GRK2 using intact human embryonic kidney 293 cells. Our results showed that agonist treatment induced GRK2 binding to DOR, GRK2 translocation to the plasma membrane, and DOR phosphorylation in cells expressing the wild-type DOR but not the mutant DOR lacking the carboxyl terminus, which contains all three GRK2 phosphorylation sites. DORs with the GRK2 phosphorylation sites modified (M3) or with the acidic residues flanking phosphorylation sites mutated (E355Q/D364N) failed to be phosphorylated in response to agonist stimulation. Agonistinduced GRK2 membrane translocation and GRK-receptor association were observed in cells expressing M3 but not E355Q/D364N. Moreover, over-expression of G␤␥ subunits promoted GRK2 binding to DOR, whereas overexpression of transducin ␣ or the carboxyl terminus of GRK2 blocked binding. Further study demonstrated that agonist stimulation induced the formation of a complex containing DOR, GRK2, and G␤␥ subunits in the cell and that agonist-stimulated formation of this complex is essential for the stable localization of GRK2 on the membrane and for its catalytic activity in vivo.
G protein-coupled receptors (GPCRs) 1 constitute a superfamily of plasma membrane receptors. More than 1,000 of 19,000 open reading frames in the genome of Caenorhabditis elegans (1) encode GPCRs, and more than 600 GPCR genes have been identified in the human genome (2). GPCRs transduce a huge number of extracellular signals from hormones, neurotransmitters, chemokines, and other environmental stimuli to the interior of cells and thus play fundamental roles in regulating a variety of cellular functions (3,4). An important feature of GPCR-mediated signal transduction is that repeated agonist stimulation triggers a negative feedback regulatory mechanism that attenuates GPCR-mediated signal transduction (desensitization). The initial event of receptor desensitization is agonist-stimulated receptor phosphorylation catalyzed by GPCR kinases (GRKs) (5). Seven members of the GRK family have been identified to date, and they have been divided on the basis of structural and functional similarities into rhodopsin kinase (GRK1 and GRK7), the ␤-adrenergic receptor (␤AR) kinase (GRK2 and GRK3), and the GRK4 (GRK4, GRK5, and GRK6) subfamilies (6 -8).
The functions of GRKs are highly regulated in the cell. GRKs preferentially catalyze phosphorylation of the activated (agonist-occupied) rather than the inactive or antagonist-occupied GPCR substrates (9). The interactions of GRKs with the activated receptor substrates in turn potently activate GRKs (10). The participation of regulatory mechanisms responsible for the membrane localization and receptor targeting of GRKs is required for agonist-induced GPCR phosphorylation. Studies revealed that the members of the ␤AR kinase subfamily of GRKs (GRK2 and GRK3) exhibit an agonist-dependent association with cell membranes and that the agonist-induced GRK2 translocation to plasma membrane precedes receptor phosphorylation and desensitization (11). Studies using purified components demonstrated that the ␤␥ subunits of the heterotrimeric G proteins (G␤␥) interact with the carboxyl tail of GRK2 and promote the association of GRK2 with lipid vesicles and rod outer segment membranes. Furthermore, this interaction stimulated the phosphorylation of rhodopsin and ␤AR in vitro (12)(13)(14)(15). The association of GRK2 and G␤␥ has been demonstrated using purified GRK2 in vitro and recently using coimmunoprecipitation in vivo (14,16). These studies proposed the following model: following the release of the free G␤␥ dimer led by agonist occupancy of a GPCR, GRK2 binds to the membrane-bound G␤␥ and subsequently targets GRK2 to the activated GPCR substrate. This model remains to be tested in vivo, and the detailed mechanism of GRK2 translocation to plasma membrane must still be developed.
Our earlier work demonstrated that GRK2 phosphorylates the ␦-opioid receptor (DOR) upon opioid challenge and results in desensitization of DOR. This work also identified the GRK2 phosphorylation sites in DOR (17)(18)(19)(20). The current study explores the mechanism of GRK2-catalyzed GPCR phosphorylation in a mammalian cellular system. We have demonstrated for the first time the formation of a stable G␤␥-GRK2-DOR complex on the membrane in response to agonist stimulation. Furthermore, we demonstrate the requirement of the agonistinduced formation of the receptor-GRK2-G␤␥ complex for the stable membrane association and function of GRK2 in vivo.
Plasmid Construction-Plasmids encoding bovine GRK2, GRK5, and HA-tagged mouse DORs including the wild-type ␦-opioid receptor (WT), the carboxyl-terminal 31-residue truncated DOR (⌬31), and the mutant T358A/T361A/S363G (M3) were prepared as described previously (18 -20). The HA-tagged mouse DOR mutants E355Q, D364N, and E355Q/ D364N were constructed by PCR mutagenesis, and the authenticity of the sequences was confirmed by DNA sequencing. Bovine GRK3 cDNA was provided by Dr. Lin Li (Shanghai Institutes of Life Sciences, Chinese Academy of Sciences) and was subcloned into pcDNA3 (Invitrogen) with a sequence encoding a FLAG tag to generate GRK3 with an amino-terminal tag (FLAG-GRK3). The GRK2-GFP pcDNA construct was a generous gift from Dr. Marc G. Caron (Duke University Medical Center).
Cell Culture and Transfection-HEK293 cells cultured in MEM containing 10% fetal bovine serum were seeded in 60-or 100-mm tissue culture dishes at 1-3 ϫ 10 6 /dish for 20 h before transfection. In immunofluorescence experiments, cells were seeded in 6-well plates or 35-mm tissue culture dishes. Plasmids (1-5 g each) were transfected into the cells using the calcium phosphate method. The cells were used 44 h after transfection, and the expression of opioid receptors was measured by a radioligand binding assay as described previously (17). The levels of the opioid receptors expressed were 2-3 pmol/mg of protein.
Receptor Phosphorylation-Measurement of opioid receptor phosphorylation was carried out as described previously (17,18). Briefly, the cells were metabolically labeled at 37°C for 60 min with 60 Ci/ml of 32 P i in phosphate-free Dulbecco's MEM, and 1 M okadaic acid or 5 M cantharidin was added 3 min before the end of labeling. Then the cells were exposed to 1 M DPDPE at 37°C for 10 min and solubilized at 4°C for 1.5 h in buffer (1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 5 mM EDTA, 10 mM NaF, 10 mM disodium pyrophosphate, 5 M cantharidin, 10 g/ml aprotinin, 10 g/ml benzamidine, and 0.2 mM phenylmethylsulfonyl fluoride). HA-DORs were immunoprecipitated with 12CA5, and the immunoprecipitation complexes were analyzed on polyacrylamide gels. After drying, the gels were subjected to quantitative analysis with a Phosphor-Imager (Amersham Biosciences).
[ 35 S]GTP␥S Binding Assay-The experiments were performed as described previously (21). Briefly, the cells were incubated in the presence or absence of 1 M DPDPE at 37°C for 10 min. The cells were then lysed in 5 mM Tris-HCl, pH 7.5, 5 mM EDTA, and 5 mM EGTA at 4°C, and the lysates were centrifuged at 30,000 ϫ g for 10 min. The membrane pellet was resuspended, and an aliquot (containing 8 g of protein) was incubated at 30°C for 1 h in 50 mM Tris-HCl, pH 7.5, 5 mM MgCl 2 , 1 mM EGTA, 100 mM NaCl, 40 M GDP, and 8 nM [ 35 S]GTP␥S (1200 Ci/mmol) in the presence or absence of 1 mM DPDPE. The reaction was terminated by dilution in cold phosphate-buffered saline (PBS) and filtered through GF/C filters under vacuum. Bound radioactivity was determined in duplicate by liquid scintillation spectrophotometry. Basal binding was determined in the absence of agonist, and nonspecific binding was determined in the presence of 10 M nonradioactive GTP␥S. The stimulation of [ 35 S]GTP␥S binding was calculated as a percentage as follows: 100 ϫ [(cpm sample Ϫ cpm nonspecific )/(cpm basal Ϫ cpm nonspecific )].
Co-immunoprecipitation and Western Procedures-HEK293 cells in 100-mm dishes were incubated at 37°C in serum-free MEM in the presence or absence of 1 M DPDPE for 3 min. At the end of the incubation, the cells were washed twice with ice-cold PBS and lysed in 1 ml of cold lysis buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% Triton X-100, 10 g/ml leupeptin, 10 g/ml aprotonin, and 1 mM phenylmethylsulfonyl fluoride) as described (22). The lysate was centrifuged, and the supernatant was incubated with 1 g of 12CA5 and 10 l of 50% slurry of protein A-Sepharose beads at 4°C for 4 h. The beads were subsequently washed three times with lysis buffer. The proteins bound to the beads were eluted in SDS-PAGE sample buffer and separated by SDS-PAGE. The presence of HA-DOR, GRK, and G␤␥ in the immunocomplexes was detected in the subsequent Western procedure with antibody specifically against HA, GRK2, FLAG, GRK5, or G␤, respectively. The immunoblots were visualized using an ECL kit (Amersham Biosciences) following the manufacturer's suggested protocol. An aliquot (2.5%) of the cell lysate was analyzed on Western blots without immunoprecipitation to quantify expression level of the protein studied.
Laser Confocal Fluorescence Microscopy-After incubation in serumfree MEM at 37°C for 60 min, the cells grown on coverslips were incubated in the presence or absence of DPDPE at 37°C for 3 min and fixed. The cells were then incubated with 12CA5 (1 g/ml) in PBS, 3% bovine serum albumin for 2 h at room temperature and washed with PBS. The cells were incubated with Texas Red-conjugated goat antimouse antibody (1 g/ml) at room temperature for 1 h and washed with PBS, and the coverslips were mounted onto microscope slides with 50% glycerol. Scanning images were recorded with a TCS NT laser confocal microscope (Leica Microsystems, Bensheim, Germany).
Statistical Analysis-Data were analyzed using the Student's t test for comparison of independent means with pooled estimates of common variances.

RESULTS
Our earlier study demonstrated that agonist-induced phosphorylation of DOR occurs on the cytoplasmic carboxyl-terminal domain of the receptor and is mediated by GRK2 (18,19). These studies suggested that the cytoplasmic tail of DOR is also important for interaction with GRK, although interaction between GRK2 and DOR has not been demonstrated directly. As shown in Fig. 1A, stimulation of the HEK293 cells expressing the wild-type DOR with DPDPE, a specific agonist of DOR, induced DOR phosphorylation. Co-expression of GRK2 strongly enhanced agonist-dependent phosphorylation of DOR. In contrast, truncated DOR lacking the carboxyl-terminal 31 residues (⌬31), although possessing unchanged surface expression and G protein coupling (18,19), failed to be phosphorylated upon DPDPE challenge even in cells over-expressing GRK2. After stimulation with the DOR-specific agonist DPDPE, GRK2 could be co-immunoprecipitated with DOR in cells expressing DOR WT and GRK2 (Fig. 1B). In contrast, no GRK2 was detected in the receptor complex in the absence of agonist challenge. Furthermore, the agonist-dependent association of GRK and DOR was time-dependent and paralleled receptor phosphorylation (data not shown). Binding of GRK2 to the receptor immunoprecipitation complex was detected 1 min following agonist stimulation, and maximal binding was reached in 3-5 min. The binding decreased 5 min after stimulation and was not detectable after 10 min. DADLE, another ␦-specific agonist, but not DAMGO, a -specific opioid agonist, could stimulate the formation of DOR-GRK2 complex in cells expressing DOR, and agonist-induced formation of DOR-GRK2 complex could be completely blocked by the opioid-specific antagonist naloxone (Fig. 1B). In contrast to observations with the wild-type DOR, no GRK2 was co-immunoprecipitated with the mutant receptor ⌬31. This mutant lacks the cytoplasmic tail and all of the GRK2 phosphorylation sites (Fig. 1B). These data demonstrate directly that agonist stimulation induces GRK2 binding to DOR and that the 31 amino acid residues in the carboxyl-terminal domain of the receptor are critical for its interaction with GRK.
HEK293 cells then were transfected with GRK2-GFP and DOR, and the subcellular distribution of GRK upon agonist stimulation was examined using a laser confocal microscope. As shown in Fig 1C, GRK2 was uniformly distributed in the cytoplasm in the absence of agonist stimulation, whereas it translocated rapidly to plasma membrane and colocalized with DOR upon challenge with DPDPE. The real-time recording of GRK2-GFP confocal fluorescence images in living cells co-expressing GRK2-GFP and the wild-type DOR showed that, in response to agonist challenge, the membrane-associated GRK2-GFP fluorescence increased quickly, whereas the cytoplasmic GRK2-GFP fluorescence decreased. This redistribution was accompanied by apparent changes in membrane shape (data not shown). However, agonist-induced GRK2-GFP translocation to cellular membranes was impaired in the cells coexpressing ⌬31, the mutant receptor incapable of binding GRK2 and of being phosphorylated (Fig. 1C). These data demonstrate that the carboxyl-terminal domain of DOR is required not only for agonist-stimulated GRK2-receptor binding but also for agonist-stimulated GRK2 membrane localization.
The agonist-induced membrane translocation of GRK2 precedes GPCR phosphorylation. Earlier studies from Loh and co-workers (23) and our group (19) revealed that GRK-related phosphorylation sites (Thr 358 and Ser 363 ) are within the carboxyl-terminal 31 amino acid residues of DOR (19, 23). As shown in Fig. 2A, DOR M3, in which all of the serine and FIG. 1. Cytoplasmic tail of DOR is required for GRK2 binding, translocation to membrane, and receptor phosphorylation. HEK293 cells were transfected with WT or ⌬31 alone or GRK2 plus WT or ⌬31 cDNAs as indicated. A, the cells were metabolically labeled with 32 P i . Left, cells were incubated in the presence or absence of 1 M DPDPE (DP) as indicated at 37°C for 10 min. right, cells co-transfected with or without GRK2 cDNA as indicated were incubated in 1 M DPDPE at 37°C for 10 min. HA-DORs were then immunoprecipitated with 12CA5, resolved on 8% SDS-polyacrylamide gels, and subjected to phosphorimaging analysis. B, the transfected cells were incubated in the presence or absence of 1 M DPDPE, DADLE, or -specific opioid agonist DAMGO or 5 M naloxone (Nal) followed by 1 M DPDPE at 37°C for 3 min, and the cell lysate was prepared. Immunoprecipitation of the receptor complex was carried out using 12CA5. Western analysis was done using antibody against GRK2 (top panel), and the same blot was reprobed with 12CA5 against HA-DOR after stripping (middle panel). Direct Western analysis of the cell lysates using GRK2 antibody for GRK2 expression is shown (bottom panel). C, GRK-GFP co-transfected cells were incubated in the presence or absence of 1 M DPDPE at 37°C for 3 min. The cells were fixed, stained with 12CA5 and Texas Red-conjugated secondary antibody, and analyzed under a laser confocal fluorescence microscope. The image overlay was obtained by superimposition of the corresponding GRK-GFP green fluorescence and HA-DOR red fluorescence images; the yellow in the image overlay represents a very close colocalization of GRK and DOR. threonine phosphorylation sites in GRK have been replaced by alanines, failed to be phosphorylated following treatment with DPDPE. The E355Q and D364N mutants of DOR bearing single point mutations at the acidic residues adjacent to two GRK phosphorylation sites (Thr 358 and Ser 363 ) showed greatly impaired phosphorylation, and a double mutant (E355Q/D364N) completely blocked receptor phosphorylation ( Fig. 2A). Overexpression of GRK2 partially rescued the impaired phosphoryl- ation capability of E355Q and D364N, but it did not have any significant effect on E355Q/D364N (Fig. 2B). Ligand (data not shown) and [ 35 S]GTP␥S (Fig. 2B) binding experiments showed that these mutated receptors were able to bind agonists, be activated, and couple to G protein normally. Thus, the impaired phosphorylation of M3, E355Q, D364N, and E355Q/D364N is not likely to be due to a deficiency in receptor activation and signaling. Furthermore, the E355Q/D364N double mutant as well as mutation of all the GRK phosphorylation sites on DOR resulted in remarkable inhibition of agonist-induced receptor desensitization (Fig. 2B). The above results indicate that the acidic residues near receptor phosphorylation sites play a critical role in GRK-mediated receptor phosphorylation.
To explore the contributions of the GRK phosphorylation sites and the adjacent acidic residues on agonist-stimulated GRK2 membrane localization, GRK2-DOR interaction, and GRK-catalyzed DOR phosphorylation, the wild-type and mutant DORs were co-expressed with GRK2 in HEK293 cells. As shown in Fig. 1C, DPDPE treatment stimulated membrane redistribution of GRK2-GFP in the cells expressing M3, which lacks all of the GRK phosphorylation sites. Furthermore, the level of GRK2 membrane translocation was comparable with that in the cells expressing wild-type DOR. Co-immunoprecipitation experiments revealed that in response to opioid agonist stimulation, M3 interacted with GRK2 as efficiently as the wild-type receptor (Fig. 2, C and D), although it could not be phosphorylated ( Fig. 2A). On the other hand, DPDPE-stimulated GRK2 membrane association and colocalization with the opioid receptor was entirely blocked in cells expressing the E355Q/D364N double mutant (Fig. 1C). Both the E355Q and D364N mutant showed a 50 -60% reduction in agonist-induced GRK2 binding in co-immunoprecipitation experiments, whereas the E355Q/D364N double mutant showed a total loss of activity in this assay (Fig. 2, C and D). The above results indicate that those GRK serine and threonine phosphorylation sites on DOR are not critically involved in the interaction with GRK2 and in the membrane association of GRK2 induced by agonist. In contrast, the glutamic and aspartic acid residues near these phosphorylation sites are critical. The data also hint that the GRK2-DOR interaction is required not only for DOR phosphorylation but also for stable GRK2 membrane localization.
Agonist-stimulated DOR phosphorylation is regulated by both the ␤AR kinase (GRK2 and GRK3) and GRK4 (GRK4, GRK5, and GRK6) subfamilies of GRKs. The function of Glu 355 and Asp 364 of DOR in the interaction with other members of the GRK family was examined in HEK293 cells co-expressing DORs and either GRK3 or GRK5. The co-immunoprecipitation experiments revealed that similar to GRK2, GRK3 and GRK5 co-immunoprecipitated with the wild-type DOR in response to opioid agonist stimulation. Both of these kinases showed reduced binding to E355Q and D364N and a total loss of binding to E355Q/D364N (Fig. 3). Furthermore, E355Q/D364N was not phosphorylated by GRK5 upon DPDPE stimulation (data not shown). The above results provide the initial evidence demonstrating the interaction of DOR with GRK3 and GRK5, in addition to GRK2. These data also indicate that the acidic residues Glu 355 and Asp 364 adjacent to phosphorylation sites on DOR are critically involved in the interaction with GRKs in both the ␤AR kinase and GRK4 families.
Binding of GRK2 to membrane-associated free G␤␥ subunits is required for the activation and membrane localization of GRK2 (12). Thus the roles of GRK2-G␤␥ interaction in agonistdependent GRK2 membrane translocation and in the interaction with the receptor further were investigated. GRK2 was co-expressed in HEK293 cells with the wild-type or mutant DORs and G␤ 2 ␥ 2 subunits, the carboxyl-terminal domain of GRK2 (GRK2ct), or transducin ␣. As shown in Fig. 4, after

FIG. 2-continued
Agonist-induced Formation of Receptor-GRK2-G␤␥ Complex incubation of HEK293 cells co-expressing GRK2 and the wildtype DOR with DPDPE, both GRK2 and G␤␥ were detected in the DOR immunoprecipitation complex. Over-expression of G␤␥ subunits in these cells increased both basal and agoniststimulated DOR-GRK2 interaction and DOR-GRK2-G␤␥ complex formation. On the other hand, over-expression of GRK2ct or the ␣ subunit of transducin, which reduces free G␤␥, greatly inhibited DOR-GRK2 interaction and DOR-GRK2-G␤␥ complex formation, (Fig. 4). These results indicate the essential role of the G␤␥ subunits in the interaction between DOR and GRK2 and in the agonist-dependent formation of a DOR-GRK2-G␤␥ complex. Moreover, the mutant receptor E355Q/ D364N, which is deficient in binding GRK2, could not form a complex with the G␤␥ subunits; this deficiency was not rescued by over-expression of G␤␥ (Fig. 4). These observations suggest that the formation of a complex of DOR and G␤␥ requires an interaction between DOR and GRK2 (i.e. GRK2 binds both receptor and G␤␥ to form a DOR-GRK2-G␤␥ complex).

DISCUSSION
The current study has demonstrated that upon agonist stimulation, GRK2 forms a complex with G␤␥ subunits and agonistoccupied opioid receptor on the membrane and that G␤␥ subunits play a critical role in the interaction of GRK with activated opioid receptor in vivo. Our results from immunoprecipitation and laser confocal fluorescent microscopy experiments indicate that agonist-stimulated formation of DOR-GRK2-G␤␥ complex is essential for the stable localization of GRK2 on the membrane and its catalytic activity toward GPCR substrate in vivo.
On the basis of these observations, a refined model for GRK2-catalyzed GPCR phosphorylation is proposed in Fig. 5. In this model, receptor activation induces GRK2 interaction with the free G␤␥ subunits associated with the membrane. This interaction is followed immediately by the binding of the GRK2 in the GRK2-G␤␥ binary complex to the activated receptor to form a stable receptor-GRK2-G␤␥ complex on the membrane. This process targets GRK2 to its substrate and results in the subsequent phosphorylation of the activated receptor. A low level of agonist-stimulated GRK2 membrane translocation was observed in cells over-expressing E355Q/D364N and G␤␥ (data not shown), suggesting the presence of GRK2-G␤␥ binary complexes on the membrane in the presence of excess free G␤␥. The agonist-induced GRK2 interaction with G␤␥ may not be as stable as the receptor-GRK-G␤␥ complex and may be converted into the receptor-GRK2-G␤␥ complex as soon as it is formed on the plasma membrane. The agonist-stimulated receptor-GRK-G␤␥ complex may contain other molecules such as scaffold protein ␤-arrestins, which have a high affinity for GPCR phosphorylated by GRKs and have been demonstrated to be associated with the activated receptors (23). The presence of additional scaffold proteins may increase the stability of the receptor-GRK-G␤␥ complex on the membrane. Furthermore, of the three functional domains of GRK2, the amino-terminal domain is proposed to mediate receptor interaction; the central domain to exert catalytic function; and the carboxyl-terminal domain, containing a pleckstrin homology domain essential for interactions with the G␤␥ subunits and phosphatidylinositol 4,5-bisphosphate, to be involved in agonist-stimulated GRK FIG. 3. Effects of acidic residues in interaction of GRK3 and GRK5 with DOR. HEK293 cells were transfected with WT, E355Q, D364N, or E355Q/E364N alone or co-transfected with or without GRK3 or GRK5 cDNA as indicated and harvested 48 h post-transfection. A, the transfected cells were incubated in the presence or absence of 1 M DPDPE (DP) at 37°C for 3 min, and the cell lysate was prepared. Immunoprecipitation of the receptor complex was carried out using 12CA5. Western analysis was done using antibody against GRK3 (top panel), and the same blot was reprobed with 12CA5 against HA-DOR after stripping (middle panel). Direct Western analysis of the cell lysates using GRK3 antibody for GRK3 expression is shown (bottom panel). B, the transfected cells were incubated in the presence or absence of 1 M DPDPE at 37°C for 3 min, and the cell lysate was prepared. Immunoprecipitation of the receptor complex was carried out using 12CA5. Western analysis was done using antibody against GRK5 (top panel), and the same blot was reprobed with 12CA5 against HA-DOR after stripping (middle panel). Direct Western analysis of the cell lysates using GRK5 antibody for GRK5 expression is shown (bottom panel).

FIG. 4.
Agonist-stimulated formation of receptor-GRK2-G␤␥ ternary complex on the membrane. HEK293 cells expressing opioid receptor with or without GRK2ct, ␤2␥2, or transducin ␣ were incubated in the presence or absence of 1 M DPDPE (DP) at 37°C for 3 min. The cells were lysed, immunoprecipitation of the receptor complex was carried out using 12CA5, and Western analysis was done using antibodies against GRK2, G␤, and 12CA5 sequentially after stripping. Direct Western analysis of the cell lysates using antibodies against GRK2, G␤, or transducin ␣ is shown. membrane translocation (24). Our observation of a receptor-GRK-G␤␥ complex in response to receptor activation is consistent with these structure-function studies on GRKs and suggests that the complex is formed through interaction of the amino-terminal domain of GRK2 with the activated receptor and of the carboxyl-terminal domain of GRK2 with the G␤␥ subunits on the membrane.
Previous work has shown that small synthetic peptides are extremely poor substrates of GRKs and that GPCR-derived peptides specifically inhibit phosphorylation of the receptor as opposed to the peptide substrate (25). These studies suggested that GRKs interact with activated GPCRs at sites distinct from their phosphorylation sites (25). Our current data show that the 31 residues in the carboxyl terminus of DOR contain sites for both GRK2 phosphorylation and interaction and that the DORs with phosphorylation sites mutated interact with GRK2 and form a stable complex with GRK2 and the G␤␥ subunits on the membrane effectively. These data demonstrate in vivo that the GRK2 phospho-receptor residues in DOR are not necessary for receptor-GRK2 interaction, although these residues are in very close proximity to residues essential for GRK2 interaction.
GRK phosphorylation sites are identified in only a few GPCR substrates, and no clear consensus substrate sequence has been found among them (25). Studies with synthetic peptide substrates and purified GRKs suggest that kinases in the GRK2 family preferentially phosphorylate peptides containing acidic residues flanking the target serines or threonines (26). Studies with the M2 muscarinic receptor and ␣1 adrenergic receptor have demonstrated that acidic amino acid residues are important in the agonist-dependent phosphorylation and desensitization of these receptors (27)(28)(29). In this study, we have found that substitution of glutamate and aspartate residues adjacent to GRK phosphorylation sties in DOR blocks agoniststimulated DOR phosphorylation and attenuated receptor desensitization and have further revealed that these two acidic residues are essential for GRK2, GRK3, and GRK5 binding. Our results have demonstrated, in a receptor context, that the negatively charged acidic residues flanking phosphorylation sites are critically involved in the interaction of the receptor with GRKs of both the ␤AR kinase and GRK4 families in vivo and thus are required for GRK-mediated receptor phosphorylation and desensitization.