Molecular Basis for Subtype-specific Desensitization of Inhibitory Adenosine Receptors ANALYSIS OF A CHIMERIC A 1 -A 3 ADENOSINE RECEPTOR*

The differing effects of short-term agonist exposure on the two inhibitory adenosine receptor (AR) subtypes have been examined using Chinese hamster ovary cells stably expressing the hemagglutinin epitope-tagged human A 1 AR and rat A 3 AR. Under conditions in which exposure of transfected cells to 5 (cid:109) M ( (cid:50) )-( R )- N 6 -(phenyli-sopropyl)adenosine resulted in the functional desensiti- zation and phosphorylation of the A 3 AR, neither prop-erty was exhibited by the A 1 AR. However, a stably expressed chimeric A 1 -A 3 AR, termed A 1 CT3AR, in which the C-terminal domain of the A 1 AR distal to its predicted palmitoylation site was replaced by the corresponding region of the A 3 AR, was able to undergo functional desensitization and agonist-stimulated phosphorylation in a manner similar to that exhibited by the A 3 AR. More-over, purified G-protein-coupled receptor kinases 2, 3, and 5 were each capable of enhancing the agonist-de-pendent phosphorylation of the A 3 AR and A 1 CT3AR in vitro . Taken together, these data demonstrate that the C-terminal domain of the A 3 AR distal to its predicted palmitoylation site is responsible for this receptor’s ability to undergo a rapid agonist-dependent desensitization and are consistent with a model in which phosphorylation of the A 3 AR within this domain by one or more G- protein-coupled

The differing effects of short-term agonist exposure on the two inhibitory adenosine receptor (AR) subtypes have been examined using Chinese hamster ovary cells stably expressing the hemagglutinin epitope-tagged human A 1 AR and rat A 3 AR. Under conditions in which exposure of transfected cells to 5 M (؊)-(R)-N 6 -(phenylisopropyl)adenosine resulted in the functional desensitization and phosphorylation of the A 3 AR, neither property was exhibited by the A 1 AR. However, a stably expressed chimeric A 1 -A 3 AR, termed A 1 CT3AR, in which the C-terminal domain of the A 1 AR distal to its predicted palmitoylation site was replaced by the corresponding region of the A 3 AR, was able to undergo functional desensitization and agonist-stimulated phosphorylation in a manner similar to that exhibited by the A 3 AR. Moreover, purified G-protein-coupled receptor kinases 2, 3, and 5 were each capable of enhancing the agonist-dependent phosphorylation of the A 3 AR and A 1 CT3AR in vitro. Taken together, these data demonstrate that the C-terminal domain of the A 3 AR distal to its predicted palmitoylation site is responsible for this receptor's ability to undergo a rapid agonist-dependent desensitization and are consistent with a model in which phosphorylation of the A 3 AR within this domain by one or more Gprotein-coupled receptor kinases initiates the desensitization process.
The phenomenon whereby a biological response plateaus and then diminishes despite the continual presence of agonist is termed desensitization. At least three different processes may be involved in the desensitization of G-protein-coupled receptor function. The earliest detectable event, occurring within seconds of agonist exposure, is that of receptor phosphorylation by either second messenger-activated kinases or G-protein-coupled receptor kinases (GRKs). 1 The latter represent a growing family of proteins that specifically phosphorylate agonist-occupied receptors (1). In the ␤ 2 -adrenergic receptor, m 2 -muscarinic receptor, and rhodopsin systems, phosphorylation increases receptor affinity for arrestin molecules, which uncouple the receptors from their associated G-proteins (2,3). Following receptor phosphorylation, many receptors undergo sequestration into a poorly defined intracellular compartment (4). While the molecular determinants controlling sequestration are being defined for an increasing number of receptors, the function of this process remains unclear since, in many cases, inhibition of sequestration does not appear to impair desensitization (5,6). Prolonged agonist exposure (several hours) can lead to the down-regulation of the receptor protein and/or its associated G-proteins (7,8).
The many physiological effects of adenosine are mediated in part by its binding to multiple cell-surface adenosine receptor (AR) subtypes termed A 1 , A 2a , A 2b , and A 3 (9). Given that both the A 1 AR and A 3 AR bind the same physiological ligand and activate G i proteins, the significance of the existence of multiple inhibitory ARs is unclear. Recent work on the regulation of multiple ␣ 2 -adrenergic (10,11) and ␤-adrenergic (12,13) receptors has demonstrated that subtype-specific adaptive desensitization mechanisms have evolved such that responses mediated by particular receptor subtypes may be turned off at distinct rates. Indeed, studies examining the agonist-induced regulation of the A 1 AR and A 3 AR have demonstrated that these receptors desensitize over quite distinct time courses (14 -18). Specifically, A 1 AR desensitization typically occurs over periods of several hours, or even days, and is temporally associated with receptor down-regulation either alone (14) or concomitantly with G i proteins (15,16). In contrast, desensitization of the rat A 3 AR expressed in the RBL-2H3 cell line undergoes a homologous functional desensitization that is apparent after only a few minutes of agonist exposure (17,18). We have recently demonstrated that the rapid functional desensitization of a recombinant rat A 3 AR expressed in CHO cells is associated with the agonist-dependent phosphorylation of the receptor by one or more GRKs (19).
To more accurately characterize subtype-specific differences in the short-term regulation of inhibitory ARs, we have characterized functional desensitization in CHO cell lines stably expressing hemagglutinin epitope-tagged human A 1 AR (20) and rat A 3 AR (21) cDNAs. In addition, we have exploited these differences to generate a chimeric A 1 -A 3 AR, which has revealed that the structural determinants conferring rapid A 3 AR desensitization kinetics and sensitivity to GRK-mediated phosphorylation reside within a small region at its C terminus. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 1 The abbreviations used are: GRKs, G-protein-coupled receptor kinases; AR, adenosine receptor; CHO, Chinese hamster ovary; IB-MECA, N 6 -(3-iodobenzyl)adenosine 5Ј-N-methyluronamide; PAGE, polyacrylamide gel electrophoresis; PBS, phosphate-buffered saline; LC, long chain alkyl spacer group; PVDF, polyvinylidene difluoride; DPCPX, 1,3-dipropyl-8-cyclopentylxanthine; (R)-PIA, (Ϫ)-(R)-N 6 -(phe-nylisopropyl)adenosine; XAC, xanthine amine congener.

EXPERIMENTAL PROCEDURES
Materials-IB-MECA (22) was the generous gift of Dr. Kenneth Jacobson (National Institutes of Health, Bethesda, MD). Cell culture supplies were from Life Technologies, Inc. Radiochemicals were from DuPont NEN. Sources of other materials have been described elsewhere (19,23).
Receptor cDNA Constructs and Expression-The CHO cell line stably expressing the hemagglutinin epitope-tagged rat A 3 AR has been previously characterized (19). The six-amino acid hemagglutinin epitope DVPDYA, recognized by monoclonal antibody 12CA5 (24), was inserted at both the amino and carboxyl termini of the human A 1 AR by insertion of polymerase chain reaction products generated using the previously described pCMV5/human A 1 AR cDNA as a template (20). Correct introduction of the epitope sequences was verified by dideoxynucleotide sequencing. The chimeric A 1 -A 3 AR, termed A 1 CT3AR, was generated by a previously described three-step polymerase chain reaction protocol using the epitope-tagged A 1 AR and A 3 AR cDNAs in pCMV5 as templates (23). The integrity of the chimeric insertion was confirmed by dideoxynucleotide sequencing.
CHO cell lines stably expressing the epitope-tagged wild-type and chimeric ARs were generated by cotransfecting cells with the appropriate cDNA subcloned into pCMV5 and pSV2Neo in a 20:1 ratio using a modified calcium phosphate precipitation/glycerol shock procedure previously described (25). After selection in G418, resistant colonies were isolated, expanded, and screened for receptor expression by radioligand binding. Cells were propagated in T-75 flasks with Ham's F-12 medium supplemented with 10% (v/v) fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin in a 37°C humidified atmosphere containing 5% CO 2 .
Intact Cell Receptor Phosphorylation-Transfected CHO cells were plated into 6-well dishes at a density of ϳ1 ϫ 10 6 cells/well and cultured overnight in regular medium. The next day, the cells were washed twice with phosphate-free Dulbecco's modified Eagle's medium and incubated for 90 min at 37°C in the same medium supplemented with 1 unit/ml adenosine deaminase and 200 Ci/ml [ 32 P]orthophosphate. After stimulation with the indicated agonists, reactions were terminated by placing the cells on ice and washing the monolayers twice with 3 ml of ice-cold phosphate-buffered saline. Cells were scraped into 1 ml of lysis buffer (50 mM Hepes, pH 7.5, 5 mM EDTA, 10 mM sodium fluoride, 10 mM sodium phosphate, 0.1 mM phenylmethylsulfonyl fluoride, and 10 g/ml each soybean trypsin inhibitor, leupeptin, and pepstatin A), transferred to microcentrifuge tubes, and lysed by vigorous vortexing. Membranes were pelleted by centrifugation (14,000 ϫ g, 10 min) and solubilized by resuspension in lysis buffer supplemented with 150 l of 1% (w/v) SDS, followed by brief sonication and heating to 65°C for 15 min. After chilling to 4°C, 600 l of a concentrated nonionic detergent mixture was added to the solubilized membranes such that the final mixture contained 1% (v/v) Triton X-100, 0.5% (w/v) sodium deoxycholate, 0.2% (w/v) SDS, and 150 mM sodium chloride in lysis buffer. Insoluble material was removed by centrifugation (14,000 ϫ g, 10 min), and the supernatant was precleared by incubation for 1 h at 4°C with protein A-Sepharose in the presence of 0.2% (w/v) IgG-free bovine serum albumin. The precleared supernatant was then incubated for 2 h at 4°C with protein A-Sepharose and 7 l of 12CA5 ascites. Immune complexes were isolated by centrifugation, washed twice with detergent buffer containing 0.2 M ammonium sulfate and once with detergent buffer alone, and eluted from the protein A-Sepharose by the addition of electrophoresis sample buffer and incubation at room temperature for 45 min. Analysis was by SDS-PAGE, using 10% (w/v) polyacrylamide resolving gels, and by autoradiography. Quantitation of phosphorylation experiments was by either densitometry or excision from the dried gel and Cerenkov counting of bands of interest.
Cell-surface Labeling with Biotin-LC-Hydrazide-This was performed on cells in 6-well dishes basically as described by Lisanti et al. (26). Briefly, cell monolayers were washed with ice-cold PBS and then treated with 10 mM sodium periodate in PBS for 30 min at 4°C in the dark. After removal of the periodate solution and further washing with PBS, cells were washed once with 0.1 M sodium acetate, pH 5.5, and then incubated for 30 min at 4°C with 2 mM biotin-LC-hydrazide in the same buffer; this procedure labels all cell-surface carbohydrate residues with biotin. Cells were then washed prior to membrane preparation and immunoprecipitation with 12CA5 as described above. After SDS-PAGE, resolved proteins were transferred to a PVDF membrane, and nonspecific protein-binding sites were blocked by a 60-min incubation in blocking buffer (5% (w/v) skimmed milk solution in PBS containing 0.2% (v/v) Triton X-100 and 0.02% (w/v) thimerosal). The membrane was then incubated for 60 min at room temperature with 1 g/ml horseradish peroxidase-conjugated streptavidin in a high detergent/skimmed milk solution. After three washes in blocking buffer and two washes in PBS, reactive proteins were visualized by an enhanced chemiluminescence protocol in accordance with the manufacturer's instructions (Renaissance, DuPont NEN).
GRK Purification-Recombinant bovine GRK2, bovine GRK3, and human GRK5 were purified from Sf9 cells by previously published procedures 48 h after infection with the appropriate baculovirus construct (27,28). Using light-activated rhodopsin as a substrate, the specific activity of each of the purified enzymes was ϳ1 mol/min/mg of protein.
In Vitro Assay of Receptor Kinase Activity-Monolayers of transfected CHO cells were washed three times with ice-cold PBS and scraped into lysis buffer (10 mM Hepes, pH 7.5, 2 mM EDTA, 0.25 M NaCl, and 0.1 mM phenylmethylsulfonyl fluoride containing 10 g/ml each soybean trypsin inhibitor and leupeptin). After Dounce homogenization on ice (20 strokes), membranes were pelleted by centrifugation and resuspended in GRK assay buffer (25 mM Hepes, pH 7.5, 2.5 mM EDTA, and 7.5 mM MgCl 2 ) supplemented with 1 unit/ml adenosine deaminase and protease inhibitors for immediate use. Assays consisted of 40 l of membrane suspension, 40 l of kinase mixture (GRK assay buffer supplemented with 0.25 mM ATP, 0.88 mM dithiothreitol, 0.15 M okadaic acid, and 10 Ci of [␥-32 P]ATP), 10 l of vehicle or purified GRK, and 10 l of vehicle or agonist. After incubation at 30°C for 5 min, reactions were terminated by placing the tubes on ice and adding 0.5 ml of ice-cold stop solution (0.1 M sodium phosphate, pH 7.5, and 10 mM EDTA). Membranes were pelleted by centrifugation (14,000 ϫ g, 10 min), and the resulting pellets were solubilized in 1% (w/v) SDS prior to dilution in nonionic detergent buffer as described above. After centrifugation to remove insoluble material, detergent extracts were equalized by protein assay prior to immunoprecipitation with 12CA5 and analysis by SDS-PAGE and autoradiography as described above.
Phosphoamino Acid Analysis-Following SDS-PAGE, proteins were transferred to a PVDF membrane. After overnight autoradiography, the region of the membrane corresponding to the phosphorylated receptor was excised, hydrated, and hydrolyzed at 110°C in 200 l of 5.7 M HCl for 90 min. The resulting hydrolysate was lyophilized and resuspended in chromatography buffer supplemented with phosphoamino acid standards. After spotting onto cellulose-coated plates, samples were subjected to ascending chromatography in an isobutyric acid, 0.5 M ammonium hydroxide (5:3, v/v) buffer system (29). Standards were visualized by ninhydrin staining, and 32 P-labeled amino acids were visualized by autoradiography.
Radioligand Binding and Adenylyl Cyclase Assays-Saturation binding experiments employing [ 3 H]DPCPX were performed and analyzed as we have described previously (23). Adenylyl cyclase assays were performed exactly as described previously using IB-MECA or (R)-PIA (30). Dose-response curves were analyzed by a previously validated curve-fitting program (31).

Functional Receptor Expression-
We have previously demonstrated that the rapid functional desensitization of the rat A 3 AR is associated with the agonist-dependent phosphorylation of the receptor in situ by a kinase of similar substrate specificity to GRK2 (19). It was also noted that the C-terminal domain of the A 3 AR distal to the predicted palmitoylation site contains multiple serine and threonine residues, several of which are surrounded by acidic amino acids (Fig. 1), a context that has been shown to enhance GRK2-mediated phosphorylation in vitro (32,33). In contrast, the A 1 AR contains no potential phosphorylation sites in the same region (Fig. 1). There-FIG. 1. Structure of A 1 AR, A 3 AR, and A 1 CT3AR. The primary sequences of the C-terminal domains proximal and distal to the predicted palmitoylation sites of the human A 1 AR, rat A 3 AR, and chimeric A 1 CT3AR are shown. Each receptor also expressed hemagglutinin epitope tags at its amino and carboxyl termini. fore, the chimeric A 1 -A 3 AR shown in Fig. 1 was constructed to determine whether this domain of the A 3 AR was sufficient to confer on the resulting predominantly A 1 -containing chimera the rapid regulatory properties of the A 3 AR.
Agonist-mediated regulation of the epitope-tagged A 1 AR, A 3 AR, and A 1 CT3AR was then studied after stable expression of each cDNA in CHO cells. We have previously demonstrated that the epitope-tagged rat A 3 AR functions comparably to the untagged rat A 3 AR in this system (19 (34) and the purified human brain A 1 AR (35).
In addition, both the A 1 AR and A 1 CT3AR were functional as determined by their ability to inhibit adenylyl cyclase activity in isolated membranes. In membranes from A 1 AR-expressing cells, (R)-PIA maximally inhibited 5 M forskolin-stimulated adenylyl cyclase activity by 52 Ϯ 6%, with an IC 50 value of 18.8 Ϯ 8.5 nM (three experiments). Under the same assay conditions, (R)-PIA maximally inhibited forskolin-stimulated adenylyl cyclase activity in membranes from A 1 CT3AR-expressing cells by 46 Ϯ 6%, with an IC 50 value of 5.5 Ϯ 2.2 nM (three experiments). Therefore, the A 1 AR and A 1 CT3AR bind antagonist radioligand with similar affinity and can sustain comparable agonist-dependent inhibition of adenylyl cyclase.
Effects of (R)-PIA Pretreatment on Receptor Function-To study the capacity of the expressed inhibitory ARs to undergo a rapid functional desensitization, transfected CHO cells were exposed to vehicle or (R)-PIA prior to membrane preparation and assayed for agonist-mediated inhibition of forskolin-stimulated adenylyl cyclase activity ( Fig. 2 and Table I). Consistent with our previous study (19), pretreatment of A 3 AR-expressing CHO cells with 5 M (R)-PIA for 10 min resulted in a functional desensitization, as manifested by a 7-fold increase in the IC 50 value for the agonist IB-MECA to inhibit forskolin-stimulated adenylyl cyclase activity, with no significant change in the maximal inhibition achieved ( Fig. 2A and Table I). In contrast, similar agonist pretreatment had no effect on the ability of the A 1 AR to inhibit adenylyl cyclase activity under the same assay conditions (Fig. 2B and Table I). Moreover, increasing the agonist exposure time to 30 min failed to unmask any detectable A 1 AR desensitization (data not shown). However, exposure of A 1 CT3AR-expressing cells to (R)-PIA for 10 min resulted in a functional desensitization that was qualitatively similar to that observed for the A 3 AR, i.e. agonist pretreatment resulted in a 5-fold increase in the IC 50 value for (R)-PIAmediated inhibition without significantly changing the maximal inhibition observed ( Fig. 2C and Table I). Therefore, while the A 1 AR and A 1 CT3AR are similar with respect to their abilities to inhibit adenylyl cyclase activity, A 1 CT3AR signaling undergoes an agonist-induced functional desensitization that is similar to that exhibited by the A 3 AR.
Agonist-stimulated Receptor Phosphorylation in Situ-Radioligand binding studies using [ 3 H]DPCPX demonstrated that levels of the A 1 AR and A 1 CT3AR were essentially unaffected by prior agonist treatment (A 1 AR levels were reduced by 1 Ϯ 13%, and A 1 CT3AR levels were decreased by 6 Ϯ 11%; three experiments). No antagonist radioligand with high affinity for the  Table I. A 3 AR is currently available. However, comparative immunoblotting of membranes with 12CA5 revealed that no significant change in the level of A 3 ARs was induced by treatment with 5 M (R)-PIA for 10 min (data not shown). Therefore, since the expression level of each of these receptor proteins in transfected cell membranes is unaffected by prior agonist exposure, a rapid loss of receptors from cell membranes cannot account for the agonist-induced functional desensitization observed for the A 3 AR and A 1 CT3AR.
To study receptor regulation in response to agonist exposure, we utilized the ability of monoclonal antibody 12CA5 to specifically recognize the hemagglutinin epitope tag sequences engineered into each receptor's coding sequence. To assess the capacity of 12CA5 to immunoprecipitate each of the expressed ARs, transfected cell monolayers were surface-labeled with biotin hydrazide and solubilized for receptor immunoprecipitation with 12CA5. After SDS-PAGE, resolved proteins were transferred to a PVDF membrane for probing with horseradish peroxidase-conjugated streptavidin as we have previously described (19). Fig. 3A demonstrates that 12CA5 could specifi-

FIG. 3. Effects of agonist pretreatment on inhibitory AR phosphorylation.
A, CHO cells stably expressing the indicated ARs were treated with or without 5 M (R)-PIA for 10 min at 37°C as indicated prior to cell-surface biotin labeling, membrane preparation, and solubilization for immunoprecipitation with 12CA5. After SDS-PAGE, immunoprecipitated proteins were transferred to a PVDF membrane, probed with horseradish peroxidase-conjugated streptavidin, and visualized by enhanced chemiluminescence as described under "Experimental Procedures." B, 32 P-labeled CHO cells stably expressing the indicated ARs were treated with or without 5 M (R)-PIA for 10 min at 37°C as indicated prior to membrane preparation, solubilization, and immunoprecipitation with 12CA5. Immunoprecipitates were analyzed by SDS-PAGE and autoradiography. C, CHO cells stably expressing chimeric A 1 CT3ARs were treated with or without 5 M (R)-PIA for 10 min at 37°C as indicated prior to immunoprecipitation with 12CA5 as described above. Following SDS-PAGE, resolved proteins were transferred to a PVDF membrane for phosphoamino acid analysis as described under "Experimental Procedures." The migrations of ninhydrin-stained standards in this thin-layer chromatography system are indicated. D, 32 P-labeled A 1 CT3AR-expressing CHO cells were treated with 1 M XAC or the appropriate vehicle as indicated prior to immunoprecipitation with 12CA5 and analysis by SDS-PAGE.

TABLE I
Effects of short-term agonist exposure on inhibitory AR function CHO cells stably expressing the indicated inhibitory ARs were incubated at 37°C for 10 min with 1 unit/ml adenosine deaminase in the absence ("Control") or presence ("Treated") of 5 M R-PIA. Incubations were terminated by placing the monolayers on ice and washing the cells extensively in ice-cold buffer prior to membrane preparation. Assay of adenylyl cyclase activity was performed in the presence of 5 M forskolin and increasing concentrations of agonist as described under "Experimental Procedures." Data are presented as means Ϯ S.E. for three experiments. cally immunoprecipitate cell surface-labeled receptor proteins of the appropriate size from each of the transfected cell lines. Both the A 1 AR and A 1 CT3AR migrate as ϳ40-kDa glycoproteins, consistent with the approximate size of the purified human brain A 1 AR as identified by radioiodination (35). We have previously demonstrated that the epitope-tagged A 3 AR can be specifically immunoprecipitated from transfected CHO cells and that it migrates as a broad 50 -70-kDa glycoprotein (Fig. 3A) (19). For the A 1 AR and A 3 AR, a 10-min agonist pretreatment appeared to produce a small decrease in the levels of cell-surface receptor compared with vehicle-treated cells (over multiple experiments, the maximal reductions observed were 20% for the A 1 AR and 24% for the A 3 AR). No decrease was consistently observed for the A 1 CT3AR at this time point. Therefore, since total receptor levels were unaffected by agonist treatment, it seems that agonist pretreatment for 10 min can induce a small but detectable internalization of the A 1 AR and A 3 AR, but not the chimeric A 1 CT3AR.
To determine whether any of the inhibitory ARs were phosphorylated in response to agonist treatment, receptors were immunoprecipitated from 32 P-labeled transfected cells following incubation with or without 5 M (R)-PIA for 10 min. Consistent with our previous report (19), the A 3 AR became phosphorylated in an agonist-dependent manner (Fig. 3B). In contrast, no phosphorylation of the A 1 AR could be detected after treatment with or without agonist, despite our ability to immunoprecipitate this receptor under these conditions (Fig. 3,  A and B). However, the A 1 CT3AR displayed a significant level of basal phosphorylation, and this was increased by some 3.3 Ϯ 1.0-fold (three experiments) after incubation with agonist. Phosphoamino acid analysis revealed that in the basal state, the A 1 CT3AR was phosphorylated exclusively on threonine residues. Agonist treatment resulted in the appearance of phosphoserine and increased the levels of phosphothreonine observed (Fig. 3C). The predominance of phosphothreonine over phosphoserine is also characteristic of the phosphoamino acid content of the A 3 AR after agonist treatment (19). To determine whether the basal phosphorylation of the A 1 CT3AR was related to a low level agonist-independent signaling capacity observed for the A 1 AR, 2 the effect of XAC, an A 1 AR "inverse agonist," on receptor function was determined. XAC appeared to have some inverse agonist activity at the A 1 CT3AR since preincubation of membranes from A 1 CT3AR-expressing CHO cells with 1 M XAC for 15 min resulted in a 15-25% increase (range over four experiments) in the stimulation of adenylyl cyclase activity produced by 5 M forskolin, suggesting that XAC is relieving a tonic inhibitory constraint on adenylyl cyclase activation. Preincubation with 1 M XAC also reduced the basal phosphorylation of the A 1 CT3AR by 38 Ϯ 10% (three experiments) (Fig. 3D). Therefore, the basal phosphorylation state of the A 1 CT3AR appears to reflect, at least in part, the existence of a population of receptors that can assume an active conformation in the absence of agonist that can be partially reversed upon the addition of an inverse agonist. Taken together, these data suggest that the ability of the A 3 AR and A 1 CT3AR to undergo agonist-induced functional desensitization is associated with their increased phosphorylation on serine and threonine residues after agonist exposure, while the inability of the A 1 AR to undergo a functional desensitization over the same time course is associated with undetectable levels of receptor phosphorylation under the same conditions. Sensitivity of Inhibitory ARs to GRK2-mediated Phosphorylation in Vitro-We previously demonstrated that the A 3 AR is a substrate for GRK2 in in vitro phosphorylation reactions (19).
In addition, it was shown that prior treatment of intact cells with agonist reduces the level of GRK2-stimulated agonist-dependent phosphorylation subsequently observed in vitro, suggesting that a GRK2-like kinase is responsible for the phosphorylation observed in the intact cell. To determine whether the distinct effects of agonist on A 1 AR and A 1 CT3AR phosphorylation in situ were reflected in their sensitivity to phosphorylation by GRK2, in vitro phosphorylation experiments were performed using membranes from transfected CHO cells and purified GRK2. Under these conditions, no agonist-dependent phosphorylation of the A 1 AR by either endogenous membraneassociated receptor kinase activity or purified GRK2 was detectable (Fig. 4). However, under the same assay conditions, the A 1 CT3AR was phosphorylated in an agonist-dependent manner by an endogenous receptor kinase activity, and the level of agonist-dependent phosphorylation was increased by some 2.2 Ϯ 0.6-fold (three experiments) in the presence of GRK2 (Fig. 4). Therefore, the distinct effects of agonist treatment on the phosphorylation of the A 1 AR and A 1 CT3AR in situ are associated with distinct sensitivities to phosphorylation by GRK2 in vitro.
Sensitivity of the A 3 AR and A 1 CT3AR to Phosphorylation by Multiple GRKs-Despite the fact that GRK2 is the most intensively studied GRK isoform with regard to its activity against a range of receptor substrates, it has become clear in the last few years that it is just one member in an expanding family of GRKs cloned to date (1). Six members of this family, termed GRK1-6, have been cloned from mammalian sources, but with the exception of GRK1 (rhodopsin kinase), the spectrum of substrates for each GRK in vivo remains unknown. One reason for this is that very few receptors have been tested for their ability to act as substrates for these kinases in vitro, and therefore, the structural determinants conferring receptor sensitivity to them are unclear.
Therefore, to determine whether the A 3 AR and A 1 CT3AR could be phosphorylated by multiple GRKs, in vitro phosphorylation experiments were performed using membranes from transfected CHO cells and each of purified GRK3 and GRK5, with GRK2 being used as a positive control (Fig. 5). Using membranes expressing the A 3 AR, it was found that GRK3 could stimulate agonist-dependent A 3 AR phosphorylation to a similar extent as GRK2, with which it exhibits an 85% identity at the amino acid level (Fig. 5, A and C) (36). It was also found that GRK5 could stimulate agonist-dependent A 3 AR phospho-2 M. E. Olah and G. L. Stiles, unpublished data.

FIG. 4. Selective agonist-dependent phosphorylation of the A 1 CT3AR by GRK2 in vitro.
In vitro phosphorylations were performed as described under "Experimental Procedures" using membranes from CHO cells stably expressing either the A 1 AR or A 1 CT3AR incubated with or without 5 M (R)-PIA in the presence or absence of 50 nM GRK2 at 30°C for 5 min. Following the addition of stop solution, membranes were pelleted for solubilization and receptor immunoprecipitation with 12CA5. Analysis was by SDS-PAGE and autoradiography. This is one of multiple experiments. rylation in vitro, despite exhibiting only a 35% amino acid identity to GRK2 (28). However, the extent to which GRK5 stimulated A 3 AR phosphorylation was slightly less than that observed for GRK2 and GRK3 (Fig. 5, A and C). Membranes expressing the A 1 CT3AR displayed similar results, with each of GRK2, GRK3, and GRK5 being capable of stimulating agonistdependent receptor phosphorylation in vitro, although the levels of phosphorylation produced by GRK2 and GRK3 were slightly higher than that displayed by GRK5 (Fig. 5, B and C). Under similar assay conditions, no agonist-dependent phosphorylation of the A 1 AR could be produced by GRK2, GRK3, or GRK5 (data not shown). Therefore, it appears that the structural determinants conferring sensitivity to agonist-dependent phosphorylation by multiple GRKs in vitro reside in the Cterminal 14 amino acids of the A 3 AR. DISCUSSION Many studies have shown that the A 1 AR expressed natively in a variety of systems desensitizes over a time course of several hours or even days (14 -16). For example, the rat adipocyte A 1 AR down-regulates and desensitizes after treatment of cultured cells with (R)-PIA for 4 days (15), while the hamster A 1 AR in the DDT 1 MF-2 cell line undergoes functional desensitization with a t1 ⁄2 of ϳ6 h (14). In contrast, the native A 3 AR expressed in the rat mast cell-derived RBL-2H3 line undergoes a rapid functional desensitization detectable within a few minutes of agonist exposure (17,18). By expressing the epitopetagged human A 1 AR and rat A 3 AR in the same cell type under control of the same expression vector, we have shown that the differences observed in the short-term regulation of native receptor function are related to receptor structure and are not simply a reflection of the use of different cell lines to study these receptors. Specifically, we have demonstrated that under conditions in which the A 3 AR undergoes agonist-dependent phosphorylation and desensitization, the A 1 AR exhibits neither of these properties. Therefore, since it appeared that the A 1 AR lacked the molecular determinants necessary to desensitize in response to short-term agonist treatment, we reasoned that it would be a useful "acceptor" molecule into which candidate regulatory cytoplasmic domains from the A 3 AR could be introduced, resulting in the formation of chimeric A 1 -A 3 ARs. The extreme C-terminal 14-amino acid segment of the A 3 AR was introduced into the A 1 AR since we had previously demonstrated that in transfected CHO cells, the A 3 AR undergoes a rapid desensitization that is associated with receptor phosphorylation by a GRK-related kinase (19). The A 3 AR contains multiple serine and threonine residues in the region of the C-terminal tail distal to the predicted palmitoylation site, and in several instances, these are flanked by acidic amino acids, which peptide phosphorylation studies have shown to be an important determinant for phosphorylation by GRK2 (32,33). Moreover, the A 1 AR does not possess any potential phosphorylation sites within the same region, perhaps suggesting that this may be a site responsible for the subtype-specific adaptive responses to short-term agonist exposure.
Expression and functional analysis of the A 1 CT3AR chimera (Fig. 1) demonstrated that its behavior in response to shortterm agonist exposure was qualitatively similar to that of the A 3 AR. Specifically, A 1 CT3AR function underwent a rapid functional desensitization in response to short-term agonist treatment, and this was associated with the increased phosphorylation of the chimera on serine and threonine residues. Neither of these properties was displayed by the A 1 AR under the same conditions. Taken together with our previous observation that a GRK-related kinase is responsible for agonist-dependent A 3 AR phosphorylation in situ (19), the current data strongly suggest that the agonist-stimulated phosphorylation of the C-terminal domain of the A 3 AR is responsible for initiating the events that lead to the rapid desensitization of receptor function.
One difference between the A 3 AR and A 1 CT3AR was the detection of a significant level of threonine phosphorylation of FIG. 5. Agonist-dependent phosphorylation of the A 3 AR and A 1 CT3AR by multiple GRKs in vitro. In vitro phosphorylations were performed as described under "Experimental Procedures" using membranes from CHO cells stably expressing either the A 3 AR (A) or A 1 CT3AR (B) incubated with or without 5 or 10 M 5Ј-N-ethylcarboxyamidoadenosine (NECA) in the presence or absence of a 50 nM concentration of the indicated GRK isoform at 30°C for 5 min. Following the addition of stop solution, membranes were pelleted for solubilization and receptor immunoprecipitation with 12CA5. Analysis was by SDS-PAGE and autoradiography. C is a quantitative analysis of three such experiments performed for each receptor. For each receptor, phosphorylation is expressed relative to that observed in the presence of agonist, but in the absence of any added GRK (set at 100%). the latter receptor in the absence of agonist (Fig. 3). It is unlikely that this is due to the presence of high amounts of adenosine in the medium since a high concentration of adenosine deaminase was included during cell incubation. Also, any effects of endogenous adenosine would be expected to similarly induce basal phosphorylation of the A 3 AR, which was not observed (Fig. 3). Instead, this phenomenon may relate to the observation that within a given population of A 1 ARs expressed, a certain percentage may be active in the absence of agonist. In a chimeric A 1 CT3AR, such low level activation would result in a basal phosphorylation since it is the activated form of the receptor that is recognized and phosphorylated by GRKs (1); the ability of the inverse agonist XAC to enhance adenylyl cyclase stimulation in isolated membranes and to reduce the basal phosphorylation of the A 1 CT3AR lends support to this hypothesis (Fig. 3D). However, a definitive answer to this question will require the identification of a compound that acts as a "neutral antagonist" at the A 1 AR. Regardless of this issue, the basal phosphorylation does not impair A 1 CT3AR function compared with the A 1 AR as determined by comparison of adenylyl cyclase inhibition in membranes from untreated cells (Fig. 2 (A  and B) and Table I). Also, (R)-PIA competition for [ 3 H]DPCPX binding in transfected cell membranes demonstrated that the fraction of receptors displaying high affinity for agonist was similar for both the A 1 AR and A 1 CT3AR, indicating that the basal phosphorylation state of the chimera does not noticeably alter the receptor-G-protein equilibrium as compared with the wild-type A 1 AR. 3 Therefore, conditions producing the additional agonist-stimulated phosphorylation on serine and threonine residues are necessary to produce the observed functional desensitization of the A 1 CT3AR.
The observation that multiple GRK isoforms could sustain the agonist-dependent phosphorylation of the C-terminal domain of the A 3 AR has important implications for the regulation of A 3 AR function in different cell types. Specifically, it would be expected that agonist-dependent phosphorylation and desensitization would be observed in many cell types natively expressing the receptor, no matter what complement of GRK isoforms may be present. In this respect, the A 3 AR resembles the ␤ 1 -and ␤ 2 -adrenergic (37,38), m 2 -muscarinic (39), and ␦-opioid (40) receptors, each of which has been shown to display an increased agonist-stimulated phosphorylation in the presence of GRK2 and GRK5 despite the low amino acid identity between these kinases (35%) (28). More important, we have established that phosphorylation by each of GRK2, GRK3, and GRK5 occurs within the same 14-amino acid domain at the C terminus of the receptor. It remains to be determined whether each kinase phosphorylates the same amino acids. However, this would seem unlikely based on the different structural requirements necessary for optimal peptide substrate phosphorylation by each kinase in vitro (32,33,39).
In conclusion, we have identified a small region of the A 3 AR C-terminal tail that can confer the abilities to undergo rapid agonist-dependent phosphorylation and desensitization on a predominantly A 1 -containing chimera, A 1 CT3AR. Like the A 3 AR, the agonist-occupied A 1 CT3AR is also a substrate in vitro for GRK2, GRK3, and GRK5. Taken together, these observations strongly support a model for rapid A 3 AR desensitization in which agonist-stimulated phosphorylation of the Cterminal domain distal to the predicted palmitoylation site by one or more GRKs initiates the events that lead to rapid recep-tor uncoupling and functional desensitization. Future analysis of cell lines stably expressing epitope-tagged wild-type and chimeric inhibitory ARs will allow a detailed examination of the structural basis for any other subtype-specific adaptive mechanisms these receptors may exhibit, such as those invoked in response to prolonged agonist treatment.