Palmitoylated cysteine 341 modulates phosphorylation of the beta2-adrenergic receptor by the cAMP-dependent protein kinase.

We previously showed that substitution of a glycine residue for the palmitoylated cysteine 341 of the human beta2-adrenergic receptor (Gly341beta2AR), increases the basal level of the receptor phosphorylation and reduces its ability to functionally interact with Gs. In the present study, we show that additional mutation of serines 345 and 346 (Ala345,346Gly341beta2AR) restored normal phosphorylation and receptor-Gs coupling, thus suggesting that the increased phosphorylation of this site, rather than the lack of palmitoylation per se, is responsible for the poor coupling of the unpalmitoylated receptor. This is supported by the observation that chemical depalmitoylation of purified beta2AR did not affect the ability of the receptor to stimulate adenylyl cyclase in reconstitution assays. Furthermore, mutation of Ser345,346 in a wild type receptor background (Ala345,346beta2AR) significantly decreased the rate of agonist-promoted desensitization of the receptor-stimulated adenylyl cyclase activity, supporting a role for this phosphorylation site in regulating the functional coupling of the receptor. Since serines 345 and 346 are located in a putative cyclic AMP-dependent protein kinase (PKA) phosphorylation site immediately downstream of the palmitoylated cysteine 341, the hypothesis that the accessibility of this site may be regulated by the receptor palmitoylation state was further assessed in vitro. In membrane phosphorylation assays, Gly341beta2AR was found to be a better substrate for PKA than the wild type receptor, thus supporting the notion that palmitoylation restrains access of the phosphorylation site to the enzyme. Taken together, the data demonstrate that palmitoylation of cysteine 341 controls the phosphorylation state of the PKA site located in the carboxyl tail of the beta2AR and by doing so modulates the responsiveness of the receptor.

Post-translational modification of transmembrane receptors has been shown to play important roles in the proper regulation of hormonal signaling. For the human ␤ 2 -adrenergic re-ceptor (␤ 2 AR), 1 a large body of evidence implicates phosphorylation in the rapid desensitization of the ␤-adrenergic receptorstimulated adenylyl cyclase activity that follows sustained stimulation (1)(2)(3). In particular, phosphorylation of the receptor by both cAMP-dependent protein kinase (PKA) and the ␤-adrenergic receptor kinase (␤ARK) has been shown to functionally uncouple the receptor from G s (4).
Another post-translational modification of the ␤ 2 AR, the palmitoylation of its cysteine 341, has also been shown to influence the ability of the receptor to functionally interact with G s and to stimulate adenylyl cyclase (5). Indeed, substitution of a glycine residue for the cysteine 341 (Gly 341 ␤ 2 AR) was found to prevent receptor palmitoylation and greatly reduced its ability to stimulate adenylyl cyclase activity. This mutation also led to a loss of the guanine nucleotide-sensitive high affinity binding state of the receptor for agonists, thus suggesting that the unpalmitoylated receptor was largely uncoupled from G s . More recently, we showed (6) that the decreased responsiveness of Gly 341 ␤ 2 AR was accompanied by a constitutive elevation of the basal level of this receptor phosphorylation. These data suggested that concerted interactions between palmitoylation and phosphorylation could play an important role in the regulation of the ␤ 2 AR function. Such a hypothesis is further supported by the observation that, as for phosphorylation, the palmitoylation state of the ␤ 2 AR is dynamically regulated by agonist stimulation (7). 2 One of the two potential PKA phosphorylation sites present in the ␤ 2 AR is located one amino acid downstream of the palmitoylated cysteine ( 341 C palm LRRSS). This observation and the fact that the unpalmitoylated receptor is already phosphorylated and uncoupled from G s raises the possibility that the phosphorylation of this site may be influenced by the palmitoylation state of the receptor and may contribute to its desensitization. The present study was designed to test this hypothesis.
Construction of Mutated ␤ 2 AR cDNAs and Cell Transfection-A human ␤ 2 AR cDNA clone encoding a glycine substitution of cysteine 341 (Gly 341 ␤ 2 AR) was constructed as in O'Dowd et al. (5) and subcloned into the eukaryotic expression vector pBC12BI (9). To generate the other ␤ 2 AR mutants, the cDNAs encoding wild type ␤ 2 AR (10), Gly 341 ␤ 2 AR (5), and Ala 261,262,345,346 ␤ 2 AR (11) were subcloned into pSP65 (Promega) by introduction in the NcoI-SalI sites. Ala 345,346 ␤ 2 AR was constructed by digestion of pSP65 plasmids containing wild type ␤ 2 AR and Ala 261,262,345,346 ␤ 2 AR cDNAs respectively, with AviII. The appropriate restriction fragments were isolated and ligated, and a resulting pSP65 plasmid containing a cDNA for Ala 345,346 ␤ 2 AR was obtained. Gly 341 Ala 345,346 ␤ 2 AR was constructed by digestion of pSP65 plasmids containing Gly 341 ␤ 2 AR and Ala 261,262,345,346 ␤ 2 AR respectively, with AviII. The appropriate restriction fragments were isolated and ligated and a resulting pSP65 plasmid containing a Gly 341 Ala 345,346 ␤ 2 AR cDNA was obtained. All of the constructs were subcloned in the expression vector pBC12BI␤ 2 AR (10) by introduction in the NcoI-SalI sites. The identities of all mutants were confirmed by direct dideoxynucleotide sequencing. The wild type ␤ 2 AR and all of the mutants were stably expressed in cultured fibroblasts (mouse LTK Ϫ cells or Chinese hamster CHW cells) by co-transfection with the neomycin resistance plasmid pSV2-Neo (Pharmacia) using the calcium phosphate precipitation procedure (12). Neomycin-resistant cells were selected by culturing in DMEM supplemented with 10% fetal bovine serum containing G418 (450 g/ml for LTK Ϫ cells or 150 g/ml for CHW cells). Clones were then screened for ␤ 2 AR expression by radioligand assays using [ 125 I]CYP as the ligand (10). Cell lines expressing a similar number of receptors were selected for the study.
Cell Culture and Membrane Preparation-LTK Ϫ and CHW cells were grown as a monolayer in 75-cm 2 Corning or Nunc plastic flasks containing DMEM supplemented with 10% fetal bovine serum, 1 mM Lglutamine, 500 units/ml penicillin, 500 units/ml streptomycin, and 0.25 g/ml amphotericin B in an atmosphere of 95% air and 5% CO 2 at 37°C. Attached cells were washed 3 times with ice-cold PBS, mechanically detached into 10 ml of buffer A (5 mM Tris-HCl (pH 7.4), 2 mM EDTA, 10 g/ml benzamidine, 5 g/ml soybean trypsin inhibitor, and 5 g/ml leupeptin) on ice. Cell suspensions were homogenized with a Polytron homogenizer (Ultra-Turrax T-25; Janke and Kunkel) using a single burst of 7 s at maximum setting. Homogenates were centrifuged at 43,000 ϫ g for 20 min at 4°C, and the pellets were rinsed twice in buffer A. Membranes were finally resuspended in buffer B (75 mM Tris-HCl (pH 7.4)), 5 mM MgCl 2 , 2 mM EDTA, 10 g/ml benzamidine, 5 g/ml soybean trypsin inhibitor, and 5 g/ml leupeptin) at a concentration of ϳ1 mg/ml and used immediately for binding or adenylyl cyclase assays. Protein concentrations were determined by the method of Bradford (Bio-Rad), using bovine serum albumin as a standard.
For agonist-induced desensitization experiments, cell medium was changed for fresh media containing isoproterenol and ascorbic acid to a final concentration of 1 M and 100 M, respectively, for the indicated time. Control cells were treated with media containing ascorbic acid only.
Radioligand Binding Assays-Radioligand binding assays were conducted essentially as described in Bouvier et al. (10) using 10 l of membrane suspension (ϳ10 g of protein) in a total volume of 0.5 ml of buffer B. For saturation experiments, concentrations of [ 125 I]CYP ranging from 5 to 250 pM were used. Competition binding assays with (Ϫ)-isoproterenol were conducted using ϳ50 pM [ 125 I]CYP as the radioligand. The concentration of the ␤-adrenergic agonist was varied from 0 to 100 M in the presence or absence of 100 M Gpp(NH)p. Binding data were analyzed by nonlinear least squares regression using the computer program Ligand (13). For routine determination of receptor number, triplicate assay tubes contained a nearly saturating concentration of [ 125 I]CYP (200 pM) in the presence or absence of 10 M (Ϫ)-alprenolol to define specific binding.
Adenylyl Cyclase Assays-Adenylyl cyclase activity was determined in crude membrane preparations according to the method of Salomon et al. (14). Membranes (ϳ4 g of protein) were incubated with a reaction mixture containing 45 mM Tris-HCl (pH 7.4), 3 mM MgCl 2 , 1. isolated by sequential chromatography on a Dowex cation exchange resin and aluminum oxide. Data calculated as pmol of cAMP produced/ min/mg were analyzed using nonlinear least squares regression.
Whole Cell Phosphorylation and Purification of the ␤ 2 ARs-Phosphorylation experiments were performed essentially as described (10). Briefly, 2 h before the labeling, cells were cultured in phosphate-free DMEM. Following this period, cells were incubated at 37°C with ϳ8 mCi of [ 32 P]H 3 PO 4 in DMEM containing 50 M NaH 2 PO 4 for 1 h. The cells were then rinsed with ice-cold PBS, mechanically detached, and resuspended in 10 ml of buffer C (20 mM Tris-HCl (pH 7.4), 5 mM EDTA, 10 mM Na 4 P 2 O 7 ⅐10 H 2 O, 10 g/ml benzamidine, 5 g/ml soybean trypsin inhibitor, and 5 g/ml leupeptin). The cells were disrupted by sonication at 4°C, and membranes were isolated by centrifugation at 43,000 ϫ g for 20 min at 4°C. The pellets were rinsed twice, and the resulting membranes were solubilized in 100 mM NaCl, 10 mM Tris-HCl (pH 7.4), 5 mM EDTA and 2% digitonin for 2 h at 4°C. Solubilized receptors were recovered in the supernatant of a 43,000 ϫ g centrifugation for 20 min at 4°C. The receptors were then purified by alprenolol-Sepharose affinity chromatography as described previously (15) and, purified receptors were concentrated by membrane filtration over Centriprep and Centricon concentrators (Amicon). The amount of purified receptors was assessed by soluble [ 125 I]CYP binding assays. Aliquots were then prepared for SDS-PAGE, and electrophoresis was conducted using 10% slab gels (16). Following electrophoresis, proteins were electrophoretically transferred to nitrocellulose in order to reduce the free phosphate background, and the membranes were exposed to Kodak XAR-5 films at Ϫ70°C for several days. Laser densitometric scanning of the autoradiographs (UltroScan, Pharmacia Biotech Inc.) was used to quantitate the incorporation of 32 P into the receptor. In some experiments, the specific activity of the cellular [ 32 P]ATP pool was determined by high pressure liquid chromatography using a reverse phase nucleotide/nucleoside 7U (Alltech Inc.) column. The stoichiometry of phosphorylation could then be calculated by liquid scintigraphy of the excised receptor band for a known quantity of ␤ 2 AR loaded on SDS-PAGE.
Phosphorylation of ␤ 2 AR in Membrane Preparations Derived from Sf9 Cells-Cells were cultured at 27°C in Grace supplemented insect medium up to a density of 1.5-2.0 ϫ 10 6 cells/ml. The cells were then infected with recombinant baculoviruses encoding wild type ␤ 2 AR, Gly 341 ␤ 2 AR, or Gly 341 Ala 345,346 ␤ 2 AR at a multiplicity of infection of 2-5. The virus encoding the wild type ␤ 2 AR was generated as described before (7), while the two other viruses were constructed by subcloning the mutant receptor coding regions into the pJV␤ 2 AR recombinant vector. Recombinant baculovirus was purified by successive plaque assays using the ␤-galactosidase assay. Cells were harvested 48 h after the infection, and membranes were prepared as described above. The final membrane pellets were resuspended in a phosphorylation buffer containing 50 mM Tris-HCl (pH 7.4), 5 mM MgCl 2 , 5 mM phosphoenolpyruvate, 0.2 unit/50 l pyruvate kinase, 0.2 mM ATP, and 0.025 mM [␥-32 P]ATP (2000 cpm/pmol). Phosphorylation assay was initiated by adding 3 ϫ 10 Ϫ5 units of protein kinase A catalytic subunit (Sigma). Maximum phosphorylation was attained following a 30-min incubation at 30°C. Thus, this incubation time was used for routine assays. At the end of the incubation, membranes were centrifuged at 43,000 ϫ g for 20 min at 4°C, and the pellets were rinsed 3 times in a buffer containing 50 mM Tris-HCl (pH 7.4), 2 mM EDTA, and 20 mM NaH 2 PO 4 . Receptors were then solubilized in 100 mM NaCl, 10 mM Tris-HCl (pH 7.4), 5 mM EDTA, and 0.3% dodecylmaltoside for 90 min at 4°C. Affinity purification was then carried out as described above. The level of phosphorylation was assessed following SDS-PAGE of the purified receptor. For this purpose, the bands corresponding to the ␤ 2 AR were excised from the gel and counted by liquid scintigraphy. The stoichiometry of phosphorylation could be calculated based on the specific activity of the [␥-32 P]ATP used in the phosphorylation assays. In some experiments, membranes were treated with agarose-bound alkaline phosphatase (600 units/ml) for 45 min at 37°C in 20 mM Tris-HCl (pH 8.0), 1 mM MgCl 2 , and 0.1 mM ZnCl 2 prior to the phosphorylation assay. Alkaline phosphatase was removed from the samples by centrifugation at 500 ϫ g for 5 min at 4°C by extensive washing and repeated centrifugation.
Hydroxylamine Treatment and Reconstitution of the ␤ 2 AR-Human recombinant ␤ 2 AR was purified from Sf9 cells as described above. The concentrated purified receptor was incubated for 1 h at room temperature in a 10 mM Tris-HCl solution (pH 7.0) containing or not containing hydroxylamine at a final concentration of 1 M. Samples were then washed extensively and desalted by filtration dialysis using Centricon 30 (Amicon). For receptor reconstitution, membranes derived from untransfected LTK Ϫ cells were prepared as described above and resuspended at a concentration of 2-3 g/l in a buffer containing 20 mM Hepes (pH 7.4), 2 mM MgCl 2 , 1 mM EDTA, 0.4% CHAPS. Purified ␤ 2 ARs treated as above were then incubated with this membrane preparation (8 fmol/l) under mild agitation at 15°C. After 1 h, an equal volume of a freshly prepared buffer containing 100 mM Hepes (pH 7.4), 10 mM MgCl 2 , 2 mM ATP, 0.2 mM GTP was added to the mixture and incubated for an additional 20 min at 30°C. The preparation was then centrifuged at 180,000 ϫ g for 40 min. The pelleted membranes were then washed twice in a buffer containing 75 mM Tris-HCl (pH 7.4), 5 mM MgCl 2 , 2 mM EDTA. The reconstituted membranes were finally diluted to 2-5 g/20 l in the same buffer and assayed for [ 125 I]CYP binding and isoproterenol-stimulated adenylyl cyclase activity as described above.
Whole Cell Palmitoylation of the ␤ 2 AR-[ 3 H]palmitate labeling experiments were performed essentially as described in Ref. 7. Briefly, Sf9 cells (2 ϫ 10 6 /ml) grown in suspension were infected with wild type ␤ 2 AR recombinant baculovirus at a multiplicity of infection of 2-5. Twenty-four hours after this treatment, [9,10-3 H]palmitic acid dissolved in dimethyl sulfoxide was added to the culture to a final concentration of 0.2 mCi/ml. After 1.5 h the labeling period was terminated by centrifugation of the cells at 500 ϫ g for 5 min at 4°C followed by two washes with ice-cold PBS. Cell membranes were then prepared, and the receptor was solubilized and purified by Sepharose-alprenolol affinity chromatography. The concentrated receptor preparation was then treated or not treated with hydroxylamine as above.

RESULTS AND DISCUSSION
Mutant forms of the human ␤ 2 AR and Gly 341 ␤ 2 AR in which serine 345 and 346 were replaced by alanine residues were constructed (see Fig. 1) and stably expressed in mouse LTK Ϫ cells. Membranes from cellular clones, selected by virtue of a co-transfected neomycin resistance marker, were prepared and assayed for [ 125 I]CYP binding activity. As seen in Table I, no significant difference in the affinity for the radioiodinated antagonist was observed among the receptors tested. The [ 125 I]CYP bound to the wild type and mutant ␤ 2 ARs with K d values in good agreement with those reported for human ␤ 2 AR expressed in other cell systems (7,17). No specific binding of [ 125 I]CYP to membranes from untransfected LTK Ϫ cells was detected (data not shown). Since apparent coupling properties can vary with receptor expression levels (10,18), cellular clones expressing similar numbers of receptors for wild type and mutant forms of human ␤ 2 AR were selected for further characterization.
To assess the functional properties of the various receptor mutants, their ability to confer an isoproterenol-sensitive adenylyl cyclase activity to LTK Ϫ cells was tested. In membranes derived from cells transfected with wild type ␤ 2 AR cDNA, isoproterenol stimulated adenylyl cyclase activity in a dose-dependent manner (Fig. 1). The EC 50 of isoproterenol was 67 Ϯ 10 nM (n ϭ 9), consistent with its potency in other cellular systems (5,7). No isoproterenol-stimulated adenylyl cyclase activity could be detected in untransfected LTK Ϫ cells (data not shown). As previously reported in CHW cells (5,6), the ability of Gly 341 ␤ 2 AR to stimulate the adenylyl cyclase was found to be markedly reduced relative to wild type ␤ 2 AR in LTK Ϫ cells. This is illustrated by the blunted isoproterenol dose-response curve observed in membranes derived from Gly 341 ␤ 2 AR expressing cells (Fig. 1). The marked difference between ␤ 2 ARand Gly 341 ␤ 2 AR-stimulated adenylyl cyclase activity did not result from the small (statistically nonsignificant) difference in receptor number expressed in the two cell lines studied (Table  I), since in several experiments the number of Gly 341 ␤ 2 ARs was higher than that of wild type ␤ 2 AR; yet, the level of stimulation  obtained for Gly 341 ␤ 2 AR never reached that of the wild type receptor.
Additional replacement of serines 345 and 346 by alanine residues into Gly 341 ␤ 2 AR almost completely restored the ability of the receptor to stimulate adenylyl cyclase activity. This is illustrated by the normal isoproterenol dose-response curve observed in membranes derived from Gly 341 Ala 345,346 ␤ 2 AR-expressing cells (Fig. 1). In fact, the maximal isoproterenol-stimulated adenylyl cyclase activities were virtually identical for wild type ␤ 2 AR and Gly 341 Ala 345,346 ␤ 2 AR (Table I). This contrasts with the much reduced isoproterenol-stimulated activity observed for Gly 341 ␤ 2 AR. These results suggest that the absence of the palmitoylated cysteine 341 decreases the ability of the receptor to stimulate the adenylyl cyclase only when the PKA phosphorylation site located downstream of the palmitoylation site is intact. Mutation of the two serines alone in the wild type receptor modestly increased the ability of the receptor to stimulate the adenylyl cyclase activity ( Fig. 1 and Table I).
Agonist binding properties of the receptors were then assessed by competition of [ 125 I]CYP binding with isoproterenol. As shown in Fig. 2 and Table II, competition of [ 125 I]CYP binding by isoproterenol was biphasic for the wild type ␤ 2 AR, and the curve could be best resolved by a two-affinity state model using iterative nonlinear least squares fitting of the nontransformed data (19). The high affinity state was guanyl nucleotide-sensitive. Indeed, when the binding was conducted in the presence of 100 M Gpp(NH)p, the curve was best fitted to a single low affinity state model. The existence of guanyl nucleotide-sensitive high affinity sites for agonist is generally believed to reflect functional coupling of the receptor to G proteins. In membranes derived from L cells expressing Gly 341 ␤ 2 AR, the displacement curves were best fitted to a single low affinity site whether the binding was conducted in the presence or the absence of Gpp(NH)p, thus suggesting that Gly 341 ␤ 2 AR is largely uncoupled from G s . This is consistent with the modest adenylyl cyclase stimulation mediated by this mutant receptor. The additional mutation of serines 345 and 346 completely restored the ability of the unpalmitoylated receptor to bind agonist with high affinity. Indeed, for Gly 341 Ala 345,346 ␤ 2 AR the competition of [ 125 I]CYP binding by isoproterenol was best resolved by a two-affinity state model (Fig. 2, Table II). As was the case for the wild type receptor, the high affinity state was sensitive to Gpp(NH)p, and a single population of low affinity sites was found when the binding was carried in the presence of the GTP analog. Mutation of the two serines alone in the wild type ␤ 2 AR did not affect the binding parameters of isoproterenol. K i values and the proportion of sites in high and low affinity states are summarized in Table II. These data suggest that the absence of serines 345 and 346 restores the ability of the unpalmitoylated receptor to interact with G s , consistent with the effects of this mutation on the capacity of the ␤ 2 AR to stimulate the adenylyl cyclase activity.
Since serines 345 and 346 represent a potential phosphorylation site for PKA, it could be hypothesized that phosphorylation of this site contributes to the uncoupled phenotype of Gly 341 ␤ 2 AR. To test this hypothesis, the basal level of phosphorylation of the wild type ␤ 2 AR, Gly 341 ␤ 2 AR and Gly 341 Ala 345,346 ␤ 2 AR was assessed in LTK Ϫ cells. Following metabolic labeling with inorganic [ 32 P]phosphoric acid ( 32 P i ), the cells were broken, membranes were prepared, and the receptors were purified by alprenolol-Sepharose affinity chromatography. An identical number of receptors, as determined by [ 125 I]CYP binding, was prepared for SDS-PAGE. Purified receptors migrated with an apparent molecular mass of ϳ70 kDa as previously reported (20). Fig. 3 illustrates the level of phosphorylation of the three receptors. The labeling intensity of purified Gly 341 ␤ 2 AR was found to be much higher than that of the wild type receptor, thus confirming that, as found in CHW cells (6), Gly 341 ␤ 2 AR has a significantly elevated basal level of phosphorylation. In contrast, the level of phosphorylation of Gly 341 Ala 345,346 ␤ 2 AR was not different from that of the wild type receptor. Thus, the mutation of serines 345 and 346 completely prevented the elevation of the basal phosphorylation state induced by the mutation of cysteine 341. Taken together, the data support the hypothesis that this PKA phosphorylation site contributes to the increased phosphorylation and uncoupling of the unpalmitoylated Gly 341 ␤ 2 AR. The low number of receptors expressed in LTK Ϫ cells and the very small basal level of phosphate incorporation into the wild type receptor makes it impossible to accurately determine the stoichiometry of phosphorylation in this system. In order to quantitate the increase in the basal phosphorylation level observed in Gly 341 ␤ 2 AR, metabolic labeling experiments with 32 P i were conducted in CHW cells expressing higher levels of wild type ␤ 2 AR and Gly 341 ␤ 2 AR (ϳ2.0 pmol/mg of protein in each case). Assessment of the specific activity of the cellular [ 32 P]ATP pool allowed us to determine the basal stoichiometry of phosphorylation. A stochiometry of 0.34 mol of phosphate/mol of receptor was found for the wild type ␤ 2 AR, while 1.1 mol of phosphate/ mol of receptor were incorporated into Gly 341 ␤ 2 AR.
The above results strongly suggest that mutation of glycine 341 does not affect receptor coupling by itself but rather as a consequence of the increased phosphorylation level resulting from the mutation. To further test the hypothesis that the absence of the palmitate moiety does not directly affect the coupling of the receptor, the effect of chemical removal of the palmitate was assessed. Recombinant human ␤ 2 AR expressed in Sf9 cells was purified, treated or not treated with hydroxylamine, and reconstituted in membranes derived from native LTK Ϫ cells devoid of endogenous ␤ 2 AR. The efficacy of the hydroxylamine treatment to remove the receptor-bound palmitate was tested using [ 3 H]palmitate-labeled ␤ 2 AR. As shown in the inset to Fig. 4, the treatment led to a complete depalmitoylation of the receptor. However, such depalmitoylation was without effect on the ability of the receptor to stimulate adenylyl cyclase activity. Indeed, reconstitution of an equal number of native or hydroxylamine-treated ␤ 2 AR conferred identical isoproterenol-sensitive adenylyl cyclase activity to the LTK Ϫ derived membranes (Fig. 4). No isoproterenol-stimulated adenylyl cyclase activity could be detected using nonreconstituted LTK Ϫ membranes (data not shown). These results confirm that the presence of receptor-bound palmitate is not required per se to allow productive interaction between the ␤ 2 AR and G s .
It has been proposed that post-translational modification of the ␤ 2 AR by palmitoylation of cysteine 341 promotes the association of the amino-terminal portion of the carboxyl tail with the plasma membrane, thus forming a fourth intracellular loop (5). Such a model was originally proposed for rhodopsin (21), and the insertion of the rhodopsin-bound palmitate moieties into the membrane lipid bilayer was experimentally confirmed using fluorescent analogues of the fatty acid (22). The results presented above clearly demonstrate that the formation of this fourth loop is not required for proper coupling of the receptor to G s . Indeed, Gly 341 Ala 345,346 ␤ 2 AR, which lacks the palmitoylated cysteine and the chemically depalmitoylated receptor, are perfectly able to stimulate the G s /adenylyl cyclase system. Nevertheless, the formation of such a loop undoubtedly modifies the topology of the carboxyl tail, a region that is known to act as substrate for regulatory kinases. In particular, serines 345 and 346 would be located very near the plasma membrane, where they might not be easily accessible to PKA. This model is consistent with the observation that only the PKA site located in the third cytoplasmic loop of the ␤ 2 AR ( 259 RRSS) and not the one in the carboxyl tail ( 341 CLRRSS) has been shown to be phosphorylated upon direct activation of PKA by dibutyryl-cAMP (23)) or following incubation of broken cell preparations with the PKA catalytic subunit (24). Following this model, mutation of Cys 341 may provide an easier access of the carboxyl tail site to PKA by preventing the formation of the putative fourth intracellular loop.
To directly test this hypothesis, the accessibility of the phosphorylation site to PKA was assessed in vitro using Sf9 cell membranes expressing wild type ␤ 2 AR, Gly 341 ␤ 2 AR, and Gly 341 Ala 345,346 ␤ 2 AR. Membranes were incubated with [␥-32 P]ATP in the presence of PKA catalytic subunit. As shown in Fig. 5, PKA catalyzed the phosphorylation of wild type ␤ 2 AR, Gly 341 ␤ 2 AR, and Gly 341 Ala 345,346 ␤ 2 AR. The phosphorylated receptors migrated with an apparent molecular mass of 45-55 kDa, consistent with the reported electrophoretic mobility of the human ␤ 2 AR expressed in Sf9 cells (7). The difference with the mobility observed in LTK Ϫ cells is due to different glycosylation patterns. As previously reported for the unpalmitoylated form of p21 N-ras (8), the unpalmitoylated Gly 341 ␤ 2 AR and Gly 341 Ala 345,346 ␤ 2 AR have a slightly lower electrophoretic mobility then the wild type receptor. When comparing the maximum level of phosphorylation, Fig. 5 clearly shows that Gly 341 ␤ 2 AR acted as a better substrate for PKA than the wild type receptor. Additional mutation of serines 345 and 346 reduced the phosphorylation to levels even lower than those reached for the wild type ␤ 2 AR, thus suggesting that phosphorylation of this site contributed to the elevated phosphorylation observed in membranes expressing Gly 341 ␤ 2 AR.
The above results are consistent with the idea that the lack of palmitoylation favors phosphorylation of the receptor by PKA, most likely by increasing the accessibility of the site 343 RRSS to the kinase. However, the stoichiometry of phosphorylation reached under the conditions studied were relatively modest (0.11 mol/mol for wild type versus 0.17 mol/mol for Gly 341 ␤ 2 AR). We reasoned that the preexisting phosphorylation state of the receptors may interfere with the in vitro assay. To make sure that this did not bias the interpretation of the data, the receptors were dephosphorylated prior to the in vitro phosphorylation assay. For that purpose, membranes were treated with alkaline phosphatase for 45 min. This treatment was found to remove more than 80% of the phosphate incorporated into the receptor following whole cell metabolic labeling with inorganic [ 32 P]phosphoric acid (data not shown). As can be seen in Fig. 6, the phosphatase treatment dramatically increased the amount of phosphate that could be transferred by PKA for the two receptors, but the higher phosphorylation observed for Gly 341 ␤ 2 AR was maintained. The stoichiometry of phosphorylation reached following pretreatment with the phosphatase was 0.31 mol/mol for wild type ␤ 2 AR and 0.70 mol/mol for Gly 341 ␤ 2 AR. The amount of [ 32 P]PO 4 incorporated in vitro represented the maximal level of phosphate transfer that could be attained for the two receptors under the condition studied, since no further phosphorylation was observed following longer incubation with the kinase (data not shown). It follows that more phosphate molecules can be transferred by PKA into Gly 341 ␤ 2 AR than into the wild type ␤ 2 AR, consistent with the idea that the unpalmitoylated receptor is a better substrate for PKA.
To determine if phosphorylation of this PKA phosphorylation site contributes to the normal development of agonist-promoted desensitization, the pattern of rapid desensitization of Ala 345,346 ␤ 2 AR was compared with that of wild type ␤ 2 AR. As shown in Fig. 7, mutation of Ser 345,346 significantly delayed the appearance of agonist-promoted desensitization. In cells expressing wild type receptor, pretreatment with isoproterenol for 1 and 2 min caused a significant decrease in the ability of the agonist to stimulate the adenylyl cyclase in membranes derived from these cells. Reduction in both the maximal stimulation (16 and 37% at 1 and 2 min, respectively) and in the efficacy of isoproterenol (EC 50 of control, 1.6 ϫ 10 Ϫ7 M; 1 min, 8.8 ϫ 10 Ϫ7 ; 2 min, 1.3 ϫ 10 Ϫ6 ) were observed. This desensitization was receptor-specific, since no significant change in either basal or forskolin-stimulated activity were observed (data not shown). In Ala 345,346 ␤ 2 AR-expressing cells, pretreatment for 1 and 2 min with isoproterenol only led to very marginal changes in the maximal adenylyl cyclase stimulation and isoproterenol efficacy. Following a longer desensitizing period (30 min), the desensitization observed for Ala 345,346 ␤ 2 AR and wild type receptor were similar (data not shown), suggesting that Ser 345 and Ser 346 are not essential for the desensitization to occur but that they contribute to its rapid onset.
The data presented in this study demonstrate that palmitoylation of Cys 341 regulates the accessibility of a PKA phosphorylation site in the carboxyl tail of the ␤ 2 AR. Our data also indicate that such a regulatory influence on the phosphorylation state of the receptor may play an important role in the rapid modulation of the receptor responsiveness.
It has been proposed that palmitoylation of GAP-43 modulates its ability to interact with G o (25). In that study, the authors proposed that the sulfhydryl functions of cysteines 3 and 4 from GAP-43 are essential to activate the G protein and that palmitoylation of the cysteines inhibits their stimulatory activity. Our study shows that this is not the case for the ␤ 2 AR. Indeed, Gly 341 Ala 345,346 ␤ 2 AR, which is lacking cysteine 341, interacts normally with G s and stimulates the adenylyl cyclase activity to levels comparable with that of wild type receptor.
For many proteins involved in signal transduction, palmitoylation serves to regulate membrane attachment. Good examples are G␣ subunits whose subcellular distribution has been shown to be modified following mutation of the palmitoylated cysteine (26,27). Interestingly, stimulation of the ␤ 2 AR has been shown to increase the turnover rate of G␣ s -bound palmitate, which may lead to a reduced palmitoylation of the protein (28 -30). This regulation of the palmitoylation state of G␣ s has been proposed to promote translocation of G␣ s from the plasma membrane to the cytosol and thus could act as a process modulating signaling efficacy (28). A similar agonist-promoted regulation of the ␤ 2 AR palmitoylation state has recently been proposed (7). 2 Based on the observations reported here, it could therefore be proposed that agonist stimulation, by promoting the depalmitoylation of the receptor, regulates membrane attachment of the proximal portion of the its carboxyl tail, providing better access for PKA to the phosphorylation site neighboring cysteine 341. The favored phosphorylation of this site may then contribute to the rapid onset of desensitization. Interestingly, the palmitoylation state of another receptor, the ionotropic glutamate receptor GluR6, has also been shown to regulate its phosphorylation by a second messenger-dependent protein kinase (32).
The contribution of the carboxyl tail PKA phosphorylation site to agonist-promoted desensitization, has previously been questioned. Indeed, studies show that this site could not be phosphorylated upon direct activation of PKA (23). Also, deletion of this PKA consensus site did not affect desensitization evoked by the PKA catalytic subunit or low doses (50 nM) of adrenaline, two conditions known to promote heterologous desensitization (24). However, the present study clearly suggest that this site becomes phosphorylated and contributes to the rapid onset of homologous desensitization upon treatment with micromolar concentrations of agonist. The easiest way to reconcile these data is to propose that serine 345 and 346 become available for phosphorylation only when receptors are agonistbound. Therefore, stimulation with low doses of agonist (which leads to occupancy of a small proportion of the receptor population) or with cAMP analogues would not promote phosphorylation of this site. These observations are consistent with a recent report by Post et al. (31), which showed in S49 cells that PKA-mediated phosphorylation is important for rapid homologous desensitization.
The data presented clearly show that phosphorylation of the PKA site closest to the palmitoylated cysteine is affected by the palmitoylation state of the receptor. However, further studies are required to determine whether or not phosphorylation of more distal sites believed to be phosphorylated by ␤ARK could also be affected. In olfactory cilia preparations, inhibition of PKA prevented the phosphorylation and desensitization of odorant receptors believed to be mediated both by PKA and a ␤ARK like enzyme (33). The authors suggested a sequential interplay between PKA and the receptor-specific kinase. Similar sequential phosphorylation of specific sites involved in desensitization have also been reported for the rhodopsin (34,35), the N-formyl peptide receptor (36), and the C5a anaphylatoxin receptor (37).
Cysteine residues located in a position similar to that of cysteine 341 in the ␤ 2 AR are found in most of the G proteincoupled receptors, and the occurrence of palmitoylation has been demonstrated for a number of these. They include rhodopsin (38), the ␣ 2A -adrenergic receptor (39), the dopamine D 1 receptor (40), the serotonin 5HT 1B (41) and 5HT 1A receptors (42), the luteinizing hormone receptor (43), and the metabotropic mGluR4 glutamate receptor (44). Although the precise site of palmitoylation has been identified only for a few receptors (5,39,44,45), a number of studies have assessed the functional effects of mutating cysteines that were either known palmitoylation sites or assumed as potential sites. Similarly to what has been observed for the ␤ 2 AR, mutation of cysteine 347 of the D 1 -dopamine receptor decreases its ability to interact with G s and to stimulate the adenylyl cyclase activity in COS cells (46). Furthermore, this mutant receptor becomes resistant to agonist-promoted desensitization, suggesting that it is constitutively desensitized. It is noteworthy that a potential PKA phosphorylation site is located four amino acids upstream of cysteine 347. In contrast, mutation of cysteine 442 in the ␣ 2A AR, which has been shown to abolish its palmitoylation, is without effect on the receptor-G protein coupling (39). Similarly, mutations of conserved cysteines in the lutropin/choriogonadotropin (43), thyrotropin-releasing hormone (47), and M 2 -muscarinic receptors (48) were found not to affect coupling to their cognate G proteins. However, these mutations markedly affected the rate of agonist-promoted internalization. These observations suggest that the functional consequences of the lack of palmitoylation may vary among different G protein-FIG. 7. Time-dependent desensitization of wild type ␤ 2 AR and Ala 345,346 ␤ 2 AR. LTK Ϫ cells stably expressing wild type ␤ 2 AR (upper panel) or Ala 345,346 ␤ 2 AR (lower panel) were incubated for 0 (circle), 1 (square), or 2 (triangle) minutes with 1 M isoproterenol. Membranes were then prepared, and adenylyl cyclase activity was measured as described under "Experimental Procedures." Data represent the mean Ϯ S.E. of three independent experiments. coupled receptors. The nature of the molecular motif(s) present in the carboxyl tail of a given receptor might determine these consequences. This is to be expected if the absence of the palmitoylated cysteines exposes the surface of the receptor carboxyl terminus and influences its interaction with various regulatory proteins.