J Biol Chem, Vol. 274, Issue 43, 31014-31019, October 22, 1999

Activation of the ₂-Adrenergic Receptor-G_s Complex Leads to Rapid Depalmitoylation and Inhibition of Repalmitoylation of Both the Receptor and G_s^*

Thomas P. Loisel^§¶, Hervé Ansanay^§, Lynda Adam^§, Stefano Marullo, Roland Seifert^**, Monique Lagacé^§, and Michel Bouvier^§§§

From the ^§ Département de Biochimie and Groupe de Recherche sur le Système Nerveux Autonome, Université de Montréal, Montréal, Québec H3C 3J7, Canada, the  Département de Pharmacologie Cellulaire et Moléculaire, CNRS-URA 1534, Institut Cochin de Génétique Moléculaire, Paris, France, and the ^** Howard Hughes Medical Institute, Stanford, California 94305

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Palmitoylation is unique among lipid modifications in that it is reversible. In recent years, dynamic palmitoylation of G protein  subunits and of their cognate receptors has attracted considerable attention. However, very little is known concerning the acylation/deacylation cycle of the proteins in relation to their activity status. In particular, the relative contribution of the activation and desensitization of the signaling unit to the regulation of the receptors and G proteins palmitoylation state is unknown. To address this issue, we took advantage of the fact that a fusion protein composed of the stimulatory  subunit of trimeric G protein (G_s) covalently attached to the ₂-adrenergic receptor (₂AR) as a carboxyl-terminal extension (₂AR-G_s) can be stimulated by agonists but does not undergo rapid inactivation, desensitization, or internalization. When expressed in Sf9 cells, both the receptor and the G_s moieties of the fusion protein were found to be palmitoylated via thioester linkage. Stimulation with the -adrenergic agonist isoproterenol led to a rapid depalmitoylation of both the ₂AR and G_s and inhibited repalmitoylation. The extent of depalmitoylation induced by a series of agonists was correlated (0.99) with their intrinsic efficacy to stimulate the adenylyl cyclase activity. However, forskolin-stimulated cAMP production did not affect the palmitoylation state of ₂AR-G_s, indicating that the agonist-promoted depalmitoylation is linked to conformational changes and not to second messenger generation. Given that, upon activation, the fusion protein mimics the activated receptor-G protein complex but cannot undergo desensitization, the data demonstrate that early steps in the activation process lead to the depalmitoylation of both receptor and G protein and that repalmitoylation requires later events that cannot be accommodated by the activated fusion protein.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Palmitoylation is a post-translational modification that is limited to a small subset of cellular proteins among which proteins involved in signal transduction are prevalent (1). This thioesterification of cysteine residues by palmitate distinguishes itself from other lipid modifications such as prenylation and myristoylation by its reversibility. Indeed, in contrast to myristoyl and prenyl moieties that are added co-translationally and generally remain attached to the proteins until the protein gets degraded, the protein-bound palmitate is added post-translationally and turns over more rapidly than the protein itself (2-4). Moreover, the palmitoylation state of several proteins has been shown to be dynamically regulated. In particular, biological regulation of the palmitoylation state of heterotrimeric G proteins and of their cognate receptors has been demonstrated (5-12).

Activation of G_s¹ through receptor stimulation, following direct activation with aluminum fluoride and cholera toxin or as a result of site-directed mutagenesis, has been shown to lead to an increased incorporation of [³H]palmitate into G_s during pulse labeling experiments. Because pulse-chase labeling experiments clearly indicated that stimulation increased the depalmitoylation rate, the enhanced incorporation was attributed to an accelerated turnover rate of the G_s-bound palmitate (10-12). Interestingly, Jones et al. (8) found that, despite the increased turnover rate, activation of G_s did not significantly affect its stoichiometry of palmitoylation, thus challenging the notion that stimulation ultimately favors the depalmitoylation reaction (12). Agonist stimulation of the ₂-adrenergic receptor (₂AR) has also been shown to increase the amount of covalently attached [³H]palmitate (5) as a result of an increased turnover rate of the receptor-bound palmitate (7). A similar agonist-promoted increase in the turnover rate of receptor-bound palmitate was observed for the _2AAR (13), the D₂-dopamine receptor (14), and the m₂-muscarinic receptor (9).

Biologically regulated changes in the palmitoylation state of either receptors or G proteins may have important functional consequences. For example, mutations that prevent palmitoylation of various G subunits have been found to inhibit their association with the plasma membrane and thus their signaling function (15-18), suggesting that biological modulation of the G protein palmitoylation state could regulate their signaling properties. Palmitoylation of G_s has also been reported to increase its affinity for G (19). For receptors, abolition of palmitoylation by site-directed mutagenesis has been shown to either decrease coupling to G proteins (9, 20-23), affect receptor internalization (24-26), or modulate receptor phosphorylation by regulatory kinases (5, 27, 28).

Despite these potentially important roles, very little is known concerning the mechanism that regulates palmitoylation of these proteins. Both enzymatic (29-32) and nonenzymatic (33, 34) acylation reactions have been proposed for G, whereas an enzyme that can catalyze the depalmitoylation of G proteins has recently been identified (35). However, the mechanisms by which activation of the signaling pathway could control the acylation/deacylation cycle remain unknown. Analysis of the effects of stimulation on the palmitoylation status of receptors and G proteins is complicated by several factors (for a review, see Ref. 36). These include the fact that, following the initial conformational changes and protein-protein interactions that are promoted by receptor stimulation, multiple processes that limit the extent of the activation and contribute to signal termination come rapidly into play. It follows that it is difficult to temporally distinguish between the early events that lead to activation from the ones involved in rapid desensitization of the signaling system. This is an important problem because these two sets of events could theoretically have opposite effects on the palmitoylation reaction. Indeed, on a time scale that is virtually indistinguishable from that of the activation of the G proteins, stimulation of the receptors leads to their progressive functional inactivation. This desensitization results largely from agonist-promoted phosphorylation, uncoupling, and internalization of the receptors (37, 38). Internalization of the G proteins has also been suggested to contribute to desensitization of the signaling unit (39-43).

In an effort to distinguish between the effects of activation and desensitization on receptor and G protein palmitoylation, we took advantage of a ₂AR-G_s fusion protein that can be activated but not desensitized, internalized, or down-regulated (44, 45). The pharmacological properties of such receptor-G protein fusion constructs has recently attracted considerable attention and many of their properties have been recently reviewed (46, 47). The agonist-bound ₂AR-G_s fusion protein presumably mimics an early intermediate in the normal activation cycle. Also of interest to the present study is the fact that complete physical dissociation between the receptor and G_s, which normally follows the initial stimulatory interaction, is not permitted in the fusion protein. These features of the fusion protein allow study of the effects of early activation events on the palmitoylation state of the receptor and G protein independently of those resulting from the inactivation processes. Furthermore, the use of fusion protein restricts the analysis to those receptors and G proteins that did physically interact in the course of the experiment. We report that stimulation of ₂AR-G_s with -adrenergic agonists promotes rapid depalmitoylation and inhibits repalmitoylation of both the receptor and the G subunit. This contrasts with the facilitated repalmitoylation that is observed when the two proteins are expressed individually and suggests that early events in the activation process lead to the depalmitoylation of the two proteins, whereas later deactivation mechanisms, that do not occur for the fusion protein, are required for the repalmitoylation reaction.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Materials-- Grace's insect medium, lactalbumin, yeastolate, penicillin, streptomycin, glutamine, fungizone, pluronic acid, and phosphate-buffered saline were from Life Technologies, Inc. Fetal bovine serum was obtained from Immunocorp. [¹²⁵I]CYP, [-³²P]ATP, [-³²P]ATP, [³H]cAMP, and (9,10)-[³H]palmitate were purchased from Mandel. Alprenolol, isoproterenol, dichloroisoproterenol, ATP, GTP, cAMP, forskolin, isobutylmethylxanthine, phosphoenolpyruvate, bovine serum albumin, myokinase, cyanogen bromide, and anti-Flag^TM M2 were obtained from Sigma. Pyruvate kinase and Geneticin were from Calbiochem. ICI118551 was from Tocris. Benzamidine, soybean trypsin inhibitor, leupeptin, and n-dodecyl--D-maltoside was from Alexis Corp.

Recombinant Baculoviruses Construction-- The recombinant c-Myc-₂AR baculovirus was generated by subcloning the cDNA of a c-Myc-tagged human ₂AR (5) into the pJVELTZ recombination plasmid (InVitrogen). The pBacPAK-pHIS-₂AR-G_s was constructed by inserting in phase the Klenow-filled NcoI-SalI cDNA fragment of the fusion protein ₂AR-G_s (45, 51) into the Klenow-filled BamHI-EcoRI pBacPAK1-polyHIS vector (CLONTECH). The constructs were confirmed by DNA sequencing. The viruses were then produced by homologous recombination in Sf9 cells according to standard procedures (52). The recombinant baculovirus encoding the ₂AR-Thr-G_s construct was generated as described previously (53). Following infection of Sf9 cells with the appropriate viruses, expression of ₂AR and ₂AR-G_s was assessed by radioligand binding assays and Western blot analysis.

Cell Culture, Metabolic Labeling, and Membrane Preparations-- Sf9 cells were cultured in Grace's supplemented media containing 10% fetal bovine serum, 0.001% pluronic acid in spinner flasks (Bellco Glass) at 27 °C. Cells (2 × 10⁶/ml) were infected with the recombinant baculoviruses at a multiplicity of infection varying between 2 and 5 for 48 to 72 h. [³H]Palmitate labeling was then carried out in cells expressing the ₂AR or ₂AR-G_s fusion proteins. Cells were harvested and placed in serum-free medium for 1 h prior to the start of metabolic labeling. [³H]palmitate dissolved in a minimal volume of dimethyl sulfoxide was then added (100 µCi/millions of cells), and the cells were incubated at 27 °C in the presence or absence of -adrenergic ligands for various periods of time as described previously (7). In some experiments, labeling was allowed to proceed for 45 min before -adrenergic ligands were added. Labeling was stopped by chilling the reaction on ice. Cells were centrifuged at 500 × g for 5 min at 4 °C, rinsed twice with ice-cold PBS and resuspended in 20 ml of an ice-cold lysis buffer containing 20 mM Tris-HCl, 5 mM EDTA, pH 7.4, and the following protease inhibitors: 5 µg/ml leupeptin, 5 µg/ml soybean trypsin inhibitor, and 10 µg/ml benzamidine. Cells were disrupted by sonication and the lysate was centrifuged 5 min at 500 × g at 4 °C. The supernatant was then centrifuged at 45,000 × g for 20 min at 4 °C. The pelleted membranes were then resuspended in 10 mM Tris-HCl, 100 mM NaCl, 2 mM EDTA, pH 7.4, containing 0.3% n-dodecyl--D-maltoside and protease inhibitors (solubilization buffer). Solubilization was carried out for 90 min at 4 °C, and solubilized receptors were purified as described below.

Receptor Affinity Purification-- Alprenolol-Sepharose affinity purification matrix was synthesized according to the method of Benovic et al. (54). This matrix was used to purify the Sf9-derived ₂AR, ₂AR-G_s, and ₂AR-Thr-G_s as described previously (5). The affinity purified preparations were concentrated using Centriprep and Centricon cartridges (Amicon), and the amount of ₂AR, ₂AR-G_s, or ₂AR-Thr-G_s in each sample was determined by [¹²⁵I]CYP soluble radioligand binding assay as described elsewhere (55).

Hydroxylamine Treatment, Chemical and Enzymatic Cleavages-- For hydroxylamine treatment, purified ₂AR or ₂AR-G_s was mixed to an equal volume of 1 M Tris, pH 7.0, containing or not 1 M NH₂OH and incubated overnight at 4 °C. Cyanogen bromide (CNBr) cleavage was carried out following a protocol described by Luo et al. (56). Briefly, affinity purified ₂AR-G_s were separated by SDS-PAGE in nonreducing condition. The proteins were then transferred electrophoretically to nitrocellulose membrane. The band corresponding to ₂AR-G_s was cut out and the strip submerged in 500 µl of 70% (v/v) formic acid. The cleavage reaction was started by adding 10 µl of 5 M CNBr in acetonitrile. The reaction was allowed to proceed for 180 min at room temperature in the dark. The thrombin cleavage of the ₂AR-Thr-G_s construct was carried out on affinity purified fusion protein using 10 NIH units/ml of thrombin from human placenta in the solubilization buffer containing only 0.03% n-dodecyl--D-maltoside for 30 min at room temperature.

SDS-PAGE and Western Blot Analysis-- Affinity purified ₂AR, ₂AR-G_s, ₂AR-Thr-G_s, ₂AR-G_s cleaved with CNBr, or ₂AR-Thr-G_s cleaved with thrombin were resolved on nonreducing (or mildly reducing, 10 mM dithiothreitol, when specified) 10-15% slab gels containing 6 M urea. The gels were then fixed, incubated in Enlightning (DuPont), or in 1 M salicylic acid, dried, and exposed to DuPont REFLECTION^TM films from 10 to 60 days at 80 °C. Fluorograms were scanned and digitized (Hewlett Packard laser scanner), and densitometric analysis was carried out using the NIH Image program. When Western blot analyses were performed, aliquots of the samples were loaded in parallel gels and transferred onto nitrocellulose membranes. The ₂AR-Thr-G_s and ₂AR moiety of the fusion protein were first visualized using the Flag M2 antibody (dilution 1:10,000) and the Renaissance chemiluminescence reagent plus (Mandel). The same membrane was then stripped using 100 mM glycine, pH 2.2 (57), and reprobed, using a rabbit polyclonal antibody against the carboxyl-terminal portion of G_s subunit (dilution 1:8, 500) (a generous gift of Dr. A. D. Strosberg, Institut Cochin de Génétique Moléculaire, Paris), to reveal the ₂AR-Thr-G_s and G_s moiety of the fusion protein.

Radioligand Binding Assay-- Sf9 cells infected with the recombinant ₂AR or ₂AR-G_s baculoviruses were harvested and rinsed twice with ice-cold phosphate-buffered saline, and the membranes were prepared according to Mouillac et al. (5). Membrane suspensions were added to obtain a concentration of 2-10 µg/ml in a final volume of 500 µl of 75 mM Tris, 12.5 mM MgCl₂, 2 mM EDTA containing a saturating concentration (250 pM) of the radiolabeled -adrenergic antagonist [¹²⁵I]CYP. Nonspecific binding was determined as the residual binding observed in the presence of 10 µM alprenolol. Binding reactions carried out at room temperature for 90 min were stopped by rapid filtration over glass-fiber filters.

Adenylyl Cyclase Assay-- Adenylyl cyclase activity was determined in membrane preparations according to the method of Salomon et al. (58). Activities were determined in the presence or absence of the following activators: 1 µM isoproterenol, 10 µM alprenolol, 10 µM dichloroisoproterenol, or 100 µM forskolin. Data were expressed as picomoles of cAMP produced per min per mg of protein. Protein concentrations were measured by the method of Bradford (Bio-Rad) using bovine serum albumin as standard (59).

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Functional Characteristics of the ₂AR-G_s Fusion-- Fig. 1 illustrates the two fusion protein constructs between ₂AR and G_s (₂AR-G_s and ₂AR-Thr-G_s) that were used in the present study. Infection of Sf9 cells with recombinant baculoviruses encoding either of the two ₂AR-G_s fusion proteins conferred both ₂AR binding (data not shown) and -adrenergic-stimulated adenylyl cyclase activities (Fig. 2) confirming that, as observed in mammalian systems (44, 45, 48), the fusion proteins are synthesized, translocated to the plasma membranes, and functional. Also in agreement with what was observed in mammalian systems, sustained stimulation of ₂AR-G_s with an agonist does not promote any desensitization of the -adrenergic-stimulated adenylyl cyclase activity. This is in sharp contrast with the rapid desensitization observed in Sf9 cells expressing the wild type ₂AR. Indeed, as seen in Fig. 2, pretreatment of ₂AR expressing cells with 1 µM isoproterenol for 30 min reduced the isoproterenol-stimulated adenylyl cyclase activity by 23% without significantly affecting the basal activity, thus leading to a desensitization of 40% of the net agonist-stimulated adenylyl cyclase activity. The same treatment was without effect on the isoproterenol-stimulated adenylyl cyclase activity in cells expressing the ₂AR-G_s fusion protein. Agonist pre-treatment for as long as 24 h was also unable to promote any desensitization of the fusion protein. This characteristic of ₂AR-G_s was interpreted by Bertin et al. (44) as an indication that the covalent complex can become activated but does not enter the deactivation path upon agonist stimulation. This contention is also supported by the observation that the formation of the nucleotide-sensitive high affinity state for agonist can be readily observed for ₂AR-G_s in both mammalian (44) and Sf9 (49, 50) cells but that no agonist-promoted internalization or down-regulation was observed upon sustained stimulation (Ref. 44, and data not shown).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   The 2AR-Gs fusion constructs. Schematic representation of the two 2AR-Gs fusion proteins used in this study. Upper panel, the 2AR-Gs fusion protein constructed as in Bertin et al. (45). This construct links the bovine Gs through its amino-terminal methionine to the carboxyl terminus of the poly-His-tagged human 2AR. Lower panel, the 2AR-Thr-Gs fusion protein constructed as in Seifert et al. (53). This construct links the human 2AR, tagged at its amino and carboxyl termini by the Flag epitope and a histidine hexamere, respectively, and bovine Gs through an engineered thrombin cleavage site. The wavy lines represent the sites of palmitoylation.

View larger version (36K): [in this window] [in a new window]   Fig. 2.   Agonist-promoted stimulation and desensitization of adenylyl cyclase activity. Membrane preparations derived from Sf9 cells infected with the wild type human 2AR or the 2AR-Gs recombinant baculoviruses were used to measure the adenylyl cyclase activity. The activity was determined in the presence of vehicle (control) (CTL) or 1 µM of the -adrenergic agonist isoproterenol (ISO). To determine the extent of agonist-promoted desensitization, membranes were prepared from cells pretreated (hatched bars) or not (dotted bars) with 1 µM isoproterenol for 30 min. Receptor expression level as assessed by [125I]CYP binding was: 8.9 ± 2 and 2.8 ± 1.1 pmol/mg of protein for 2AR and 2AR-Gs, respectively, and was not affected by the agonist pre-treatment. Data represent the mean ± S.E. of five independent experiments. * indicates a statistically significant difference (p < 0.05).

It follows that ₂AR-G_s provides a convenient model to study the early events involved in the activation of the ₂AR-Gs complex without the confounding effects of the regulatory events leading to deactivation. This may be particularly important when considering that, following their initial activating interaction, the receptor and G_s may be targeted to distinct cellular compartments upon dissociation. Therefore, ₂AR-G_s allows study of the early events linked to the activation of a single G protein by a unique receptor molecule at equimolar ratio in a common cellular compartment.

Palmitoylation of ₂AR-G_s-- Based on the premises described above, we undertook study of the dynamics of ₂AR-G_s palmitoylation and the effect of agonist activation on the palmitoylation state of this complex. As shown in Fig. 3, metabolic labeling of Sf9 cells expressing either ₂AR or ₂AR-G_s led to the incorporation of [³H]palmitate in the two proteins. The major radiolabeled bands of ~45 and ~96 kDa obtained following alprenolol-Sepharose affinity purification corresponded to the expected molecular masses for the ₂AR and ₂AR-G_s, respectively, when expressed in Sf9 cells (5). Molecular species with identical electrophoretical mobilities were detected in Western blot analysis using anti-₂AR and anti-_s antibodies (data not shown), thus confirming the identity of the fusion protein. In Fig. 3A, identical numbers of ₂AR and of ₂AR-G_s, as assessed by radioligand binding, were loaded but densitometric analysis revealed around 1.7 times higher [³H]palmitate incorporation into the ₂AR-G_s than ₂AR, consistent with the fact that two palmitoylation sites are present in the fusion protein as compared with only one in the receptor. Fig. 3B illustrates the sensitivity of the labeling to hydroxylamine treatment, indicating that the [³H]palmitate was covalently attached to both ₂AR and ₂AR-G_s via thioester bonds.

View larger version (37K): [in this window] [in a new window]   Fig. 3.   Palmitoylation of 2AR and 2AR-Gs via hydroxylamine-sensitive thioester linkage. Sf9 cells infected with the 2AR or the 2AR-Gs recombinant baculoviruses were labeled for 45 min with [3H]palmitate. Receptor and fusion proteins were then purified by alprenolol-Sepharose chromatography and prepared for nonreducing SDS-PAGE. Panel A, 0.3 pmol of 2AR and 2AR-Gs, as assessed by [125I]CYP binding, were loaded in lanes 1 and 2, respectively. Panel B, the purified receptor and fusion protein were treated (+) or not () with NH2OH (1 M; pH 7.0) prior to SDS-PAGE. 2.2 pmol of 2AR were loaded in lanes 1 and 2, and 0.3 pmol of 2AR-Gs was loaded in lanes 3 and 4. The fluorogram shown is representative of three independent experiments.

Agonist-promoted Depalmitoylation of ₂AR-G_s-- To assess the effect of receptor stimulation on the dynamics of palmitoylation, pulse labeling experiments were carried out in the presence or absence of agonists for periods varying between 5 and 60 min. In the absence of agonist, incorporation of palmitate into ₂AR-G_s increased almost linearly for the first 30 min of labeling and remains stable thereafter (Fig. 4). The presence of isoproterenol during the labeling period greatly inhibited the incorporation of [³H]palmitate in the fusion protein. This unexpected result contrasts sharply with the agonist-promoted increase in palmitate incorporation observed on ₂AR and G_s when these proteins are expressed individually (5, 7, 10-12).

View larger version (30K): [in this window] [in a new window]   Fig. 4.   Effects of isoproterenol on the incorporation of [3H]palmitate into 2AR-Gs. Sf9 cells infected with the 2AR-Gs recombinant baculovirus were labeled with [3H]palmitate for the indicated times in presence of 1 µM isoproterenol (ISO and +) or the vehicle alone (control) (CTL and ). The fusion protein was then purified by alprenolol-Sepharose chromatography and resolved by SDS-PAGE. Identical numbers of 2AR-Gs (0.5 pmol), as assessed by [125I]CYP binding, were loaded in each lane. Relative incorporation of [3H]palmitate into 2AR-Gs was estimated by densitometric analysis of the fluorograms. The graph shown is the mean ± S.E. of three independent experiments.

For the ₂AR and G_s expressed separately, the increase in palmitate turnover was linked to a faster rate of depalmitoylation upon agonist stimulation (7, 12). The apparent increase in [³H]palmitate incorporation was thus attributed to a concomitant acceleration of the repalmitoylation reaction. It follows that the agonist-promoted reduction of [³H]palmitate incorporation into the ₂AR-G_s fusion protein could result from a slower depalmitoylation or reflect an inhibition of the repalmitoylation reaction. To distinguish between these two hypotheses, cells were metabolically labeled with [³H]palmitate in the absence of agonist. Following a 45-min pulse period, corresponding to the period required to attain steady state labeling, isoproterenol was added or not in the continued presence of [³H]palmitate and incubated for an additional 5 or 15 min. As seen in Fig. 5, incubation with isoproterenol rapidly reduced the extent of ₂AR-G_s palmitoylation, thus suggesting that agonist stimulation promotes its rapid depalmitoylation and that repalmitoylation of the active complex cannot occur. This is in sharp contrast with the increased repalmitoylation that is observed when identical treatment is carried out in cells expressing the wild type ₂AR as an individual protein (Fig. 5B).

View larger version (32K): [in this window] [in a new window]   Fig. 5.   Isoproterenol-promoted depalmitoylation of 2AR-Gs. Sf9 cells infected with 2AR or 2AR-Gs recombinant baculovirus were labeled with [3H]palmitate for 45 min. Cells were then treated with 1 µM isoproterenol (ISO and +) or the vehicle (control) (CTL and ) alone, for 5 or 15 min in the continued presence of [3H]palmitate. The fusion protein was then purified by alprenolol-Sepharose chromatography and resolved by SDS-PAGE. Identical numbers of receptor, as assessed by [125I]CYP binding, were loaded in each lane, and the relative incorporation of [3H]palmitate into 2AR-Gs was estimated by densitometric analysis of the fluorograms. The bar graph shown in panel A represents the mean ± S.E. of four independent experiments. The fluorogram in panel B shows the effect of a 5-min treatment with isoproterenol following a 45-min pulse labeling on [3H]palmitate incorporation into wild type 2AR (lanes 1 and 2) and into 2AR-Gs (lanes 3 and 4). Equivalent numbers of receptor (0.9 pmol) were loaded in each lane.

Because the ₂AR-G₂ fusion protein can be activated but that later processes of inactivation such as G protein dissociation, desensitization, or internalization do not occur, it could be hypothesized that early events leading to activation of the receptor-G protein complex promote depalmitoylation but that later processes are required for repalmitoylation. The fact that the depalmitoylation and repalmitoylation reactions may occur with very similar kinetics when the ₂AR and G_s are expressed as individual proteins may explain why Jones et al. (8) did not observe any change in the stoichiometry of palmitoylation of G_s upon activation despite the universally observed increase in the turnover rate of the G_s-bound palmitate. The stabilization of the activated receptor-G protein complex using ₂AR-G_s allowed isolation of the effects that resulted solely from the activation process.

The effect of -adrenergic ligands of various levels of intrinsic activity was then assessed on the palmitoylation of ₂AR-G_s. As shown in Fig. 6, the addition of all ligands caused a significant reduction in the incorporation of the labeled fatty acid into ₂AR-G_s. Interestingly, the extent of the decrease in labeling was directly correlated (r² = 0.992) to the intrinsic activity of the compounds toward ₂AR-G_s as assessed in a membrane adenylyl cyclase assay (Fig. 6B). However, direct stimulation of cAMP production by forskolin did not affect the palmitoylation of ₂AR-G_s (Fig. 6C), thus suggesting that the agonist-promoted depalmitoylation is linked to conformational changes imposed by the agonists and not to second messenger generation.

View larger version (36K): [in this window] [in a new window]   Fig. 6.   Effects of -adrenergic ligands of various intrinsic efficacy on the palmitoylation of 2AR-Gs. Sf9 cells infected with the 2AR-Gs recombinant baculovirus were labeled with [3H]palmitate for 1 h in presence of alprenolol (ALP), dichloroisoproterenol (DCI), isoproterenol (ISO), forskolin (FK), or the vehicle (CTL). Affinity chromatography purified 2AR-Gs was then resolved by SDS-PAGE. In each experiment, identical amount of fusion protein obtained from the experimental and control conditions was loaded. Relative incorporation of [3H]palmitate into 2AR-Gs was estimated by densitometric analysis of the fluorograms. In panels A and C, the bar graphs shown are the mean ± S.E. of five independent experiments. Panel B, linear regression between the extent of agonist-stimulated adenylyl cyclase activity observed for the various ligands in membranes derived from cells expressing 2AR-Gs and the relative [3H]palmitate incorporation observed into 2AR-Gs in the presence of the same ligands. Data represent the mean of six independent experiments.

[³H]Palmitate Incorporation into the ₂AR and G_s Moieties of the Fusion Protein-- Cysteine 341 of ₂AR and cysteine 3 of G_s, corresponding to position 358 and 428 in ₂AR-G_s, respectively, represent the confirmed palmitoylation sites of these two proteins (20, 29). In the experiments described above, palmitoylation of ₂AR-G_s was studied as a whole with no specific consideration of the individual palmitoylation sites. To determine whether the two sites were indeed palmitoylated and to assess if agonist treatment had similar effects on the palmitoylation state of the two proteins, we took advantage of another fusion protein construct in which a thrombin cleavage site was engineered between the receptor and G_s (₂AR-Thr-G_s; see Fig. 1). As a control, thrombin treatment was performed on wild type ₂AR without any effect on the palmitoylation state nor the integrity of the receptor (data not shown). Fig. 7A shows that thrombin treatment of the purified fusion protein, following metabolic labeling, generated two labeled proteins corresponding to the expected mobility for ₂AR and G_s, indicating that the two proteins were palmitoylated within the fusion construct. The identity of the cleaved fragments was further confirmed by Western blot analysis using the anti-_s antibody to detect G_s and the anti-Flag M2 antibody to detect the Flag epitope-bearing ₂AR. The apparently higher [³H]palmitate incorporation observed into the ₂AR band when compared with G_s most likely reflects the presence of some background labeling observed in this region of the gel even in the absence of thrombin.

View larger version (44K): [in this window] [in a new window]   Fig. 7.   Thrombin cleavage of 2AR-Thr-Gs. Sf9 cells were infected with a baculovirus encoding the 2AR-Thr-Gs fusion protein. Metabolic labeling with [3H]palmitate was then carried out for 45 min before adding 1 µM isoproterenol (ISO) or not (control) (CTL) for an additional 15 min. The fusion protein was purified by alprenolol-Sepharose affinity chromatography and treated (+) or not () with thrombin (Thr, 10 NIH units/ml) for 30 min. The reactions were resolved under nonreducing conditions (panel A) or mildly reducing (10 mM dithiothreitol) conditions (panel B) by SDS-PAGE containing 6 M urea. 1.2 pmol of receptor, as assessed by [125I]CYP soluble radioligand binding, were loaded in each lane in both panels A and B. Relative incorporation of [3H]palmitate into the fusion 2AR-Thr-Gs or the 2AR and the Gs moieties were estimated by densitometric analysis of the fluorograms. The fluorograms shown are representative of four independent experiments. In each case, parallel SDS-PAGE were carried out for Western blot analysis, as described under "Experimental Procedures," to confirm the identity of each protein.

As previously observed for ₂AR-G_s, the presence of isoproterenol during the metabolic labeling induced a significant reduction of the [³H]palmitate incorporation into ₂AR-Thr-G_s. Thrombin cleavage revealed that the overall decrease in the radiolabeling of the fusion protein was the consequence of a reduction of [³H]palmitate incorporation into both the receptor and G_s. The agonist-induced G_s depalmitoylation that we observed could be mediated, in part, by the receptor-promoted dissociation of subunits from the activated fusion protein. Indeed, as reported by Iiri et al. (19), did protect GDP-bound _s but not _s-GTP[S] from depalmitoylation by a recombinant esterase. Because, nonreducing SDS-PAGE conditions could lead to aggregation of some proteins, including the receptor, reducing conditions were also used. As shown in Fig. 7B, identical results were obtained when receptor and G_s were resolved under mildly reducing conditions (10 mM dithiothreitol) that diminished aggregation and promoted only partial chemical depalmitoylation.

Palmitoylation of both receptor and G_s and the effect of isoproterenol on the two proteins was further confirmed using CNBr hydrolysis of the ³H-palmitoylated ₂AR-G_s construct. The primary sequence of ₂AR-G_s containing 18 methionines, complete cleavage should generate 19 fragments (Fig. 8A). Given that the two palmitoylation sites are located on two distinct fragments, two peptides distinguishable by their size are expected to be ³H-palmitoylated. Calculated masses for the expected palmitoylated fragments are 16.4 and 7.2 kDa corresponding to the ₂AR- and the G_s-derived peptides, respectively. As shown in Fig. 8B, CNBr treatment yielded two peptides of the expected electrophoretic mobility, confirming that both the receptor and G_s were palmitoylated within the fusion protein. The difference in the labeling intensity of the two bands most likely results from quantitatively different elution and recovery of the two fragments from the nitrocellulose membrane during the hydrolysis. Weakly labeled bands at 21 and 44 kDa represent partial cleavage products. Addition of isoproterenol following a 45-min pulse labeling with [³H]palmitate promoted the depalmitoylation of both the 16.4 and 7.2 kDa fragments, confirming once more that agonist activation favored the depalmitoylation of both the receptor and G_s.

View larger version (44K): [in this window] [in a new window]   Fig. 8.   CNBr cleavage of 2AR-Gs. Panel A, expected sizes of the 19 fragments obtained after complete CNBr cleavage of the fusion protein, using the program PROLYSIS. Because the two palmitoylation sites are located on the carboxyl-terminal portion of the 2AR moiety and the amino-terminal portion of the Gs moiety, two distinct palmitoylated fragments of 16.4 and 7.2 kDa and corresponding to the 2AR- (*) and the Gs (**)-derived peptides, respectively, should be produced. Panel B, infected Sf9 cells expressing 2AR-Gs were labeled for 45 min with [3H]palmitate and then treated with the vehicle (CTL, lane 1) or with isoproterenol (ISO, lane 2) for 15 min in the continued presence of labeled palmitate. The radiolabeled fusion protein was purified, resolved by SDS-PAGE, and electrophoretically transferred to a nitrocellulose membrane. Bands corresponding to the fusion protein were then treated with CNBr for 3 h. Generated peptides were separated by SDS-PAGE. The fluorogram shown is representative of two independent experiments.

Taken together, the data presented in this study show that the initial events leading to activation of the ₂AR-G_s complex promote the rapid depalmitoylation of both receptor and G_s and that sustained activation prevents repalmitoylation occurring. The extent of depalmitoylation is directly proportional to the intrinsic activity of the agonist and most likely depends on conformational changes and protein-protein interactions that are stabilized by the activating ligands. Also, this study demonstrates for the first time that depalmitoylation and repalmitoylation occur during distinct phases of the receptor-G protein activation/inactivation cycle. Further studies are now required to determine how receptor activation leads to depalmitoylation of the receptor-G protein complex and why deactivation is required for repalmitoylation to occur. In particular, the role played by the newly characterized acyl-protein thioesterase (35) in the activation driven depalmitoylation will need to be assessed.

	FOOTNOTES

^* This work was supported by grants from the Medical Research Council of Canada (MRCC), NATO and by MRCC studentships (to T. P. L. and L. A.), a Heart and Stroke Foundation of Canada fellowship (to H. A.), and a Deutsche Forschungsgemeinschaft research fellowship (to R. S.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

These authors contributed equally to this work.

^¶ Present address: LEBS-CNRS, 91198 Gif-sur-Yvette, France.

Present address: Dept. Pharmacology & Toxicology, the University of Kansas, Lawrence, KS 66046.

^§§ MRCC scientist. To whom correspondence should be addressed: Département de Biochimie, Faculté de Médecine, P. O. Box 6128, Succ. Centre-Ville, Montréal, Québec H3C 3J7, Canada. Tel.: 514-343-6372; Fax: 514-343-2210; E-mail: bouvier@bcm.umontreal.ca.

	ABBREVIATIONS

The abbreviations used are: G_s, stimulatory  subunit of trimeric G protein; ₂AR, ₂-adrenergic receptor; Sf9, Spodoptera frugiperda; G protein, guanine nucleotide-binding protein; ₂AR-G_s, fusion protein linking G_s and a histidine hexamere to the ₂AR carboxyl terminus and amino terminus, respectively; ₂AR-Thr-G_s, fusion protein linking the ₂AR tagged at its amino and carboxyl termini by the Flag epitope and histidine hexamere, respectively, and G_s through an engineered thrombin cleavage site; PAGE, polyacrylamide gel electrophoresis; [¹²⁵I]CYP, radio labeled iodocyanopindolol.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES