Characterization of the Phosphorylation Sites Involved in G Protein-coupled Receptor Kinase- and Protein Kinase C-mediated Desensitization of the α1B-Adrenergic Receptor*

Catecholamines as well as phorbol esters can induce the phosphorylation and desensitization of the α1B-adrenergic receptor (α1BAR). In this study, phosphoamino acid analysis of the phosphorylated α1BAR revealed that both epinephrine- and phorbol ester-induced phosphorylation predominantly occurs at serine residues of the receptor. The findings obtained with receptor mutants in which portions of the C-tail were truncated or deleted indicated that a region of 21 amino acids (393–413) of the carboxyl terminus including seven serines contains the main phosphorylation sites involved in agonist- as well as phorbol ester-induced phosphorylation and desensitization of the α1BAR. To identify the serines invoved in agonist- versus phorbol ester-dependent regulation of the receptor, two different strategies were adopted, the seven serines were either substituted with alanine or reintroduced into a mutant lacking all of them. Our findings indicate that Ser394 and Ser400 were phosphorylated following phorbol ester-induced activation of protein kinase C, whereas Ser404, Ser408, and Ser410 were phosphorylated upon stimulation of the α1BAR with epinephrine. The observation that overexpression of G protein-coupled kinase 2 (GRK2) could increase agonist-induced phosphorylation of Ser404, Ser408, and Ser410, strongly suggests that these serines are the phosphorylation sites of the α1BAR for kinases of the GRK family. Phorbol ester-induced phosphorylation of the Ser394 and Ser400 as well as GRK2-mediated phosphorylation of the Ser404, Ser408, and Ser410, resulted in the desensitization of α1BAR-mediated inositol phosphate response. This study provides generalities about the biochemical mechanisms underlying homologous and heterologous desensitization of G protein-coupled receptors linked to the activation of phospholipase C.

Desensitization is a general regulatory phenomenon of G protein-coupled receptors resulting in the attenuation of the receptor-mediated response. Two major patterns of desensitization referred to as homologous and heterologous desensitization can be distinguished. Homologous desensitization is defined as a rapid loss of responsiveness for a receptor repeatedly exposed to its specific agonist, whereas in heterologous desensitization stimulation of a receptor by an agonist can attenuate the response mediated by other receptors eliciting similar cellular effects (1).
In the G protein-coupled receptor family (2), receptor desensitization has been extensively characterized for rhodopsin mediating phototransduction in retinal rod cells and for the ␤ 2adrenergic receptor (␤ 2 AR) 1 which mediates catecholamineinduced stimulation of adenylyl cyclase. The second messengerdependent cAMP-dependent protein kinase can phosphorylate and desensitize the ␤ 2 AR both in response to its agonist as well as to other agents increasing the cellular content of cAMP. On the other hand, a prominent role in homologous desensitization of rhodopsin and ␤ 2 AR is played by the second messengerindependent rhodopsin kinase (3) and ␤-adrenergic receptor kinase (␤ARK) (4), respectively. Once the receptor is occupied by the agonist, it is recognized by the kinase and becomes phosphorylated. The subsequent uncoupling of the receptor and G protein is then mediated by arrestin proteins, which specifically bind to the phosphorylated receptor (5,6). Rhodopsin kinase and ␤ARK are members of the newly discovered family of G protein-coupled receptor kinases (GRK) (7). These protein kinases have the unique ability to recognize and phosphorylate their G protein-coupled receptor substrates predominantly in their active (i.e. agonist-occupied) conformations. Recently, we have provided evidence that the ␣ 1B AR coupled to Gq-mediated activation of phospholipase C can be phosphorylated by at least two types of protein kinases, a protein kinase C (PKC) upon its stimulation by phorbol esters (8) and protein kinases belonging to the GRK family once the receptor is stimulated by its agonist (9). The observation that agonist-induced phosphorylation of the ␣ 1B AR is insensitive to inhibitors of PKC suggests that a PKC is involved mainly in heterologous desensitization, whereas a GRK mediates agonist-induced desensitization (homologous) of the ␣ 1B AR (8).
The sites involved in GRK-mediated phosphorylation have been unambiguously identified only for rhodopsin and more recently for the ␤ 2 AR. Rhodopsin is phosphorylated in a lightdependent manner by rhodopsin kinase (GRK1). Both in vitro and in vivo studies have demonstrated that the sites for GRK1mediated phosphorylation are represented by specific serines located in the distal part of the C terminus of rhodopsin (10 -15), whereas for the ␤ 2 AR the phosphorylation sites for ␤ARK (GRK2) and GRK5 have been identified only in vitro (16). For few other G protein-coupled receptors, including ␣ 2A AR (17), N-formyl peptide receptor (18) and C5a receptor (19) potential phosphorylation sites for GRK2-mediated phosphorylation have been proposed on the basis of mutagenesis studies.
The phosphorylation sites for the second messenger-depend-ent PKC have been characterized only for a few G proteincoupled receptors, including rhodopsin (20), ␤ 2 AR (21), and more recently for the GLP1 receptor (22), using site-directed mutagenesis. Recently, we have demonstrated that the sites for phorbol ester-as well as agonist-induced phosphorylation of the ␣ 1B AR are located in the C-tail of the receptor as suggested by the fact that a truncated mutant (T368) lacking the last 147 amino acids was totally impaired in its ability to undergo phophorylation (8). In this study, we have constructed a large number of receptor mutants to identify the phosphorylation sites of the C-tail involved in agonist-versus phorbol esterinduced desensitization of the ␣ 1B AR. Our findings demonstrate that GRK and PKC can phosphorylate the ␣ 1B AR on different serine residues of the C-tail, triggering desensitization of the receptor-mediated response.

EXPERIMENTAL PROCEDURES
Construction of Mutated ␣ 1B ARs-The cDNA encoding the hamster ␣ 1B AR (23) was mutated by a polymerase chain reaction mutagenesis technique using Taq DNA polymerase (Boehringer Mannheim). The mutated DNA fragments obtained were digested with BssHII and ApaI and cloned into the expression vector pRK5 (24) containing the ␣ 1B AR cDNA. Recombinant clones were isolated and sequenced by cycle sequencing using Exo Ϫ Pfu polymerase (Stratagene).
Cell Culture and Transfection-COS-7 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum and transfected with different DNA by the DEAE-dextran method. The cDNA encoding the bovine GRK2 (25) was subcloned in pCMV5 (26). For phosphorylation experiments, COS-7 cells (5 ϫ 10 6 ) were grown in 100-mm dishes, and the transfected DNA was 1 g/ million cells for the receptors and for GRK2 3 g/million cells. The total amount of DNA transfected was kept constant (4 g/million cells) under different conditions adding pCMV. For inositol phosphate determination, COS-7 cells (0.25 ϫ 10 6 ) grown in 20-mm dishes were transfected with 0.2 g of DNA/million cells for the receptors and 0.6 g of DNA/ million cells for GRK2. The total amount of DNA transfected was kept constant (1 g/million cells) under different conditions adding pCMV. COS-7 cells were harvested 48 h after their transfection.
Western Blot Analysis of GRK 2-The expression of bovine GRK2 in transfected COS-7 cells was assessed as previously described (9) using an antiserum raised against a glutathione S-transferase fusion protein encoding residues 467-688 of rat GRK3 (27).
Ligand Binding-Membrane preparations derived from cells expressing the wild type or mutated ␣ 1B AR and ligand binding assays using [ 125 I]HEAT were performed as described previously (8). Prazosin (10 Ϫ6 M) was used to determine nonspecific binding. [ 125 I]HEAT concentration was 300 pM for saturation binding and 80 pM for competition binding analysis of epinephrine. Data were analyzed by nonlinear leastsquare regression analysis (28). 32 P-Labeling and Immunoprecipitation of the Receptors-COS-7 cells expressing the different ␣ 1B AR in the absence or presence of bovine GRK2 were grown in 100-mm dishes, equilibrated in phosphate-free DMEM for 2 h, and then incubated in the same buffer containing 32 P i (0.2 mCi/ml) for 2 h at 37°C. The incubation was then continued in the presence of epinephrine (10 Ϫ4 M) or PMA (10 Ϫ7 M). Solubilization and immunoprecipitation of the receptors was as described previously (9). After autoradiography, the 32 P content of gel slices containing the receptor was quantified by liquid scintillation spectroscopy and normalized to the receptor amount loaded on the gel.
Two-dimensional Phosphoamino Acid Analysis-The portion of gel encompassing the phosphorylated receptor was excised, and the protein was extracted and hydrolyzed with 6 N HCl. After thin layer chromatography using using a HTLE 7000 electrophoresis system (C.B.S. Scientific Company, Del Mar, CA), the phosphoamino acid content of the phosphorylated receptor was analyzed by autoradiography.
Inositol Phosphate Determination-For the determination of labeled total inositol phosphates, cells expressing the different receptors in the absence or presence of bovine GRK2 were grown in 20-mm dishes and labeled with [ 3 H]inositol at 4 mCi/ml for 15-18 h in inositol-free DMEM supplemented with 1% fetal bovine serum. After labeling, cells were stimulated for 20 min with epinephrine (10 Ϫ4 M) in the presence of 20 mM LiCl. Inositol phosphates were extracted as described previously (23)

RESULTS
Regions of the C-tail Involved in Agonist-versus PMA-induced Phosphorylation and Desensitization of the ␣ 1B AR-We previously showed (8) that the carboxyl portion of the ␣ 1B AR plays a crucial role in receptor regulation. This was demonstrated by the fact that truncation of the last 147 amino acids abolished both agonist-and phorbol ester-induced receptor phosphorylation and desensitization. To assess which amino acids could be acting as phosphate acceptors, we performed a phosphoamino acid analysis of the phosphorylated ␣ 1B AR immunoprecipitated from cells treated with epinephrine or PMA. As shown in Fig. 1, the phosphoamino acid analysis revealed that both exposure to epinephrine and PMA results in the phosphorylation of serine, but not of threonine or tyrosine, residues of the ␣ 1B AR.
To identify the amino acids of the ␣ 1B AR involved in agonistversus phorbol ester-induced phosphorylation, we compared the phosphorylation properties of five receptor mutants: four truncated mutants T469, T412, T387, and T368, lacking the last 46, 103, 128, and 147 amino acids, respectively, and a deletion mutant (⌬393-413) missing a region of 21 amino acids (393-413) containing seven serines (Fig. 2). Ligand binding of [ 125 I]HEAT on membranes from transfected COS-7 cells indicated that the expression levels of the different ␣ 1B ARs were similar. The B max ranged from 1 to 2 pmol/mg of protein, and the K d of [ 125 I]HEAT binding was approximately 80 pM for all the receptors. The K i values of epinephrine ranged from 10 Ϫ5 to 3 ϫ 10 Ϫ5 M for all receptors (data not shown).
To assess the phosphorylation properties of the receptors, the wild type ␣ 1B and mutated receptors were immunoprecipitated from transfected COS-7 cells metabolically labeled with [ 32 P]orthophosphate as described previously (8). Exposure of cells expressing the ␣ 1B AR to epinephrine (10 Ϫ4 M) or PMA (10 Ϫ7 M) for 15 min resulted in a maximal increase of receptor phosphorylation of 60 and 80% above basal, respectively (Fig.  3). As previously reported (8), the T368 receptor mutant could not be detected as a phosphorylated protein. Similarly, the T387 mutant did not display any basal or stimulated phosphorylation (Fig. 3B). On the other hand, the T469 and T412 truncated receptors could be detected as phosphorylated proteins, migrating at ϳ75 and ϳ70 kDa, respectively (Fig. 3A). When compared with the wild type ␣ 1B AR, the T469 mutant did not display any significative difference in receptor phosphorylation (Fig. 3B), suggesting that the serines located in the last 46 amino acids of the C terminus of the receptor are involved in neither agonist-nor in PMA-induced phosphorylation. In contrast, for the T412 receptor PMA-induced phosphorylation was reduced of about 30% as compared with that of the wild type receptor, whereas epinephrine-mediated phosphorylation was not significantly different (Fig. 3B). Altogether, these results indicate that the serines involved in the phosphorylation of the ␣ 1B AR are located in two different regions of the C terminus; the first including amino acids 387-412 contains the majority of the sites for both agonist-and PMA-induced phosphorylation, whereas the second, located between residues 412 and 469, includes sites involved in part of PMA-induced phosphorylation of the ␣ 1B AR. A crucial role of serine residues located between amino acids 387 and 412 was confirmed by the results obtained with the deletion mutant ⌬393-413 which was completely impaired in its ability to undergo receptor phosphorylation either upon stimulation with epinephrine or PMA (Fig. 3).
We next investigated the correlation between agonist-induced phosphorylation and desensitization of the inositol phosphate (IP) response mediated by the wild type and mutated ␣ 1B AR. In our previous work, we demonstrated that overexpression of GRK2 and GRK3 enhanced both agonist-induced phosphorylation and desensitization of the ␣ 1B AR (9). In agreement with our previous findings, overexpression of GRK2 caused an inhibition of 52% of the ␣ 1B AR-mediated IP response as compared with that of the receptor expressed alone (Fig. 4). The inhibitory effect of GRK2 on the receptor-mediated IP increase for the T469 and T412 receptor mutants was similar to that observed for the wild type ␣ 1B AR (Fig. 4). In contrast, the IP response mediated by the T368, T387, and ⌬393-413 receptors, which cannot undergo agonist-induced phosphorylation, was not significantly decreased by overexpression of bovine GRK2 (Fig. 4). This strongly suggests that the serines located between residues 393 and 413 are involved in both GRK-mediated phosphorylation and desensitization of the ␣ 1B AR.
In a second set of experiments, we assessed the correlation between PMA-induced phosphorylation and desensitization of the wild type and mutated receptors. Cells expressing the ␣ 1B AR or T368 receptor were treated with increasing concentrations of PMA for 10 min. Fig. 5 shows that at a concentration of 10 Ϫ8 M PMA inhibited the ␣ 1B AR-mediated response of about 45%, without any significant effect on the response mediated by the T368 mutant (Fig. 5A). However, at a concentration of 10 Ϫ7 M PMA could also impair the T368-mediated response, even if at a smaller extent as compared with the wild type receptor (25% versus 68% of impairment for the T368 and ␣ 1B AR, respectively). Thus, high concentrations of PMA (Ͼ10 Ϫ8 M) might inhibit the receptor-mediated IP response by a mechanism independent from receptor phosphorylation. This was confirmed by experiments exploring the effect of PMA on the IP response mediated by Gq␣ expressed in COS-7 cells following its stimulation with AlF 4 Ϫ . Fig. 5 shows that 10 Ϫ7 M PMA could inhibit the Gq␣-mediated response of about 30%, whereas no significant inhibition could be observed at a concentration of 10 Ϫ8 M (Fig. 5A).
Taken together, these results suggest that PMA can impair the ␣ 1B AR-mediated response by at least two different mechanisms, depending on its concentration. The first occurring at low PMA concentrations (10 Ϫ8 M) might be truly mediated by receptor phosphorylation, because it was not observed with the T368 receptor mutant lacking all the phosphorylation sites. On the basis of these observations, we compared the effect of a low concentration of PMA (10 Ϫ8 M) on the IP response mediated by different receptor mutants expressed in COS-7 cells. Fig. 5B shows that treatment of cells with PMA (10 Ϫ8 M) reduced the IP response of 40% for the wild type and T469 receptors and of 30% for the T412, without any significant effect on the response mediated by the T368, T387, and ⌬393-413 receptor mutants.
In conclusion, the lack of phorbol ester-induced phosphorylation observed for the T368, T387, and ⌬393-413 receptor mutants (Fig. 3) correlates well with the loss of PMA-mediated receptor desensitization. For the T412 mutant, PMA-induced phosphorylation and desensitization of the receptor were also closely related being both reduced as compared with the wild type ␣ 1B AR (Figs. 3 and 4).
Identification of the Serines Involved in Agonist-versus Phorbol Ester-induced Phosphorylation of the ␣ 1B AR-To identify the individual serine residues involved in agonist-versus phorbol ester-induced phosphorylation and desensitization of the ␣ 1B AR, we mutated all seven serines located between amino acids 393 and 413 (Ser 394 , Ser 398 , Ser 400 , Ser 404 , Ser 408 , Ser 410 , and Ser 413 ) (Fig. 2). The serine residues were substituted with alanine either individually or in different combinations as indicated in Fig. 6. For all the mutant receptors the B max and the K d of [ 125 I]HEAT as well as the K i for epinephrine binding were similar to those of the wild type ␣ 1B AR (data not shown). All the receptor mutants could mediate the epinephrine-induced IP response at a similar extent to that of the wild type ␣ 1B AR (results not shown).
As indicated in Fig. 6, mutagenesis of either Ser 394 or Ser 400 caused a decrease of the PMA-induced receptor phosphorylation of 31 and 38%, respectively, without any significant effect on agonist-induced phosphorylation. On the other hand, substitution of either Ser 404 or of the doublet Ser 408 -Ser 410 , resulted in a decrease of epinephrine-induced phosphorylation of 36 and 64%, respectively, without affecting PMA-induced receptor phosphorylation. Substitution of either Ser 398 or of the doublet Ser 413 -Ser 415 did not alter receptor phosphorylation. Finally, the substitution of all the serines located between residues 393 and 415 (S394 -415A mutant) resulted in an almost complete loss of the agonist-induced phosphorylation and in a reduction of 76% of the PMA-induced phosphorylation.
These findings identify, on one hand, Ser 394 and Ser 400 as sites of PMA-dependent phosphorylation and, on the other, Ser 404 , Ser 408 , and Ser 410 as sites of agonist-induced phosphorylation of the ␣ 1B AR. The fact that the mutagenesis of all the serines between amino acids 393 and 415 (S394 -415A mutant) did not completely abolish PMA-induced receptor phosphorylation suggests the existence of additional sites of phosphorylation for PKC. In agreement with this finding, pretreatment of the cells expressing the S394 -415A mutant with the specific PKC inhibitor RO-318220, completely abolished the residual agonist-as well as PMA-induced receptor phosphorylation (results not shown).
As mentioned above, the T412 truncated mutant, but not the T469, displayed a 30% decrease of the PMA-induced receptor phosphorylation (Fig. 3). This suggests that a PMA-dependent phosphorylation site is located between amino acids 413 and 469 of the ␣ 1B AR. To identify the residue which contributes to this residual phosphorylation, all the serines and threonines of the S394 -415A mutant located between amino acids 415 and 469 (Ser 421 , Ser 423 , Ser 425 , Ser 447 , Ser 452 , and Ser 465 ) as well as Ser 369 , Thr 385 , and Thr 390 , were individually mutated into alanine. None of these substitutions was able to abolish or reduce the residual PMA-induced phosphorylation of the S394 -415A receptor mutant (results not shown). This might reflect the fact that the mutation of one serine or threonine can be compensated by the phosphorylation of another.
To further characterize the role of individual serines in the phosphorylation of the ␣ 1B AR, serines 394, 398, 400, 404, 408, 410, 413, and 415 were reintroduced individually or in combination in the S394 -415A mutant, which lacks all of them. As shown in Fig. 6, the epinephrine-and PMA-induced phosphorylation of the S394 -415A mutant was only 12 and 24% of that of the wild type ␣ 1B AR, respectively. Reintroduction of Ser 394 (S394i) or Ser 400 (S400i) restored the PMA-, but not agonistinduced phosphorylation of the S394 -415A mutant to 64 and 71% of that of the the wild type ␣ 1B AR, respectively. The effect of Ser 394 and Ser 400 seemed to be additive as demonstrated by the fact that reintroduction of both (S394i/S400i) almost completely restored the PMA-dependent phosphorylation of the S394 -415A mutant to the level of the wild type ␣ 1B AR. The individual or combined reintroduction of Ser 404 , Ser 408 , and Ser 410 (S404i, S408i, S410i, S408i/S410i, and S404i/S408i/ S410i) did not have any effect on the phosphorylation of the Ser394 -415A mutant induced by PMA, but could restore that induced by epinephrine. In particular, reintoduction of Ser 404 , Ser 408 , and Ser 410 increased the phosphorylation of the S394 -415A mutant to levels ranging from 41 to 46% of that of the wild type ␣ 1B AR. The role of these serines in epinephrineinduced phosphorylation is additive as demonstrated by the fact that for the mutant S408i/S410i and S404i/S408i/S410i, the epinephrine-induced phosphorylation was 66 and 94% of that of the wild type ␣ 1B AR, respectively.
Involvement of Ser 404 , Ser 408 , and Ser 410 in GRK2-mediated Phosphorylation and Desensitization of the ␣ 1B AR-To investigate whether serines 404, 408, and 410 could be substrates of GRK2-mediated phosphorylation of the ␣ 1B AR, the wild type as well as the S404i, S408i/S410i and S404i/S408i/S410i receptors, carrying serines involved in agonist-induced phosphorylation, were coexpressed with bovine GRK2 in COS-7 cells. Overexpression of GRK2 significantly increased epinephrineinduced phosphorylation of the ␣ 1B AR as well as that of the S404i, S408i/S410i, and S404i/S408i/S410i receptors as compared with that of the receptors expressed alone (Fig. 7). These findings indicate that these three serines might be independ- ently phosphorylated by GRK2 following stimulation of the receptor with the agonist. Since the level of epinephrine-induced phosphorylation of the S404i/S408i/S410i mutant was comparable to that of the wild type ␣ 1B AR either in the absence or presence of GRK2, it is likely that Ser 404 , Ser 408 , and Ser 410 represent the only sites for agonist-induced phosphorylation which can be mediated by GRK2.
As shown in Fig. 8, overexpression of GRK2 in COS-7 cells decreased the ␣ 1B AR-mediated IP response by 52% as compared with that of the receptor expressed alone, without any significant effect on the response mediated by the T368 and S394 -415A receptor mutants, which were impaired in agonistinduced phosphorylation. However, overepression of GRK2 could reduce the IP response mediated by the S404i, S408i, and S410i mutants of about 30 -35% and that mediated by the S404i/S408i/S410i receptor of 49%. On the other hand, the IP response mediated by receptors carrying Ser 394 (S394i) and Ser 400 (S400i) was insensitive to the overexpression of GRK2. Altogether, these findings provide strong evidence that Ser 404 , Ser 408 , and Ser 410 , confer to the ␣ 1B AR the ability to undergo both GRK2-mediated phosphorylation and desensitization.
Involvement of Ser 394 and Ser 400 in PMA-induced Desensitization of the ␣ 1B AR-To assess whether the PMA-induced phosphorylation of Ser 394 and Ser 400 induced by PMA could impair the ␣ 1B AR-mediated IP response, cells expressing wild type ␣ 1B AR or different mutated receptors were treated with low concentration of PMA (10 Ϫ8 M) for 10 min. As shown in Fig.  9, in cells treated with PMA (10 Ϫ8 M) the IP response mediated by the ␣ 1B AR was inhibited by 45%. On the other hand, that mediated by the S394 -415A mutant, which was impaired in its ability to undergo PMA-induced phosphorylation, was inhibited by 15%. The individual or combined reintroduction of serines 394 or 400 in the S394 -415A mutant could restore the inhibitory effect of PMA on the receptor-mediated IP response (Fig. 9). On the other hand, the reintroduction of Ser 404 , Ser 408 , and Ser 410 did not confer any sensitivity of the receptor mutant to PMA (Fig. 9). The fact that the IP response mediated by the S394 -415A mutant was still impaired by 15% by PMA as compared with that of the T368 mutant is in agreement with FIG. 7. Effect of GRK2 on phosphorylation of the wild type and mutated ␣ 1B AR. A, COS-7 cells were transfected with the DNA encoding the wild type or mutated ␣ 1B AR alone (lanes 1, 2, 5, 6, 9, 10, 13, and 14) or in combination with the DNA encoding GRK2 (lanes 3, 4, 7, 8, 11, 12, 15, and 16). After labeling with 32 P i , cells were incubated for 15 min in the absence (B) or presence (E) of 10 Ϫ4 M epinephrine. The experimental conditions are as in Fig. 3. B, COS-7 cells expressing the wild type or mutated ␣ 1B AR alone or in combination with GRK2 were labeled with 32 P i and incubated for 15 min in the absence (basal) or presence of 10 Ϫ4 M epinephrine (EPI). Receptor numbers measured in membrane preparations were similar under the different conditions for all receptors ranging from 1 to 2 pmol/mg of protein (200 -300 fmol/million cells). The 32 P content of the phosphorylated receptor was quantified as indicated under "Experimental Procedures." Epinephrine-induced phosphorylation has been calculated by subtracting the value of basal phosphorylation for each receptor Results are the mean Ϯ S.E. of three independent experiments. Statistical significance was analyzed by a paired Stutent's t test. *p Ͻ 0.05 as compared with epinephrine-induced phosphorylation of the receptors expressed alone. the previous observation that the Ser394 -415A mutant displayed a small level of PMA-induced phosphorylation (Fig. 6). Altogether these findings provide strong evidence that Ser 394 and Ser 400 mediate a large portion of the PMA-dependent phosphorylation and desensitization of the ␣ 1B AR.
Effect of Different Mutations on the Basal Level of Phosphorylation of the ␣ 1B AR-In the absence of any stimulus, the ␣ 1B AR expressed in COS-7 cells, HEK293 (9), or Rat-1 cells (8) displays a basal level of phosphorylation which is approximately similar in the different cell systems expressing the recombinant receptor. Treatement of cells with the PKC inhibitor RO318220 did not significantly decrease the basal level of receptor phosphorylation (8). On the other hand, for several of the ␣ 1B AR mutants described in this study the basal level of phosphorylation was clearly lower than that of the wild type ␣ 1B AR (Fig. 6).
In general, mutations of one or two serines included in the sequence 393-413 resulted in a decrease of basal phosphorylation ranging from 13 to 30% as compared with the wild type ␣ 1B AR (Fig. 6). On the other hand, mutation of all the serines (S394 -415A mutant) (Fig. 6) or their removal in mutants T368, T387, and ⌬393-413 (Fig. 3) dramatically reduced the basal level of receptor phosphorylation. The reintroduction of any of the serines included in the sequence 393-413 into the S394 -415A mutant lacking all of them could restore basal phosphorylation to levels ranging from 20 to 64% that of the wild type ␣ 1B AR (Fig. 6).
Altogether, these findings indicate that in the ␣ 1B AR different serines of the C-tail might be phosphorylated under basal conditions. However, there seems to be no clear relationship between the role of distinct serines in maintaining basal phosphorylation and their involvement in agonist-or phorbol esterinduced phosphorylation of the receptor. In addition, whether the basal phosphorylation of the ␣ 1B AR has any functional correlate remains to be explored. In our experiments, the mutation of different serines did not result in any significant change in either basal or agonist-induced IP response of the mutated receptor as compared with the wild type under the experimental conditions used in this study (results not shown).

DISCUSSION
In this study we provide strong evidences that GRK and PKC can phosphorylate different subsets of serines in the C-tail of the ␣ 1B AR. Using site-directed mutagenesis, we have identified a first cluster of two serines (Ser 394 and Ser 400 ) phosphorylated following PMA-induced activation of PKC from a second cluster of three serines (Ser 404 , Ser 408 , and Ser 410 ) phosphorylated upon agonist-induced stimulation of the ␣ 1B AR. The phosphorylation of these two clusters of serines by PKC and GRK2, respectively, resulted in desensitization of the ␣ 1B AR.
Phosphorylation of Ser 404 , Ser 408 , and Ser 410 by GRK2 Is Involved in Agonist-induced Desensitization of the ␣ 1B AR-We recently demonstrated the involvement of GRK2 and GRK3 in agonist-induced phosphorylation and desensitization of the ␣ 1B AR (9). In this study, we have identified Ser 404 , Ser 408 , and Ser 410 as the phosphorylation sites involved in GRK2-induced desensitization (Fig. 6).
This was mainly demonstrated by two observations. First, receptor phosphorylation was reduced 36 and 64% by the mutation of Ser 404 alone and by that of Ser 408 and Ser 410 in combination, respectively (Fig. 6). Second, the reintroduction of all the three serines in a receptor mutant lacking all of them (S394 -415A mutant) could almost completely restore agonistinduced phosphorylation of the receptor. Since the extent of receptor phosphorylation correlated with the number of serines reintroduced, it seems that Ser 404 , Ser 408 , and Ser 410 can be phosphorylated independently following agonist stimulation of the receptor. Overexpression of GRK2 was able to specifically increase agonist-induced phosphorylation of those receptor mutants carrying Ser 404 , Ser 408 , and Ser 410 individually or in different combinations (Fig. 7). This argues for a nonsequential mechanism of receptor phosphorylation by GRK2.
The fact that Ser 404 , Ser 408 , and Ser 410 are in proximity of a pair of acidic amino acids (Fig. 2) is in agreement with the notion that both GRK1 and GRK2 are acidotropic kinases (29). Interestingly, rhodopsin and the ␤2AR possess a doublet of negatively charged residues located on the proximal side of the residue phosphorylated by GRK1 and GRK2, respectively. The N-formyl peptide receptor (for which the GRK2 phosphorylation sites have been deduced from in vitro studies using a fusion protein corresponding to the whole C-tail) as well as the ␣ 2A AR (for which the sites for GRK2 phosphorylation have been proposed on the basis of mutagenesis studies) also possess a doublet of acidic amino acids in proximity to the phosphorylated residues (17,18). Moreover, substitution of the two acidic residues with alanine seems to strongly reduce GRK2-induced phosphorylation of the N-formyl peptide receptor (18).
For rhodopsin (12), the N-formyl peptide receptor (18) as well as the C5a receptor (19), GRK-mediated phosphorylation seems to proceed sequentially. On the contrary, our findings suggest that GRK2 can phosphorylate Ser 404 , Ser 408 , and Ser 410 of the ␣ 1B AR independently (Fig. 6). Similar results were reported for the ␣ 2A AR, for which phosphorylation of four consecutive serines in the third intracellular loop seems to occur nonsequentially as demonstrated by the fact that mutation to alanine of any single serine reduced the extent of phosphorylation by about 25% (17). Thus, it appears that the sequential or nonsequential nature of GRK phosphorylation can vary depending on the receptor substrate and could be dictated by the structural or conformational properties of the phosphorylation site.
Our findings provide strong evidence that GRK2-mediated phosphorylation of Ser 404 , Ser 408 , and Ser 410 might trigger the desensitization of the ␣ 1B AR. This is supported by the fact that the combined mutation of these three serines (S394 -415A mutant) impaired the receptor's ability to undergo GRK2-mediated receptor desensitization (Fig. 8). On the other hand, the reintroduction of any of these serines in the S394 -415A mutant, lacking all of them, partially restored the GRK2-induced receptor desensitization (Fig. 8). Indeed, the presence of one of the three serines is sufficient to confer to the receptor some ability to undergo partial desensitization.
Phorbol Ester-induced Desensitization of the ␣ 1B AR Involves Multiple Mechanisms Including Phosphorylation of Ser 394 and Ser 400 -Our finding identify Ser 394 and Ser 400 as the main phosphorylation sites of the ␣ 1B AR for PKC. This was demonstrated by the observation that mutation of either Ser 394 or Ser 400 reduced PMA-induced receptor phosphorylation, whereas the reintroduction of Ser 394 and Ser 400 in the S394 -415A mutant, lacking both of them, could restore receptor phosphorylation to the levels observed in the wild type ␣ 1B AR (Fig. 6).
A large number of proteins have been shown to be phosphorylated by PKC (30). Among the sites phosphorylated by PKC, the most recurrent motif is (S/T)X(K/R) (30), where X indicates any amino acid. Of the two PMA phosphorylation sites identified in the C-tail of the ␣ 1B AR, Ser 400 is a classical phosphorylation site (SRK) for PKC (Fig. 2), while Ser 394 is included in the motif RXXSXXR, which is a less commun consensus site.
In addition to Ser 394 and Ser 400 , a third site seems to be phosphorylated upon stimulation of cells expressing the ␣ 1B AR with PMA. This is demonstrated by two findings. First, the Ser394 -415A mutant lacking both Ser 394 and Ser 400 retained a small level of PMA-induced phosphorylation, which was 24% that of the wild type ␣ 1B AR. Second, the truncated receptor T412, in which both Ser 394 and Ser 400 are conserved, displayed a 30% decrease of PMA-induced phosphorylation. However, using site-directed mutagenesis, we were unable to identify the amino acid residues involved in this residual PKC-mediated phosphorylation of the ␣ 1B AR.
The cellular response mediated by most G protein-coupled receptors, linked to the activation of phospholipase C, has been shown to desensitize rapidly following phorbol ester-induced activation of PKC. Moreover, for many receptors, including the ␣ 1B AR (8), the M3-mAchR (31), the receptor of thrombin (32), the CCK receptor (31), and the NK2 receptor (33), the activation of PKC by phorbol ester can increase receptor phosphorylation. However, a clear demonstration for a direct causal relationship between PKC-mediated phosphorylation and desensitization is still missing for many G protein-coupled receptors.
In this study we provide evidence that PKC-mediated desensitization of the ␣ 1B AR results, at least in part, from receptor phosphorylation. This was demonstrated by three main observations: (a) the individual mutation of both Ser 394 and Ser 400 partially impaired PMA-induced desensitization, (b) the reintroduction of either Ser 394 or Ser 400 in the S394 -415A mutant lacking both of them could restore PMA-induced receptor desensitization, and (c) the S394 -415A mutant still retained its ability to undergo PMA-induced phosphorylation and desensitization, even if at a smaller extent as compared with the wild type ␣ 1B AR. Similar evidence for a causal relationship between PKC-induced phosphorylation and desensitization has been provided for the 5-HT 1A (34) and NK2 (33) receptors.
However, our findings indicate that, in addition to receptor phosphorylation, other mechanisms might underlie PMA-induced desensitization of the ␣ 1B AR. Low concentrations of PMA (10 Ϫ8 M) could exclusively desensitize the wild type ␣ 1B AR-mediated IP response (Fig. 5A). On the other hand, concentrations higher then 10 Ϫ8 M could impair the IP response mediated by the T368 receptor, which could not undergo phosphorylation, as well as by Gq␣ upon its stimulation with AlF 4 Ϫ (Fig. 5A). Thus, PKC-mediated desensitization of G proteincoupled receptor linked to phospholipase C might involve phosphorylation of both receptors and other signaling proteins including Gq␣. Our findings are in agreement with recent work showing that PKC-mediated desensitization of the receptor for platelet-activating factor involves both phosphorylation of the receptor and that of the phospholipase C ␤3 (35).
Conclusions-Our study has several implications. First it provides strong evidence that both GRK2-and PKC-mediated desensitization of the ␣ 1B AR might result from the phosphorylation of distinct serines in the C-tail of the receptor. Second, it contributes to the elucidation of the consensus sites for the phosphorylation mediated by kinases of the GRK family. Finally, it might provide generalities about the biochemical mechanisms involved in homologous and heterologous regula-tion of G protein-coupled receptors linked to the activation of phospholipase C.