Phosphorylation of Ser 363 , Thr 370 , and Ser 375 Residues within the Carboxyl Tail Differentially Regulates m -Opioid Receptor Internalization*

Prolonged activation of opioid receptors leads to their phosphorylation, desensitization, internalization, and down-regulation. To elucidate the relationship between m -opioid receptor (MOR) phosphorylation and the regulation of receptor activity, a series of receptor mutants was constructed in which the 12 Ser/Thr residues of the COOH-terminal portion of the receptor were substituted to Ala, either individually or in combination. All these mutant constructs were stably expressed in human embryonic kidney 293 cells and exhibited similar expression levels and ligand binding properties. Among those 12 Ser/Thr residues, Ser 363 , Thr 370 , and Ser 375 have been identified as phosphorylation sites. In the absence of the agonist, a basal phosphorylation of Ser 363 and Thr 370 was observed, whereas [ D -Ala 2 ,Me-Phe 4 ,Gly 5 - ol ]en-kephalin (DAMGO)-induced receptor phosphorylation occurs at Thr 370 and Ser 375 residues. Furthermore, the role of these phosphorylation sites in regulating the internalization of MOR was investigated. The mutation of Ser 375 to Ala reduced the rate and extent of receptor internalization, whereas mutation of Ser 363 and Thr 370 to Ala accelerated MOR internalization kinetics. The present data show that the basal phosphorylation of MOR could play a role in modulating agonist-induced receptor internalization kinetics. Furthermore,

-opioid receptor (MOR) serves as the principle physiological target for most clinically important opioid analgesics, such as morphine and fentanyl (2,3). Although many opioid alkaloids exert their pharmacological effects via MOR, their binding affinity for the receptor and potency to activate the receptor do not always correspond to their abilities to induce tolerance (4 -8). This suggests that other cellular processes that modulate MOR responsiveness, such as receptor desensitization and internalization, may contribute to opioid tolerance and dependence. Like many other GPCRs, opioid receptors are regulated by agonist-dependent processes and undergo receptor phosphorylation, desensitization, internalization, and down-regulation (1). Interestingly, in addition to the subtype-specific regulation of opioid receptors (9 -12), individual opioid receptors are differentially regulated by distinct opioid agonists (5-8, 13, 14). In the case of MOR, opioid agonists demonstrating equivalent ability to activate receptor signaling exhibit remarkable differences in their abilities to functionally desensitize (5,6) and induce internalization of the receptor in both transfected cells and neurons (7,8,(13)(14)(15). However, the detailed molecular events underlying this differential regulation of MOR by distinct agonists remain unclear.
Using chimeric, truncated, or mutated opioid receptors, several studies reported the crucial role of the C-tail of opioid receptors in regulating their activities and trafficking (11, 16 -20). Whereas several potential phosphorylation sites were suggested to be involved in regulating the activity and trafficking of opioid receptors, the actual phosphorylation of these residues was not demonstrated. Only recently, Maestri-El Kouhen et al. (21) have reported the identification of phosphorylation sites of the ␦-opioid receptor (DOR) in human embryonic kidney (HEK293) cells. These results show the implication of precise phosphorylation sites (Thr 358 and Ser 363 ) in regulating the internalization and desensitization of DOR (21). Additionally, agonist-induced internalization of opioid receptors seems to be cell type-dependent (22), indicating a recruitment of specific cellular factors in regulating opioid receptor trafficking.
From our own studies and other reports, agonist-induced opioid receptor phosphorylation has been shown to be time-and dose-dependent and could be blocked by antagonist naloxone (23)(24)(25). We have previously reported that rapid phosphorylation of MOR does not directly correlate with the relatively slow rate of receptor desensitization as measured by the loss of inhibition of adenylyl cyclase activity (23). We recently reconciled that such a difference could be attributed to the expres-sion level of the receptor as well as the rapid recycling and resensitization of MOR after etorphine treatment (26). By limiting the recycling of the receptor and receptor density on the cell surface, rapid desensitization of MOR was observed (26). Whether MOR phosphorylation plays a role in these cellular events needs to be established unequivocally. Therefore, it is imperative to understand the phosphorylation mechanisms and identify specific phosphorylation sites and their involvement in the cellular regulation of MOR.
In the present work, we examine the role of phosphorylation of individual Ser/Thr residues in agonist-induced MOR internalization. Our mutational analysis indicates that Ser 363 , Thr 370 , and Ser 375 residues are phosphorylation sites, and we clearly show that these sites are differentially phosphorylated in basal and agonist-induced conditions. Subsequently, we provided evidence that phosphorylation of these specific Ser/Thr residues differentially regulates agonist-induced MOR internalization. This study reveals that agonist-induced internalization of MOR is modulated by site-specific phosphorylation.  (27) subcloned into the expression vector pcDNA 3 (Invitrogen) was used to generate most point mutations. Additional point mutations, alone or in combination, were constructed in multiple steps using the QuickChange TM site-directed mutagenesis method as outlined by Stratagene (La Jolla, CA). Forward and reverse oligonucleotides with the desired mutation and a designed endonuclease site were synthesized. The nucleotide sequences of all mutants were confirmed by dideoxynucleotide termination reactions using Sequenase II.

Materials-Oligonucleotides
Stable Transfection of HEK293 Cells with Wild Type and Mutant -Opioid Receptors-HEK293 cells were cultured in minimal essential medium supplemented with 10% fetal bovine serum, 100 g/ml streptomycin, and 100 IU/ml penicillin under humidified atmosphere at 5% CO 2 . Cells were stably transfected by the Ser/Thr mutants or the wild type rat -opioid receptor-1 cDNAs using the calcium phosphate precipitation method (28). To avoid any position effect because of random integration of cDNAs into the chromosomes, colonies of transfected HEK293 cells surviving 1 mg/ml Geneticin (G-418) selection were pooled. Confirmation of -opioid receptor expression was determined by whole cell binding assay using [ 3 H]diprenorphine in 25 mM HEPES buffer, 5 mM MgCl 2 , pH 7.6. Specific binding is defined as the difference between the radioactivity bound to the cells in the presence and absence of naloxone (100 M).
Receptor Phosphorylation-Phosphorylation of HEK293 cells stably expressing the Ser/Thr mutant or wild type -opioid receptors was determined as described previously (23). Cultured cells in 6-well plates were washed with phosphate-free Dulbecco's modified Eagle's medium and incubated with 100 Ci/ml [ 32 P]orthophosphate for 2 h at 37°C in 10% CO 2 . 5 M DAMGO was added as indicated. The reactions were terminated on ice at which point the cells were washed with ice-cold phosphate-buffered saline and subsequently lysed in lysis buffer (25 mM HEPES, pH 7.4, 1%,v/v, Triton X-100, 5 mM EDTA with 100 g/ml bacitracin, 10 g/ml leupeptin, 0.1 mM phenylmethylsulfonyl fluoride, 100 g/ml soybean trypsin inhibitor, 10 g/ml pepstatin A, and 20 g/ml benzamidine as protease inhibitors and with 50 mM sodium fluoride, 10 mM sodium pyrophosphate, and 0.1 mM sodium vanadate as phosphatase inhibitors). After solubilization, insoluble debris was removed by centrifugation at 14,000 ϫ g for 15 min at 4°C. Supernatants were eluted to twice the volume with lysis buffer (without Triton X-100) before loading onto 1-ml wheat germ lectin affinity columns pre-equilibrated with buffer A (25 mM HEPES, pH 7.4, 100 mM NaCl, and 0.1 Triton X-100). The columns were then washed with 10 ml of buffer A to remove nonbound radioactive proteins, and the MOR-containing fraction was eluted with 3 ml of buffer A containing 0.5 M N-acetylglucosamine and the protease/phosphatase inhibitors as indicated above. These purified samples were incubated in the presence of 3 g of rat hemagglutinin-monoclonal antibody 3F10 (Roche Molecular Biochemicals) and prewashed immunopure protein G-agarose beads (Pierce) overnight at 4°C. The final samples were eluted from the agarose beads and separated on a 10% SDS-PAGE (23). After electrophoresis, the gels were dried, and phosphorylated bands were visualized and quantified using the PhosphorImager Storm 840 system (Molecular Dynamics, Sunnyvale, CA). Phosphorylation levels obtained for mutant and wild type receptors were normalized according to protein amount as well as receptor expression level within the starting cell materials.
Opioid Receptor Binding-Characterization of the -opioid receptor binding sites in HEK293 cells expressing wild type and mutated receptors was carried out with membranes (100,000 ϫ g ϫ 60-min membrane preparations minus nuclei) in 25 mM HEPES buffer, pH 7.6, containing 5 mM MgCl 2 at 24°C for 90 min as described previously (29). The K d and B max values for [ 3 H]diprenorphine were determined for each mutant by saturation binding. IC 50 and K i values for DAMGO were determined by competition binding studies using increasing concentrations of DAMGO to compete for the binding of 1 nM [ 3 H]diprenorphine. Data from saturation and competition binding were fitted by nonlinear curve fitting using the data analysis program GraphPad Prism.
Cyanogen Bromide Cleavage of -Opioid Receptor-Radiolabeled -opioid receptor bands were excised from the gel, and receptor protein was eluted in Tris-glycine buffer (25 mM Tris, 192 mM glycine, and 0.1% w/v SDS) at 4°C. Receptor fractions were first dialyzed and lyophilized and then reduced and alkylated before treatment in the presence of CNBr (160 mg/ml in 70% formic acid) (30). CNBr cleavage products were separated on a Tricine-SDS-PAGE (16.5% acrylamide, 6% bisacrylamide) (31), and phosphopeptide bands were visualized by PhosphorImager.
Quantitative Analysis of -Opioid Receptor Internalization by Fluorescence Flow Cytometry-A pool of stably transfected HEK293 cells expressing wild type or mutant hemagglutinin epitope-tagged receptors was incubated in the presence of 5 M DAMGO for the indicated time intervals. Cells were then chilled on ice to stop membrane trafficking, and all subsequent incubations were performed at 4°C. Residual cell surface receptors were monitored by incubating cells in the presence of monoclonal anti-hemagglutinin antibody (clone 16B12) (Babco, Richmond, CA) for 2 h followed by an incubation with the Alexa 488conjugated anti-IgG antibodies (Molecular Probes, Inc., Eugene, OR) for 1 h. Surface receptor staining intensity of antibody-labeled cells was analyzed using fluorescence flow cytometry (FACScan, Becton Dickinson, Palo Alto, CA). Each data point represents the average of mean fluorescence values Ϯ S.E. of receptor internalization determined in an analysis of 10,000 cells performed in triplicate.

RESULTS
Previously, we reported that DAMGO-induced MOR phosphorylation occurs rapidly with maximum phosphorylation obtained after 20 min of agonist treatment (23). To localize the intracellular domains containing the phosphorylation sites, phosphorylated MOR was purified and chemically cleaved in the presence of cyanogen bromide as described under "Experimental Procedures." The theoretical CNBr cleavage after Met residues generates several fragments with different sizes including the C-tail domain of the receptor (Fig. 1A). To perform total protein digestion, phosphorylated receptors were incubated with CNBr for up to 24 h. After CNBr cleavage, samples containing generated peptides were resolved on 16.5% Tricine-SDS-PAGE, and phosphorylated bands were visualized using a PhosphorImager (Fig. 1B). Only one phosphorylated band migrating at ϳ13 kDa, a size corresponding to the C-tail domain of MOR, could be detected (Fig. 1B). This result strongly indicated that DAMGO-induced phosphorylation occurs only within the cytoplasmic tail of MOR. A direct sequencing of the purified C-tail of MOR to determine the exact phosphorylation sites was not feasible because of the insufficient amount of purified peptide. Therefore, we used a site-directed mutagenesis approach to identify phosphorylated Ser/Thr residues within the C-tail of MOR.
There are 12 Ser/Thr residues within the C-tail of rat MOR that are potential phosphorylation sites. To identify which Ser/Thr residues are phosphorylated, we constructed a number of mutated receptors in which these residues were substituted to Ala, either individually or in combination. Each of these mutated receptors as well as wild type MOR were stably expressed in HEK293 cells. To avoid any phenotype effect as a result of random integration of the plasmid into the chromosomes, a mixture of cell clones stably expressing either receptor was used in the present study as described under "Experimental Procedures." Receptor expression levels (B max ) obtained were high for all mutated and wild type receptors ranging from 0.9 to 6.9 pmol/mg protein (Table I). K d values of [ 3 H]diprenorphine were similar among all these receptors, whereas small variations in the affinities of certain mutated receptors for DAMGO were observed (Table I).
To define the region within the C-tail domain of MOR containing the phosphorylation sites, we first used a group of mutated receptors in which we progressively substituted clusters of Ser/Thr residues to Ala (Fig. 2C). After 20 min of treatment in the presence of a saturating concentration of DAMGO (5 M), phosphorylated receptors were purified and resolved on 10% SDS-PAGE as described under "Experimental Procedures" ( Fig. 2A). Phosphorylated bands were quantified and normalized according to total protein amount as well as receptor expression levels measured in the presence of [ 3 H]diprenorphine. As shown in Fig. 2, A and B, the substitution of the first four Ser/Thr residues (Thr 354 -Thr 357 ) to Ala (mutant 354 -357) did not affect agonist-induced MOR phosphorylation when compared with that of wild type receptor. Substitution of the six residues Thr 354 -Thr 357 , Ser 363 , and Thr 364 to Ala (mutant 354 -364) reduced the level of phosphorylation to approximately 60% of wild type receptor. Phosphorylation of mutant 354 -376, in which the nine residues Thr 354 -Thr 357 , Ser 363 , Thr 364 , Thr 370 , Ser 375 , and Thr 376 were replaced by Ala, was completely abolished. As expected, no phosphorylation signal could be detected with mutant 354 -394 in which all 12 Ser/Thr residues of the C-tail were changed to Ala (Fig. 2). These results clearly support our finding from CNBr experiments and show that phosphorylation occurs only at the C-tail of MOR. In addition, these results indicate that the phosphorylation sites are localized within the region Ser 363 -Thr 376 .
With the next group of mutated receptors, we focused on the region between Ser 363 and Thr 376 by substituting the five potential phosphorylation sites (Ser 363 , Thr 364 , Thr 370 , Ser 375 , and Thr 376 ) to Ala, either individually or in combination (Fig.  3C). The phosphorylation level of mutant 363-376, in which all of the five potential sites were replaced by Ala, was nearly abolished (Fig. 3, A and B). However, when compared with mutant 354 -376 that completely abolishes receptor phosphorylation (Fig. 2), mutant 363-376 still presents a weak phosphorylation signal (Ͻ5% of maximal phosphorylation). This slight difference in phosphorylation signals between these two mutants could be because of a minor phosphorylation site present outside of the Ser 363 -Thr 376 region. Nevertheless, this result indicates that major phosphorylation sites are located within the Ser 363 -Thr 376 region. Mutation of the doublet Ser 363 -Thr 364 to Ala (mutant 363/ 364) significantly reduced phosphorylation level of the receptor to approximately 60%. The single mutation S363A (mutant 363) also reduced the level of phosphorylation, but the T364A mutation (mutant 364) did not, suggesting that Ser 363 is the only phosphorylation site within this doublet (Fig. 3). Similarly, the substitution of Ser 375 -Thr 376 (mutant 375/376) reduced the phosphorylation level to approximately 60%. Only the S375A (mutant 375) but not the T376A (mutant 376) mutation reduced the phosphorylation level of the receptor (Fig.  3). This indicates that Ser 375 is also a phosphorylation site. Finally, the substitution of Thr 370 into Ala (mutant 370) reduced the phosphorylation level to approximately 70%, suggesting that this residue is a phosphorylation site as well (Fig.  3). Taken together, these results indicate that among the 12 potential phosphorylation sites present in the C-tail of MOR, three residues, Ser 363 , Thr 370 , and Ser 375 , are actually involved in MOR phosphorylation.
The possibility that the observed reduction in phosphorylation level with these mutated receptors could be a result of conformational changes of the C-tail domain or a modification of a kinase recognition motif cannot be ruled out. Therefore, to  visualize the phosphorylation of each of these three residues (Ser 363 , Thr 370 , and Ser 375 ), we constructed a new set of mutated receptors in which we replaced two of these three Ser/Thr residues with Ala, leaving the third residue wild type (Fig. 4B). Thus, we were able to analyze both qualitatively and quantitatively the phosphorylation of those three sites in the absence (basal) and in the presence of DAMGO. Substantial basal phosphorylation of MOR in the absence of agonist activation has been reported by several groups including ours (13,23,24). As shown in Fig. 4, substantial basal phosphorylation of wild type receptor in the absence of agonist was observed. This basal phosphorylation was increased to approximately 3-fold in the presence of DAMGO (Fig. 4). Phosphorylation of mutant 363/370/375, in which the three phosphorylation sites were mutated to Ala, is completely abolished in the basal condition, whereas a weak phosphorylation signal can still be observed in the presence of DAMGO (Fig. 4, lanes 4  and 9). This remaining signal is not significant and is comparable with that observed with mutant 363-376 (Fig. 3), indicating the presence of a minor phosphorylation site outside of the Ser 363 -Thr 376 region. However, the majority of DAMGOinduced phosphorylation signal (Ͼ95%) is abolished with this mutant (363/370/375). Interestingly, when Ser 363 is left wild type (mutant 370/375), this mutant shows a significant basal phosphorylation, and this observed pattern of phosphorylation remains identical to that obtained in the presence of DAMGO. Thr 370 (mutant 363/375 leaving Thr 370 wild type) was weakly phosphorylated in basal conditions, and this phosphorylation pattern was increased significantly to approximately 2-fold in the presence of agonist. Finally, in the case of Ser 375 (mutant 363/370 leaving Ser 375 wild type), no basal phosphorylation could be detected, although this site is strongly phosphorylated in the presence of DAMGO (Fig. 4). The levels of DAMGOinduced phosphorylation observed with all these mutants were dramatically reduced in the presence of the antagonist naloxone (10 M) to comparable levels with these obtained in basal conditions, whereas incubation with naloxone alone had no effect on basal phosphorylation itself (data not shown). Taken together, these results confirm that Ser 363 , Thr 370 , and Ser 375 are phosphorylated, and they indicate that these residues are phosphorylated in an independent manner. Additionally, our data clearly point out that those three sites are phosphorylated differentially: Ser 363 is phosphorylated only in the basal condition, Thr 370 is phosphorylated both in the absence and in the presence of DAMGO, and Ser 375 is phosphorylated only in the presence of DAMGO (Fig. 4). Therefore, it is tempting to propose that these three sites may participate to different extents in regulating MOR activity.
Previously, we reported that the rapid receptor phosphorylation (min) does not directly correlate with the relatively slow rate of desensitization (h) of MOR (23). Subsequently, we demonstrated that the relatively slow rate of desensitization was because of rapid recycling and resensitization of MOR (26). The rapid desensitization of MOR was observed only when the receptor expression level was low and receptor recycling was blocked. This study suggests that MOR internalization participates in receptor desensitization. Several studies have suggested a role of agonist-induced phosphorylation in opioid re- ceptor internalization (8,14,32,33). Therefore, it is important to clearly delineate the role of phosphorylation in regulating the loss of cell surface receptors. In this regard, it is important to examine the potential role of each individual phosphorylated Ser/Thr in regulating MOR internalization.
DAMGO-induced loss of cell surface wild type or mutated receptors was monitored using flow cytometry as described under "Experimental Procedures." Kinetic analysis showed that MOR internalizes rapidly in response to DAMGO activation (t1 ⁄2 ϳ16 min) and that approximately 50% of the total cell surface receptors were internalized after 1 h of treatment (Fig.  5). The S375A mutation (mutant 375) significantly reduced the rate and extent of DAMGO-induced internalization of receptor compared with wild type MOR. In this case, only approximately 20% of the receptors were internalized after 1 h of DAMGO activation, p Ͻ 0.001 (Fig. 5B). However, internalization kinetics of the T370A mutant (t1 ⁄2 ϳ18 min, extent internalization ϳ50%) were similar to those observed for the wild type receptor (Fig. 5A). Interestingly, the S363A mutation accelerated the rate (t1 ⁄2 ϳ10 min, p Ͻ 0.01) but did not increase the extent of DAMGO-induced receptor internalization (Fig. 5A). As a mixture of HEK293 clones stably expressing wild type or mutated receptors was used for this study, the observed differences in receptor internalization among those mutated receptors could not be attributed to a phenotype effect. Thus, these results clearly show that phosphorylation plays an important role in controlling MOR internalization. Additionally, it appears clearly that phosphorylation of these three sites contributes differentially in regulating MOR internalization.
We next examined the effects of mutating either two or three of those phosphorylation sites to Ala (Fig. 4B) on receptor internalization. Mutation of both Ser 363 and Thr 370 (mutant 363/370) further increases the rate and extent of internalization compared with that of single mutant S363A. This result suggests that the phosphorylation of both residues (Ser 363 and Thr 370 ) plays a role in attenuating receptor internalization (Fig. 5A). However, all mutants containing S375A mutation (mutants 363/375, 370/375, and 363/370/375) showed a reduced rate and extent of MOR internalization when compared with wild type receptor (Fig. 5B). These results show clearly that the phosphorylation of Ser 375 plays a critical role in promoting the internalization of activated MOR. It should be noted that internalization kinetics observed in the case of mutants 363/375, 370/375, and 363/370/375 were slightly faster than those of the S375A single mutant (Fig. 5B). This finding confirms that phosphorylation of Ser 363 and Thr 370 has an opposite effect to that of the Ser 375 mutation in regulating MOR internalization. Additionally, 363/370/375 mutations did not completely block the internalization of MOR (Fig. 5B), indicating that other signals or motifs could participate in these cellular events.
To avoid any differential rate of recycling of mutant receptors, which may interfere with measured internalization kinetics, a receptor recycling blocking agent monensin was included in this study to monitor only internalized receptors. A 1-h pretreatment with monensin (50 M) prior to the addition of DAMGO had no significant effect on the internalization kinetics of mutant or wild type -opioid receptors in HEK293 cells (data not shown). Thus, in our present study, measurement of DAMGO-induced loss of cell surface receptors is mainly because of the internalization phenomena. Our results clearly show differential involvement of phosphorylation sites in regulating -opioid receptor internalization.

DISCUSSION
The present study represents a significant step toward understanding the molecular events associated with the regulation of MOR. We first demonstrated that of the 12 Ser/Thr residues present within the C-tail of MOR, 3 residues (Ser 363 , Thr 370 , and Ser 375 ) are phosphorylated. The role of these specific phosphorylated sites in receptor internalization was delineated. Systematic studies of the corresponding Ser/Thr mutants revealed differential regulation of MOR internalization by phosphorylation.
Several studies have reported agonist-induced phosphorylation of opioid receptors, which has been shown to be time-dependent and agonist concentration-dependent in different cell systems (23)(24)(25)33). Whereas all these studies suggest that phosphorylation plays a role in opioid receptor regulation, i.e. desensitization, internalization, and down-regulation, molecular events related to receptor phosphorylation/regulation are still poorly understood. In addition, differential phosphorylation/regulation of opioid receptors by opioid agonists has been observed (6,11,14,33). These reports suggest that different agonist-activated receptor conformations may exist and that specific phosphorylation site(s) may participate in regulating receptor activity and trafficking. Thus, several Ser/Thr residues within the COOH terminus of opioid receptors have been suggested to be involved in the regulation of these receptors. Thr 394 , Ser 353 , and Ser 369 have been shown to be important for the regulation of -opioid receptors (20,34,35), ␦-opioid receptors (18,19), and -opioid receptors (36), respectively, although phosphorylation of these residues has not been determined. Recently, a T394A mutation was shown to significantly block DAMGO-induced MOR phosphorylation to approximately 10% of the wild type receptor and impair receptor desensitization in Chinese hamster ovary cells (37), indicating that this residue is a phosphorylation site. However, in HEK293 cells, the mutation of Thr 394 to Ala reduced the DAMGO-induced phosphorylation level to approximately 80% of maximal phosphorylation obtained with wild type receptor (data not shown). This finding clearly shows that major phosphorylation sites of MOR expressed in HEK293 cells are located in residues other than Thr 394 . DAMGO-induced phosphorylation of mutant 354 -376 (leaving Thr 394 wild type) is completely abolished (Fig. 2). These results clearly indicate that DAMGO does not induce phosphorylation of Thr 394 in HEK293 cells, but this site might participate in the agonist-induced phosphorylation of MOR as a recognition binding motif for a kinase or a regulatory protein.
Taken together, these results suggest that cell type-specific differences may exist in the agonist-induced receptor phosphorylation. Our current mutational analysis indicated that at least in HEK293 cells Thr 370 and Ser 375 are the residues phosphorylated in the presence of agonist, whereas Thr 394 is the residue that is phosphorylated in Chinese hamster ovary cells (37). Considering that the T394A mutation did not completely block agonist-induced MOR phosphorylation in Chinese hamster ovary cells (37), other sites involved in receptor phosphorylation in this cell line need to be identified.
Several kinases have been postulated to be involved in opioid receptor phosphorylation including protein kinase A (38,39), protein kinase C (33,40), GRKs (25,34), mitogen-activated protein kinase (41,42), calmodulin kinase II (43,44), and tyrosine kinase (32). Recently, our group has reported a [D-Pen 2 -D-Pen 5 ]enkephalin-induced hierarchical phosphorylation mechanism of DOR in HEK293 cells (21). Ser 363 has been shown to be the primary agonist-induced phosphorylation site allowing subsequent phosphorylation of Thr 358 . Thr 361 was shown to be involved in regulating phosphorylation of DOR without being phosphorylated. In addition, mutation of Thr 361 to Ala impairs phosphorylation of Thr 358 , suggesting that Thr 361 could participate in a recognition binding motif for a kinase or regulatory protein (21). Although MOR and DOR have differences in their C-tail primary sequence, it is interesting to note that a common motif encompassing agonistinduced phosphorylated sites exists for both receptor subtypes. The organization of motif ( 357 VTACTPSD 364 ) containing Thr 358 and Ser 363 of DOR is analogous to the motif ( 369 NTRE-HPST 376 ) containing Thr 370 and Ser 375 of MOR. The presence of acidic or charged amino acids surrounding those agonistinduced phosphorylated sites suggests that they are potential GRK-phosphorylation sites. Furthermore, a conserved Pro residue directly present upstream of Ser 363 and Ser 375 in DOR and MOR, respectively, may confer a specific spatial conformation of the C-tail, and these phosphorylation sites may be accessible to specific kinase(s). Whether the same kinase(s) is involved in the phosphorylation of both receptor subtypes remains to be demonstrated. Clearly, the two sites being phosphorylated in the presence of the agonist do not represent a mitogen-acti-vated protein kinase site. Several groups including our own have reported basal phosphorylation of opioid receptors in the absence of agonist (13,23,24,33), suggesting a correlation between this basal phosphorylation and the observed basal activity of these receptors. Our present results show that Ser 363 is specifically phosphorylated in the absence of DAMGO treatment. Thr 370 appears to be phosphorylated in both basal and agonist-induced conditions. However, Ser 375 is phosphorylated only in the presence of DAMGO. The present identification of phosphorylation sites in MOR will greatly facilitate determination of the kinase responsible for phosphorylation of each specific site as well as associated protein(s) involved in regulating receptor activities.
Previous reports show no effects of mutated or truncated COOH termini of opioid receptors on desensitization and/or internalization of these receptors (22,36,45,46). These studies raise the fundamental question of whether or not phosphorylation is actually required for agonist-induced desensitization and internalization of opioid receptors. The present identification of MOR phosphorylation sites enabled us not only to dissect the complex phosphorylation mechanisms but also to investigate the role of each individual phosphorylation site in receptor internalization. For the first time, we show that the rate and extent of MOR internalization are differentially modulated by site-specific phosphorylation. Phosphorylation of Ser 375 plays an important role in promoting the internalization of activated receptor, whereas phosphorylation of Ser 363 and Thr 370 may synergistically attenuate this receptor internalization. Segredo et al. (14) have shown that truncation of -opioid receptor at positions 363 and 354 differentially regulates DAMGO-induced receptor internalization in HEK293 cells. This study suggests that phosphorylation sites located in the C-tail domain play distinct roles both in promoting agonistinduced endocytosis of activated receptors and preventing receptor endocytosis in the absence of the agonist (14). The same study shows that a truncated mutant at position 354 internalized constitutively and that receptor internalization is enhanced in presence of DAMGO (14). Similarly, Trapaidze et al. (46) have shown that a truncated MOR lacking Thr 394 continued to internalize in response to DAMGO but not the mutant lacking the last COOH-terminal 24 amino acids. Consistent with our findings, these results indicate the presence of positive and negative endocytotic signals within the C-tail of MOR.
In addition, the current study provides evidence that a single phosphorylated Ser/Thr mutation could differentially alter MOR internalization. Although the vast majority of endocytotic signals described previously are positive signals, the presence of endocytotic negative signals within the C-tail of the receptors has been reported for several GPCRs including somatostatin (47), parathyroid hormone/parathyroid hormone-related receptors (48), and luteinizing hormone/human choriogonadotropin (49) receptors. Thus, these results and our findings indicate that receptor-mediated endocytosis is a highly complex and regulated process with multiple receptor domains playing facilitatory as well as inhibitory roles.
As for many GPCRs, opioid receptors are endocytosed in a dynamin-dependent manner via clathrin-coated pits (9,33,36,40). Our findings suggest that at least two mechanisms of phosphorylation may exist and contribute differentially to MOR endocytosis in HEK293 cells. The first mechanism of phosphorylation is agonist-induced GRK-mediated phosphorylation of Ser 375 , and maybe Thr 370 , which involves the binding of arrestin and subsequent internalization of MOR. This is consistent with a recent report that blocking DAMGO-induced phosphorylation but not basal phosphorylation of MOR in HEK293 cells prevents the receptor from being internalized (42). The second mechanism is the phosphorylation of Ser 363 , and maybe Thr 370 , in the absence of agonist that may involve kinase(s) other than GRKs and could play a regulatory role in modulating basal or constitutive MOR internalization. As shown by Segredo et al. (14), both constitutive and agonistinduced internalization of MOR could be mediated by the clathrin-coated pit pathway. Receptor recycling to the cell surface after DAMGO treatment was negligible in the current study (data not shown) in contrast to the substantial receptor recycling after etorphine treatment as reported previously (26). This finding reflects the differential regulation of MOR by opioid agonists including desensitization, internalization, and down-regulation of the receptors (5,6,8,50). Structurally, distinct domains of the rat MOR were shown to be involved in the binding of receptor selective alkaloids and peptides (51,52). In addition, the level of overall MOR phosphorylation was reported to be agonist-dependent (6,33). Taken together, these findings strongly indicate that different agonist-activated receptor conformations exist and that receptor phosphorylation plays an important role in the structure-function of this receptor. Thus, identification of phosphorylation sites induced by other agonists will help considerably to delineate molecular mechanisms controlling the structure-function of MOR.
In conclusion, the present studies have brought new insights into the complex function of MOR phosphorylation. We have identified three phosphorylated sites within the C-tail of MOR and delineated their role in regulating receptor endocytosis. Interestingly, we have found that phosphorylation of Ser 375 is critical for DAMGO-induced MOR internalization. However, phosphorylation of Ser 363 and Thr 370 represents a negative regulatory signal in this internalization process. Whether these regulation mechanisms are conserved among other members of the GPCR family remains to be demonstrated. Nevertheless, the present work makes a significant step toward the elucidation of molecular events related to opioid receptor regulation and thus to opioid tolerance and dependence.