Proteasome Involvement in Agonist-induced Down-regulation of μ and δ Opioid Receptors*

This study investigated the mechanism of agonist-induced opioid receptor down-regulation. Incubation of HEK 293 cells expressing FLAG-tagged δ and μ receptors with agonists caused a time-dependent decrease in opioid receptor levels assayed by immunoblotting. Pulse-chase experiments using [35S]methionine metabolic labeling indicated that the turnover rate of δ receptors was accelerated 5-fold following agonist stimulation. Inactivation of functional Gi and Go proteins by pertussis toxin-attenuated down-regulation of the μ opioid receptor, while down-regulation of the δ opioid receptor was unaffected. Pretreatment of cells with inhibitors of lysosomal proteases, calpain, and caspases had little effect on μ and δ opioid receptor down-regulation. In marked contrast, pretreatment with proteasome inhibitors attenuated agonist-induced μ and δ receptor down-regulation. In addition, incubation of cells with proteasome inhibitors in the absence of agonists increased steady-state μ and δ opioid receptor levels. Immunoprecipitation of μ and δ opioid receptors followed by immunoblotting with ubiquitin antibodies suggested that preincubation with proteasome inhibitors promoted accumulation of polyubiquitinated receptors. These data provide evidence that the ubiquitin/proteasome pathway plays a role in agonist-induced down-regulation and basal turnover of opioid receptors.

amino termini located on the extracellular side of the plasma membrane that are linked to seven-transmembrane helices connected by relatively short intracellular and extracellular loops, and contain carboxyl termini that face the interior of the cell. Ligands approach and engage GPCRs from the extracellular space, and receptor activation results in coupling to heterotrimeric G proteins on the intracellular face of membrane. The amino termini of nearly all G protein-coupled receptors contain consensus amino acid sequences for asparagine-linked glycosylation; two sites for N-linked glycosylation are present in ␦ and receptors, while five are found in the receptor (3).
Three types of opioid receptor, ␦, , and , have been cloned and characterized extensively (4 -7). Opioid receptors have unique ligand specificities, anatomical distributions, and physiological functions (8 -15). Opioid receptors exhibit ϳ60% identity in their amino acid sequences, however, marked differences in sequence conservation are evident within receptor subdomains. The amino acid sequences of putative transmembrane spanning segments and the three intracellular loops are highly conserved among opioid receptor types, whereas sequences in the extracellular amino termini, second and third extracellular loops, and the intracellular carboxyl termini are considerably more divergent (1). Upon activation, opioid receptors interact with multiple G proteins to regulate adenylyl cyclase, the MAP kinase pathway, phosphatidylinositol 3-kinase, Ca 2ϩ channels, and K ϩ channels (16 -20).
Short-term exposure of opioid receptors to agonists, like most GPCRs, leads to receptor desensitization, while chronic exposure leads to receptor down-regulation (reviewed in Refs. 1 and 21). Homologous desensitization appears to involve phosphorylation of the agonist-activated GPCR by members of the G protein-coupled receptor kinase family, and subsequent binding of arrestin proteins, which effectively uncouple the interaction of activated GPCRs from heterotrimeric G proteins (21). In addition to preventing receptor/G protein interactions, arrestins initiate receptor endocytosis via clathrin-coated pits (22,23). Endosome-associated receptors can be resensitized by protein phosphatases and recycled back to the plasma membrane, or be degraded intracellularly. The mechanism for GPCR proteolysis has been generally assumed to involve fusion of endosomes with lysosomes, based on studies of epidermal growth factor receptor down-regulation (24).
Acute agonist-induced desensitization of opioid receptors appears to involve similar mechanisms as described for the ␤ 2adrenergic receptor. It has been shown that ␦ and opioid receptors are phosphorylated by G protein-coupled receptor kinases following agonist treatment (25,26), and are endocytosed in a dynamin-dependent process via clathrin-coated pits (27). Overexpression of arrestin or G protein-coupled receptor kinase leads to enhanced agonist-induced internalization (26,28). ␦ Opioid receptors do not appear to recycle to the plasma membrane efficiently, while coexpressed ␤ 2 -adrenergic recep-tors do (29), and following treatment of cells for several hours with agonist, ␦ opioid receptors are apparently associated with lysosomes (30).
This study was designed to test the hypothesis that lysosomal proteases are responsible for opioid receptor down-regulation. Contrary to this hypothesis, these data implicate a prominent role for the ubiquitin/proteasome pathway in agonistinduced down-regulation and basal turnover of opioid receptors.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-Human embryonic kidney (HEK) 293 cells were cultured at 37°C in a humidified atmosphere containing 5% CO 2 in Dulbecco's modified Eagle's medium, supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin sulfate. HEK 293 cells were transfected using electroporation, as described previously (31), with expression plasmids encoding murine or ␦ opioid receptors tagged at the amino termini with the FLAG epitope (kindly provided by Dr. M. Van Zastrow). Cells stably expressing epitope-tagged opioid receptors were selected in media containing 1 mg/ml G418 (Life Technologies, Gaithersburg, MD), as described (32,33).
Membrane Preparation and Radioligand Binding Assays-HEK 293 cells expressing epitope-tagged or ␦ opioid receptors were grown to near confluence in 100-mm diameter dishes. For membrane preparations, the culture medium was aspirated and cells were harvested using a cell scraper in 5 ml of 50 mM Tris-HCl buffer, pH 7.5, per 100-mm dish. The cell suspension was homogenized with a Tekmar tissuemizer (Cincinnati, OH), then centrifuged at 35,000 ϫ g for 20 min. The membrane pellet was washed 3 times in Tris buffer and then resuspended by homogenization in 1 ml of 0.32 M sucrose, 50 mM Tris-HCl, pH 7.5, per dish, and the crude membrane preparation was stored at Ϫ80°C.
Opioid receptor binding assays were conducted in duplicate on membrane preparations diluted 20 -40- Rockville, MD). Binding assays were conducted at 22°C in a volume of 0.25 ml (ϳ20-40 g of protein/ml); 10 M cyclazocine was used to define specific binding. Following a 1-h incubation, assays were terminated by filtration through Whatman GF/B filters. Filters were soaked in Ecoscint liquid scintillation mixture (National Diagnostics, Somerville, NJ) prior to determination of filter bound radioactivity using a Beckman LS 1701 scintillation counter. Receptor binding data was analyzed by nonlinear regression using Prism 2.0 (GraphPad Software, San Diego, CA). Protein concentrations were determined with the Bio-Rad protein assay (Hercules, CA), using bovine serum albumin as the standard.
Immunoblotting and Immunoprecipitation-For immunoblotting experiments, cells were grown to near confluence in 60-mm dishes. Cell extracts were prepared by incubating the cells in 0.2 ml of lysis buffer consisting of 150 mM Tris-HCl, pH 7.5, 300 mM NaCl, 1 mM MgCl 2 , 1 mM CaCl 2 , 1% Triton X-100, 10% glycerol and protease inhibitor mixture (containing 4-(2-aminoethyl)benzenesulfonyl fluoride, pepstatin A, E-64, bestatin, leupeptin, and aprotinin, Sigma) for 1 h on ice. Cell debris was pelleted by centrifugation and the supernatant was used for further analysis. Protein concentrations in supernatants were determined using the Bio-Rad assay with bovine serum albumin as standard. Cell extracts containing ϳ20 g of protein were mixed with an equal volume of gel loading buffer (Bio-Rad) and heated at 40°C for 5 min. Proteins were resolved using 12 or 15% SDS-PAGE and transferred to Immobilon P SQ PVDF membranes (Millipore, Bedford, MA). Membranes were blocked for 1 h in 2.3% dried milk, 0.5% bovine serum albumin, 0.1% Nonidet P-40, 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 0.1% Tween 20, followed by overnight incubation at 4°C with mouse anti-FLAG M1 monoclonal antibody (Sigma). Membranes were then washed and incubated with anti-mouse IgG conjugated with alkaline phosphatase (Santa Cruz Biotechnology, Santa Cruz, CA) for 1 h at room temperature and developed using CDP-Star Western blot chemiluminescence reagent (PerkinElmer Life Sciences) as described previously (32). Kodak Biomax MR film was used to capture chemiluminescence.
Receptor glycosylation was investigated by incubating cell lysates in 150 mM Tris-HCl, pH 7.5, 300 mM NaCl, 1 mM MgCl 2 , 1 mM CaCl 2 , 1% Triton X-100, 10% glycerol, and protease inhibitor mixture with or without N-glycosidase F (40 units/mg of membrane protein, Roche Molecular Biochemicals, Indianapolis, IN) at 37°C for 3 h. For immunoblotting, treated and untreated lysates were diluted with an equal volume of gel loading buffer, proteins were resolved using 10% SDS-PAGE, and assayed by immunoblotting as described above.
For receptor immunoprecipitation, cell extracts were incubated for 2 h at 4°C with 30 l of lysis buffer containing a suspension of anti-FLAG M1 monoclonal antibody conjugated to agarose (Sigma). Following incubation, antibody gel beads were washed three times by centrifugation with lysis buffer and incubated for 1 h at room temperature with 40 l of elution buffer (50 mM Tris-HCl, pH 7.5, 300 mM NaCl, 1% Triton X-100, 10% glycerol, and 5 mM EDTA). Following brief centrifugation, supernatants containing immunopurified receptor protein were mixed with an equal volume of gel loading dye and heated at 40°C for 5 min. Proteins were resolved using 12% SDS-PAGE and transferred to PVDF membranes. Receptor proteins were detected using anti-FLAG M1 monoclonal antibodies, as described above. Ubiquitinated receptor proteins were visualized with rabbit anti-ubiquitin antibodies (Dako, Carpinteria, CA) followed by goat anti-rabbit IgG secondary antibodies conjugated with alkaline phosphatase (Santa Cruz). Blots were developed using CDP Star chemiluminescence reagent.
Pulse-Chase Metabolic Labeling-For metabolic labeling studies, stably transfected HEK 293 cells expressing FLAG-tagged ␦ opioid receptors were incubated in serum-and methionine-free minimal essential media containing 90 Ci/ml [ 35 S]methionine for 3 h at 37°C, then chased with methionine-containing minimal essential media in the absence and presence of the peptide agonist, DADL (1 M), for 2, 4, 6, 8, and 18 h. Cell lysates were immunoprecipitated with anti-FLAG monoclonal antibody agarose beads, and resolved using SDS-PAGE. Gels were fixed and dried prior to autoradiography. ␦ Receptor bands were quantified by excising the receptor band from the gel followed by liquid scintillation counting, and by scanning densitometry of autoradiographic bands using NIH Image software, version 1.61; results were comparable.
Agonist-induced Phosphorylation of MAP Kinase-MAP kinase assays were conducted as described previously (33). Briefly, HEK 293 cells expressing the FLAG-tagged ␦ receptor were incubated with or without 25 M ZLLL for 18 h, and then treated with or without DADL (1 M) for an additional 10 min. Cells were lysed in 150 mM Tris-HCl, pH 7.5, 300 mM NaCl, 1 mM MgCl 2 , 1 mM CaCl 2 , 1% Triton X-100, 10% glycerol, containing mixtures of phosphatase inhibitors and protease inhibitors (Sigma). Cell debris was pelleted by centrifugation and protein concentration in the supernatant was determined using the Bio-Rad protein assay. Equal amounts of protein from each sample were resolved using 12% SDS-PAGE and transferred to Immobilon P SQ PVDF membranes. Membranes were blocked for 1 h in 2.3% dried milk, 0.5% bovine serum albumin, 0.1% Nonidet P-40, 50 mM Tris-HCl, pH 7.5, 150 mM NaCl and 0.1% Tween 20, followed by overnight incubation with mouse monoclonal anti-phospho-MAPK antibody or rabbit anti-MAP kinase antibody (New England Biolabs, Beverly, MA). Membranes were then washed and incubated with goat anti-mouse IgG conjugated with alkaline phosphatase or goat anti-rabbit IgG conjugated with alkaline phos-phatase (Santa Cruz Biotechnology, Santa Cruz, CA) for 1 h at room temperature. Blots were washed and developed using CDP Star Western blot chemiluminescence reagent (PerkinElmer Life Sciences) and Kodak Biomax MR film. Blots were quantified using NIH Image software, version 1.61.
In Vitro Transcription and Translation of Opioid Receptors-[ 35 S]Methionine-labeled opioid receptor proteins were expressed using a coupled in vitro transcription and translation system (TNT rabbit reticulocyte extracts, Promega, Madison, WI). Plasmids encoding and ␦ opioid receptor cDNAs were transcribed with T7 RNA polymerase and translated with reticulocyte extracts in the presence of [ 35 S]methionine according to conditions recommended by the manufacturer. Luciferase cDNA transcribed by T7 RNA polymerase served as a positive control reaction. Translated proteins were resolved using 15% SDS-PAGE. Gels were fixed, dried, and then exposed to Kodak Biomax MR film for autoradiography.

Properties of FLAG-tagged Opioid Receptors and Kinetics of
Down-regulation-As reported recently (32,33), we have established stable HEK 293 cell lines that express and ␦ opioid receptors with a FLAG epitope at the amino termini. Using [ 3 H]bremazocine as radioligand, the B max values for cell lines expressing epitope-tagged receptors used in the present experiments were ϳ5 and 13 pmol/mg of protein for and ␦ opioid receptors, respectively, and apparent dissociation constants (K d ) for bremazocine were 1.0 and 0.8 nM, respectively (see Table I).
SDS-PAGE and immunoblot analysis with the anti-FLAG M1 monoclonal antibody revealed that the ␦ opioid receptor was expressed as two major heterogeneous forms, with apparent molecular masses of ϳ55-65 and 130 -145 kilodaltons (kDa) in HEK 293 cells (Fig. 1). As expected due to its larger mass, the opioid receptor migrated more slowly as two major diffuse bands with apparent molecular masses of ϳ70 -80 and 150 -160 kDa. Evidence that ␦ and opioid receptors are glycoproteins was provided by digestion with N-glycosidase F, an amidase that hydrolyzes nearly all types of N-glycan chains from the asparagine in N-linked glycoproteins. N-Glycosidase F treatment of ␦ and opioid receptors increased the mobilities of both bands to species with apparent molecular masses of 34 and 93 kDa for the ␦ receptor, and 36 and 82 kDa for the receptor. The predicted molecular masses for the FLAG-tagged receptors are ϳ41 and 45 kDa for the ␦ and opioid receptors, respectively. The reason for the aberrant electrophoretic mobilities of the receptors, even after deglycosylation, is not known at present. The immunoreactive bands with slower mobilities may represent opioid receptor dimers that have been described recently (34,35).
Prolonged agonist treatment of HEK 293 cells expressing either or ␦ opioid receptors resulted in decreased steady-state receptor levels, as measured by determination of B max using saturation binding analysis or by immunoblotting (see below). This agonist-induced decrease in receptor number is the process referred to in this report as receptor down-regulation. We have recently reported that the magnitude of ␦ receptor downregulation was dependent on which agonist was employed (33). In the present study, it was found that the extent of receptor down-regulation was also dependent on the agonist used ( Fig.  2). Overnight treatment with DAMGO resulted in a Ͼ80% decrease in the receptor B max , while morphine induced only a 45% decrease. Neither ligand altered the apparent dissociation constant of [ 3 H]bremazocine for the receptor (untreated K d ϭ 1.1 Ϯ 0.2 nM, DAMGO-treated K d ϭ 1.2 Ϯ 0.1 nM, morphinetreated K d ϭ 0.9 Ϯ 0.2 nM). Overnight treatment with etorphine, DADL, fentanyl, and methadone decreased the receptor B max by 70 -80%, suggesting that these agonists had an intrinsic efficacy that was similar to that of DAMGO for inducing down-regulation of the receptor (data not shown).
The kinetics of ␦ receptor down-regulation was investigated by pulse-chase analysis following metabolic labeling of cells Effect of pertussis toxin on agonist-induced down-regulation of ␦ and opioid receptors HEK 293 cells expressing FLAG-tagged ␦ or opioid receptors were preincubated at 37°C for 3 h in serum-free Dulbecco's modified Eagle's medium with or without 100 ng/ml pertussis toxin (Ptx), then incubated for 18 h in the absence and presence of 1 M DAMGO ( receptors) or DADL (␦ receptors). Membrane fractions were prepared and saturation analysis was performed using [ 3 H]bremazocine as radioligand. Saturation curves were analyzed by non-linear regression to generate apparent dissociation constants, K D (nM), and maximum number of receptors, B max (fmol/mg of protein). B max ratios represent quotients ϫ 100 of agonist/control, Ptx/control, and Ptx ϩ agonist/Ptx. Data represent mean Ϯ S.E. of three to four experiments conducted in duplicate.

FIG. 1. ␦ And opioid receptors expressed in HEK 293 cells are N-linked glycoproteins.
Cell lysates were prepared by extracting monolayers of HEK 293 cells expressing ␦ or opioid receptors in 0.2 ml of lysis buffer consisting of 150 mM Tris-HCl, pH 7.5, 300 mM NaCl, 1 mM MgCl 2 , 1 mM CaCl 2 , 1% Triton X-100, 10% glycerol and protease inhibitor mixture for 1 h on ice. Cellular debris was pelleted by centrifugation and the supernatants were treated with or without N-glycosidase F (protease-free, 40 units/mg of membrane protein) at 37°C for 3 h, then resolved using 10% SDS-PAGE. FLAG-tagged receptors were assayed by immunoblotting using the anti-FLAG M1 monoclonal antibody. The mobilities of molecular mass standards (in kDa) are indicated to the right. ␦ OR (gly), glycosylated (untreated) ␦ receptors; ␦ OR (degly), deglycosylated (treated with N-glycosidase F) ␦ receptors; OR (gly), glycosylated (untreated) receptors; OR (degly), deglycosylated (treated with N-glycosidase F) receptors. N-Glycosidase F treatment of ␦ receptors increased the mobilities of the 55-65-and 130 -145-kDa bands to species with apparent molecular masses of 34 and 93 kDa, respectively. N-Glycosidase F treatment of receptors shifted the mobilities of the 70 -80-and 150 -160-kDa bands to species with apparent molecular masses of 36 and 82 kDa, respectively. This experiment was replicated three times with similar results.
with [ 35 S]methionine. HEK 293 cells expressing FLAG-tagged ␦ receptors were preincubated at 37°C with [ 35 S]methionine in serum-and methionine-free Dulbecco's modified Eagle's medium for 3 h, and then chased with unlabeled methioninecontaining medium in the absence and presence of 1 M DADL. Cell lysates were prepared at various times following the unlabeled methionine chase, immunoprecipitated with the anti-FLAG M1 monoclonal antibody, and then analyzed by SDS-PAGE and autoradiography (Fig. 3). In the absence of agonist, receptor degradation exhibited first-order exponential decay with a rate constant of 0.08 Ϯ 0.02, corresponding to a half-life of 8.7 h. In the presence of DADL, the receptor degradation rate was accelerated significantly with a rate constant of 0.42 Ϯ 0.07 and a half-life of 1.6 h.
The kinetics of receptor down-regulation was also examined by immunoblot analysis (Fig. 4). Stable HEK 293 cells expressing epitope-tagged receptors were incubated in serum-free medium at 37°C with 1 M DADL for various time intervals ranging from 0.5 to 18 h, and cell lysates were then analyzed by immunoblot analysis using the M1 monoclonal antibody. The level of ␦ receptor protein remained steady until 2-4 h after DADL treatment. Levels then decreased progressively at 4, 6, and 18 h after agonist administration to 77, 38, and 4% of zero time controls, respectively. The time course of receptor downregulation was similar, although a decrease in receptor protein was already apparent at the 2-h time point.
The rate of recovery of ␦ receptor expression following agonist-induced down-regulation was measured by treating ␦ receptor-expressing HEK 293 cells with 1 M DADL for 18 h, then assaying receptor binding at various time points following removal of the agonist from the media. As displayed in Fig. 5, down-regulation was readily reversible upon removal of agonist. After an 18-h exposure of cells to DADL, receptor levels were reduced to 13% of untreated control cells. At 2, 4, 8, and 24 h following agonist removal, receptor levels were 31, 50, 69, and 95% of untreated control cells, respectively.
Role of Functional G Protein-To evaluate the role of G i and G o heterotrimeric GTP-binding proteins in the receptor downregulation pathway, cells expressing FLAG-tagged or ␦ opioid receptors were pretreated with or without 100 ng/ml pertussis toxin in serum-free media for 3 h at 37°C, followed by overnight incubation in the absence and presence of 1 M DADL. Receptor levels in cell lysates were compared by immunoblot analysis and quantified using NIH Image software. Marked differences between and ␦ opioid receptors were observed regarding the effects of pertussis toxin on receptor down-regulation (Fig. 6). Incubation of cells expressing the ␦ receptor for 18 h with DADL resulted in a 90% loss of receptor protein.
Pertussis toxin treatment alone in the absence of agonist resulted in a nearly 2-fold increase in ␦ receptor expression. The efficacy of DADL toward inducing down-regulation of the ␦ receptor was not changed appreciably by pretreatment with pertussis toxin: overnight treatment with the peptide still resulted in a 75% decrease in receptor protein in cells that were preincubated with pertussis toxin.
Markedly different results were obtained regarding the sensitivity of receptor down-regulation to pertussis toxin. Chronic exposure of HEK 293 cells expressing the receptor with agonist led to an 85% decrease in the level of receptors, as expected. Treatment with pertussis toxin alone increased receptor protein by 50% over control levels. In striking contrast to the results with the ␦ receptor, however, preincubation with pertussis toxin completely blocked DADL-induced down-regulation of the opioid receptor (Fig. 6).
Qualitatively similar results were obtained when the effect of pertussis toxin on receptor down-regulation was determined by radioligand saturation analysis (Table I). Prolonged incubation of HEK 293 cells expressing ␦ and opioid receptors with peptide agonists decreased the number of receptors by 80 -90%. Incubation of cells with pertussis toxin in the absence of agonist resulted in a 2.7-and 1.6-fold increase in B max of ␦ and receptors, respectively. Pertussis toxin pretreatment had no effect on the extent of ␦ receptor down-regulation, but reversed significantly the decrease in the B max of the receptor induced by DAMGO. Following an 18-h incubation with DAMGO, the receptor B max was 20% of the untreated sample; in the presence of pertussis toxin the DAMGO-treated B max was 70% of the control level.
Proteases Involved in Agonist-induced Down-regulation of and ␦ Opioid Receptors-To attempt to identify which proteases were responsible for agonist-induced down-regulation of opioid receptors, immunoblotting and radioligand receptor binding assays were performed using cells that were preincubated with several different cell-permeable protease inhibitors prior to agonist stimulation. To test the prevailing hypothesis that receptors are degraded in the lysosome, the protease inhibitor examined initially was E64d ((2S,3S)-trans-epoxysuccinyl-L-leucylamido-3-methylbutane ethyl ester, loxastatin). E64d readily enters cells and irreversibly inhibits thiol proteases, and is particularly active against lysosomal cathepsins B, H, and L (36). E64d also inhibits effectively the proteolytic activities of calpains I and II, which are cytosolic, calcium-dependent cysteine proteases. Stable HEK 293 cells expressing ␦ opioid receptors were pretreated at 37°C in serum-free media for 1 h with or without 25 M E64d, followed by 18 h treatment with or without 1 M DADL. Saturation analysis of [ 3 H]bremazocine binding to washed membrane preparations indicated that E64d treatment had minimal effect on DADL-induced down-regulation (Fig. 7A) and had no effect on steady-state ␦ receptor levels in the absence of agonist (Fig. 7B).
Attention was then focused on the proteasome, a large multisubunit protease complex that degrades the majority of shortlived intracellular proteins (37,38). Pretreatment with the cell-permeable proteasome inhibitor, ZLLL, under the same conditions as those described above for E64d, yielded strikingly different results. ZLLL significantly attenuated the extent of ␦ receptor down-regulation. Prolonged treatment with DADL induced an 86% decrease in B max , while in the presence of ZLLL there was only a 45% decrease (Fig. 7A). Incubation with ZLLL alone also increased steady-state ␦ receptor levels by 70% (Fig. 7B).
The effect of lactacystin was tested to further support the involvement of the proteasome in ␦ receptor turnover. Lactacystin is a Streptomyces product that acts as a highly selective irreversible inhibitor of the proteasome (39). The effects of lactacystin were very similar to those observed using ZLLL. Lactacystin significantly blocked DADL-induced down-regulation of the ␦ opioid receptor with an efficacy that was similar to ZLLL (Fig. 7A), and increased steady-state levels of the ␦ receptor by 85% (Fig. 7B).
The proteasome inhibitor, ZLLL, also effectively inhibited ␦ opioid receptor down-regulation in neuroblastoma ϫ glioma NG108-15 cells that endogenously express the ␦ opioid receptor gene (Table II). Overnight treatment of NG108-15 cells at 37°C with 1 M DADL in serum-free medium resulted in a 91% decrease in the receptor B max , based on saturation analysis with [ 3 H]bremazocine as radioligand (B max ratio was 9.1%). In the NG108-15 cell line, incubation with the proteasome inhibitor alone did not increase the receptor density. This suggests that the proteasome is not involved in basal turnover of the ␦ receptor in this cell line, unlike its activity in HEK 293 cells (Fig. 7B, and see below). When NG108-15 cells were preincubated at 37°C for 1 h with 25 M ZLLL prior to agonist treatment, the B max ratio increased to 39% (Table II), indicating that the extent of DADL-induced down-regulation was reduced significantly by the proteasome inhibitor (ANOVA, Tukey's Multiple Comparison Test, p Ͻ 0.05).
To probe further the activity of ZLLL, experiments were performed to estimate the lowest effective concentration necessary to block agonist-induced down-regulation of the ␦ opioid receptor and increase steady-state receptor levels in HEK 293 cells. Using saturation analysis with [ 3 H]bremazocine to deter-FIG. 5. Recovery of agonist-induced ␦ opioid receptor downregulation assayed by radioligand binding. HEK 293 cells expressing the FLAG-tagged ␦ receptor were incubated in serum-free medium at 37°C for 18 h in the absence and presence of 1 M DADL, cells were washed three times, and then replenished with serum-free medium lacking agonist. Cells were harvested at 0, 2, 4, 8, and 24 h following removal of DADL, homogenized, and assayed for specific receptor binding using 3 nM [ 3 H]ethylketocyclazocine, and 10 M cyclazocine to assess nonspecific binding. Data are expressed as percentages of the binding to homogenates from control cells not treated with agonist, normalized for protein content, and represent the means and standard errors of three to five independent assays conducted in duplicate.

FIG. 6. Effect of pertussis toxin on agonist-induced down-regulation of the ␦ and opioid receptor. HEK 293 cells expressing FLAG-tagged ␦ receptors (upper panel) or receptors (lower panel)
were preincubated at 37°C in the absence or presence of 100 ng/ml pertussis toxin (Ptx) for 3 h in serum-free media followed by overnight treatment with or without DADL (1 M), as indicated above the immunoblot. Cell lysates were prepared and resolved by SDS-PAGE, followed by immunoblotting with the anti-FLAG M1 monoclonal antibody. The mobilities of 61-and 85-kDa molecular mass standards are indicated to the right. Pertussis toxin pretreatment had little effect on the extent of ␦ receptor down-regulation, but completely reversed agonist-induced down-regulation of the receptor. This experiment was repeated three times with similar results. mine B max , it was observed that 2.5 M ZLLL in the cell culture medium was sufficient for significant attenuation of ␦ receptor down-regulation and to inhibit basal turnover (Fig. 8). The effects did not reach statistical significance using the lowest concentration of ZLLL tested (250 nM). As a further control for specificity, the effect of the related dipeptide aldehyde, ZLL, at 25 M was evaluated. This reagent, unlike the tripeptide analog, displays little inhibitory activity toward the proteasome, but does inhibit calpain and lysosomal thiol proteases (38). Pretreatment with the dipeptide aldehyde ZLL had no effect on agonistinduced down-regulation of opioid receptors (data not shown).
Immunoblot analysis was also utilized to evaluate the effect of the tripeptide ZLLL on and ␦ opioid receptor down-regulation and steady-state levels in HEK 293 cells. As illustrated in the left panel of Fig. 9, chronic DADL treatment caused a 70% decrease in the level of opioid receptor protein. DADLinduced down-regulation of the receptor was completely blocked in the presence of the proteasome inhibitor, and ZLLL treatment alone (in the absence of agonist) increased the steady-state level of receptor protein. Similar results were observed in immunoblots of the ␦ opioid receptor (Fig. 9, right   panel). The DADL-induced decrease in the level of ␦ receptor protein was completely reversed by preincubation with ZLLL, and treatment with ZLLL alone increased steady-state ␦ receptor levels. Thus, the results obtained utilizing immunoblot FIG. 7. Effect of protease inhibitors on agonist-induced downregulation and basal levels of the ␦ opioid receptor. HEK 293 cells expressing the ␦ receptor were pretreated for 1 h at 37°C with or without 25 M E64d (a lysosomal protease and calpain inhibitor), ZLLL (proteasome inhibitor), or lactacystin (proteasome inhibitor), then incubated in the absence and presence of 1 M DADL for an additional 18 h at 37°C. Membrane fractions were prepared and washed thoroughly to remove agonist, and then used for saturation analysis using [ 3 H]bremazocine as radioligand and cyclazocine to assess nonspecific binding. The maximum number of binding sites (B max ) was determined by nonlinear regression analysis of the curves using GraphPad Prism software and were normalized for protein content. A, effect of protease inhibitors on DADL-induced down-regulation. DADL, cells incubated for 18 h with agonist without preincubation with a protease inhibitor; ϩE64d, cells preincubated with E64d followed by 18 h DADL treatment; ϩZLLL, cells preincubated with ZLLL followed by 18 h DADL treatment; ϩlactacyst, cells preincubated with lactacystin followed by 18-h DADL treatment. B, effect of protease inhibitors on steady state levels of opioid receptors. Cells were incubated for 19 h with protease inhibitors (in the absence of DADL) as indicated. Data are expressed as percentages of the control B max in the absence of agonist, and represent the means and standard errors of three to five independent assays conducted in duplicate. In A, asterisks indicate significant differences from DADL treatment alone; in B, asterisks indicate significant differences from untreated controls (ANOVA, Tukey's Multiple Comparison Test, p Ͻ 0.05). Proteasome inhibitors attenuated agonist-induced down-regulation and increased steady state ␦ receptor levels.  analysis agreed well with radioligand binding data regarding the ability of the proteasome inhibitor to block agonist-induced down-regulation and to elevate steady-state receptor levels in HEK 293 cells.
It should be noted that for both the and ␦ receptor, the proteasome inhibitor induced the appearance or increased the levels of several immunoreactive receptor bands compared with untreated samples (Fig. 9). It was particularly evident that higher molecular weight species of receptors were induced in cells treated with proteasome inhibitors. In addition, lower molecular weight species in and ␦ receptor samples were stabilized in cells treated with ZLLL.
One of the downstream effectors regulated by opioid receptor activation is MAP kinase, which becomes rapidly activated by phosphorylation on Thr 202 and Tyr 204 following agonist treatment (40). It was of interest to determine whether opioid receptors were still capable of activating the MAP kinase signal transduction pathway following incubation of cells with proteasome inhibitors. HEK 293 cells expressing FLAG-tagged ␦ receptors were incubated at 37°C in serum-free media for 18 h in the absence and presence of 25 M ZLLL, and then subsequently treated with and without 1 M DADL for 10 min. Protein extracts were resolved by SDS-PAGE then assayed by immunoblotting with antibodies specific for the phosphorylated form of MAP kinase, or with antibodies that recognize total MAP kinase protein as a control. As shown in Fig. 10, DADL induced phosphorylation of MAP kinase to a similar extent in both untreated and ZLLL-treated cells, indicating that disruption of opioid receptor turnover did not have a significant impact on agonist-dependent signal transduction.
Another series of experiments was performed utilizing immunoblotting to assess the ability of other protease inhibitors to modulate the extent of agonist-induced down-regulation of ␦ and receptors. It was consistently observed that proteasome inhibitors, but not lysosomal protease inhibitors, calpain inhibitors, or caspase inhibitors, were effective at blocking agonistinduced down-regulation of both ␦ and receptors (Figs. 11 and 12, respectively). As discussed earlier using ZLLL (Fig. 9), all of the additional proteasome inhibitors tested, including lactacystin, PSI, and ALLN, increased steady-state ␦ and receptor levels when cells were preincubated with the proteasome in-hibitors in the absence of agonist (Figs. 11 and 12). Accumulation of higher molecular weight receptor species was also observed with all proteasome inhibitors, although this is not evident in the cropped figures. As shown in Figs. 11 and 12, preincubation of HEK 293 cells with lactacystin, PSI, and ALLN completely blocked DADL-induced down-regulation of FLAG-tagged ␦ and receptors. In contrast, inhibition of lysosomal cathepsins and thiol proteases (by E64d and leupeptin), trypsin-like serine proteases (by leupeptin), cytoplasmic calpain (by E64d and leupeptin), and caspases (by caspase-3 in- hibitor III) had little or no effect on the extent of agonistinduced down-regulation of the ␦ opioid receptor (Fig. 11) or opioid receptor (Fig. 12). Our observation that leupeptin did not block opioid receptor down-regulation differed from a previous study in which it was reported that high concentrations of leupeptin (100 g/l) attenuated down-regulation of the ␦ receptor (29). We have observed recently, however, that leupeptin inhibits agonist binding to and ␦ receptors in vitro at concentrations greater than 100 M. 2 Ubiquitination of Opioid Receptors-Proteins that are degraded by the proteasome are first tagged by covalent attachment of the 8.5-kDa protein, ubiquitin, which is recognized by the 19 S regulatory complex of the proteasome (37,38). Based on our observations regarding the effects of proteasome inhibitors on agonist-dependent opioid receptor down-regulation and basal turnover, we sought evidence for ubiquitination of and ␦ receptors. Suggestive evidence that opioid receptors were ubiquitinated came from experiments in which FLAG-tagged and ␦ receptor cDNAs were transcribed and translated in vitro using rabbit reticulocyte lysates, in comparison with transcription and translation of luciferase cDNA as a control. Following SDS-PAGE and autoradiography of the [ 35 S]methionine-labeled translation products, luciferase migrated as a sharp band of 61 kDa, along with several minor bands with slightly greater mobilities (Fig. 13). It is important to note that no bands were evident in the luciferase sample that corresponded to molecular weights greater than the expected luciferase protein of 61,000. In contrast, translation of the receptor yielded a major band migrating with an apparent molecular mass of 42 kDa, but also included a diffuse ladder of bands with greater molecular masses that spanned the length of the gel (Fig. 13). Similar results were obtained when the ␦ receptor was translated. The predominant band migrated with an apparent molecular mass of 37 kDa, however, also evident was a spectrum of bands migrating with slower mobilities extending the entire length of the gel lane. Rabbit reticulocyte lysates contain ubiquitin and the enzymes required for conjugation of ubiquitin to proteins (37,38), and close scrutiny of the autoradiograms of in vitro translated opioid receptors revealed that many bands were spaced at intervals of ϳ7-10 kDa, consistent with the molecular mass of ubiquitin. Although alternative explanations exist, it was plausible that the higher molecular weight receptor species evident following in vitro translation or that accumulate in cells incubated with proteasome inhibitors represent polyubiquitinated receptors.
Furthermore, direct evidence for ubiquitination of opioid receptors, utilizing immunoprecipitation of opioid receptors followed by immunoblotting with anti-ubiquitin antibodies is provided in Fig. 14. HEK 293 cells expressing FLAG-tagged ␦ receptors (Fig. 14, left panel) or receptors (Fig. 14, right  panel) were pretreated with or without 25 M ZLLL for 1 h at 37°C in serum-free medium, then incubated in the absence and presence of 1 M DADL for 18 h. Protein extracts were prepared using lysis buffer containing a mixture of protease inhibitors. Lysates were immunoprecipitated with anti-FLAG monoclonal antibodies conjugated to agarose beads, resolved by SDS-PAGE, and then transferred to PVDF membranes, as described under "Experimental Procedures." To validate the immunoprecipitation procedure, the blotted PVDF membranes were initially probed with anti-FLAG M1 monoclonal antibodies to verify the presence of FLAG-tagged opioid receptors. The results (data not shown) were very similar to those shown in Fig. 9, which correspond to cell lysates assayed directly by 2 A. Khokhar, K. Chaturvedi, and R. D. Howells, manuscript in preparation. Luciferase cDNA was transcribed by T7 RNA polymerase to serve as a control. Translated proteins were resolved using 15% SDS-PAGE, gels were fixed, dried, and exposed to Kodak Biomax MR film for autoradiography. The mobilities of molecular mass standards (in kDa) are indicated to the left. In vitro translated and ␦ receptors migrated with apparent molecular masses of 42 and 37 kDa, respectively, however, unlike the luciferase control, a spectrum of bands migrating with slower mobilities was also present, some of which are labeled to the right of the lanes (Ͻ). This experiment was replicated three times with similar results.
immunoblotting, indicating that opioid receptors were efficiently immunoprecipitated using the conditions employed. When the four ␦ receptor immunoprecipitated samples were blotted with anti-ubiquitin antibodies, untreated (ϪZLLL, ϪDADL) and DADL-treated (ϪZLLL, ϩDADL) samples displayed very low levels of polyubiquitinated receptors (Fig. 14,  left panel). In contrast, it was clearly evident that receptors were polyubiquitinated in samples preincubated with the proteasome inhibitor, ZLLL, either in the absence and presence of DADL (ϩZLLL, ϪDADL and ϩZLLL, ϩDADL, respectively, Fig. 14, left panel). In both of these samples, polyubiquitinated receptor proteins were evident with apparent molecular masses ranging from 42 kDa to greater than 187 kDa.
Analysis of receptor immunoprecipitates indicated that in untreated samples, there were readily detectable receptor proteins conjugated to ubiquitin, which had apparent molecular masses of ϳ122 and 230 kDa. No polyubiquitinated receptor was detectable in the DADL-treated, down-regulated receptor immunoprecipitate (lane 2, right panel). The absence of any signal in this lane highlighted the lack of nonspecific binding of the anti-ubiquitin antibody under these conditions. Blank lanes were also observed when wild type HEK 293 cells that did not express opioid receptors were immunoprecipitated and immunoblotted using the same procedures as a negative control (data not shown). In contrast, a spectrum of polyubiquitinated receptor protein bands was clearly evident in samples preincubated with the proteasome inhibitor ZLLL, corresponding to molecular masses of 79 kDa and species with molecular masses from 120 kDa and greater that accumulated at the top of the gel. It seems reasonable to assume that inhibition of the proteasome stabilized the polyubiquitinated receptors that, in the absence of inhibitors, would normally be rapidly degraded. DISCUSSION The major conclusions to be drawn from this study are as follows: 1) proteolysis of and ␦ opioid receptor protein is evident 2-4 h following DADL treatment and there is a near complete loss of receptor protein following 18 h of exposure. 2) Based on pulse-chase experiments, the ␦ receptor degradation rate is accelerated 5-fold by agonist treatment. 3) Functional pertussis toxin-sensitive G proteins are required for agonistinduced down-regulation of the opioid receptor but not for ␦ opioid receptor down-regulation. 4) Agonist-induced down-regulation of opioid receptors can be blocked by proteasome inhibitors but not by lysosomal, calpain, or caspase protease inhibitors. 5) Proteasome inhibitors, but not lysosomal protease inhibitors, also increase steady-state levels of and ␦ opioid receptors. 6) Ubiquitinated and ␦ opioid receptors accumulate in cells treated with proteasome inhibitors.
In eukaryotic cells, a wide variety of proteins with roles in cell cycle progression, transcriptional control, signal transduction, and metabolic regulation are degraded by the ubiquitinproteasome pathway (reviewed in Refs. 37, 38, and 41). The 26 S proteasome is a 2.5 MDa complex consisting of a 20 S proteolytic core complex and two 19 S regulatory complexes, and is capable of cleaving at basic, acidic, and hydrophobic amino acids within proteins. Proteins are targeted to the proteasome by covalent ligation to ubiquitin, a highly conserved 76-amino acid protein. Ubiquitin-protein ligation involves the sequential action of at least two, and often three, enzymes. The C-terminal glycine of ubiquitin is first activated through formation of an ATP-dependent thioester linkage with a cysteine in the E1 activating enzyme. The ubiquitin is next transferred to an active site cysteine of a ubiquitin-conjugating protein, E2, prior to formation of an isopeptide bond between the C terminus of ubiquitin, and an ⑀-amino group of a lysine within the target protein. Target protein recognition is often mediated by E3 ubiquitin-ligases. After the linkage of ubiquitin to the substrate protein, a polyubiquitin chain is usually formed, in which the C terminus of ubiquitin becomes linked to one of several lysine residues in the previous ubiquitin. The polyubiquitinated proteins are recognized by the 19 S regulatory subunits of the 26 S proteasome, the ubiquitin moieties are recycled through the action of ubiquitin hydrolases, and the targeted protein substrates are degraded by the 20 S catalytic core complex.
Membrane-permeable inhibitors of the proteasome have contributed greatly to our understanding of the involvement of the ubiquitin/proteasome system in protein degradation. The initial proteasome inhibitors were hydrophobic peptide aldehydes that can enter mammalian cells and inhibit proteasome function in vivo (42)(43)(44)(45). The most selective proteasome inhibitor is lactacystin, a Streptomyces metabolite. Lactacystin inhibits the trypsin-like, chymotrypsin-like, and peptidylglutamyl-peptide hydrolyzing activities of the proteasome by covalent modification of the amino-terminal threonine of the mammalian proteasome subunit X, confirming that this residue is essential for catalytic activity (39). In this study, we have used four different proteasome inhibitors, including lactacystin, to confirm the role of the proteasome in and ␦ opioid receptor turnover and agonist-induced down-regulation.
In addition to the selective destruction of cytosolic and nuclear regulatory proteins, it has become apparent that several yeast and mammalian membrane proteins also undergo ubiquitination and degradation by the proteasome (41). In mammalian cells, studies using proteasomal inhibitors have implicated the proteasome in the down-regulation of the plateletderived growth factor receptor (46), low density lipoprotein receptor (47), Met tyrosine kinase receptor (48), mannose phosphate receptor (49), and growth hormone receptor (50). In Saccaromyces cerevisiae, Ste2p, which is the G protein-coupled receptor for the pheromone ␣ mating factor, is ubiquitinated in a ligand-dependent manner while the receptor resides in the plasma membrane, and ubiquitination is required for endocy- were pretreated with or without 25 M ZLLL for 1 h at 37°C in serum-free medium, then incubated in the absence and presence of 1 M DADL for 18 h, as indicated above the immunoblots. Opioid receptors were immunoprecipitated by incubating cell extracts for 2 h at 4°C with 30 l of anti-FLAG M1 monoclonal antibody agarose gel. Immunoprecipitated receptors were resolved on 12% SDS-PAGE and transferred to PVDF membranes. Ubiquitinated receptor proteins were detected with rabbit anti-ubiquitin antibody followed by goat anti-rabbit IgG conjugated with alkaline phosphatase. Blots were developed using CDP Star chemiluminescence reagent. The mobilities of molecular mass standards (in kDa) are indicated to the right of the immunoblots. This experiment has been replicated three times with similar results. tosis (51). In this case, the receptor is monoubiquitinated rather than being polyubiquitinated, and the monoubiquitin serves as an internalization signal but does not target the receptor to the proteasome (52), which generally requires a ubiquitin chain at least four units in length for recognition (53). Regarding mammalian G protein-coupled receptors, it has been reported that rhodopsin and its G protein, transducin, are ubiquitinated and subject to ubiquitin-dependent proteolysis in vertebrate rod outer segments (54). Evidence has also been presented that another component of GPCR signaling, namely the G protein-coupled receptor kinase 2, is degraded by the ubiquitin/proteasome pathway (55).
Regarding early signals that trigger agonist-induced downregulation, in this study we found that pretreatment of cells with pertussis toxin blocked down-regulation of the receptor, while ␦ receptor down-regulation remained unaltered. These results are in agreement with the previous studies (56) using Neuro2A cells expressing cloned opioid receptors, but contrast with another report in which it was stated that in C6 cells, down-regulation of the expressed receptor by full agonists was independent of G protein coupling (57). The reason for the discrepant results is not known, but it may be related to the use of different cell lines to express opioid receptors. The insensitivity of ␦ receptor down-regulation to pertussis toxin may indicate that receptor proteolysis is independent of G protein coupling, or that the agonist-stimulated receptor couples to Ptx-insensitive G proteins that trigger pathways involved in receptor proteolysis. Indeed, evidence in support of ␦ receptor coupling to G␣ q and G␣ z has been reported (58).
Results derived both from direct receptor immunoblots and from immunoprecipitation of opioid receptors followed by immunoblotting with ubiquitin antibodies are consistent with the proposal that and ␦ receptor proteins are ubiquitinated directly. It appears that the majority of ubiquitin-immunoreactive bands present in receptor immunoprecipitates are also recognized by the FLAG monoclonal antibody. Ubiquitination sites on the receptors are currently being investigated by substituting lysine residues with arginine. It is also possible, although less likely, that the ubiquitin-immunoreactive bands in Fig. 14 represent other proteins that are tightly associated with the opioid receptors. The protein-protein interactions giving rise to such complexes would obviously have to withstand the relatively harsh conditions of detergent extraction, immunoprecipitation, and SDS-PAGE. It is also possible that opioid receptors are not targeted directly to the proteasome complex. For example, it is conceivable that inhibition of receptor downregulation by proteasome inhibitors is due to stabilization of a protein which acts to prevent receptor down-regulation, such as a protease inhibitor, that is degraded by the ubiquitin/proteasome pathway.
We do not wish to dispute recent observations that following endocytic trafficking, opioid receptors bound to endosomes fuse eventually with lysosomes (29,30). We interpret our data as suggesting that the ubiquitin/proteasome pathway operates upstream of trafficking to lysosomes to initiate receptor downregulation. A similar proposal has been made regarding the down-regulation of the platelet-derived growth factor receptor ␤: after ligand stimulation the receptor is polyubiquitinated, degraded by the proteasome, and resultant peptide fragments are delivered to and further degraded in lysosomes (46). The intracellular location where the proteasome-mediated proteolysis of the receptors is occurring is not known at present. It is of interest to note that mutations have been made in the ␦ receptor that block internalization but do not block down-regulation (34,59). It was also reported recently that the Src kinase inhibitor, PP1, inhibited ␦ receptor internalization but not receptor down-regulation (60). In addition, morphine induces down-regulation of the receptor, albeit with less efficacy than DAMGO, but does not trigger receptor internalization (61,62). Taken together, these observations suggest that opioid receptor proteolysis may be occurring at the plasma membrane. In accord with this proposal, it has been reported recently that down-regulation of the ␤ 2 -adrenergic receptor does not require endocytosis and does not involve the lysosomal degradation pathway (63). In summary, results of this study implicate a prominent role for the ubiquitin/proteasome pathway in agonist-induced down-regulation and basal turnover of opioid receptors.