Protein Kinase Cζ Mediates μ-Opioid Receptor-induced Cross-desensitization of Chemokine Receptor CCR5*

We have previously shown that the μ-opioid receptor (MOR) is capable of mediating cross-desensitization of several chemokine receptors including CCR5, but the biochemical mechanism of this process has not been fully elucidated. We have carried out a series of functional and biochemical studies and found that the mechanism of MOR-induced cross-desensitization of CCR5 involves the activation of PKCζ. Inhibition of PKCζ by its pseudosubstrate inhibitor, or its siRNA, or dominant negative mutants suppresses the cross-desensitization of CCR5. Our results further indicate that the activation of PKCζ is mediated through a pathway involving phosphoinositol-dependent kinase-1 (PDK1). In addition, activation of MOR elevates the phosphorylation level and kinase activity of PKCζ. The phosphorylation of PKCζ can be suppressed by a dominant negative mutant of PDK1. We observed that following MOR activation, the interaction between PKCζ and PDK1 is immediately increased based on the analysis of fluorescent resonance energy transfer in cells with the expression of PKCζ-YFP and PDK1-CFP. In addition, cells expressing PKCζ kinase motif mutants (Lys-281, Thr-410, Thr-560) fail to exhibit full MOR-induced desensitization of CCR5 activity. Taken together, we propose that upon DAMGO treatment, MOR activates PKCζ through a PDK1-dependent signaling pathway to induce CCR5 phosphorylation and desensitization. Because CCR5 is a highly proinflammatory receptor, and a critical coreceptor for HIV-1, these results may provide a novel approach for the development of specific therapeutic agents to treat patients with certain inflammatory diseases or AIDS.

motactic responses and chemokine expression in human and murine cells (6 -9). Similarly, Tyr-D-Ala-Gly-N-Me-Phe-Gly-ol (DAMGO), 3 a highly selective ligand for the -opioid receptor (MOR), suppresses chemokine receptor function in vitro (7, 10 -12). Recent work from our laboratories, and others, suggest that heterologous desensitization is the primary mechanism for the inhibitory activity of opioids for chemokine receptor function. Studies reported by Grimm et al. (13,14) show that activation of and ␦ opioid receptors (MOR and DOR) induced cross-desensitization of the chemokine receptors CCR1, CCR2, CXCR1, and CXCR2, but not FPR (the high affinity receptor for fMLF), resulting in a failure of monocytes and neutrophils to manifest chemotactic responses to chemokines such as CCL2, CCL3, and CCL5 in a dose-dependent manner. Further studies have shown that activation of MOR in HEK293 cells stably expressing both MOR and CCR1, failed to migrate in response to CCL3, a CCR1 ligand (15). More recently, using both primary cells as well as stably transfected cell lines, we have shown that MOR activation induces heterologous desensitization of CCR5 but not CXCR4 (16), a finding which is consistent with the notion that cross-talk among G protein-coupled receptors (GPCRs) is selective (17).
The mechanisms mediating heterologous desensitization between opioid and chemokine receptors are not clearly understood. Our laboratories have previously shown that the MORinduced cross-desensitization of CCR5 does not result in target receptor endocytosis (10,13,14,16). Evidence from studies of cross-talk between a number of chemokine, opioid, and formyl peptide receptors suggest that multiple processes may contribute to heterologous desensitization. However, the biochemical mechanism of cross-desensitization in most cases involves target receptor phosphorylation (10,13,(17)(18)(19). In addition, studies from several laboratories have shown that second messenger-dependent kinases are required for heterologous desensitization of a number of G protein-coupled receptors (reviewed in Ref. 19), and in particular, the function of protein kinase C (PKC) is typically required for the cross-talk between G i -coupled receptors.
We are particularly interested in the cross-talk consequences following the activation of MOR, since this receptor is the major receptor, which is activated following heroin or morphine abuse. Moreover, we have given our primary attention to the capacity of this opioid receptor to cross-desensitize CCR5, because this receptor is a critical major HIV co-receptor, and a major participant in inflammatory responses. The biochemical basis for the cross-talk between MOR and CCR5 has not previously been examined, and in the present study, we report that PKC, an atypical PKC, plays an important role in the MORinduced cross-desensitization of CCR5. We employed a specific PKC pseudosubstrate inhibitor, a PKC siRNA, or PKC dominant negative mutants to show that the inhibition of PKC reduces the DAMGO-induced desensitization of CCR5. We also demonstrate that the activation of MOR promotes the association between CCR5 and PKC, and induces both the kinase activity of PKC, as well as the phosphorylation of CCR5 in both transfected cell lines and primary macrophages. Finally, we find that the MOR-induced activation of PKC is dependent on the activation of PDK1. Moreover, the activation of PKC is associated with the formation of a complex between PKC and PDK1.
Isolation of Monocytes and Generation of Monocyte-derived Macrophages-Human peripheral blood mononuclear cells (PBMC) were obtained from healthy donors and isolated from buffy coats by density gradient centrifugation using Ficoll-Paque plus (Amersham Biosciences). Monocytes were isolated from PBMCs using the Midi-MACS magnetic separation system and monocyte isolation kit II (Miltenyi Biotec, Auburn, CA) according to the manufacturer's protocol. Briefly, cells were incubated with 30 l of MACS buffer (PBS containing 2 mM EDTA and 0.5% BSA) and 10 l of FcR blocking reagents per 10 7 cells. After incubation at 4°C for 10 min, 30 l of buffer and 20 l of anti-biotin microbeads per 10 7 cells were added. After an additional incubation at 4°C for 15 min, the cells were washed in 2 ml of buffer, then loaded onto a magnetic separation column. The unlabeled, negatively selected, monocytes were eluted off the column and collected. Monocytes were seeded in 6-well plates and macrophages were generated by incubating these monocytes for 7 days in R10 medium (RPMI 1640 containing 10% fetal bovine serum, 25 mM HEPES, 2 mM L-glutamine, and 0.1% gentamycin) supplemented with 100 ng/ml recombinant human macrophage colony-stimulating factor (M-CSF; PeproTech, Rocky Hill, NJ).
M13 Transfection-Cell transfection was performed using a Nucleofector following the manufacturer's instructions (Amaxa Inc., Gaithersburg, MD). In brief, cells were passaged for 2 days in 300.19 medium without antibiotics. Cells were then resuspended at 1 ϫ 10 6 cells/100 l in nucleofector solution V and mixed with 2.5 g construct DNA, or siRNA (Santa Cruz Biotechnology Inc. Santa Cruz, CA), followed by electroporation using the Amaxa Nucleofector with the U-015 program.
Preparation of Lysates, Immunoprecipitation, and Western Blot Analyses-Cells were serum starved in starvation medium (300.19 medium containing 1% bovine serum albumin without serum). Cells were then treated with DAMGO for the designated times and concentrations for 5 min, and were then lysed in Lysis Buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 2 mM EDTA, complete mini protease inhibitors, phosphatase inhibitor mixture II (Calbiochem, La Jolla, CA) and 2% Triton X-100). After centrifugation at 13,000 rpm for 15 min, the protein concentration of the supernatants was determined using the BCA protein assay reagent (Thermo Fisher Scientific, Pierce). Supernatants of the cell lysates were pre-cleared with protein G-Sepharose beads by incubation at 4°C for 2 h, followed by centrifugation at 13,000 rpm for 1 min. The supernatants were then incubated with 2 g of the indicated antibodies and 30 l of protein G-Sepharose beads overnight at 4°C. The immunoprecipitated protein complexes were resolved by SDS-PAGE. Proteins were detected using Western blot with the indicated primary antibodies and horseradish peroxidase-conjugated secondary antibodies and SuperSignal West Pico chemiluminescent substrate (Thermo Fisher Scientific, Pierce).
Isolation of Membrane Fractions-Cells treated with 1 M DAMGO, were suspended in homogenization buffer (10 mM HEPES, 50 mM mannitol, 1 mM EDTA, and protease inhibitors) at 4°C for 10 min. Cells were mechanically sheared with a 25 gauge needle. Nuclei and cell debris were removed by centrifugation at 3,000 rpm for 5 min. Membranes were collected by ultra-centrifugation at 100,000 ϫ g for 1 h at 4°C, and dissolved in Lysis Buffer for immunoprecipitation and Western blot analyses.
Flow Cytometry-M13 cells were treated with goat serum at 4°C for 30 min, washed in FACS buffer (1% BSA in PBS), resuspended in either FACS buffer only or FACS buffer containing biotinylated anti-FLAG (Sigma-Aldrich Corp.) and anti-CCR5-APC (BD Biosciences Pharmingen, Palo Alto, CA) antibodies, and incubated at 4°C for 30 min. For cells labeled with biotinylated anti-FLAG antibody, a second labeling step was carried out using streptavidin-PE (BD Biosciences Pharmingen) for 30 min at 4°C. Cells were then washed and resuspended in 400 l of FACS buffer for flow cytometric analysis using a FACS Calibur flow cytometer (BD Biosciences).
Calcium Flux-M13 cells were incubated in serum-free 300.19 Medium with 5 M Fura-2 AM (Invitrogen, Molecular Probes, Eugene, OR) for 30 min at room temperature. Calcium flux was measured with a Fluomax-3 fluorescent spectrophotometer (Horiba Jobin Yvon, Inc.) (20). The ratio of fluorescence at 340 and 380 nm was calculated using the DataMax 2.20 software (Horiba Jobin Yvon, Inc.). For heterologous desensitization experiments, the cells were first incubated with 1 M DAMGO for 5 min before addition of CCL4. For experiments with CTAP (D-Phe-Cys-Tyr-D-Trp-Arg-Thr-Pen-Thr-NH 2 , Tocris Bioscience, Ellisville, MO) and specific inhibitors, cells were pretreated for 5 min before addition of agonists.
Chemotaxis Assay-Chemotaxis was performed using a 48-well chemotaxis chamber (Neuro Probe Inc., Gaithersburg, MD) as previously described (16,17). In brief, M13 cells were resuspended in 300. 19 Binding Medium (RPMI 1640 with 1% bovine serum albumin, 0.1 mM non-essential amino acids, 1 mM sodium pyruvate, 2 mM L-glutamine, 0.5 mM ␤-mercaptoethanol, and 20 mM HEPES). Cells were pretreated with DAMGO and/or kinase inhibitors for 30 min at 37°C and loaded into the upper chemotaxis chambers with 5 m polycarbonate filter coated with fibronectin (Sigma-Aldrich). CCL4, at the designated concentrations, was loaded into the lower chambers. The chambers were incubated at 37°C for 5 h. The filters were then washed, fixed, and stained. Cells that had migrated through the filter were counted under 100ϫ magnification in 4 different fields, and results are expressed as the number of cells migrated in response to the chemokine per high power field.
Receptor Phosphorylation-Cells were incubated with phosphate-free DMEM medium at 37°C for 2 h followed by the treatment with 250 Ci/ml [ 32 P]orthophosphoric acid (H 3 PO 4, PerkinElmer Life Sciences, Boston, MA) for an additional 2 h. CCL5 or DAMGO were added to cells to a final concentration of 10 nM or 1 M, respectively, and incubated at 37°C for 25 min. In designated experiments, cells were treated with specific inhibitors for 30 min prior to addition of DAMGO. Cells were then treated with Lysis Buffer at 4°C for 15 min, and the supernatants were treated with 5 l of anti-CCR5 antibodies and 25 l of protein-G-Sepharose beads (GE Healthcare Life Science) for phospho-CCR5 immunoprecipitation. After 24 h, the affinity gel was spun down, washed twice with Lysis Buffer, and subjected to SDS-PAGE. The phosphorylated CCR5 was assessed by autoradiography.
[␥-32 P]ATP-PKC Kinase Assay-Cells treated with DAMGO were lyzed in Lysis Buffer. CCR5 was immunoprecipitated from 350 g of the protein extract with anti-CCR5 antibodies or anti-FLAG affinity gel. 100 M of ATP was then added to the immunoprecipitated protein complexes and incubated for 1 h at 4°C. The immune complexes were washed twice with Lysis Buffer and subjected to a kinase reaction analysis using 5 l 10ϫ H1 Kinase Buffer (0.5 M Tris-HCl pH 7.5, 0.1 M MgCl 2 , and 10 mM DTT), 10 g histone I protein (Sigma) and 5 Ci of [␥-32 P]ATP (Amersham Biosciences/GE Healthcare, Piscataway, NJ). The reaction mixtures were incubated at 30°C for 45 min, SDS-loading buffer was added and heated at 95°C for 5 min. The proteins were resolved using SDS-PAGE. The levels of phosphorylated histone H1 protein were evaluated using radioautography or a Cyclone TM phosphorimager (Perkin Elmer Life Sciences, Boston, MA).
Colorimetric PKC Kinase Assay-PKC kinase activity was assessed using the HTscan PKC kinase assay kit (Cell Signaling Technology Inc., Danvers, MA) with modification. In brief, cells were treated with the Kinase Lysis Buffer (20 mM Tris-HCl pH 8.0, 15% glycerol, 1% Triton X-100, 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 5 g/ml aprotinin, 5 g/ml leupeptin, 1 mM sodium orthovanadate, and 1 mM EGTA). The protein complexes immunoprecipitated as described above were used as the source of enzymes for the analysis of PKC kinase activity. Two reactions were performed for each sample in parallel: reaction A, the sample mixed with 40 M PKC pseudosubstrate inhibitor (PSI), and reaction B, without PSI. Samples were incubated at 30°C for 30 min in Reaction Buffer (25 mmol/liter Tris-HCl pH 7.5, 10 mmol/liter MgCl 2 , 5 mM ␤-glycerophosphate, 0.1 mM Na 3 VO 4 , 2 mM DTT, 200 M ATP, 1.5 M biotinylated kinase peptide substrate). Reactions were stopped with 50 mM EDTA, and reaction mixtures were then transferred to streptavidin-coated strips and incubated at room temperature for 1 h. The phosphorylated peptides were detected by incubation with rabbit anti-phosphorylated peptide antibodies, followed by incubation with HRP-labeled antirabbit antibodies, and developed with TMB substrate (R&D Systems, Minneapolis, MN). The TMB reaction was terminated by adding 3 N HCl, and the absorbance was measured using a microplate reader (Thermo Fisher) at 450 nm. The kinase activity of PKC is reported as the difference between absorbance of reaction B and A. The data presented are the average of triplicate determinations with standard deviation.
DNA Constructs-The PKC and PDK1 coding sequences were amplified by polymerase chain reaction using either PKC-GFP plasmid (Clontech, Mountain View, CA) or pCMV-SPORT6-PDK1 (Invitrogen, Full-Length Mammalian Gene Collection Gene ID 4778360), employing flanking primers (PKC: forward primer: 5Ј-CGGAATTCAGATCTCGAGA-TGG-3Ј and reverse 5Ј-GCTCTAGAACCGACTC CTC-3Ј; PDK1 forward: 5Ј-GGAATTCCATATGGCCAGGACCACC-3Ј and reverse 5Ј-GCTCTAGATGCACAGCGGCGTC-3Ј, respectively, the underlined sequences indicate EcoRI and XbaI restriction sites). These genes were both cloned into the pDNR-Dual donor vector using the Creator DNA cloning kit (Clontech). The PKC and PDK1 constructs were transferred from the donor vector to the acceptor vectors pLPS-ECFP or pLPS-EYFP using the Cre recombination reaction. These acceptor vectors were generated from pLPS-3Ј-EGFP (Clontech) with the substitution of EGFP with an ECFP or EYFP DNA fragment excised from pECFP-N1 or pEYFP-N1 plasmids (Invitrogen) using AgeI and NotI. The PKC-EYFP and PDK1-ECFP constructs were transfected into cells using the Nucleofector kit as described above. The CFP and YFP double-positive stable cell lines were obtained following flow cytometry cell sorting. The molecular weight of PKC and PDK1 fusion proteins were verified using Western blot with anti-PKC (Santa Cruz Biotechnologies) or PDK1 (Cell Signaling Technologies) antibodies.
Fluorescence Microscopy-For M13 cells, images were captured at room temperature using Olympus inverted fluorescent microscope 1X51 at 40ϫ using a SPOT Insight QE CCD microscope digital camera (Diagnostic Instruments, Sterling Heights, MI). Detection of CFP and YFP was performed using CFP (ET 436/20, Dichroic T455LP, Emitter ET 480/40, Chroma Technology, Rockingham VT) and FITC filters. For CMC cells, images were captured using Leica TCS SP5 confocal microscopy in a time lapse manner using 63ϫ oil objective lens with an Argon laser excitation for CFP and YFP at 458 nm and 514 nm, emission at 465-505 nm and 525-600 nM, respectively. The fluorescence intensity in CMC cells was measured using Image J 1.43u. The ratio of fluorescence intensity between cell membrane, and cytosol was calculated and plotted.
Measurement of the Interaction between PKC and PDK1 using FRET Analysis-Cells were washed twice with Hank's Buffered Salt Solution (HBSS) containing calcium and magnesium, resuspended in HBSS, and transferred to 4.5 cm disposable polymethyl-methacrylate cuvettes (VWR). The YFP emission was measured at 525 nm by the CFP excitation at 435 nm using a Fluoromax-3 fluorescence spectrophotometer with a 450 watt xenon lamp.
Statistical Analysis-Experimental procedures were performed at least three times, and results were analyzed for statistical significance using the unpaired Student's t test.

A PKC Inhibitor Blocks the DAMGO-induced Desensitization of CCR5 in Mediating Calcium Flux and Chemotaxis
Responses in M13 Cells-We previously showed that DAMGO can induce CCR5 cross-desensitization in monocytes and CMC cells. CMC cells are derived from the Chinese hamster ovary (CHO) cell line expressing fully functional MOR and Flagtagged CCR5 receptors, based on radiolabeled binding analysis and receptor-mediated activation of G protein activity (10). However, the molecular mechanism of MOR-induced desensitization of CCR5 is still unclear. To better understand this mechanism, we established the M13 cell line, a mouse pre-B 300.19 cells transfected to stably express both the human MOR and human CCR5 receptors. The reasons for using M13 cells for the mechanistic study in addition to CMC cells include: 1) the mechanism revealed in M13 cells could be more similar to immune cells than CMC cells; 2) M13 cells can be efficiently transfected and highly express the transfected MOR and other genes; 3) the availability of reagents such as RNAi and antibodies that are commercially available for murine cells; 4) M13 cells exhibit strong calcium flux and chemotaxis responses to cytokine stimuli. We performed flow cytometric analysis to examine receptor expression in M13 cells, and the results showed that virtually all (Ͼ95%) of these cells co-expressed both MOR and CCR5 (Fig. 1, A and B). The analysis also showed that only a small (Ͻ5%) portion of these cells expressed CCR5 in the absence of MOR. These receptors appeared to be fully functional based on the ability of MOR and CCR5 to induce a calcium flux response (Fig. 1, C and D) and chemotaxis (Fig. 1E). Pretreatment with H-D-Phe-Cys-Tyr-D-Trp-Arg-Thr-Pen-Thr-NH 2 (CTAP), a MOR-selective antagonist, blocked the calcium flux induced by DAMGO, confirming that the calcium response to DAMGO was mediated by activation of MOR (Fig.  1C). To determine whether activation of MOR led to a deficit in CCR5 function, cells were treated with DAMGO for 2 min prior to the treatment with CCL4. The results show that pretreatment with 1 M DAMGO led to a reduction in the calcium flux response to CCL4 (Fig. 1D). The CCL4 induced chemotaxis of M13 was also suppressed by DAMGO (Fig. 1F). DAMGO pretreatment reduced the chemotactic response to CCL4 by 45% with 100 nM DAMGO, and by more than 60% with 1 M DAMGO. These results showed that activation of MOR induced heterologous desensitization of CCR5.
To explore the role of PKC in mediating MOR-induced desensitization of CCR5, we used a selective PKC inhibitor to assess whether PKC was required for the DAMGO-induced desensitization of CCR5 in mediating a calcium flux induced by CCL4. Cells were pretreated with a myristoylated PKC PSI (PKC positions 113-129 [SIYRRGARRWRKLYRAN]) (21,22) for 5 min prior to pretreatment with DAMGO, followed by analysis of the the calcium response to CCL4. The results show that the PKC PSI significantly reduced the DAMGO-induced inhibition of the calcium response to CCL4 ( Fig. 1G; peak relative fluorescence ratio for controls 2.82 Ϯ 0.1, compared with 1.76 Ϯ 0.07 for DAMGO-treated versus 2.31 Ϯ 0.12 for DAMGOϩPSI. This represents a calcium response relative to baseline of 0.12 Ϯ 0.06 versus 0.67 Ϯ 0.05 for DAMGO-treated versus DAMGOϩPSI (p Ͻ 0.01 for DAMGO versus DAMGOϩPSI). The PSI did not affect the CCR5 mediated calcium flux in M13 cells at the 0.5 M dose (supplemental Fig. S1). We also found the PSI reversed the ability of DAMGO pretreatment to inhibit the chemotaxis response to CCL4. The results show that at an optimal PSI concentration, 0.5 M, induced a 2-fold increase in the chemotactic response of DAMGO pretreated M13 cells to CCL4 (Fig. 1H). These results suggest that PKC plays an important role in mediating the heterologous desensitization between MOR and CCR5.
An RNAi for PKC Blocks DAMGO-induced Cross-desensitization of CCR5-To further examine the role of PKC in the heterologous desensitization pathway between MOR and CCR5, we transfected M13 cells with PKC RNAi to "knockdown" endogenous PKC protein expression. After treatment with the PKC RNAi, the PKC protein level was significantly decreased in comparison to the housekeeping protein GAPDH ( Fig. 2A). The PKC kinase activity was also reduced in these cells (Fig. 2B). The cells with reduced PKC protein expression demonstrated a greater calcium flux (Fig. 2C) and chemotaxis ( Fig. 2D) response to CCL4 after pretreatment with DAMGO in comparison to cells transfected with the scrambled RNAi or control cells without transfection. These results indicate that PKC is required for the biochemical process leading to heterologous desensitization between MOR and CCR5. We also noted a modest reduction in the activity of CCR5 in the presence of the PKC-RNAi (Fig. 2D), which may indicate a possible role for PKC in CCR5 function.
The Phosphorylation of CCR5 Induced by DAMGO Treatment Can Be Suppressed by a PKC Inhibitor-We have previously shown that MOR-induced desensitization of CCR5 is associated with phosphorylation of CCR5 (16). To examine whether DAMGO induced cross-phosphorylation of CCR5 is mediated by PKC, we first examined the serine phosphorylation state of CCR5 in DAMGO-pretreated M13 cells using a Western blotting method (23). M13 cells were treated with CCL4, CCL5, or DAMGO in the presence or absence of PSI, then immunoprecipitated with anti-CCR5 antibodies and blotted with anti-phosphoserine antibodies. We found that CCL4 or CCL5 treatment induced CCR5 serine phosphorylation, which was not substantially changed by PSI (Fig. 3A, upper  panel). In contrast, DAMGO treatment induced CCR5 serine The CCR5 and GAPDH in the cell lysates were probed as loading controls, and the results showed that they were expressed at similar levels in all treatment groups (Fig. 3A, lower panel). We also confirmed this finding using a [ 32 P]orthophosphate labeling method (Fig. 3B). Similarly, we found that DAMGO-induced phosphorylation of PKC in CMC cells (supplemental Fig. S2). These results suggest that cross-talk between MOR and CCR5 results in PKC-dependent phosphorylation of CCR5.
Phosphorylated PKC Associates with CCR5-We found that the level of phosphorylated PKC increased in M13 cells after DAMGO pretreatment (Fig. 4A, upper panel). To examine whether the phosphorylated PKC forms a complex with CCR5, we immunoprecipitated CCR5 complexes from the M13 cell lysates, then probed with monoclonal antibodies, which recognize PKC at the phosphorylated Thr-410 site. We found that the association between this phosphorylated form of PKC and CCR5 increased after DAMGO treatment (Fig. 4A, middle  panel). In addition, the phosphorylated PKC is increased in the membrane fraction, and associated with CCR5 in both M13 cells (Fig. 4B) and CMC cells (supplemental Fig. S3) after 5 min of treatment with DAMGO. These results indicated an increase in the association of CCR5 with phospho-PKC after MOR activation at the cell membrane.   DAMGO Stimulates PKC Kinase Activity-The association between PKC and CCR5 suggests that PKC is in close proximity to CCR5 and enables PKC to phosphorylate CCR5. Because the PKC phosphorylation, particularly at the Thr-410 residue site, is required for its activation (24), and we found that Thr-410 is phosphorylated in cells after DAMGO treatment, to confirm that the activity of PKC is indeed increased in cells, we further analyzed the kinase activity of PKC using a colorimetric PKC kinase assay. This assay specifically measures the kinase activity of PKC, and the absorbance correlates with PKC protein concentration, and shows a linear dose curve to recombinant PKC protein (supplemental Fig. S4). By analyzing the kinase activities of immunoprecipitated PKC, we found the kinase activity of PKC was elevated within 30 s following treatment with DAMGO in M13 cells, and gradually increased further during the 5 min analysis (Fig. 5A). Moreover, our analysis showed that the DAMGO-induced activation of PKC was dose dependent (Fig. 5B).
To test whether the PKC associated with CCR5 is also in an activated state, we analyzed the PKC kinase activity in immu-noprecipitated complexes with CCR5. After M13 cells were treated with 100 nM DAMGO for 5 min, CCR5 protein complexes were immunoprecipitated with anti-CCR5 antibody, and the PKC kinase activity was determined (Fig. 5C). We found that the kinase activity of PKC associated with CCR5 was increased, based on comparison with the anti-GFP antibody immunoprecipitation control, indicating that the CCR5 associated-PKC is activated in M13 cells after DAMGO treatment.
We further performed experiments using a [␥-32 P]ATPase kinase assay. M13 cells were treated with 1 M DAMGO for 2 min and the kinase activity of PKC in the immunoprecipitated CCR5 complexes was analyzed using histone-I as a substrate and detected by autoradiography. Here again, we found that PKC kinase activity significantly increased with DAMGO treatment for 2 min (Fig. 5F). The kinase activity was substantially suppressed by the treatment with the PKC PSI, consistent with the specificity of PKC activity. Thus, these experiments suggest that treatment with DAMGO activates PKC.
To confirm the phosphorylation findings just described describe a general regulatory mechanism of cross-talk between MOR and CCR5, we analyzed the kinase activities in CCR5 immunoprecipitated complexes from both CMC cells and primary macrophages. These primary cells express both endogenous MOR and CCR5. Here again, we found that the kinase activities of CCR5 are elevated in both CMC and primary macrophages using the colorimetric assay (Fig. 5, D and E). Collectively, the results suggest a common mechanism in which MOR activation induces PKC activation, and activated PKC is associated with CCR5.
The PDK1 Pathway Is Involved in the MOR-induced Activation of PKC-To determine the pathway involved in PKC activation, we examined whether DAMGO-induced PKC activation is dependent on PDK1, a kinase with the reported capacity to activate PKC (25). We examined the phosphorylation status of PDK1 in DAMGO-treated cells, and we found that the level of PDK1 phosphorylation was increased within 1 min following DAMGO administration (Fig. 6A). To examine the involvement of PDK1 in PKC activation, we established a PDK1-deficient M13 cell line by transfecting M13 cells with a molecular construct that yields expression of kinase-dead (KD-PDK1 which bears a mutatation in the PIF pocket (L155E). Western blot analysis (Fig. 6B) shows that KD-PDK1 is highly expressed. Moreover, the calcium flux and chemotaxis results (Fig. 6, C and D) show that the cells expressing KD-PDK1 failed to exhibit the typical DAMGO-induced cross-desensitization observed in the parental M13 cells, or M13 cells expressing empty vector.
Activation of MOR Promotes the Interaction between PKC and PDK1-To examine the molecular action of PDK1 in activating PKC upon DAMGO treatment, we measured the physical interaction of PDK1 and PKC using M13 and CMC cells, which stably express PDK1-CFP and PKC-YFP. These fusion proteins permit interaction studies using fluorescence resonance energy transfer (FRET) analysis. Cell imaging showed that PDK1-CFP ubiquitously distributes in cells. PKC-YFP predominately distributes in the cytoplasm (Fig. 7, A and D), which is similar to the distribution pattern of endogenous PDK1 and PKC (26,27). The results suggest that following the DAMGO treatment, PKC initiates translocation to the cellular membrane (Fig. 7, A and D), and the translocation is apparent as early as 10 min following opioid treatment (Fig. 7, D and E). These distribution patterns were also exhibited in transfected CMC cells (Fig. 7, D and E). Western blot analysis showed that PKC-YFP (105 kDa) or PDK1-CFP (92 kDa) were highly expressed in M13 cells with transfection of either PKC-YFP or PDK1-CFP but not in the untransfected cells or cells with transfection of CFP, or YFP (Fig. 7B). We measured the interaction between PKC-YFP and PDK1-CFP using FRET analysis. The FRET signal was substantially increased within 45 s of DAMGO treatment (Fig. 7C), while the fluorescence intensity in M13 cells with CFP and YFP expression did not show detectable elevation. These results clearly indicate that there was an enhanced interaction between PDK1 with PKC following MOR activation.
Kinase-deficient PKC Mutants Fail to Support Cross-desensitization of CCR5-We examined the ability of three kinasedeficient PKC mutants to participate in the heterologous desensitization of CCR5. We generated mutants for three critical residues for these studies including Lys-281, an essential amino acid for kinase activity in the ATP binding region of the kinase domain, Thr-410 in the activation loop and the phosphorylation site for PDK1, and Thr-560 in the turn motif which is the autophosphorylation site and its phosphorylation is also crucial for the activation (reviewed in Ref. 25). We transfected M13 cells to stably express PKC mutants K281M, T410A, and T560A. We also included the T410E mutation, since this substitution results in an enzyme which retains most of the activity of the wild-type form. Cells were also transfected with wildtype PKC or pcDNA3.0 empty vector as controls. Western blot analysis showed that these proteins were highly expressed (Fig.  8A). The kinase activities (Fig. 8B) of cells bearing PKC K281M, T410A, or T560A mutations in M13 cells were lower than those of cells without transfection or transfected with vector control, while wild type and T410E-expressing cells demonstrated kinase activities, which were greater than control. The results (Fig. 8C) showed that after DAMGO treatment, the calcium flux response of M13 cells with K281M, T410A, and T560A mutations were restored to normal (or elevated) levels in response to CCL4, while the T410E mutant and wild type expressing cells exhibited the typical reduced calcium flux responses following DAMGO pre-treatment (Fig. 8C). Consistent with these calcium flux results, DAMGO treatment of cells expressing K281M, T410A, and T560A mutants failed to result in the typical inhibition of the CCR5 chemotactic response, when compared with wild type, T410E, non-transfected, or vector transfected control cells (Fig. 8D). These results show that PKC residues Lys-281, Thr-410, and Thr-560, with established roles in the activity of this kinase, are also necessary for PKC to mediate DAMGO-induced cross-desensitization of CCR5.

DISCUSSION
In the present study, we systematically analyzed the role of PKC in the MOR-induced desensitization of CCR5. Our results demonstrate that activated MOR up-regulates the activity of PKC through PDK1. PDK1 interacts with and phosphorylates PKC, and the activated PKC then phosphorylates CCR5 through its interaction with CCR5 at the cell membrane (Fig. 9). In the present study immunoprecipitated CCR5 from lysates, or cell membrane preparations, of cells treated with DAMGO, were analyzed for PKC-specific kinase activities in the protein complexes by either a radioisotope method, or by a non-radioactive colorimetric assay. The results show that the activation of MOR induces an elevation in PKC kinase activity, which can be inhibited by a PKC specific pseudosubstrate inhibitor. Immunoprecipitation analysis indicates that the phosphorylated and activated form of PKC directly associates with CCR5. These results suggest that PKC participates in the the heterologous desensitization of CCR5 through a phosphorylation mechanism. To our best knowledge, this is the first report showing that PKC mediates cross-desensitization between GPCRs by this mechanism.
In this study, we propose a mechanism of PKC-mediated CCR5 desensitization in which MOR is activated by DAMGO, and this is followed by induction of PDK1 activation. We found that overexpression of KD-PDK1, a kinase-dead mutant of PDK1, decreases MOR-induced CCR5 desensitization. Previous studies have demonstrated that PDK1 can directly phosphorylate PKC (28). Finally, we also demonstrated a rapid association of PDK1-CFP and PKC-YFP following MOR activation, based on FRET analysis. Collectively, these results suggest a signaling pathway in which MOR induces PDK1 activation, leading to association between PDK1 and PKC, and this results in PKC activation.
As a general mechanism, our evaluation of both M13 and CMC transfected cell lines, as well as primary monocyte-derived macrophages, shows that the association between CCR5 and PKC is increased after MOR activation. The activated PKC may translocate to the cell membrane following MOR activation. PKC has previously been shown to translocate to the cellular membrane upon activation of this kinase with either ceramide in astocytes (27), or by activation of opioid receptors using met-enkephalin (a non-selective opioid agonist), in HEK293 cells (15).
The PKC family includes twelve serine-threonine kinases, which are subdivided into three groups based on structural homology. Classical PKCs (␣, ␤1, ␤2, and ␥) can be regulated by both calcium and diacylglycerol (DAG), while novel (␦, ⑀, , and ) and atypical PKCs ( and /) are calcium independent, and only the activity of novel but not atypical PKCs can be activated by DAG.
Structurally, PKC consists of a Phox-Bim1 (PB1) domain in the N terminus, a pseudosubstrate (PS) region, a C1 domain, and a Ser/Thr kinase domain in the C terminus. The kinase domain includes an ATP binding region, an activation loop, a turn motif and a hydrophobic motif. In the ATP binding region, Lys-281 is essential for kinase activity. Amino acid substitutions of Lys-281 have been used as kinase-defective dominant-negative forms of PKC. Thr-410 in the activation loop is phosphorylated by PDK1, and T410A loses enzymatic activity, while the Glu mutant T410E, which may mimick phsophorlated Thr, retains kinase activity (25,29). PKC Thr-560 in the turn motif is the autophosphorylation site, and its phosphorylation is also crucial for kinase activation (24).
Previous reports have suggested that PKC family members are involved in the cross-desensitization of CCR5 following activation of the high affinity formyl peptide receptor (FPR) in monocyte-derived immature dendritic cells (30). In these studies, W peptide, a potent agonist of FPR, induced phosphorylation of CCR5 in a PKC-dependent (staurosporine-inhibitable) manner (31). Work carried out using various partially selective chemical inhibitors of the PKCs have suggested that either  PKC␦ or PKC appear to be necessary for opioid receptor-induced cross-desensitization of CCR1 (15). In contrast, the CXCR1-induced cross-desensitization of CCR5 appears to be dependent on PKC⑀ (31). Moreover, DAMGO treatment stimulates PKC⑀ translocation to the cell membrane and induces MOR homologous desensitization in SH-SY5Y neuroblastoma cells (32). The variable use of PKC isoforms may enable the GPCRs to selectively cross-talk with other GPCRs, depending on the precise PKC family member that is induced, and the ability of the potential target receptor to serve as a substrate for that PKC member.
PKC can be involved in some of the signaling responses mediated via other GPCRs. For example, PKC is involved in the polymorphonuclear leukocyte integrin-dependent adhesion and chemotaxis in responses to formyl peptide and CXCL8 (33). Indeed, it would appear that PKC may participate in integrin activation pathways, based on studies showing that integrin-dependent signaling can result in PKC activation through the recruitment of small GTPases including Rho and cdc42 and subsequent recruitment and activation of this PKC (34). However, the results in the present report show that PKC-PSI fails to inhibit the calcium and chemotactic responses through CCR5. Taken together, these studies suggest that CCR5 may carry out signaling processes, which are less dependent on PKC than are the pathways initiated by the receptors for CXCL8 or formyl peptide.
It should be acknowledged that signal transduction pathways other than the PDK1 and PKC pathway described here, may also play a role in MOR induced CCR5 desensitization. For example, other PKCs can also activate MAP kinases through phosphorylation of the serine/threonine protein kinase c-Raf (35), and activated MAP kinases can in turn activate G proteincoupled receptor kinases (GRKs) (36). It is conceivable that GRKs may contribute to the desensitization of CCR5 by crossphosphorylation (37), which would provide a pathway for PKCmediated indirect phosphorylation of CCR5.
It has been found that adaptors or scaffold proteins are required for PKC to achieve functional specificity and plasticity in various cellular processes. For example, PKC indirectly phosphorylates the IL-4 receptor (IL-4R) by being recruited to the IL-4R complex through an interaction with the tyrosine kinase Janus kinase-1 (Jak1). Once recruited, Jak1 phosphorylates the IL-4R␣ chain, creating a docking site for the SH2 domain of Stat6, which is then recruited and phosphorylated by the receptor-bound Jak1 (38).
In controlling cell polarity, a ternary complex of PAR-2, PAR-6, and PKC or PKC is formed. Disruption of the protein complexes by overexpression of kinase-negative PKC and PAR-6 suppresses cell polarity (reviewed in Ref. 24). In addition, PKC associates with the nerve growth factor p75 receptor, and this association appears to require the participation of a scaffold protein, designated sequestosome-1 (p62) (39). The binding of PKC to p62 involves the interactions between PB1 domains within both PKC and p62. PB1 domains have also been identified in other scaffold proteins such as Partitioning Defective-6 (Par-6) and MAPK/ERK kinase (MEK)-5 (40). In this way, p62 may potentially bind to a number of proteins involved in various biological processes. Recent evidence sug-gests that p62 plays a role in inflammation, neurogenesis, and T cell differentiation, by virtue of the heterodimerization of p62 with proteins such as PKC, PKC/, receptor-Interacting protein (RIP-1), and TNF receptor-associated factor-6 (TRAF-6) (41). p62 interacts with the atypical PKCs (PKC and PKC/) but not with any of the other closely related PKC family members. However, currently it is not clear whether the process of binding of phosphorylated PKC to CCR5 may require p62, or another scaffold protein(s) or kinases to mediate receptor phosphorylation.
The activation of MOR is relevant to the situation that occurs with heroin abuse, and the consequences for CCR5 function in this setting is an important issue for further analysis. It should be pointed out that morphine, a natural heroin metabolite, exhibits agonist activity which is distinct in some respects from some other MOR-selective agonists such as DAMGO. For example, morphine fails to induce a significant degree of MOR internalization, in contrast to methadone, fentanyl, endomorphin 2, or DAMGO (42,43). On the other hand, morphine is able to activate GIRK channels via MOR at a level, which is not significantly different than methadone or DAMGO (44). These studies suggest that these agonists exhibit differential signaling activities, or biased agonism (45).
We believe our findings have implications for the possible development of PKC-based therapeutic strategies for the treatment of HIV-1 infections. We have previously shown that DAMGO-induced desensitization of CCR5 inhibits the capacity of human macrophages to be infected with R5 strains of HIV-1, in the absence of detectable internalization of CCR5 (16). Based on the present results, it appears that the PKCmediated phosphorylated form of CCR5 may be unable to function as a fully active co-receptor for this virus. Additional experiments to address the biochemical basis for the failure of macrophages to be infected with HIV-1 following MOR-induced cross-desensitization of CCR5 are currently underway in our laboratories. Moreover, given the clear role of CCR5 in inflammatory responses, and inflammatory disease conditions, we suggest that the present studies take on additional significance. The down-regulation of the functional activity of CCR5 following MOR activation, and the predominant role of PKC in mediating this process, suggest that a PKC-related therapeutic may be developed, which could be used to treat inflammatory disease states.