Abstract
The µ-opioid receptor (OPRM1) signals as a classic G protein–coupled receptor by activating heterotrimeric Gi/Go proteins resulting in adenylyl cyclase (AC) inhibition. Such AC inhibition is desensitized after prolonged agonist treatment. However, after receptor desensitization, the intracellular cAMP level remains regulated by OPRM1, as demonstrated by the intracellular cAMP level increase or AC superactivation upon removal of an agonist or addition of an antagonist. We now demonstrate that such intracellular cAMP regulation is mediated by a novel noncanonical signaling pathway resulting from OPRM1 being converted to a receptor tyrosine kinase (RTK)-like entity. This noncanonical OPRM1 signaling is initiated by the receptor recruiting and activating Src kinase within the receptor complex, leading to phosphorylation of the OPRM1 Tyr336 residue. Phospho-Tyr336 serves as the docking site for growth factor receptor-bound protein/son of sevenless, leading to the recruitment and activation of the Ras/Raf-1 and subsequent phosphorylation and activation of AC5/6 by Raf-1. Such sequence of events was established by the absence of Ras/Raf1 recruitment and activation by the OPRM1-Y336F mutant, by the presence of Src kinase inhibitor 4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d]pyrimidine (PP2) or the absence of Src activity, by the presence of specific Raf-1 inhibitor GW5074 (5-iodo-3-[(3,5-dibromo-4-hydroxyphenyl) methylene]-2-indolinone) or the absence of Raf-1, or by the dominant negative RasN17 mutant. Src together with Ras activates Raf1 which was established by the inability of the Raf1-Tyr340/341 mutant to activate AC. Hence, the phosphorylation of OPRM1 at Tyr336 by Src serves as the trigger for the conversion of a classic Gi/Go-coupled receptor into an RTK-like entity, resulting in a noncanonical pathway even after the original Gi/Go signals are blunted.
Introduction
In G protein–coupled receptor (GPCR) canonical signaling, agonist stimulation leads to a rapid exchange of GDP for GTP on the α subunit of the heterotrimeric G proteins and the subsequent dissociation of α and βγ subunits. Both α and βγ subunits are involved in the activation of different effectors, including adenylyl cyclase (AC), phospholipase C, and ion channels (Pierce and Lefkowitz, 2001). However, this linear model does not adequately address the full range of GPCR stimulation. There is evidence to suggest that the overall cellular responses depend on GPCR-mediated stimulation of a large number of complex signaling cascades. Therefore, it is not surprising that cross-regulation occurs in the GPCR signaling pathways. For example, β-arrestins have been demonstrated to not only blunt the GPCR signal but also serve as signal transducers for the same GPCR. As observed with the signaling of β2-adrenergic receptors, β-arrestins mediate agonist-induced receptor desensitization and internalization, and also mediate the recruitment of c-Src to the receptor, which facilitates the activation of mitogen-activated protein kinase/extracellular signal-regulated kinase (MAPK/ERK) cascades (Luttrell et al., 1999a; McDonald et al., 2000; Lefkowitz and Shenoy, 2005). Cross-regulation can also occur by the modulation of either upstream or downstream events of other receptors’ signaling pathways on GPCR activation, which results in even more complicated cellular responses. An example for such cross-regulation is the growth-promoting effects of many GPCRs that are mediated through transactivation of the epidermal growth factor receptor (EGFR), which elicits cellular responses mainly through its intrinsic tyrosine kinase activity, as in the case of bradykinin receptors activation in PC12 cells leading to transactivation of EGFR and the activation of the Ras/MAP kinase pathway (Zwick et al., 1997).
In the case of opioid receptors, activation of the receptors results in the inhibition of AC activity, activation of MAPK and Kir3 K+ channels, inhibition of voltage-dependent Ca2+ channels, and modulation of other effectors such as phospholipase C (Law et al., 2000). However, prolonged agonist exposure not only blunts these cellular responses but also results in a compensatory increase in the intracellular cAMP level, which is particularly significant upon the removal of an agonist or the addition of an antagonist such as naloxone (Sharma et al., 1977; Law et al., 1983; Pineyro and Archer-Lahlou, 2007). This compensatory increase in AC activity or AC superactivation phenomenon has been postulated to be the molecular basis of drug tolerance and dependence (Koob and Bloom, 1988). However, the classic linear GPCR signaling pathway cannot account for such change from the initial receptor-mediated AC inhibition to the eventual receptor-mediated AC activation because the Gi/Go-mediated signaling is blunted after prolonged agonist treatment.
Although the exact mechanism for such signal changes is yet to be elucidated, activation of specific protein kinases and subsequent phosphorylation of AC isoforms (Avidor-Reiss et al., 1996, 1997) and other signaling molecules, such as G protein–coupled receptor kinases (GRK) 2/3 (Chakrabarti et al., 2001) have been suggested to be the key for the observed AC activation. Among all the protein kinases studied, protein kinase C (PKC), MAP kinases and Raf-1 have been implicated in the AC superactivation (Li and Chang, 1996; Varga et al., 2002; Schallmach et al., 2006). Alternative mechanisms, such as agonist-induced receptor internalization and the increase in the constitutive activities of the receptor or the switching from Gi/Go-coupled to Gs-coupled, also have been suggested to play a role in AC superactivation (Szucs et al., 2004; Walwyn et al., 2007). Regardless of the mechanism, the exact molecular events that lead to the switching of opioid receptor from an inhibitory response to a stimulatory response remain elusive.
In one of our previous studies, we demonstrated that AC superactivation required the recruitment and activation of Src and subsequent Src-mediated phosphorylation of the µ-opioid receptor (OPRM1) at the Tyr336 residue (Zhang et al., 2009). In our current study, we further detail the cellular events after Tyr336 phosphorylation that lead to AC superactivation by demonstrating that phosphorylation of OPRM1 at Tyr336 serves as the docking site for growth factor receptor-bound protein (Grb)/son of sevenless (SOS), leading to the recruitment and activation of Ras/Raf-1 and the subsequent phosphorylation and activation of AC5/6 by Raf-1, resulting in the ultimate AC superactivation. Hence, the phosphorylation of OPRM1 at Tyr336 by Src kinase serves as the switch for the conversion of a classic Gi/Go-coupled receptor into an receptor tyrosine kinase–like (RTK-like) entity, resulting in a noncanonical signal pathway even after the canonical Gi/Go signaling has been attenuated.
Materials and Methods
Cell Culture.
Human embryonic kidney 293 (HEK293) cells stably expressing hemagglutinin (HA)-OPRM1 (HA-OPRM1HEK293; HEKMT) and HEK293 cells stably expressing HA-OPRM1 with tyrosine mutants (HEKMT-Y336F) were cultured at 37°C in advanced minimum Eagle’s medium (Invitrogen/Life Technologies, Carlsbad, CA) supplemented with 5% fetal bovine serum (HyClone, Logan, UT), 2 mM GlutaMAX (GIBCO/Life Technologies, Grand Island, NY), 100 U/ml penicillin, 100 μg/ml streptomycin, and 0.1 mg/ml G418 (geneticin; GIBCO) in a 5% CO2 incubator. Mouse embryonic fibroblasts (MEF)-Raf-1 wild type and cells isolated from Raf-1 knockout mice were cultured at 37°C in advanced Dulbecco’s modified Eagle's medium (Invitrogen) supplemented with 5% fetal bovine serum, 2 mM GlutaMAX, 100 units/ml penicillin and 100 μg/ml streptomycin in a 10% CO2 incubator. HEKMT or MEF-Raf1 wild-type or MEF cells isolated from Raf-1 knockout mice were transiently transfected with various cDNA plasmids using Nucleofector transfection kits designed for HEK293 cells or for MEF cells (Lonza, Basel, Switzerland).
cAMP Assay.
Cells were seeded in 96-well plates the day before and morphine was added to the medium 4 hours before the assays. The culture medium was removed and cells were washed with medium at 37°C once and then 100 μl treatment buffer (0.5 mM isobutylmethylxanthine and 10 μM forskolin in Krebs-Ringer HEPES buffer) with different concentration of naloxone was added to each well. The cells were incubated for 15 minutes at 37°C and then the assay was terminated by placing the plate in an 85°C water bath for 6 minutes to lyse the cells and release the intracellular cAMP. The cAMP level was measured by using AlphaScreen cAMP Detection Kit (PerkinElmer Life and Analytical Sciences, Waltham, MA) and Biomek 2000 Laboratory Automation Workstation (Beckman Coulter, Brea, CA). The results were analyzed and summarized with Prism 4.0 software (GraphPad Software, San Diego, CA).
Immunoprecipitation and Western Blotting.
Cells were treated with or without morphine and/or naloxone, then the lipid raft fraction was prepared following the method described elsewhere (Zhang et al., 2009). The prepared lipid raft was incubated with mouse monoclonal anti-HA antibody (1:200; Covance, Princeton, NJ) in 1 ml of buffer A (100 mM NaCl, 10 mM Tris, pH 7.4) in the presence of 0.1% digitonin (Sigma-Aldrich, St. Louis, MO) at 4°C overnight. Proteins in the samples were resolved by 10% SDS-polyacrylamide gel electrophoresis and were electrophoretically transferred to polyvinylidene difluoride membranes (0.45 μm; Amersham Biosciences/GE Healthcare, Little Chalfont, UK). The presence of pRaf-1, pSrc, AC5/6, and OPRM1 in the SDS-PAGE-resolved immunoprecipitates was determined with anti–pRaf1-Tyr340/341 (1:500; Millipore, Billerica, MA), anti-pSrc-Tyr416 (1:500; Millipore), anti-AC5/6 (1:500; Santa Cruz Biotechnology, Santa Cruz, CA) or rabbit anti-HA (1:1000; Covance) antibody respectively. The band intensities were quantified and analyzed with the ImageQuant software (Amersham Biosciences/GE Healthcare).
Ras Activation Assay.
The Ras activity was determined by using a Ras activation assay kit (Millipore) as described by the manufacturer’s protocol. Briefly, the cells were lysed by adding 0.5 ml of MLB buffer (Mg lysis/wash buffer: 25 mM HEPES, pH 7.5, 150 mM NaCl, 1% Igepal CA-630 (octylphenoxypolyethoxyethanol), 10 mM MgCl2, 1 mM EDTA, and 10% glycerol). The solution was cleared, and the protein concentration was determined by bicinchoninic acid assay. The cell lysis solution from each treatment with same amount of protein was mixed with 5 μg of Raf-1 Ras-binding domain agarose and incubated at 4°C for 45 minutes with slow rotation. The agarose beads were pelleted down by brief centrifugation and washed 3 times with 0.5 ml of MLB (Mg lysis/wash buffer) each. The agarose beads were resuspended in 40 μl 2× Western sample loading buffer and were boiled for 5 minutes. The samples were loaded on a 12% SDS-PAGE gel, and Western blot was performed as described; the blot was probed with mouse anti-Ras antibody (1:1000; Millipore). A portion of the cell lysis solution from each sample with an equal amount of protein was loaded to another gel, and the result was used to determine the amount of input Ras.
Raf-1 Activity Assay.
Raf-1 kinase activity was determined with the assay kit supplied by Millipore. Briefly, after various treatments, cells were lysed in lysis buffer containing 1% Triton X-100, 10 mM Tris, pH 7.5, 120 mM NaCl, 25 mM KCl, 100 μM Na3VO4, and 1× complete protease inhibitor (Roche Applied Science, Indianapolis, IN). The cell lysate was cleared by centrifuging at 16,000g for 5 minutes, and the supernatant was incubated with 2 μg anti–Raf-1 polyclonal antibody (Santa Cruz Biotechnology) overnight at 4°C, followed by another 3 hours of incubation after the addition of 20 μl of protein G agarose beads. Afterward, the agarose beads were washed 3 times with buffer A. Then, the agarose beads pellets were resuspended in 20 μl assay dilution buffer [20 mM 3-(N-morpholino)propanesulfonic acid, pH 7.2, 25 mM β-glycerol-phosphate, 5 mM EDTA, 1 mM sodium orthovanadate, 1 mM dithiothreitol], 10 μl Mg/ATP cocktail (75 mM MgCl2, 500 μM ATP), and 1.6 μl (0.4 μg) inactive mitogen-activated protein kinase kinase 1 (MEK1). The mixture was incubated for 30 minutes at 30°C with shaking. The reaction was terminated by adding 10 μl 6× Western sample loading buffer and boiled for 5 minutes. The samples were loaded on a 10% SDS-PAGE gel, and the blot was probed with anti-pMEK1 (1:1000; Millipore) or anti-Raf-1 (1:500; Santa Cruz Biotechnology) antibody. The kinetics of Raf-1 kinase activation were determined by the use of [γ-32P]ATP.
The immunoprecipitation was performed in a similar manner except the agarose beads pellets were suspended in 50-μl kinase buffer (50 mM Tris-HCl, pH 7.5; 10 mM MgCl2, 10 mM MnCl2, 25 μM ATP, 1 mM dithiothreitol, and 100 μM Na3VO4) containing 5 μCi of [γ-32P]ATP (New England Nuclear/PerkinElmer Life and Analytical Sciences, Waltham, MA). The mixture was incubated at 30°C, and 10-μl aliquots were removed at the time points of 0, 5, 10, and 20 minutes. The reaction was terminated by addition of 10 μl of 40% trichloroacetic acid. The sample was then spotted onto a P81 phosphocellulose paper square (Millipore). The paper was washed extensively with 1% phosphoric acid 3 times and with acetone 1 time. The radioactivity retained on the P81 paper was quantified by liquid scintillation counting.
Reverse-Transcription Polymerase Chain Reaction.
The total RNA of cultured cells was isolated using TRI Reagent (Molecular Research Center, Cincinnati, OH) according to the manufacturer’s instructions. The RNA samples were then treated with DNase I (Ambion/Life Technologies, Austin, TX). The reverse-transcription polymerase chain reaction (RT-PCR) reaction was performed by using One Step RT-PCR Kit (Qiagen, Valencia, CA) following the manufacturer’s recommendations, and 1.0 μg of DNase I-treated total RNA was added to each reaction. The RT-PCR conditions were as follows: 30 minutes at 50°C followed by 15 minutes at 95°C and then 35 cycles of 1 minute at 94°C, 1 minute at 60°C (54°C for β-actin), and 1 minute at 72°C followed by 1 cycle of 10 minutes at 72°C. The PCR primers for Raf-1 were forward: AAGAAAGCACGCTTAGATTGG, reverse: CAAGAGCTGTCTGATATT; for OPRM1, forward: CATCAAAGCACTGATC ACGATTCC, reverse: TAGGGCAATGGAGCAGTTTCTGC; and for β-actin, forward: AGACCTCTATGCCAACACAGT, reverse: ATCTCGGTGGTTAGGTGTGTC.
Results
Our previous study suggested that during the chronic morphine treatment OPRM1 recruited and activated Src, which in turn phosphorylated the receptor at Tyr336. This is the pivotal event that eventually leads to AC superactivation (Zhang et al., 2009). However, the function of this phosphorylated Tyr336 in OPRM1 signaling leading to AC superactivation has not been addressed. Classically, phosphorylated tyrosine residue often serves as a docking site for proteins with Src homology 2 domain (SH2) leading to the recruitment of the downstream signaling proteins to the receptor’s proximity (Karoor et al., 1998; Fan et al., 2001; Tobin, 2008). Among them, the Grb/SOS complex has been shown to bind to the phosphorylated Tyr of RTK and transduce the signal to the downstream Ras/MAP pathway (Satoh et al., 1990). If the Src-mediated recruitment of Grb/SOS to the phosphorylated OPRM1 Tyr336 is the key for AC superactivation, then the disruption of such a complex formation should affect AC superactivation. When HEK293 cells stably expressing OPRM1 (HEKMT cells) were treated with the SOS1 Src homology 3 domain (SH3) domain-binding peptide, which specifically blocks the interaction of SOS1 and Grb2, the chronic morphine-induced AC superactivation was significantly blunted (155.0% ± 15.6% compared with 309.1% ± 19.6% in presence of 10−4 M naloxone; Fig. 1A). Because Grb2/SOS1 usually binds and activates Ras, we examined whether Ras is involved in the AC superactivation process. Transient expression of the dominant negative RasN17 mutant reduced morphine-induced AC superactivation to 223.6% ± 20.7% as compared with the control of 336.3% ± 25.6% in presence of 10−4 M naloxone (Fig. 1B). A similar result was obtained by pretreating the cells with the Ras inhibitor farnesylthiosalicylic acid, which decreased AC superactivation to 194.4.1% ± 21.7% (Fig. 1B).
Measurements of Ras activation revealed there was a time-dependent increase in Ras activity during chronic morphine treatment (Supplemental Fig. 1). Addition of the antagonist naloxone after chronic morphine (4 hours) further potentiated Ras activation (193.7 ± 15.8% compared with 161% ± 14.9% for 4-hour morphine without naloxone; Fig. 2A). Furthermore, pretreatment of cells with the Src kinase inhibitor PP2 (4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d]pyrimidine) or Ras inhibitor farnesylthiosalicylic acid totally abolished chronic morphine-induced Ras activation (106.0% ± 22.5% and 83.7% ± 25.1%, respectively, compared with morphine and naloxone treatment only; Fig. 2A). The ability to blunt the morphine-induced Ras activation by these inhibitors correlates with the ability of these inhibitors to block morphine-induced AC superactivation (Fig. 1B) (Zhang et al., 2009).
More importantly, Ras is observed to be within the OPRM1 signaling complex during chronic morphine treatment, as revealed by coimmunoprecipitation (co-IP) studies (166.0% ± 22.1% compared with control, Fig. 2B). Coincidently, addition of naloxone further increased the amount of Ras co-IP with OPRM1 by 271.0% ± 34.4% as compared with control (Fig. 2B). This Ras and OPRM1 interaction can be prevented by transient expression of the dominant negative RasN17 mutant (39.3% ±14.7% compared with the morphine and naloxone treatment only; Fig. 2C). As discussed earlier, phosphorylated Try336 could act as a docking site for downstream signal molecules to facilitate AC superactivation. Hence, it is not unexpected that when morphine-induced, Src-mediated Tyr-phosphorylation was not observed (Zhang et al., 2009) in HEK293 cells expressing the OPRM1 Y336F mutant (HEKMT-Y336F), Ras activation could not be detected (Fig. 2D) and co-IP of Ras with OPRM1 was not observed either (L. Zhang, unpublished observations).
Because Ras activation is often the consequence of RTK activation and transactivation of RTK by opioid receptors has been reported (Belcheva et al., 2002), whether OPRM1-mediated Ras activation and subsequent AC superactivation is the consequence of RTK transactivation was examined. HEKMT cells were incubated with the EGFR inhibitor lavendustin C at different concentrations as indicated (Fig. 3A). Even at the highest concentration of lavendustin C tested (100 nM), attenuation of AC superactivation was not observed (IC50, 12 nM). These results were confirmed by using various concentrations of another EGFR inhibitor, AG-1478 (N-(3-chlorophenyl)-6,7-dimethoxy-4-quinazolinamine; IC50, 3 nM; Fig. 3B). It was also reported that opioid receptors could transactivate fibroblast growth factor receptor (Belcheva et al., 2002); however, pretreatment of the cells with various concentrations of the fibroblast growth factor receptor inhibitor PD-173074 (N-[2-[[4-(diethylamino)butyl]amino]-6-(3,5-dimethoxyphenyl)pyrido[2, 3-d]pyrimidin-7-yl]-N-(1,1-dimethylethyl)-urea; IC50, 21.5 nM) did not result in any inhibition of chronic morphine-induced AC superactivation (Fig. 3C). Based on these observations, it is unlikely that the Ras activation observed in the presence of naloxone after chronic morphine treatment is the consequence of RTK transactivation.
Upon activation, the active Ras binds to Raf kinases with high affinity and causes Raf kinases to translocate to the cell membrane where Raf activation takes place. The involvement of Raf-1 in chronic opioid administration had been implicated based on the fact that pretreatment with the Raf-1 specific inhibitor GW5074 (5-iodo-3-[(3,5-dibromo-4-hydroxyphenyl) methylene]-2-indolinone) significantly blunted AC superactivation (Varga et al., 2002). By performing co-IP studies, we could demonstrate that activated Raf-1 (pRaf-1), activated c-Src (pSrc), and AC5/6 formed a signal complex with OPRM1 in a chronic morphine-dependent manner (264.6% ± 43.0%, 215.3% ± 22.8% and 155.6% ± 21.6%, respectively, compared with the control; Fig. 4, A and B). Pretreatment with the Raf-1 inhibitor GW5074 totally blocked the co-IP of pRaf-1 with OPRM1 but did not alter morphine-induced Src activation (124.7% ± 39.1% and 187.2% ± 22.0% respectively). However, pretreatment of Src kinase inhibitor PP2 not only eliminated pSrc from the complex but also prevented the appearance of pRaf-1 in the complex (92.5% ± 23.0% and 113.0% ± 36.7%, respectively), suggesting that the presence of pRaf-1 within the OPRM1 complex requires Src kinase activation. Both PP2 and GW5074 pretreatment also reduced the presence of AC5/6 in the OPRM1 complex in the presence of morphine and naloxone from 155.6% ± 21.6% to 82.0% ± 13.2% and 73.3% ± 6.5%, respectively (Fig. 4, A and B).
The measurements of Raf-1 kinase activity revealed that Raf-1 activity was significantly enhanced within the OPRM1 signaling complex as determined by the phosphorylation of the Raf-1 substrate MEK1 (pMEK1) after chronic morphine treatment (168.0% ± 26.7%) and after chronic morphine followed by naloxone treatment (191.0 ± 36.0%) when compared with the control (Fig. 4, C and D). A similar result was obtained from a Raf-1 kinase kinetics assay using [γ-32P]ATP to label the substrate (Fig. 4E). All these results demonstrate that the phosphorylation of OPRM1 Tyr336 is the key for recruiting Ras to the receptor complex, leading to the translocation and activation of Raf-1 to cell membranes via the binding of protein kinase to Ras. If this is the scenario, then it should follow that without Tyr336 being phosphorylated, Raf-1 translocation and activation should not occur. Indeed, in cells expressing the OPRM1 Y336F mutant, the phosphorylation of Raf-1’s substrate MEK1 was not observed after chronic morphine treatment (Fig. 5, A and B).
To further confirm the involvement of Raf-1 in chronic morphine-induced AC superactivation, MEF cells from both wild-type (WT) and Raf-1 knockout (KO) mice were used to transiently express OPRM1 and/or Raf-1. Figure 6A shows the RT-PCR results of Raf-1 levels that were similar in various transfection combinations as indicated. The cAMP assays using these transfected cells suggested that only when cells expressed both OPRM1 and Raf-1, AC superactivation was significantly potentiated (Raf-1WT + OPRM1, 194.8% ± 15.6% and Raf-1 KO + OPRM1 + Raf-1, 278.5% ± 19.7%). Other combinations with the absence of either OPRM1 or Raf-1 did not exhibit AC superactivation (Fig. 6, B and C). These MEF cells studies support the Raf-1 inhibitor GW5074 studies and clearly indicate that Raf-1 activation is critical for the observed AC superactivation.
At present, the exact mechanism for Raf-1 kinase activation remains elusive. However, the Src kinase phosphorylation at 340 and/or 341 tyrosine residues is one of the keys for Raf-1 activation (Marais et al., 1995). Because activated Raf-1 and activated Src can be observed within the OPRM1 signaling complex in co-IP experiments (Fig. 4A), it is likely that Src within the receptor complex could phosphorylate and thus activate Raf-1, leading to AC superactivation. To test this hypothesis in AC superactivation, Tyr340 or Tyr341 was mutated to Ala singly or in combination to generate the Raf1Y340A or Raf1Y341A single mutant or Raf1Y340/341AA double mutant. To mimic the phosphorylated state at these Tyr residues, they were mutated to Asp to generate the constitutively active Raf1Y340/341DD mutant. These mutants or WT Raf-1 were transiently expressed in MEF-Raf1 KO cells, and their expression levels were similar as determined by RT-PCR (Fig. 7A). To test their effect on chronic morphine induced AC superactivation, WT or mutant Raf-1 was cotransfected with OPRM1 to the MEF-Raf-1 KO cells. Both single mutant Raf-1Y340A and Raf-1Y341A and double mutant Raf1Y340/341AA significantly blunted the AC superactivation to 155.2% ± 14.0%, 177.5% ± 16.0%, and 131.4% ±12.1%, respectively, as compared with the cells expressing WT Raf-1, where AC superactivation was at 286.3 ± 21.9% above the control level. In contrast, expression of the constitutively active mutant Raf-1Y340/341DD significantly potentiated AC superactivation to 436.9% ± 29.7% above the control level (Fig. 7, B and C). These studies suggest that the phosphorylation of Raf-1 at these Tyr residues is another key for the observed AC superactivation.
Discussion
In past decades, the classic view of GPCRs as simple heterotrimeric G protein activators has gradually evolved into a concept of intracellular signaling pathways with complex modulations. Emerging evidence suggests that the efficiency, specificity, and fidelity of GPCR signaling can be achieved by a more confined and organized system, which may involve crosstalk, fine-tuning, and specific regulation at multiple levels, even within the same GPCR signal pathway. For example, β-arrestins have been shown to recruit different downstream signal molecules to control the outcome of the signal transduction in addition to their roles as signal terminators (Cao et al., 2000; Lefkowitz and Shenoy, 2005). Crosstalk can also happen between different GPCRs’ pathways, leading to distinct signal pathways that are different from the originals. For example, activation of phospholipase C–coupled P2Y2 receptors specifically inhibited β2-adrenergic receptor-mediated cAMP production via Gi proteins in pGT-β cells (Suh et al., 2001). The cross-regulations may also happen between GPCRs and other signal pathways; among them, the transactivation between GPCR and RTK has been intensively studied (Daub et al., 1996). It has been reported that stimulation of various kinds of GPCRs, including the receptors for bradykinin, angiotensin, thrombin, lysophosphatidic acid, bombesin, endothelin, and muscarinic acetylcholine, can transactivate the EGFR system (Prenzel et al., 1999; Kalmes et al., 2000; Gschwind et al., 2001; Pierce et al., 2001).
For opioid receptors, agonist-induced acute inhibition of AC activity is a classic GPCR signaling event (Zhang et al., 2006). However, after prolonged treatment, the inhibition effect is blunted; it gradually converts to a compensatory increase in intracellular cAMP level, and is particularly significant upon the removal of the agonist or the addition of an antagonist (Sharma et al., 1977; Law et al., 1982; Pineyro and Archer-Lahlou, 2007). This observed change from receptor-mediated AC inhibition to receptor-mediated AC activation cannot be explained by the classic GPCR signal paradigm and may reflect intracellular signal switching. Our previous studies suggested that Src kinase activation and subsequent Src-mediated phosphorylation of OPRM1 at Tyr336 were the keys for such signal switching. Similar GPCR phosphorylation-induced signal switching had been intensively studied in β-arrestin–mediated signaling. Lefkowitz and coworkers were the first to report that β-arrestins bound to the GRK-phosphorylated receptor were not only a signal termination event, but also the β-arrestins could serve as adaptors to recruit Src to the receptor complex, leading to Src-dependent, receptor-stimulated activation of downstream effectors such as the MAP kinases (Luttrell et al., 1999b). Because β-arrestins can recruit a broad spectrum of signaling molecules, the GRK/β-arrestin system has been demonstrated to function not solely in the receptor desensitization process but also in signaling pathways that participate in biased agonism (Lefkowitz and Shenoy, 2005).
In our current study, identifying various molecular events after chronic morphine administration has revealed a novel signal pathway for the opioid receptor. Distinct from the canonical Gi/Go-mediated receptor signaling, the phosphorylation of OPRM1 Tyr336 serves as the docking site for Grb/SOS (Supplemental Fig. 2), leading to the recruitment and activation of Ras/Raf-1 and the subsequent phosphorylation and activation of AC5/6 by Raf-1. A number of GPCRs have been shown to activate or inhibit the MAPK cascade by either a Ras-dependent or Ras-independent pathway requiring GPCR transactivation of RTKs such as EGFR (Daub et al., 1996). For example, expression of polypeptides that sequester free Gβγ subunits blocks Ras-dependent MAPK/ERK activation via the lysophosphatidic acid receptor (Herrlich et al., 1998), α2A-adrenergic receptor (van Biesen et al., 1995), M2-muscarinic receptor (Koch et al., 1994), and D2-dopamine receptor (Faure et al., 1994). Our data suggest that Ras activation occurs during chronic morphine treatment and is even more pronounced upon the addition of the antagonist naloxone, which mimics the drug withdrawal state (Fig. 2A). This Ras activation does not require the transactivation of RTKs (Fig. 3); instead, it occurs within the OPRM1 signal complex in a Src-dependent manner (Fig. 2A). Furthermore, in cells expressing the Tyr336 mutant receptor, Ras activation was not observed, nor could Ras be detected in the OPRM1 signal complex (Fig. 2D; L. Zhang, unpublished observations). Together with the observation that disruption of SOS1 and Grb2 interaction significantly blunted AC activation (Fig. 1), we conclude that phospho-Tyr336 is a docking site for the recruitment of Grab2/SOS1/Ras to OPRM1, thus forming a signal complex in which Ras is activated.
In the classic view, activated Ras will bind to Raf-1 and recruit protein kinase to the cell membrane, where Raf-1 is activated (Baccarini, 2005). Based on the observation that the Raf-1 inhibitor GW5074 attenuates AC activation, Raf-1 has been implicated in opioid receptor–mediated chronic morphine-induced AC superactivation (Varga et al., 2002). Our current study is the first to suggest that Raf-1 is recruited to the OPRM1 signal complex and activated during chronic morphine/naloxone treatment, and that this Raf-1 activation is Src kinase dependent (Fig. 4, A and B). Based on our previous studies, which indicated that OPRM1 could recruit and activate Src during chronic morphine treatment, together with the fact that Raf-1 activation is not observed in cells expressing Tyr336 mutant OPRM1 (Fig. 5), we conclude that phospho-Tyr336 recruits Raf-1 to the OPRM1 signal complex via activated Ras, where Raf-1 is activated by Src. By using MEF-Raf-1 WT and KO cells, we demonstrated Raf-1 activation as a prerequisite for AC superactivation; this activation is mediated by OPRM1 directly, not by other receptor signal pathways (Fig. 6). The currently accepted model for Raf-1 activation involves recruitment to the plasma membrane by activated Ras via the Ras-binding domain in Raf-1, dephosphorylation of the inhibitory protein kinase A (PKA)/protein kinase B site Ser259 at Raf-1 by phosphatase PP2A, and eventual phosphorylation at Tyr340 and/or Tyr341 probably by Src kinase (Marais et al., 1995; Kolch, 2002; Baccarini, 2005). Our data have shown that Src is required for Raf-1 activation (Fig. 4, A and B); mutation at Tyr340 or Tyr341 significantly attenuates AC activation, and double mutation at these sites almost totally abolishes the AC activation, suggesting that these two sites are equally important for Raf-1 activation. The observation that the constitutively active mutant Tyr340/341DD significantly enhanced AC activation further substantiates this conclusion (Fig. 7, B and C). One may expect that the constitutively active Raf-1-Tyr340/341DD mutant itself is enough to cause AC superactivation, but without the chronic morphine/naloxone treatment, Src activation and subsequent phosphorylation at Tyr336 of OPRM1 were not observed. Furthermore, without recruitment of Ras to and activation of Ras by OPRM1, Raf-1 was not recruited to the OPRM1 signal complex. Hence, constitutively active Raf-1 itself cannot lead to AC superactivation. In addition to phosphorylation and activation by Src, other protein kinases such as PKC (Kolch et al., 1993), Akt (Zimmermann and Moelling, 1999), PAK3 (King et al., 1998), and PKA (Dhillon et al., 2002) have been suggested to be involved in Raf-1 regulation either positively or negatively. Whether these protein kinases are involved in Raf-1 activation during chronic morphine administration remains to be examined.
In the RTK/Ras/MAPK cascade, activated Raf-1 will directly phosphorylate MEK1/2, leading to the activation of MAPK/ERK, which in turn can activate AC. However, during chronic morphine treatment, MAPK activation is not observed (L. Zhang, unpublished observation). Instead, several studies have suggested that the Raf-1 can phosphorylate several AC isoenzymes directly, such as AC2, AC5, and AC6 (Ding et al., 2004). Avidor-Reiss et al. (1997) reported that AC activation can be reproduced in COS-7 cells with cotransfection of AC5 and OPRM1. It also has been shown that Raf-1-mediated phosphorylation of AC6 leads to the sensitization of this isoenzyme to stimulators (Tan et al., 2001) such as forskolin (i.e., AC activation). Furthermore, mutagenesis studies from the same investigators have suggested that Ser750 on AC6 is phosphorylated by Raf-1 and that mutation of Ser750 to Ala significantly impaired AC6 activity (Ding et al., 2004). Our current data clearly indicate that AC5/6 was recruited to the OPRM1 signal complex during chronic morphine treatment in a Src- and Raf-1–dependent manner (Fig. 4, A and B). Because other protein kinases such PKA and PKC and also some regulatory proteins such as RGS2 (Roy et al., 2006), SNAPIN (SNAP-25/Snare complex) (Chou et al., 2004), and PAM (protein associated with myc) (Gao and Patel, 2005) have been suggested to be capable of modulating the AC5/6 activity, chronic morphine-induced AC activation could involve multiple protein kinases in addition to Src and Raf-1 and the regulatory proteins as discussed earlier.
In conclusion, our current study provides a new model for OPRM1-mediated chronic morphine-induced AC superactivation. The phosphorylation at Tyr336 by Src kinase serves as a docking site to recruit Grb/SOS/Ras and Raf-1. This chronic agonist-induced, site-specific phosphorylation results in a distinct signaling outcome. As a result, OPRM1 is converted from a classic Gi/Go-coupled receptor into an RTK-like entity, resulting in a noncanonical signaling pathway even when canonical Gi/Go signaling has been attenuated (summarized in Fig. 8). It is noteworthy that a similar model has been suggested for RTK signaling in which RTKs exert some of their effects by engagement of GPCR signaling molecules, including heterotrimeric G proteins and β-arrestins, suggesting that RTKs can also act as GPCR-like receptors (Luttrell et al., 1990; Liang and Garrison, 1991; Dalle et al., 2001; Povsic et al., 2003; Cao et al., 2009). Because chronic agonist-induced AC superactivation has been implicated in drug addiction processes, if the Src-mediated Tyr336 phosphorylation can be prevented while still maintaining the overall cellular Src activity, then a specific treatment paradigm for opioid drug addiction might be attainable.
Acknowledgments
The authors thank Dr. J. P. Stork (Vollum Institute, Portland, OR) for RasN17, Dr. J. F. Hancock (University of Texas, Houston) for Raf-1 and Raf-1 340/341DD cDNA constructs, and Dr. M. Baccarini (University of Vienna, Austria) for MEF-Raf1 wild-type and knockout cells.
Authorship Contributions
Participated in research design: Zhang, Law.
Conducted experiments: Zhang.
Performed data analysis: Zhang, Loh, Law.
Wrote or contributed to the writing of the manuscript: Zhang, Law.
Footnotes
- Received July 3, 2013.
- Accepted September 23, 2013.
This work was supported by National Institutes of Health National Institute on Drug Abuse [Grants DA00056439 and DA01180615].
↵This article has supplemental material available at molpharm.aspetjournals.org.
Abbreviations
- AC
- adenylyl cyclase
- AG-1478
- N-(3-chlorophenyl)-6,7-dimethoxy-4-quinazolinamine
- co-IP
- coimmunoprecipitation
- EGFR
- epidermal growth factor receptor
- ERK
- extracellular signal-regulated kinase
- GPCR
- G protein–coupled receptor
- Grb
- growth factor receptor-bound protein
- GRK
- G protein–coupled receptor kinase
- GW5074
- 5-iodo-3-[(3,5-dibromo-4-hydroxyphenyl) methylene]-2-indolinone
- HA
- hemagglutinin
- HEK293
- human embryonic kidney 293 cells
- HEKMT
- HEK293 cells stably expressing HA-OPRM1
- HEKMT-Y336F
- HEK293 cells stably expressing HA-OPRM1 with Tyr336 mutant
- Igepal CA-630
- octylphenoxypolyethoxyethanol
- KO
- knockout
- MAPK
- mitogen-activated protein kinase
- MEF
- mouse embryonic fibroblasts
- MEK1
- mitogen-activated protein kinase kinase 1
- OPRM1
- µ-opioid receptor
- PD-173074
- N-[2-[[4-(diethylamino)butyl]amino]-6-(3,5-dimethoxyphenyl)pyrido[2, 3-d]pyrimidin-7-yl]-N-(1,1-dimethylethyl)-urea
- PKA
- protein kinase A
- PKC
- protein kinase C
- PP2
- 4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d]pyrimidine
- RT-PCR
- reverse-transcription polymerase chain reaction
- RTK
- receptor tyrosine kinase
- SH2/3 domain
- Src homology 2/3 domain
- SOS
- son of sevenless
- WT
- wild-type
- Copyright © 2013 by The American Society for Pharmacology and Experimental Therapeutics