κ-Opioid Receptor Signals through Src and Focal Adhesion Kinase to Stimulate c-Jun N-Terminal Kinases in Transfected COS-7 Cells and Human Monocytic THP-1 Cells

Heptahelical opioid receptors can be classified into three subtypes, δ, μ, and κ, which are found on immunocytes. For instance, early pharmacological studies on murine T cell lymphoma cell line R1.1 indicated the presence of κ-opioid receptors (Lawrence and Bidlack, 1992). Furthermore, significant levels of κ- but not δ- or μ-transcripts have been detected in human peripheral blood lymphocytes and monocytes (Gavériaux et al., 1995). Indeed, opioid receptors in these cell lines act as molecular targets for their agonists to modulate the immune system. Considerable evidence has demonstrated that κ-opioids can elicit both pro- and antiinflammatory actions. The endogenous κ-opioid dynorphin induces chemotaxis of human monocytes and stimulates superoxide production and tumoricidal activity of peritoneal macrophages (reviewed in McCarthy et al., 2001). Additionally, dynorphin A dose dependently up-regulates interleukin (IL)-2 production and proliferation of rat splenocytes in response to concanavalin A (Ni et al., 1999). On the contrary, a κ-opioid receptor-selective agonist, U-50,488H, reduces the expression of proinflammatory cytokines such as IL-1 and tumor necrosis factor-α by macrophages (Belkowski et al., 1995). With regard to the regulation of cytokine and chemokine receptors, administration of U-50,488H on thymocytes significantly stimulated CCR2 expression, but inhibited the IL-7 receptor (Zhang and Rogers, 2000).

Although opioid-mediated immunomodulations are well described, the underlying signal transduction mechanisms required for these activities remain elusive. It is well known that κ-opioid receptors can stimulate mitogen-activated protein kinases (MAPKs). Indeed, MAPKs are one of the most versatile signaling kinases that modulate numerous biological responses ranging from neuronal differentiation to immune responses. In the regulation of extracellular signal-regulated protein kinases (ERKs), the κ-opioid receptor appears to utilize a Gβγ- and Ras-dependent mechanism in both C6 glioma cells (Bohn et al., 2000) and transfected COS-7 cells (Belcheva et al., 1998). With regard to the regulation of immune responses, the c-Jun N-terminal kinase (JNK), another member of the MAPK family, can be potently activated by inflammatory cytokines such as tumor necrosis factor-α and IL-1, bacterial endotoxins, and environmental stress. JNK is responsible for phosphorylating various transcription factors including c-Jun, JunB, Elk-1, activating transcription factor 2, and nuclear factor of activated T cells (NF-AT) (reviewed in Davis, 2000). Behrens et al. (2001) reported that c-Jun stimulation by JNK was essential for thymocyte apoptosis, whereas JNK-enhanced-NF-AT activity is involved in T cell proliferation. Similar to κ-opioids, a role of JNK in IL-2 production is complicated and controversial. In mixed lymphocyte populations of JNK1 or two knockout mice, a clear suppression on the level of IL-2 has been revealed (Sabapathy et al., 2001), supporting the importance of JNK in modulating IL-2 expression. However, an opposite observation has emerged wherein CD4⁺ T cells completely lacking JNK have elevated IL-2 levels (Dong et al., 2000). Recently, morphine and [D-Ala²,N-MePhe⁴,Gly-ol⁵]-enkephalin (DAMGO, a synthetic μ-opioid) have been shown to induce the apoptosis of macrophage and T lymphocytes through p38 MAPK and JNK, respectively (Singhal et al., 2001, 2002). Activation of JNK by μ- and δ-opioid receptors have been reported (Kam et al., 2003, 2004), raising the possibility that κ-opioid receptors may modulate immune functions via JNK.

Ample evidence indicates that G protein-coupled receptors (GPCRs) can transactivate the epidermal growth factor (EGF) receptor (Daub et al., 1996), as well as activate nonreceptor tyrosine kinases such as the focal adhesion kinase (FAK; Schlaepfer et al., 1999) and Src family kinases (Luttrell et al., 1996). These tyrosine kinases are implicated in the transmission of signals to the MAPK cascades. Indeed, constitutive activation of JNK has been demonstrated in Rat-2 fibroblasts expressing v-Src oncoprotein and in human embryonic kidney 293T cells expressing membrane-bound forms of FAK (Li and Smithgall, 1998; Igishi et al., 1999). Moreover, small GTPases of the Ras and Rho families appear to form an integral part of the signaling route linking GPCRs to MAPK. Members of the Rho family (Rac and Cdc42) act as upstream activators of the JNK pathway (reviewed in Davis, 2000). The activities of the small GTPases are in turn regulated by guanine nucleotide exchange factors (GEFs). Thus, a multitude of signaling molecules may relay the activation signal from GPCRs to JNK. In the present study, we explored the ability of κ-opioid receptors to stimulate JNK in human monocytic THP-1 cells and transfected COS-7 cells. Upon stimulation of κ-opioid receptors, increases in Src and FAK autophosphorylations were detected. Our study also revealed that the κ-opioid receptor required Rac and Cdc42, as well as a GEF, Son-of-sevenless (Sos), to stimulate JNK.

Materials and Methods

Materials. The cDNA of mouse κ-opioid receptor (in the pCMV6 vector) was kindly provided by Dr. G. Bell (University of Chicago, Chicago, IL). JNK-HA cDNA was provided by Dr. T. A. Voyno-Yasenetskaya (University of Illinois, Chicago, IL). The plasmids encoding Src and its dominant-negative mutant (SrcK295R/527F) as well as dominant-negative Ras (RasS17N), Rac (RacT17N), Cdc42 (Cdc42T17N), and RhoA (RhoAT19N) were obtained as previously described (Kam et al., 2003). The cDNAs encoding the p21-binding domain (PBD) from human PAK1 and Sos-Pro were kind donations from Dr. G. M. Bokoch (Scripps Research Institute, La Jolla, CA) and Dr. R. J. Lefkowitz (Duke University Medical Center, Durham, North Carolina), respectively. The HA-tagged cDNAs of FAK site-directed mutants, including FAKY397F, FAKY925F, and FAKK454R, were kindly donated by Dr. T. Hunter (Salk Institute for Biological Studies, La Jolla, CA). [γ-³²P]ATP was purchased from PerkinElmer Life and Analytical Sciences (Boston, MA). Anti-phospho-JNK, anti-JNK, and anti-phospho-Src-Y416 antibodies were obtained from Cell Signaling Technology Inc. (Beverly, MA). Antiserum against Gα_t was obtained from Transduction Laboratories (Lexington, KY). Anti-phospho-FAK (Y397), anti-FAK, and anti-Rac antibodies were purchased from Upstate Biotechnology (Lake Placid, NY), and rabbit polyclonal antibody against human Cdc42 was obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Pertussis toxin (PTX) and 12CA5 (anti-HA) antibody were purchased from List Biological Laboratories Inc. (Campbell, CA) and Roche Applied Science (Indianapolis, IN), respectively. Tyrphostin AG1478 and pyrazolopyrimidine PP1 were purchased from Calbiochem-Novabiochem Co. (La Jolla, CA). Cell culture reagents, including LipofectAMINE PLUS, were obtained from Invitrogen (Carlsbad, CA). (±)-trans-U-50,488 methanesulfonate (U-50,488H), nor-binaltorphimine dihydrochloride (nor-BNI,) and all other chemicals were purchased from Sigma-Aldrich (St. Louis, MO).

Cell Culture and Transfection. COS-7 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) fetal bovine serum, 50 units/ml penicillin, and 50 μg/ml streptomycin in a humidified atmosphere containing 5% CO₂ at 37°C. One day before the transfection, cells were seeded onto 6-well plates at a density of 3 × 10⁵ cells/well. Transfection was performed by means of LipofectAMINE PLUS reagents according to the supplier's instructions, and the transfected cells were kept in the growth medium for 36 h. Under the same physical conditions, human monocytic THP-1 cells were cultured in RPMI 1640 medium supplemented with 50 μM 2-mercaptoethanol, 10% (v/v) fetal bovine serum, 50 units/ml penicillin, and 50 μg/ml streptomycin.

In Vitro JNK Assay. Transfected COS-7 cells were serumstarved for 18 h in the presence or absence of PTX (100 ng/ml) before drug treatment. Where appropriate, additional treatments of PP1 (25 μM, 30 min) and AG1478 (500 nM, 30 min) were applied to the starved cells. Cells were then treated with the assay medium (Dulbecco's modified Eagle's medium with 20 mM HEPES) in the absence or presence of 100 nM U-50,488H for 15 min at 37°C and lysed in 500 μl of lysis buffer (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 5 mM EDTA, 40 mM NaP₂O₇, 1% Triton X-100, 1 mM dithiothreitol, 200 μMNa₃VO₄, 100 μM phenylmethylsulfonyl fluoride, 2 μg/ml leupeptin, 4 μg/ml aprotinin, and 0.7 μg/ml pepstatin) and shaken on ice for 30 min. The lysates were centrifuged at 14,000g for 5 min at 4°C. Fifty microliters of each sample were used for the detection of JNK-HA expression, and 450 μl were incubated for 1 h at 4°C with anti-HA antibody (2 μg/sample), followed by incubation with 30 μlof protein A-agarose (50% slurry) at 4°C for 1 h. The resulting immunoprecipitates were washed twice with lysis buffer and twice with kinase assay buffer (40 mM HEPES, pH 8.0, 5 mM Mg(C₂H₃O₂)₂, 1 mM EGTA, 1 mM dithiothreitol, 200 μM Na₃VO₄). Washed immunoprecipitates were resuspended in 40 μl of kinase assay buffer containing 5 μg of glutathione S-transferase (GST)-c-Jun per reaction, and the kinase reactions were initiated by the addition of 10 μl of ATP buffer (50 μM ATP with 2 μCi of [γ-³²P]ATP per sample). After a 30-min incubation at 30°C with occasional shaking, the reactions were terminated by adding 10 μl of 6× Laemmli sample buffer and boiled for 5 min, and the samples were subjected to 12% SDS-polyacrylamide gel electrophoresis (PAGE). The radioactivity incorporated into GST-c-Jun was detected by autoradiography, and the signal intensity was quantified using a PhosphorImager SI 445 (Amersham Biosciences, Sunnyvale, CA).

Western Blot. THP-1 cells were plated at 1 × 10⁶ cells/ml in serum-free media and maintained for 18 to 24 h with or without PTX (100 ng/ml). If necessary, THP-1 cells were preincubated with different inhibitors before the addition of U-50,488H. Cells were then lysed in 200 μl of lysis buffer. Supernatants of cell lysates were collected by centrifugation at 14,000g for 5 min, mixed with 40 μl of 6× sample buffer, and boiled for 5 min. Proteins of each sample (150 μg) were resolved by 12% SDS-PAGE, and then transferred to nitrocellulose membranes. Phosphorylated and total kinases (JNK, Src, and FAK) were detected by specific antibodies, followed with horseradish peroxidase-conjugated secondary antibody. Immunoblots were developed in the presence of enhanced chemiluminescence reagents, and the images detected in X-ray films were quantified by densitometric scanning using the Eagle Eye II still video system (Stratagene, La Jolla, CA).

Immunoprecipitation. Quiescent THP-1 cells (1 × 10⁶ cells/ml) were treated with 100 nM U-50,488H at different durations (Fig. 5C). The stimulation was terminated by adding cold lysis buffer (500 μl). The cell lysates were cleared by centrifugation at 14,000g for 5 min at 4°C and then were incubated with 2 μg of monoclonal anti-Src antibodies for 3 h at 4°C, followed by incubation with 30 μl of protein A-agarose (4°C for 1 h). Immunoprecipitates were washed three times with lysis buffer and collected by centrifugation. Proteins were boiled in 40 μl of 2× sample buffer, resolved on a 12% SDS-PAGE, and transferred to nitrocellulose membranes. The membranes were analyzed by Western blotting using anti-FAK antibody.

GTPase Pull-Down Assay. GTPase pull-down assay was performed as described previously (Kam et al., 2003). The cDNAs of PAK-PBD in pGEX-4T3 were expressed in Escherichia coli as a fusion protein with GST. The fusion proteins were purified from glutathione-Sepharose beads. Serum-starved THP-1 cells were lysed with 500 μl of Mg²⁺-containing lysis buffer (MLB; 25 mM HEPES, pH 7.5, 150 mM NaCl, 1% Triton X-100, 0.25% sodium deoxycholate, 10% glycerol, 25 mM NaF, 10 mM MgCl₂, 1 mM EDTA, 1 mM sodium orthovanadate, 10 μg/ml leupeptin, 10 μg/ml aprotinin). Cell lysates were centrifuged at 4°C for 10 min at 14,000g. The supernatant (50 μl) was used for detecting total Rac and Cdc42. Cell lysates (450 μl) were incubated with 10 μg of GST-PAK-PBD and 15 μl of 50% slurry of glutathione-Sepharose beads at 4°C for 60 min with constant rotation. Bound proteins were collected by centrifugation, and pellets were washed three times in Mg²⁺-containing lysis buffer and finally suspended in 2× sample buffer (40 μl). Proteins were resolved by 12% SDS-PAGE, and the bound Rac or Cdc42 were analyzed by immunoblotting using antiserum against Rac or Cdc42, respectively.

Results

Stimulation of JNK by κ-Opioid Receptor in Human Monocytic THP-1 Cells. We began our study by examining the ability of an endogenous κ-opioid receptor to stimulate JNK in human monocytic THP-1 cells. The activation of JNK was monitored by immunoblotting using an antiserum, which recognizes the p46/p54 JNK isoforms dually phosphorylated at threonine 183 and tyrosine 185. Incubation of THP-1 cells with the κ-opioid receptor-selective agonist, U-50,488H (100 nM), increased JNK phosphorylation in a time-dependent manner, with a peak at around 5 to 15 min. The JNK activity gradually declined to the basal level in the next 30 to 45 min of the agonist incubation (Fig. 1A). The dose-response relationship of U-50,488H on the JNK phosphorylation in THP-1 cells was also investigated. As shown in Fig. 1, a significant JNK stimulation was detected with as low as 10 nM U-50,488H, and the maximum was attained by around 0.1 to 1 μM. To further map the signaling pathway by which the κ-opioid receptor induced JNK stimulation, heterologous expression assays in COS-7 cells were employed as described previously (Kam et al., 2003). Suppression of intracellular signaling components can be conveniently achieved by cotransfecting their dominant-negative mutants, without affecting the expression of receptors (unpublished observations). Plasmids expressing the κ-opioid receptor and HA-tagged JNK (HA-JNK) were cotransfected into COS-7 cells, and the JNK activity was measured by an in vitro kinase assay using GST-c-Jun as a substrate. In comparison with time course studies in THP-1 cells, the kinetics of JNK activation in transfected COS-7 cells was slower and reached a maximum after 15 min of incubation of U-50,488H (Fig. 1B). Although the dose-response curve in COS-7 cells was similar to that of THP-1 cells, at least 100 nM U-50,488H was required to elicit a significant JNK activation in the transfected cells (Fig. 1B). This small discrepancy between the transfected cells and native cells might be due to their fundamental difference in compartmentalization of receptors and G proteins (Shapira et al., 2000). In all subsequent experiments, THP-1 cells and COS-7 cells were exposed to 100 nM U-50,488H for 5 and 15 min, respectively.

κ-Opioid Receptor-Induced JNK Activation Is Mainly Mediated by Gβγ Derived from G_i/G_o Protein. Opioid receptors are prototypical G_i/G_o-coupled receptors, because their cellular signals can be effectively blocked by PTX. For example, opioid-induced activation of ERK1/2 in transfected COS-7 cells is mediated via PTX-sensitive G proteins (Belcheva et al., 1998). Yet, it has been found that opioid receptors can additionally utilize numerous PTX-insensitive G proteins, such as G_z and G₁₆, to regulate downstream effectors (Lai et al., 1995; Lee et al., 1998). Since G₁₆ is implicated in JNK stimulation and its expression is restricted in hematopoietic cells, we asked whether the κ-opioid receptor stimulated JNK via PTX-sensitive or PTX-insensitive G proteins in THP-1 cells. THP-1 cells were incubated with 100 ng/ml PTX for 18 h prior to stimulation with U-50,488H. As illustrated in Fig. 2A, the JNK activation in the κ-opioid-stimulated THP-1 cells was completely abrogated by pretreatment with PTX. Likewise, a complete blockade of the JNK stimulation by PTX was also observed in COS-7 cells (Fig. 2A). These results thus suggested a preferential coupling between the κ-opioid receptor and G_i/G_o protein to regulate JNK in both cell lines. Additionally, the JNK activation in U-50,488H-treated THP-1 cells was reversed by cotreatment with the κ-opioid receptor-selective antagonist nor-BNI (10 μM; data not shown), indicating a specific involvement of κ-opioid receptor-mediated response.

It is generally believed that Gβγ dimers act as potent activators of JNK (Coso et al., 1996). The contribution of Gβγ dimers in linking the κ-opioid receptor to the JNK pathway was evaluated using the α subunit of transducin (Gα_t), which functions as a Gβγ scavenger protein. As shown in Fig. 2B, overexpression of Gα_t in COS-7 cells markedly attenuated the JNK stimulation by U-50,488H, whereas it did not affect the total amount of HA-JNK. Hence, the κ-opioid receptor regulates the JNK pathway mainly through a Gβγ-dependent mechanism.

κ-Opioid Receptor-Initiated JNK Activation Requires Src Family Tyrosine Kinases but Not EGF Receptor Transactivation. The capability of the v-Src oncoprotein to stimulate JNK has been demonstrated in Rat-2 fibroblasts (Li and Smithgall, 1998). Our recent study has also found that overexpression of wild-type Src kinase is sufficient to stimulate the JNK activity in COS-7 cells (Kam et al., 2003). Additionally, Src kinase can in turn be activated upon overexpression of Gβγ or by stimulation of lysophosphatidic acid (LPA) (Luttrell et al., 1996), suggesting that the kinase probably participates in a number of GPCR-regulated signaling pathways. We therefore investigated the role of Src family tyrosine kinases in the activation of JNK by κ-opioid receptors. In THP-1 cells, U-50,488H not only stimulated the JNK activity, but also led to the phosphorylation of c-Src kinase at residue Y416, as detected in an anti-phospho-Tyr416 c-Src immunoblot (Fig. 3A, top panel). Furthermore, the U-50,488H-increased JNK and Src activities were abolished by pretreatment with a potent inhibitor of Src family tyrosine kinases PP1 (10 μM, 30 min; Fig. 3A, bottom panel). These results thus indicate that Src family tyrosine kinases are capable of linking the κ-opioid receptor to JNK pathway.

In agreement with the results obtained in Fig. 3A, pretreatment of transfected COS-7 cells with PP1 attenuated the JNK activation in response to the application of U-50,488H (Fig. 3B). To investigate the specific contribution of Src kinase in the pathway, cDNA encoding a kinase-deficient mutant of Src was cotransfected into COS-7 cells together with the κ-opioid receptor and HA-JNK. Figure 3B depicts the inhibitory effect of kinase-inactive Src overexpression on the JNK activation by U-50,488H, thus confirming the importance of Src kinase in the pathway. Additional studies are necessary to identify whether other members of the Src family are also involved in the JNK regulation by κ-opioid receptor.

Many studies support the idea that GPCRs such as those for endothelin-1, LPA, and α-thrombin are able to activate the ERK pathway by transactivating EGF receptors (Daub et al., 1996). Similarly, the μ-opioid receptor-induced ERK activation requires a Src-dependent transactivation of EGF receptor (Belcheva et al., 2001). Therefore, we asked whether EGF receptor transactivation is required for JNK activation by the κ-opioid receptor. In control experiments, 100 ng/ml EGF treatment of THP-1 cells clearly increased JNK activity, which was suppressed in the presence of AG1478 (an inhibitor for EGF receptor; 500 nM, 30 min; Fig. 4A). The inhibitory effect of AG1478 on EGF-induced JNK activation was also observed in COS-7 cells expressing HA-JNK (data not shown). These studies confirmed that there were functional EGF receptors in both THP-1 cells and COS-7 cells, and the inhibitory effect of AG1478 on the EGF receptor was verified. As shown in Fig. 4, pretreatment of THP-1 cells and transfected COS-7 cells with AG1478 had no effect on the JNK activation following stimulation by U-50,488H. The EGF receptor is apparently not involved in the signaling route linking the κ-opioid receptor to JNK.

A Src-Binding Site Mutant of FAK Blocks the JNK Stimulation by the κ-Opioid Receptor. Similar to Src tyrosine kinases, ample evidence supports a role for integrin-linked FAK in mediating mitogenic responses induced by GPCR, including those for vasopressin, endothelin, and bombesin (reviewed in Schlaepfer et al., 1999). As membrane-bound forms of FAK have been shown to exhibit a remarkable increase in JNK activity (Igishi et al., 1999), we were prompted to examine whether FAK participates in the κ-opioid receptor-initiated JNK pathway. Several site-directed mutants of FAK, including FAKY397F, FAKY925F, and FAKK454R (Schlaepfer and Hunter, 1997), were employed. Y397 is a major site for FAK autophosphorylation and binds the SH2 domain of Src family kinases after stimulation. The bound Src subsequently phosphorylates FAK on Y925, which creates a binding site for a Grb2 SH2 domain. Once a FAK-Grb2 complex is formed, downstream signaling pathways will be initiated (reviewed in Schlaepfer et al., 1999). Moreover, the K454 residue is essential for the kinase activity of FAK (reviewed in Schlaepfer et al., 1999). Figure 5A revealed the effects of FAK site-directed mutants on the JNK activation following stimulation of COS-7 cells with U-50,488H. Compared with the JNK stimulation in vector control, only overexpression of the Src-binding site mutant FAKY397F in COS-7 cells blocked the JNK activation (Fig. 5A). Despite coexpression of kinase-inactive FAKK454R, the ability of U-50,488H to induce JNK activation was unaffected (Fig. 5A). We also found no effect of Grb2-binding site mutant FAKY925F on the JNK activity (Fig. 5A). As determined by immunodetection with anti-HA antibody, coexpression of FAK-interfering mutants did not affect the total amount of JNK. Next, we investigated whether the activated κ-opioid receptor would lead to the autophosphorylation of FAK. After THP-1 cells were challenged with U-50,488H, a time-dependent FAK phosphorylation at Tyr397 was observed and reached a maximum at 3 to 15 min (Fig. 5B, left panel). To assess the relative contribution of Gβγ and Gα_i in the receptor signaling to FAK autophosphorylation, COS-7 cells were transfected with vector, Gβ₁, Gγ₂, Gβ₁γ₂, and wild type of Gα_i2 or its constitutively activated mutant. As shown by the anti-phospho FAK Tyr397 immunoblot in Fig. 5B (right panel), overexpression of Gβ₁γ₂ was able to increase FAK autophosphorylation, but the activated mutant of Gα_i2 was ineffective, suggesting that the κ-opioid receptor promoted FAK autophosphorylation through a Gβγ-dependent pathway. Since activated FAK is known to associate with Src, we further attempted to detect the FAK-Src complex in response to the κ-opioid receptor by coimmunoprecipitation. Stimulation of THP-1 cells with U-50,488H not only contributed to FAK phosphorylation at Tyr397, but it also triggered a transient formation of FAK-Src complex (peaking at 3 min; Fig. 5C). These experiments confirmed that the κ-opioid receptor required Tyr397 autophosphorylation on FAK and probably the subsequent association with Src kinase to stimulate JNK activity.

Small GTPases and a GEF Are Necessary for κ-Opioid-Stimulated JNK Activity. Small GTPases of the Ras and Rho families are believed to represent an integral part of the signaling route linking many cell surface receptors to MAPK. In GPCR-mediated JNK pathways, members of the Rho family, Rac and Cdc42, have been proposed as upstream regulators of MAPK kinase kinase in the JNK cascade but operate downstream of nonreceptor tyrosine kinases. Yamauchi et al. (2002) have recently reported that endothelin induces JNK stimulation by a sequential mechanism through Src kinases and Rho family GTPases Rac1 and Cdc42. To clarify which GTPases participate in the signaling cascade connecting the κ-opioid receptor to JNK, dominant-negative mutants of small GTPases including RasS17N, RacT17N, Cdc42T17N, and RhoT19N were cotransfected together with the receptor and HA-JNK into COS-7 cells. The dominant-negative mutants have a high affinity for GDP and compete with endogenous GTPases for coupling to their specific GEFs. Overexpression of RacT17N and Cdc42T17N resulted in the abolition of JNK activity evoked by κ-opioid receptors. In contrast, RasS17N and RhoT19N did not affect the JNK activation (Fig. 6A). To confirm the role of Rac and Cdc42 in the signaling pathway, we measured the GTP-bound form of endogenous Rac and Cdc42 following stimulation of THP-1 cells with U-50,488H by using the GST-PAK-PBD pull-down assay (Kam et al., 2003). As illustrated in Fig. 6B, stimulation of the κ-opioid receptor by U-50,488H led to a transient activation of Rac and Cdc42, which peaked at 5 min. Taken together, Rac and Cdc42 are required for JNK activation by the κ-opioid receptor.

Stimulation of small GTPases requires the replacement of GDP with GTP, a process promoted by GEFs, and so we next explored the identity of the GEFs involved in signaling from κ-opioid receptor toward JNK. Ubiquitously expressed Sos belongs to the Dbl family of GEFs and possesses catalytic activity on Ras and Rac. Overexpression of its Dbl homology (DH) domain in COS-1 cells can trigger a Rac-dependent JNK activation (Nimnual et al., 1998). Additionally, the μ-opioid receptor requires Sos to mediate ERK activation (Hawes et al., 1998). We therefore investigated the contribution of Sos to the κ-opioid receptor-stimulated JNK activity. As the activity of endogenous Sos could be inhibited in the presence of its proline-rich motif (Sos-Pro; Hawes et al., 1998), the cDNAs encoding Sos-Pro was utilized here. As illustrated in Fig. 7, the interfering mutant Sos-Pro markedly reduced the JNK activity in transfected COS-7 cells following stimulation of the κ-opioid receptor. These findings support the idea that Sos is an important Rac-GEF for transmitting signals from the κ-opioid receptor to the JNK cascade.

Discussion

Opioid receptors are detected in a variety of immunocytes including platelets, lymphocytes, macrophages, and monocytes (reviewed in McCarthy et al., 2001), implicating a pivotal role for opioids in the modulation of the immune system. Increasing evidence has suggested that opioid-elicited immunomodulations may occur through MAPK cascades. For instance, morphine-induced stimulations of JNK and p38 MAPK are accompanied by apoptosis of T cells and macrophages, respectively, in a caspase-dependent manner (Singhal et al., 2001, 2002). Herein, we showed that exposure of human monocytic THP-1 and transfected COS-7 cells to U-50,488H effectively stimulated JNK. The key signaling intermediates in the JNK pathway in response to stimulation by the κ-opioid receptors were also delineated (Fig. 8). This pathway required Gβγ subunits that are released upon G_i/o activation to stimulate JNK. Essential involvements of Src family tyrosine, FAK, Sos, Rac, and Cdc42 in the pathway were also demonstrated.

Lai et al. (1995) overexpressed Gα_z in COS-7 cells transfected with the G_i-coupled κ-opioid receptors and observed a PTX-insensitive inhibition of adenylyl cyclase. Similar transfection studies revealed that upon coexpression with Gα₁₆, the κ-opioid receptor became capable of stimulating phospholipase C in the presence of PTX (Lee et al., 1998). These findings clearly indicate functional couplings of opioid receptors to G proteins (including G_z and G₁₆) other than G_i to regulate downstream signals. Therefore, it is reasonable expect that G₁₆ endogenously expressed in THP-1 cells could mediate κ-opioid receptor-activated JNK. However, our results did not support the participation of G₁₆ in the pathway since U-50,488H-stimuated JNK was completely suppressed by pretreating THP-1 cells with PTX (Fig. 2A). As G₁₆ and G_z are not found in COS-7 cells, the JNK activation was predictably abrogated by PTX (Fig. 2A). G_i signaling may also be predominant in other cellular responses such as κ-opioid receptor induced-lymphocyte chemotaxis (Arai et al., 1997). Upon stimulation of κ-opioid receptors, Gβγ instead of Gα_i appeared to be the primary signaling component because overexpression of Gα_t in COS-7 cells totally abolished the JNK activation (Fig. 2B and 8). These results were in agreement with an earlier report that COS-7 cells overexpressing Gβγ, but not constitutively activated Gα_i2, efficiently increased JNK activity (Coso et al., 1996). Additionally, Gβγ appears to be crucial for JNK activation by many G_i-coupled receptors, including m2 muscarinic (Coso et al., 1996) and δ-opioid (Kam et al., 2003) receptors. However, it should be noted that in certain cell types such as Rat-1 and human embryonal kidney 293 cells, a stimulatory effect of Gα_i on JNK has been demonstrated (Edamatsu et al., 1998; Yamauchi et al., 2000).

Src kinase, a downstream target of Gβγ, is generally believed to be a key component in GPCR-mediated MAPK signaling. PP1 or dominant-negative Src mutant can suppress Gβγ-dependent mitogen-activated protein kinase kinase 4 (Yamauchi et al., 1999) and JNK stimulation by endothelin (Yamauchi et al., 2002). Similarly, our results support the involvement of Src family tyrosine kinases in κ-opioid receptor-initiated JNK pathway. U-50,488H-induced Src phosphorylation at Y416 and JNK stimulation were abolished by PP1 or kinase-deficient Src (Figs. 3 and 8). To date, the precise mechanism by which GPCR stimulates Src activity has yet to be elucidated. In a reconstituted system, Src can be directly stimulated by activated Gα_i but not by Gβγ (Ma et al., 2000). However, Luttrell et al. (1996) found a potent stimulatory effect of Gβγ on Src activity in COS-7 cells, in which intracellular adaptor proteins might be involved. In κ-opioid-induced JNK activation, Gβγ is presumably more important than Gα_i because its removal by Gα_t effectively suppressed the activating signal. Apart from Src, phosphatidylinositol 3-kinase (PI3K) and the EGF receptor have been demonstrated to participate in GPCR-mediated stimulation of MAPK cascades (Lopez-Ilasaca et al., 1998; Belcheva et al., 2001). Nevertheless, pretreatment of THP-1 cells and transfected COS-7 cells with inhibitors for PI3K (wortmannin) or EGF receptor (AG1478) did not affect the κ-opioid receptor-induced JNK responses (Fig. 4; Kam et al., 2004), so their requirements in this signaling pathway were excluded.

The involvement of FAK in κ-opioid receptor signaling to JNK was also implicated by cotransfecting its site-directed mutants into COS-7 cells. We found that only the Tyr397 on FAK was essential for the JNK activation (Fig. 5A). Moreover, in response to U-50,488H, FAK was autophosphorylated on Tyr397 site, and it also formed a complex with Src (Fig. 5, B and C). These observations favor an alternative mechanism for enzymatic activation of Src, in which intramolecular interaction between its SH2 domain and a negative-regulatory Y527 residue is disrupted upon its association with Tyr397-phosphorylated FAK. Therefore, κ-opioid receptor-induced FAK phosphorylation on Tyr397 not only facilitated the formation of a complex between FAK and Src, but also increased Src activity, which in turn stimulated the JNK cascade (Fig. 8). The exact mechanism of κ-opioid receptor-induced FAK phoshosphorylation remains unclear. Nevertheless, we found an increased autophosphorylation of FAK in COS-7 cells overexpressing Gβ₁γ₂, but not activated Gα_i2 mutant (Gα_i2QL; Fig. 5C). Moreover, PP2 failed to inhibit FAK autophosphorylation in Swiss 3T3 cells treated with several GPCR agonists including vasopressin, bradykinin, endothelin, and LPA (Salazar and Rozengurt, 2001). The κ-opioid receptor presumably requires Gβγ, but not Src family tyrosine kinases to induce FAK autophosphorylation. K454 is conserved at catalytic subdomain II of all eukaryotic protein kinase and is responsible for ATP-binding and regulation of catalytic activity. Instead, K454 on FAK was not necessary for κ-opioid receptor-stimulated JNK pathway due to the failure of FAKK454R to inhibit the JNK activation (Fig. 5A). The inhibitory effect of FAKK454R was possibly compensated by another site located in the kinase catalytic domain of FAK such as Y576 and Y577. Indeed, the formation of FAK-Src complex permits Src to further phosphorylate both Y576 and Y577, resulting in the promotion of the maximal FAK catalytic activity (reviewed in Schlaepfer et al., 1999). Similar to K454R mutant, FAKY925F was without effect on the JNK activation by U-50,488H (Fig. 5A), implicating that the interaction of FAK with Grb2 was not required for the pathway. But it does not mean that Grb2 is not important for κ-opioid receptors to stimulate JNK due to the presence of an alternative mechanism for recruitment of Grb2, such as Src/Shc/Grb2 pathway. Additional studies are required to determine the functional significance of these tyrosine residues in mediating GPCR-induced JNK activation.

It was not surprising that the κ-opioid receptor-initiated JNK pathway was primarily regulated by small GTPases Rac and Cdc42, but not Ras and Rho (Fig. 6 and 8), because constitutively activated Rac and Cdc42 have higher potency to stimulate JNK than the active mutants of Ras and Rho (reviewed in Davis, 2000). It has previously been shown that dominant-negative Rac and Cdc42 can abolish Src-dependent JNK activation, whereas dominant-negative Ras is ineffective (Li and Smithgall; 1998). In the present study, we show that at least one GEF, Sos, appears to provide the linkage from the κ-opioid receptor to Rac. Recently, Innocenti et al. (2002) illustrated the importance of adaptor proteins on the GEF activity of Sos; Sos exclusively displays Rac-GEF activity upon formation of a tricomplex by association with Eps8 (a SH3 domain-containing adaptor protein) and E3b1 (a scaffold protein). Since complete suppression of the κ-opioid receptor-stimulated JNK was observed in the presence of Sos-Pro (Fig. 7), Sos-Pro can probably down-regulate the Cdc42 signal as well. Since the proline-rich domains of Sos and Vav (a Rho family GEF) can associate with the SH3 domain of Grb2 (reviewed in Whitehead et al., 1997), Sos-Pro may also bind to the upstream adaptor of the Cdc42 GEF. On the other hand, the DH domain of Sos contributes to Rac-dependent JNK stimulation (Nimnual et al., 1998) but is down-regulated by its adjacent PH domain via their intramolecular interaction (Nimnual et al., 1998). Inhibition on the DH domain can be relieved by the binding of PI3K products (PI-3,4,5-P3 and PI-3,4-P3; Nimnual et al., 1998) to the PH domain, suggesting that PI3K can regulate Sos activity. However, κ-opioid receptor did not require PI3K to stimulate JNK (Kam et al., 2004), and so the receptors presumably use a mechanism independent of the PI3K pathway to regulate Sos. In this regard, an association of Gβ₁γ₂ with the PH-Sos (Sawai et al., 1999) is a possible means for regulating Sos.

In conclusion, we have demonstrated that JNK activation could be induced in monocytic THP-1 cells and transfected COS-7 cells following stimulation of κ-opioid receptors. The putative signaling pathway was also defined, in which κ-opioid receptors utilized Gβγ released from PTX-sensitive G_i/G_o to invoke JNK cascade. Additionally, the pathway appeared to be mediated by several signaling intermediates including Src, FAK, Sos, Rac, and Cdc42. The contribution of opioids in the regulation of inflammatory responses, ranging from chemotaxis to cytokine production, has already been established (reviewed in McCarthy et al., 2001). Given that migration of immunocytes necessitates rearrangement of actin cytoskeleton, which may be correlated with the Rac/Cdc42-dependent JNK pathway, it remains to be determined whether the demonstrated ability of κ-opioid receptors to stimulate JNK in THP-1 cells is important for monocytic chemotaxis.