Rac and Cdc42-dependent regulation of c-Jun N-terminal kinases by the δ-opioid receptor

Authors

  • Angel Y. F. Kam,

    1. Department of Biochemistry, the Molecular Neuroscience Center and the Biotechnology Research Institute, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong
    Search for more papers by this author
  • Anthony S. L. Chan,

    1. Department of Biochemistry, the Molecular Neuroscience Center and the Biotechnology Research Institute, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong
    Search for more papers by this author
  • Yung H. Wong

    1. Department of Biochemistry, the Molecular Neuroscience Center and the Biotechnology Research Institute, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong
    Search for more papers by this author

Address correspondence and reprint requests to Yung H. Wong, Department of Biochemistry, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China. E-mail: boyung@ust.hk

Abstract

Heptahelical opioid receptors utilize Gi proteins to regulate a multitude of effectors including the classical adenylyl cyclases and the more recently discovered mitogen-activated protein kinases (MAPKs). The c-Jun NH2-terminal kinases (JNKs) belong to one of three subgroups of MAPKs. In NG108-15 neuroblastoma × glioma hybrid cells that endogenously express δ-opioid receptors, δ-agonist dose-dependently stimulated JNK activity in a pertussis toxin-sensitive manner. By using COS-7 cells transiently transfected with the cDNAs of δ-opioid receptor and hemagglutinin (HA)-tagged JNK, we delineated the signaling components involved in this pathway. Sequestration of Gβγ subunits by transducin suppressed the opioid-induced JNK activity. The possible involvement of the small GTPases was also examined. Expression of dominant negative mutants of Rac and Cdc42 blocked the opioid-induced JNK activation, and a partial inhibition was observed in the presence of the dominant negative mutant of Ras. In contrast, the dominant negative mutant of Rho did not affect the opioid-induced JNK activation. In addition, the receptor-mediated JNK activation was dependent on Src family tyrosine kinases, but independent of phosphatidylinositol-3 kinase and EGF receptor tyrosine kinases. Collectively, these results demonstrate functional regulation of JNK by the δ-opioid receptor, and this pathway requires Gβγ, Src kinases and the small GTPases Rac and Cdc42.

Abbreviations used
ADP

adenosine diphosphate

ATP

adenosine triphosphate

DMEM

Dulbecco's modified Eagle's medium

DPDPE

[D-Pen2,D-Pen5] enkephalin

EGF

epidermal growth factor

ERK

extracellular signal-regulated protein kinases

G protein

guanine nucleotide-binding regulatory protein

GDP

guanosine 5′-diphosphate

GTP

guanosine 5′-triphosphate

HA

hemagglutinin

HEPES

4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid

JNK

c-Jun N-terminal kinase

LPA

lysophosphatidic acid

MAPK

mitogen-activated protein kinase

MLB

Mg2+-containing lysis buffer

PAK

p21-activated kinase

PBD

p21-binding domain

PI3K

phosphatidylinositol-3 kinase

PTX

pertussis toxin

SDS–PAGE

sodium dodecyl sulfate polyacrylamide gel electrophoresis

Opioids and narcotic opiates are highly effective in relieving severe pain. They also affect a number of physiological functions such as hormone secretions, neurotransmitter release, feeding, gastrointestinal motility, respiratory depression, and immunosuppression. These biological effects are mediated via opioid receptors that are classified into three major types (µ, δ, and κ). In mammals, the G protein-coupled opioid receptors are mainly expressed in the CNS, but can also be found in other peripheral tissues and cells such as the lymphocytes. The pharmacological features of opioid receptors have been well characterized at both cellular and molecular level. Opioid receptors inhibit adenylyl cyclase and Ca2+ channels, but stimulate K+ channels and mitogen-activated protein kinases (MAPK) through pertussis toxin (PTX)-sensitive Gi/Go proteins (reviewed in Law et al. 2000). MAPKs are serine/threonine kinases, capable of phosphorylating various transcription factors and so regulating transcription events. The ability of opioid receptors to stimulate MAPK activities provides the molecular basis for opioid actions on cell proliferation and differentiation.

MAPKs are grouped into three subfamilies including the extracellular signal-regulated kinases (ERKs), c-Jun NH2-terminal kinases (JNKs), and p38. It has been demonstrated that the δ-opioid receptors expressed in Rat-1 fibroblasts and human embryonic kidney (HEK) 293 cells stimulate ERK1/2 activity in a Gi-dependent manner (Burt et al. 1996; Tso et al. 2000). Zhang et al. (1999) also reported p38 and ERK activation by endogenous δ-opioid and opioid receptor-like (ORL1) receptors in NG108-15 neuroblastoma × glioma hybrid cells. As deltorphin augmented accumulation of AP-1 complexes and the activity of NF-AT transcription factors in Jurkat T cells (Hedin et al. 1997), it seems likely that activation of opioid receptors may lead to the stimulation of JNK activity.

JNKs are efficiently activated by inflammatory cytokines, environmental stress, and protein synthesis inhibitors (Kyriakis et al. 1994), and they are involved in cell cycle arrest, apoptosis, and differentiation. JNKs are widely expressed in the nervous system. In prior studies on PC-12 pheochromocytoma cells, withdrawal of nerve growth factor activates JNK and causes apoptosis (Xia et al. 1995), supporting the role of JNK in neuronal apoptosis. Opioids are also involved in the inhibition of neuronal growth, based on earlier reports. Opioid peptides inhibit proliferation of astrocytes, and promote their morphological differentiation (Stiene-Martin et al. 1991). Activation of µ-opioid receptors in the astrocyte inhibits DNA synthesis through a Ca2+-dependent mechanism (Hauser et al. 1996). Because JNK has been implicated in suppressing neuronal growth, it may play a role in opioid-induced antiproliferation, inhibition of DNA synthesis, and differentiation. The demonstrated ability of the δ-opioid receptor to regulate ERK and p38 lends further support to this notion. However, little is known with regard to the activation of JNK by opioid receptors.

In terms of regulation of ERK, opioid receptors appear to utilize Gβγ subunits and Ras (Belcheva et al. 1998), and may involve protein kinase C (PKC) and tyrosine protein kinases (Fukuda et al. 1996). Several studies have revealed that Gβγ can also stimulate JNK activity (Lopez-Ilasaca et al. 1998; Yamauchi et al. 1999). Gβγ is known to directly activate the γ-isoform of phosphatidylinositol 3-kinase (PI3Kγ), which in turn increases the activities of Src family tyrosine kinases (Lopez-Ilasaca et al. 1997). Therefore, both Gβγ and PI3Kγ are candidates in mediating the Gi-regulated signaling toward JNK. In addition, small guanosine 5′-triphosphate (GTP)ases of the Rho family including Rac and Cdc42 are connected to JNK via intermediates of MAPK kinase kinase (MAPKKK/MEKK) 1–4, and MAPK kinase (MKK/MEK) 4/7. One direct target of activated Rac and Cdc42 is the p21-activated kinase (PAK), which functions upstream of MEKK and may be involved in the JNK cascade (Robinson and Cobb 1997). Extensive and complicated cross-talks between G proteins and MAPKs are known to exist (Lowes et al. 2002) and they could provide multiple routes for opioid receptor to regulate JNK.

In this report we explored the possible regulation of JNK activity by the δ-opioid receptor. The effects of δ-selective ligands on JNK activity were examined in NG108-15 cells that endogenously express the δ-opioid receptor. Further characterizations of the opioid-induced JNK activity were conducted with heterologous expression assays in COS-7 cells. Participations of Gβγ, Src family tyrosine kinases, and small GTP-binding proteins in opioid-induced activation of JNK were delineated.

Materials and methods

Materials

The cDNAs of rat δ-opioid receptor (in the pCMV6 vector) and JNK-HA (hemagglutinin) were kindly provided by Christopher Evans (Department of Psychiatry and Biobehavioral Science, University of California, Los Angeles, CA, USA) and Tatyana A. Voyno-Yasenetskaya (University of Illinois, Chicago, IL, USA), respectively. The cDNAs of wild-type and dominant negative mutant of PI3Kγ (PI3KγK832R in the pcDNA3 vector) were donated by Dr Matthias P. Wymann (University of Fribourg, Switerland). The cDNAs of wild-type and dominant negative mutant of Src kinase (SrcK295R/527F) were obtained from Dr S. Lin (Hong Kong University of Science and Technology, Hong Kong, China). The plasmids encoding the dominant negative mutants of Ras (RasS17N) and Rac (RacT17N) were generous gifts from Eric J. Stanbridge (University of California, Irvine, CA, USA). Plasmids encoding the dominant negative mutants of Cdc42 (Cdc42T17N) and RhoA (RhoAT19N) were kindly provided by Marc Symons (Picower Institute for Medical Research, New York, NY, USA). [γ-32P]ATP (adenosine triphosphate) was purchased from DuPont-NEN (Boston, MA, USA). Anti-phospho-JNK and anti-JNK antibodies were obtained from New England BioLabs (Beverly, MA, USA). Antiserum against Gαt was from Transduction Laboratories (Lexington, KY, USA). PAK-1 PBD glutathione agarose (GST-PAK-PBD) and mouse monoclonal antibody against human Rac were purchased from Upstate Biotechnology (Lake Placid, NY, USA). Rabbit polyclonal antibody against human Cdc42 was from Santa Cruz Biotechnology (Santa Cruz, CA, USA). PTX and 12CA5 (anti-HA) antibody were purchased from List Biological Laboratories (Campbell, CA, USA) and Roche Molecular Biochemicals (Indianapolis, IN, USA), respectively. [D-Pen2,D-Pen5] enkephalin (DPDPE) and naltrindole hydrochloride were from Research Biochemicals Inc. (Natrick, MA, USA). Wortmannin, AG1478, and radicicol were purchased from Calbiochem-Novabiochem Co. (La Jolla, CA, USA). Cell culture reagents, including LipofectAMINE PLUS, were obtained from Life Technologies (Gaithersburg, MD, USA), and all other chemicals were purchased from Sigma (St Louis, MO, USA).

Cell culture and transfection

COS-7 cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (v/v) fetal calf serum, 50 units/mL penicillin and 50 µg/mL streptomycin in 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 following the supplier's instructions, and then the transfected cells were kept in the growth medium for 36 h. In the case of NG108-15 cells, they were cultured in the same growth medium and conditions, except that fatal calf serum was reduced to 5%.

In vitro JNK assay

COS-7 cells were serum starved for 18 h in the presence or absence of PTX (100 ng/mL) before drug treatment. In case of the inhibitors’ sensitivity assays, additional treatments of wortmannin (100 nm, 15 min), AG1478 (500 nm, 30 min), and radicicol (10 µm, 3 h) were applied to the starved cells. Then, the cells were treated with the assay medium (DMEM with 20 mm HEPES) in the absence or presence of 100 nm DPDPE 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 µm Na3VO4, 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 14 000 g for 5 min at 4°C. Then, 50 µL of each sample was used for the detection of JNK-HA expression, the remaining was incubated for 1 h at 4°C with anti-HA antibody (2 µg/sample), followed by incubation with 30 µL of 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 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 30 min incubation at 30°C with occasional shakings, the reactions were terminated by adding 10 µL of 6 × sample buffer and boiling for 5 min, and the samples were subjected to 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS–PAGE). The radioactivity incorporated into GST-c-Jun was detected by autoradiogram, and the signal intensity was quantified by PhosphorImager (Molecular Dynamics 445 SI, Sunnyvale, CA, USA).

Western blot

NG108-15 cells were seeded onto 6-well plates at a density of 1.5 × 105 cells/well and were kept in the growth medium overnight. The cells were then serum starved for 18 h in the absence or presence of PTX (100 ng/mL), followed by pre-treatments with inhibitors if necessary. The cells were stimulated with DPDPE (100 nm, 15 min) and then lysed in 300 µL of lysis buffer. Supernatants were collected by centrifugation at 14 000 g for 5 min 80 µL of each sample was resolved by 12% SDS–PAGE, and then transferred to nitrocellulose membranes. The phosphorylated and total JNK were detected by the specific antibodies as mentioned under Materials, followed with horseradish peroxidase-conjugated secondary antibody. The blot was 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, USA).

PAK-PBD binding assay

GTPase pull-down assay was performed according to the manufacturer's protocol (Upstate Biotechnology). In brief, COS-7 cells were serum starved and 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 14 000 g. Then, 50 µL of the supernatant was used for detecting total Rac and Cdc42. A total of 5–10 µg of GST-PAK-PBD agarose were added to the remaining cell lysates (450 µL) and incubated at 4°C for 60 min with constant rotation. Bound proteins were collected by centrifugation and pellets were washed three times in MLB and finally suspended in 2 × Laemmli sample buffer (20 µL). Proteins were resolved by 12% SDS–PAGE and the bound Rac and Cdc42 were analyzed by immunoblotting using antiserum against Rac and Cdc42, respectively.

Results

Opioid-induced activation of JNK in NG108-15 cells

Because a previous study has demonstrated that the δ-selective agonist, DPDPE, is capable of stimulating the p38 MAPK in NG108-15 cells via endogenous δ-opioid receptors (Zhang et al. 1999), we asked if JNK could also be activated in these cells. Activation of JNK was assessed by antiphospho-JNK antiserum. As shown in Fig. 1, DPDPE treatment of NG108-15 cells significantly increased JNK activity in a time-dependent manner. The JNK activity peaked at around 15–30 min after the addition of 100 nm DPDPE, and then gradually decreased but remained elevated up to 60 min. The time course of opioid-induced JNK activation was slower than that observed for DPDPE-induced p38 activation, where maximal stimulation was attained by 1 min (Zhang et al. 1999). Using a 15-min incubation time, NG108-15 cells were challenged with increasing concentrations of DPDPE (0.1 nm to 1 µm). The dose–response curve revealed that an increase in phosphorylation of JNK was detectable with as low as 10 nm DPDPE and reached the maximum at 1 µm (Fig. 1), while the total amount of JNK was not altered (as determined by anti-JNK antiserum; Fig. 1, lowest panel).

Figure 1.

Stimulation of endogenous δ-opioid receptors in NG108-15 cells resulted in JNK activation. NG108-15 cells were serum-starved for 18 h before application of 100 nm DPDPE for the indicated periods (a), or before stimulation with different concentrations of DPDPE for 15 min (b). Phosphorylated and total JNK were detected by different specific antibodies and quantified by densitometry. *DPDPE significantly stimulated the phosphorylation of JNK (Bonferroni t-test, p < 0.05).

Opioid receptors are prototypical Gi/Go-coupled receptors and their signal responses including inhibition of adenylyl cyclase and Ca2+ channels, as well as stimulation of K+ channels and ERK are efficiently inhibited by PTX (reviewed in Law et al. 2000). PTX modifies Gi/Go proteins by adenosine diphosphate (ADP)-ribosylation and thus inhibits their coupling with the corresponding receptors. We therefore examined the role of PTX-sensitive G proteins in the δ-opioid receptor-mediated stimulation of JNK. The DPDPE-induced JNK activation was significantly inhibited in NG108-15 cells that were pre-treated with 100 ng/mL of PTX for 18 h (Fig. 2a), indicating the functional involvement of Gi/Go proteins. Moreover, the JNK response was blocked in the presence of 10 µm naltrindole hydrochloride, a specific δ-opioid antagonist, confirming the requirement of activated δ-opioid receptors.

Figure 2.

δ-Opioid receptor mediates JNK activation in a PTX-sensitive but wortmannin-insensitive manner. NG108-15 cells were serum-starved in the absence or presence of PTX (100 ng/mL) before stimulation with 100 nm DPDPE in the absence or presence of naltrindole (10 µm; NTI) for 15 min (a), or accompanied by an additional wortmannin pre-treatment (100 nm, 15 min) before the application of DPDPE (b). *DPDPE significantly stimulated the phosphorylation of JNK (Bonferroni t-test, p < 0.05). **Wortmannin pre-treatment significantly stimulated the basal JNK activity as compared to untreated cells (Bonferroni t-test, p < 0.05).

As it has been reported that JNK activation may be mediated via PI3K (Lopez-Ilasaca et al. 1997), NG108-15 cells were pre-treated with 100 nm wortmannin (a specific inhibitor of PI3K) for 15 min before stimulation by DPDPE. As shown in Fig. 2(b), the δ-opioid receptor-induced JNK phosphorylation was insensitive to wortmannin. This wortmannin insensitivity indicates a lack of involvement of PI3K in δ-opioid receptor-induced JNK activation. Interestingly, the basal level of JNK activity was enhanced upon wortmannin pre-treatment in NG108-15 cells (Fig. 2b). This result is consistent with the observation that activated Akt (a downstream target of PI3K) inhibits Rac-GTP binding, and in turn negatively regulates JNK (Kwon et al. 2000). Hence, pre-treatment of wortmannin might stimulate JNK via the inhibition of Akt.

δ-Opioid receptor-induced activation of JNK in COS-7 cells is primarily mediated via PTX-sensitive G proteins and Gβγ subunits

In order to map further the signaling pathway by which the δ-opioid receptors stimulate JNK, heterologous expression assays in COS-7 cells were utilized. We introduced plasmids encoding the δ-opioid receptor and the HA-tagged JNK into COS-7 cells. Using anti-HA antibody, the HA-JNK was immunoprecipitated from the lysates of the transfected cells. In vitro kinase activity was measured by using the recombinant GST-c-Jun as a substrate. In each experiment, the expression level of HA-JNK was determined by immunodetection to compare the transfection efficiency. As illustrated in Fig. 3(a), DPDPE-induced JNK stimulation was time-dependent and the maximal JNK activity was observed at 15 min after stimulation by the agonist. The JNK activity remained relatively high even at 30 min, but returned to the basal level after 45 min of ligand incubation. The δ-opioid receptor-stimulated JNK activity was slow and transient in comparison with opioid-induced stimulation of ERK. Belcheva et al. (1998) have shown that δ-opioid receptor-induced ERK activity peaks at 5–10 min after addition of the agonist in transfected COS-7 cells. This ERK stimulation is sustained for 30–60 min and gradually decreases in the next 4–48 h of ligand incubation. A dose–response curve was obtained from a series of the ligand concentration (0.1 nm to 1 µm) with a 15-min incubation time, the stimulatory effect of DPDPE on JNK was dependent on the agonist concentration (Fig. 3b). Phosphorylation of JNK was significantly increased even at a concentration as low as 10 nm of DPDPE and reached the maximum at 100 nm (about 2.5-fold). These results were similar to those obtained with the NG108-15 cells (Fig. 1). However, DPDPE appeared more potent in activating ERK than JNK in COS-7 cells. As low as 0.1 nm of DPDPE has been shown to induce approximate fourfold stimulation of ERK activity (Belcheva et al. 1998). Based on the time course and dose curve studies shown in Fig. 3, all subsequent experiments were carried out by incubating the cells for 15 min with 100 nm DPDPE.

Figure 3.

DPDPE-induced activation of JNK. COS-7 cells were co-transfected with the cDNAs of JNK-HA (0.5 µg) and δ-opioid receptor (0.5 µg). JNK assay was performed as described under Materials and methods. Transfected cells were incubated with DPDPE at the indicated times and doses. (a) Time course for JNK activation by 100 nm DPDPE; (b) JNK activity determined at 15 min after the addition of increasing concentrations of DPDPE. Values shown represent the mean ± SEM from four separate experiments. The phosphorylation of GST-c-Jun and the expression of JNK-HA in the cell lysates are shown on the bottom. *DPDPE significantly stimulated JNK activity (Bonferroni t-test, p < 0.05).

To confirm the involvement of Gi/Go proteins, transfected COS-7 cells were pre-treated with PTX (100 ng/mL for 18 h) before the addition of DPDPE. The increased JNK activity in response to stimulated δ-opioid receptor was completely abolished by the pre-treatment of PTX (Fig. 4). In the regulation of ERK, opioid receptors appear to utilize Gβγ subunits that are released upon the activation of PTX-sensitive G proteins to mediate the response (Belcheva et al. 1998). Moreover, Gβγ are also involved in the JNK stimulation generated by other Gi-coupled receptors such as the m2 muscarinic receptors (Coso et al. 1996). We therefore asked if Gβγ is involved in the signal transduction pathway from δ-opioid receptor to JNK. A plasmid encoding the α subunit of transducin (Gαt) was co-transfected with HA-JNK and the receptor into COS-7 cells. Because Gαt can associate efficiently with free Gβγ dimer, it can suppress Gβγ-dependent responses (Federman et al. 1992). As shown in Fig. 5, the activation of JNK by the δ-opioid receptor was clearly reduced by co-expressing Gαt, indicating the involvement of Gβγ dimer in mediating the stimulatory effect of DPDPE on JNK. The expression of Gαt in transfected COS-7 cells was confirmed by immunodetection with a Gαt-specific antiserum, and the expression of Gαt did not affect the total amount of JNK (Fig. 5, middle and lowest panels).

Figure 4.

The effect of PTX on JNK activity. COS-7 cells were co-transfected with the cDNAs of JNK-HA and δ-opioid receptor as in legend to Fig. 3. The JNK activity was determined at 15 min in the absence (□, basal) or presence of 100 nm DPDPE (▪) with or without PTX pretreatment (100 ng/mL, 18 h). Values shown represent the mean ± SEM from four separate experiments. *DPDPE significantly stimulated JNK activity (Bonferroni t-test, p < 0.05).

Figure 5.

Activation of JNK by δ-opioid receptor is decreased in the presence of βγ-scavenging proteins. COS-7 cells were co-transfected with the cDNAs of JNK-HA (0.5 µg) and δ-opioid receptor (0.5 µg), together with vector or Gαt (0.5 µg for each). The JNK activity was determined at 15 min in the absence (□, basal) or presence of 100 nm DPDPE (▪) as indicated. Data represent the mean ± SEM from five independent experiments. *DPDPE significantly stimulated JNK activity (Bonferroni t-test, p < 0.05).

Stimulation of JNK by δ-opioid receptor is independent of PI3K but dependent on Src family tyrosine kinases

Stimulation of ERK by Gβγ is mediated via PI3Kγ, which increases the activities of Src family tyrosine kinases. The activated Src induces tyrosine phosphorylation of Shc and thus permits the recruitment of Grb–Sos complex to plasma membrane, resulting in the promotion of GDP–GTP exchange on Ras, and the subsequent activation of the ERK cascade (Hawes et al. 1996; Luttrell et al. 1996; Lopez-Ilasaca et al. 1997). A recent study also showed that overexpression of Gβγ significantly increased JNK activity, and this stimulation could also be suppressed by wortmannin (Lopez-Ilasaca et al. 1998). Furthermore, Yamauchi et al. (1999) have reported that inhibitions of Src family tyrosine kinases can block Gβγ-induced MKK4 activation and subsequently inhibit JNK activity.

These observations prompted us to investigate whether PI3K and Src family tyrosine kinases are involved in transducing signals from the δ-opioid receptor to JNK. COS-7 cells overexpressing δ-opioid receptor and HA-JNK were pre-incubated with 100 nm wortmannin for 15 min before stimulation by DPDPE. Similar to the results obtained with NG108-15 cells (Fig. 2b), wortmannin had no demonstrable effect on the JNK activation by the δ-opioid receptor in COS-7 cells (Fig. 6a). Unlike in NG108-15 cells, wortmannin did not affect the basal activity of JNK in COS-7 cells. Moreover, a dominant negative mutant of PI3Kγ was utilized to assess the involvement of PI3Kγ in the δ-opioid receptor-mediated JNK stimulation. Stoyanova et al. (1997) have reported that substitution of Arg for Lys-832 in PI3Kγ completely abrogates the lipid kinase and protein kinase activities. Consistent with the above observation, PI3KγK832R mutant did not affect the JNK activity upon stimulation of δ-opioid receptor (Fig. 6b). Therefore, PI3K does not contribute to δ-opioid receptor-mediated JNK activation. In separate control experiments, we were able to demonstrate that wortmannin effectively inhibited PI3K and Akt phosphorylation under identical experimental conditions (data not shown).

Figure 6.

Lack of involvement of PI3Kγ in δ-opioid-induced stimulation of JNK. (a) COS-7 cells were co-transfected with the cDNAs of δ-opioid receptor and JNK-HA. The transfectants were incubated for 15 min with DPDPE (100 nm) with or without wortmannin (100 nm, 15 min) pre-treatment. (b) The δ-opioid receptor and JNK-HA were co-expressed with wild-type (WT) or dominant negative mutant (DN) of PI3Kγ in COS-7 cells (0.5 µg for each plasmid). The JNK activity was determined at 15 min in the absence (□, basal) or presence of 100 nm DPDPE (▪). Results are the mean ± SEM from five separate experiments. *DPDPE significantly stimulated JNK activity (Bonferroni t-test, p < 0.05).

Next, we used radicicol (an inhibitor of Src family tyrosine kinases) to explore the involvement of endogenous Src kinases in this pathway. Radicicol can reverse the transformed phenotype of src-transformed cells and has been reported to inhibit the phosphorylation and protein kinase activity of pp60v–src (Kwon et al. 1992a, 1992b). However, it does not inhibit the serine/threonine kinases, such as MAPKK or MAPK in vitro (Kwon et al. 1995). Transfected COS-7 cells were treated with radicicol (200 ng/mL) and then stimulated with DPDPE. Radicicol significantly attenuated the δ-opioid receptor-induced JNK activation (Fig. 7a).

Figure 7.

Suppression of δ-opioid-induced JNK activation by radicicol and dominant negative mutant of Src kinase. (a) The δ-opioid receptor and JNK-HA were co-expressed in COS-7 cells. The transfected cells, with or without radicicol pre-treatment (200 ng/mL, 3 h), were incubated in the absence (□, basal) or presence of DPDPE (▪, 100 nm) for 15 min. (b) Plasmids encoding δ-opioid receptor and JNK-HA were co-transfected into COS-7 cells, together with either vector, wild-type (WT) or dominant negative mutant (DN) of Src kinase. The JNK activity was determined at 15 min in the absence or presence of 100 nm DPDPE (▪). Results are the mean ± SEM from five separate experiments. *DPDPE significantly stimulated JNK activity (Bonferroni t-test, p < 0.05).

This inhibitory effect on JNK activity was also obtained with another Src kinase inhibitor, PP1 (10 μm, 3 h treatment; data not shown). As shown in Fig. 7(b), upon co-expression of dominant negative mutant of Src with δ-opioid receptor, the JNK activity in response to DPDPE was also blocked. Collectively, these findings demonstrate that Src family tyrosine kinases are able to provide a relay from δ-opioid receptors to JNK. Noticeably, there was an increase in basal JNK activity in COS-7 cells co-expressing the wild-type of Src kinase as compared to the vector control, suggesting that the overexpression of Src kinases can facilitate the spontaneous activation of JNK.

δ-Opioid receptor stimulates JNK activity through the small GTP-binding proteins Rac and Cdc42, but not Rho

Opioid receptor-mediated activation of ERK involves Gβγ dimers and is Ras-dependent (Belcheva et al. 1998). Although transient expression of v-ras in COS-7 cells efficiently stimulates ERK, it only weakly stimulates JNK. In contrast, two members of the Rho family, Rac and Cdc42, in their activated forms can potently stimulate JNK but not ERK (Coso et al. 1995). These observations suggest that Ras mediates the signaling from various stimuli to ERK, while members of the Rho family of small GTP-binding proteins regulate JNK activity. We attempted to identify the endogenous small GTP-binding proteins that participate in linking the δ-opioid receptor to JNK by co-expressing dominant negative mutants including RasS17N, RacT17N, Cdc42T17N, or RhoT19N. These mutants have high affinity for GDP and inhibit the function of endogenous small G proteins presumably by blocking their access to exchange factors. As shown in Fig. 8(a), JNK activation by the δ-opioid receptor was completely blocked by RacT17N and Cdc42T17N but not by RhoT17N. Ras might also contribute to the DPDPE-induced activation of JNK because the DPDPE-induced response was attenuated in the presence of RasS17N (Fig. 8a). Hence, Rac and Cdc42 appear to be major intermediates in the pathway of JNK activation elicited by the δ-opioid receptor. In the case of co-expression of RacT17N and Cdc42T17N, basal JNK activity was also suppressed, suggesting that if highly expressed, these dominant negative mutants might be sufficient to inhibit the downstream effectors themselves. Preliminary experiments suggest that co-expressions of the wild-type of these four GTPases do not significantly affect basal- or agonist-induced JNK activity. To determine whether δ-opioid receptor stimulated Rac or Cdc42, the GST-PAK-PBD binding assay was used essentially as described previously (Bagrodia et al. 1995). As PAK specifically binds to the activated, GTP-bound Rac and Cdc42 via interaction with its p21-binding domain (PBD; Manser et al. 1994), activation of both GTPases can be assessed by their bindings to the GST-PAK fusion proteins. Stimulation of δ-opioid receptor time-dependently activated binding of endogenous Rac and Cdc42 to GST-PAK-PBD, peaking at around 15–30 min (Fig. 8b). In order to demonstrate the functional integrity of GST-PAK-PBD, GTPγS was added to the cell lysate to induce binding of Rac or Cdc42 to the fusion protein (Fig. 8b).

Figure 8.

δ-Opioid receptor-stimulated JNK is dependent on Rac and Cdc42. (a) COS-7 cells were transfected with the cDNAs of JNK-HA and δ-opioid receptor in the absence or presence of RasS17N, RacT17N, N17Cdc42 or N17Rho (0.5 µg for each). The JNK activity was determined at 15 min in the absence (□, basal) or presence of 100 nm DPDPE (▪). Values represent the mean ± SEM from seven independent experiments. *DPDPE significantly stimulated JNK activity (Bonferroni t-test, p < 0.05). (b) COS-7 cells expressing δ-opioid receptors were stimulated by 100 nm DPDPE for the indicated times, and binding of activated Rac-GTP and Cdc42-GTP to GST-PAK-PBD was determined by western blotting. Positive control was loaded with GTPγS as indicated. Then, 50 µL cell lysates were probed for total Rac or total Cdc42.

δ-Opioid receptor-mediated JNK activation does not involve EGF receptor

Daub et al. (1996) have demonstrated that in Rat 1 cells, endothelin-1, lysophosphatidic acid (LPA), and α-thrombin receptors stimulate epidermal growth factor (EGF) receptor/HER2 phosphorylation and ERK activity, which could be significantly attenuated upon inhibition of EGF receptor, by either an EGF receptor-selective tyrphostin, AG1478, or expression of a dominant negative mutant of EGF receptor. Presumably, the EGF receptor mediated the signal transduction pathway of the G protein-coupled receptors to ERK. Thus, we performed experiments to assess the possible involvement of EGF receptor in JNK activation by the δ-opioid receptor. AG1478 failed to inhibit the JNK activation in response to the addition of DPDPE (Fig. 9), excluding an involvement of EGF receptor in δ-opioid receptor signaling toward JNK. In control experiments, EGF (100 ng/mL, 30 min) elicited a threefold increase in JNK activity in COS-7 cells, but this JNK stimulation was diminished almost to the basal level after AG1478 pre-treatment (data not shown).

Figure 9.

Effect of EGF receptor inhibitors on δ-opioid receptor signaling toward JNK. The JNK-HA and δ-opioid receptor were expressed together in COS-7 cells. The transfected cells were stimulated for 15 min with DPDPE (100 nm) in the presence or absence of AG1478 pre-treatment (500 nm, 30 min). Results are the mean ± SEM from four separate experiments. *DPDPE significantly increased JNK activity (Bonferroni t-test, p < 0.05). □, Basal; ▪, DPDPE.

Discussion

As opioid receptors can regulate neural development and synaptic plasticity, while MAPKs are involved in growth, differentiation, and even apoptotic events in neurons, it seems plausible that JNK is involved in mediating opioid signals. Indeed, the present study demonstrated that in NG108-15 cells and transfected COS-7 cells, the δ-opioid receptor could stimulate JNK activity in a PTX-sensitive manner. By using heterologous expression assays in COS-7 cells, we further showed that the δ-opioid receptor-induced JNK activation was mediated by Gβγ dimers and involved the participation of Rac and Cdc42 GTPases as well as a Src family tyrosine kinase. On the contrary, Rho, PI3Kγ, and EGF receptors seemingly were not important intermediates in the pathway.

The involvement of PTX-sensitive Gi/Go proteins in the δ-opioid-induced stimulation of JNK in both NG108-15 cells and the transfected COS-7 cells (Figs 2 and 4) was not surprising, because the opioid receptors are classical Gi-coupled receptors. However, it should be noted that the opioid receptors could interact with a variety of PTX-insensitive G proteins, such as Gz (Chan et al. 1995) and G16 (Offermanns and Simon 1995; Lee et al. 1998) to inhibit adenylyl cyclase and stimulate phospholipase C, respectively. Presumably, opioid receptors may also act through PTX-insensitive pathways to regulate JNK activity. This presumption is supported by a recent study where the activated ORL1 receptor can functionally interact with PTX-insensitive Gz, G12, G14, and G16 to increase JNK activity in COS-7 cells, despite the inactivation of Gi/Go proteins in these cells by PTX (Chan and Wong 2000). Furthermore, the ORL1 receptor-induced JNK activation is just partially inhibited by PTX in both NG108-15 cells and transfected COS-7 cells, implicating a potential difference between the δ-opioid and ORL1 receptors.

Under the current G protein signaling paradigm, both the α-subunit and the Gβγ dimers of activated G proteins could regulate effectors. In terms of regulation of JNK, Gβγ appears to be more effective than Gαi subunit in activating JNK (Yamauchi et al. 2000). This is consistent with our results where co-transfection of Gαt almost completely abolished the JNK activation by δ-opioid receptor in COS-7 cells (Fig. 5), implicating an essential role for Gβγ as a link between the δ-opioid receptor and JNK. As a downstream effector of Gβγ, PI3Kγ has been shown to activate JNK in COS-7 cells (Lopez-Ilasaca et al. 1997), and thus represents a possible candidate linking the δ-opioid receptor to the activation of JNK. In COS-7 cells, Gβγ-induced JNK activation is effectively blocked by wortmannin and partially suppressed by co-transfecting a dominant negative mutant of PI3Kγ (Lopez-Ilasaca et al. 1998). However, an opposite perspective on the role of PI3K in mitogenic signaling has indicated that pre-treatment of wortmannin and LY294002 failed to inhibit JNK activation by mastoparan in HEK 293 cells (Yamauchi et al. 2000). Moreover, Akt (a downstream target of PI3K) has been implicated as a negative regulator for JNK activity (Kwon et al. 2000). Our findings also did not support the functional role of PI3K on the δ-opioid receptor-mediated JNK activation due to its insensitivity to wortmannin and PI3KγK832R (Figs 2 and 6). This discrepancy may be caused by differences in receptors and cell types. Cell type specificity might also account for the elevation of basal JNK activity in NG108-15 cells, but not in COS-7 cells, following inhibition of PI3K by wortmannin (Figs 2b and 6a). Presumably, Akt constitutively suppresses the basal JNK activity in NG108-15 cells and this suppression could be relieved upon wortmannin inhibition of PI3K. Nevertheless, DPDPE treatment still elicited a significant JNK activation after wortmannin treatment (Fig. 2), suggesting that the signaling pathways from the stimulatory δ-opioid receptor and inhibitory Akt to JNK might occur independently.

The Gi-coupled receptors can activate Src family kinases such as Src, Fyn, Yes, and Lyn in a PTX-sensitive manner, suggesting that Gi proteins can diversify signals through Src tyrosine kinases (Chen et al. 1994; Ptasznik et al. 1995). Luttrell et al. (1996) have shown that LPA and Gβγ mediate Shc tyrosine phosphorylation and then activate the Ras-MAPK pathway through Src kinases. Several studies subsequently reported the involvement of other non-receptor tyrosine kinases in G protein-coupled receptor-induced ERK activation, including Fyn, Lyn, and Yes (Ptasznik et al. 1995; Wan et al. 1996). We found that Src family tyrosine kinases might indeed link δ-opioid receptors to the activation of JNK (Fig. 7B). These results suggest a model of JNK activation by δ-opioid receptors. The stimulation of δ-opioid receptor induces release of Gβγ, resulting in the activation of Src family tyrosine kinases. Once activated, Src kinases promote tyrosine phosphorylation of some adaptor proteins and the recruitment of specific guanine nucleotide exchange factors, followed by activation of JNK cascade. On the other hand, a putative member of the focal adhesion kinase family, Pyk is associated with c-Src (Dikic et al. 1996) and is activated upon stimulation of κ-opioid receptors in C6 glioma cells (Bohn et al. 2000). Pyk2 has also been reported to stimulate p38 MAPK activity (Pandey et al. 1999), so the importance of Pyk2 in the δ-opioid receptor signaling toward JNK should be considered.

Experiments utilizing dominant negative mutants of Ras and Rho family GTPases presented here provide further insight into the molecular mechanism of JNK activation induced by the δ-opioid receptor. The regulatory effect of the δ-opioid receptor on JNK activation was significantly inhibited by dominant negatives of Rac and Cdc42, but only partially reduced by dominant negative Ras. The dominant negative Rho did not affect the JNK activation (Fig. 7a). GST-PAK-PBD pull-down assays (Fig. 7b) suggest that agonist-induced stimulation of the δ-opioid receptor can indeed activate Rac and Cdc42. These results reveal that the major pathway from δ-opioid receptors to JNK requires Rac or Cdc42, and these two small GTP-binding proteins are likely to have an overlapping function in the regulation of JNK. These data are consistent with the earlier reports that constitutively activated mutants of Rac and Cdc42 effectively stimulate JNK activity, whereas the activated form of Rho poorly activates JNK. Moreover, the dominant negative mutants of Rac and Cdc42 significantly inhibit both EGF and TNFα-induced JNK stimulation, suggesting that both Rac and Cdc42 are critical intermediates in the signaling of growth factors toward JNK (Coso et al. 1995).

Several studies have implicated Src family tyrosine kinase as an upstream regulator of Ras and Rho family small GTPases in the G protein-coupled receptor signaling to MAPKs. For example, in HEK 293 cells, m1 muscarinic acetylcholine receptor-induced p38, MKK3, and MKK6 activations are diminished by treatment with PP1 (an inhibitor preferential for Src family tyrosine kinase; Hanke et al. 1996). In addition, activations of p38 and MKK3 by v-Src are suppressed by co-expression of dominant negative mutants of Rac and Cdc42 (Yamauchi et al. 2001). One possible candidate linking v-Src to Rho family small GTP-binding proteins is guanine-nucleotide exchange factors (GEFs). This idea is supported by the observation that Vav, a Rho family GEF specific for Rac and Cdc42, requires its tyrosine phosphorylation to promote the exchange of GDP to GTP on the small G proteins. Vav can be activated directly by Src family tyrosine kinases, Lck (Han et al. 1997). However, Vav is restrictively expressed in hematopoietic and trophoblast cells, where it is essential for T cell development and activation. Nevertheless, other GEFs specific for Rac and Cdc42 might exist in other cell types and might serve as targets of Src kinases.

A ligand-independent transactivation of EGF receptor by G protein-coupled receptors has been recognized as an important event in mitogenic signaling (Daub et al. 1996). Our experiments did not support the involvement of EGF receptor in δ-opioid receptor-mediated activation of JNK. AG1478 had no demonstrable effect on the opioid-induced JNK activation (Fig. 8) despite the fact that signaling via the EGF receptor was suppressed. This observation indicates that transactivation of the EGF receptor is not necessary for the stimulation of JNK by δ-opioid receptors, and that different G protein-coupled receptors may utilize disparate pathways to regulate JNK. Taken together, the present study provides a putative model in which δ-opioid receptor stimulates JNK activity via Gβγ, Src family tyrosine kinase, Rac, and Cdc42. Further studies are required to identify which GEFs are specific for transmitting signals from opioid receptors to MAPKs and to characterize how the receptors regulate the GEFs and the Rho family small GTPases.

Acknowledgements

We are extremely grateful to the following individuals for kindly providing the cDNAs: C. Evans for rat δ-opioid receptor, T. Voyno-Yasenetskaya for the JNK-HA, Dr S. Lin for wild-type and dominant negative mutant of Src, Dr Matthias P. Wymann for wild-type and dominant negative mutant of PI3Kγ, E. J. Stanbridge for RasS17N and RacT17N, M. Symons for Cdc42T17N and RhoT17N. We also thank Dr M. K. C. Ho for invaluable discussions. This work was supported in part by grants from the University Grants Committee of Hong Kong (AoE/B-15/01), the Research Grants Council of Hong Kong (HKUST 6115/00M and 2/99C) and the Hong Kong Jockey Club.

Ancillary