Distinct opioid receptor agonists have been proved to induce differential patterns of ERK activation, but the underlying mechanisms remain unclear. Here, we report that Ser363 in the δ-opioid receptor (δOR) determines the different abilities of the δOR agonists DPDPE and TIPP to activate ERK by G-protein- or β-arrestin-dependent pathways. Although both DPDPE and TIPP activated ERK1/2, they showed different temporal, spatial and desensitization patterns of ERK activation. We show that that DPDPE employed G protein as the primary mediator to activate the ERK cascade in an Src-dependent manner, whereas TIPP mainly adopted a β-arrestin1/2-mediated pathway. Moreover, we found that DPDPE gained the capacity to adopt the β-arrestin1/2-mediated pathway upon Ser363 mutation, accompanied by the same pattern of ERK activation as that induced by TIPP. Additionally, we found that TIPP- but not DPDPE-activated ERK could phosphorylate G-protein-coupled receptor kinase-2 and β-arrestin1. However, such functional differences of ERK disappeared with the mutation of Ser363. Therefore, the present study reveals a crucial role for Ser363 in agonist-specific regulation of ERK activation patterns and functions.
Opioid receptors belong to the superfamily of G-protein-coupled receptors (GPCRs), which couple to the Gi/o proteins. Like many other GPCRs, opioid receptors undergo internalization and desensitization following agonist exposure, manifested by a decrease in the receptor number at the plasma membrane and loss of the receptor functions. Interestingly, the desensitization and internalization of opioid receptors are agonist-selective. For example, although both morphine and etorphine are agonists of opioid receptors, they exhibit remarkable differences in the internalization and desensitization of μ-opioid receptor (μOR) and δ-opioid receptor (δOR).
It is now widely accepted that extracellular signal-regulated kinases 1 and 2 (ERK1/2) are downstream effectors of GPCRs, including δOR. Recently, agonist-specific desensitization of ERK signaling pathway has been reported for the δOR (Eisinger et al., 2002; Hong et al., 2009). In addition, differential kinetic patterns of ERK activation by individual δOR agonists have also been observed (Eisinger and Schulz, 2004; Audet et al., 2005). For example, the duration of ERK activated by morphine and etorphine is quite different. ERK1/2 activated by etorphine is quite transient, whereas ERK1/2 activated by morphine is quite persistent (Eisinger and Schulz, 2004). Moreover, such different patterns of ERK activation induced by morphine and etorphine have been linked to their ability to differently desensitize and internalize δOR (Eisinger and Schulz, 2004; Eisinger et al., 2002; Audet et al., 2005). However, the molecular mechanisms underlying this differential regulation of ERK activation by distinct agonists remain unclear.
Accumulating evidence shows that GPCRs can activate ERK through multiple mechanisms, including G-protein-dependent and G-protein-independent pathways. In the G-protein-dependent pathway, GPCRs can initiate a Gβγ-triggered activation of component kinases of ERK cascade in a phosphoinositide 3-kinase (PI3K)-dependent manner (Hawes et al., 1996; Lopez-Ilasaca et al., 1997). Moreover, the μ-opioid receptor was found to activate ERK in a manner dependent on EGFR transactivation (Belcheva et al., 2001). Recently, increasing evidence suggests that β-arrestin2 can function as a scaffold for component kinases of the ERK cascade to activate ERK in a G-protein-independent mechanism (DeFea et al., 2000; Luttrell et al., 2001; Lefkowitz and Shenoy, 2005; Zheng et al., 2008) (supplementary material Fig. S1). ERK activated by the G-protein-dependent pathway is transient, and undergoes nucleus translocation. However, ERK activated by the β-arrestin-dependent pathway is more persistent, and distributes in the cytoplasm (Lefkowitz and Shenoy, 2005; Ahn et al., 2004), though sometimes also in the nucleus (Zheng et al., 2008).
The findings that ERK can be activated by G-protein- or β-arrestin-dependent pathways imply the existence of distinct active configurations of GPCRs responsible for the two pathways, which can be stabilized by unique ligands. It has been shown that for the β2 adrenergic receptor (β2AR) and angiotensin type 1A receptor (AT1aR), the mutation of several residues in the C-terminal can shift these two receptors to activate ERK from a G-protein-dependent mechanism to a β-arrestin-dependent mechanism (Wei et al., 2003; Shenoy et al., 2006). This suggests that certain residues in the C-terminal of GPCRs might be crucial for the formation of distinct active receptor configurations in response to distinct agonist stimulations.
Another crucial mechanism of receptor regulation is phosphorylation. It has been shown that δOR phosphorylation confers its selectivity for β-arrestin2, and subsequently regulates receptor internalization and adenylyl cyclase desensitization (Qiu et al., 2007). For the δOR, Ser363 is a primary phosphorylation site targeted by G-protein-coupled receptor kinase-2 (GRK2). The phosphorylation of residue Ser363 in the C-terminal tail of δOR plays an important role in the regulation of receptor trafficking, desensitization and internalization (Trapaidze et al., 1996; Kouhen et al., 2000). Our recent research has found differential phosphorylation of Ser363 upon stimulation by distinct agonists DPDPE [(D-Pen2, D-Pen5)enkephalin)] and TIPP (H-Tyr-Tic-Phe-Phe-OH) (Hong et al., 2009). The phosphorylation, however, did not contribute to differential receptor desensitization (supplementary material Fig. S2). It has also been reported that differential temporal patterns of ERK activation can be detected upon activation of the δORs by distinct agonists (Eisinger et al., 2002; Eisinger and Schulz, 2004). However, the role of Ser363 in the C-terminal for the δOR adoption of distinct pathways to activate ERK remains to be elucidated.
DPDPE and TIPP are both selective ligands for δOR. Our recent study has found that although both ligands can activate ERK1/2, only DPDPE treatment induces the desensitization of ERK signaling (Hong et al., 2009). Our preliminary study also found that DPDPE and TIPP showed distinct kinetic and spatial patterns of ERK activation. This study, therefore, was undertaken to investigate whether G-protein- and β-arrestin-dependent pathways are differently involved in mediating DPDPE- and TIPP-stimulated ERK activation, to determine the effects of ERK activated by distinct mechanisms on the phosphorylation of GRK2 and β-arrestin1, and to explore the role of the Ser363 residue for the δOR employment of distinct pathways to activate ERK in response to DPDPE and TIPP stimulations.
Differential duration and subcellular distribution of DPDPE- and TIPP-mediated ERK1/2 activation
Our previous and recent studies have shown that TIPP might represent a novel class of δOR agonists with unique properties. To determine whether it also differs from the classic δOR agonist DPDPE in regulation of ERK1/2 signaling, we examined whether the stimulation of δOR with DPDPE and TIPP could result in differential patterns of ERK1/2 activation. We checked the effects of these agonists in CHO cells stably coexpressing δOR (CHOδOR cells) and in NG108-15 cells endogenously expressing δOR. First, we determined the time course of ERK activation by DPDPE and TIPP. CHOδOR and NG108-15 cells were treated with saturating concentrations (1 μM) of each drug for the times shown in Fig. 1, and then phospho-ERK1/2 was measured by western blot analysis of cell lysates using phospho-specific anti-ERK1/2 antibody. Differential durations were observed for phospho-ERK1/2 stimulated by DPDPE and TIPP (Fig. 1A–D). TIPP-stimulated activation was more sustained and showed no declination over 60 minutes, whereas DPDPE-stimulated activation was transient, and declined rapidly.
Because DPDPE exhibits greater ability to activate ERK1/2 relative to TIPP, we next determined whether such distinct ERK activation durations of DPDPE and TIPP were due to their differential ability to activate ERK. To this end, we tested the duration of ERK activation induced by 1 nM DPDPE and 1 μM TIPP. These concentrations were chosen because 1nm DPDPE produced a similar level of ERK activation as 1 μM TIPP (see supplementary material Fig. S3). The duration of ERK phosphorylation was not changed after adjustment of the DPDPE dose (Fig. 1E,F).
We next used confocal laser microscopy to examine the subcellular distribution of phospho-ERK1/2 stimulated by DPDPE and TIPP in CHO cells that stably coexpressed δOR. As shown in Fig. 2A (top row), in the absence of agonists there was minimal staining of phospho-ERK1/2 in the cytoplasm and nucleus Treatment with 5 μM DPDPE for 5 minutes induced a sharp rise in phospho-ERK1/2 immunoreactivity, and the staining of phospho-ERK1/2 was predominantly localized in the nucleus (Fig. 2A, middle row). TIPP treatment also led to striking enhancement of phospho-ERK1/2 immunoreactivity, but the staining of phospho-ERK1/2 was mainly distributed in cytoplasm (Fig. 2A, bottom row). This differential subcellular distribution of phospho-ERK1/2 activated by DPDPE and TIPP was further validated using the Cellomics phospho-ERK1/2 nucleus translocation assay. Stimulation with increasing doses of DPDPE resulted in a significantly greater phospho-ERK1/2 nucleus translocation than found with TIPP (Fig. 2B,C). Taken together, these results suggest that differential mechanisms might be involved in DPDPE- and TIPP-stimulated ERK1/2 activation.
Differential desensitization of DPDPE- and TIPP-mediated ERK1/2 activation
An intriguing feature of δOR-mediated ERK signaling is the differential desensitization properties following activation by distinct agonists (Eisinger et al., 2002; Eisinger and Schulz, 2004; Hong et al., 2009). To further confirm the distinct regulation of ERK1/2 activity by DPDPE and TIPP, we also examined the effects of either DPDPE or TIPP pretreatment on subsequent ERK1/2 activation by a second DPDPE stimulus, in CHO cells that stably coexpressed δORs. Pretreatment of cells with 1 μM DPDPE for 4 hours led to robust attenuation of subsequent ERK1/2 activation by a second DPDPE stimulus (Fig. 3A,B). By contrast, the ability of DPDPE to activate ERK1/2 remained unaffected after pretreatment with 1 μM TIPP for 4 hours. Similar results were obtained in NG108-15 cells that endogenously express δOR (Fig. 3C,D). These results indicate that DPDPE, but not TIPP, treatment desensitizes stimulation of ERK1/2 by DPDPE. Similar results were obtained in NG108-15 cells treated with 1 nM DPDPE and 1 μM TIPP (Fig. 3E,F).
DPDPE, but not TIPP, induced association of Gβγ with PLCβ3 and c-Src, which was inhibited by expression of S363A mutant δOR
Differential temporal and spatial patterns of ERK activation by DPDPE and TIPP suggest that G-protein- and β-arrestin-dependent mechanisms might be differentially involved in DPDPE and TIPP-stimulated ERK activation. To verify this hypothesis, we tested the early signals that represent the G-protein-dependent pathway. It has been established that upon the activation of GPCRs, dissociation of Gβγ subunits can directly regulate a variety of effectors, including phospholipases (Murthy and Makhlouf, 1996), and c-Src (Gentili et al., 2006) and can function as the primary mediator of Ras activation in an Src-dependent manner and of subsequent signaling via mitogen-activated protein (MAP) kinases (Belcheva et al., 1998; Koch et al., 1994). Therefore, we next determined whether differential associations of Gβγ subunits with phospholipases and c-Src occurred in response to DPDPE and TIPP stimulation in CHOδOR cells. As shown in Fig. 4, minor associations of Gβγ with phosphoinositide-specific phospholipase Cβ3 (PLCβ3) and c-Src were found in the absence of drug stimulation. However, stimulation with DPDPE led to a robust but transient increase in the associations of Gβγ with PLCβ3 and c-Src. By contrast, stimulation with TIPP failed to induce a significant association of Gβγ with either PLCβ3 or c-Src, indicating that DPDPE induces formation of the complex including Gβγ, PLCβ3 and c-Src, whereas TIPP does not.
GRK2-mediated phosphorylation of the Ser363 residue in the C-terminal tail of δOR is important for agonist-mediated receptor internalization and desensitization downstream of the G-protein-mediated pathway (Trapaidze et al., 1996; Kouhen et al., 2000). Our recent study also found that DPDPE, but not TIPP, could induce the phosphorylation of Ser363 in the δOR (Hong et al., 2009). To determine whether the Ser363 residue in δOR is a crucial site in the G-protein-dependent signaling pathway, we tested whether DPDPE could induce the associations of Gβγ with PLCβ3 and c-Src in CHO cells transiently transfected with the mutant δOR in which Ser363 was replaced by alanine (S363A). As shown in Fig. 4, this mutation significantly inhibited the DPDPE-induced associations of Gβγ with PLCβ3 and c-Src. The results indicate that the Ser363 in the C-terminal tail of δOR is essential for the G-protein-mediated pathway activated by DPDPE, but is not involved in the TIPP-activated pathway.
To further determine whether the G-protein-dependent mechanism was differentially involved in DPDPE- and TIPP-induced ERK activation, we inhibited the expression of Gβ1 subunit by small interference RNA (siRNA), and detected ERK activation induced by DPDPE and TIPP. As shown in Fig. 5, downregulation of Gβ1 expression remarkably reduced ERK1/2 activation stimulated by 1 nM DPDPE compared with control siRNA-transfected cells, indicating that DPDPE activates ERK1/2 in a G-protein-dependent manner. However, silencing Gβ1 expression resulted in no significant effect on 1 μM TIPP-induced ERK1/2 activation as compared with control siRNA-transfected cells (Fig. 5), indicating that TIPP activates ERK1/2 in a manner independent of G proteins.
Because Ser363 in the δOR was shown to be essential for the G-protein-mediated pathway activated by DPDPE, we therefore determined whether expression of S363A mutant δOR could affect the pathway by which DPDPE activates ERK. Interestingly, the inhibitory effect of Gβ1-specific siRNA on DPDPE-stimulated ERK activation was abolished in cells expressing S363A mutant δOR, indicating that Ser363 is a crucial site for DPDPE adaptation of the G-protein-dependent mechanism to activate ERK. However, TIPP-induced ERK activation was not affected by the expression of either Gβ1-specific siRNA or S363A δOR (Fig. 5). This result further confirmed the crucial role of Ser363 in G-protein-dependent activation of ERK by individual agonists.
S363A mutation of δOR inhibited the association of Src with the ERK cascade induced by DPDPE
Because DPDPE but not TIPP could induce the association of Gβγ with Src, we further examined whether there was difference between DPDPE and TIPP treatment on the Src association with Raf-1, MEK1 and ERK1/2. CHOδOR cells were treated with 1 μM DPDPE or TIPP for 5 minutes, and the association of Src with Raf-1, MEK1 and ERK1/2 was tested by co-immunoprecipitation and immunoblotting. As shown in Fig. 6, minor associations of Raf-1, MEK1 and ERK1/2 with c-Src were found in the absence of DPDPE or TIPP. After treatment with DPDPE, the levels of Raf-1, MEK1 and ERK1/2 association with c-Src were greatly increased. However, TIPP could not induce such associations. Combined with the findings above, the results suggest that DPDPE, but not TIPP, stimulation of ERK1/2 is implicated in Gβγ subunits as the primary mediator of Raf activation in an Src-dependent manner.
Because Ser363 in the δOR was shown to be crucial for the association of Gβγ with Src, we therefore determined whether expression of S363A mutant δOR could inhibit the ability of DPDPE to induce the association of Src with the ERK cascade. As shown in Fig. 6, DPDPE could not induce the association of c-Src with Raf-1, MEK1 or ERK1/2 in CHO cells transiently transfected with S363A δOR. The results indicate that the Ser363 site of δOR is a major determinant for G-protein-dependent ERK activation.
Effect of β-arrestin1/2 levels on DPDPE- and TIPP-stimulated ERK1/2 activation in cells expressing wild-type or mutant δOR
To determine whether the β-arrestin-dependent pathway is involved in activation of ERK1/2 by TIPP, we manipulated the levels of endogenous β-arrestin1 and β-arrestin2 (β-arrestin1/2) in CHOδOR cells using either siRNA to inhibit expression or DNA plasmids to induce overexpression, and then detected ERK1/2 activation by TIPP and DPDPE. As shown in Fig. 7, silencing β-arrestin1/2 expression using siRNAs targeting β-arrestin1 and β-arrestin2 remarkably reduced ERK1/2 activation stimulated by TIPP compared with control siRNA-transfected cells, indicating that TIPP activates ERK1/2 in a β-arrestin-dependent manner. However, silencing β-arrestin1/2 expression by siRNA resulted in no significant effect on DPDPE-stimulated ERK1/2 activation as compared with control siRNA-transfected cells, indicating that DPDPE activates ERK1/2 in a manner dependent on G proteins, but not on β-arrestins. To further determine the role of Ser363 of δOR in adopting the pathway of ERK activation, we examined whether the expression of S363A mutant δOR could affect DPDPE- or TIPP-induced ERK activation in cells expressing either wild-type or S363A δOR. As shown in Fig. 7, ERK activated by both DPDPE and TIPP was significantly inhibited by siRNA targeting β-arrestin1/2 in cells expressing S363A δOR, indicating that the mutation of Ser363 of δOR could change the pathway of DPDPE-induced ERK activation from G-protein-dependent to β-arrestin-dependent.
To further confirm the distinct involvement of β-arrestin1/2 in TIPP- and DPDPE-mediated ERK1/2 activation, we next examined the effect of overexpression of β-arrestin1/2 on TIPP- and DPDPE-mediated ERK1/2 activation. In cells expressing β-arrestin1/2–GFP and wild-type δOR, stimulation with TIPP for 5 minutes significantly increased phospho-ERK1/2, whereas no significant enhancement of phospho-ERK1/2 was detected in the cells overexpressing β-arrestin1/2–GFP following stimulation with DPDPE for 5 minutes (Fig. 8). This supports the idea that β-arrestin1/2 might not play an important role in DPDPE-stimulated ERK1/2 activation. On the other hand, in cells expressing β-arrestin1/2–GFP and S363A mutant δOR, stimulation with both DPDPE and TIPP for 5 minutes significantly increased ERK1/2 phosphorylation, agreeing with the results shown in Fig. 7. Overall, these results indicate that TIPP, but not DPDPE, activates ERK1/2 mainly in a β-arrestin-dependent manner, and that the Ser363 site of δOR is a major determinant of the pathway leading to ERK activation.
TIPP took β-arrestin as scaffold for kinases of the ERK cascade, and DPDPE had a similar effect in cells expressing mutant δOR
The finding that ERK1/2 activated by TIPP and DPDPE is differentially sensitive to siRNAs targeting β-arrestin1/2 and to overexpression suggests that β-arrestins might act as scaffolds for the components of the ERK cascade following activation of the δOR by DPDPE or TIPP. To confirm that β-arrestins can function as scaffolds for the component kinases of the ERK cascade, we tested the effect of DPDPE or TIPP stimulation on the assembly of complexes containing β-arrestin1/2 and Raf-1, MEK1 and ERK1/2 by immunoprecipitation and immunoblotting. CHO cells coexpressing hemagglutinin (HA)-tagged δOR and β-arrestin1–GFP or β-arrestin2–GFP, were treated for 5 minutes with 1 μM DPDPE or TIPP, and then the complexes containing β-arrestins were immunoprecipitated using anti-β-arrestin antibodies, separated by SDS-PAGE, and immunoblotted by antibodies against Raf-1, MEK1, ERK1/2 and δOR. As shown in Fig. 9, TIPP treatment significantly increased the levels of Raf-1, MEK1 and ERK1/2 association with β-arrestin1/2–GFP. However, only a minor association of Raf-1, MEK1 and ERK1/2 with β-arrestin1/2–GFP was found in the cells treated with DPDPE, indicating that TIPP was more likely to activate ERK in a β-arrestin-dependent manner. However, only DPDPE was able to induce the association of β-arrestin1/2–GFP with Src, which agrees with the finding that TIPP was less capable of initiating membrane translocation of β-arrestins or Src than was DPDPE (Hong et al., 2009).
Furthermore, we also tested the effect of transient transfection of S363A δOR on β-arrestin scaffolding. In CHO cells that expressed S363A δOR, DPDPE lost its ability to induce the association of β-arrestin1/2–GFP with Src. However, scaffolding of the ERK cascade molecules by β-arrestin1/2–GFP in response to DPDPE treatment was enhanced in cells expressing mutant δOR. The immunoblotting of total phospho-Ser363 δOR shows that Ser363 was phosphorylated by DPDPE, but not by TIPP, in cells expressing wild-type δOR, whereas such phosphorylation by DPDPE was not observed in cells transfected with mutant δOR. These results indicate that Ser363 phosphorylation is crucial for the δOR adoption of signal pathways for activating ERK1/2 in response to individual agonist stimulations.
Cytoplasmic distribution of phospho-ERK1/2 after TIPP stimulation resulted in GRK2 and β-arrestin1 phosphorylation, and DPDPE had similar effects in cells expressing mutant δOR
It has been increasingly accepted that the mechanism of MAPK activation is a major determinant of MAPK function. The data above show that TIPP could facilitate the formation of complexes composed of β-arrestin and ERK cascade molecules and generate a larger pool of cytosolic phospho-ERK1/2, which suggests that the substrates of ERK1/2 activated by TIPP are cytoplasmic. It has been reported that cytosolic ERK1/2 can inactivate β-arrestin1 and GRK2 by phosphorylating Ser412 of β-arrestin1 and Ser670 of GRK2 (Lin et al., 1999; Pitcher et al., 1999). We therefore tested whether ERK activated by TIPP could phosphorylate β-arrestin1 and GRK2. CHOδOR cells were treated for 5 minutes with 1 μM DPDPE or TIPP before nuclei and cytoplasm were separated. Consistent with the observations using confocal microscopy, phospho-ERK1/2 activated by DPDPE was mainly detected in the nucleus, whereas phospho-ERK1/2 activated by TIPP was mainly detected in the cytoplasm (Fig. 10A,B). As expected, GRK2 was found to be robustly phosphorylated at Ser670 in response to treatment with 1 μM TIPP, but not DPDPE, for 5 minutes. The phosphorylation of β-arrestin1 was markedly decreased in response to treatment with DPDPE but not with TIPP (Fig. 10A,C,D). Cytosolic β-arrestin1 is constitutively phosphorylated and rapidly dephosphorylated when it is recruited to the plasma membrane in response to agonist stimulation (Lin et al., 1997; Lin et al., 1999). At the plasma membrane, β-arrestin1 is rapidly dephosphorylated. In the case of DPDPE treatment, β-arrestin1 cannot be phosphorylated due to the relatively lower level of phospho-ERK1/2 in the cytoplasm, thereby leading to a significant decrease of phosphorylated β-arrestin1. However, the phosphorylation of β-arrestin1 Ser412 sustained in response to treatment with 1 μM TIPP, because of the high level of phospho-ERK1/2 and thus phosphorylated β-arrestin1, was maintained. Taken together, these results indicate that TIPP, but not DPDPE, treatment can form a cytoplasmic pool of phospho-ERK, which in turn results in phosphorylation (inactivation) of β-arrestin1 and GRK2.
Because Ser363 phosphorylation of δOR is a determinant of β-arrestin scaffolding of the ERK cascade, we also tested the cytoplasmic pool of β-arrestin-bound phospho-ERK upon stimulation with DPDPE and TIPP in CHO cells transiently transfected by S363A δOR. As shown in Fig. 10, upon TIPP stimulation, phospho-ERK1/2 presented a similar distribution to that in cells expressing wild-type δOR. However, unlike the distribution found mainly in the nucleus in response to DPDPE stimulation in cells expressing wild-type δOR, ERK activated by DPDPE remained in the cytoplasm in cells transfected with mutant δOR and could also induce robust phosphorylation of GRK2 and maintain the phosphorylation status of β-arrestin1 in cells expressing mutant δOR. Overall, these results further indicate that δOR Ser363 phosphorylation is a major determinant of ERK distribution, which in turn determines MAPK function.
Mutation of Ser363 of δOR rescued ERK signaling desensitization and prolonged DPDPE-induced ERK activation
The results shown above indicate that the Ser363 residue in the C-terminal tail of δOR is a crucial site for the δOR adoption of G-protein- or β-arrestin1/2-dependent pathways to activate ERK, leading to differential kinetic and spatial patterns of phopho-ERK1/2 activation in response to DPDPE or TIPP stimulation. To determine whether differential desensitization and duration of ERK activation by DPDPE or TIPP are attributed to their adaptations of distinct mechanisms (G-protein-dependent versus β-arrestin-dependent) to activate ERK, we next examined the effect of S363A mutation on the desensitization and short duration of DPDPE-mediated ERK1/2 activation. As shown in Fig. 11A,B (also see Fig. 3), DPDPE stimulation was unable to induce a significant increase of ERK1/2 phosphorylation in CHOδOR cells pretreated for 2 hours with 1 nM or 1 μM DPDPE, but not with 1 μM TIPP, indicating that the δOR was desensitized in modulation of ERK1/2 activation upon DPDPE treatment. However, the introduction S363A mutant δOR significantly rescued DPDPE-induced ERK1/2 desensitization (Fig. 11A,B). As shown in Fig. 11C,D (also see Fig. 1), in cells expressing wild-type δOR, ERK activation stimulated by 1 nM or 1 μM DPDPE but not by 1 μM TIPP was transient and declined rapidly. However, in cells expressing S363A δOR, DPDPE stimulation (like TIPP) also resulted in prolonged ERK activation. These results suggest that mutation of the Ser363 residue of δOR could promote DPDPE to adopt a β-arrestin1-depedent mechanism for activation of ERK, thereby leading to non-desensitization and prolonged duration.
The present study clearly demonstrates that DPDPE and TIPP produced distinct temporal and spatial patterns of ERK activation mediated by δORs. ERK1/2 activated by DPDPE is more intense but transient and distributes in the nucleus, consistent with the pattern of G-protein-dependent activation. However, ERK1/2 activated by TIPP is less robust but more persistent and distributes largely in the cytoplasm, consistent with pattern of β-arrestin-dependent activation. Differential involvement of G-protein- and β-arrestin-dependent mechanisms in DPDPE- and TIPP-stimulated ERK activation is supported by several observations. First, DPDPE but not TIPP treatment results in the dissociation of heterotrimeric G proteins into Gβγ subunits that are associated with PLCβ3 and c-Src, indicative of activation of G protein upon occupancy of its receptor by DPDPE, but not TIPP. Second, TIPP but not DPDPE exhibited robust ability to employ β-arrestin1/2 as a scaffold for the component kinases of the ERK cascade. Third, manipulation of the levels of Gβ1 and β-arrestin1/2 significantly alter the ability of TIPP, but not DPDPE, to stimulate ERK activation. Finally, the cytosolic colocalization of phospho-ERK1/2 with GFP–β-arrestin was visualized overwhelmingly in cells treated with TIPP but not in those treated with DPDPE. These observations support the conclusion that G-protein- and β-arrestin-dependent pathways might mediate DPDPE- and TIPP-stimulated ERK activation, respectively. Moreover, the present study further revealed that the Ser363 residue in the C-terminal tail of δOR is crucial for the δOR adoption of G-protein- or β-arrestin-dependent mechanisms to activate ERK1/2 in response to DPDPE and TIPP stimulations. This was shown by the finding that after mutation of δOR Ser363, DPDPE also adopted β-arrestin1/2 as scaffolds to assemble a complex with kinases of the ERK cascade instead of employing the Gβγ subunits as the primary mediator to activate the kinases of the ERK cascade in an Src-dependent manner.
Although increasing evidence indicates that GPCRs can activate ERK1/2 in a β-arrestin-dependent mechanism, the functions of ERK activated by such mechanism are poorly understood. The present study suggests that β-arrestin-dependent activation of ERK might exert an inhibitory effect on the desensitization and internalization of the δORs by controlling the subcellular distribution and duration of active ERK. ERK is known to phosphorylate multiple substrates, including nuclear and non-nuclear substrates, such as plasma membrane, and cytoplasmic and cytoskeletal proteins (Pearson et al., 2001). The subcellular distribution of active ERK might be crucial for determining the substrates that it phosphorylates. Cytosolic GRK2 and β-arrestin1/2 have been reported to be phophorylated and inactivated by ERK1/2 (Lin et al., 1999; Pitcher et al., 1999; Elorza et al., 2000). Because ERK1/2 stimulated by TIPP mainly colocalizes in the cytosol, this might ensure that phospho-ERK1/2 is targeted to β-arrestin1/2 and GRK2, and inactivates them by phosphorylation. Indeed, our data clearly show that TIPP treatment facilitated the phosphorylation at Ser670 of GRK2 and maintained the phosphorylation of β-arrestin1, corresponding to the inability of TIPP to cause δOR internalization and desensitization, as shown by our recent study (Hong et al., 2009).
The duration of ERK1/2 activation has been shown to play a role in the determination of δOR internalization or desensitization (Eisinger and Schulz, 2004). The formation of a stable cytoplasmic pool of β-arrestin-bound phospho-ERK by TIPP stimulation might restrict phospho-ERK translocation into the nucleus, where it is dephosphorylated and inactivated (Ahn et al., 2004); this could lead to a prolonged duration. On the other hand, a stable cytoplasmic pool of β-arrestin-bound phospho-ERK might also restrict the translocation of cytosolic β-arrestin1/2 to membrane receptors. Overall, these actions of β-arrestin-bound phospho-ERK in the cytoplasm might provide the molecular basis underlying the regulatory effects of ERK on the functions of β-arrestin1/2 and GRK2. The implication of β-arrestin-dependent ERK activation in inhibition of δOR desensitization is evidenced by the observations that when Ser363 was mutated, DPDPE gained the ability to utilize β-arrestin1/2 as scaffolds to assemble a complex with kinases of the ERK cascade (Fig. 9), accompanied by a remarkable decrease in the desensitization of ERK signaling (Fig. 11A,B). Accordingly, there is evidence that morphine, a unique agonist that fails to desensitize and internalize opioid receptors (Keith et al., 1996; Sternini et al., 1996), has been shown to display a similar pattern of ERK activation as TIPP, with longer duration and without desensitization. Such a pattern of ERK activation is linked to the incapability of morphine to induce the internalization and desensitization of the δORs (Eisinger et al., 2002; Eisinger and Schulz, 2004). We also found that, like TIPP, morphine could employ β-arrestin1/2 as a scaffold for the component kinases of the ERK cascade (see supplementary material Fig. S1). Taken together, the findings of our studies and those of others suggest that β-arrestin-dependent ERK activation might play a role in preventing δOR desensitization.
Currently, the mechanisms underlying the different involvement of G protein and β-arrestins in DPDPE- and TIPP-mediated ERK1/2 activation are unclear. Distinct receptor configurations formed upon occupancy of receptors by DPDPE and TIPP might be crucially responsible for such differences. Accumulating evidence suggests that the GPCRs could adopt differential active configurations, which might couple distinct signal pathways (Azzi et al., 2003; Wei et al., 2003; Shenoy et al., 2006). Individual ligands might stabilize distinct active configurations (Ghanouuni et al., 2001) and thus prefer the coupling of one signal pathway to another (Berg et al., 1998). For the system considered herein, DPDPE would stabilize a receptor configuration that could adopt G-protein-dependent pathway to activate ERK, whereas the configuration stabilized by TIPP could not adopt a G-protein-dependent mechanism but favor employment of a β-arrestin-dependent mechanism to activate ERK. The phosphorylation of the Ser363 residue in the C-terminal tail of δOR might be crucial for receptor adaptation of differential active configurations that could respectively employ G-protein- or β-arrestin-dependent signal pathways to activate ERK in response to DPDPE and TIPP stimulations. DPDPE stimulation induced Ser363 phosphorylation but formed negligible complexes of β-arrestin1/2 and the component kinases of the ERK cascade (Fig. 9). By contrast, TIPP stimulation failed to induce Ser363 phosphorylation but formed robust complexes of β-arrestin1/2 and the component kinases of the ERK cascade. More importantly, when the Ser363 residue was mutated, DPDPE gained the ability to employ β-arrestin1/2 as scaffolds to assemble a complex with kinases of the ERK cascade.
The paradoxical phenomena observed in this study and in our recent study are that although the β-arrestin scaffold for component kinases of ERK cascade was clearly detected upon TIPP treatment, little recruitment of β-arrestin1/2 to the membrane receptors could be detected by confocal laser microscopy upon TIPP stimulation (Hong et al., 2009). We reason that the small amount of β-arrestin recruitment could be caused by a combination of factors. First, because TIPP was unable to induce the release of Gβγ subunits that are known to play a major role in GRK2 membrane targeting, which is essential for β-arrestin recruitment, TIPP might lead to insufficient β-arrestin recruitment to these receptors (Shenoy et al., 2006). Second, the receptor configuration induced by TIPP might not favor deubiquitylation of β-arrestin, which would decrease the propensity of β-arrestin for membrane association, because the ubiquitylation status of β-arrestin determines the stability of the δOR–β-arrestin complex as well as the trafficking pattern of β-arrestin (Shenoy et al., 2003; Shenoy et al., 2005).
Alternatively, in light of the findings of our studies and those of others, it seems that β-arrestins are not exclusively receptor-binding adapters as thought previously. It has been shown that in COS1 cells coexpressing HA-tagged β2AR and β-arrestin–GFP, translocation of β-arrestin to the membrane could not be detected upon stimulation with ICI118551 or propranolol (Azzi et al., 2003), although they robustly activated ERK1/2 by an β-arrestin-dependent pathway. In addition, Song and co-workers have shown that the scaffolding function is not limited to receptor-bound arrestin; a receptor binding-impaired β-arrestin mutant can also act as a scaffold for the JNK3 activation cascade (Song et al., 2009). Moreover, recent studies also demonstrated that β-arrestin can directly scaffold c-Raf, MEK and ERK (Song et al., 2009; Meng et al., 2009). Therefore, direct association of the δOR with β-arrestin1/2 might not be absolutely indispensable for scaffolding component kinases of the ERK cascade by β-arrestin1/2. Additionally, β-arrestin1/2 can serve as a scaffold for assembly of the complexes containing the component kinases of the ERK, not only at the membrane but also at the cytoplasm. Taken together, how a receptor that binds relatively little β-arrestin can robustly engage a β-arrestin-dependent signaling pathway to activate ERK remains to be addressed.
In summary, the present study clearly demonstrates that the δOR could employ either a G-protein- or β-arrestin-dependent mechanism to activate ERK, dependent on distinct agonist stimulations. Moreover, we further revealed that the Ser363 residue in the C-terminal tail of δOR was crucial for the δOR adoption of a G-protein- or β-arrestin-dependent mechanism to activate ERK1/2 in response to DPDPE and TIPP stimulations. More importantly, we found that ERK activated by a β-arrestin-dependent pathway could phosphorylate and inactivate β-arrestin1 and GRK2, thereby exerting a negative regulation on receptor desensitization. Therefore, this study reveals a novel molecular mechanism underlying the regulation of the δOR trafficking and desensitization.
Materials and Methods
Chemicals and reagents
ECL Plus Western Blotting Detection Reagents were purchased from GE Healthcare (Little Chalfont, Buckinghamshire, UK). The HA-tag, p363-δ-opioid receptor, phospho-β-arrestin1, GRK2, c-Src, MEK1 and PLCβ3 antibodies were purchased from Cell Signaling Technology (Beverly, MA). The Alexa-Fluor-488-conjugated secondary antibody was purchased from Invitrogen (Carlsbad, CA). The phospho-ERK, ERK2, Raf-1 and lamin B antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). TIPP (H-Tyr-Tic-Phe-Phe-OH) was purchased from Tocris Cookson (Ballwin, MO). The anti-Gβ antibody and ReBlot Plus Mild Antibody Stripping Solution (10×) was from Millipore (Bedford, MA). Fugene 6 transfection reagent was from Roche. The QuikChange Lightning Site-Directed Mutagenesis Kit was purchased from Stratagene (La Jolla, CA). All other reagents and antibodies were purchased from Sigma (St Louis, MO).
Mutagenesis of δOR
The serine codon (TCC) for amino acid 363 of the mouse δOR was mutated into an alanine codon (GCC) using the QuikChange Lightning Site-Directed Mutagenesis Kit. The wild-type cDNA clone in pCDNA3.0 was used as a template with the following sense and antisense primers: 5′-TGTCACTGCCTGCACCCCCGCC-GACGGCCC-3′ (sense), and 5′-GGGCCGTCGGCGGGGGTGCAGGCAGTGACA-3′ (antisense). A polymerase chain reaction (PCR) with 1 μM QuikChange Lightning Enzyme was run for 18 cycles (denaturing 20 seconds at 95°C, annealing 10 seconds at 60°C, extending 3 minutes at 68°C). Digestion of the reacted solution with Dpn I destroyed the transfection capacity of the template plasmid. Escherichia coli colonies transfected with the digested PCR solution were grown to produce plasmid for transfecting the CHO cells.
Cell culture and transfection
The cDNA of pcDNA3.0 containing the mouse δOR with the HA epitope tag inserted to the N-terminus of δOR or the mutated δOR were transfected into CHO cells by using Fugene 6 transfection reagent (Roche). Cells stably expressing HA–δOR were obtained by MACS cell selection kits (Miltenyi Biotech, Germany) and maintained by 0.5 mg/ml G418. The CHOδOR cells were transfected with GFP–β-arrestin1 or GFP–β-arrestin2, followed by flow cytometry sorting.
Immunoprecipitation and immunoblotting
Cells were solubilized in 1 ml of glycerol lysis buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 10% glycerol v/v, 0.5% Nonidet P-40 v/v, 2 mM EDTA, 100 μM Na3VO4, 1 mM phenylmethylsulfonylfluoride, 10 μg/ml leupeptin and 10 μg/ml aprotinin), and clarified by centrifugation at 10,000 g for 10 minutes. Immunoprecipitation was performed using 10 μl of 50% slurry of protein A/G agarose, with constant agitation overnight at 4°C. Immune complexes were washed three times with glycerol lysis buffer and boiled in Laemmli sample buffer. Immunoprecipitated proteins were resolved by SDS-PAGE and transferred to nitrocellulose membrane for immunoblotting. Chemiluminescence detection was performed by using the ECL Plus Western Blotting Detection Reagent (GE Healthcare) and immunoblots were quantified by densitometry using Quantity One (Bio-Rad). For repeated immunoblotting, membranes were stripped in ReBlot Plus Mild Antibody Stripping Solution for 15 minutes (Millipore).
Cellomics phospho-ERK assay
Cells growing on standard high-density microplates were stimulated, and then fixed and immunofluorescently labeled to identify activated ERK in the cell nucleus. The antibody bound to dually phosphorylated ERK and was validated for non-crossreactivity to other MAPK (p38, JNK/SAPK) members as well as unphosphorylated ERK. Prepared cells were analyzed using standard fluorescence microscopy or using a Cellomics fully automated HCS Reader to facilitate the quantification of ERK1/2 phosphorylation and translocation.
Chemical synthesis of double-stranded siRNAs targeting β-arrestin1 and β-arrestin2 was performed as described (Zhang et al., 2005). The siRNA sequences targeting β-arrestin1 and β-arrestin2 are 5′-GGAAGCTCAAGCACGAAGACA-3′ and 5′-GGCTTGTCCTTCCGCAAAGAC-3′, corresponding to the positions 872–892 and 216–236 relative to the start codons, respectively. A non-silencing RNA duplex (5′-AAUUCUCCGAACGUGUCACGU-3′), as indicated by the manufacturer, was used as control. For siRNA experiments, CHOδOR cells that were 40–50% confluent on 60-mm dishes were transfected with 20 μg of siRNA by using the Oligofectamine Transfection Reagent (Invitrogen), and were split into 24-well plates after 48 hours for phospho-ERK assays.
Confocal laser microscopy
Cells were seeded onto poly-D-lysine-coated coverslips placed in a 24-well plate and were exposed to 1 μM DPDPE or 1 μM TIPP in serum-free medium for time periods up to 30 minutes at 37°C. After incubation, cells were washed with ice-cold PBS for termination and fixed with 4% paraformaldehyde in PBS for 15 minutes. For the detection of cytosolic proteins, cells were fixed, washed with PBS and permeabilized with 1% Triton X-100 in PBS for 10 minutes at room temperature before staining to allow labeling of intracellular molecules. Nonspecific sites were blocked with 2% BSA and 0.5% FBS in PBS (containing 0.1% Tween) for 1 hour at room temperature. Subsequently, cells were incubated with primary and secondary antibodies diluted in PBS containing 2% BSA for 30 minutes, washed with PBS (containing 0.1% Tween) three times (15 minutes each) and mounted on slides with pure glycerin (1:1 with Milipore water). Fluorescence was observed using a Zeiss LSM 510 Laser scanning microscope and a Zeiss 63× 1.4 numerical aperture oil immersion lens with dual line switching excitation (488 nm for GFP and Alexa Fluor 488, 568 nm for TRITC) and emission (515–540 nm for GFP and Alexa Fluor 488, 590–620 nm for TRITC) filter sets.
Cells seeded on 24-well plates were starved for at least 4 hours in serum-free medium prior to stimulation. After stimulation with ligands of various concentrations for the indicated time periods, cells were solubilized by directly adding 1× SDS-sample buffer, followed by boiling at 100°C for 5 minutes. Equal micrograms of cellular extracts were separated on 12% Tris-glycine polyacrylamide gels and transferred to nitrocellulose membranes for immunoblotting. Phospho-ERK1/2 and total ERK1/2 were detected by immunoblotting with mouse polyclonal anti-phospho-ERK antibody and anti-ERK antibody (Santa Cruz Biotechnology, 1:1000). Phosphorylated ERK1/2 immunoblots were quantified by densitometry using Quantity One (Bio-Rad).
Nuclear extracts preparation
Nuclear extracts were prepared as described previously (Dignam et al., 1983) with minor modifications. After 12 hoursof serum starvation, cells were incubated with 1 μM DPDPE or TIPP for 5 minutes, washed and resuspended in 400 ml of hypotonic buffer (50 mM Tris-HCl pH 7.5, 10 mM NaCl and 2 mM EDTA). After incubation on ice for 10 minutes, 3 ml of 1% NP-40 was added. After 5 minutes, the cytoplasm was collected and the nuclei (pellet) were resuspended in hypertonic buffer and shaken for 1 hour at 4°C. After centrifugation at 10,000 g for 10 minutes, the supernatant (nuclear extracts) were saved.
All statistical and curve-fitting analyses were performed using the GraphPad Prism 5.0 software. Data represent mean ± s.e.m. of at least three separate experiments. Statistical significance was determined by one-way ANOVA followed by post hoc comparison using Dunnett's tests. When only two groups were compared, statistical significance was determined using an unpaired Student's t-test.
We thank Gang Pei for providing plasmids of β-arrestin1 and β-arrestin2. This work was supported by the Ministry of Science and Technology of China (2008BAI49B05; 2009CB522000; 2009ZX09301-001) and by National Science Fund for Distinguished Young Scholar from the National Natural Science Foundation of China (30425002), the National Natural Science Foundation (30873050) and fund granted by Chinese Academy of Sciences (KSCX2-YW-R-253).
↵* These authors contributed equally to this work
Supplementary material available online at http://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.073742/-/DC1
- Accepted August 26, 2010.
- © 2010.