The Agonist of Adenosine A1 Receptor Induced Desensitization of delta Opioid receptor-mediated Raf-1/MEK/ERK Signaling by Feedback Phosphorylation of Raf-1-Ser289/296/301

Our previous study found that activation of adenosine A1 receptor (A1R) induced phosphorylation of delta opioid receptor (DOR) and desensitization of its downstream signaling molecules, cAMP and Akt. To further investigate the effect of A1R agonist on DOR signaling and the underlying mechanism, we examined the effect of A1R activation upon binding of its agonist N6-cyclohexyl-adenosine (CHA) on DOR-mediated Raf-1/MEK/ERK activation, and found that prolonged CHA exposure resulted in downregulation of DOR-mediated Raf-1/MEK/ERK signaling pathway. CHA-treatment time dependently attenuated Raf-1-Ser338 phosphorylation induced by [D-Pen2,5] enkephalin (DPDPE), a specific agonist of DOR, and further caused downregulation of the Raf-1/MEK/ERK signaling pathway activated by DOR agonist. Moreover, CHA exposure time-dependently induced the phosphorylation of Raf-1-Ser289/296/301, the inhibitory phosphorylation sites that were regulated by negative feedback, thereby inhibiting activation of the MEK/ERK pathway, and this effect could be blocked by MEK inhibitor U0126. Finally, we proved that the heterologous desensitization of the Raf-1/MEK/ERK cascade was essential in the regulation of anti-nociceptive effect of DOR agonists by confirming that such effect was inhibited by pretreatment of CHA. Therefore, we conclude that the activation of A1R inhibits DOR-mediated MAPK signaling pathway via heterologous desensitization of the Raf-1/MEK/ERK cascade, which is a result of ERK-mediated Raf-1-Ser289/296/301 phosphorylation mediated by activation of A1R.


Introduction
During the past three decades, the adenosine receptors have been found to interact with opioid receptors, both directly [1][2][3] or indirectly [4], in clinical therapeutics such as analgesia [2] and cardio protection [3]. Our previous study found that the agonist of adenosine A 1 receptor (A 1 R), N6-cyclohexyl-adenosine (CHA), induced phosphorylation of delta opioid receptor (DOR), along with heterologous desensitization of DOR-mediated inhibition of intracellular cAMP and phosphorylation of Akt in CHO cells stably expressing A 1 R and DOR [5], thus revealing a mechanism underlying DOR desensitization induced by A 1 R-DOR interaction. However, the signaling pathway supporting such effect remains unclear. To further study the effect of A 1 R agonist on DOR signaling and its possible mechanism, we focused on the role of mitogen-activated protein kinase (MAPK) pathways mediated by DOR in our present study.
Both A 1 R and DOR belong to the G protein-coupled receptor (GPCR) superfamily, the activation of which may give rise to transient activation of mitogen-activated protein kinase (MAPK) pathways, and especially the extracellular signal-regulated kinase (ERK) signaling pathway [6][7][8], which consists of a three-tier hierarchical signaling cascade of protein kinases: v-raf-1 murine leukemia viral oncogene homolog 1 (Raf), mitogen-activated protein kinase or ERK kinase (MEK) and ERK [9,10]. As homologs of the v-Raf oncogene, three isoforms of the Raf kinase have been discovered in vertebrates: A-Raf, B-Raf and C-Raf (Raf-1), the most intensively studied of all isoforms [10]. All three isoforms are activated by binding of their upstream regulators, the small G proteins of the rat sarcoma (Ras) family, to their N-terminal regions, and share three common conservative regions (CRs), namely CR1, CR2 and CR3 [10]. CR1 contains the RAS-binding domain (RBD) and the cysteine-rich domain (CRD) that are required for membrane recruitment, whereas the hinge region (CR2) and catalytic region (CR3) contain numerous phosphorylation sites for kinase activity regulation that are targeted by a variety of other kinases and proteins, such as protein kinase A and 14-3-3, through direct binding [10,11].
As a serine/threonine protein kinase with multiple phosphorylation sites, Raf-1 is regulated via a series of phosphorylation status that contribute to either activation or inhibition of itself, which are induced by a number of kinases and proteins, including upstream and downstream kinases of the MAPK cascade itself, or even other regulating proteins. For example, the Ser338 residue is a more important phosphorylation site for activation of the MAPK cascade, since it is indispensable in introducing negative charges into the 4-amino-acid motif (SSYY), which is essential for mediating Ras-dependent Raf-1 activation. Therefore, the phosphorylation of Raf-1 Ser338 is considered as an index of MAPK activation [12]. Other phosphorylation sites include Ser43, Ser233 and Ser259, which are targets of protein kinases that are responsible for Raf-1 inhibition, and Ser233, Ser259 and Ser621 that may result from binding of 14-3-3, thereby activating Raf-1 [10,13,14]. On the other hand, Ser289, Ser296 and Ser301, the three adjacent phosphorylation sites are targets of ERK, the downstream kinase of the MAPK cascade, in the key regulatory step on negative feedback. Phosphorylation of these sites contribute to negative regulation of Raf-1, while phosphorylation of Ser338 may block ERK-induced inhibitory effect [15]. U0126, the specific inhibitor of MEK1/2, was able to inhibit phosphorylation of Ser289/296/301 in COS-7 and NIH3T3 cells after prolonged exposure to platelet derived growth factor (PDGF) and epidermal growth factor (EGF), whereas activation or overexpression of ERK resulted in enhanced Ser289/296/301 phosphorylation of Raf-1 [16,17].
The negative feedback mechanism induced by Ser289/296/301 phosphorylation has revealed an essential manner of Raf-1 regulation, and thereby the modulation of the entire Raf-1/MEK/ERK cascade upon GPCR activation. However, is remains elusive whether CHA-induced heterologous desensitization of DOR-mediated ERK signaling pathway is associated with the triple phosphorylation sites of Raf-1, although we have previously proven that activation of A 1 R may induce heterologous desensitization by facilitating DOR phosphorylation, while inhibiting cAMP accumulation and Akt phosphorylation [5]. Thus, in our present study, we focus on the mechanism by which prolonged CHA treatment induce heterologous desensitization of DOR-induced MAPK cascade, along with the role of Raf-1-Ser289/296/301 phosphorylation in such regulation.

Cell Culture
CHO cells stably expressing HA-tagged A 1 R and DOR-CFP fusion protein were maintained in F12 medium (Thermo-Fisher Scientific, MA, USA) with 10% fetal calf serum, and incubated in a humidified atmosphere consisting of 5% CO2 at 37 °C. 0.5 mg/ml G418 were added to maintain expression of both receptors.

Western Blot Analysis
Cells were seeded in 24-well plates, incubated at 37 °C for 24 h, and starved in serum free media overnight. After treated with indicated chemicals, cells were lysed immediately by RIPA extraction buffer, and then boiled for 10 min.
Cell extracts were subjected to 10%-SDS polyacrylamide gel electrophoresis, and transferred on to nitrocellulose membrane (GE Healthcare, UK). Membranes were blocked with 5% non-fat dried milk dissolved in PBS/0.1% Tween 20 (PBS/T) for 1 h, and incubated overnight at 4 °C with primary antibodies diluted in PBS/T containing 5% non-fat dried milk. Next, membranes were subjected to 4 washes with PBS/T before incubating for 1 h at room temperature with a horseradish peroxidase (HRP)-conjugated secondary antibody (Merck, Germany). Chemiluminescence detection was performed with ECL Plus Western Blotting Detection Reagent (GE Healthcare, UK). Immunoblots were quantified by densitometry with Quantity One software (Bio-Rad, CA, USA). For repeated immunoblotting, membranes were stripped in ReBlot Plus Mild Antibody Stripping Solution for 15 min (Merck, Germany).

Animals
Eight-week-old male C57BL/6J mice were obtained from Shanghai Laboratory Animal Center, Chinese Academy of Sciences (Shanghai, China) two weeks before experiments, and housed in the Laboratory Animal Center of Zhejiang Chinese Medical University accredited by the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC). Animals were cared under standard environmental conditions (12 h light-dark cycle and 24 ± 2 °C) with access to food and water provided ad libitum. All

Tail-flick Test
The tail-flick test was adopted to evaluate drug-induced antinociception according to the procedure previously described [18]. Briefly, the light intensity of tail-flick apparatus (Ugo Basile, Italy) was set within the range of 2.5-3.5 s to allow the basal tail-flick latencies, with a cut-off time of 10 s to avoid tissue damage. The latencies were recorded at 5, 15, 30, 45, 60, 90, 120 and 180 min after DPDPE administration, and the differences between curves were calculated by two-way ANOVA. The area under curve (AUC) of the timeresponse curves obtained by calculating the area under the time-response curve of the tail flick test (test latency -basal latency) from 5 to 180 min were also regarded as an index of the antinociceptive effects.

Statistical and Data Analysis
All statistical and curve-fitting analysis was performed by the GraphPad Prism 7.0 software (GraphPad software, San Diego, CA, USA). Data represent mean ± S.E.M. of from at least four separate experiments. Statistical differences between values from two groups were determined using unpaired Student's t tests, whereas those between three or more groups were analyzed using analyses of variance (ANOVAs) followed by Newman-Keuls post hoc comparisons, with independent or repeated-measures in accordance with the experimental design. For in vivo experiments, data were analyzed by two-way ANOVA to determine the statistical differences among different groups of drug treatments. A difference of p < 0.05 was considered statistically significant.

Both CHA and DPDPE Induced Phosphorylation of Raf-1-Ser338 and ERK1/2 in a time-dependent Manner
In order to prove the roles of A 1 R and its agonist in heterologous desensitization of DOR-induced MAPK cascade, we first examined the abilities of agonists of A 1 R and DOR, CHA and DPDPE, respectively, in activation of Raf-1 and ERK1/2, by detecting the crucial phosphorylation sites that control their activation (Fig. 1). By western-blot assays, we found that both CHA and DPDPE induced ERK1/2 phosphorylation in a dose-dependent manner (0-10 µM for CHA, and 0-100 nM for DPDPE) after the CHO cells were imposed to agonists for 5 min, and determined the saturating concentrations of both agonists (1 µM for CHA, and 10 nM for DPDPE) when ERK1/2 phosphorylation reached the peak levels ( Fig. 1, A and B). Next, we tested the phosphorylation of Raf-1 and ERK1/2 upon treatment with 1 µM CHA and 10 nM DPDPE for different time periods as indicated (0-240 min, Fig. 1, C-F). Treatment with both agonists resulted in rapid rise of phospho-Raf-1-Ser338 and phospho-ERK1/2 immunoreactivities within 5 min of agonist exposure, which gradually declined in proportion with the duration of treatment, from 15 to 240 min. Thus, it is clear that CHA and DPDPE activated Raf-1 and ERK1/2 in identical time-dependent manners. The consistency in the time course of their phosphorylation may further infer that both CHA and DPDPE activated ERK via Raf-1-dependent pathways.
for 5 min. As a result, ERK phosphorylation was attenuated by CHA pretreatment in a time-dependent manner, as CHA pretreatment with over 10 min was able to drastically inhibit subsequent ERK phosphorylation induced by DPDPE. On the other hand, pretreatment with 10 nM DPDPE was not able to affect ERK phosphorylation induced by the following incubation with CHA ( Fig. 2, C and D), indicating the unidirectional regulation of DOR by the A 1 R agonist. Moreover, when cells were pretreated with 100 µM DPCPX, a specific antagonist of A 1 R, for 15 min, followed by treatment with 1 µM CHA for an additional 60 min, the effect of CHA on heterologous desensitization was significantly blocked, as DPDPE was able to induce ERK phosphorylation despite pretreatment with CHA. However, DPCPX

A 1 R Activation Induced Heterologous Desensitization of DOR-mediated ERK Phosphorylation
Now that we have proven that A 1 R and DOR activate ERK1/2 in identical time-dependent manners, and due to our previous finding that CHA was able to induce desensitization of DOR-mediated cAMP/Akt signaling [5], a phenomenon referred to as heterologous desensitization, we further examined whether CHA also inhibited DOR-mediated ERK phosphorylation, in our current study. As shown in Fig. 2, A and B, CHO cells stably expressing A 1 R and DOR were pretreated with 1 µM CHA for different time periods (0-360 min), washed, and incubated with 10 nM DPDPE Fig. 1 Both CHA and DPDPE induced phosphorylation of Raf-1-Ser338 and ERK1/2 in a time-dependent manner (A) Dose-dependent phosphorylation of ERK1/2 stimulated by CHA (0-10 µM, 5 min) and DPDPE (0-100 nM, 5 min), with immunoblots of total ERK2 used as the internal reference (B) Quantification of ERK1/2 phosphorylation levels obtained in A. The results were normalized against the level of ERK2, and further normalized against the result obtained from the control group. Error bars represent SEM, n = 5 (C and E) Time-dependent phosphorylation of Raf-1-Ser338 (top panel) and ERK1/2 (bottom panel) stimulated by CHA (1 µM, 0-240 min) and DPDPE (10 nM, 0-240 min), with immunoblots of actin and total ERK2 used as the internal reference, respectively. Immunoblots were tested with anti-phospho-Raf-1-Ser338, antiactin, anti-phospho-ERK1/2, and anti-ERK2 antibodies (D and F) Quantification of pRaf-1-338 (D) and pERK1/2 levels (F) obtained in C and E, respectively. The results were normalized against the level of actin, and further normalized against the result obtained from the control group. Error bars represent SEM, n = 5 expressing A 1 R and DOR were pretreated with 1 µM CHA for different time periods (0-60 min), washed, and incubated with 10 nM DPDPE for 5 min. Phosphorylation of Raf-1-Ser338 and MEK1/2 stimulated by DPDPE were blocked by prolonged CHA pretreatment (15-60 min), suggesting the heterologous desensitization of both Raf-1 and MEK1/2 downstream of DOR induced by activation of A 1 R. Further investigation with different doses (0.1-10 nM) of DPDPE showed that DPDPE could dose-dependently activate the phosphorylation of Raf-1-Ser338 and MEK1/2, in a manner similar to that of ERK, which decreased after pretreatment with 1 µM CHA for 60 h (Fig. 3, C-E), indicating that desensitization of DOR-mediated MAPK cascade occurs treatment alone showed no effect on DPDPE-induced ERK activation (Fig. 2E). These results suggest that CHA induced the heterologous desensitization of ERK by activating A 1 R.

A 1 R Activation Induced Heterologous Desensitization of DOR-mediated Raf-1 and MEK Activation
After demonstrating the heterologous desensitization of ERK induced by A 1 R agonist, we next tested whether the two upstream kinases of the MAPK cascade, Raf-1 and MEK, were also desensitized heterologously upon A 1 R activation. As shown in Fig. 3, A and B, CHO cells stably pretreatment, with immunoblots of total ERK2 used as the internal reference, respectively (B) Quantification of ERK1/2 phosphorylation levels obtained in A at different time points. The results were normalized against the level of ERK2, and further normalized against the result obtained from the control group. Error bars represent SEM, n = 5 (C) Time-dependent phosphorylation of ERK1/2 with DPDPE pretreatment, with immunoblots of total ERK2 used as the internal refer-ence, respectively (D) Quantification of ERK1/2 phosphorylation levels obtained in C at different time points. The results were normalized against the level of ERK2, and further normalized against the result obtained from the control group. Error bars represent SEM, n = 5 (E) Cells were pretreated with or without 100 µM DPCPX for 15 min, followed by treatment with 1 µM CHA or not for an additional 60 min, washed, then incubated with 10 nM DPDPE or not for an additional 5 min. Immunoblots were tested with anti-phospho-ERK1/2 and anti-ERK2 antibodies, and represent four independent experiments. Data represent means ± SEM, n = 4. *p < 0.05; t tests Fig. 4, A and B, phosphorylation of Raf-1-Ser289/296/301 significantly increased in a time-dependent manner after treatment with 1 µM CHA, in a time-course compatible with the desensitization of DOR-mediated Raf-1/MEK/ ERK activation. However, treatment of 10 nM DPDPE was not able to result in significant Raf-1-Ser289/296/301 phosphorylation (Fig. 4, C and D), thus ruling out negative feedback triggered by DOR activation itself. On the other hand, this phosphorylation effect could be inhibited by A 1 R antagonist DPCPX (100 µM, 15 min) but not DOR antagonist naloxone (100 µM, 15 min), indicating that prolonged CHA treatment resulted in inhibition of Raf-1 activity via Ser289/296/301 phosphorylation via a A 1 R-dependent and DOR-independent mechanism. We could also rule out the probability that the antagonists might directly affect Raf-1 phosphorylation by proving that neither DPCPX nor naloxone was able to alter the phosphorylation of Raf-1-Ser289/296/301 alone (Fig. 4E). Thus, we may conclude that CHA induced desensitization of the Raf-1/MEK/ERK pathway via phosphorylation of Raf-1-Ser289/296/301, in a A 1 R-dependent manner. regardless of the binding doses of its agonist. On the other hand, pretreatment with 10 nM DPDPE was not able to affect the phosphorylation status of either MEK1/2 or Raf-1-Ser338 induced by the following incubation with CHA (Fig. 3, F and G), indicating the unidirectional regulation of DOR by the A 1 R agonist. Thus, these results suggest that the Raf-1/MEK/ERK pathway undergoes significant heterologous desensitization upon activation of A 1 R, another member of the GPCR family.

CHA Induced Phosphorylation of Raf-1-Ser289/296/301 via a A 1 R-dependent Mechanism
Although we have shown the desensitization of DORmediated MAPK cascade after A 1 R activation initiated by CHA binding, the underlying mechanism remains unclear. To further probe into the detailed mechanism that controlled Raf-1 activity, we tested the phosphorylation of Raf-1-Ser289/296/301, which were considered as major inhibitory controlling phosphorylation sites that were regulated by negative feedback, and phosphorylation of these sites gives rise to inhibition of Raf-1 activity [15,17]. As shown in The results were normalized against the level of actin, and further normalized against the result obtained from the control group. Error bars represent SEM, n = 5. ***p < 0.001 compared with the pMEK1/2 level of the 0 min group; #p < 0.05, ###p < 0.001 compared with the Raf-1-pSer338 level of the 0 min group; t tests (C) Dose-dependent phosphorylation of Raf-1-Ser338 and MEK1/2 with CHA pretreatment. Cells were pretreated with or without 1 µM CHA or not for 1 h, washed, then incubated with or without increasing concentrations of DPDPE for an additional 5 min. Immunoblots were tested with anti-phospho-Raf-1-Ser338, anti-phospho-MEK1/2 and anti-actin antibodies. Immunoblots of total ERK2 was used as the internal reference. Error bars represent SEM, n = 4 (D and E) Quantification of pMEK1/2 (D) and pRaf-1-Ser338 (E) levels obtained in C with different DPDPE concentrations. The results were normalized against the level of actin, and further normalized against the result obtained from the control group. Error bars represent SEM, n = 5. *p < 0.05, **p < 0.01, ***p < 0.001 compared with the groups treated with the same concentration of DPDPE, but without CHA pretreatment; #p < 0.05, ###p < 0.001 compared with the control group without CHA pretreatment; t tests (F) Time-dependent phosphorylation of Raf-1-Ser338 and MEK1/2 with DPDPE pretreatment. Cells were pretreated with or without 10 nM DPDPE for indicated time periods, washed, then incubated with 1 µM CHA or not for an additional 5 min (G) Quantification of Raf-1-Ser338 and MEK1/2 phosphorylation levels obtained in F at different time points. The results were normalized against the level of actin, and further normalized against the result obtained from the control group. Error bars represent SEM, n = 5 phosphorylation, whereas CHA pretreatment was able to induce significant Raf-1-Ser289/296/301 phosphorylation, accompanied by attenuated ERK phosphorylation upon DPDPE treatment, which is considered a sign of heterologous desensitization of DOR-mediated ERK signaling. However, with the pretreatment of U0126, the ability of CHA to induce such desensitization was inhibited drastically, as the capability of DPDPE to stimulate ERK phosphorylation was restored. These results suggest that CHA could phosphorylate Raf-1-Ser289/296/301 through negative feedback of MEK/ERK activation, which in turn gives rise to desensitization of DOR-mediated ERK phosphorylation.
Since we have proven that U0126 could inhibit CHAinduced phosphorylation of Raf-1-Ser289/296/301 and desensitization of DPDPE-induced ERK phosphorylation, we infer that U0126 pretreatment might be able to inhibit desensitization of DOR-mediated activation of the entire Raf-1/MEK/ERK cascade, which was induced by CHA exposure. To test our hypothesis, we examined the activation of Raf-1, MEK1/2 and ERK1/2, after pretreatment of U0126 and CHA (Fig. 5, C-E). Cells were incubated with 10 µM U0126 for 30 min, followed by treatment with 1 µM CHA for 60 min, washed, and treated with 10 nM DPDPE for 5 min. The results showed that CHA pretreatment alone was able to inhibit phosphorylation of Raf-1-Ser338, MEK1/2 and ERK1/2, indicating desensitization of all three kinases of the MAPK cascade. However, this effect was inhibited markedly by U0126, suggesting that U0126 was able to inhibit CHA-induced heterologous desensitization of DOR-mediated Raf-1/MEK/ERK cascade.

U0126 Inhibited CHA-induced Phosphorylation of Raf-1-Ser289/296/301 and Heterologous Desensitization of DOR-mediated Raf-1/MEK/ERK Cascade
Since phosphorylation of Raf-1-Ser289/296/301 could be the direct consequence of negative feedback induced by MEK/ERK activation, we then detected whether U0126, the specific inhibitor of MEK which functionally antagonize AP-1 transcriptional activity via noncompetitive inhibition [19], was able to affect the phosphorylation status of Raf-1-Ser289/296/301 after CHA treatment. As shown in Fig. 5 A, cells were pretreated with 10 µM U0126 for 30 min, followed by incubation with 1 µM CHA for 60 min, and the phosphorylation status of Raf-1-Ser289/296/301 was detected by the western-blot assay. The result showed that although U0126 had no significant effect on its own, its pretreatment significantly blocked Raf-1-Ser289/296/301 phosphorylation, which was induced by CHA when U0126 was not present, indicating that the activity of MEK/ERK kinases were required for CHA-induced Raf-1-Ser289/296/301 phosphorylation, which is a prerequisite for ERK desensitization. Thus, in order to determine the final effect of U0126 on DPDPE-induced ERK activation, we next examined phosphorylation of ERK after the sequential pretreatments of U0126 and CHA (Fig. 5B). Cells were incubated with 10 µM U0126 for 30 min, followed by treatment with 1 µM CHA for 60 min, washed, and treated with DPDPE for 5 min. As a result, 1 µM CHA, but not 10 nM DPDPE, was able to induce Raf-1-Ser289/296/301 phosphorylation. U0126 alone had no significant effect on Raf-1-Ser289/296/301 Thus, we have proven that the heterologous desensitization was initiated by A 1 R-mediated activation of the Raf-1/MEK/ERK cascade. Activated MEK/ERK was able to inhibit Raf-1 activity via phosphorylation of
Our present study complies with the previous finding that prolonged A 1 R stimulation resulted in heterologous desensitization of DOR-mediated inhibition of intracellular cAMP levels and Akt phosphorylation [5]. It is worth noting that in this previous work, we found that desensitization of DOR-mediated signaling was not via direct downregulation of gene expression, but via modification of phosphorylation of DOR-Ser363, a phosphorylation site targeted by G-protein-coupled receptor kinase 2 (GRK2) and is essential for regulation of ERK activation patterns and functions [22]. In combination with our present study, we have demonstrated the mechanisms underlying desensitization of all levels in the GPCR-mediated signal transduction, including the receptor (DOR), the second messenger (cAMP), the protein kinase (Akt) and the downstream MAPK pathway (Raf-1/ MEK/ERK). We may thus infer that GPCR-mediated desensitization of signaling pathways are constantly modified by phosphorylation modification of proteins, rather than on a higher level, such as gene expression, thereby accounting for the rapidness and reversibility of desensitization induced by the interaction between GPCRs.
Although we have proven that DPDPE-induced Raf-1-Ser338 phosphorylation was inhibited by increased Ser289/298/301 phosphorylation, which is dependent on the Ser289/296/301 sites as a mechanism of negative feedback, thus undermining DOR-induced activation of Raf-1, along with its downstream kinases, MEK1/2 and ERK1/2, which finally contributed to physiological and behavioral effects, such as the nociceptive behavior (Fig. 7).

Discussion
In the current study, we found that prolonged CHA exposure resulted in downregulation of DOR-mediated Raf-1/ MEK/ERK signaling pathway, since exposure to 1 µM CHA for over 10 min was able to attenuated DPDPE-stimulated phosphorylation of Raf-1-Ser338/MEK/ERK. The phenomenon of heterologous desensitization relied on the inhibitory phosphorylation of Raf-1-Ser289/296/301, the crucial targets of the negative feedback induced by MEK/ ERK kinases, which were activated as a consequence of A 1 R activation upon binding of CHA. The desensitization of the MAPK pathway, however, could be blocked by MEK inhibitor U0126, which inhibited ERK activity and thereby cutting off the negative feedback circuit. With the presence of U0126, phosphorylation of Raf-1-Ser289/296/301 was inhibited, thus restoring the activity of Raf-1 stimulated by DOR, along with the downstream kinases of the MAPK cascade, MEK1/2 and ERK1/2. These results revealed a new mechanism that regulate heterologous desensitization of GPCR-mediated MAPK signaling, thus contributing to our activation of the MEK/ERK cascade, the underlying molecular mechanisms remain unclear. According to the findings that the phosphorylation of Ser338 and Ser289/298/301 may not occur simultaneously [15,23], we infer that conformational alterations of the Raf-1 kinase might account for the reciprocal inhibitory mechanism underlying the switching between the activated and the inhibited statuses of the kinase. The activated status features phosphorylation at Ser338 with unphosphorylated Ser289/298/301, whereas Ser338 phosphorylation is inhibited by Ser289/298/301 phosphorylation during the inhibition of Raf-1. Such regulation might be controlled by a certain upstream molecule, but the detailed mechanism remains elusive and should be revealed in the future.
Another issue worth noting is that although A 1 R and DOR are both members of the GPCR family, they present entirely different downstream effects, as ERK activated by A 1 R is able to induce Raf-1-Ser289/296/301 phosphorylation, but ERK activated by DOR has no such effect. In a similar case, we found in our previous study that DPDPE and H-Tyr-Tic-Phe-Phe-OH (TIPP), two agonists of DOR, also induce different downstream effects upon DOR activation [22]. A similar phenomenon was also discovered in the case of the µ-opioid receptor (MOR), as morphine and fentanyl may induce totally different patterns of cell signaling and physiological effects [24,25]. According to our analyses of the mechanisms underlying such biased agonism, the different pathways leading to ERK activation induced by distinct agonists, either PKC-dependent or β-arresindependent, may account for the localization and function of phosphorylated ERK [25]. Thus, we may hypothesize that different patterns of ERK activation downstream of A 1 R and DOR might also contribute to their different effects on Raf-1-Ser289/296/301 phosphorylation, and the direct underlying evidence should be confirmed in the future.
Despite the consistency in mechanisms underlying the desensitization of distinct signaling molecules downstream of DOR, we may nevertheless find discrepancies in the time course and extent of their desensitization. For example, ERK and Akt were desensitized rapidly, as their phosphorylation started to decline with CHA exposure of less than 10 min, and vanished with CHA pretreatment of 60 min. In comparison, desensitization of cAMP was relatively slow, with a half-life period of 2.56 h. Moreover, the heterologous desensitization of cAMP was incomplete [5]. The diversity of desensitization of distinct signaling molecules indicates that mechanisms underlying DOR desensitization are pathway-specific, suggesting that the desensitization of a certain signaling molecule not only depends on the functional declination of the receptor itself, but also on the status of other correlative signaling molecules [26]. Therefore, although heterologous desensitization of a GPCR and its downstream Fig. 7 A schematic diagram summarizing the signaling pathways underlying heterologous desensitization of DOR-mediated MAPK cascade induced by A 1 R activation The black solid lines summarize the conclusions of the present study revealing the mechanisms underlying the heterologous desensitization of the Raf-1/MEK/ERK cascade. CHA binds and activates A1R, which in turn activates the Raf-1/MEK/ERK pathway. Activated ERK1/2 in turn promotes phosphorylation of Raf-1-Ser289/296/301, which negatively modulates the activity of Raf-1, thereby inhibiting the capability of DOR in activating the Raf-1/MEK/ERK signaling pathway The gray dashed lines illustrate the hypotheses indicating mechanisms underlying DOR desensitization at the receptor level, which could be deduced from previous studies. (1) CHA could induce Ser363 phosphorylation of DOR, suggesting that the DOR and A1R may undergo cross-talk at receptor levels [5]; (2) MEK/ERK could induce Ser670 phosphorylation of G-protein-coupled receptor kinase 2 (GRK2) [22], the kinase that regulates DOR phosphorylation, and this effect was blocked by U0126, suggesting the regulation of DOR phosphorylation by MEK.
represents negative controlling sites that inhibit the activity of Raf-1. This negative feedback undermines the capability of DOR in activating the Raf-1/MEK/ERK signaling pathway, thereby inhibiting DOR-induced activation of MAPK cascade. Thus, activation of A 1 R inhibits DOR-mediated MAPK signaling pathway via heterologous desensitization. signaling pathways are considered as an entirety due to their functional integrity, we should nevertheless investigate their unique characteristics in regard to the detailed molecular mechanisms.
Our findings regarding the crosstalk between A 1 R-and DOR-mediated signaling pathways represent a new mechanism that control physiologic functions related to opioid receptors, including their nociceptive and non-nociceptive roles. For example, it has been discovered that activation of adenosine receptors was able to modulate the effect and tolerance of the opioid receptor system, which result in potentiation of morphine antinociception [27], visceral antinociception [28] and spinal cord nociceptive reflexes [29]. Moreover, since DOR has emerged as a promising target for pain treatment [30] due to the analgesic potency of its own agonists [31] or contribution to the antinociception effects of other drugs [18], investigations on mechanisms that control DOR activity, such as its interaction with A 1 R, is of great value in development of novel strategies for pain management. The non-nociceptive roles of opioid receptors in neurogenesis and cognitive functions of opioid receptors [32] have added extra significance to studies on mechanisms that modulate their signal transduction. For instance, DOR antagonists were able to promote neuronal differentiation and inhibit glial differentiation, suggesting its involvement in differentiation of adult neural stem/progenitor cells (NSPCs) [33]. Besides, the DOR system was also found to be involved in regulation of negative emotions including anxiety [34] and depression [35]. Based on all findings discussed above, we may infer that the factors modulating DOR functioning, including A 1 R, might also take part in the effects known as DOR-mediated consequences, both nociceptive and non-nociceptive. On the other hand, since A 1 R activity is also modulated in central nervous system diseases [36] and emotional abnormalities such as depression [37,38] and fear [39], we may also conjecture the potential roles of DOR in the modulating mechanisms underlying such disorders. Thus, extensive studies on A 1 R-DOR interaction may help expand our perspectives in searching novel targets for treatment of diseases that were originally considered as being mediated by one receptor alone. However, since the current study were carried out on the in vitro CHO cell system, further work is needed to test the conclusions by in vivo behavioral studies using animal models of drug addiction and cognitive disorders.

Conclusion
We demonstrate that CHA is able to activate the MEK/ERK pathway upon binding to A 1 R. Activated ERK1/2 in turn promotes phosphorylation of Raf-1-Ser289/296/301, which