Research ArticleGPCR SIGNALING

Plasma membrane localization of the μ-opioid receptor controls spatiotemporal signaling

See allHide authors and affiliations

Science Signaling  09 Feb 2016:
Vol. 9, Issue 414, pp. ra16
DOI: 10.1126/scisignal.aac9177

Spatiotemporal opioid receptor signaling

The μ-opioid receptor (MOR) is a GPCR that mediates the effects of endogenous opioids and opioid analgesics, such as morphine. Different MOR agonists produce different biological effects, in part by differentially regulating receptor phosphorylation and internalization. In cells transfected with MOR, Halls et al. examined downstream signaling in the absence of receptor internalization. Whereas the synthetic opioid DAMGO stimulated receptor movement within the plasma membrane and transiently increased ERK activity in both the cytosol and nucleus, morphine stimulated a protein kinase C–dependent pathway that restricted MOR movement and produced prolonged cytosolic ERK activity. Similar effects were observed in mouse dorsal root ganglion neurons, suggesting that the differences in plasma membrane mobility or clustering of MOR may underlie the differential effects of its agonists in vivo.

Abstract

Differential regulation of the μ-opioid receptor (MOR), a G protein (heterotrimeric guanine nucleotide–binding protein)–coupled receptor, contributes to the clinically limiting effects of opioid analgesics, such as morphine. We used biophysical approaches to quantify spatiotemporal MOR signaling in response to different ligands. In human embryonic kidney (HEK) 293 cells overexpressing MOR, morphine caused a Gβγ-dependent increase in plasma membrane–localized protein kinase C (PKC) activity, which resulted in a restricted distribution of MOR within the plasma membrane and induced sustained cytosolic extracellular signal–regulated kinase (ERK) signaling. In contrast, the synthetic opioid peptide DAMGO ([d-Ala2,N-Me-Phe4,Gly5-ol]-enkephalin) enabled receptor redistribution within the plasma membrane, resulting in transient increases in cytosolic and nuclear ERK activity, and, subsequently, receptor internalization. When Gβγ subunits or PKCα activity was inhibited or when the carboxyl-terminal phosphorylation sites of MOR were mutated, morphine-activated MOR was released from its restricted plasma membrane localization and stimulated a transient increase in cytosolic and nuclear ERK activity in the absence of receptor internalization. Thus, these data suggest that the ligand-induced redistribution of MOR within the plasma membrane, and not its internalization, controls its spatiotemporal signaling.

INTRODUCTION

Heterotrimeric guanine nucleotide–binding protein (G protein)–coupled receptors (GPCRs) are the largest family of cell surface signaling proteins encoded by the human genome. These receptors enable cells to respond to structurally diverse endogenous and environmental signals, and they are the targets of more than 30% of marketed drugs. It is increasingly recognized that the uniform increase in the abundance of second messengers throughout the cell cannot explain the diversity of GPCR-mediated effects. Rather, spatial (location) and temporal (duration) control of signaling play an important role (1, 2). Spatial compartmentalization of signaling can be achieved by the formation of GPCR-dependent protein complexes, which restrict the diffusion of second messengers to induce extremely localized signals (3). In addition, multiple regulatory mechanisms (including receptor phosphorylation, desensitization, and internalization) control the duration of GPCR activation. Therefore, the spatial and temporal distribution of both receptors and signaling effectors are critical for the generation of distinct and highly specialized GPCR-mediated responses.

The μ-opioid receptor (MOR) has been extensively studied because of its physiological importance in mediating the effects of endogenous opioids as well as its prominence as the target of opioid analgesics, such as morphine. However, chronic use of opioid analgesics is still clinically limited by the development of tolerance, addiction, constipation, and respiratory depression (4). At the cellular level, stimulation of MOR by all opioids activates the same G protein–dependent signaling pathways. MOR activates Gαi/o proteins, leading to inhibition of cyclic adenosine monophosphate (cAMP) generation, increased extracellular signal–regulated kinase (ERK) phosphorylation and activation, activation of G protein–regulated inwardly rectifying potassium channels (GIRKs), and inhibition of voltage-gated calcium channels (5). However, different MOR agonists induce distinct patterns of receptor regulation and internalization. In particular, morphine causes limited receptor phosphorylation and recruitment of the scaffolding protein β-arrestin, which results in compromised receptor internalization and resensitization (610). These observations have prompted intensive studies of the ability of MOR ligands to differentially activate G proteins and β-arrestins in an effort to explain their divergent biological effects (1113).

It is now apparent that the spatiotemporal characteristics of a signal can specify the outcome of receptor activation (1, 2). Most opioids, including morphine, result in the phosphorylation and activation of cytosolic ERK (1416); however, unlike other opioids, morphine is unable to promote nuclear ERK activation (15). Together with its impaired ability to induce the internalization of MOR, this finding suggests that morphine may stimulate a spatiotemporal cellular response distinct from those induced by other opioids. To investigate this hypothesis, we used complimentary biophysical techniques and superresolution microscopy. We report that morphine and DAMGO activated distinct spatial and temporal signaling profiles that were controlled by the plasma membrane localization of MOR induced by the two ligands. Subcellular targeted Förster resonance energy transfer (FRET) biosensors showed that only morphine-dependent stimulation of MOR induced sustained cytosolic ERK phosphorylation and plasma membrane–localized protein kinase C (PKC) activation, which restricted the localization of MOR. In contrast, DAMGO caused the redistribution of MOR within the plasma membrane as well as the transient activation of both cytosolic and nuclear ERK. Thus, not only did morphine and DAMGO stimulate different signaling pathways, but they also activated signals in distinct subcellular compartments with distinct temporal profiles. Furthermore, we altered the spatiotemporal signaling profile of morphine to mimic that of DAMGO by enabling the redistribution of MOR within the plasma membrane in the absence of β-arrestin recruitment or receptor internalization. Thus, these data suggest that receptor localization within the plasma membrane determines the spatiotemporal signals activated by MOR in response to different ligands.

RESULTS

Ligand-dependent spatiotemporal signaling of MOR

To gain spatial and temporal resolution of MOR signaling in live cells, we used FRET biosensors for ERK and PKC (EKAR and CKAR, respectively), which localized to different subcellular compartments (17, 18). In human embryonic kidney (HEK) 293 cells cotransfected with plasmids encoding MOR and either a cytosolic or a nuclear ERK biosensor (cytoEKAR and nucEKAR, respectively), DAMGO and morphine at EC50 concentrations (concentration that provokes 50% of the maximal response of the receptor; 10 nM DAMGO and 100 nM morphine, fig. S1A) caused distinct temporal profiles of ERK activation, shown as changes in the corresponding FRET signal. Whereas DAMGO caused a transient increase in cytosolic ERK activity, morphine induced a sustained increase (Fig. 1, A and B). Moreover, only DAMGO caused a transient increase in nuclear ERK activity (Fig. 1, C and D). Ligand-dependent responses were also observed when assessing the direct activation of PKC. In cells cotransfected with plasmids encoding MOR and a plasma membrane–localized biosensor of PKC activity (pmCKAR), only morphine caused a sustained increase in PKC activity (Fig. 1E). DAMGO did not affect the activity of plasma membrane–localized PKC, even at maximal concentrations (1 μM, fig. S1B), and neither ligand stimulated the activity of cytosolic PKC (Fig. 1F).

Fig. 1 Ligand-dependent spatiotemporal signaling of MOR.

(A to D) Spatiotemporal activation of ERK in transfected HEK 293 cells treated with vehicle, DAMGO, or morphine for the indicated times. (A) Analysis of cytosolic ERK activity. Data are means ± SEM of 416 to 606 cells from five experiments. (B) Representative pseudocolor ratiometric images of cytoEKAR. (C) Analysis of nuclear ERK activity. Data are means ± SEM of 561 to 810 cells from five experiments. (D) Representative pseudocolor ratiometric images of nucEKAR. Pseudocolor scale as in (B). (E and F) Spatiotemporal activation of PKC in transfected HEK 293 cells after treatment with vehicle, DAMGO, or morphine for the indicated times. (E) Analysis of plasma membrane–localized PKC activity. Data are means ± SEM of 155 to 220 cells from three experiments. (F) Analysis of cytosolic PKC activity. Data are means ± SEM of 45 to 115 cells from three experiments.

The distinct internalization profiles of MOR in response to DAMGO and morphine (6, 10) were quantified with a bioluminescence resonance energy transfer (BRET) assay that detects the proximity between BRET partners in defined subcellular compartments in live cells (19, 20). Consistent with previous reports, incubation of HEK 293 cells with 1 μM DAMGO (concentration corresponding to its EC50 of β-arrestin recruitment using BRET; fig. S2A) induced MOR internalization, as shown by the increase in the BRET signal between a Renilla luciferase–tagged MOR (MOR-RLuc) and a Venus-tagged marker of early endosomes (Rab5a-Venus) (Fig. 2A). In contrast, morphine produced no substantial change in BRET (Fig. 2A and fig. S2B). These results were validated by automated, high-content image analysis (fig. S2C). DAMGO-mediated MOR endocytosis was unaffected by the inhibition of Gαi/o activity with NF023 or pertussis toxin (PTx) (21, 22) but was abolished by the clathrin-dependent endocytosis inhibitor PitStop2 (23), by the expression of a dominant-negative dynamin mutant (K44E) (24), or by knockdown of β-arrestins [combined small interfering RNAs (siRNAs) specific for β-arrestin-1 and β-arrestin-2] (Fig. 2, A and B, and fig. S2, D to H). These data suggest that β-arrestin recruitment and MOR endocytosis are independent of Gαi/o coupling.

Fig. 2 Effect of Gαi/o protein inhibition, β-arrestin knockdown, or inhibition of endocytosis on MOR-stimulated cytosolic and nuclear ERK activities.

(A and B) The trafficking of MOR to early endosomes in transfected HEK 293 cells in response to treatment with vehicle, DAMGO, or morphine for 30 min was determined by BRET analysis between MOR-RLuc and Rab5a-Venus. (A) Cells were treated with the clathrin-mediated endocytosis inhibitor PitStop2 (PS2) or its inactive control, or transfected to express wild-type (WT) dynamin or a dominant-negative dynamin K44E mutant. (B) Cells were treated with or without siRNAs specific for β-arrestins or were preincubated with the indicated Gαi/o protein inhibitors. Data are means ± SEM from five experiments. (C to F) Analysis of the spatial activation of ERK in HEK 293 cells in response to vehicle, DAMGO, or morphine in the presence or absence of β-arrestin–specific siRNA, Gαi/o protein inhibitors, PitStop2 or its inactive control, or upon expression of WT or K44E mutant dynamin. (C) Analysis of cytosolic ERK activity. Data are means ± SEM of 96 to 168 cells from three experiments. (D) Analysis of cytosolic ERK activity in cells in which endocytosis was inhibited. Data are means ± SEM of 35 to 606 cells from three experiments. (E) Analysis of nuclear ERK activity in cells treated with Gαi/o protein inhibitors or β-arrestin–specific siRNAs. Data are means ± SEM of 52 to 258 cells from three experiments. (F) Analysis of nuclear ERK activity in cells in which endocytosis was inhibited. Data are means ± SEM of 51 to 306 cells from three experiments. *P < 0.05, **P < 0.01, and ***P < 0.001 versus vehicle control. Data were analyzed by two-way analysis of variance (ANOVA) with Tukey’s multiple comparison test. AUC, area under the curve; scram., scrambled.

Previous studies have linked PKC activation to cytosolic ERK activity and β-arrestin activation to increased nuclear ERK activity to conclude that G protein– and β-arrestin–dependent pathways activate distinct modes of ERK signaling (15). By inhibiting Gαi/o proteins, we demonstrated that cytosolic ERK activation in response to DAMGO and morphine was dependent on Gαi/o (Fig. 2C). In agreement with previous studies, cytosolic ERK activity was unaffected by knockdown of β-arrestins (Fig. 2C). However, inhibition of receptor endocytosis by PitStop2 or by expression of the dynamin K44E mutant transformed the profile of DAMGO-induced cytosolic ERK activity from a transient to a sustained signal, consistent with the retention of MOR at the plasma membrane (Fig. 2D and fig. S2, I and J). As expected, the increase in nuclear ERK activity in response to DAMGO was dependent on β-arrestins and receptor internalization (Fig. 2, E and F). Thus, our results suggest that Gαi/o activation by MOR mediates increases in cytosolic ERK activity in response to DAMGO and morphine, and confirm that the increases in nuclear ERK activity in response to DAMGO are dependent on β-arrestins and receptor endocytosis.

Role of PKC activation in the spatiotemporal profile of ERK in response to morphine

Inhibition of the activity of Gαi/o subunits (with NF023 or PTx) or of Gβγ subunits [with the cell-permeable N-myristoylated Gβγ-selective peptide mSIRK or by expression of βARKct, a GPCR kinase 2 (GRK2) C-terminal peptide that interferes with Gβγ function] (25, 26) abolished the response of plasma membrane–localized PKC to morphine (Fig. 3A); however, knockdown of β-arrestins and negative controls (inactive mSIRK L9A and scrambled siRNA) had no such effects (Fig. 3A and fig. S3A). Thus, the sustained increase in plasma membrane–localized PKC activity that was stimulated by morphine was mediated by Gαi/o and Gβγ subunits.

Fig. 3 The role of PKC activation by morphine in the spatiotemporal control of ERK activity.

(A) The effects of the indicated G protein inhibitors or inactive controls on plasma membrane PKC activity in HEK 293 cells treated with vehicle, DAMGO, or morphine were determined with the pmCKAR FRET biosensor. Data are means ± SEM of 39 to 229 cells from three experiments. (B to D) Analysis of the MOR-stimulated spatiotemporal activation of ERK in response to vehicle, DAMGO, or morphine in cells in which Gβγ or PKC signaling was inhibited. (B) Analysis of cytosolic ERK activity over time. Data are means ± SEM of 31 to 101 cells from three experiments. (C) Analysis of nuclear ERK activity over time. Data are means ± SEM of 74 to 126 cells from three experiments. (D) Nuclear ERK activity was analyzed as the AUC. Data are means ± SEM of 22 to 360 cells from three experiments. (E and F) MOR trafficking in response to vehicle, DAMGO, or morphine was monitored in HEK 293 cells in the presence or absence of the indicated Gβγ and PKC inhibitors. (E) Analysis of BRET between MOR-RLuc and Rab5a-Venus. Data are means ± SEM of three to seven experiments. (F) Analysis of BRET between MOR-RLuc and KRas-Venus. Data are means ± SEM of three to seven experiments. (G to I) Effect of the indicated phosphorylation site mutations on MOR trafficking and nuclear ERK activity. (G) Analysis of BRET between MOR-RLuc8 and β-arrestin-2–YFP (yellow fluorescent protein). Data are means ± SEM of three to seven experiments. (H) Analysis of BRET between MOR-RLuc8 and Rab5a-Venus. Data are means ± SEM of three or four experiments. (I) Analysis of nuclear ERK activity over time. Data are means ± SEM of 87 to 359 cells from three to five experiments. P < 0.05, P < 0.01, and P < 0.001 versus vehicle control. Data were analyzed by two-way ANOVA with Tukey’s (A and D) or Dunnett’s (E to I) multiple comparison tests. GFx, GF109203X.

Previous studies reported that PKC activity mediates the increased activity of cytosolic ERK in response to morphine (15). We therefore investigated whether the Gαi/o-Gβγ-PKC pathway influenced the distinct ERK spatiotemporal signaling profiles of MOR. Rather than decreasing ERK activity, and in contrast to previous reports, inhibition of Gβγ subunits or of PKC (with GF109203X or Gö6983) (27, 28) transformed the temporal profile of morphine-stimulated cytosolic ERK activity to resemble the transient response induced by DAMGO (Fig. 3B and fig. S3, B and C). Moreover, inhibition of the Gβγ-PKC pathway also enabled morphine to increase the activity of nuclear ERK (Fig. 3, C and D). Previous studies implicated PKCα, PKCγ, and PKCε as the PKC isoforms that contribute to morphine signaling and to the development of morphine tolerance (16, 2932). Of these, only the mRNAs of PKCα and PKCε were present in our HEK 293 cell line (fig. S3D). Inhibition of PKCα (with Gö6976, which targets PKCα and PKCβ1) (33), but not PKCε (with iPKCε, a cell-permeable PKCε inhibitory peptide) (34), transformed the temporal profile of morphine-stimulated cytosolic ERK activity from being sustained to be being transient and also facilitated an increase in nuclear ERK activity (fig. S3, E and F). There were no effects of inactive controls or these inhibitors on the responses of cells to DAMGO (Fig. 3, B to D, and fig. S3, C and F).

As expected, the inhibition of Gβγ subunits or PKC did not substantially affect the recruitment of β-arrestin-2 or MOR internalization in response to DAMGO, as determined by BRET analysis and high-content imaging (Fig. 3E and fig. S3, G to I). In contrast, upon inhibition of Gβγ subunits or PKC, the activation of MOR by morphine resulted in a decrease in BRET between MOR-RLuc and the plasma membrane marker KRas-Venus (Fig. 3F), which suggested that there was an increase in the distance between these two proteins. In the absence of MOR internalization (Fig. 3E and fig. S3, G and H), the morphine-stimulated change in BRET between MOR and KRas may indicate a movement of the receptor away from KRas within the plasma membrane. Thus, the transient activation of cytosolic and nuclear ERK elicited by morphine did not require MOR internalization but may instead depend on the translocation of MOR within the plasma membrane.

The importance of the localization of MOR within the plasma membrane for the control of spatiotemporal signaling was also supported by the effects of the expression of a phosphorylation-deficient MOR mutant (S375A) (35). DAMGO still induced the recruitment of β-arrestin-2 to MOR S375A; however, the receptor was not internalized as determined by high-content imaging or analysis of BRET between the receptor and Rab5a (Fig. 3, G and H, and fig. S3, G and H). There was no change in the BRET between MOR S375A and KRas in response to DAMGO or morphine (fig. S3J); however, stimulation of MOR S375A by either DAMGO or morphine induced transient increases in cytosolic and nuclear ERK activity (Fig. 3I and fig. S3K).

To confirm that receptor phosphorylation was key for the control of the plasma membrane localization of MOR and its spatiotemporal signaling, we used a phosphorylation-deficient MOR mutant in which all of the C-terminal serine and threonine residues were mutated to alanines (11ST/A) (9). Consistent with previous reports, MOR 11ST/A was not internalized, as determined by measurement of BRET with Rab5a, nor did it recruit β-arrestin-2 in response to DAMGO (Fig. 3, G and H). However, stimulation of MOR 11ST/A by either DAMGO or morphine induced a transient increase in nuclear ERK activity, with no accompanying change in BRET between the receptor and KRas (Fig. 3I and fig. S3J). Phosphorylation of Ser375 therefore appears to be critical for the control of the spatiotemporal signaling by MOR in response to morphine. Together, these data suggest that the impaired trafficking of MOR mutants results in an altered signaling profile and support the hypothesis that the plasma membrane localization of MOR, and not β-arrestin recruitment or receptor internalization, plays a key role in the spatiotemporal control of receptor signaling.

Ligand-dependent redistribution of MOR within the plasma membrane

To investigate the changes in MOR distribution elicited by morphine upon inhibition of the Gβγ-PKCα pathway, we assessed receptor localization at the plasma membrane by confocal microscopy and subcellular fractionation. After 10 min of stimulation of the MOR (when all signaling pathways are activated), there was no substantial colocalization between the receptor and immunolabeled clathrin as determined by confocal microscopy analysis under any condition tested (fig. S4, A and B). However, 60 min of stimulation with DAMGO, but not morphine, caused substantial colocalization between MOR and clathrin (fig. S4C). In contrast, activation of the fast internalizing β2-adrenergic receptor (β2AR) by isoprenaline caused substantial receptor-clathrin colocalization after only 10 min of stimulation (fig. S4, A to C). Similarly, there was no effect of DAMGO or morphine on the location of FLAG-tagged MOR (FLAG-MOR) within non–lipid-rich (that is, Triton X-100–soluble) plasma membrane domains through basic lipid fractionation analysis (fig. S4D). Therefore, the distinct spatiotemporal signaling profiles of morphine and DAMGO do not reflect either ligand-dependent MOR clustering in clathrin-coated pits or translocation of the receptor to different lipid domains.

To investigate the localization of MOR within the plasma membrane with increased resolution, we used ground-state depletion (GSD) superresolution microscopy in total internal reflection fluorescence (TIRF) mode. GSD-TIRF enables the detection of events within the plane of the plasma membrane to an axial resolution of 100 nm. This approach can measure the distance between an event (for example, a receptor or receptor clusters) and its nearest neighbor across a population. Stimulation of FLAG-MOR with DAMGO for 10 min increased the average distance between detected events (Fig. 4, A and B), which suggested the redistribution of MOR within the plasma membrane. This increase in distance occurred before, and was independent of, receptor internalization, because the dominant-negative dynamin K44E mutant had no effect (fig. S4, E and F).

Fig. 4 DAMGO induces a distinct MOR distribution at the plasma membrane.

GSD-TIRF microscopy was used to monitor the plasma membrane distribution of FLAG-MOR in HEK 293 cells in response to treatment with vehicle, DAMGO, or morphine for 10 min. (A) Representative GSD-TIRF images and Euclidean distance maps (EDMs) from cells under the indicated conditions. Scale bars, 1 μm. (B) Average distances to nearest neighbors in the cells shown in (A). Data are means ± SEM of three to nine experiments. (C) Average distances to nearest neighbors in cells subjected to Gβγ inhibition by preincubation with mSIRK and then treated with vehicle, DAMGO, or morphine. Data are means ± SEM of three to nine experiments. (D) Representative GSD-TIRF images and EDMs from cells preincubated with mSIRK to inhibit Gβγ and then subjected to the indicated treatments. Scale bars, 1 μm. Pseudocolor scale is as described in (A). (E) Representative GSD-TIRF images and EDMs of unstimulated cells expressing either WT MOR or the MOR S375A mutant. Scale bars, 1 μm. Pseudocolor scale is as described in (A). (F) Average distances to nearest neighbors in the cells shown in (E). Data are means ± SEM of three to nine experiments. P < 0.01 and P < 0.001 versus vehicle control. Data were analyzed by one-way ANOVA with Dunnett’s multiple comparison test (B and C) or unpaired t test (E).

The stimulation of cells expressing FLAG-MOR with morphine for 10 min did not change the average distance between events (Fig. 4, A and B); however, after the inhibition of Gβγ subunits, morphine increased the distance between detected MOR events (Fig. 4, C and D, and fig. S4, E and G), suggesting that activation of this pathway by morphine normally restricts MOR localization. Furthermore, the distance between MOR events under basal conditions in cells expressing MOR S375A was also increased when compared to that in cells expressing the wild-type receptor (Fig. 4, E and F). This increase in distance between events was not a result of decreased receptor abundance at the plasma membrane (MOR S375A 510,000 sites per cell, MOR wild-type 121,000 sites per cell measured by whole-cell [3H]diprenorphine binding; table S1), confirming that MOR S375A was differentially distributed compared to the wild-type receptor.

Thus, our results suggest that the activation of MOR by morphine restricts receptor localization, whereas DAMGO stimulates the redistribution of MOR within the plasma membrane. Disruption of the Gβγ-PKCα pathway enabled morphine to stimulate a DAMGO-like redistribution of MOR but did not result in receptor internalization. Receptor redistribution preceded endocytosis (in the case of DAMGO) or occurred independently of endocytosis (in the case of morphine), and it appeared to control the ability of MOR to transiently activate cytosolic and nuclear ERK.

Effect of plasma membrane organization on MOR spatiotemporal signaling

To confirm the importance of membrane organization in the control of compartmentalized MOR signaling, we depleted cholesterol from the plasma membrane with methyl-β-cyclodextrin (MβCD) (36) or filipin III (37). Neither reagent had any substantial effect on MOR internalization, as determined by high-content imaging (fig. S5, A and B); however, both MβCD and filipin III abolished the distinct spatiotemporal signaling profiles of morphine and DAMGO (Fig. 5 and fig. S5). Upon cholesterol depletion, both morphine and DAMGO increased PKC activity at the plasma membrane and caused a transient increase in both cytosolic and nuclear ERK activity (Fig. 5 and fig. S5, C to F). Indeed, replenishment of membrane cholesterol by incubation of the cells with MβCD-cholesterol complexes completely restored the original spatiotemporal signaling profiles of DAMGO and morphine (Fig. 5 and fig. S5). Thus, disruption of membrane organization altered the spatiotemporal signaling profiles of MOR, with no change in the ability of the receptor to be internalized, suggesting that the plasma membrane localization of MOR plays an important role in determining its spatiotemporal signaling.

Fig. 5 Disruption of membrane architecture alters MOR signaling profiles.

(A to F) HEK 293 cells were pretreated with vehicle, MβCD, or MβCD-cholesterol complexes (MβCD/choles.) before being treated with vehicle, DAMGO, or morphine. The spatiotemporal activation of plasma membrane–localized PKC and either cytoplasmic or nuclear ERK was then measured. (A) Plasma membrane–localized PKC activity in response to DAMGO. Data are means ± SEM of 40 to 174 cells from three experiments. (B) Cytosolic ERK activity in response to DAMGO. Data are means ± SEM of 30 to 167 cells from three experiments. (C) Nuclear ERK activity in response to DAMGO. Data are means ± SEM of 68 to 230 cells from three experiments. (D) Plasma membrane–localized PKC activity in response to morphine. Data are means ± SEM of 41 to 195 cells from three experiments. (E) Cytosolic ERK activity in response to morphine. Data are means ± SEM of 32 to 194 cells from three experiments. (F) Nuclear ERK activity in response to morphine. Data are means ± SEM of 80 to 217 cells from three experiments.

MOR compartmentalized signaling in dorsal root ganglion neurons

To confirm the physiological relevance of the spatiotemporal signaling patterns of MOR that we determined in HEK 293 cells, we used nucleofection to transfect isolated neurons from mouse dorsal root ganglions (DRGs) with the FRET biosensors. DRG neurons are the principal mediators of nociception from the periphery to the spinal cord, and activation of endogenous MOR in these neurons partially mediates the analgesic actions of opioids (38).

Activation of MOR in DRG neurons stimulated ERK and PKC activity with spatiotemporal profiles that were identical to those observed in HEK 293 cells. DAMGO caused a transient increase in both cytosolic and nuclear ERK activities, whereas morphine elicited a sustained increase in the activities of cytosolic ERK and plasma membrane–localized PKC (Fig. 6, A to C). Inhibition of PKC decreased the percentage of neurons (from 75 to 49%) that exhibited sustained cytosolic ERK activity in response to morphine and increased the percentage of neurons (from 25 to 51%) that exhibited a transient increase in cytosolic ERK activity (Fig. 6, D and E). There was no effect of PKC inhibition on the temporal profile of cytosolic ERK activity after stimulation with DAMGO (Fig. 6, D and E). As was observed in HEK 293 cells, inhibition of PKC enabled morphine to stimulate nuclear ERK activity in DRG neurons (Fig. 6F).

Fig. 6 Spatiotemporal signaling of endogenous MOR in DRG neurons.

(A to F) Analysis of the spatiotemporal activation of ERK and PKC in DRG neurons treated with vehicle, DAMGO, or morphine. (A) Cytosolic ERK activity. Data are means ± SEM of 56 to 120 neurons from three experiments. (B) Nuclear ERK activity. Data are means ± SEM of 45 to 64 neurons from three experiments. (C) Plasma membrane–localized PKC activity. Data are means ± SEM of 40 to 55 neurons from three experiments. (D) Effect of PKC inhibition on cytosolic ERK activity. Data are means ± SEM of 86 to 99 neurons from three experiments. (E) Population analysis of the temporal profile of cytosolic ERK activity. The numbers of neurons in each group are indicated. (F) Effect of PKC inhibition on nuclear ERK activity. Data are means ± SEM of 25 to 73 neurons from three experiments. (G to I) GSD-TRIF–based analysis of the plasma membrane distribution of endogenous MOR in DRG neurons in response to treatment with vehicle, DAMGO, or morphine for 10 min. Data are means ± SEM of 9 to 15 cells from three experiments. (G) Isolated DRG neuron immunostained for MOR (green) and tubulin βIII (magenta). Scale bar, 10 μm. (H) Representative GSD-TIRF images and EDMs of DRG neurons under the indicated conditions. Scale bars, 1 μm. (I) Average distances to nearest neighbors. Data are means ± SEM of 9 to 15 cells from three experiments. P < 0.05, P < 0.01, and P < 0.001 versus vehicle control. Data were analyzed by two-way ANOVA with Tukey’s multiple comparison test (F) or one-way ANOVA with Dunnett’s multiple comparison test (I).

We also assessed the distribution of endogenous MOR at the plasma membrane of DRG neurons by GSD-TIRF microscopy (Fig. 6G). As was observed in HEK 293 cells, stimulation of endogenous MOR in DRG neurons with DAMGO increased the distance between detected events at the plasma membrane (Fig. 6, H and I). In contrast, there was no change in the distance between MOR events in response to morphine. Thus, in DRG neurons, as in HEK 293 cells, receptor redistribution at the plasma membrane correlated with transient increases in cytosolic and nuclear ERK activities in response to DAMGO. Moreover, inhibition of PKC enabled morphine to cause transient increases in cytosolic and nuclear ERK activities. As such, the spatiotemporal regulation of MOR activation and signaling identified in recombinant expression systems also occurred in DRG neurons endogenously expressing this receptor.

DISCUSSION

The use of biophysical approaches to assess MOR signaling in real time and in live cells revealed a previously uncharacterized mechanism that contributes to the control of differential MOR activation. Here, we showed that the activation of MOR by DAMGO stimulated the translocation of the receptor within the plasma membrane. This translocation preceded receptor trafficking to clathrin-containing domains and internalization, and is likely dependent on receptor phosphorylation (Fig. 7A). This translocation, but not internalization, of MOR determined the transient cytosolic ERK activity profile and the activation of nuclear ERK (Fig. 7A). In contrast, morphine activated plasma membrane–localized PKCα through Gβγ subunits, which prevented receptor translocation within the plasma membrane. This resulted in the sustained activity of cytosolic ERK, but not nuclear ERK (Fig. 7B). Inhibition of this Gβγ-PKCα pathway enabled the morphine-activated MOR to translocate within the plasma membrane, thus transforming its spatiotemporal signaling profile (Fig. 7B). Furthermore, this altered signaling profile mimicked that of the internalizing ligand DAMGO (that is, it was characterized by transient cytosolic and nuclear ERK activity) but occurred in the absence of β-arrestin-2 recruitment and receptor internalization.

Fig. 7 Plasma membrane localization controls MOR spatiotemporal signaling.

(A) DAMGO causes recruitment of GRK2 and β-arrestin-2 (i), facilitating MOR redistribution across the plasma membrane and transient activation of Gαi/o-mediated cytosolic ERK and Gαi/o-independent nuclear ERK (ii). Upon prolonged stimulation of MOR, DAMGO stimulates MOR clustering and receptor internalization through clathrin-coated pits (iii) to early endosomes (iv). (B) Morphine stimulates the plasma membrane–localized Gβγ-PKCα pathway, which prevents receptor translocation within the plasma membrane. This causes a sustained activation of Gαi/o-mediated cytosolic ERK (i). Inhibition of the Gβγ-PKCα pathway or alteration of the organization of the plasma membrane facilitates MOR translocation and the activation of nuclear ERK by morphine (ii) in the absence of receptor internalization.

These results add details to previous descriptions of ligand-dependent differences in ERK signaling (1416). Previous studies used Western blotting analysis to show that etorphine-induced ERK phosphorylation was dependent on β-arrestins, whereas morphine activated ERK through a PKC-dependent pathway (15). However, we showed that upon PKC inhibition, morphine stimulated ERK phosphorylation, although this signal had different temporal dynamics and occurred in both the cytosol and the nucleus (Figs. 3 and 7B). Therefore, the activation of cytosolic ERK by morphine is not PKC-dependent, but rather PKC, by controlling the localization of MOR, likely determines the dynamics and location of this response. It is interesting to consider that in the context of a whole cell after its solubilization (with a relatively greater contribution of cytosolic compared to nuclear ERK), this altered temporal profile could appear as an apparent decrease in morphine-stimulated ERK activity. This illustrates the extra mechanistic detail that can be obtained by resolving spatial and temporal signaling dynamics in live cells. We therefore propose that the plasma membrane organization of MOR, and not just the recruitment of β-arrestin and subsequent receptor internalization, determines the spatiotemporal outcome of receptor activation. Furthermore, these mechanisms appear to operate in nociceptive neurons and may thus contribute to the analgesic actions of opioids.

The ability of DAMGO, but not morphine, to cause receptor redistribution may relate to differential patterns of MOR phosphorylation. Although all opioids cause the phosphorylation of MOR at Ser375, this event is mediated by different kinases depending on the ligand (9, 39). Previous studies showed that the DAMGO-activated MOR is phosphorylated by GRK2 and GRK3 and that internalizing ligands stimulate the higher-order phosphorylation of flanking residues, which results in efficient β-arrestin recruitment and receptor internalization (9). Here, we showed that the recruitment of β-arrestin-2, translocation of MOR, and activation of nuclear ERK in response to DAMGO preceded receptor internalization. As such, we hypothesize that the differential recruitment of regulatory proteins (including GRKs and β-arrestins) to MOR may underlie receptor redistribution at the plasma membrane, and thus indirectly control spatiotemporal signaling. This hypothesis is supported by the finding that mutation of the key hierarchical phosphorylation site of MOR (to generate the MOR S375A mutant) affected the localization of the receptor within the plasma membrane and its spatiotemporal signaling. In this context, β-arrestins are increasingly recognized as scaffolding proteins for signaling complexes, in addition to their traditional roles in the regulation of receptor desensitization and internalization (40). Furthermore, evidence suggests that GRKs can also have important scaffolding functions, particularly in the control of ERK activation (41, 42). We hypothesize that the differential assembly of receptor kinases and other signaling mediators in response to morphine versus DAMGO determines MOR redistribution, transient signaling profiles, and the activation of nuclear ERK. Furthermore, this hypothesis entails that the responses of opioid ligands will be highly dependent on the specific protein content of opioid-responsive cells (6, 7, 43, 44).

Our results also highlight the importance of PKCα in governing the spatiotemporal signaling profiles of MOR. Previous studies showed that the phosphorylation and desensitization of MOR after stimulation with morphine is partially dependent on PKC (39, 45, 46). Moreover, there are indications that PKC plays a substantial role in the initiation and maintenance of tolerance to morphine analgesia (47, 48). To date, evidence for morphine-induced activation of PKC comes from coimmunoprecipitation studies showing the recruitment of overexpressed PKCε to MOR (16) and increased PKC activity in cell lysates (49). By measuring endogenous PKC activity at the subcellular level, we demonstrated that morphine, but not DAMGO, stimulated the sustained activation of PKC at the plasma membrane. Whereas PKC can phosphorylate MOR directly (32, 50), it can also phosphorylate proteins that participate in MOR signaling, such as Gαi (51) or GRK2 (52), and could therefore restrict receptor redistribution by modulating the function, association, or both of such signaling and scaffolding proteins with MOR.

It is clear that plasma membrane organization plays a critical role in the control of MOR spatiotemporal signaling. Whether MOR resides within biochemically defined, lipid-rich plasma membrane regions is controversial (5355). However, and consistent with our findings, previous studies provided evidence for the restricted plasma membrane localization of the receptor as well as the agonist-regulated diffusion of MOR within the plasma membrane (5659). Protein-protein interactions were hypothesized to mediate the restricted and slow diffusion of agonist-stimulated, noninternalizing MOR (60). Together with this hypothesis, the results presented here suggest that the dynamic organization of MOR within the plasma membrane, rather than the association of MOR with a predefined lipid-rich domain, may control ligand-dependent receptor redistribution and distinct spatiotemporal signaling profiles. The dependence of MOR signaling on plasma membrane localization extends previous studies that demonstrated distinct control of spatiotemporal signaling by endosomally localized GPCRs (2, 61). In the context of MOR, mechanistic insight into the actions of morphine at the cellular level is of particular therapeutic relevance because of the severe side effects induced by this opiate. Whether chronic exposure to opiates differentially alters the spatiotemporal signaling, plasma membrane distribution, or both of MOR remains to be investigated.

MATERIALS AND METHODS

Reagents

DAMGO was obtained from Mimotopes. Morphine and anti–FLAG M2 were from Sigma-Aldrich. Coelenterazine h was obtained from Promega. β-Arrestin-1– and β-arrestin-2–specific siRNAs were purchased from GE-Dharmacon. Mouse anti–early endosome antigen 1 was from BD Transduction Laboratories; rabbit anti-MOR (UMB-3) was from Abcam; mouse anti–tubulin βIII isoform for confocal imaging was from Merck Millipore; Alexa Fluor–conjugated goat anti-mouse secondary antibodies were from Jackson ImmunoResearch; anti–caveolin 1, anti–β-actin, and anti–clathrin heavy chain were from Abcam; anti–β-tubulin for Western blotting was from Santa Cruz Biotechnology; antibodies against β-arrestin-1/2 were from Cell Signaling Technology; and fluorescent IRDye–conjugated goat anti-rabbit (800 channel) and anti-mouse (680 channel) secondary antibodies were from LI-COR Biotechnology.

Complementary DNAs

Plasmids encoding KRas-Venus, Rab5a-Venus, green fluorescent protein (GFP)–dynamin, and GFP-dynamin K44E have been previously described (19, 61, 62). MOR-RLuc was from L. Bohn (Scripps, Jupiter, FL); FLAG-MOR was from M. Christie (University of Sydney, Sydney, New South Wales, Australia); β-arrestin-2–YFP was from M. Caron (University of North Carolina); and FLAG-MOR 11ST/A was from S. Schulz (Friedrich Schiller University, Jena, Germany). The following constructs were obtained from Addgene: cytoEKAR GFP/RFP (plasmid 18680), cytoEKAR Cerulean/Venus (plasmid 18679), nucEKAR GFP/RFP (plasmid 18682) and nucEKAR Cerulean/Venus (plasmid 18681) (17), and cytoCKAR (plasmid 14870) and pmCKAR (plasmid 14862) (18). MOR S375A has a mutation of the essential site governing hierarchical phosphorylation (human, S377A; mouse, S375A) (9) and was generated with the QuikChange site-directed mutagenesis kit. RLuc8-tagged MOR was generated by subcloning MOR into the pcDNA3-RLuc8 vector.

Cell culture and inhibitors

HEK 293 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 5% (v/v) fetal bovine serum (FBS). Cells were transfected with linear polyethyleneimine (63). For siRNAs, cells were transfected with scrambled or combined β-arrestin-1 and β-arrestin-2 SMARTpool ON-TARGETplus siRNAs (25 nM) with Lipofectamine 2000, 24 hours before the cells were transfected with plasmids encoding receptor and biosensors. Cells were pretreated with inhibitors for 30 min at 37°C, except for filipin III, MβCD, or MβCD-cholesterol complexes (45-min pretreatment) or PTx (16-hour pretreatment). MβCD-cholesterol complexes were formed as described previously (64). Inhibitors were used at the following concentrations: 30 μM PitStop2 or inactive PitStop2, 10 μM NF023, PTx at 100 ng/ml, 5 μM mSIRK or mSIRK L9A, 1 μM GF109203X or Gö6983, 10 nM Gö6976, 10 μM Myr-EAVSLKPT-OH (a PKCε translocation inhibitor peptide), filipin III at 1 μg/ml, 10 mM MβCD, and 2 mM MβCD with 0.2 mM cholesterol (MβCD-cholesterol complexes). All experiments were performed in live cells at 37°C. For all regulation and trafficking experiments, cells were stimulated with an EC50 concentration of DAMGO or morphine (both 1 μM) as defined by β-arrestin-2 concentration-response curves (fig. S2A). For all signaling experiments, cells were stimulated with an EC50 concentration of DAMGO (10 nM) or morphine (100 nM) as defined by AlphaScreen pERK assays (fig. S1A).

RNA sequencing

RNA was extracted from two passages of HEK 293 cells (P0 and P37) with the RNeasy Mini Kit (Qiagen). Transcriptome sequencing was performed by the Beijing Genomics Institute.

DRG isolation and culture

All procedures involving mice were approved by the Monash Institute of Pharmaceutical Sciences animal ethics committee. DRG neurons were isolated and transfected with 600 ng of cytoEKAR Cerulean/Venus, nucEKAR Cerulean/Venus, or pmCKAR with the Nucleofector system (Lonza). Detailed protocols of DRG isolation and nucleofection were described previously (63).

BRET assays

HEK 293 cells were transfected with 1 μg of MOR-RLuc and 4 μg of KRas-Venus, Rab5a-Venus, or β-arrestin-2-YFP. For coexpression, cells were transfected with an additional 2 μg of βARKct, GFP-dynamin, or GFP-dynamin K44E. After 24 hours, cells were plated in poly-d-lysine–coated 96-well plates (CulturPlate, PerkinElmer) and allowed to adhere. Forty-eight hours after transfection, the cells were equilibrated in Hanks’ balanced salt solution (HBSS) and then were incubated with vehicle [0.1% dimethyl sulfoxide (DMSO)], DAMGO, or morphine for 30 min. Coelenterazine h (Promega) was added at a final concentration of 5 μM, and the cells were incubated for a further 10 min. BRET measurements were obtained with the PHERAstar Omega microplate reader (BMG Labtech), which enabled sequential integration of the signals detected at 475 ± 30 nm and 535 ± 30 nm with filters with the appropriate band pass. Data were presented as a BRET ratio (calculated as the ratio of the YFP signal to the Renilla luciferase signal) corrected for vehicle.

High-content image analysis

HEK 293 cells were plated in poly-d-lysine–coated, black, optically clear 96-well plates (ViewPlate, PerkinElmer) and transfected with plasmid encoding MOR-GFP (20 ng per well). Forty-eight hours after transfection, the cells were incubated with inhibitors and treated with vehicle (0.1% DMSO), DAMGO (1 μM), or morphine (1 μM) for 30 min. Cells were fixed with 4% paraformaldehyde (PFA) and washed three times with phosphate-buffered saline (PBS). Nuclei were stained with Hoechst (1 μg/ml). Images of four fields of view were collected with a GE Healthcare INCell 2000 Analyzer with a Nikon Plan Fluor ELWD 40× [numerical aperture (NA), 0.6] objective. Analysis was performed with the granularity application module in MetaMorph imaging software (v7.8.6, Molecular Devices). Granule detection was set at 4 to 8 μm, nuclei detection was set at 35 to 60 μm, and the total number of cytosolic granules per cell was calculated. The effect of vehicle was subtracted, and the data were expressed relative to the DAMGO-stimulated response (in the absence of inhibitors).

Förster resonance energy transfer

HEK 293 cells were transfected with plasmid encoding MOR (55 ng per well) and with plasmid encoding cytoEKAR GFP/RFP, nucEKAR GFP/RFP, cytoCKAR, or pmCKAR (40 ng per well). For coexpression, cells were transfected with plasmid encoding βARKct, GFP-dynamin, or GFP-dynamin K44E (50 ng per well). Experiments in which GFP-dynamin or GFP-dynamin K44E were coexpressed used the Cerulean-Venus FRET sensors. FRET was measured with a high-content GE Healthcare INCell 2000 Analyzer as described previously (63). Briefly, fluorescence imaging was performed with a Nikon Plan Fluor ELWD 40× (NA, 0.6) objective and FRET module. For GFP-RFP emission ratio analysis, cells were sequentially excited with a fluorescein isothiocyanate (FITC) filter (490/20) with emission measured with dsRed (RFP from Discosoma sp.) (605/52) and FITC (525/36) filters and a polychroic mirror, optimized for the FITC-dsRed filter pair (Quad4). For cyan fluorescent protein (CFP)–YFP or Cerulean-Venus emission ratio analysis, cells were sequentially excited with a CFP filter (430/24) with emission measured with YFP (535/30) and CFP (470/24) filters and a polychroic mirror, optimized for the CFP-YFP filter pair (Quad3). HEK 293 cells were imaged every 1 min, which enabled image capture at 14 wells/min; DRG neurons were imaged every 1 min with four fields of view per well, which enabled the capture of 3 wells/min. At the end of every experiment, the same cells were stimulated for 10 min with the positive control [200 nM phorbol 12,13-dibutyrate for ERK or 200 nM phorbol 12,13-dibutyrate with phosphatase inhibitor cocktail 2 (Sigma Aldrich) for PKC] to generate a maximal FRET change, and positive emission ratio images were captured for 4 min. Data were analyzed with the FIJI distribution of ImageJ (65). The three emission ratio image stacks (baseline, stimulated, and positive) were collated and aligned with the StackCreator script (63). Cells were selected, and fluorescence intensity was measured over the combined stack. Background intensity was subtracted, and then the FRET data were plotted as the change in FRET emission ratio relative to the maximal response for each cell [FRET ratio/maximum FRET ratio (F/Fmax)]. For HEK 293 cells, only cells that showed more than a 10% change relative to baseline after stimulation with the positive control were considered for analysis. For DRG neurons, all cells that showed more than a 3% change relative to baseline after stimulation with the positive control were considered for analysis. Ratiometric pseudocolor images were generated as previously described (66). The Green Fire Blue LUT was applied, and the brightness and contrast range was set to the minimum and maximum FRET ratios within the image stack (0.13 to 0.23).

GSD-TIRF microscopy

HEK 293 cells and DRG neurons were treated with vehicle (0.1% DMSO), DAMGO, or morphine as indicated in the figure legends; fixed in 4% PFA for 20 min at 4°C; washed for 15 min with PBS; blocked in PBS with 1% normal goat serum and 0.1% saponin for 1 hour at room temperature; and incubated overnight at 4°C with mouse anti-FLAG (at a 1:1000 dilution) for HEK 293 cells or with rabbit anti-MOR (UMB-3, 1:250) and anti–tubulin βIII (1:1000) for DRG neurons. Cells were washed and incubated with Alexa Fluor 568– or Alexa Fluor 647–conjugated goat anti-mouse or anti-rabbit secondary antibodies (1:400; for 2 hours at room temperature). Coverslips were mounted on a concave slide containing 100 mM cysteamine (MEA) and sealed. Cells were observed with a Leica GSD microscope with HCX PL APO 160× (NA, 1.43) objective, SuMo stage, Andor iXon Ultra 897 camera, and LAS AF software. Pumping occurred at 100% laser power until the frame correlation dropped to 0.25. Data were acquired at 50% laser power, and up to 30,000 frames were captured. TIRF penetration was at 110 nm. Only neurons with positive staining for β-tubulin were analyzed. Images were analyzed with FIJI software (65). Individual particles were selected with Find Maxima (noise tolerance 5) to generate a binary output of the single points. The average distance between events was calculated by generating a centroid list with the Analyze Particles command and was processed by the Nearest Neighbor Distance (NND) macro (Yuxiong Mao). EDMs were generated from the single-point binary image with the Euclidean distance option.

Whole-cell radioligand binding assays

Cells were plated in a 96-well isoplate (20,000 cells per well; PerkinElmer) and allowed to adhere for 24 hours. The cells were then washed three times with assay buffer [146 mM NaCl, 10 mM d-glucose, 5 mM KCl, 1 mM MgSO4, 2 mM CaCl2, 1.5 mM NaHCO3, 10 mM Hepes (pH 7.45)] and incubated for 3 hours at 4°C with increasing concentrations of [3H]diprenorphine (specific activity, 36.1 Ci/mmol). Nonspecific binding was determined by the coaddition of 1 μM naloxone. After being washed in cold saline, the cells were solubilized in Optiphase scintillant, and radioactivity was measured in a MicroBeta counter (PerkinElmer).

AlphaScreen pERK signaling assays

ERK1/2 phosphorylation was detected with the AlphaScreen ERK1/2 SureFire protocol (TGR Biosciences). Briefly, cells were seeded in clear 96-well plates (40,000 cells per well) and allowed to adhere for 8 hours. Cells were washed twice with PBS and incubated in serum-free DMEM overnight at 37°C and 5% CO2. Cells were stimulated for 5 min, and then the medium was replaced with lysis buffer in which the cells were incubated for 5 min at room temperature with agitation. Lysates were transferred to a 384-well white ProxiPlate (PerkinElmer), and then a mixture of activation buffer, reaction buffer, and AlphaScreen beads (100:600:3) was added to generate a final lysate/mixture ratio of 5:8. Plates were incubated for 1.5 hours at 37°C and then were read on a Fusion-α plate reader (PerkinElmer). Data were expressed relative to the responses of cells to vehicle (0%) and 10% (v/v) FBS (100%).

Confirmation of β-arrestin knockdown

Cells were transfected with combined β-arrestin-1– and β-arrestin-2–specific siRNAs or scrambled siRNA (25 nM) in a six-well plate. After 72 hours, cells were washed with ice-cold PBS, harvested by scraping in 1 ml of PBS, and then centrifuged at 500g for 5 min at 4°C. Cells were then lysed in 100 μl of lysis buffer [50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.1 mM EDTA, 0.6% Tween 20, protease inhibitor cocktail] for 30 min on ice by passing through a 26-gauge needle 10 times. Lysates were then centrifuged at 700g for 5 min at 4°C; the supernatants were incubated with 50 U of deoxyribonuclease I for 1 hour at 37°C, and then Laemmli sample buffer was added.

Basic membrane fractionation

Cells transfected with 5 μg of plasmid encoding FLAG-MOR (in confluent 10-cm dishes) were treated with 0.1% DMSO (vehicle), 10 μM DAMGO, or 10 μM morphine for 10 min at 37°C; washed with ice-cold MES buffer [25 mM MES (pH 6.5), 150 mM NaCl]; and then incubated in 1 ml of ice-cold MBS buffer (MES containing 1 mM EDTA and protease inhibitor cocktail) for 10 min on ice. The cells were scraped, centrifuged at 500g for 5 min at 4°C, and then resuspended in 0.5 ml of MBS buffer containing 1% Triton X-100. After a 20-min incubation on ice, samples were homogenized with a 2-ml Dounce homogenizer and centrifuged at 16,000g for 5 min at 4°C. The supernatants were collected as the Triton X-100–soluble fractions. The pellets (the Triton X-100–insoluble fractions) were resuspended in 225 μl of MBS containing 1% Triton X-100 by sonication three times (30 s, 20% amplitude, 2-mm microprobe) with a Q125 sonicator (Qsonica). Laemmli sample buffer was added, and the samples were incubated at 37°C for 10 min.

Western blotting analysis

Samples were resolved by SDS-PAGE with 10% tris-glycine gels and then were transferred onto 0.45-μm Bio-Rad low-fluorescence polyvinylidene difluoride membranes with a Trans-Blot SD Semi-Dry Transfer Cell (Bio-Rad). Membranes were blocked for 1 hour at room temperature with Odyssey Blocking Buffer (LI-COR Biotechnology) and then were incubated with primary antibody overnight at 4°C. The membranes were washed three times in PBS containing 0.1% Tween 20, incubated with secondary antibody for 1 hour at room temperature, and then were washed three times in PBS containing 0.1% Tween 20. Secondary antibody fluorescence was detected with an Odyssey Classic Infrared Imager (LI-COR Biotechnology) with the intensity setting adjusted to be in the linear range for infrared fluorescence detection. Antibodies were used at the following dilutions: mouse anti–FLAG M2, 1:1000; rabbit anti–caveolin 1, 1:1000; rabbit anti–β-tubulin, 1:5000; rabbit anti–β-arrestin-1/2, 1:2000; mouse anti–β-actin, 1:2000; goat anti-rabbit IRDye800, 1:5000; and goat anti-mouse IRDye680, 1:10,000.

Localization of MOR or β2AR by confocal microscopy

Cells expressing MOR-GFP were treated with vehicle (0.1% DMSO), DAMGO (1 μM), or morphine (1 μM), whereas cells expressing β2AR-GFP were treated with vehicle (0.1% ascorbic acid) or 1 μM isoprenaline for the times indicated in the figure legends. After stimulation, the cells were fixed with 4% PFA in PBS for 20 min at 4°C and then were washed for 15 min with PBS. Cells were blocked in PBS containing 1% normal goat serum and 0.1% saponin for 1 hour at room temperature and then were incubated with rabbit anti–clathrin heavy chain (1:1000) overnight at 4°C. Cells were washed and incubated with Alexa Fluor 568–conjugated goat anti-rabbit secondary antibody (1:400) for 2 hours at room temperature. To visualize the colocalization of MOR-GFP or β2AR-GFP with clathrin in HEK 293 cells or of MOR with β-tubulin in DRG neurons, cells were observed with a Leica SP8 confocal microscope and HCX PL APO 63× CS2 (NA, 1.40) oil objective. Images were collected at a zoom of 1 to 3, and three to five optical sections were taken at intervals of 0.39 μm. Colocalization between receptor and clathrin was quantified with the colocalization threshold command in FIJI (65) after background intensity was subtracted from the corresponding images.

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/9/414/ra16/DC1

Fig. S1. Ligand-dependent spatiotemporal signaling of MOR.

Fig. S2. MOR regulation and its effect on cytosolic ERK activity.

Fig. S3. Roles of PKC and receptor phosphorylation on the MOR-dependent spatiotemporal profiles of ERK signaling.

Fig. S4. MOR localization within the plasma membrane is ligand-dependent.

Fig. S5. Effect of membrane disruption on MOR trafficking and spatiotemporal signaling.

Table S1. Expression of wild-type MOR and the MOR S375A mutant.

REFERENCES AND NOTES

Acknowledgments: We thank D. D. Jensen for technical assistance; M. J. Christie for critical discussion; R. J. Summers, J. R. Lane, and A. M. Ellisdon for careful review of this manuscript; and L. Bohn, S. Schulz, and M. Caron for providing DNAs. Funding: This work was supported by a Monash Fellowship to M.C.; National Health and Medical Research Council (NHMRC) RD Wright Fellowship to M.L.H. (1061687); NHMRC Australia Fellowship to N.W.B. (63303); NHMRC Project Grants (1011796, 1047633, 1049682, and 1031886) to M.C., M.L.H., and N.W.B.; ARC Centre of Excellence in Convergent Bio-Nano Science and Technology, Monash Institute of Pharmaceutical Sciences Large Grant Support Scheme grants to M.C. and M.L.H.; and Monash University support to N.W.B. G.L.T. is funded by DSTO (Defence Science and Technology Organisation) Australia. Author contributions: M.L.H. and M.C. designed the study; M.L.H., H.R.Y., C.J.N., G.L.T., A.B.G., S.C., D.P.P., and M.C. performed the experiments; M.L.H. and M.C. analyzed and interpreted the data; M.L.H. and M.C. drafted and wrote the manuscript; all authors reviewed the manuscript; and N.W.B. and N.A.L. provided reagents. Competing interests: The authors declare that they have no competing interests.
View Abstract

Stay Connected to Science Signaling

Navigate This Article