Receptor heteromerization expands the repertoire of cannabinoid signaling in rodent neurons

Abstract

A fundamental question in G protein coupled receptor biology is how a single ligand acting at a specific receptor is able to induce a range of signaling that results in a variety of physiological responses. We focused on Type 1 cannabinoid receptor (CB₁R) as a model GPCR involved in a variety of processes spanning from analgesia and euphoria to neuronal development, survival and differentiation. We examined receptor dimerization as a possible mechanism underlying expanded signaling responses by a single ligand and focused on interactions between CB₁R and delta opioid receptor (DOR). Using co-immunoprecipitation assays as well as analysis of changes in receptor subcellular localization upon co-expression, we show that CB₁R and DOR form receptor heteromers. We find that heteromerization affects receptor signaling since the potency of the CB₁R ligand to stimulate G-protein activity is increased in the absence of DOR, suggesting that the decrease in CB₁R activity in the presence of DOR could, at least in part, be due to heteromerization. We also find that the decrease in activity is associated with enhanced PLC-dependent recruitment of arrestin3 to the CB₁R-DOR complex, suggesting that interaction with DOR enhances arrestin-mediated CB₁R desensitization. Additionally, presence of DOR facilitates signaling via a new CB₁R-mediated anti-apoptotic pathway leading to enhanced neuronal survival. Taken together, these results support a role for CB₁R-DOR heteromerization in diversification of endocannabinoid signaling and highlight the importance of heteromer-directed signal trafficking in enhancing the repertoire of GPCR signaling.

Figure 1. Increased CB₁R activity in the absence of DOR.

A, Basal [³⁵S]GTPγS binding was measured in cortical membranes from wild-type and DOR −/− mice. Membranes from cortices were prepared as described –, treated with vehicle or 1 µM SR141716 (SR) for 1 hour and subjected to [³⁵S]GTPγS binding as described in Methods. Basal [³⁵S]GTPγS binding/10 µg protein in vehicle treated membranes is taken as 100%. Data represent Mean ± SEM (n = 3 individual animals in triplicate). Statistically significant differences between vehicle and SR141716 treatment are indicated *p<0.05, (t test). B, [³⁵S]GTPγS binding assay in cortical membranes from wild-type and DOR −/− mice. Membranes were treated with increasing concentrations of the CB₁R agonist Hu210 (0–1 µM) and [³⁵S]GTPγS binding was measured as described in Methods. EC₅₀ and E_max values were calculated using GraphPad Prism software. Data represent Mean ± SEM (n = 3 individual animals in triplicate). *p<0.05; **p<0.01 for DOR−/− vs wild-type (t test). C, Localization of endogenous CB₁R and DOR in mouse primary cortical cells, 14DIV. Cells fixed with 4%PFA in PBS and permeablized with 0.1% Triton, were immunostained with the goat polyclonal anti-CB₁R(N-terminal) antibody (1∶500; green) and rat polyclonal anti-DOR antibody (1∶500; red) and visualized using Alexa 488-conjugated anti-goat (1∶1000) and Alexa 594-conjugated anti-rat (1∶1000) secondary antibodies using confocal microscopy as described in Methods. Representative figure from 3 independent experiments shown.

Figure 2. Association between CB₁R and DOR alters CB₁R localization.

A, Lysates (100–200 µg) from N2A^CB1R and N2A^CB1RDOR cells were subjected to immunoprecipitation with 1 µg of anti-CB₁R (C-terminal) antibody, the immunoprecipitates were resolved on 10% SDS-PAGE and probed for the presence of myc-DOR using mouse monoclonal anti-myc antibody (1∶1000) and for CB₁R using rabbit polyclonal anti-CB₁R (C-terminal) antibody (1∶500) as described in Methods. IRDye 680 anti-rabbit and IRDye 800 anti-mouse were used as secondary antibodies (1∶10,000). Representative of 3 independent experiments shown. B, CB₁R-DOR complexes exhibit greater interaction with AP-2 than AP-3. Lysates (100–200 µg) from N2A^CB1R and N2A^CB1RDOR cells were subjected to immunoprecipitation using 1 µg of an anti-CB₁R (C-terminal) antibody as described in Methods. The immunoprecipitates were resolved on 10% SDS-PAGE and probed for the presence of AP-3 (1∶1000), AP-2 (1∶1000) and CB₁R (C-term) (1∶500) using specific antibodies as described in Methods. IRDye 680 anti-rabbit and IRDye 800 anti-mouse were used as secondary antibodies (1∶10,000). Representative of 3 independent experiments shown. C, Localization of endogenous CB₁R in N2A^CB1R and of CB₁R and DOR in N2A^CB1RDOR cells. Cells fixed with 4%PFA in PBS and permeablized with 0.1% Triton, were stained with the rabbit polyclonal anti-CB₁R (C-terminal) antibody (1∶500; green) and the mouse monoclonal anti-myc antibody (1∶1000; red) and visualized using Alexa 488-coupled anti-rabbit or Alexa 594-coupled anti-mouse secondary antibodies (1∶1000) using confocal microscopy as described in Methods. Representative of 3 independent experiments shown. D, Cell surface staining of endogenous CB₁R and stably expressed DOR in N2A^CB1RDOR cells. N2A^CB1R and N2A^CB1RDOR cells were stained with a goat polyclonal anti-CB1R (N-terminal) antibody (1∶500) and mouse monoclonal anti-myc antibodies (1∶1000) prior to fixation of the cells to label cell surface receptors, as described . After fixation, cells were visualized with Alexa 594-coupled anti-goat and Alexa 488-coupled anti-mouse secondary antibodies (1∶1,000) using confocal microscopy as described in Methods. Representative of 3 independent experiments shown.

Figure 3. PLC-dependent arrestin3 association with CB1R-DOR complex.

A, [³⁵S]GTPγS binding assay in membranes from N2A^CB1R and N2A^CB1RDOR cells. Membranes (10 µg) were treated with indicated concentrations of the CB₁R agonist Hu210. [³⁵S]GTPγS binding was measured as described in Methods. EC₅₀ and E_max values were calculated using GraphPad Prism software. Data represent Mean ± SEM (n = 3 independent experiments in triplicate). B, Dose-response of Hu210-mediated ERK phosphorylation in N2A^CB1R and N2A^CB1RDOR cells. Starved N2A^CB1R and N2A^CB1RDOR cells seeded in 24 well-plates were treated with indicated concentrations of Hu210 for 5 minutes. Cell lysates (30 µg protein) were analyzed by Western blotting and probed for the levels of pERK (1∶1000) and ERK (1∶1000) as described in Methods. IRDye 680 anti-rabbit and IRDye 800 anti-mouse were used as secondary antibodies (1∶10,000). EC₅₀ and E_max values were calculated using GraphPad Prism software. Data represent Mean ± SEM (n = 3 independent experiments). *p<0.05 for N2A^CB1RDOR vs N2A^CB1R (t test). C, Effect of DOR down-regulation on ERK phosphorylation. F11 cells transduced with the DOR shRNA expressing lentivirus were starved for 4–6 h and treated with Hu210 (100 nM) for 5 min. Cell lysates (30 µg protein) were analyzed by Western blotting and probed for the levels of pERK (1∶1000) and ERK (1∶1000) as described in Methods. IRDye 680 anti-rabbit and IRDye 800 anti-mouse were used as secondary antibodies (1∶10,000). Data from 3 independent experiments is shown. *p<0.05 (t test). D, Examination of arrestin3 interaction with CB₁R after Hu210 treatment. N2A^CB1R and N2A^CB1RDOR, starved for 4 hours were stimulated with 100 nM Hu210 for 5 minutes and cell lysates prepared as described in Methods. Lysates (30 µg protein) were subjected to either Western blotting using rabbit anti-CB1R (C-terminal 1∶500) and mouse anti-arrestin 3 antibodies (1∶500) or to immunoprecipitation using 1 µg of agarose-coupled anti-CB₁R (C-terminal) antibody. Immunoprecipitates were probed for arrestin3 levels by Western blot using the mouse anti-arrestin 3 antibody. IRDye 680 anti-rabbit and IRDye 800 anti-mouse were used as secondary antibodies (1∶10,000). Representative of 3 independent experiments shown. E, Effect of Hu210 on arrestin recruitment. U2OS cells co-expressing ProLink/Enzyme Donor (PK)-tagged DOR and the Enzyme Activator (EA)-tagged arrestin3 fusion protein without or with CB₁R were treated with indicated concentrations of Hu-210. Arrestin3 recruitment was determined using the PathHunter Detection Kit as described in Methods. Data represent Mean ± SEM (n = 4). F, Effect of PLC inhibitor (U73122) on DOR phosphorylation at serine 363 after Hu210 treatment. N2A^CB1RDOR cells were starved for 4–6 hours, and incubated with vehicle (DMSO) or U73122 (1 µM) for 30 minutes, then stimulated with 100 nM Hu210 for 5 minutes. Cell lysates (30 µg protein) were subjected to Western blotting using rabbit polyclonal phosphoDOR Ser 363 (1∶1000), mouse monoclonal anti-myc (1∶1000) antibodies and IR Dye 680 anti-rabbit and IR Dye 800 anti-mouse secondary antibodies (1∶10,000) as described in Methods. Data represent Mean ± SEM (n = 3).

Figure 4. Engagement of arrestin3-dependent signaling in N2A^CB1RDOR.

Time course of Hu210-mediated ERK phosphorylation in A, N2A^CB1R; B, N2A^CB1RDOR; and C, N2A^CB1RDORΔ15 cells transfected with control or arrestin3-targeting siRNA. N2A^CB1R alone or stably expressing either DOR or DORΔ15, transfected with a control or arrestin3-targeting siRNA, were starved for 4 hours, then stimulated with 100 nM Hu210 for the indicated times. Cell lysates (30 µg protein) were subjected to Western blotting for the levels of ERK (1∶1000), pERK (1∶1000), and arrestin3 (1∶500) as described in Methods. IRDye 680 anti-rabbit and IRDye 800 anti-mouse were used as secondary antibodies (1∶10,000). Data represent Mean ± SEM (n = 3); *p<0.05; **p<0.01; ***p<0.001, for control vs Arr3 siRNA (t test).

Figure 5. Role of arrestin3 and ERK substrates in cannabinoid signaling by the CB₁R-DOR heteromer.

A–B, Effect of CB1R-DOR heteromerization on subcellular localization of pERK. A, N2A^CB1R and N2A^CB1RDOR cells were treated with Hu210 (100 nM; 0, 5 or 10 min), cytoplasmic and nuclear extracts were prepared as described in Methods and analyzed (30 µg protein) by Western blotting with pERK (1∶1000), ERK (1∶1000), lamin A/C (1∶2000), and GAPDH (1∶2000) antibodies. IRDye 680 anti-rabbit and IRDye 800 anti-mouse were used as secondary antibodies (1∶10,000). B, N2A^CB1R and N2A^CB1RDOR cells treated with Hu210 (100 nM; 5 min) were immunostained with pERK (1∶1000, red) and pericentrin (1∶1000, green) antibodies and visualized using Alexa 488-conjugated anti-rabbit (1∶1000) and Alexa 594-conjugated anti-mouse (1∶1000) secondary antibodies using confocal microscopy as described in Methods. Representative of 3 independent experiments shown. C, Time course of phosphorylation of pERK substrates. Lysates (30 µg protein) from N2A^CB1R and N2A^CB1RDOR cells treated with Hu210 (100 nM; 0, 5, 10 or 30 min) were analyzed by Western blotting with STAT3 (1∶1000), phosphoSTAT3 (1∶1000), phospho-p70S6K (1∶1000), phospho-p90rsk (1∶1000) and phosphoBAD (1∶1000) antibodies. IRDye 680 anti-rabbit and IRDye 800 anti-mouse were used as secondary antibodies (1∶10,000). ERK (1∶1000) is used as a loading control. Representative of 3 independent experiments shown. D, Involvement of PLC and MEK in the phosphorylation of STAT3 and BAD. Lysates (30 µg protein) from N2A^CB1R and N2A^CB1RDOR cells treated with 100 nM Hu210 for 5 min in the absence or presence of U73122 (1 µM) or PD98059 (PD, 10 µM) were analyzed by Western blotting with STAT3 (1∶1000), phosphoSTAT3 (1∶1000), and phosphoBAD (1∶1000) antibodies. IRDye 680 anti-rabbit and IRDye 800 anti-mouse were used as secondary antibodies (1∶10,000). ERK (1∶1000) is used as a loading control. Representative of 3 independent experiments shown. E, Involvement of arrestin3 in BAD phosphorylation. Arrestin3 was down-regulated in N2A^CB1RDOR cells by transfection with a siRNA. These cells were stimulated with Hu210 (100 nM) for 5 min, in the absence or presence of PD (10 µM). Lysates (30 µg protein) were analyzed by Western blotting with phospho-p90rsk (1∶1000) and phosphoBAD (1∶1000) antibodies. IRDye 680 anti-rabbit and IRDye 800 anti-mouse were used as secondary antibodies (1∶10,000). ERK (1∶1000) is used as a loading control. Representative of 3 independent experiments shown.

Figure 6. CB₁R-DOR heteromerization promotes cell survival.

A, Hu210-treated N2A^CB1RDOR cells exhibit increased survival compared to N2A^CB1R cells. N2A^CB1R or N2A^CB1RDOR cells were treated with 1 µM Hu210 for the indicated days and survival measured by trypan blue exclusion as described in Methods. Data represent Mean ± SEM (n = 4 in triplicate). B, Hu210-treated N2A^CB1RDOR cells exhibit lower apoptosis as compared to N2A^CB1R cells. Apoptosis of N2A^CB1R or N2A^CB1RDOR treated for 3 or 8 days with 1 µM Hu210 was measured using caspase-3 activity assay as described in Methods. Data represent Mean ± SEM (n = 4 in triplicate); ***p<0.001 N2A^CB1RDOR vs N2A^CB1R (t test). C, CB₁R antagonist treatment decreases neuronal survival of striatal neurons from wild-type but not DOR−/− mice. Primary striatal neurons from wild-type or from DOR−/− mice were prepared as described in Methods. The CB₁R antagonist AM251 (10 µM) was added to the growth media at DIV7 and cellular viability assessed at DIV10 as described in Methods. Data represent Mean ± SEM (n = 2–4) D–E, A schematic of the signaling pathways emanating from CB₁R in N2A^CB1R (D) and N2A^CB1RDOR (E) cells. Activation of CB₁R in N2A^CB1RDOR cells leads to differential activation of signaling molecules and phosphorylation of ERK substrates.