The orphan receptor GPR55 is a novel cannabinoid receptor

Cloning of hGPR55

hGPR55 (EMBL accession no. BC032694) was amplified from human genomic DNA by polymerase chain reaction (PCR) and sub-cloned into mammalian expression plasmids pIRESneo2 and pcDNA3 using standard techniques.

Expression profiling

GPR55 mRNA levels in human and mouse tissues were analysed by quantitative real-time PCR analysis using ABI PRISM 7900 HT sequence detection system (Applied Biosystems, Hercules, CA, USA). Primer/probe sets for hGPR55 were: 5′-TCTACATGATCAACCTGGCAGTCT-3′, 5′-CTGGGACAGGACCATCTTGAA-3′ and 5′-FAM-TGACCTGCTGCTGGTGCTCTCCC-TAMRA-3′, and for mGPR55 were: 5′-CTATCTACATGATCAACTTGGCTGTTT-3′, 5′-TGTGGCAGGACCATCTTGAA-3′ and 5′-FAM-CGATTTACTGCTGGTGCTCTCCCTCCC-TAMRA-3′. To determine relative mRNA levels of GPR55, results were normalized to its content of the mRNA encoding the ribosomal protein 36B4 (used as an internal standard).

Cell transfection and membrane preparation

Human embryonic kidney—HEK293s cells (5 × 10⁶) were seeded in T75 flasks and after 24 h, cells were transiently transfected with 10 μg of relevant plasmid using Lipofectamine Plus (Invitrogen, Carlsbad, CA, USA). Membranes were prepared after 48 h using standard methods and stored at −80 °C. Protein concentration was measured according to the method of Bradford (Bio-rad Laboratories, Foster City, CA, USA) (Bradford, 1976). CB₁ and CB₂ membranes were commercially available (PerkinElmer).

Radioligand binding assays

Radioligand binding was initiated by the addition of 5 μg of membrane protein to each well of a 96-well plate containing 50 nM [³H]-(−)-cis-3-[2-hydroxy-4-(1,1–dimethylheptyl)phenyl]-trans-4-(3-hydroxypropyl)cyclohexanol (CP55940) (Tocris, Ellisville, Missouri, USA), [³H]-SR141716 (Amersham, Piscataway, NJ, USA) or [³H]-R(+)-[2,3-di-hydro-5-methyl-3-[(morpholinyl)methyl] pyrrolo[1,2,3-de] -1,4-benz-oxazinyl]-(1-naphthalenyl)-methanone-mesylate (WIN55,212-2) (Amersham), sufficient volume of buffer (50 mM Tris–HCl, 5 mM MgCl₂, 50 mM NaCl, pH 7.4, 0.1% bovine serum albumin (BSA)) to bring the total volume of each well to 200 μl. Non-specific binding was determined in the presence of 10 μM CP55940 (Tocris), SR141716 and WIN55,212-2 (Tocris). The membranes were incubated at 30 °C for 90 min and the reaction was then terminated by the addition of ice-cold wash buffer (50 mM Tris–HCl, 5 mM MgCl₂, 50 mM NaCl, pH 7.4) followed by rapid filtration under vacuum through Printed Filtermat B glass fibre filters (Wallac, Turku, Finland) (0.05% polyethylenimine (PEI)-treated) using a Micro 96 Harvester (Skatron Instruments, Lier, Norway). The filters were dried for 30 min at 50 °C, then a paraffin scintillant pad was melted onto the filters and the bound radioactivity was determined using a 1450 Microbeta Trilux (Wallac) scintillation counter.

[³⁵S]-GTPγS binding assay

[³⁵S]-Guanosine 5′-[γ-³⁵S]-triphosphate (GTPγS) binding assays were conducted at 30 °C for 45 min in membrane buffer (100 mM NaCl, 5 mM, 1 mM EDTA, 50 mM HEPES, pH 7.4) containing 0.025 μg μl⁻¹ of membrane protein with 0.01% BSA (fatty-acid free) (Sigma, St Louis, MO, USA), 10 μM guanosine 5′-diphosphate (GDP) (Sigma), 100 μM dithiothreitol (DTT) (Sigma) and 0.53 nM [³⁵S]-GTPγS (Amersham) in a final volume of 200 μl. Non-specific binding was determined in the presence of 20 μM unlabelled GTPγS (Sigma). The reaction was terminated by addition of ice-cold wash buffer (50 mM Tris–HCl, 5 mM MgCl₂, 50 mM NaCl, pH 7.4) followed by rapid filtration under vacuum through Wallac GF/B glass-fibre filters using a cell harvester (Skatron). The filters were left to dry for 30 min at 50 °C, then a paraffin scintillant pad was melted onto the filters and the bound radioactivity was determined using a microbeta scintillation counter (Wallac). Antagonist potency was determined versus an EC₈₀ concentration of CP55940 that was determined empirically on the day of the experiment. Data were fitted using the equation y=A+((B−A)/1+((C/x)∧D))) and the EC₅₀ estimated where A is the non-specific binding, B is the total binding, C is the IC₅₀ and D is the slope.

Peptide and antibody blocking of [³⁵S]-GTPγS binding assays

[³⁵S]-GTPγS binding assays were performed as above with additional pre-incubation of membranes with and without peptides or antibodies for the G-protein subunits Gα₁₃, Gα_i and Gα_s for 15 min at 30 °C (Santa Cruz Biotechnology, Santa Cruz, CA, USA). Data were analysed using paired t-test (**P<0.05; ***P<0.01).

Pertussis toxin treatment

Cell transfections were conducted as described above with the exception that the cells prior to harvesting were pre-incubated with Pertussis toxin (Sigma) overnight (0.1 μg ml⁻¹ final concentration). The cells were then harvested and membranes were prepared as described above.

Plate-based FLIPR Ca²⁺ assays

In brief, 1 day before the assay was performed, HEK293 cells expressing GRP55 were plated in 96-well, black-walled, assay plates, at a density of 25 000 cells per well. These plates were then returned to the cell-culture incubator until 1.5 h before the assay when they were removed and the cells were loaded with the Ca²⁺ reporter dye Fluo4 (Invitrogen) for 1 h in a cell-culture incubator. After this, the plates were placed into a fluorescent imaging plate reader (FLIPR) to monitor fluorescence (λ_ex=488 and λ_EM=540 nm) before and after the addition of ligands of interest.

Determination of rhoA, rac1 and cdc42 activity

RhoA, rac1 and cdc42 activity was measured according to the manufacturer's instructions (Upstate Biotechnology Inc., Charlottesville, Virginia, USA). HEK293s GPR55-transfected cells were seeded on six-well plates, grown to 80% confluence, and serum-starved for 24 h. Following treatment with selected compounds at 37 °C for 15 min, the cells were washed with phosphate-buffered saline and harvested with 500 μl of lysis buffer provided by the manufacturer, with the addition of a mixture of protease inhibitors (Roche Molecular Biochemicals, Basel, Switzerland). The cell lysates were clarified by centrifugation at 15 000 g for 1 min. For a negative control, cell lysate was incubated with 1 mM GDP for 15 min at 30 °C. The cell lysates were then incubated with 10 μg of GST-RBD-agarose (Rho-binding domain of rhotekin) or GST-PBD-agarose (p21-binding domain of human PAK-1) to precipitate GTP-bound rhoA and GTP-bound rac1 and cdc42, respectively. The beads were then washed three times with lysis buffer and samples were prepared for electrophoresis by adding 1 × sodium dodecyl sulphate loading dye. Samples were boiled for 5 min and resolved by 12% sodium dodecyl sulphate-polyacrylamide gel electrophoresis. Bound rhoA, rac1 and cdc42 were detected by western blot using the appropriate polyclonal antibodies specific for rac1, cdc42 (1:1000; Upstate Biotechnology) and rhoA (1:200; Santa Cruz Biotechnology).

Cloning and sequence determination of GPR55

Nucleotide primers designed against the 5′- and 3′-ends of human, mouse and rat gpr55 were used to isolate open reading frames for GPR55 from the three species using genomic DNA. The sequence of the human gene was similar but not identical to that already described (Sawzdargo et al., 1999), however it was consistent with the human genome sequence. The sequence of all clones isolated differed in that there was a nucleotide insertion and deletion at positions 393 and 427 respectively, resulting in a frame shift of the translated sequence, consequently changing 11 amino acids at the predicted junction of intracellular loop 2 and transmembrane helix 4 (Figure 1). Since we could find no evidence for the existence of the previously published sequence, we concluded that the difference originated from a sequencing error by the authors (Sawzdargo et al., 1999).

The rat and mouse genes were cloned using a similar approach and their sequences were found to be identical to those found in GenBank (AC119315 (position 129078–130085) AC107707 (position 31198–32181)) demonstrating 75 and 78% identity to the human sequence respectively (Figure 1). Both the rat and mouse sequences are consistent with the human genome sequence in the region of the intracellular loop 2—transmembrane helix 4 region rather than the published sequence (Sawzdargo et al., 1999) containing the insertion and deletion (Figure 1). Despite the low level of identity between the human and rodent forms of GPR55, the genomic linkage confirms that the rodent genes are orthologues of the human gpr55. Phylogenetically, the GPR55 sequence belongs to a cluster of receptors that are either orphans (GPR35, GPR92, P2Y5) or have been recently deorphanized (P2Y9 (Noguchi et al., 2003), GPR40 (Briscoe et al., 2003), GPR41 and GPR43 (Brown et al., 2003)).

Expression profile of GPR55

We next investigated the expression pattern of GPR55 in a panel of mouse tissues using quantitative PCR (Figure 2). GPR55 mRNA is found in a number of tissues with the highest mRNA levels detected in the adrenals, parts of the gastrointestinal tract, as well as in the CNS. As seen with CB₁ receptors, a broad distribution of GPR55 mRNA is found in brain tissue, however the levels are significantly lower than those for CB₁ (Figure 2, inset).

GPR55 binds and is activated by cannabinoid ligands

To test the possibility that GPR55 maybe a cannabinoid receptor, we generated an N terminus FLAG-tagged human GPR55 and transiently transfected the plasmid containing the cDNA into HEK293s cells. Cell-surface expression of the recombinant receptor was confirmed using an anti-FLAG antibody (Figure 3). We then examined the ability of the cannabinoid receptor radioligands [³H]-CP55940, [³H]-SR141716 and [³H]-WIN55,212-2 to bind to membranes prepared from the transiently transfected cells. No specific binding was observed using 50 nM of each radioligand in membranes prepared from untransfected HEK293s cells. However, in membranes prepared from HEK293s cells transfected with the FLAG-tagged cDNA for human GPR55, a clear specific binding for [³H]-CP55940 was observed (Figure 4). In addition, a small specific binding for [³H]-SR141716 was seen whereas there was no binding for [³H]-WIN55,212-2 (Figure 4). Subsequent experiments repeated these findings using alternative ligands as unlabelled competitors to confirm specificity. As a consequence of these observations, we generated a HEK293s cell line stably expressing the FLAG-tagged human GPR55. Cell-surface expression was confirmed using an anti-FLAG antibody and this cell line was used for further studies.

We, next determined whether the interaction of CP55940 with GPR55 had a functional consequence. Since GTPγS has the potential to pick up activation of most heterotrimeric G proteins if the experimental conditions are appropriate, we tested membranes expressing GPR55 using a factorial design strategy with and without 1 μM CP55940 varying GDP, MgCl₂, NaCl and saponin. A number of the conditions tested generated an increased GTPγS binding in the GPR55-containing membranes, but not with control membranes, in the presence of CP55940 (data not shown). Using the optimum condition identified (see Methods), we found that CP55940 stimulated GTPγS binding with an EC₅₀ of 5 nM (Figure 5a and Table 1). With this finding we went on to evaluate other cannabinoid ligands for their ability to promote GTPγS binding via GPR55.

A number of endogenous cannabinoid ligands have been identified and characterized to date and we therefore examined their effect upon GPR55. The endocannabinoid anandamide stimulated GTPγS binding with an EC₅₀ of 18 nM (Figure 5b and Table 1). The other endocannabinoids, 2 arachidonoylglycerol (2-AG), noladin ether, palmitoylethanolamide, virodhamine and OEA all stimulated GTPγS binding with EC₅₀ values of 3, 10, 4, 12 and 440 nM respectively (Table 1). In parallel experiments, all these compounds generated the expected activities at CB₁ and CB₂ receptors (Table 1). None of these ligands had any effect when tested under identical conditions against membranes prepared from untransfected cells. Of note is the efficacy of virodhamine which under the assay conditions used is approximately 160% that of the other endocannabinoid ligands, noladin ether and 2-AG and double the efficacy of anandamide.

We next tested Δ⁹-THC, the psychoactive component of the cannabis plant C. sativa, for its activity at GPR55. Δ⁹-THC activated GTPγS binding with an EC₅₀ of 8 nM (Figure 5a and Table 1). We also examined the effect of cannabinol, cannabidiol and related compounds. Cannabidiol was without effect as an agonist in the GTPγS assay. However, cannabidiol was able to antagonize the agonist effect of CP55940 with an IC₅₀ of 445 nM (Figure 5b and Table 2). Abnormal cannabidiol functioned as an agonist with an EC₅₀ of 2.5 μM while a similar compound O1602, was significantly more potent at 13 nM. (−) 11-OH-8-Tetrahydrocannabinol-dimethylheptyl (HU210) is a highly potent CB₁ agonist and also demonstrated agonist activity at GPR55 with a potency of 26 nM, which is more than a 100-fold less potent than that found in parallel experiments at the CB₁ receptor (Table 1). A commonly used tool ligand of the cannabinoid system is WIN55,212-2. Consistent with the demonstrated lack of binding activity of this compound in our initial experiments, we observed no functional activity of WIN55,212-2 as either an agonist or antagonist (Figure 5c and Table 1). Finally, we tested the ability of known antagonists of CB₁ receptors for their effect at GPR55. 1-(2,4-Dichlorophenyl)-5-(4-iodophenyl)-4-methyl-N-4-morpholinyl-1H-pyrazole-3-carboxamide (AM281) was without effect as either an agonist or antagonist whereas 1-(2,4-dichlorophenyl)-5-(4-iodophenyl)-4-methyl-N-1-piperidinyl-1H-pyrazole-3-carboxamide (AM251) behaved as an agonist with an EC₅₀ of 39 nM. In all the experiments described above, the data were the same whether the receptor was expressed with or without the FLAG epitope.

G-protein coupling of GPR55

We next investigated the nature of the signalling pathway activated by GPR55 by examining the G-protein coupling. In the first instance, we examined the effect of Pertussis toxin on the ability of GPR55 to mediate GTPγS binding. Membranes prepared from cells treated with toxin were still able to mediate a robust response to compounds shown to be agonists of GPR55 (data not shown), suggesting that G_i G-proteins are not involved downstream of GPR55. We also tested GPR55-expressing HEK293s cells using FLIPR to determine whether there was evidence of a calcium signal that could be indicative of G_q coupling. No agonist-mediated calcium signalling was detected when compared to untransfected cells suggesting that G_q was not coupling to GPR55. To further investigate the G-protein signalling pathway downstream of GPR55 we took an antibody and peptide blocking approach in the GTPγS assay. Peptides equivalent to the last 12 amino acids of Gα_i1/2, Gα_i3, Gα_s and Gα₁₃ were incubated with GPR55-containing membranes for 15 min prior to performing GTPγS assays. The peptides equivalent to G_i1/2, G_i3 and G_s had no effect upon the GTPγS signal consistent with the lack of effect of Pertussis toxin (Figure 6a). However, the G₁₃ peptide dose dependently inhibited GTPγS binding suggesting that this peptide makes a specific interaction with GPR55 and prevents the receptor coupling to and activating G₁₃ (Figure 6a). A similar experiment was then performed using antibodies raised against the C-terminal peptides of the different G proteins. Consistent with the peptide studies anti-G_i1/2, anti-G_i3 and anti-Gα_s had no effect upon GTPγS binding mediated by GPR55 (Figure 6b). At the same time, anti-Gα₁₃ prevented GTPγS binding in a dose-dependent manner, demonstrating that the GTPγS signal being measured as a consequence of agonist activity at GPR55 was a result of G₁₃ activation (Figure 6b).

To further demonstrate that the signalling of GPR55 was Gα₁₃–mediated, we performed additional studies. Cells stably expressing human GPR55 were transiently transfected with plasmid DNA containing the human Gα₁₃ gene or with vector control. As shown in Figure 7, while the vector control did not change GTPγS readout, the membranes prepared from the Gα₁₃-transfected cells showed an augmented signal in response to cannabinoid ligands, indicative of increased expression of the coupling G protein.

Downstream signalling by GPR55

Assuming therefore that GPR55 is Gα₁₃-coupled, it is reasonable to expect that downstream signalling pathways of the G protein will be activated in a GPR55-dependent manner. To this effect, we looked at the activation of rhoA, cdc42 and rac1 in response to various ligands in GPR55-expressing and control HEK293s cells. Figure 8 shows that both anandamide and O1602 but not WIN55,212-2 treatment induced the activation of rhoA, cdc42 and rac1. This effect was blocked by the GPR55 antagonist, cannabidiol.