N-Arachidonyl Glycine Does Not Activate G Protein–Coupled Receptor 18 Signaling via Canonical Pathways

Abstract

Recent studies propose that N-arachidonyl glycine (NAGly), a carboxylic analogue of anandamide, is an endogenous ligand of the Gα_i/o protein–coupled receptor 18 (GPR18). However, a high-throughput β-arrestin–based screen failed to detect activation of GPR18 by NAGly (Yin et al., 2009; JBC, 18:12328). To address this inconsistency, this study investigated GPR18 coupling in a native neuronal system with endogenous signaling pathways and effectors. GPR18 was heterologously expressed in rat sympathetic neurons, and the modulation of N-type (Ca_v2.2) calcium channels was examined. Proper expression and trafficking of receptor were confirmed by the “rim-like” fluorescence of fluorescently tagged receptor and the positive staining of external hemagglutinin-tagged GPR18-expressing cells. Application of NAGly on GPR18-expressing neurons did not inhibit calcium currents but instead potentiated currents in a voltage-dependent manner, similar to what has previously been reported (Guo et al., 2008; J Neurophysiol, 100:1147). Other proposed agonists of GPR18, including anandamide and abnormal cannabidiol, also failed to induce inhibition of calcium currents. Mutants of GPR18, designed to constitutively activate receptors, did not tonically inhibit calcium currents, indicating a lack of GPR18 activation or coupling to endogenous G proteins. Other downstream effectors of Gα_i/o-coupled receptors, G protein–coupled inwardly rectifying potassium channels and adenylate cyclase, were not modulated by GPR18 signaling. Furthermore, GPR18 did not couple to other G proteins tested: Gα_s, Gα_z, and Gα₁₅. These results suggest NAGly is not an agonist for GPR18 or that GPR18 signaling involves noncanonical pathways not examined in these studies.

Introduction

Seven-transmembrane G protein–coupled receptors (GPCRs) are the single largest family of receptors localized to the cell surface and the most common target for currently available therapeutics (Jacoby, 2006). These receptors are defined by their ability to activate heterotrimeric guanine nucleotide binding proteins upon agonist stimulation. GPCRs without an identified endogenous ligand are considered “orphan receptors” and represent novel therapeutic targets. One such orphan receptor, G protein–coupled receptor 18 (GPR18), is found at high levels in the testis, small intestines and cells associated with the immune system including lymphocytes, thymus and spleen (Gantz et al., 1997).

Recently, N-arachidonyl glycine (NAGly) was identified as the endogenous ligand for GPR18 (Kohno et al., 2006). NAGly is a lipoamino acid found most abundantly in the spinal cord and brain (Huang et al., 2001). The chemical structure of NAGly is similar to that of anandamide (N-arachidonylethanolamide, AEA), but NAGly shows no activity at the two identified cannabinoid receptors, CB₁ and CB₂ (Huang et al., 2001). Anandamide and abnormal cannabidiol (Abn-Cbd), a synthetic cannabinoid, show agonist activity in GPR18-expressing cells (McHugh et al., 2010, 2012). Of note, anandamide exerts vasodilatory effects independent of CB₁- and CB₂R activity (Jarai et al., 1999), suggesting the existence of a third member of the cannabinoid receptor family, possibly GPR18 (McHugh et al., 2010).

GPR18 coupling to pertussis toxin–sensitive Gα_i/o proteins has been suggested in studies using cell line expression systems (Kohno et al., 2006; McHugh et al., 2010, 2012; Takenouchi et al., 2012). However, a high-throughput screening of orphan GPCRs and lipid ligands failed to detect activation of GPR18 by NAGly (Yin et al., 2009). This discrepancy could arise from the different assays used to detect G protein activation. The PathHunter assay used by Yang’s group assesses G protein–coupling using β-arrestin–mediated internalization, which is a common but not universal desensitization pathway for GPCRs. The mitogen-activated protein kinase phosphorylation assay used by Bradshaw’s group is a distant downstream effector of GPCR activation that can also be activated by multiple types of receptors (e.g., tyrosine kinase receptors). Here, we further examine the G protein signaling pathways of GPR18. We used rat superior cervical ganglion neurons to heterologously express GPR18 for study and measured inhibition of Ca²⁺ current as an assay for G protein activation. Gα_i/o protein–coupled receptor modulation of Ca²⁺ channels is mediated directly by liberated Gβγ (Herlitze et al., 1996; Ikeda, 1996) and occurs within a time course that can be observed during an electrophysiology experiment. The system is robust—to our knowledge, all Gα_i/o-coupled GPCRs that traffic to the plasma membrane produce a Ca²⁺ current inhibition when activated with a cognate agonist.

Surprisingly, we found a lack of GPR18 activation by NAGly and other proposed agonists, including anandamide and abnormal cannabidiol. Other commonly associated downstream effectors of Gα_i/o protein–coupled receptors were examined and no response to NAGly in GPR18-expressing cells was observed. This study brings into question the identity of the endogenous ligand for GPR18 and the renaming of these receptors to the NAGly-receptor.

Materials and Methods

Molecular Cloning and Mutagenesis.

Oligonucleotide primers for cloning were designed on the basis of NM_182806.1 (Mus musculus GPR18 RefSeq accession number) and commercially synthesized (IDT, Coralville, IA). GPR18 was amplified from marathon-ready mouse brain cDNA (Clontech, Mountain View, CA) using polymerase chain reaction and PfuUltra DNA polymerase (Stratagene, La Jolla, CA). To subclone into the mammalian expression vector pCI (Promega, Madison, WI), we used the following primers: forward 5′-GATCGAATT CACCATGGCC ACCCTGAGCA ATCACAACC-3′ (EcoRI site underlined) and reverse 5′-GATCGATCGC GGCCGCTCAA AGCATCTCAC TGTTCATGTT GC-3′ (NotI site underlined). A fusion construct with GPR18 and enhanced green fluorescent protein (EGFP) was designed with the following primer set: forward 5′-GATCCAAGCT TTGACCATGG CCACCCTGAG CAATCAC-3′ (HindIII site underlined) and reverse 5′-GATCGATCCC GCGGAAGCAT CTCACTGTTC ATGTTGC-3′ (SacII site underlined), and subcloned into EGFP-N1 (Clontech). A cloning error was detected within the junction of the GPR18-EGFP construct that caused a frameshift mutation and dim expression of EGFP. QuikChange (Stratagene) site-directed mutagenesis using the following primers, forward 5′-GCAACATGAA CAGTGAGATG CTTGATATCC CGCGGGCCCG GGATCC-3′ and reverse 5′-GGATCCCGGG CCCGCGGGAT ATCAAGCATC TCACTGTTCA TGTTGC-3′ (SacII site underlined, EcoRV site italicized), were used to return the EGFP sequence in-frame and introduce an EcoRV restriction site. Hemagglutinin-tagged versions of GPR18 were produced from custom-designed vectors built on a pcDNA3.1+ backbone. GPR18 was subcloned into an N-terminal 3xHA-tagging vector with the following primers: forward 5′-GCCACCCTGA GCAATCAC-3′ and reverse 5′-GATCGATCGC GGCCGCTCAA AGCATCTCAC TGTTCATGTT GC-3′ (NotI site underlined), using EcoRV and NotI restriction enzymes. A C-terminal 3xHA-tagged construct of GPR18, GPR18-3xHA, was produced by cutting and subcloning from the GPR18-EGFP construct with HindIII and EcoRV. Point mutations of GPR18 and the α_2A-adrenergic receptor (ADRA2A) were generated using QuikChange site-directed mutagenesis and the primers listed in Table 1. All generated constructs were confirmed by DNA sequencing (Supplemental Fig. 1, A and B; Macrogen, Rockville, MD).

TABLE 1.

Primers for QuikChange site-directed mutagenesis

Point mutations of GPR18 and ADRA2A were generated using a QuikChange site-directed mutagenesis strategy and the primer sets listed in table, which were commercially synthesized (IDT). All mutant constructs were confirmed by DNA sequencing.

Point Mutation	Forward Primer (5′– 3′ End)	Reverse Primer (5′–3′ End)
GPR18 A108N	GGTGGTGTTTTACCCAAGCCTCAATCTGTGGCTTCTTGC	GCAAGAAGCCACAGATTGAGGCTTGGGTAAAACACCACC
GPR18 N40A	GGGCTGTTTGTTGCTGTCACTGCGTTGTGGG	CCCACAACGCAGTGACAGCAACAAACAGCCC
GPR18 D118A	GCTTTCATTAGTGCTGCCAGATACATGGCCATCG	CGATGGCCATGTATCTGGCAGCACTAATGAAAGC
GPR18 D118T	GCTTTCATTAGTGCTACCAGATACATGGCCATCG	CGATGGCCATGTATCTGGTAGCACTAATGAAAGC
GPR18 I231E	GGTCAAGGAGAAGTCCGAACGGATCATCATGACC	GGTCATGATGATCCGTTCGGACTTCTCCTTGACC
ADRA2A N51A	CCGTGTTCGGCGCCGTGCTTGTCATCATTGCCG	CGGCAATGATGACAAGCACGGCGCCGAACACGG
ADRA2A D130A	CCATCAGCTTGGCTCGTTACTGGTCC	GGACCAGTAACGAGCCAAGCTGATGG
ADRA2A D130T	CCATCAGCTTGACTCGTTACTGGTCC	GGACCAGTAACGAGTCAAGCTGATGG
ADRA2A T373E	CCGCGAGAAGCGCTTCGAATTCGTGCTGGCGG	CCGCCAGCACGAATTCGAAGCGCTTCTCGCGG

Live- and Fixed-Cell Staining.

HeLa cells (ATCC, Manassas, VA) were cultured (2.0 × 10⁴ cells per ml) in minimal essential medium supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 100 μg/ml streptomycin (MEM+/+, Gibco, Grand Island, NY) on poly-l–lysine coated glass-bottomed dishes (MatTek, Ashland, MA). Cells were transfected with a mixture of 0.5 μg cDNA and 7 μl fully deacylated polyethylenimine (PEI) at 7.5 mM in 100 μl MEM-/− overnight.

For live-cell staining, dishes were gently washed with Dulbecco’s phosphate-buffered saline (DPBS) with Ca²⁺ and Mg²⁺ (DPBS+/+) to remove culture medium before a blocking solution, 2% bovine serum albumin (BSA) in DPBS+/+, was added for 30 minutes at 37°C. HeLa cells were then incubated with primary antibody (biotin-labeled anti-HA.11 clone 16B12, 1:200; Covance, Berkeley, CA) in blocking solution for 1 hour at 37°C. After gentle washing of cells with blocking solution, a streptavidin conjugated to quantum dot 655 (Qdot655) secondary antibody (1:200; Molecular Probes, Eugene, OR) was added for 1 hour at 37°C. Cells were washed with blocking solution for 10 minutes and then exchanged for DPBS+/+ before imaging.

For immunocytochemistry, dishes were gently washed with DPBS+/+ before fixing with 4% paraformaldehyde for 20 minutes at room temperature. After washing out fixative with DPBS+/+, HeLa cells were permeabilized with 0.5% Tween-20 in DPBS+/+ for 30 minutes at room temperature. Following solution exchange with DPBS+/+, a blocking solution, 2% BSA in Tris buffered saline with 0.05% Tween-20 (TBS-T; 10 mM Tris base, 250 mM NaCl, pH 7.5), was added for 1 hour at room temperature. Primary antibody (biotin anti-HA, 1:500; Covance) in blocking solution was incubated with cells overnight at 4°C. After cells were washed with TBS-T for 10 minutes, secondary antibody (streptavidin Qdot655, 1:500; Molecular Probes) incubation was for 2 hours at room temperature. Cells were washed with TBS-T for 10 minutes and then exchanged for DPBS+/+ before imaging.

All staining experiments included parallel negative controls, which received the same treatment except without primary antibody incubation.

Imaging.

HeLa cells and neurons expressing EGFP-constructs were imaged using a 63× (1.2 numerical aperture) or 40× (1.2 numerical aperture) objective mounted on a Zeiss LSM510 Meta confocal microscope with ZEN 2008 acquisition software (Carl Zeiss, Jena, Germany). For EGFP fluorescence, the excitation wavelength was 488 nm and emission wavelength was band-pass filtered between 500 and 550 nm. Qdot655 fluorescent images were acquired with 488-nm excitation and a 650- and 710-nm band-pass–filtered emission.

Western Blotting.

Unless otherwise indicated, all reagents for Western blotting were from Thermo Scientific (Rockford, IL). Transfected HeLa cells were lysed with mammalian protein extraction reagent with protease inhibitors for 5 minutes at room temperature. A reducing SDS loading buffer was added to lysates and heated, 85°C for 5 minutes, before electrophoresing samples on a 4–15% Tris-glycine precast gel (Bio-Rad, Hercules, CA) with Laemmli running buffer (Laemmli, 1970). Proteins were transferred to a polyvinylidene difluoride membrane at 280 mA for 1 hour. The membrane was blocked with 5% BSA in TBS-T for 3 hours before incubating overnight with a mouse anti-GFP antibody (1:2000; UC Davis/NIH NeuroMab Facility, Davis, CA) at 4°C. A goat anti-mouse horseradish peroxidase (HRP)–conjugated secondary antibody (1:1000, 1 hour) in blocking solution was applied before chemiluminescent detection of blots with SuperSignal West Femto substrate and a Kodak Image Station 4000R (Carestream Molecular Imaging, Woodbridge, CT). Membranes were stripped with Restore PLUS stripping buffer and reprobed for loading controls. Rabbit anti-tubulin (1:2000, overnight; Cell Signaling, Boston, MA) and rabbit anti–cyclophilin B (1:5000, overnight; Abcam, Cambridge, MA) primary antibodies were used, followed by a goat anti-rabbit HRP-conjugated secondary antibody (1:1000, 1 hour) and chemiluminescent detection.

Superior Cervical Ganglion Neuron Dissociation and Intranuclear Microinjection of cDNA.

All animal studies were conducted in accordance with the National Institutes of Health’s Guide for the Care and Use of Laboratory Animals.

Superior cervical ganglion (SCG) neurons from adult (6–12 weeks old) male Wistar rats were dissected and dissociated as described previously (Ikeda, 2004; Ikeda and Jeong, 2004). Briefly, animals were anesthetized by CO₂ inhalation and decapitated before dissection. Two SCGs per rat were removed, desheathed, cut into small pieces, and incubated in modified Earles’ balanced salt solution containing 2 mg/mL collagenase (CLS4; Worthington Biochemical, Lakewood, NJ), 0.6 mg/mL trypsin (Worthington Biochemical) and 0.1 mg/mL DNase I at 36°C for 1 hour in a water bath shaker oscillating at 110 rpm. The Earles’ balanced salt solution was supplemented with 3.6 g/L d-glucose and 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES). After incubation, neurons were mechanically dissociated by vigorously shaking the flask for 10 seconds. Neurons were centrifuged at 570 rpm for 6 minutes and resuspended in MEM+/+ twice before being plated on poly-l-lysine–coated tissue culture dishes. Cells were maintained in a humidified 95% air/5% CO₂ incubator at 37°C.

Three to 6 hours after dissociation, plasmid constructs were injected directly into the nucleus of SCG neurons as described previously (Ikeda, 2004; Ikeda and Jeong, 2004; Lu et al., 2009). Briefly, cDNA was injected with a FemtoJet microinjector and 5171 micromanipulator (Eppendorf, Hamburg, Germany) using an injection pressure and duration of 140–160 hPa and 0.3 second, respectively. Injected plasmids were diluted in elution buffer (10 mM Tris-HCl, pH 8.5) and centrifuged in capillary tubes at 10000 rpm for at least 30 minutes. GPR18 constructs were injected at a concentration of 50–100 ng/μl, and ADRA2A mutants were injected at a lower concentration (10 ng/μl). To identify successfully injected neurons, pEGFP-N1 cDNA (Clontech) was coinjected at a concentration of 5 ng/μl. After injections, neurons were incubated overnight at 37°C and electrophysiologic experiments were performed the following day.

Electrophysiology.

Ca²⁺-channel currents (I_Ca) and G protein–coupled inwardly rectifying K⁺ currents (I_GIRK) were recorded using conventional whole-cell patch-clamp techniques (Hamill et al., 1981). Patch electrodes were pulled from borosilicate glass capillaries (1.65 mm outer diameter, 1.20 mm inner diameter; King Precision Glass, Claremont, CA) using a Model P-97 micropipette puller (Sutter Instrument, Novato, CA). The patch electrodes were coated with silicone elastomer (Sylgard 184; Dow Corning, Midland, MI) and fire-polished. An Ag/AgCl pellet connected to the bath solution via a 0.15 M NaCl/agar bridge was used as a ground. The cell membrane capacitance was canceled and series resistance was compensated (>85% prediction and correction; lag set to 5 microseconds) with a patch-clamp amplifier (Axopatch 200A/B; Molecular Devices, Sunnyvale, CA). Voltage protocol generation and data acquisition were performed using custom-designed software (S5) on a Macintosh G4 computer (Apple, Cupertino, CA). Current traces were filtered at 2 kHz (−3 dB; four-pole Bessel), digitized at 10 kHz with a 16-bit analog-to-digital converter board (ITC-18, HEKA, Bellmore, NY) and stored on the computer for later analyses.

For recording I_Ca, patch pipettes were filled with an internal solution containing (in mM) 120 N-methyl-d-glucamine, 20 tetraethylammonium hydroxide (TEA-OH), 11 ethylene glycol tetraacetic acid (EGTA), 10 HEPES, 10 sucrose, 1 CaCl₂, 14 Tris-creatine phosphate, 4 MgATP and 0.3 Na₂GTP, pH 7.2 with methanesulfonic acid. External I_Ca recording solution consisted of (in mM) 140 methanesulfonic acid, 145 TEA-OH, 10 HEPES, 10 glucose, 10 CaCl₂ and 0.0003 tetrodotoxin (TTX), pH 7.4 with TEA-OH. A Tris-based external I_Ca solution was also tested containing (in mM) 155 Tris-base, 20 HEPES, 10 glucose, 10 CaCl₂ and 0.0003 TTX, pH 7.4 with methanesulfonic acid.

To measure G protein modulation of Ca²⁺-channels, a double-pulse protocol consisting of two 25-millisecond test pulses to +10 mV separated by a 50-millisecond conditioning pulse to +80 mV (Elmslie et al., 1990) was evoked every 10 seconds from a holding potential of −80 mV. To measure low- and high-voltage activated (LVA and HVA)-I_Ca in the same cell, two 25-millisecond pulses, the first to −40 mV and the second to +10 mV, separated by a 60-millisecond pulse to −60 mV, to inactivate LVA-I_Ca, was applied every 10 seconds from a holding potential of −80 mV. I_Ca-voltage relationships were studied by applying a series of 70-millisecond depolarizing voltage steps from a holding potential of −80 mV.

For recording I_GIRK, patch pipettes were filled with an internal solution containing (in mM) 135 KCl, 11 EGTA, 10 HEPES, 2 MgCl₂, 1 CaCl₂, 4 MgATP, and 0.3 Na₂GTP, pH 7.2 with KOH. External I_GIRK recording solution consisted of (in mM) 140 NaCl, 5.4 KCl, 10 HEPES, 15 glucose, 15 sucrose, 2 CaCl₂, 0.8 MgCl₂ and 0.0003 TTX, pH 7.4 with NaOH.

I_GIRK were elicited from 200-millisecond voltage ramps from −140 to −40 mV and the holding potential was set to −60 mV.

Live-Cell Bioluminescence Resonance Energy Transfer–based Assay for cAMP.

Human embryonic kidney (HEK)-293 cells (ATCC) were plated (5.0 × 10⁵ cells per ml) on 24-well plates in MEM+/+. Cells were transfected overnight with a mixture of 150 ng cAMP sensor using YFP-Epac-RLuc (CAMYEL) cDNA (Jiang et al., 2007), 100 ng empty vector or selected G protein–coupled receptor cDNA and 4 μl PEI in 50 μl MEM−/− per well. Approximately 16 hours after transfection, HEK cells were removed from 24-well plates with TrypLE Express (Gibco) and washed twice in DPBS+/+ before being loading onto black 96-well microplates (Berthold, Bad Wildbad, Germany).

Light intensity was measured using a Tristar LB941 luminometer (Berthold) controlled by MikroWin 2000 acquisition software (Berthold). Net bioluminescence resonance energy transfer (BRET) was calculated from the light intensity measured alternately from donor and acceptor channels in 1-second intervals using the emission filters 460/60 nm and 542/27 nm, respectively (Semrock, Rochester, NY). For testing the kinetic responses of GPCRs coupled to Gα_i/o pathways, 14 seconds of baseline recording, followed by injection of 5 μM h-coelenterazine (Nanolight Technology, Pinetop, AZ) substrate, 120 seconds of recording, injection of 1 μM forskolin, 240 seconds of recording, injection of agonist (10 μM NAGly or 100 μM glutamate), and 540 seconds of recording. For testing the kinetic responses of GPCRs coupled to Gα_s pathways, 14 seconds of baseline recording, followed by injection of 5 μM h-coelenterazine, 120 seconds of recording, injection of agonist (10 μM NAGly or 10 μM dopamine), 240 seconds of recordings, injection of 10 μM forskolin, and 240 seconds of recording.

Drugs and Chemicals.

Unless otherwise stated, all chemicals were purchased from Sigma-Aldrich. TTX was purchased from Alomone Laboratories (Jerusalem, Israel); pertussis toxin (PTX) and cholera toxin (CTX) were purchased from List Biologic Laboratories (Campbell, CA); NAGly, N-arachidonoyl-l-serine, AEA, Abn-Cbd, and O-1602 were purchased from Cayman Chemical (Ann Arbor, MI); h-coelenterazine was purchased from Nanolight Technology (Pinetop, AZ), and dopamine hydrochloride was purchased from Tocris (Bristol, UK). PTX or CTX was added (0.5 μg/ml) to the culture medium, bathing SCG neurons or HEK cells during overnight incubation (>16 hours).

Drugs for electrophysiologic experiments were diluted to final concentrations from stock solutions on the day of experiment and applied directly onto neurons using a custom-made gravity-fed perfusion system with separate perfusion lines feeding into a four-bore glass capillary tube (VitroCom, Mountain Lakes, NJ) connected to a fused silica capillary tube. A constant flow of external solution was applied onto cells during baseline recordings and switched to a drug solution during drug applications to avoid flow-induced artifacts. All recordings were performed at room temperature (20–24°C).

For BRET experiments, drugs were injected into each well by the luminometer’s injection system. Before the start of experiments, all lines were primed with distilled water and then reprimed with the drug solution to be injected. Chemicals were made the day of experiments at a concentration so the total volume (injected volume + volume in well) would produce the final desired concentration.

Data Analysis and Statistical Testing.

ImageJ software, version 1.45I (National Institutes of Health, Bethesda, MD), was used to analyze and adjust contrast of images for presentation in figures. Igor Pro, version 6 (WaveMetrics, Portland, OR), was used to analyze current traces. I_Ca amplitude was measured isochronally 10 milliseconds after the initiation of a test pulse to 10 mV or at the maximum peak I_Ca during test pulses to −40 mV. The facilitation ratio (FR) was determined as the ratio of postpulse to prepulse I_Ca. The peak I_GIRK was taken 3 milliseconds after the start of the ramp. Drug responses were normalized to baseline I_Ca or I_GIRK using the equation I_drug/I_baseline × 100, where I_drug and I_baseline are the current amplitudes during and before drug application, respectively. Net BRET was calculated as A/D − d, where A is the acceptor channel intensity, D is the donor channel intensity, and d is the background or spectral overlap (calculated previously as the A/D for donor, Rluc8, alone). A single net BRET value, 4 minutes after injection of agonist, was used for comparison of net BRET values between groups.

Statistical tests were performed with GraphPad Prism 5 for Mac OS X (GraphPad Software, La Jolla, CA). Individual data points were represented on graphs with the mean ± S.E.M. Statistical significance between two groups was determined using an unpaired t test. To compare three or more groups, a one-way analysis of variance (ANOVA) test followed by Newman-Keuls post-test was performed. P < 0.05 was considered to represent a statistically significant finding, except when a multiple comparison Bonferroni correction was applied as indicated.

Results

Heterologous Expression of Full-length GPR18 at the Membrane.

To determine the cellular localization of heterologously expressed receptor, GPR18-EGFP cDNA was injected into SCG neurons and compared with injected EGFP and EGFP-KRas tail cDNA expression, which label the cytoplasm and plasma membrane of the cell, respectively. The GPR18 construct displayed a “rim-like” fluorescence (Fig. 1A, top panel). This pattern was unlike that of cytoplasmic EGFP (Fig. 1B, top panel) but was similar to the membrane-bound EGFP-KRas tail (Fig. 1C, top panel). Line plots of fluorescence intensity (Fig. 1, A–C, top panel insets) show the highest intensity values along the edge of neurons injected with GPR18-EGFP or EGFP-KRas tail, whereas fluorescence intensities of EGFP-injected SCG neurons were uniform across the cell. HeLa cells were also transfected with EGFP-labeled constructs to compare expression in a cell line expression system (Fig. 1, A–C, bottom panels). In expressing HeLa cells, GPR18-EGFP and EGFP-KRas tail displayed a “rim-like” fluorescence pattern and EGFP displayed fluorescence throughout the cell, similar to SCG neuron expression.

Fig. 1.

Expression and localization of heterologously expressed GPR18-EGFP. Confocal images of (A) GPR18-EGFP, (B) EGFP, and (C) EGFP-KRas tail constructs expressed in SCG neurons (top panels) or HeLa cells (bottom panels). Insets of top panels: line plots of fluorescence intensity from dashed line across injected SCG neuron. Note the “rim-like” fluorescence of the GPR18-EGFP construct, similar to the membrane-bound EGFP-KRas tail-expressing cells and different from cytosolic EGFP-expressing cells. Scale bar is 20 µm. (D) Sample Western blot of EGFP-tagged constructs. Primary anti-GFP antibody (1:2000, NeuroMab) and secondary anti-mouse HRP-conjugated antibody (1:1000, Thermo) were used. Band in GPR18-EGFP lane at approximately 80 kDa, band in EGFP lane at 27 kDa, and band in EGFP-KRas tail lane at 29 kDa. (E) Stripped and reprobed blot for loading controls α-tubulin (1:2000, Cell Signaling) and cyclophilin B (1:5000, Abcam) corresponding to bands at 51 and 19 kDa, respectively. MW, molecular weight.

Western blotting was used to confirm expression of full-length GPR18. An antibody against GFP was used to detect EGFP-tagged receptor because commercially available antibodies for GPR18 have not been validated against a GPR18-knockout animal. An approximately 80-kDa band was detected in the GPR18-EGFP lane (Fig. 1D). This band is larger than the predicted size of the protein, 66 kDa, which suggests post-translational modification of the receptor. The GFP antibody also detected EGFP and EGFP-KRas tail constructs, and their bands corresponded to the predicted protein mass, 27 and 29 kDa, respectively. The same blot was reprobed for loading controls α-tubulin and cyclophilin B, which corresponded to bands of 51 and 19 kDa, respectively (Fig. 1E).

To confirm GPR18 expression in the plasma membrane, live-cell staining of epitope-tagged GPR18 was performed. The perimeter of HeLa cells expressing the external epitope-tagged version of GPR18, 3xHA-GPR18, was stained after the live-cell staining procedure (see Materials and Methods and Fig. 2A), whereas staining was absent in HeLa cells expressing the internal epitope-tagged version of GPR18, GPR18-3xHA (Fig. 2B). To confirm expression of receptors, GPR18-3xHA-expressing cells were stained after fixation and permeabilization to allow antibody access to the interior of cells. Fixed and permeabilized GPR18-3xHA cells stained positive for the HA-epitope along the edge and within HeLa cells (Fig. 2C). Staining was absent in negative controls (data not shown).

Fig. 2.

Live-cell staining and imaging of hemagglutinin-tagged GPR18. Confocal images of stained HeLa cells transfected with (A) 3xHA-GPR18 or external HA-tagged GPR18, (B and C) GPR18-3xHA or internal HA-tagged GPR18. EGFP was cotransfected to label transfected cells. (A and B) Live-cell staining with biotin-labeled anti-HA antibody (1:200; Covance) and streptavidin conjugated Qdot655 secondary antibody (1:200; Molecular Probes). (C) HA staining followed fixation and permeabilization of HeLa cells. First panels are images obtained from the GFP channel (500- to 550-nm emission band-pass filter), second panels are fluorescent images from the far-red channel (650- to 710-nm emission band-pass filter), and the last panels are merged images pseudo-colored green and purple. Scale bar is 20 µm.

Taken together, heterologously expressed GPR18 inserts with appropriate topology into the plasma membrane, where it is accessible to agonists and G proteins.

Direct Potentiation of HVA-Ca²⁺ Channels by NAGly but No GPR18-Mediated Inhibition of I_Ca by NAGly.

To study coupling of GPR18 to G proteins, we used sympathetic neurons as a heterologous expression system. We have expressed non-native GPCRs in SCG neurons by intranuclear microinjection (Ikeda, 2004; Ikeda and Jeong, 2004; Lu et al., 2009) or RNA transfection (Williams et al., 2010) and reliably recapitulated endogenous G protein signaling pathways for study (Ikeda et al., 1995; Guo and Ikeda, 2004, 2005,Guo et al., 2008b). Furthermore, we can examine G protein activity in SCG neurons by measuring GPCR-mediated inhibition of endogenous Ca²⁺ current (I_Ca). We have previously shown that Ca²⁺-channels in SCG neurons are modulated by GPCRs coupled through various G protein families: Gα_i/o (Ikeda et al., 1987, 1995; Ikeda, 1992; Zhu and Ikeda, 1993; Guo and Ikeda, 2004, 2005), including Gα_z (Jeong and Ikeda, 1998), Gα_s (Zhu and Ikeda, 1994), and Gα_q/11 (Kammermeier et al., 2000).

We first examined the effect of NAGly application on I_Ca recorded from uninjected neurons. I_Ca was evoked at 0.1 Hz using a double-pulse voltage protocol (Fig. 3Ai, inset) in solutions designed to isolate I_Ca. The facilitation ratio (FR, ⊡), defined as the ratio of I_Ca evoked in the second test pulse (postpulse, Fig. 3, ⬤) to I_Ca in the first test pulse (prepulse, Fig. 3, ○), was used as a measure of Gβγ-mediated I_Ca modulation. NAGly (10 μM) potentiated I_Ca (Fig. 3Ai), increasing both pre- and postpulse I_Ca and resulting in no change in the FR (Fig. 3, Aii and Aiii). A lower concentration of NAGly (1 μM) did not produce a change in I_Ca. As a positive control for G protein activation, norepinephrine (NE; 10 μM) was applied to the same uninjected neuron. NE, activating endogenous α₂-adrenergic receptors, produced a robust decrease in I_Ca and an increase in the FR (Fig. 3, Aii and Aiii). Our laboratory has previously documented a direct effect of lipoamino acids on voltage-gated Ca²⁺ channels (Guo et al., 2008a). Here, we reproduced the result of enhanced I_Ca, in a voltage-dependent manner and a hyperpolarized I_Ca-voltage curve with NAGly (Fig. 3B). NAGly had an effect on untagged GPR18-expressing SCG neurons similar to that of uninjected controls (Fig. 3C). A lower dose of NAGly (1 μM) produced no change in I_Ca, and a higher dose increased I_Ca (Fig. 3Ci) both pre- and postpulse currents (Fig. 3Cii) with no change in FR (Fig. 3Ciii). NE-mediated inhibition of I_Ca persisted in untagged GPR18-expressing neurons (Fig. 3, Cii and Ciii). Drug responses, expressed as I_Ca amplitude during drug application normalized to baseline I_Ca, were compared between both groups in Figure 3D. Responses to low or high concentrations of NAGly were not significantly different between uninjected controls and untagged GPR18-expressing neurons (unpaired t test, P > 0.05). NE responses were also not significantly different between uninjected controls and untagged GPR18-expressing neurons (unpaired t test; P > 0.05). In every cell tested, both concentrations of NAGly failed to inhibit I_Ca, whereas all exhibited NE-mediated inhibition of I_Ca. The enhancement of I_Ca by NAGly was not mediated by endogenous Gα_i/o protein–coupled receptors because overnight incubation of uninjected and untagged GPR18-injected SCG neurons with PTX, which uncouples Gα_i/o protein signaling from GPCRs, did not affect NAGly responses. After PTX treatment, NAGly still potentiated I_Ca in uninjected SCG neurons (106.9% ± 2.3% baseline I_Ca, n = 8) and untagged GPR18-expressing SCG neurons (112.4% ± 9.7% baseline I_Ca, n = 3).

Fig. 3.

NAGly-mediated potentiation of N-type (Ca_v2.2) Ca²⁺ channel currents (I_Ca) in rat SCG neurons. Data are from whole cell patch-clamp recordings of rat sympathetic neurons obtained at room temperature (20–24°C). (Ai) Sample superimposed I_Ca traces evoked from an uninjected SCG neuron using the double-pulse I_Ca protocol, shown as inset. Two 25-millisecond test pulses to 10 mV from a holding potential of −80 mV, separated by a 50-millisecond conditioning pulse to 80 mV. For both sample traces displayed in the figure, solid black trace is the baseline I_Ca, solid gray trace is the I_Ca during application of 10 μM NAGly, y-axis scale bar is 0.5 nA, and x-axis scale bar is 10 milliseconds. (Aii) Time course of I_Ca amplitude in an uninjected neuron during exposure to 1 and 10 μM NAGly (solid gray line) and 10 μM norepinephrine (NE, dashed black line). ○ represents the prepulse I_Ca, ● represents the postpulse I_Ca. (Aiii) Time course of the FR of the same sample cell in Aii during exposure to 1 and 10 μM NAGly and 10 μM NE. (B) Ca²⁺ current-voltage relationship of uninjected neuron before (○) and during 10 μM NAGly () application. Note the voltage-dependent enhancement of Ca²⁺ currents and the hyperpolarizing shift in the IV curve peak. (Ci) Sample superimposed I_Ca traces evoked from an untagged GPR18-injected SCG neuron using the double-pulse protocol. (Cii) Time course of I_Ca amplitude in an untagged GPR18-injected neuron during exposure to 1 and 10 μM NAGly (solid gray line) and 10 μM norepinephrine (NE, dashed black line). (Ciii) Time course of the FR of the same sample cell in Cii during exposure to 1 and 10 μM NAGly and 10 μM NE. (D) Changes in I_Ca amplitude produced by NAGly (1 or 10 μM) and NE (10 μM) from each cell are represented as individual points in the dot plot graph. Drug responses were normalized to baseline I_Ca using the equation I_drug/I_baseline × 100, where I_drug and I_baseline are I_Ca amplitudes during and before drug application, respectively. Mean ± S.E.M. drug responses are represented as lines on graph. The n values for each group are indicated on graph in parentheses. Means values between uninjected and untagged GPR18-injected neurons were not significantly different (unpaired t test, P > 0.05).

NAGly and NE responses were also tested in N- and C-terminal tagged versions of GPR18. I_Ca were potentiated by NAGly 114.6% ± 5.7% (n = 7), 105.1% ± 1.6% (n = 7), and 105.1% ± 2.8% (n = 5) from baseline I_Ca in GPR18-EGFP–, 3xHA-GPR18–, and GPR18-3xHA–expressing neurons, respectively. NE inhibited I_Ca 41.3% ± 3.7%, 43.1% ± 4.5%, and 42.9% ± 5.1% of baseline I_Ca in GPR18-EGFP–, 3xHA-GPR18–, and GPR18-3xHA–injected neurons, respectively. No significant difference in NAGly or NE responses was observed between untagged and tagged versions of GPR18 (one-way ANOVA, P > 0.05). Untagged GPR18 was heterologously expressed in cells for the rest of this study, unless otherwise stated.

Inhibition of LVA Ca²⁺ channels by NAGly (Barbara et al., 2009) was used as a positive control for agonist activity. SCG neurons do not endogenously express T-type Ca²⁺ channels (Fig. 4A), so we injected cDNA encoding Ca_v3.1 or Ca_v3.2 into cells. LVA-I_Ca elicited by a test pulse to −40 mV (Fig. 4A, inset) was transiently activated showing fast inactivation during the 25-millisecond pulse (Fig. 4B). Currents were also reversibly inhibited by application of 100 μM Ni²⁺ (Ni²⁺ inhibition of Ca_v3.1-expressing neurons, 17% ± 4.5% baseline I_Ca, n = 5; Ni²⁺ inhibition of Ca_v3.2-expressing neurons, 9.0% ± 1.1% baseline I_Ca, n = 5). NAGly applied to Ca_v3.1-injected SCG neurons inhibited LVA-I_Ca while potentiating HVA-I_Ca (Fig. 4B). In SCG neurons injected with Ca_v3.2, NAGly-induced inhibition of LVA-I_Ca (64.1% ± 4.8% baseline I_Ca, n = 3) and potentiated HVA-I_Ca (106.1% ± 10.6% baseline I_Ca, n = 3). In SCG neurons coinjected with Ca_v3.1 and GPR18, NAGly inhibited LVA-I_Ca and potentiated HVA-I_Ca (Fig. 4C). In cells injected with Ca_v3.2 and GPR18, NAGly inhibited LVA-I_Ca (52.1% ± 7.6% baseline I_Ca, n = 8) and potentiated HVA-I_Ca (109.3% ± 7.8% baseline I_Ca, n = 8). No significant difference in NAGly inhibition of LVA-I_Ca was observed between Ca_v3.1-expressing neurons with or without GPR18 coexpressed (Fig. 4D; unpaired t test, P > 0.05). NAGly enhancement of HVA-I_Ca was also not significantly different between Ca_v3.1 alone and Ca_v3.1 with GPR18-expressing neurons (Fig. 4D; unpaired t test, P > 0.05).

Fig. 4.

NAGly-mediated inhibition of LVA-I_Ca heterologously expressed in rat SCG neurons. (A) Sample I_Ca trace from an uninjected SCG neuron, elicited by the low-voltage I_Ca protocol illustrated as inset. Two 25-millisecond test pulses, the first to −40 mV and the second to 10 mV, separated by a 60-millisecond pulse to −60 mV to inactivate LVA-I_Ca. (B) Sample superimposed I_Ca traces from a Ca_v3.1-injected SCG neuron, elicited by the low-voltage I_Ca protocol. For all sample traces displayed in the figure, solid black trace is the baseline I_Ca, solid gray trace is the I_Ca during application of 10 μM NAGly, y-axis scale bar is 0.5 nA, and x-axis scale bar is 10 milliseconds. Note the inhibition of LVA-I_Ca elicited during the first test pulse and the potentiation of HVA-I_Ca elicited during the second test pulse by application of NAGly. (C) Sample superimposed I_Ca traces from an SCG neuron coinjected with Ca_v3.1 and GPR18, elicited by the low-voltage I_Ca protocol. In this set of experiments, an untagged version of GPR18 was used. (D) Changes in I_Ca amplitude, both LVA- and HVA-I_Ca, produced by NAGly (10 μM) from each cell are represented as individual points in the dot plot graph. NAGly responses were normalized to baseline I_Ca. Mean ± S.E.M. NAGly responses are represented as lines on graph. The n values for each group are indicated on graph in parentheses. Mean values between Ca_v3.1 alone and Ca_v3.1 with GPR18-injected groups were not significantly different (unpaired t test, P > 0.05).

To ensure the external recording solution was not interfering with drug activity at the receptor, a Tris-based external I_Ca solution was also tested. No significant difference in NAGly responses was observed between uninjected and GPR18-expressing SCG neurons (for controls, 122.4% ± 3.2% baseline I_Ca, n = 9; for GPR18-expressing neurons, 115.5% ± 2.2% baseline I_Ca, n = 10; unpaired t test, P > 0.05) or NE responses (for controls, 44.1% ± 4.2%, n = 9; for GPR18-expressing neurons, 40.8% ± 2.7%, n = 10, unpaired t test, P > 0.05).

Although a positive effect from GPR18 has been observed from mouse-derived cell lines endogenously expressing GPR18 (McHugh et al., 2010; Burstein et al., 2011; Takenouchi et al., 2012), most GPR18 studies have used human GPR18 (Gantz et al., 1997; Kohno et al., 2006; Qin et al., 2011; McHugh et al., 2012). To assess the variability of GPR18 across species, protein sequences from different species were analyzed by protein alignment. On the basis of GPR18 protein sequence alignment (Supplemental Fig. 1C), mouse GPR18 is 95.2% identical, 97% similar to rat GPR18 and 85.8% identical, 92.1% similar to human GPR18. The greatest divergence in GPR18 protein appears in the N terminus. Because the sequences are highly similar in the transmembrane domains and important signaling regions of GPCRs, we are confident our mouse GPR18 clone is similar to studies of human GPR18. Also, because mouse and rat GPR18 protein sequences are highly similar, we do not anticipate difficulty heterologously expressing the mouse GPR18 clone in rat neurons.

Thus, we observed direct effects of NAGly only on LVA-I_Ca and HVA-I_Ca that are not G protein mediated. Furthermore, we did not observe NAGly-mediated inhibition of HVA-I_Ca in GPR18-injected neurons.

Other Potential Agonists of GPR18 Do Not Inhibit I_Ca.

Other proposed agonists of GPR18 were tested on SCG neurons expressing GPR18, and inhibition of I_Ca was used as a measure of G protein activation (Fig. 5). N-arachidonoyl-l-serine (10 µM), another lipoamino acid, potentiated I_Ca in all groups tested: uninjected, GPR18-expressing, and CB₁R-expressing SCG neurons. No significant difference between groups was observed (one-way ANOVA, P > 0.05). AEA (10 µM), the endocannabinoid neurotransmitter, did not inhibit I_Ca in uninjected or GPR18-injected cells. AEA inhibited I_Ca in CB₁R-expressing neurons, which was significantly different from all groups (one-way ANOVA, P < 0.05). Synthetic cannabinoids, abnormal cannabidiol and O-1602, had no effect on baseline I_Ca in uninjected or GPR18-expressing neurons. NE was applied to all cells as a positive control for G protein modulation and was found to inhibit I_Ca. NE responses from CB₁R-injected neurons are significantly less than the other groups tested (one-way ANOVA, P < 0.001), possibly because of the overexpression of exogenous receptor and sequestration of available G proteins from other GPCRs.

Fig. 5.

Effectiveness of various proposed agonists of GPR18 to inhibit I_Ca in GPR18- or CB₁R-expressing SCG neurons. In this set of experiments, an untagged version of GPR18 was used. N-arachidonoyl-l-serine, AEA, and NE were applied to uninjected, GPR18- and CB₁R-injected neurons. Because of the lack of a confirmed endogenous receptor, Abn-Cbd and O-1602 were applied to uninjected and GPR18-injected neurons only. Changes in I_Ca amplitude in response to various agonists from each cell are represented as individual points in the dot plot graph, normalized to baseline I_Ca. Mean ± S.E.M. drug responses are represented as lines on graph. The n values for each group are indicated on graph in parentheses. For N-arachidonoyl-l-serine, AEA, and NE treatment groups, a one-way ANOVA followed by Newman-Keuls post-test was used to compare groups. For Abn-Cbd and O-1602, an unpaired t test was used. *P < 0.05, ***P < 0.001.

Mutations in GPCRs that Induce Tonic Receptor Activity Do Not Activate GPR18.

To bypass the need for agonists to activate GPR18, mutations that induce constitutive activity of receptors were introduced. These mutations were based on mutagenesis studies of the α_1B-adrenergic receptor (Cotecchia et al., 1990; Kjelsberg et al., 1992; Scheer et al., 1996, 1997), as illustrated in Supplemental Fig. 2 with snake plot diagrams (Supplemental Fig. 2A) and sequence alignments (Supplemental Fig. 2B). Site-directed mutagenesis and primers listed in Table 1 were used to generate mutants of GPR18 and a Gα_i/o protein–coupled receptor, ADRA2A.

The properties of I_Ca inhibition induced by constitutively active Gα_i/o-coupled receptors are analogous to the inhibition induced by agonist application: a high FR and kinetic slowing during prolonged voltage depolarization. Other GPCR responses may also be inhibited by expression of constitutively active receptors because excess free-Gβγ is loaded onto downstream effectors, thereby effectively reducing the dynamic range of Gβγ-mediated responses. Constitutively active GPCRs that inhibit I_Ca through voltage-independent mechanisms are more difficult to assess because inhibited currents display similar kinetics as uninhibited currents and the mechanisms responsible for I_Ca inhibition can be quite diverse. The overall reduction in I_Ca amplitude, or I_Ca density, may indicate tonic receptor activity for such mechanisms of I_Ca inhibition. Thus, basal FR, NE-mediated I_Ca inhibition, and I_Ca density were used as measures of tonic receptor activity.

The D/ERY motif is common in all GPCRs, and mutations in the first residue of the motif can induce constitutive activity of receptors (Scheer et al., 1996, 1997). Expression of a GPR18 version of this mutation, GPR18 D118T-EGFP, was localized primarily in a membrane network inside the cell, reminiscent of the endoplasmic reticulum, and not in the plasma membrane (Fig. 6A, left). A substitution mutation of the glutamate residue to alanine, an amino acid that confers less constitutive activity than threonine (Scheer et al., 1997), did not alter expression of the GPR18 mutant from the endoplasmic reticulum (data not shown). I_Ca from GPR18 D118T-expressing neurons was elicited using the double-pulse voltage protocol (Fig. 6A, right) and were found to be indistinguishable from controls. The increase in I_Ca amplitude in the postpulse, following a long depolarizing conditioning pulse, represents relief of basal Gβγ-mediated inhibition of I_Ca and is responsible for a basal FR > 1. The basal FR measured in GPR18 D118T- and GPR18 D118A-injected neurons was similar to that in uninjected controls (Table 2). NE responses were also similar between GPR18 D118T and uninjected controls (Fig. 6A, right, and Table 2). In contrast, I_Ca measured from SCG neurons expressing ADRA2A D130T had a significantly larger basal FR than its GPR18 mutant counterpart (Table 2, unpaired t test, P < 0.05), which was clearly observed with the double-pulse voltage protocol (Fig. 6B). The kinetic slowing of I_Ca was present during the first test pulse of ADRA2A D130T mutants (Fig. 6B). The increased basal FR in ADRA2A D130T mutants was blocked by overnight PTX treatment (ADRA2A D130T basal FR after PTX, 1.3 ± 0.06, n = 7), indicating tonically active ADRA2A receptors coupled to Gα_i/o proteins.

Fig. 6.

Mutant class A GPCRs designed to induce or alter tonic activity of receptors. (A) Left: Confocal image of GPR18 D118T-EGFP construct expressed in HeLa cells. Note the fluorescence primarily inside of the cell in the internal membrane network, reminiscent of the endoplasmic reticulum. Scale bar is 20 µm. Right: Sample superimposed I_Ca traces from an untagged GPR18 D118T-injected SCG neuron, elicited by the double-pulse I_Ca protocol. For all sample traces displayed in the figure, solid black trace is the baseline I_Ca, dashed black trace is the I_Ca during application of 10 μM NE, solid gray trace is the I_Ca during application of 10 μM NAGly, y-axis scale bar is 0.5 nA, and x-axis scale bar is 10 milliseconds. Horizontal dashed gray line from the peak of the postpulse I_Ca highlights the relief of tonic Ca²⁺ channel inhibition by the conditioning pulse. (B) Sample superimposed I_Ca traces from an ADRA2A D130T-injected SCG neuron, elicited by the double-pulse I_Ca protocol. Note the kinetic slowing and inhibition of prepulse I_Ca and the large relief of tonic Ca²⁺ channel inhibition by the conditioning pulse in the baseline I_Ca trace. (C) Responses of a proposed nonconstitutively active mutant of GPR18, GPR18 A108N (Qin et al., 2011), to NAGly and other agonists of GPR18. Left: Confocal image of GPR18 A108N-EGFP construct expressed in HeLa cells. Note the “rim-like” fluorescence in transfected cells. Scale bar is 20 µm. Right: Sample superimposed I_Ca traces from an untagged GPR18 A108N-injected SCG neuron, elicited by the double-pulse I_Ca protocol. Below: Time course of I_Ca amplitude in an untagged GPR18 A108N-expressing neuron during exposure to 10 μM NAGly, 10 μM AEA, and 10 μM Abn-Cbd. ○ represents the prepulse I_Ca, ● represents the postpulse I_Ca. Note the potentiation of pre- and postpulse I_Ca after application of NAGly and lack of I_Ca response to AEA and Abn-Cbd application. (D) Changes in I_Ca amplitude, normalized to baseline I_Ca, from each cell, are represented as individual points in the dot plot graph. Mean ± S.E.M. drug responses are represented as lines on graph. The n values for each group are indicated on graph in parentheses. To compare groups, a one-way ANOVA followed by Newman-Keuls post-test was used. *P < 0.05, ***P < 0.001.

TABLE 2.

Summary of constitutively active mutants

Mutants and ADRA2A alone were injected into SCG neurons and Ca²⁺ currents (I_Ca) were elicited using the double-pulse protocol. Basal FR, a sensitive indicator of tonic receptor activity, was determined from the ratio of postpulse to prepulse I_Ca of the first recording obtained from each cell. NE responses were normalized to baseline I_Ca using the equation I_drug/I_baseline × 100, where I_drug and I_baseline are Ca²⁺ current amplitudes during and before NE application, respectively. The I_Ca density was determined from the postpulse I_Ca divided by the capacitance of the cell, calculated from integrating the area under the current trace obtained from a 10-mV step applied before cell capacitance compensation. Units in picoamperes per picofarad, or pA/pF. Mean ± S.E.M. values are indicated with number of cells in parentheses. Unpaired t tests were performed to compare each untagged GPR18 mutant with the analogous ADRA2A mutant. To account for multiple comparisons, a Bonferroni correction was used and P < 0.017 was considered statistically significant.

Construct	Basal FR	NE Response (% Baseline ICa)	ICa Density (pA/pF)
GPR18 D118T	1.3 ± 0.04 (13)	36.7 ± 3.1 (13)	−28.5 ± 3.5 (13)
ADRA2A D130T	2.2 ± 0.23 (18)a	42.9 ± 5.0 (16)	−22.8 ± 3.1 (18)
GPR18 D118A	1.2 ± 0.03 (23)	46.2 ± 3.2 (22)	−23.6 ± 2.6 (23)
ADRA2A D130A	1.8 ± 0.19 (12)a	35.3 ± 4.5 (12)	−25.1 ± 3.4 (12)
GPR18 N40A	1.3 ± 0.04 (14)	43.6 ± 2.3 (13)	−29.4 ± 3.3 (14)
ADRA2A N51A	2.6 ± 0.21 (11)a	74.8 ± 11.1 (6)a	−23.9 ± 4.8 (11)
GPR18 I231E	1.2 ± 0.04 (5)	40.8 ± 6.8 (5)	−17.7 ± 3.8 (5)
ADRA2A T373E	2.1 ± 0.3 (7)a	32.7 ± 3.5 (6)	−22.1 ± 3.3 (7)
Uninjected	1.3 ± 0.03 (14)	39.9 ± 2.8 (13)	−34.1 ± 4.1 (14)
ADRA2A	1.6 ± 0.2 (14)	35.6 ± 5.5 (14)	−20.8 ± 2.8 (13)

^aP < 0.017.

Mutations in the highly conserved asparagine residue within the first transmembrane domain of GPCRs (Supplemental Fig. 2, A and B) can also induce constitutive activity (Scheer et al., 1996). In addition, mutations in the C-terminal end of the third intracellular loop have been described for ADRA1B (Cotecchia et al., 1990; Kjelsberg et al., 1992) and ADRA2A (Ren et al., 1993) to induce constitutive activity. Analogous mutations were generated in GPR18 and ADRA2A, based on sequence alignment to ADRA1B (Supplemental Fig. 2B) and topographic location (Supplemental Fig. 2A), and constitutive activity was assessed. Mutations introduced into GPR18 failed to tonically active receptor, as measured by basal FR, NE response, and I_Ca density (Table 2). On the other hand, mutations designed to constitutively activate the ADRA2A receptor produced significantly larger basal FR than did similar mutations in GPR18 (Table 2; unpaired t test with multiple comparison correction, P < 0.017) and the ADRA2A N51A mutant significantly reduced the effectiveness of NE-induced inhibition of I_Ca.

Thus, mutations predicted to induce constitutive receptor activity did not activate GPR18 but were effective when introduced into ADRA2A receptors, which tonically activate Gα_i/o signaling pathways.

Proposed GPR18 Agonists Do Not Inhibit I_Ca in SCG Neurons Expressing a Nonconstitutively Active Mutant of GPR18.

A recent study found endogenous GPR18 is constitutively active and differentially expressed in melanoma metastases (Qin et al., 2011). However, we found no indication GPR18 expressed in SCG neurons was constitutively active (basal FR, 1.3 ± 0.04, n = 25; NE-mediated I_Ca inhibition, 47.4% ± 3.6% baseline I_Ca, n = 17; I_Ca density, −25.4 ± 2.2 picoamperes per picofarad, pA/pF, n = 25). This paper also describes a single amino acid residue responsible for conferring constitutive activity of GPR18 (A108) and mutating the residue to an asparagine restored NAGly-induced G protein signaling. We tested this constitutively active null mutant by monitoring I_Ca in SCG neurons during agonist application.

HeLa cells transfected with GPR18 A108N-EGFP displayed a “rim-like” fluorescence pattern (Fig. 6C, left). In SCG neurons expressing GPR18 A108N, NAGly potentiated pre- and postpulse I_Ca elicited with the double-pulse voltage protocol (Fig. 6C, right). Other proposed agonists of GPR18, AEA and abnormal cannabidiol, were also tested in GPR18 A108N-injected cells, but neither induced I_Ca inhibition (Fig. 6C, lower panel). Only the positive control, NE, produced a significant reduction in I_Ca in GPR18 A108N-expressing neurons (Fig. 6D, one-way ANOVA, P < 0.001).

No Evidence of GPR18 Coupling to Gα_z.

The lack of GPR18 activation by NAGly prompted us to explore possible coupling of GPR18 to other Gα proteins not endogenously expressed in SCG neurons. We have previously expressed Gα_z in SCG neurons (Jeong and Ikeda, 1998) and found coupling of Gα_z to endogenous GPCRs (Fig. 7A). In Gα_z-expressing cells, NE-induced I_Ca inhibition by a Gβγ-mediated mechanism: there were kinetic slowing of I_Ca during NE application in the first test pulse (Fig. 7Ai), substantial relief of prepulse I_Ca inhibition by the conditioning pulse, and thus increase in FR during NE application (Fig. 7Aiii). The kinetics of NE activation and deactivation in Gα_z-injected cells (Fig. 7Aii) were slower than endogenous Gα_i/o protein coupling (Fig. 3Aiii), which is consistent with the slower intrinsic GTPase activity of Gα_z compared with Gα_i or Gα_o. Together with the persistence of NE-mediated inhibition of I_Ca in Gα_z-expressing neurons after overnight PTX treatment (Fig. 7C) suggests coupling of endogenous α₂-adrenergic receptors to Gα_z.

Fig. 7.

Functional coupling of endogenous α₂-adrenergic receptors to Gα_z but not untagged GPR18. (Ai) Sample superimposed I_Ca traces evoked from a Gα_z-injected SCG neuron using the double-pulse I_Ca protocol. For both sample traces displayed in Figure, solid black trace is the baseline I_Ca, dashed black trace is the I_Ca during application of 10 μM NE, solid gray trace is the I_Ca during application of 10 μM NAGly, y-axis scale bar is 0.5 nA, and x-axis scale bar is 10 milliseconds. (Aii) Time course of I_Ca amplitude in a Gα_z-injected neuron during exposure to 10 μM NAGly (solid gray line) and 10 μM NE (dashed black line). ○ represents the prepulse I_Ca, ● represents the postpulse I_Ca. Note the slower onset and off-rate of NE compared with uninjected SCG neurons (Fig. 3Aii). (Aiii) Time course of the facilitation ratio (FR) of the same sample cell in Aii during exposure to 10 μM NAGly and 10 μM NE. (Bi) Sample superimposed I_Ca traces evoked from an SCG neuron coexpressing Gα_z and untagged GPR18, using the double-pulse I_Ca protocol. (Bii) Time course of I_Ca amplitude in a Gα_z + untagged GPR18-injected neuron during exposure to 10 μM NAGly (solid gray line) and 10 μM NE (dashed black line). (Biii) Time course of FR of the same sample cell as in Bii. (C) Changes in I_Ca amplitude produced by NAGly (10 μM) and NE (10 μM) from each cell are represented as individual points in the dot plot graph. Mean ± S.E.M. drug responses are represented as lines on graph. NE-mediated responses were elicited from SCG neurons after overnight PTX treatment. The n values for each group are indicated on graph in parentheses. To compare groups, a one-way ANOVA followed by Newman-Keuls post-test was performed. ***P < 0.001. (D) Testing possible coupling of constitutively active GPR18 with Gα_z in SCG neurons. (Di) The basal FR, a sensitive indicator of tonic receptor activity, was determined from the ratio of postpulse to prepulse I_Ca of the first recording obtained from each cell. Basal FR values from each cell are represented as individual points in the dot plot graph. Mean ± S.E.M. basal FR are represented as lines on graph. To compare groups, a one-way ANOVA followed by Newman-Keuls post-test was performed. ***P < 0.001. (Dii) Changes in I_Ca amplitude in response to NE, normalized to baseline I_Ca, from each cell, are represented as individual points in the dot plot graph. Mean ± S.E.M. NE response are represented as lines on graph. No significant difference was observed between groups, as determined by one-way ANOVA followed by Newman-Keuls post-test (P > 0.05).

NAGly slightly potentiated I_Ca in Gα_z-injected SCG neurons (Fig. 7Ai), increasing both pre- and postpulse I_Ca (Fig. 7Aii) resulting in no change of the FR (Fig. 7Aiii). In SCG neurons injected with Gα_z and GPR18, NE, but not NAGly, induced inhibition of I_Ca (Fig. 7, Bi and Bii) and increased the FR during NE application (Fig. 7Biii). NAGly responses were not significantly different between uninjected, Gα_z alone, GPR18 coexpressing Gα_z and GPR18 A108N coexpressing Gα_z groups (Fig. 7C, one-way ANOVA, P > 0.05). NE responses in PTX-treated neurons were significantly lower in Gα_z alone, GPR18 coexpressing Gα_z, and GPR18 A108N coexpressing Gα_z groups compared with uninjected controls (Fig. 7C; one-way ANOVA, P < 0.001).

Mutants of GPR18 designed to induce constitutive activity were also coexpressed with Gα_z to test possible coupling. Basal FR was significantly reduced in SCG neurons expressing Gα_z compared with uninjected controls (one-way ANOVA, P < 0.001), but no significant difference in basal FR was found between the various GPR18 mutants when coexpressed with Gα_z (Fig. 7Di). NE-induced inhibition of I_Ca was also unchanged across all groups tested (Fig. 7Dii).

Therefore, we found no evidence of GPR18 coupling to Gα_z.

No Evidence of GPR18 Coupling to Gα₁₅.

Potential coupling of GPR18 to Gα₁₅ was also tested. Gα₁₅ is highly expressed in hematopoietic cells (Giannone et al., 2010), similar to GPR18 expression. Although Gα₁₅ is considered a promiscuous G protein, capable of coupling various classes of GPCRs, reconstitution of Gα₁₅ signaling and functional coupling of this G protein to endogenous GPCRs in SCG neurons has yet to be demonstrated.

To demonstrate functional coupling of a GPCR to Gα₁₅ in SCG neurons, we used I_Ca inhibition as an assay for G protein activity. Expression of Gα₁₅ alone suppressed endogenous NE-mediated signaling (Supplemental Fig. 3A). This artifact of Gα protein overexpression, which sequesters available Gβγ protein, could be alleviated by coinjecting Gβ₁ and Gγ₂ with Gα₁₅ to restore the stoichiometric balance of heterotrimeric G protein signaling (Supplemental Fig. 3B). The basal FR of Gα₁₅ was also restored to uninjected control levels after coexpressing Gβ₁γ₂ (for Gα₁₅ alone, basal FR, 1.0 ± 0.01, n = 16; for Gα₁₅β₁γ₂, basal FR, 1.3 ± 0.02, n = 36). Endogenous α₂-adrenoceptors do not couple to PTX-insensitive Gα₁₅ because there was no NE-mediated inhibition of I_Ca in Gα₁₅β₁γ₂-expressing cells after overnight PTX treatment (Supplemental Fig. 3C). Because mGluR2 has been shown to couple to Gα₁₅ in vitro (Gomeza et al., 1996), mGluR2 and Gα₁₅β₁γ₂ were coexpressed in SCG neurons and glutamate-induced inhibition of I_Ca after PTX treatment was measured (Fig. 8A). The properties of I_Ca inhibition by mGluR2 coupled to Gα₁₅ were distinct from mGluR2 coupled to Gα_i/o (Supplemental Fig. 3, D and E): There was no kinetic slowing during the first test pulse of the double-pulse protocol, almost equal inhibition of the pre- and postpulse I_Ca and therefore a smaller change in the FR, and insensitivity to PTX. Gα_i/o-mediated pathways still dominated mGluR2 signaling because glutamate induced a change in FR in cells coexpressing Gα₁₅β₁γ₂ and mGluR2 (Supplemental Fig. 3F).

Fig. 8.

Functional coupling of heterologously expressed mGluR2 receptors to Gα₁₅ but not untagged GPR18. (A) Demonstration of functional coupling of mGluR2 to Gα₁₅. For controls of functional coupling to Gα₁₅, see Supplemental Fig. 3. (Ai) Sample superimposed I_Ca traces evoked from an SCG neuron coexpressing Gα₁₅β₁γ₂ and mGluR2 using the double-pulse I_Ca protocol, following overnight PTX treatment. For both sample traces displayed in the figure, solid black trace is the baseline I_Ca, dashed black trace is the I_Ca during application of 10 μM NE, dashed gray trace is the I_Ca during application of 100 μM glutamate, y-axis scale bar is 0.5 nA, and x-axis scale bar is 10 milliseconds. (Aii) Time course of I_Ca amplitude in a Gα₁₅β₁γ₂ + mGluR2-injected neuron during exposure to 10 μM NE (dashed black line) and 100 μM glutamate (dashed gray line). ○ represents the prepulse I_Ca, ● represents the postpulse I_Ca. (Aiii) Time course of the FR of the same sample cell in Aii during exposure to 10 μM NE and 100 μM glutamate. (Bi) Sample superimposed I_Ca traces evoked from an SCG neuron coexpressing Gα₁₅β₁γ₂ and untagged GPR18, using the double-pulse I_Ca protocol, following overnight PTX treatment. (Bii) Time course of I_Ca amplitude in an SCG neuron coexpressing Gα₁₅β₁γ₂ and untagged GPR18 during exposure to 10 μM NAGly (solid gray line) and 10 μM NE (dashed black line). (Biii) Time course of the FR of the same sample cell in Bii. (C) Changes in I_Ca amplitude produced by NAGly (10 μM) and NE (10 μM) from each cell are represented as individual points in the dot plot graph. Mean ± S.E.M. drug response are represented as lines on graph. Drug responses were tested in SCG neurons after overnight PTX treatment. The n values for each group are indicated on graph in parentheses. To compare groups, a one-way ANOVA followed by Newman-Keuls post-test was performed, but no significant difference was observed (P > 0.05). (D) To test possible coupling of constitutively active GPR18 with Gα₁₅ in SCG neurons, basal I_Ca density was measured. Basal I_Ca density was determined from the postpulse I_Ca divided by the capacitance of the cell, calculated from integrating the area under the current trace obtained from a 10-mV step applied before cell capacitance compensation. I_Ca density values from each cell are represented as individual points in the dot plot graph. Mean ± S.E.M. I_Ca density is represented as lines on graph. I_Ca density was measured in SCG neurons after overnight PTX treatment. To compare groups, a one-way ANOVA followed by Newman-Keuls post-test was performed. *P < 0.05.

In PTX-treated SCG neurons injected with Gα₁₅β₁γ₂ and GPR18, NAGly potentiated pre- and postpulse I_Ca evoked by the double-pulse voltage protocol (Fig. 8, Bi and 8Bii), with no change in the FR (Fig. 8Biii). No NAGly-induced inhibition of I_Ca was observed in PTX-treated SCG neurons coexpressing Gα₁₅β₁γ₂ and GPR18 or GPR18 A108N (Fig. 8C), and NAGly responses were not significantly different between all groups tested (one-way ANOVA, P > 0.05). No NE-induced inhibition of I_Ca was observed in any PTX-treated group tested (Fig. 8C).

Mutants of GPR18 designed to induce constitutive activity were also coexpressed with Gα₁₅β₁γ₂ to test possible coupling. I_Ca density was significantly reduced in Gα₁₅β₁γ₂ and mGluR2-injected neurons during glutamate application (one-way ANOVA, P < 0.05) but mutants of GPR18 coexpressed with Gα₁₅β₁γ₂ did not change I_Ca density compared with uninjected controls (Fig. 8D).

Therefore, we found no evidence of GPR18 coupling to Gα₁₅.

Potentiation of I_GIRK by NAGly but No GPR18-mediated Potentiation of I_GIRK.

Previous work has suggested GPR18 couples to Gα_i/o protein signaling pathways (Kohno et al., 2006; McHugh et al., 2010, 2012; Takenouchi et al., 2012), but we have not observed any GPR18-mediated inhibition of I_Ca, which is a primary downstream effector of Gα_i/o. We tested other effectors of Gα_i/o proteins, including G protein–coupled inwardly-rectifying K+ currents (I_GIRK), to further investigate GPR18 signaling. GIRK channels expressed in SCG neurons open upon G protein activation by a Gβγ-mediated mechanism.

SCG neurons do not endogenously express GIRK channels, so we injected cDNA encoding homomeric GIRK channel subunits, GIRK4 S143T (Vivaudou et al., 1997). I_GIRK was evoked at 0.1 Hz using a voltage ramp (Fig. 9Ai, inset) in solutions designed to isolate I_GIRK. In SCG neurons expressing GIRK4 S143T, peak I_GIRK increased during NAGly treatment (Fig. 9, Ai and Aii). As a positive control for Gα_i/o protein activation, NE was applied and peak I_GIRK also increased (Fig. 9, Ai and Aii). In SCG neurons expressing GIRK4 S143T and GPR18, both NAGly and NE potentiated I_GIRK (Fig. 9, Bi and Bii). NE significantly increased I_GIRK in SCG neurons expressing GIRK4 S143T with or without GPR18 (Fig. 9C; one-way ANOVA, P < 0.001). But the modulation of I_GIRK by NAGly was modest and only reached significance in SCG neurons expressing GIRK4 S143T alone with and without PTX treatment (Fig. 9C), suggesting that the effect of NAGly on I_GIRK is independent of GPR18 expression. Furthermore, NAGly’s effect on I_GIRK is not Gα_i/o protein mediated because overnight PTX treatment did not block NAGly responses on I_GIRK but did block NE responses (Fig. 9C).

Fig. 9.

Testing modulation of I_GIRK channels by untagged GPR18. (Ai) Sample superimposed I_GIRK traces evoked from a GIRK4 S143T-injected SCG neuron using the voltage ramp, shown as inset. A 200-millisecond voltage ramp from −140 to −40 mV, from a holding potential of −60 mV, was used to elicit I_GIRK. For both sample traces displayed in the figure, solid black trace is the baseline I_GIRK, dashed black trace is the I_GIRK during application of 10 μM NE, solid gray trace is the I_GIRK during application of 10 μM NAGly, and y-axis scale bar is 1 nA. (Aii) Time course of I_GIRK amplitude in a GIRK4 S143T-injected neuron during exposure to 10 μM NAGly (solid gray line) and 10 μM NE (dashed black line). (Bi) Sample superimposed I_GIRK traces evoked from an SCG neuron coinjected with GIRK4 S143T and untagged GPR18 using the I_GIRK voltage ramp protocol. (Bii) Time course of I_GIRK amplitude in a GIRK4 S143T + GPR18-injected neuron during exposure to 10 μM NAGly (solid gray line) and 10 μM NE (dashed black line). (C) Changes in I_GIRK amplitude produced by NAGly (10 μM) and NE (10 μM) from each cell are represented as individual points in the dot plot graph. Drug responses were normalized to baseline I_GIRK. Mean ± S.E.M. drug responses represented as lines on graph. The n values for each group are indicated on graph in parentheses. To compare groups, a one-way ANOVA followed by Newman-Keuls post-test was performed. *P < 0.05, ***P < 0.001.

Thus, NAGly activation of GPR18 was not observed with G protein modulation of I_GIRK currents as an assay.

No NAGly-induced Change in cAMP Levels in GPR18-Expressing HEK Cells.

Another downstream effector of G proteins is adenylate cyclase, which converts ATP to cAMP. Live-cell cAMP levels were monitored in HEK cells transfected with the BRET-based cAMP sensor, CAMYEL (Jiang et al., 2007), loaded onto a multiwell luminescence plate reader. Net BRET was calculated from measurements of light intensity from the acceptor and donor channels after application of enzyme substrate, h-coelanterizene. High net BRET values indicate low intracellular cAMP levels, and low net BRET values indicate high intracellular cAMP levels.

The ability of Gα_i/o protein–coupled receptors to inhibit adenylate cyclase and reduce intracellular cAMP levels after forskolin stimulation was used as a measure of Gα_i/o protein activation. Application of forskolin (1 μM) produced a decrease in net BRET, and NAGly application failed to change net BRET levels in empty vector- or GPR18-expressing HEK cells (Fig. 10Ai). A comparison of net BRET values 4 minutes after NAGly application shows no significant difference between empty vector and GPR18-expressing cells (Fig. 10Aii; unpaired t test, P > 0.05). Samples preloaded with NAGly 3 hours before experiments did not significantly reduce the forskolin-induced decrease in net BRET (for empty vector, net BRET, 0.19 ± 0.004, n = 10; for GPR18-expressing HEKs, net BRET, 0.20 ± 0.005, n = 10; unpaired t test, P > 0.05). On the other hand, glutamate treatment of mGluR2-transfected HEK cells increased net BRET values after forskolin stimulation, which was blocked by overnight PTX treatment (Fig. 10Bi). The increase in net BRET values 4 minutes after glutamate application was significant for the mGluR2-transfected group (Fig. 10Bii; one-way ANOVA, P < 0.05).

Fig. 10.

Monitoring modulation of adenylate cyclase using the BRET-based cAMP sensor CAMYEL. Data are obtained from live HEK cells loaded in a microplate luminometer. HEK cells were transfected with empty vector or selected G protein–coupled receptor cDNA, CAMYEL cDNA, and PEI. Approximately 16 hours after transfection, cells were transferred to a black 96-well microplate and loaded into a luminometer for recording. Net BRET was calculated as A/D – d, where A is the light intensity measured from the acceptor channel (542/27 nm) for 1 second, D is the light intensity measured from the donor channel (460/60 nm) for 1 second, and d is the background or spectral overlap, calculated previously as the A/D value for Rluc8 alone. Net BRET values are inversely related to cAMP levels (high net BRET = low intracellular cAMP levels, low net BRET = high intracellular cAMP levels). (A and B) Test of Gα_i/o-mediated inhibition of forskolin stimulated cAMP production. (Ai) Time course of net BRET readings from empty vector and untagged GPR18-expressing HEK cells. F, forskolin (1 μM); NG, NAGly (10 μM). (Aii) Net BRET values 4 minutes after injection of NAGly from each sample well are represented as individual points in the dot plot graph. Mean ± S.E.M. net BRET values are represented as lines on graph. The n values for each group are indicated on graph in parentheses. No significant difference was observed between empty vector and GPR18-transfected cells (unpaired t test, P > 0.05). (Bi) Time course of net BRET readings from empty vector and mGluR2-expressing HEK cells. F, forskolin (1 μM); G, glutamate (100 μM). PTX was applied to HEK cells during overnight incubation. (Bii) Net BRET values 4 minutes after injection of glutamate from each sample well are represented as individual points in the dot plot graph. Mean ± S.E.M. net BRET values are represented as lines on graph. A significant increase in mean net BRET value after glutamate application was observed in the mGluR2-transfected group (one-way ANOVA followed by Newman-Keuls post-test, ***P < 0.001). (C and D) Test of Gα_s-mediated stimulation of cAMP production. (Ci) Time course of net BRET readings from empty vector and untagged GPR18-expressing HEK cells. F, forskolin (10 μM). (Cii) Net BRET values 4 minutes after injection of NAGly from each sample well are represented as individual points in the dot plot graph. Mean ± S.E.M. net BRET values are represented as lines on graph. No significant difference was observed between empty vector and GPR18-transfected cells (unpaired t test, P > 0.05). Di) Time course of net BRET readings from empty vector and D₁R-expressing HEK cells. D, dopamine hydrochloride (10 μM); F, forskolin (10 μM). CTX was applied to HEK cells during overnight incubation. (Dii) Net BRET values 4 minutes after injection of dopamine from each sample well are represented as individual points in the dot plot graph. Mean ± S.E.M. net BRET values are represented as lines on graph. A significant decrease in mean net BRET value after dopamine application was observed in the D₁R-transfected group (one-way ANOVA followed by Newman-Keuls post-test, ***P < 0.001).

Gα_s-coupled pathways stimulate adenylate cyclase and the ability of GPCRs to increase intracellular cAMP levels was used as a measure of Gα_s-protein activation. NAGly did not significantly change net BRET levels in empty vector- or GPR18-expressing HEK cells (Fig. 10C; unpaired t test, P > 0.05). But dopamine applied to D₁R-transfected HEK cells was able to decrease net BRET values, which was blocked by overnight CTX treatment (Fig. 10Di). Only the D₁R-transfected group significantly reduced net BRET values 4 minutes after dopamine application (Fig. 10Dii; one-way ANOVA, P < 0.001).

With the BRET-based cAMP sensor, we found no evidence of NAGly-mediated GPR18 coupling to Gα_i/o or Gα_s signaling pathways.

Discussion

In the present study, we could not activate GPR18 that was heterologously expressed in a native neuronal system. NAGly is the proposed endogenous ligand for GPR18 (Kohno et al., 2006) and mediates microglia migration through a PTX-sensitive pathway (McHugh et al., 2010); however, we did not observe activation of Gα_i/o protein signaling after NAGly application on heterologously expressed GPR18 in SCG neurons (Fig. 3). Moreover, other putative agonists, including AEA, abnormal cannabidiol and O-1602 did not induce I_Ca inhibition in GPR18-injected SCG neurons (Fig. 5). Likewise, a high-throughput screen of 43 lipid ligands, which included AEA and NAGly, also failed to activate GPR18 (Yin et al., 2009). This was surprising because most, if not all, endogenous or heterologously expressed Gα_i/o protein–coupled receptors in SCG neurons negatively couple to N-type Ca²⁺ channels via a PTX-sensitive pathway. Instead, NAGly produced a consistent increase in I_Ca (Fig. 3), similar to what has previously been reported by our laboratory (Guo et al., 2008a). As positive controls, NAGly potently inhibited heterologously expressed T-type Ca²⁺ channels confirming agonist activity (Fig. 4), and full-length GPR18 protein was expressed and trafficked to the plasma membrane, as demonstrated with the “rim-like” fluorescence pattern of a GFP-tagged version of the receptor (Fig. 1) and positive staining of external HA-tagged GPR18 (Fig. 2). However, in the absence of an actual receptor response, expression of functional GPR18 receptors in this study remains to be determined. A protein sequence alignment of GPR18 (Supplemental Fig. 1C) suggests the mouse GPR18 clone used in our study is similar to functional GPR18 in other studies (Kohno et al., 2006; Qin et al., 2011; McHugh et al., 2012) and capable of proper protein expression. Other downstream effectors of Gα_i/o were examined, but NAGly failed to activate GPR18. I_GIRK was potentiated by NAGly (Fig. 9), but this effect was independent of GPR18 expression and Gα_i/o protein coupling because 1) NAGly-induced potentiation of current was present in GIRK4 S143T alone injected SCG neurons and 2) augmentation of I_GIRK by NAGly persisted after overnight PTX treatment. In GPR18-expressing cells, cAMP levels were not altered by NAGly after forskolin-induced cAMP production (Fig. 10A), suggesting no inhibition of adenylate cyclase by GPR18. Furthermore, GPR18 failed to couple to other Gα proteins, Gα_z (Fig. 7) and Gα₁₅ (Fig. 8), heterologously expressed in SCG neurons. We have previously demonstrated GPCR coupling to Gα_z using agonist-mediated I_Ca inhibition (Jeong and Ikeda, 1998), but this is the first demonstration of negative coupling of activated Gα₁₅ to HVA-Ca²⁺ channels. Thus, modulation of N-type Ca²⁺ channels is a versatile assay of G protein activity because it is a common downstream effector of multiple G protein families.

Mutants of GPR18 that are predicted to confer constitutive activity, failed to tonically activate (i.e., in the absence of overt agonist) the receptor according to the measures of tonic G protein activity used in this study (Fig. 6; Table 2). On the other hand, analogous mutations in ADRA2A (of the same class A of GPCRs as GPR18) were tonically active when expressed in rat sympathetic neurons (Fig. 6B; Table 2). After expression of ADRA2A mutants, the basal FR, a sensitive indicator of tonic G protein activation (Ikeda, 1991), was significantly elevated (Table 2). This elevation was abolished after PTX treatment implicating tonic activation of endogenous Gα_i/o proteins. It should be noted that some mutations designed to confer tonic activity of GPCRs also increase agonist potency (Cotecchia et al., 1990), suggesting tonic G protein activity may actually represent increased sensitivity to endogenous agonist levels. This possibility cannot be excluded before testing receptor antagonists. Other possible consequences of tonic GPCR activity, such as activation of β-arrestin–mediated internalization, are not quantifiable with the electrophysiological techniques used in this study but may nevertheless occur. Redistribution of mutant GPR18 receptors into the endoplasmic reticulum of cells (Fig. 6A) may reflect this mechanism, or mutations in the receptor may affect receptor trafficking. The lack of receptor expression at the plasma membrane may confound measurements of G protein activity. However, we have described other heterologously expressed GPCRs where GFP-fusion constructs are located primarily inside the cell but receptors maintain functional G protein signaling (Guo et al., 2008b). Thus, we infer that a small amount of GPCR expressed in the plasma membrane, which may not be noticeable with GFP-fusion constructs, is sufficient to carry out G protein activity. Assuming the mutations introduced into GPR18 were successful in producing a tonically active receptor, the inability to measure changes in basal G protein activity may indicate the existence of a mediator that couples GPR18 to G proteins, which is missing in SCG neurons or lost by cellular dialysis during whole-cell recordings. Heterodimerization of GPR18 with another GPCR may facilitate G protein signaling of GPR18 by heterodimer directed signaling (for reviews see Hudson et al., 2010; Marshall and Foord, 2010) and studies in endogenously expressing GPR18 cells may be necessary to find its GPCR binding partner and to recapitulate G protein signaling of GPR18.

The only positive responses of NAGly observed in this study were its direct effects on voltage-gated Ca²⁺ channels (Figs. 3 and 4). Similar effects of lipoamino acids and arachidonic acid, a product of fatty acid amide hydrolase (FAAH) hydrolysis of NAGly (Grazia Cascio et al., 2004), on ion channel function have been documented (Chemin et al., 2007; Guo et al., 2008a; Barbara et al., 2009). However, it is unlikely that the effects observed in this study are a result of NAGly breakdown to arachidonic acid and glycine because arachidonic acid reduces N-type Ca²⁺ channel amplitude (Liu and Rittenhouse, 2000) and the effect of NAGly on T-type Ca²⁺ channels is independent of FAAH activity (Barbara et al., 2009). NAGly has many reported actions: Gα_i/o-mediated activation of large-conductance Ca²⁺-sensitive K⁺ (BK) channels (Begg et al., 2003; Parmar and Ho, 2010), partial agonism of GPR92 (Oh et al., 2008), and blockade of glycine uptake via the glycine transporter, GLYT2 (Wiles et al., 2006), to name a few. Of note, NAGly is also a potent competitive substrate with AEA for FAAH (Huang et al., 2001) and inhibiting FAAH can increase endogenous AEA levels (Burstein et al., 2002). It is unclear which target of NAGly is responsible for the NAGly-induced cell migration or apoptosis observed in other recombinant systems (McHugh et al., 2010, 2012; Takenouchi et al., 2012) or whether GPR18 signaling is cell-type dependent. For instance, if NAGly-induced signaling is intimately related to resting endocannabinoid levels, the lack of endogenous endocannabinoid production in SCG neurons (Won et al., 2009) may be responsible for this study’s inability to reconstitute GPR18 signaling pathways.

GPR18-mediated signaling directly in neurons has not been demonstrated. GPR18 activity has been implicated in neuronal function by virtue of its regulation of microglial function (McHugh, 2012), but no detectable levels of GPR18 transcript are found in human brain (Gantz et al., 1997). The lack of receptors in neuronal tissues contradicts the abundance of endogenous NAGly in the spinal cord and brain (Huang et al., 2001) and the ability of NAGly to affect neuronal activity. NAGly can modulate nociceptive signaling in the spinal cord (Huang et al., 2001; Vuong et al., 2008), although it is unclear whether GPR18 signaling in microglia is responsible for this effect or whether other targets of NAGly, such as GLYT2, play a more dominant role. NAGly can also indirectly affect neuronal excitability and synaptic transmission by altering circulating levels of endogenous endocannabinoids (Burstein et al., 2002), which can signal through CB₁Rs located throughout the CNS. The disconnect between localization of GPR18 and endogenous NAGly supports the presence of another receptor for NAGly, which is responsible for signaling in neurons. Other identified targets of NAGly may fulfill this role, but there remains a GPCR responsive to NAGly that can regulate neuronal activity. Two separate studies (Begg et al., 2003; Parmar and Ho, 2010) have implicated a NAGly-sensitive Gα_i/o-coupled receptor in the regulation of BK-channels, a channel that repolarizes the membrane potential and is an important regulator of action potential duration. This receptor has yet to be identified.

In a native neuronal system, heterologously expressed GPR18 is not activated by NAGly, thus corroborating the result obtained from the high-throughput screen of lipid ligands (Yin et al., 2009). This would argue against the de-orphanization of GPR18 until we better understand all of the elements involved in GPR18 signaling. Perhaps studies examining NAGly-mediated responses from endogenously expressing GPR18 cells will shed some light on this signaling pathway.

Abbreviations

Abn-Cbd

abnormal cannabidiol

ADRA2A

α_2A-adrenergic receptor

AEA

anandamide, N-arachidonylethanolamide

ANOVA

analysis of variance

BRET

bioluminescence resonance energy transfer

BSA

bovine serum albumin

CAMYEL

cAMP sensor using YFP-Epac-RLuc

CTX

cholera toxin

DPBS

Dulbecco’s phosphate buffered saline

EGFP

enhanced green fluorescent protein

FAAH

fatty acid amide hydrolase

FR

facilitation ratio

GPR18

G protein–coupled receptor 18

GPCR

G protein–coupled receptors

HEK

human embryonic kidney

HEPES

4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

HRP

horseradish peroxidase

HVA

high voltage activated

I_GIRK

G protein–coupled inwardly rectifying K⁺ currents

LVA

low voltage activated

MEM

minimum essential medium, NAGly, N-arachidonyl glycine

NE

norepinephrine

PEI

polyethylenimine

PTX

pertussis toxin

Qdot655

quantum dot 655

SCG

superior cervical ganglion

TBS-T

Tris buffered saline with 0.05% Tween-20

TEA-OH

tetraethylammonium hydroxide

TTX

tetrodotoxin

Authorship Contributions

Participated in research design: Lu, Puhl, Ikeda

Conducted experiments: Lu

Contributed new reagents or analytic tools: Puhl

Performed data analysis: Lu

Wrote or contributed to writing of the manuscript: Lu, Puhl, Ikeda