D₁-D₂ Dopamine Receptor Synergy Promotes Calcium Signaling via Multiple Mechanisms

Abstract

The D₁ dopamine receptor (D₁R) has been proposed to form a hetero-oligomer with the D₂ dopamine receptor (D₂R), which in turn results in a complex that couples to phospholipase C–mediated intracellular calcium release. We have sought to elucidate the pharmacology and mechanism of action of this putative signaling pathway. Dopamine dose-response curves assaying intracellular calcium mobilization in cells heterologously expressing the D₁ and D₂ subtypes, either alone or in combination, and using subtype selective ligands revealed that concurrent stimulation is required for coupling. Surprisingly, characterization of a putative D₁-D₂ heteromer-selective ligand, 6-chloro-2,3,4,5-tetrahydro-3-methyl-1-(3-methylphenyl)-1H-3-benzazepine-7,8-diol (SKF83959), found no stimulation of calcium release, but it did find a broad range of cross-reactivity with other G protein–coupled receptors. In contrast, SKF83959 appeared to be an antagonist of calcium mobilization. Overexpression of G_qα with the D₁ and D₂ dopamine receptors enhanced the dopamine-stimulated calcium response. However, this was also observed in cells expressing G_qα with only the D₁R. Inactivation of G_i or G_s with pertussis or cholera toxin, respectively, largely, but not entirely, reduced the calcium response in D₁R and D₂R cotransfected cells. Moreover, sequestration of G_βγ subunits through overexpression of G protein receptor kinase 2 mutants either completely or largely eliminated dopamine-stimulated calcium mobilization. Our data suggest that the mechanism of D₁R/D₂R–mediated calcium signaling involves more than receptor-mediated G_q protein activation, may largely involve downstream signaling pathways, and may not be completely heteromer-specific. In addition, SKF83959 may not exhibit selective activation of D₁-D₂ heteromers, and its significant cross-reactivity to other receptors warrants careful interpretation of its use in vivo.

Introduction

Dopamine is a neurotransmitter that functions in the central nervous system to regulate neural processes that include motor control, cognition, and memory. Dysregulation of the dopamine (DA) system is associated with neurologic disorders such as Parkinson disease, schizophrenia, addiction, and attention deficit hyperactivity disorder. Five DA receptor (DAR) genes exist in mammals, each of which encodes a DAR subtype (D₁R–D₅R); these genes are grouped by structure and function into the D1-like (D₁R and D₅R) and D2-like (D₂R, D₃R, and D₄R) DAR families. The D1-like receptors couple to the G_s/olf proteins to activate adenylyl cyclase–mediated formation of cAMP, whereas the D2-like receptors couple to the G_i/o proteins to inhibit adenylyl cyclase (Sibley and Monsma, 1992; Missale et al., 1998). Several studies, however, have proposed DAR-mediated signaling pathways that do not involve activation of either G_i/o or G_s/olf proteins.

The first evidence for alternate signaling pathways came from multiple studies reporting “D1-like” receptor stimulation of intracellular calcium mobilization, which was suggested to be a result of G_q-mediated activation of phospholipase C (PLC) (Mahan et al., 1990; Undie and Friedman, 1990; Wang et al., 1995; Pacheco and Jope, 1997). Subsequently, it was shown that in vitro cell cultures coexpressing the D₁R and D₂R could couple to intracellular calcium mobilization through the G_q-PLC-diacylglycerol pathway (Lee et al., 2004; Rashid et al., 2007a). This calcium response required both coexpression and coactivation of both receptor subtypes. This led to the proposal of a “noncanonical” mechanism for DAR-mediated signaling wherein the D₁R forms a heteromeric complex with the D₂R and induces PLC-mediated intracellular calcium mobilization (Lee et al., 2004; Rashid et al., 2007b; Hasbi et al., 2011). The precise mechanism for this type of signaling and its prevalence in vivo, however, remain unclear.

In vivo, there is evidence both for (Surmeier et al., 1992, 1996; Lester et al., 1993; Ariano et al., 1997; Aizman et al., 2000; Lee et al., 2004) and against (Gerfen et al., 1990; Le Moine et al., 1991; Hersch et al., 1995; Le Moine and Bloch, 1995; Bertran-Gonzalez et al., 2008) the existence of neural cells coexpressing both D₁R and D₂R. Interestingly, some neurons that appear to coexpress D₁R and D₂R have neuronal projections that express only D₁R or only D₂R (Lee et al., 2004). This finding, along with the different methods of detection and visualization, may partially explain the incongruent reports of D₁R and D₂R colocalization. However, several recent studies using confocal FRET techniques argue for direct demonstration of the existence of D₁-D₂ heteromers in 10–20% of the cell bodies and presynaptic terminals of medium spiny neurons within the nucleus accumbens (Hasbi et al., 2009; Perreault et al., 2011, 2012a), and the two DARs have been shown to cointernalize after selective activation of either receptor (O’Dowd et al., 2005; So et al., 2005).

Interestingly, several agonists of the benzazepine family seem to exhibit differential effects on the D₁R monomer compared with the proposed D₁-D₂ heteromer (Rashid et al., 2007b). One such compound, 6-chloro-2,3,4,5-tetrahydro-1-(3-methylphenyl)-3-(2-propenyl)-1H-3-benzazepine-7,8-diol (SKF83822), has been proposed to selectively activate D₁R-mediated cAMP production but have no effect on calcium mobilization (Rashid et al., 2007a,b). In contrast, another benzazepine, 6-chloro-2,3,4,5-tetrahydro-3-methyl-1-(3-methylphenyl)-1H-3-benzazepine-7,8-diol (SKF83959), has been proposed to selectively activate the heteromer-mediated calcium release and have no effect on cAMP production (Rashid et al., 2007a,b; Hasbi et al., 2011). More recent studies have used this finding to interpret the results of systemic SKF83959 injections in mice, which resulted in increased Ca²⁺/calmodulin-dependent protein kinase IIα phosphorylation and increased brain-derived neurotrophic factor expression in striatal neurons (Hasbi et al., 2009; Ng et al., 2010). It was also shown that expression of glutamate decarboxylase-67 and the vesicular glutamate transporters 1 and 2 in striatal neurons, when injected into rats, was altered by SKF83959 (Perreault et al., 2012b), which, again, was interpreted to be due to selective D₁-D₂ heteromer activation.

In the current study, we further investigated the biology and pharmacology of the proposed D₁-D₂ heteromer and the mechanism of calcium mobilization in heterologous expression systems. Although we found that coactivation of both D₁R and D₂R protomers is required for calcium mobilization to occur, there appear to be multiple mechanisms besides G_q activation through which this pathway is elicited. We also studied the functional characteristics of SKF83959 to determine its viability as a heteromer-selective in vivo ligand and found that it was significantly less selective than previously appreciated. In fact, we were not able to provide evidence for selective activation of the D₁-D₂ heteromer. These results indicate that D₁Rs and D₂Rs can synergize to induce calcium mobilization, although the mechanisms of activation are multiple and complex and there is not, as yet, a selective pharmacology.

Materials and Methods

Human embryonic kidney 293-tsa201 (HEK293T) cells were a gift from Dr. Vanitha Ramakrishnan. A D₁R expressing stable cell line was purchased from Codex Biosolutions, Inc. (Gaithersburg, MD). [³H]N-methyl-(R)-(+)-7-chloro-8-hydroxy-3-methyl-1-phenyl-2,3,4,5-tetrahydro-1H-3-benzazepine (SCH23390) (80.5 Ci/mmol) and [³H]N-methylspiperone (85.5 Ci/mmol) were obtained from PerkinElmer Life Sciences (Waltham, MA). Cell culture media and reagents were purchased from MediaTech/Cellgro (Manassas, VA). Cell culture flasks and materials and all assay plates were purchased from Greiner Bio-One (Monroe, NC). SKF83959 and SKF83822 were purchased from Tocris Bioscience/RD Systems (Minneapolis, MN). All other compounds and buffer components were purchased from Sigma-Aldrich (St. Louis, MO) except where indicated.

Cell Culture and Transfection.

HEK293T cells and D₁R CODEX cells were maintained in Dulbecco’s modified Eagle’s medium supplemented with a final concentration of 10% fetal bovine serum, 100 U/ml penicillin, 100 μg/ml streptomycin, 1 mM sodium pyruvate, and 10 μg/ml gentamicin. Cells were incubated at 37°C, 5% CO₂, and 90% humidity. They were passaged and plated mechanically using calcium-free Earle’s balanced salt solution and pelleted by centrifugation at 1000g for 10 minutes. For transfection studies, HEK293T cells were seeded in 150-mm plates at 10 × 10⁶ cells per plate. After 24 hours, cells were transfected according to the manufacturer’s recommendations using Clontech’s CalPhos transfection kit (Clontech Laboratories, Inc., Mountain View, CA). The DAR plasmid constructs were FLAG-tagged rat D₁R, D_2SR (D₂ short splice variant), or D_2LR (D₂ long splice variant) in the pCD-SRα vector (Takebe et al., 1988; Monsma et al., 1990; Zhang et al., 1994) and D₄R in pcDNA3.1(+) vector (Schetz and Sibley, 2001). Additional experiments were done using the G_q protein in the pcDNA3.1(+) vector (Missouri S&T cDNA Resource Center, Rolla, MO) and various functionally dominant negative G protein receptor kinase 2 (GRK2) mutants: GRK2 C-terminus 495–689 in pcDNA3(+), GRK2 K220R in pcDNA3(+), and empty pcDNA3.1(+) (Koch et al., 1994; Freedman et al., 1995). For all transfections, 5 μg of each DNA construct was used to transfect cells, with the exception of D₁R, in which 10 μg was used.

Radioligand Binding Assays.

Forty-eight hours after transfection, cells were dissociated from plates using calcium-free Earle’s balanced salt solution, and intact cells were collected by centrifugation at 900g for 10 minutes. Cells were resuspended and lysed using 5 mM Tris-HCl and 5 mM MgCl₂ at pH 7.4 at 4°C. Cell lysate was pelleted by centrifugation at 20,000g for 30 minutes and resuspended in 5 mM Tris-HCl at pH 7.4; 100 μl of cell lysate (containing 8 μg of protein for D₂R assays or 10 μg of protein for D₁R assays) was incubated for 90 minutes at room temperature with various concentrations of [³H]N-methyl-SCH23390 (D₁R binding) or [³H]N-methylspiperone (D₂R binding) in a final reaction volume of 250 μl. Nonspecific binding was determined in the presence of 4 μM (+)-butaclamol. Bound ligand was separated from the unbound by filtration through a PerkinElmer Unifilter-96 GF/C 96-well microplate using the PerkinElmer Unifilter-96 Harvester, washing three times, 1 ml per well in ice-cold assay buffer. After drying, 50 μl of liquid scintillation cocktail (MicroScint PS; PerkinElmer) was added to each well, plates were sealed, and the plates were analyzed on a PerkinElmer Topcount NXT. For competition binding assays, a fixed concentration of 0.5 nM [³H]N-methyl-SCH23390 was incubated with various concentrations of SKF83959, and the remainder of the assay was performed as described already herein. K_i values were calculated from observed IC₅₀ values using the Cheng-Prusoff equation and a K_d value of 0.5 nM for SCH23390, as determined in independent saturation isotherms (unpublished data). Expression of the D₄R was determined in an identical assay format as that for the D₂R.

Competition Radioligand Binding Screen.

A primary, single-point radioligand competition binding assay was performed to assay for radioligand binding inhibition by SKF83959 (10 μM). Forty-three G protein–coupled receptors (GPCRs) and neurotransmitter-related proteins were screened in the primary assay using radioligands with known binding properties. The percentage of inhibition was calculated by subtracting the percentage of specific binding in the presence of the test compound from the percentage of specific binding in the absence of the test compound (n = 4). Receptors whose corresponding radioligands had greater than 50% inhibition at 10 μM SKF83959 underwent secondary radioligand competition binding assays to generate full competition curves. K_i determinations and receptor binding profiles were provided by the National Institute of Mental Health (NIMH) Psychoactive Drug Screening Program (PDSP), Contract HHSN-271-2008-00025-C. The NIMH PDSP is directed by Dr. Bryan L. Roth (University of North Carolina, Chapel Hill, NC) and by Project Officer Jamie Driscol (NIMH, Bethesda, MD). For experimental details, including radioligands used and associated K_d values for each individual receptor, please refer to the PDSP website: http://pdsp.med.unc.edu/.

Calcium Mobilization Assays.

HEK293T cells were transiently transfected as described; 24 hours after transfection, cells were plated in 384-well, optical, clear-bottom, black-walled plates (20 μl/well, 30,000 cells/well; Greiner Bio-One). Forty-eight hours after transfection, cells were incubated for 60 minutes at room temperature in the dark with Fluo-8 NW calcium dye and an extracellular signal quencher to block any signal from extracellular calcium (Screen Quest Fluo-8 NW Calcium Assay Kit; AAT Bioquest, Inc., Sunnyvale, CA), as recommended by the manufacturer. The plates were then treated with various concentrations of antagonist or agonists (diluted in the presence of 0.2 mM sodium metabisulfite) as indicated in the Results and figure legends. For agonist reads, plates were read kinetically in real-time (every 0.6 second) by recording a baseline read for 14 seconds before the addition of an agonist compound and then continually measured for 2 minutes after agonist addition. For antagonist reads, plates were read kinetically in real-time (every 0.6 second) by recording a baseline reading for 20 seconds before the addition of that antagonist. Then, 3 minutes later, agonist compound was added, and the plates were read for an additional 3 minutes. All compound additions were done in unison using the 384-tip onboard robotics on a Functional Drug Screening System (FDSS) μCell (Hamamatsu, Bridgewater, NJ), and plates were continuously read using the FDSS μCell from the bottom throughout the assay with an excitation wavelength of 480 nm and an emission wavelength of 540 nm. Data were recorded and quantified as maximum minus minimum (max−min) relative fluorescence units within the assay window using FDSS software. Data are expressed as a percentage of the control max−min relative fluorescence units for given studies as indicated in the figure legends. In these experiments, D₁R and D₂R receptor expression levels typically varied between 1 and 3 pmol/mg protein. We found that coexpressing both receptors sometimes affected their expression compared with expressing them alone (unpublished data). However, this did not affect the calcium mobilization response, which, although not studied in detail, appeared to require simply a minimum level of dual receptor expression.

Statistical Analysis.

Data are expressed as a percentage of control values for individual experiments. Nonlinear regression of all data was conducted on GraphPad Prism 5.01 (GraphPad Software, Inc., La Jolla, CA). Results are expressed as mean ± S.E.M.

See Supplemental Materials and Methods section for additional procedures.

Results

Previous studies have suggested that the D₁-D₂ receptor complex may signal as a heteromer and have implicated SKF83959 as a compound that may selectively activate this signaling complex (Lee et al., 2004; So et al., 2005; Rashid et al., 2007a,b). However, these findings have not been corroborated, and the mechanisms by which the D₁-D₂ receptor complex signals remain unclear. To investigate the apparent ability of D₁-D₂ receptor oligomerization to alter the G protein coupling of component receptors, we first transiently expressed the D₁R either alone or concurrently with either the short (D_2SR) or long (D_2LR) isoforms of the D₂R and measured intracellular calcium mobilization via kinetic fluorescence imaging. Preliminary coimmunoprecipitation experiments revealed that D₁-D₂ hetero-oligomers were indeed capable of forming under these expression conditions (Supplemental Fig. 1). When cells were transfected with the D₁R and D_2LR or the D₁R with D_2SR, a clear dose-dependent activation of calcium mobilization was observed in response to DA (Fig. 1). Importantly, we observed no difference in coupling efficacy or agonist potency between the short and long isoforms of the D₂R. However, when cells were transfected with any of the subtypes alone, the receptors failed to couple to calcium mobilization (Fig. 1). These data suggest that expression and activation of both the D₁R and D₂R are essential for coupling to calcium mobilization and signaling.

Fig. 1.

Agonist-induced calcium mobilization in DA receptor–transfected cells. HEK293T cells were transiently transfected with D₁R, D_2LR, D_2SR, D₁R + D_2LR, or D₁R + D_2SR, as indicated and described in Materials and Methods. Twenty-four hours later, cells were plated in 384-well plates and assayed the following day for calcium mobilization after stimulation by DA (D₁R + D_2LR EC₅₀ = 73.8 nM, D₁R + D_2SR, EC₅₀ = 58.2 nM). Data are representative of three independent experiments done with the same assay conditions on different days. Data are expressed as percentage of control, normalized to the maximum signal seen via DA stimulation of D₁R + D_2LR transfected cells. Error bars indicate S.E.M. from multiple wells within the representative experiment.

To investigate further that the activation of both receptor subtypes is required to stimulate calcium mobilization, we used receptor subtype–selective antagonists. Concentration response inhibition curves for the D₁R-selective (SCH23390) and the D₂R-selective (sulpiride) antagonists were generated for cells transfected with the D₁R and D₂R (Fig. 2, A and B). Cells were simultaneously stimulated with 1 μM DA and examined for calcium mobilization. We observed complete inhibition of the calcium signal with either SCH23390 or sulpiride treatment. The potencies of the antagonists (SCH23390 IC₅₀ ∼8.0 nM, sulpiride IC₅₀ ∼0.7 nM) are consistent with their known affinities for their selective subtypes as determined in our laboratory (unpublished data) as well as by other groups (Seeman and Van Tol, 1993; Millan et al., 2001). More importantly, complete inhibition of the calcium response is seen at antagonist concentrations that have no effect on the opposite receptor subtype. Thus, selectively blocking DA activation of either receptor subtype is sufficient to prevent calcium mobilization, further suggesting that both receptor protomers must be activated for this signaling to occur.

Fig. 2.

Inhibition of D₁R + D_2LR–mediated calcium mobilization by either D₁R- or D₂R-selective antagonists. HEK293T cells were transfected with D₁R + D_2LR as described and 24 hours later were plated in 384-well plates. Cells were incubated with the indicated concentrations of the D₁R-selective antagonist SCH23390 (A) or the D₂R-selective antagonist sulpiride (B) and then stimulated with an ∼EC₈₀ of DA (1 μM; SCH23390 IC₅₀ = 8.0 nM, sulpiride, IC₅₀ = 0.7 nM). Data are expressed as a percentage of the control (10 μM) DA response and are representative of two independent experiments performed with the same assay conditions on different days. Error bars indicate S.E.M. from multiple wells within the representative experiment.

Whereas the studies using subtype-selective antagonists suggested that both D₁R and D₂R are required for calcium signaling, it might be possible that stabilizing one subtype into an inactive state within a heteromer might alter the conformation of the corresponding partner. Thus, to elucidate further the coupling mechanism, subtype-selective agonists were used to determine whether indeed activation of both protomers is required for calcium mobilization. As seen in Fig. 3, concurrent administration of a D₁R-selective (SKF83822) and a D₂R-selective (quinpirole) agonist to cells cotransfected with D₁R and D₂R resulted in a calcium mobilization response that nearly matched that of DA. In contrast, when D₁R plus D₂R–cotransfected cells were stimulated with quinpirole alone, no calcium mobilization was observed. Furthermore, when the cotransfected cells were stimulated with SKF83822, no calcium mobilization was seen at concentrations selective for D₁R. A small response was observed at 10 μM, but this was at a concentration where SKF83822 loses receptor subtype selectivity and can begin to stimulate the D₂R as well. Previous studies showed that SKF83822 has an affinity for D₁R in the ∼2 nM range and D₂R in the ∼200 nM range (O’Sullivan et al., 2004). Experiments done in our laboratory have demonstrated a D₂R affinity that is greater than 10 μM (unpublished data), supporting the idea that the SKF83822-mediated calcium response seen at high concentrations is due to nonselective receptor activation. In addition, when cells were transfected with any of the subtypes individually, no signal was seen from any of the agonists (unpublished data). Taken together, these data indicate that stimulation of both receptor subtypes is necessary for calcium mobilization.

Fig. 3.

Stimulation of D₁R + D_2LR–mediated calcium mobilization by either D₁R- or D₂R-selective agonists. HEK293T cells were transfected with D₁ R+ D_2LR as described, plated 24 hours later in 384-well plates, and assayed for calcium accumulation the following day. Cells were stimulated with one of the following agonists as indicated: DA, the D₁R-selective agonist SKF83822, the D₂R-selective agonist quinpirole, or both SKF83822 and quinpirole (D₁R + D_2LR EC₅₀ = 610.8 nM) combined. Control cells expressing the D₁R, D_2SR, or D_2LR individually did not show a significant calcium response to concurrent agonist administration. Data are expressed as a percentage of control maximum DA-stimulated response and are representative of two independent experiments performed with the same assay conditions on different days. Error bars indicate S.E.M. from multiple wells within the representative experiment.

Previous studies suggested that SKF83959 may be a D₁-D₂ heteromer-selective compound, and a significant calcium response to this ligand has been reported in cells coexpressing the D₁R and D₂R (Lee et al., 2004; Rashid et al., 2007a,b; Beaulieu and Gainetdinov, 2011). This compound has also been reported to have seemingly paradoxical effects on the D₁R, exhibiting both antagonist and agonist properties, depending on the system (Panchalingam and Undie, 2001; Cools et al., 2002; Zhang et al., 2005). In our current studies, we treated D₁R and D₂R cotransfected cells with SKF83959 and, surprisingly, were unable to elicit a calcium response (Fig. 4A). Furthermore, when SKF83959 was added in concert with the D₂R selective agonist quinpirole, we were still unable to observe a significant calcium response. It should be noted that SKF83959 consistently failed to stimulate calcium mobilization even when this experiment was performed using different lots of compound from different vendors on separate days, as well as with different drug solvents (unpublished data). We also had one lot of compound chemically analyzed to verify its purity (unpublished data). To demonstrate that the SKF83959 compound was pharmacologically active in our hands, we performed two separate experiments. As shown in Fig. 4B, we stimulated calcium mobilization with DA and then dose dependently added either the D₁R-selective antagonist SCH23390 as a control (see Fig. 2A) or SKF83959 to see whether it might function as an antagonist in this system. In fact, it did, exhibiting even higher potency than SCH23390, although its efficacy of antagonism was less, exhibiting a maximum inhibition of ∼50%. Finally, we performed a radioligand binding competition assay with SKF83959 and cells transfected with the D₁R (Fig. 4C). SKF83959 was able to compete potently and fully for radioligand binding to the D₁R. These experiments (Fig. 4, B and C) demonstrate that SKF83959 is active in binding to the monomeric D₁R, as well as active as a partial antagonist of the calcium response observed in D₁R and D₂R cotransfected cells. In contrast, it does not appear to function as an agonist with respect to stimulating calcium mobilization in the D₁R and D₂R cotransfected cells.

Fig. 4.

Pharmacological characterization of SKF83959 on D₁R + D_2LR–mediated calcium mobilization. HEK293T cells were transfected with D₁R + D_2LR as described, plated 24 hours later in 384-well plates, and assayed for calcium accumulation the following day. (A) Cells were stimulated with one of the following conditions as indicated: DA, SKF83959, the D₂R-selective agonist quinpirole, or both SKF83959 and quinpirole combined. (B) Cells were incubated with SKF83959 or the D₁R-selective antagonist SCH23390, then stimulated with an ∼EC₈₀ of DA (1 μM). Data are expressed as a percentage of control maximum DA-stimulated response and are representative of two or three independent experiments performed with the same assay conditions on different days. Error bars indicate S.E.M. from multiple wells within the representative experiment. (C) HEK293 cells stably transfected with D₁R (Codex Biosolutions, Inc., Gaithersburg, MD) were grown and membranes harvested as described in Materials and Methods. Membranes were incubated with various concentrations of SKF83959 and 0.5 nM [³H]SCH23390 as indicated. Graph is representative of two independent experiments done on different days. Data are expressed as specific binding in units of fmol/mg. K_i value was calculated using the Cheng-Prushoff equation and a radioligand K_d value of 0.5 nM as determined via saturation binding isotherms (unpublished data). Average K_i for SKF83959 on D₁R was 2.6 nM ± 0.7.

Given the apparent discrepancies of our findings with some previous studies (Lee et al., 2004; Rashid et al., 2007b; Hasbi et al., 2011) and the possibility that SKF83959 may not be as selective as previously thought, we sought to screen its selectivity against various GPCRs. This was accomplished through collaboration with the NIMH Psychoactive Drug-Screening Program (http://pdsp.med.unc.edu). For the primary screen, a single-point radioligand binding competition experiment was performed with 10 μM SKF83959 as a competitor against an appropriate receptor-specific radioligand of known properties. Forty-three GPCRs and signaling proteins were screened this way, and 20 of them resulted in >50% inhibition at 10 μM SKF83959 (Table 1). In contrast, 23 GPCR targets were found to have <50% inhibition at 10 μM SKF83959 and were therefore considered relatively “inactive/low affinity” for SKF83959 (Supplemental Table 1). The 20 “active” receptors/proteins underwent secondary radioligand competition binding experiments to generate full competition curves for SKF83959 and K_i values for these receptors were determined and are shown in Table 1. Of note is that the serotonin 5-HT2A, 5-HT2B, 5-HT2C, 5-HT5A, and 5-HT6 receptors; the adrenergic α2A, α2B, and α2C receptors; the D₁ D₂, and D₅ DARs; and the serotonin transporter all have nanomolar K_i values. SKF83959 demonstrated very high (sub-100 nM) affinity for four of these GPCRs: the serotonergic receptor subtypes 5-HT2C, the adrenergic receptor subtype α2C, the D₁, and D₅ DAR subtypes, and the serotonin transporter. Notably, SKF83959 has also recently been shown to be a potent allosteric modulator of the σ-1 receptor (Guo et al., 2013). Taken together, these data indicate that SKF83959 has significantly high affinities for a wide number of receptors and thus caution should be taken when interpreting in vivo experimentation and the selectivity of this agent.

TABLE 1.

SKF93959 competition binding experiments against various G protein–coupled receptors

K_i values were derived from radioligand binding competition curves generated against each of the above targets (n = 2) as described in Materials and Methods.

Target	SKF83959 Ki	S.E.M.
	nM
5-HT1A	1648.0	352.3
5-HT2A	246.6	32.1
5-HT2B	405.0	145.1
5-HT2C	32.8	13.3
5-HT5A	277.8	141.8
5-HT6	546.0	56.0
α1A	1290.5	154.5
α1D	1115.5	232.4
α2A	323.7	120.6
α2B	163.1	17.8
α2C	31.1	7.6
D1R	1.7	0.8
D2R	567.0	150.0
D3R	1018.3	109.8
D4R	1975.7	756.4
D5R	4.0	0.1
H2	1699.3	640.3
M4	5238.5	1985.5
M5	3484.0	114.0
SERT	365.6	79.2

α1A, α-adrenergic receptor subtype 1A; α1D, α-adrenergic receptor subtype 1D; α2A, α-adrenergic receptor subtype 2A; α2B, α-adrenergic receptor subtype 2B; α2C, α-adrenergic receptor subtype 2C; 5-HT1A, serotonergic receptor subtype 1A; 5-HT2A, serotonergic receptor subtype 2A; 5-HT2B, serotonergic receptor subtype 2B; 5-HT2C, serotonergic receptor subtype 2C; 5-HT5A, serotonergic receptor subtype 5A; 5-HT6, serotonergic receptor subtype 6; M4, muscarinic receptor subtype 4; M5, muscarinic receptor subtype 5; SERT, serotonin transporter.

Whereas D₁ and D₂ receptors appear capable of signaling through calcium mobilization when both receptors are stimulated, the mechanism of transduction remains unclear. To understand more clearly the mechanisms involved, we tested the hypothesis that the receptors, perhaps within the context of a heteromer, may switch G protein-coupling selectivity and gain the ability to activate G_q. We first examined this possibility by overexpressing G_qα in cells expressing the D₁R + D₂R. Interestingly, the resulting DA-stimulated calcium signal was increased by 200% compared with cells transfected with the D₁R + D₂R alone (Fig. 5A). Expression of only the G_qα protein in the absence of either receptor did not enable the ability of DA to stimulate calcium mobilization (Fig. 5A). In parallel studies, we examined how overexpression of G_qα with the D₁R or D₂R alone could couple to intracellular calcium mobilization. Consistent with Fig. 1, cells transfected with D₁R or D₂R alone did not give a calcium response. However, when G_qα was overexpressed, the D₁R was able to elicit a DA-stimulated calcium signal in the absence of the D₂R (Fig. 5B), although the calcium response was not as large as that seen with the D₁R + D₂R + G_qα transfection (cf. Fig. 5, A and B). No such phenomenon was observed with the D₂R. Taken together, these data suggest that the G_q protein may be involved in calcium mobilization mediated by a D₁-D₂ heteromer, but this interpretation is complicated by the fact that overexpression of G_qα can also lead to monomeric D₁R coupling.

Fig. 5.

Influence of G_qα protein overexpression on D₁R + D_2LR–mediated calcium mobilization. (A) HEK293T cells were transfected with D₁R + D_2LR with and without G_qα or with G_qα alone (D₁R + D₂R EC₅₀ = 168.3 nM, EC_max = 100%; D₁R + D₂R + G_q EC₅₀ = 16.8 nM, EC_max = 300.1%). (B) HEK293T cells were transfected with D₁R + D_2LR, D₁R, or D₂R with and without G_qα (D₁R + G_q EC₅₀ = 10.3 nM, EC_max = 152.2%). Twenty-four hours later, cells were plated in 384-well plates and assayed the following day for calcium mobilization after stimulation by the indicated concentrations of DA. Data are expressed as a percentage of control maximum DA stimulation for D₁R + D_2LR alone and are representative of two or three independent experiments performed with the same assay conditions on different days. Error bars indicate S.E.M. from multiple wells within the representative experiment.

Given our results with G_qα overexpression, we re-evaluated SKF83959 stimulation of calcium mobilization under these conditions in the D₁R and D₂R coexpressed cells. We found that with G_qα overexpression, SKF83959 is able to stimulate calcium mobilization in a manner similar to that of DA (Fig. 6A), whereas it is unable to stimulate such a response in cells lacking G_qα overexpression (Figs. 4 and 6A). Interestingly, SKF83959 was also able to stimulate calcium mobilization in cells expressing the D₁R and overexpressing G_qα, but not D₁R alone (Fig. 6A). These results led us to test the antagonist sensitivity of the SKF83959 responses, as shown in Fig. 6B. We found that the D₁R-selective antagonist SCH23390 could completely ablate SKF83959 stimulation of calcium mobilization in both D₁R + G_qα transfected and D₁R + D₂R + G_qα transfected cells. However, in contrast to what we observed for DA stimulation of D₁R + D₂R cotransfected cells, the D₂R-selective antagonist sulpiride was unable to block SKF83959 stimulation of calcium mobilization. These results suggest that overexpression of G_qα enables SKF83959 to stimulate monomeric D₁R present in the D₁R and D₂R cotransfected cells, rather than enabling it to gain function as a D₁-D₂ heteromeric-selective agonist.

Fig. 6.

SKF83959 stimulates D₁R-dependent calcium mobilization in the presence of G_qα. HEK293T cells were transfected with D₁R + D_2LR, G_qα or with G_qα alone as described, plated 24 hours later in 384-well plates, and assayed for calcium accumulation the following day. (A) Cells were stimulated with SKF83959. The line at 100% denotes the maximal DA response of D₁R + D_2LR cells. (B) Cells were incubated with the D₁R-selective antagonist SCH23390 (1 μM) or the D₂R-selective antagonist sulpiride (1 μM) and then stimulated with an ∼EC₈₀ of SKF83959 (100 nM). Error bars indicate S.E.M. from multiple wells within the representative experiment, which was replicated twice with similar results. DMSO, dimethylsulfoxide.

Although the extant hypothesis, which our overexpression data support, is that G_q is central to the stimulation of calcium mobilization, the central question is whether direct coupling with a D₁-D₂ heteromer may be involved. An alternative hypothesis is that the D₁R and D₂R signal through downstream pathways that converge on the G_q protein or other components of the calcium mobilization process. To test whether D₁-D₂ synergistic signaling is independent of G_i or G_s protein function, we interfered with the activity of G_i and G_s by treatment with toxins. D₁R and D₂R cotransfected cells were incubated overnight in media containing pertussis toxin (PTX) to inhibit G_i protein function (Namkung et al., 2009) or cholera toxin (CTX) to interfere with G_s protein function (Mannoury la Cour et al., 2011). Cells were then assayed for calcium mobilization in response to DA stimulation. We found that treatment with CTX or PTX drastically, but not entirely, reduced the calcium response (Fig. 7). These data support the involvement of D₁R-G_s– and D₂R-G_i–mediated mechanisms that majorly contribute to the calcium response in the D₁R and D₂R cotransfected cells.

Fig. 7.

G protein dependency of D₁R + D_2LR–mediated calcium mobilization. HEK293T cells were transfected with D₁R + D_2LR. Cells were incubated overnight in 1 μg/ml PTX or 1 μg/ml CTX; 48 hours post transfection, cells were assayed for calcium mobilization by stimulation with the indicated concentrations of DA (CTX EC_max = 14%, inhibition = 86% control, PTX EC_max = 24%, inhibition = 76% control). Data are expressed as a percentage of control maximum DA stimulation seen in untreated D₁R + D_2LR cells and are representative of two or three independent experiments performed with the same assay conditions on different days. Error bars indicate S.E.M. from multiple wells within the representative experiment.

Another possibility, however, may be that general G_i-G_q “cross-talk” is occurring after receptor activation, which leads to PLC activation. Multiple cases of G_i-G_q cross-talk in other receptor systems and cell types have been documented (Okajima et al., 1989; Carroll et al., 1995; Toms and Roberts, 1999; Rebres et al., 2011), and G_i-G_q cross-talk in the D₁-D₂ receptor system could account for the PTX sensitivity of the calcium signal. In this model, any G_i-linked GPCR, not just the D₂R, would be able to support a G_q-mediated calcium response. To test this possibility, we used the D₄R, a G_i-linked DAR, which has not been found to form hetero-oligomers with the D₁R (González et al., 2012). We cotransfected the D₁R and D₄R and compared the DA response with that in the D₁R + D₂R transfected cells (Fig. 8). In fact, the D₄R did not support a calcium response in the presence of coexpressed D₁R, indicating that nonspecific G_i-G_q cross-talk, at least as previously described (Okajima et al., 1989; Carroll et al., 1995; Toms and Roberts, 1999; Rebres et al., 2011). does not explain the D₁-D₂ heteromer-mediated calcium response.

Fig. 8.

Dopamine does not elicit a calcium response in cells co-expressing the D₁R and D₄R. HEK293T cells were transiently transfected with D₁R + D_2LR or D₁R + D₄R, as indicated and described in Materials and Methods. Twenty-four hours later, cells were plated in 384-well plates and assayed for calcium mobilization through stimulation by the indicated concentrations of DA. Data are expressed as a percentage of control maximum DA stimulation seen in cells transfected with D₁R + D_2LR only (EC₅₀ = 162.0 nM) and are representative of two or three independent experiments done with the same assay conditions on different days. Expression of the D₄R was confirmed using radioligand binding assays as described in Materials and Methods and was similar to that of the D₂R. Error bars indicate S.E.M. from multiple wells within the representative experiment.

The potential involvement of multiple G_α-proteins led us to also investigate other mechanisms by which D₁R and D₂R activation could stimulate calcium mobilization. Notably, G_βγ subunits have been shown to increase cytoplasmic calcium concentrations by stimulating PLCβ (Beaulieu and Gainetdinov, 2011). A recent publication reported that the ghrelin receptor-D₂R dimer-linked calcium response was PTX sensitive, required PLC activity, and could be ablated by sequestering the G_βγ subunits (Kern et al., 2012). To see whether G_βγ plays a role in the D₁-D₂ heteromer-mediated calcium release, we cotransfected the D₁R and D₂R with two different functionally dominant negative GRK2 mutants. The mutants we used were GRK2 K220R and the GRK2 C-terminal 495–689 peptide fragment (GRK2 c-term), both of which are unable to phosphorylate GPCRs but can bind to and sequester G_βγ subunits (Koch et al., 1994; Freedman et al., 1995). We found that overexpression of GRK2 K220R was able to ablate completely DA-stimulated calcium mobilization in the D₁R and D₂R cotransfected cells (Fig. 9A). Similarly, overexpression of GRK2 c-term drastically reduced, but did not completely ablate, the DA-stimulated calcium response (Fig. 7B). These data suggest that the observed calcium mobilization occurring in response to D₁R and D₂R activation is largely dependent on free G_βγ subunits.

Fig. 9.

GRK2 influence on DA-mediated D₁R + D_2LR calcium mobilization. HEK293T cells were transiently transfected with D₁R + D_2LR and either empty pcDNA vector the GRK2 catalytically inactive mutant GRK2 K220R (A) (D₁R + D₂R EC₅₀ = 269.1 nM) or the GRK2 C-terminal 495–689 fragment (B; GRK2 c-term; D₁R + D₂R EC₅₀ = 90.4 nM, EC_max = 100% control; D₁R+D₂R + pcDNA EC₅₀ = 188.5 nM, EC_max = 106%; D₁R + D₂R +GRK2 c-term EC₅₀ = 288.1 nM, EC_max = 30% control, 70% inhibition), as indicated and described in Materials and Methods. Twenty-four hours later, cells were plated in 384-well plates and assayed the following day for calcium mobilization after stimulation by the indicated concentrations of DA. Data are expressed as a percentage of control maximum DA stimulation seen in cells transfected with D₁R + D_2LR only and are representative of two or three independent experiments done with the same assay conditions on different days. Error bars indicate S.E.M. from multiple wells within the representative experiment.

Discussion

Receptor oligomers of many different GPCR types have been proposed to form homo- or hetero-oligomers with biochemical and functional characteristics that are unique to their oligomeric conformations (Ferre et al., 2009). These GPCR oligomers have been found not only to occur within a type of GPCR but also across different classes, families, types, and subtypes (Prinster et al., 2005). In addition to signaling, internalization and degradation of GPCRs in homo- and hetero-oligomers have been found to differ from their monomeric activities (Milligan, 2004; Terrillon and Bouvier, 2004; Prinster et al., 2005; Ferre et al., 2009; Missale et al., 2010). Like previously described receptor oligomers, it has been shown that the D₁R and D₂R can coimmunoprecipitate with each other (Lee et al., 2004; Pei et al., 2010; Supplemental Fig. 1), and fluorescence imaging has shown that the two receptors cointernalize when one or the other receptor is stimulated (O’Dowd et al., 2005, 2012; So et al., 2005; Dziedzicka-Wasylewska et al., 2006; Łukasiewicz et al., 2009). We have demonstrated that the calcium response is unique to cells that coexpress both D₁ and D₂ DARs and that the DARs must be costimulated, as an antagonist to either receptor blocks the transduction. However, the mechanism of action and whether heteromers or homomers form the functional units for calcium signaling remain unclear.

It has been suggested that the coactivation of the D₁-D₂ complex causes a conformational change that results in the direct interaction between the C terminus of the D₁R and the third intracellular loop (ICL3) of the D₂R (O’Dowd et al., 2012). The ICL3 is the only region of difference between D_2LR and D_2SR, and there is evidence that it results in differences in the G protein coupling and signaling capabilities of each D₂R isoform (Kendall and Senogles, 2011). Recently, it was proposed that the ICL3 of D_2LR, but not the D_2SR, could form a complex with the D₁R (Pei et al., 2010), but the findings were based on the use of glutathione S-transferase and trans-activator of transcription–fused D₂R ICL3 fragments, which may not accurately mimic native receptor conformations and interactions. Later, it was shown that both D₂R splice isoforms were able to cointernalize with the D₁R (O’Dowd et al., 2012). Our results show that both D_2SR and D_2LR can couple with the D₁R to mobilize calcium (Fig. 1), and we have found that this is also true for both human (unpublished data) and rat DARs. We have also confirmed that both receptors must be expressed in the same cell and coactivated to induce a calcium response in HEK293T cells.

Our data also suggest that G_q protein signaling may play a role in the calcium response elicited by the D₁-D₂ complex. This was demonstrated by observing increased calcium mobilization in response to DA in cells transfected with the D₁R and D₂R plus G_qα. However, we also observed that the D₁R alone may couple to G_qα when the α subunit is expressed in significantly high amounts. This is likely due to the D₁R having a relatively low affinity for G_qα; however, it may activate G_q-mediated calcium mobilization under conditions where G_q expression is very high. This is also supported by the enhanced calcium response we observed when the D₁R and D₂R are coexpressed in the presence of high levels of G_q protein, where the D₁R is the protomer within the heteromer that likely activates G_qα (Rashid et al., 2007b). In this model, it is hypothesized that the D₂R allosterically modulates the D₁R (Rashid et al., 2007b; Hasbi et al., 2011). We believe, however, that the enhanced calcium mobilization seen in the D₁R + D₂R + G_qα transfected cells is not due solely to D₁R monomer activation of G_q, as the degree of calcium mobilization (300% of control, Fig. 5A) is twice that seen in the D₁R-G_qα transfected cells (Fig. 5B). Interestingly, another study has also reported D₁R-mediated calcium release from internal stores in mouse cells lacking thymidine kinase transfected with the human D₁R (Liu et al., 1992), indicating that this is not an event particular to our experimental paradigm. Thus, although G_q may play a role in the apparent ability of the D₁-D₂ heteromer to couple to calcium signaling, this role may be dependent on the level of G_q protein expression, either on a total cellular basis, which would thus be cell-type dependent, or this signaling may be localized to specific membrane microdomains (see discussion to follow).

It has also been suggested that SKF83959 may act as a D₁-D₂ heteromer-selective agonist, and it has been used as a putative heteromer-selective probe in vivo. However, these studies are not without controversy, as SKF83959 has a history of unusual pharmacology. Panchalingam and Undie (2001) found that SKF83959 inhibited D₁R-stimulated cAMP formation and also induced striatal intracellular calcium mobilization in rats and monkeys. It lacked the side effects typical to D₁R agonists that stimulate cAMP production but paradoxically seemed to cause typical D₁R agonist-like behaviors in rats (Perreault et al., 2010) and is an effective anti-Parkinsonian agent in MPTP (1-methyl-4-phenyl-1,2,3,6-tetrahydrapyridine)-lesioned monkeys unresponsive to l-Dopa (l-3,4,-dihydroxyphenylaline) (Andringa et al., 1999). In our hands SKF83959 did not stimulate a calcium response in cells transfected with both the D₁R and D₂R, despite the fact that it was active in binding to the D₁R. In fact, it appeared to act as an antagonist of the DA-stimulated calcium response in D₁R and D₂R cotransfected cells. In contrast, when G_qα was overexpressed, SKF83959 stimulated a calcium response in cells cotransfected with the D₁R and G_qα, as well as cells cotransfected with D₁R, D₂R, and G_qα. However, we observed that whereas the D₁R-selective antagonist SCH23390 completely blocked the SKF83959-stimulated calcium response in both transfection conditions, the D₂R-selective antagonist sulpiride was ineffective in the D₁R and D₂R cotransfection condition. This contrasts with sulpiride’s ability to block completely DA-stimulated calcium mobilization in the D₁R and D₂R cotransfected cells (cf. Figs. 2B and 6B). This finding suggests that SKF83959 is not activating the D₁-D₂ heteromer but rather is activating only D₁R monomers that exist in the D₁R and D₂R cotransfected cells. This could be explained by the functionally selective or biased agonist properties of SKF83959 in that it can selectively activate D₁R-G_q signaling, provided there is sufficient G_qα present, but our current results do not support its ability to activate the D₁-D₂ heteromer.

It has also been proposed that D₁-D₂ heteromer activation via SKF83959 in vivo and in vitro results in increased Ca²⁺/calmodulin-dependent protein kinase IIα levels in the striatum and nucleus accumbens, further resulting in enhanced brain-derived neurotrophic factor expression and increased neuronal maturation and differentiation (Rashid et al., 2007a; Hasbi et al., 2009; Ng et al., 2010; Perreault et al., 2012b). Given that our experiments indicated that SKF83959 could not induce D₁-D₂ heteromer–selective calcium mobilization in a controlled cell environment, we conducted a single-point competition-binding screen against an array of 43 GPCRs and additional signaling proteins (Supplemental Table 1; Table 1). We observed that SKF83959 demonstrated considerably high affinity for multiple receptors and other signaling proteins, and we conducted secondary competition binding experiments on the ones for which it showed the highest affinity. Surprisingly, SKF83959 showed nanomolar affinities for many different GPCRs, including several serotonergic, adrenergic, dopaminergic, and muscarinic receptor subtypes (Table 1). This result, as well as our functional data, questions whether SKF83959 may be useful as a selective probe to study D₁-D₂ heteromer or even D1-like receptor signaling in vivo.

Our data also suggest that calcium signaling through the D₁-D₂ receptor complex is largely sensitive to G_i and G_s inhibition by PTX and CTX, respectively. This led us to investigate additional hypotheses for the mechanism of D₁-D₂ calcium signaling. Recently, Kern et al. (2012) showed that the ghrelin receptor could hetero-oligomerize with the D₂R. This heteromer induced calcium release from internal cellular stores in a PLC-dependent and PTX-sensitive manner and seemed to require G_βγ subunit activation. Previous studies have shown that GRK2 can bind to and sequester G_βγ subunits (Koch et al., 1994), and catalytically inactive GRK2 mutants that retain G_βγ binding have been used as tools to block G_βγ signaling without the complication of added receptor desensitization (Koch et al., 1994; Freedman et al., 1995). Our data demonstrated that the catalytically inactive GRK2 K220R mutant completely ablated the DA-stimulated calcium response in the D₁R and D₂R transfected cells, whereas GRK2 c-term (a truncated GRK2 protein that includes only the G_βγ binding domain) largely decreased the calcium response. Since activated G_βγ subunits can stimulate PLCβ activity (Camps et al., 1992), our results are consistent with the hypothesis that the DA-stimulated calcium response significantly involves G_βγ activation of PLCβ. Additionally, the N-terminal RGS domain of GRK2 has been shown to facilitate weak GTPase-activating protein-like activity on G_q, inhibiting PLC activation. This may explain the difference in degree of calcium signal inhibition between the GRK2 K220R mutant and the truncated GRK2 c-term mutant (Carman et al., 1999). Therefore, the activation of PLCβ may be G_qα- as well as G_βγ-dependent and due largely to synergistic cross-talk between the D₁R and D₂R.

Figure 10 represents several hypothetical signaling pathways for D₁-D₂ receptor-calcium signaling in HEK293T cells. Pathway A represents D₁-D₂ heterodimer activation of G_q leading to G_qα activation of PLCβ, as has been hypothesized in the literature (Rashid et al., 2007b). Pathway B represents G_βγ activation of PLCβ, where free β/γ subunits could arise through activation of either G_s, G_i, or G_q. Pathway C represents coactivation of D₁R and D₂R monomers and cross-talk between G_s and G_i protein–mediated downstream signaling pathways, ultimately leading to PLCβ activation. Given that PTX and CTX can nearly eliminate the DA-stimulated calcium signaling, we believe that pathway A is largely inoperative in our system under basal conditions. Pathway C could readily account for the requirement for dual receptor activation, but the fact that G_βγ sequestration largely eliminates the DA calcium response suggests that pathway B is critically important. The PTX/CTX results further implicate G_s or G_i; however, the requirement for dual receptor activation in pathway B is not completely clear. Certainly, additional work is required to answer these questions, but it is clear from these studies that D₁-D₂ receptors can dually activate calcium signaling through more than a single mechanism.

Fig. 10.

Various mechanisms of PLCβ activation that may occur when the D₁R and D₂R are coexpressed and coactivated.

One additional consideration for D₁-D₂-calcium signaling, which does not necessarily exclude the possibility of heteromer formation, may involve the aggregation of the two DARs and their associated proteins in lipid rafts. Lipid rafts are a well-known but poorly understood platform for modulating certain protein-protein interactions in neurons as well as affecting GPCR ligand sensitivity, membrane trafficking, and signaling (Allen et al., 2007; Korade and Kenworthy, 2008; Björk and Svenningsson, 2011; Kong et al., 2011; Sebastião et al., 2011; Celver et al., 2012). Lipid rafts would readily enable cross-talk between the D₁R and D₂R and could assist in the multifaceted signaling profile of the D₁-D₂ receptor complex. In addition, differences in lipid raft composition, cell background, and assay detection may explain some of the differences observed between our data and the data generated by other groups. Despite the seeming complexity of the D₁-D₂ receptor signaling mechanisms, it may yet be useful to study how synergistic concurrent activation of the D₁R and D₂R may induce effects not seen when either receptor is expressed alone. This can be examined by coexpressing mutants of the the D₁R and D₂R, which have been reported to be unable to form dimers (O’Dowd et al., 2012), and studying the effect of coactivation on the generation of a calcium signal. Additionally, a compound that can selectively bias both receptors toward a conformation that promotes PLC activation may be useful in providing a clearer understanding of the DAR system in vivo.

Abbreviations

CTX

cholera toxin

D₁R

D₁ dopamine receptor subtype

D₂R

D₂ dopamine receptor subtype

D_2LR

D₂R long splice variant

D_2SR

D₂R short splice variant

DA

dopamine

DAR

dopamine receptor

FDSS

Functional Drug Screening System

GRK2

G protein receptor kinase 2

GPCR

G protein–coupled receptor

HEK293T

human embryonic kidney cells 293-tsa201

ICL3

third intercellular loop

L-Dopa

l-3,4,-dihydroxyphenylaline

MPTP

1-methyl-4-phenyl-1,2,3,6-tetrahydrapyridine

NIMH

National Institute of Mental Health

PDSP

Psychoactive Drug Screening Program

PLC

phospholipase C

PTX

pertussis toxin

SCH23390

(R)-(+)-7-chloro-8-hydroxy-3-methyl-1-phenyl-2,3,4,5-tetrahydro-1H-3-benzazepine

SKF83822

6-chloro-2,3,4,5-tetrahydro-1-(3-methylphenyl)-3-(2-propenyl)-1H-3-benzazepine-7,8-diol

SKF83959

6-chloro-2,3,4,5-tetrahydro-3-methyl-1-(3-methylphenyl)-1H-3-benzazepine-7,8-diol

Authorship Contributions

Participated in research design: Chun, Free, Doyle, Sibley, Rankin, Huang.

Conducted experiments: Chun, Doyle, Rankin, Huang.

Contributed new reagents or analytic tools: Free.

Performed data analysis: Chun, Free, Doyle, Rankin, Huang.

Wrote or contributed to the writing of the manuscript: Chun, Free, Sibley, Rankin, Huang.