ACCELERATED COMMUNICATION Ligand-Dependent Oligomerization of Dopamine D2 and Adenosine A2A Receptors in Living Neuronal Cells

Adenosine A2A and dopamine D2 receptors (A2A and D2) associate in homoand heteromeric complexes in the striatum, providing a structural basis for their mutual antagonism. At the cellular level, the portion of receptors engaging in homoand heteromers, as well as the effect of persistent receptor activation or antagonism on the cell oligomer repertoire, are largely unknown. We have used bimolecular fluorescence complementation (BiFC) to visualize A2A and D2 oligomerization in the Cath.a differentiated neuronal cell model. Receptor fusions to BiFC fluorescent protein fragments retained their function when expressed alone or in A2A/A2A, D2/D2, and A2A/D2 BiFC pairs. Robust fluorescence complementation reflecting A2A/D2 heteromers was detected at the cell membrane as well as in endosomes. In contrast, weaker BiFC signals, largely confined to intracellular domains, were detected with A2A/dopamine D1 BiFC pairs. Multicolor BiFC was used to simultaneously visualize A2A and D2 homoand heteromers in living cells and to examine drug-induced changes in receptor oligomers. Prolonged D2 stimulation with quinpirole lead to the internalization of D2/D2 and A2A/D2 oligomers and resulted in decreased A2A/D2 relative to A2A/A2A oligomer formation. Opposing effects were observed in cells treated with D2 antagonists or with the A2A agonist 5 -N-methylcarboxamidoadenosine (MECA). Subsequent radioreceptor binding analysis indicated that the drug-induced changes in oligomer formation were not readily explained by alterations in receptor density. These observations support the hypothesis that long-term drug exposure differentially alters A2A/D2 receptor oligomerization and provide the first demonstration for the use of BiFC to monitor drugmodulated GPCR oligomerization. A growing number of G protein-coupled receptors (GPCRs) have been shown to exist as oligomers with unique functional properties and physiological relevance (Pin et al., 2007). Evidence suggests that A2A and D2 form receptor heteromers. Both receptors are highly expressed in the striatum, where they colocalize on spiny neurons (Fink et al., 1992). The receptors have opposing actions on adenylyl cyclase activity, through coupling to G s (A2A) or G i/o (D2) proteins. Biochemical and behavioral evidence also indicates antagonistic A2A/D2 interactions (Ferre et al., 1991; Agnati et al., 2003; Fuxe et al., 2007). Moreover, persistent D2 activation sensitizes A2A receptor-stimulated cAMP accumulation (Vortherms and Watts, 2004). A2A and D2 have been shown to oligomerize in resonance energy transfer as well as coimmunoprecipitation experiments (Hillion et al., 2002; Canals et al., 2003; Kamiya et al., 2003). Therefore, a direct A2A/D2 interaction may account for the antagonism between the two receptors. In addition to forming heteromers, A2A and D2 also exist as homomers (Lee et al., 2000; Armstrong and Strange, 2001; Gazi et al., 2003; Canals et al., 2004; Guo et al., 2005). The stoichiometry of A2A and D2 in A2A/D2 heteromers is unknown, as is the relative proportion of A2A and D2 recepThis project was funded by Purdue University and by National Institute of Mental Health grant MH060397 (to V.J.W.). Article, publication date, and citation information can be found at http://molpharm.aspetjournals.org. doi:10.1124/mol.108.047472. □S The online version of this article (available at http://molpharm. aspetjournals.org) contains supplemental material. ABBREVIATIONS: GPCR, G protein-coupled receptor; BiFC, bimolecular fluorescence complementation; CAD, Cath.a differentiated; HA, hemagglutinin; YFP, yellow fluorescent protein; V, Venus; C, Cerulean; VN, Venus N-terminal fragment; CN, Cerulean N-terminal fragment; VC, Venus C-terminal fragment; CC, Cerulean C-terminal fragment; D2L, long isoform of the dopamine D2 receptor; MECA, 5 -N-Methylcarboxamidoadenosine; CGS15943, 9-chloro-2-(2-furyl)(1,2,4)triazolo(1,5-c)quinazolin-5-amine; Ro 20-1724, 4-(3-butoxy-4-methoxybenzyl)imidazolidin-2one; ZM 241-385, 4-(2[7-amino-2-(2-furyl)[1,2,4]triazolo[2,3-a][1,3,5]triazin-5-ylamino]ethyl)-phenol; ANOVA, analysis of variance; HEK, human embryonic kidney; ER, endoplasmic reticulum. 0026-895X/08/7403-544–551$20.00 MOLECULAR PHARMACOLOGY Vol. 74, No. 3 Copyright © 2008 The American Society for Pharmacology and Experimental Therapeutics 47472/3372215 Mol Pharmacol 74:544–551, 2008 Printed in U.S.A. 544 http://molpharm.aspetjournals.org/content/suppl/2008/06/04/mol.108.047472.DC1 Supplemental material to this article can be found at: at A PE T Jornals on A ril 2, 2017 m oharm .aspeurnals.org D ow nladed from tors engaging in heteroor in homomers (or existing as monomeric receptors). Although A2A and D2 homoand heteromerization was shown to be constitutive and was not affected by acute receptor activation (Canals et al., 2003; Gazi et al., 2003; Canals et al., 2004), the effect of persistent receptor activation or antagonism on the relative homo-/heteromer population has not been investigated. Bimolecular fluorescence complementation (BiFC) is an emerging technique to monitor protein-protein interactions (Hu et al., 2002; Shyu et al., 2006). Whereas most currently available techniques are restricted to the detection of two interacting proteins, multicolor BiFC (i.e., the reconstitution of distinct spectral GFP variants) allows the simultaneous detection of two distinct protein-protein interactions in living cells (Hu and Kerppola, 2003). We have applied multicolor BiFC to simultaneously visualize A2A/D2 heteromers and A2A homomers in the Cath.a differentiated (CAD) neuronal cell model (Qi et al., 1997). The results indicate that A2A/D2 heteromers coexist and colocalize with A2A homomers. Prolonged (18-h) treatment with the selective D2 agonist quinpirole or the D2 antagonist sulpiride had opposing effects on the proportion of A2A/D2 heteromers relative to A2A homomers. These observations have clinical implications in the management of Parkinson’s disease and schizophrenia, which rely on long-term treatment with drugs targeting dopamine receptors. Materials and Methods Materials. D2L, A2A, and D1 cDNAs were obtained from the Missouri S&T cDNA Resource Center. Growth media and reagents (unless otherwise stated) were purchased from Sigma-Aldrich (St. Louis, MO). [H]cAMP (25 Ci/mmol) was purchased from PerkinElmer Life and Analytical Sciences (Waltham, MA). [H]Spiperone (85 Ci/mmol) was from GE Healthcare (Chalfont St. Giles, Buckinghamshire, UK). Cell Culture. CAD cells were maintained as described previously (Vortherms and Watts, 2004). Expression Vectors. Full-length human D2L, A2A, or D1 cDNAs were amplified by polymerase chain reaction using oligonucleotides incorporating EcoRI and XbaI or XhoI restriction sites and omitting stop codons. Polymerase chain reaction fragments digested with EcoRI/XbaI or EcoRI/XhoI were ligated into the corresponding sites from pBiFC vectors (Shyu et al., 2006). These vectors contain fragments from the yellow Venus [V (Nagai et al., 2002)] or the cyan Cerulean [C (Rizzo et al., 2004)] enhanced fluorescent proteins. NTerminal fragments (VN or CN) include residues 1 to 172, whereas C-terminal fragments (VC or CC) include residues 155 to 238. Cloning into pBiFC vectors incorporates MYC (pBiFC-VN), HA (pBiFC-VC and pBiFC-CC), or FLAG (pBiFC-CN) N-terminal epitope tags to the fusion proteins to ease their detection. Receptor fusions to Venus or Cerulean were obtained by swapping BiFC fragments with Venus or Cerulean coding sequences. Constructs were verified by DNA sequencing. Imaging and Image Analysis. CAD cells were grown to 70% confluence in four-well Lab-Tek coverslips (Nalge Nunc International, Rochester, NY) and transfected using 1 l/well Lipofectamine 2000 (Invitrogen, Carlsbad, CA), according to the manufacturer’s recommendations. DNA amounts per well were 500 ng (D2L and D1 constructs), 200 ng (A2A-VN), 100 ng (A2A-CC, A2A-CN), or 20 ng (mCherry-Mem, YFP-Endo, YFP-Golgi, and YFP-ER). Twenty-four hours after transfection, the growth media was replaced with phosphate-buffered saline, and images were captured using a chargecoupled device camera mounted on a TE2000-U inverted fluorescence microscope (Nikon Instruments Inc., Melville, NY) equipped with a 100-W mercury lamp and band-pass filters (Chroma, Rockingham, VT) for Venus (excitation at 500/20 nm; emission at 535/30 nm), Cerulean (excitation at 430/25 nm; emission at 470/30 nm), or mCherry (excitation at 572 nm/23 nm). Fluorescent images were acquired using the MetaMorph software (Molecular Devices, Sunnyvale, CA) and AutoDeblur (MediaCybernetics, Bethesda, MD) was used for three-dimensional deconvolution. Blind selection and analysis of the cells avoided experimental bias. Quantification of BiFC signals was performed as described previously (Hu et al., 2002), using the ImageJ software (http://rsb.info.nih.gov/ij/). Stacks of fluorescent images were analyzed as follows. Background fluorescence intensities were determined by measuring areas devoid of cells and were subtracted from each pixel intensity measurement. After background removal, pixel intensities were scaled by a factor equal to the inverse of the exposure time. Images from the mCherry-Mem membrane marker were used to select cells for analysis and to normalize BiFC signals. As an approximation of plasma membrane signals, maximal pixel intensities along lines traced across plasma membranes were measured. Intracellular signals were measured by tracing regions of interest and determining average pixel intensities. Cells with saturated signals, as well as cells with signals lower than 1.5 times background values, were not considered for analysis. Because Venus/mCherry fluorescence ratios exhibited non-Gaussian distributions, median values were calculated and averaged between different experiments. In multicolor BiFC experiments, median Venus/Cerulean fluorescence ratios were measured. For each condition, approximately 40 cells were analyzed. Median values from at least three independent experiments were averaged and used for statisti-