METABOTROPIC GLUTAMATE TYPE 5, DOPAMINE D₂ AND ADENOSINE A_2A RECEPTORS FORM HIGHER-ORDER OLIGOMERS IN LIVING CELLS

Abstract

G protein-coupled receptors are known to form homo- and heteromers at the plasma membrane, but the stoichiometry of these receptor oligomers are relatively unknown. Here, by using bimolecular fluorescence complementation, we visualized for the first time the occurrence of heterodimers of metabotropic glutamate mGlu₅ receptors (mGlu₅R) and dopamine D₂ receptors (D₂R) in living cells. Furthermore, the combination of bimolecular fluorescence complementation and bioluminescence resonance energy transfer techniques, as well as the sequential resonance energy transfer (SRET) technique, allowed us to detect the occurrence receptor oligomers containing more than two protomers, mGlu₅R, D₂R and adenosine A_2A receptor (A_2AR). Interestingly, by using high-resolution immunoelectron microscopy we could confirm that the three receptors co-distribute within the extrasynaptic plasma membrane of the same dendritic spines of asymmetrical, putative glutamatergic, striatal synapses. Also, co-immunoprecipitation experiments in native tissue demonstrated the existence of an association of mGlu₅R, D₂R and A_2AR in rat striatum homogenates. Overall, these results provide new insights into the molecular composition of G protein-coupled receptor oligomers in general and the mGlu₅R/D₂R/A_2AR oligomer in particular, a receptor oligomer that might constitute an important target for the treatment of some neuropsychiatric disorders.

Introduction

The striatum receives the densest dopamine innervation and contains the highest concentration of dopamine receptors in the brain (Gerfen, 2004; Lindvall and Bjorklund, 1978). As a consequence, dopamine plays a key role in motor activity and goal-directed behaviors and also in the pathophysiology of diverse disorders, including Parkinson’s disease and drug addiction (Ferre et al., 2004). Dopamine acts on G-protein-coupled receptors (GPCRs) that, based on sequence homologies and pharmacological profiles, have been classified in D₁-like, which include D₁ and D₅ receptors, and D₂-like receptors, which comprise the long and short isoforms of the D₂ receptor (termed D_2L and D_2S, respectively), the D₃ and the D₄ receptors (Missale et al., 1998). The striatum also receives a dense glutamatergic innervation from cortical, thalamic and limbic areas (Gerfen, 2004). Several classes of glutamate receptors, widely distributed in the CNS, have been characterized: three subtypes of ionotropic glutamate (iGlu) receptors and a family of GPCRs, called metabotropic glutamate (mGlu) receptors, which act through different second messenger pathways. Eight members of the mGlu receptors family have been identified and categorized into three subgroups on the basis of their sequence homology, agonist selectivity and signal transduction pathway. Group I contains mGlu₁ and mGlu₅ receptor subtypes, which are coupled to phospholipase C in transfected cells and have quisqualic acid (Quis) as the most potent agonist.

In addition to dopamine and glutamate, adenosine plays a key role in basal ganglia function (Ferre et al., 1997). Adenosine seems to be mainly produced by metabolism of ATP co-released with glutamate from nerve terminals and astrocytes (Ferre et al., 2007) and mediates its actions by specific GPCRs, which are currently classified in A₁, A_2A, A_2B, and A₃ subtypes (Fredholm et al., 2001). Compared to the other adenosine receptor subtypes, A_2A receptors (A_2ARs) are particularly concentrated in the striatum (Ferre et al., 1997; Rosin et al., 1998), where they are mostly expressed by γ-aminobutyric acid (GABA)ergic striatopallidal neurons (Schiffmann et al., 1991; Schiffmann et al., 2007). Interestingly, ultrastructural analysis has demonstrated that A_2ARs are localized mostly postsynaptically in the dendrites and dendritic spines of striatal GABAergic neurons (Hettinger et al., 2001). A_2AR immunoreactivity was observed primarily at glutamatergic (asymmetric) synapses, predominantly at the postsynaptic but also at the presynaptic site (Ciruela et al., 2006a; Hettinger et al., 2001). Therefore, it has been suggested that A_2AR plays a key role in the fine-tuning modulation of glutamatergic neurotransmission in striatal GABAergic neurons both at the postsynaptic and presynaptic level (Ciruela et al., 2006a; Ciruela et al., 2006b; Ferre et al., 2007; Ferre et al., 2007; Hettinger et al., 2001).

In the GABAergic striatopallidal neurons A_2AR are co-localized with D₂ receptors (D₂R), and establish functional A_2AR-D₂R heteromers (Ferre et al., 2007; Ferre et al., 2007) and, in the striatopallidal complex in primates, mGlu₅ receptor (mGlu₅R) showed a very similar localization to that described for A_2AR and D₂R in rats (Paquet and Smith, 2003). Furthermore, evidence has been provided for the existence of A_2AR-mGlu₅R heteromerization in transfected cells and in rat striatum (Ferre et al., 2002). Significantly, multiple interactions between striatal A_2AR, D₂R and mGlu₅R were described at the biochemical and behavioural level (Ferre et al., 2007; Ferre et al., 2007; Ferre et al., 2008; Fuxe et al., 2003; Popoli et al., 2001), suggesting the possible existence of higher-order mGlu₅R-D₂R-A_2AR oligomers. In the present study, for the first time, we provide evidence for the existence of mGlu₅R/D₂R/A_2AR oligomers in HEK-293 cells by two different experimental approaches, namely the combined use of bimolecular fluorescence complementation (BiFC) and bioluminescence energy transfer (BRET) (Gandia et al., 2008a) and by sequential resonance energy transfer (SRET) (Carriba et al., 2008). The existence of these higher-order receptor oligomeric complexes might be relevant to striatal function both in normal and pathological conditions (Ferre et al., 2003).

Materials and methods

Plasmid constructs

The cDNA encoding the mGlu_5b receptor was subcloned into the EcoRI/BamHI sites of pGFP²-N3 (Perkin-Elmer, Waltham, MA, USA) and pEYFP-N1 (Clontech, Mountain View, CA, USA) vector as previously described (Cabello et al., 2007). Also, the cDNA encoding dopamine D₂ receptor was amplified by PCR using the following primers: FD2Nhe (5′-CCGCGCTAGCATGGATCCACTGAATCTGTCC-3′) and RD2Xho (5′-CCGCTCGAGCCGCAGTGGAGGATCTTCAGG -3′) and cloned into the NheI/XhoI sites of pRluc-N1 vector (Perkin-Elmer). The other constructs used were previously described (Canals et al., 2003). For the bimolecular complementation experiments a C-terminal truncated version of YFP, named N-YFP (amino acids 1 to 155), was made by PCR amplification and cloning into the XhoI site of pcDNA3.1 using the following primers: FnYFP (5′-CCGCTCGAGACCATGGTGAGCAAGGGCGAGGAGC-3′) and RnYFP (5′-CCGTCTAGATCAGGCCATGATATAGACGTTG-3′). Also, an N-terminal truncated version of YFP, named C-YFP (amino acids 155 to 231), was made using the same strategy and using the following primers: FcYFP (5′-CCGCTCGAGACCATGGACAAGCAGAAGAACGGC-3′) and RcYFP (5′-CCGTCTAGATTACTTGTACAGCTCGTCCAT-3′). The cDNAs encoding the D₂R, mGlu_5bR, GABA_B1bR and GABA_B2R without their stop codon were subcloned into the NheI/EcoRI or KpnI/EcoRI sites of pcDNA3.1 vector (Invitrogen, Carlsbad, CA, U.S.A.), thus containing in frame the sequences for N-YFP or C-YFP. Finally, the cDNA encoding A_2AR and GABA_B2R without their stop codon were amplified using sense and antisense primers harboring unique restriction enzyme sites and subcloned to be in-frame with Rluc in the pRluc-N1 vector or with EYFP in the pEYFP-N1 vector. Finally, besides the YFP constructs contained the ATG initiation codon when these were compared to the same constructs lacking the initiation codon neither apparent difference in molecular size nor in the subcellular distribution was observed, thus suggesting that no apparent translation took place from this second initiation codon (data not shown).

Antibodies

The primary antibodies used were: rabbit anti-A_2AR whole serum (Ciruela et al., 2004), rabbit anti-A_2AR polyclonal antibody (Alomone Labs, Jerusalem, Israel), mouse anti-A_2AR monoclonal antibody (clone 7F6-G5-A2; Millipore, Billerica, MA, USA), rabbit anti-D₂R whole serum (Bjelke et al., 1996; Canals et al., 2003; Jansson et al., 1999; Levey et al., 1993), rabbit anti-D₂R polyclonal antibody (Millipore), goat anti-D₂R polyclonal antibody (Santa Cruz, CA, USA), rabbit anti-D₂R polyclonal antibody (Narushima et al., 2006), rabbit anti-mGlu₅R polyclonal antibody (Millipore), guinea pig anti-mGlu₅R polyclonal antibody (Uchigashima et al., 2007b), mouse anti-GFP monoclonal antibody (Sigma-Aldrich, St. Louis, MO, USA). The secondary antibodies were: horseradish-peroxidase (HRP)-conjugated goat anti-rabbit IgG (Pierce, Rockford, IL, U.S.A.), HRP-conjugated rabbit anti-mouse IgG (Dako, Denmark), HRP-conjugated anti-rabbit IgG TrueBlot (eBioscience, San Diego, CA, USA), donkey anti-rabbit-Cy5^™ (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA), donkey anti-mouse-Cy3^™ (Jackson ImmunoResearch Laboratories, Inc.), donkey anti-goat-Cy2^™ (Jackson ImmunoResearch Laboratories, Inc.), goat anti-mouse IgG conjugated to 10 nm colloidal gold (1/100; Aurion, Wageningen, The Netherlands), goat anti-rabbit IgG conjugated to 15 nm colloidal gold (1/100; Aurion) and goat anti-guinea pig IgG conjugated to 20 nm colloidal gold (1/100; Aurion).

Cell culture, transfection and membrane preparation

Human embryonic kidney (HEK-293) cells were grown in Dulbecco’s modified Eagle’s medium, DMEM (Sigma-Aldrich) supplemented with 1 mM sodium pyruvate, 2 mM L-glutamine, 100 U/ml streptomycin, 100 μg/ml penicillin and 10% (v/v) foetal bovine serum (FBS) at 37°C and in an atmosphere of 5% CO2. HEK-293 cells growing in 25 cm² flasks or in 6-well plates, containing 20 mm coverslips when used for confocal microscopy, were transiently transfected with the DNA encoding the specified proteins by the calcium phosphate precipitation technique (Jordan et al., 1996). The cells were harvested at either 24 or 48 hours after transfection. Membrane suspensions from transfected HEK-293 cells or from rat striatum were obtained as described previously (Burgueno et al., 2003).

Immunoprecipitation and Immunostaining

For immunoprecipitation membranes from transiently transfected HEK cells or from rat striatum were solubilized in ice-cold RIPA buffer (50 mM Tris·HCl pH 7.4, 100 mM NaCl, 1% Triton-X100, 0.5% sodium deoxycholate, 0.2% SDS and 1 mM EDTA) for 30 min on ice. The solubilized preparation was then centrifuged at 13.000xg for 30 min. The supernatant (1mg/ml) was processed for immunoprecipitation, each step of which was conducted with constant rotation at 0–4 °C. The supernatant was incubated overnight with the indicated antibody. Then 40 μl of either a suspension of protein G cross linked to agarose beads or TrueBlot^™ anti-Rabbit Ig IP Beads (eBioscience) were added and the mixture was incubated overnight. Subsequently, the beads were washed twice with ice-cold RIPA buffer, twice with ice cold RIPA buffer diluted 1:10 in TBS (50 mM Tris·HCl pH 7.4, 100 mM NaCl) and once with TBS and aspirated to dryness with a 28-gauge needle. Subsequently, 30μl of SDS-PAGE sample buffer (8M Urea, 2% SDS, 100mM DTT, 375mM Tris, pH 6.8) was added to each sample. Immune complexes were dissociated by heating to 37°C for 2 h and resolved by SDS-polyacrylamide gel electrophoresis in 6.5% or 10% gels and immunoblotted as described above.

For immunocytochemistry transiently transfected HEK-293 cells were fixed in 4% paraformaldehyde for 15 min, and washed with phosphate buffered saline (PBS) containing 20 mM glycine (buffer A) to quench the remaining free aldehyde groups. Cells were permeabilized with buffer A containing 0.2% Triton X-100 for 5 minutes. Blocking was done using buffer A containing 1% BSA (buffer B). Cells were labeled for 1 h at room temperature with the indicated primary antibody, washed for 30 min in buffer B and stained with the corresponding secondary antibodies for another hour. Samples were rinsed and observed in a confocal microscope (Lujan and Ciruela, 2001; Sarrio et al., 2000).

For electron microscopy, ultrathin sections (70–90 nm) from three Lowicryl-embedded blocks were incubated for 45 min on pioloform-coated nickel grids with drops of blocking solution consisting of 2 % albumin in 0.05 M TBS, 0.9 % NaCl and 0.03 % Triton X-100. The grids were transferred to a mixture solution A2AR, D2R and mgluR5 antibodies at a final protein concentration of 10 μg/mL diluted in blocking solution overnight at room temperature. After several washes in TBS, grids were incubated for 2 h in drops of a mixture of goat anti-mouse IgG conjugated to 10 nm-colloidal gold particles, goat anti-rabbit IgG conjugated to 15 nm-colloidal gold particles and goat anti-guinea pig IgG conjugated to 20 nm-colloidal gold particles (Aurion, Wageningen, The Netherlands), each diluted 1:100 in a 0.05 M TBS solution containing 2 % normal human serum and 0.5 % polyethylene glycol. Grids were then washed in TBS for 30 min and counterstained for electron microscopy with saturated aqueous uranyl acetate and lead citrate. Ultrastructural analyses were performed in a Jeol-1010 electron microscope. Electron photomicrographs were captured with CCD camera (Mega View III; Soft Imaging System, Germany). Digitized electron images were then modified for brightness and contrast by using Adobe PhotoShop CS1 (Mountain View, CA) to optimize them for printing.

Gel electrophoresis and immunoblotting

Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS/PAGE) was performed using 7.5 or 10 % polyacrylamide gels. Proteins were transferred to PVDF membranes using a semi-dry transfer system and immunoblotted with the indicated antibody and then HRP-conjugated goat anti-rabbit IgG (1/30000), HRP-conjugated rabbit anti-mouse IgG (1/5000) or HRP-conjugated anti-rabbit IgG TrueBlot (1/1000). The immunoreactive bands were developed using a chemiluminescent detection kit (Pierce) (Ciruela and McIlhinney, 1997).

Resonance energy transfer-based methods

For bioluminescence resonance energy transfer (BRET) experiments HEK cells transiently transfected with a constant amount of cDNA encoding the D₂R^Rluc and increasing amounts of A_2AR^YFP or CD₄R^YFP were detached, washed and resuspended in HBSS buffer containing 10 mM glucose. To maintain the same ratio of DNA in cotransfections, we used empty vector, pcDNA3.1, to equilibrate the amount of total DNA transfected. To control for cell number, protein concentration in the samples was determined using a Bradford assay kit (Bio-Rad, Hercules, CA, USA) using bovine serum albumin dilutions as standards. Cell suspension (20 μg protein) was distributed in duplicate into 96-well microplates (black plates with a transparent bottom for fluorescence measurement or white plates with white bottom for BRET determination). Fluorescence and bioluminescence readings were collected using a Mithras LB 940 (Berthold Technologies, DLReady, Germany) that allows the integration of the signals detected in the filter at 485 nm (440–500 nm, maximum in bioluminescence emission) and 530 nm (510–590 nm, maximum in YFP emission). YFP fluorescence was defined as the fluorescence of the sample minus the fluorescence of cells expressing only Rluc-tagged receptor. For BRET measurement, 5 μM h-coelenterazine (Molecular Probes, Eugene, OR, USA) was added to the samples and readings were performed after 1 min (net BRET determination) and 10 min (Rluc luminiscence quantification). BRET signal was determined by calculating the ratio of the light emitted by YFP (510–590 nm) over the light emitted by the Rluc (440–500 nm). The net BRET values were obtained by subtracting the BRET background signal detected when Rluc-tagged construct was expressed alone. Curves were fitted using nonlinear regression and one-phase exponential association fit equation (GraphPad Prism, San Diego, CA, USA).

Displacement studies were performed at a constant BRET ratio, around the BRET₅₀ of the A_2AR^Rluc/D₂R^YFP pair, and increasing amounts of mGlu₅R or CD₄R constructs. The BRET values were expressed as a percentage of the BRET₅₀ of the mGlu₅R nontransfected cells and better fitted to a two site competition curve using nonlinear regression (GraphPad Prism). The combined BRET-bimolecular fluorescence complementation (BiFC) assay was performed as described recently (Gandia et al., 2008a). Briefly, cells transfected with indicated plasmids were placed in DMEM medium for 24 hours at 37oC, followed by another 24 hours at 30oC in order to allow proper the maturation of complemented fluorophore (Schmidt et al., 2006). After transfection cells were processed for BRET measurements as described above.

For Sequential Ressonance Energy Transfer (SRET) experiments HEK-293 cells were transiently co-transfected with plasmid encoding D_2LR^Rluc (2 μg), mGlu₅R^GFP2 (4 μg) and A_2AR^YFP (6 μg for a single SRET experiment or increasing amounts for SRET curves), as described previously (Carriba et al., 2008). Fluorescence and luminescence were determined as described above. For SRET measurements cells were distributed in 96-white well microplates and 5 μM DeepBlueC (Molecular Probes) was added. SRET signal was collected in a Mithras LB 940 using the detected filters at short-wavelength at 485 nm (440–500 nm) and at long-wavelength at 530 nm (510–590 nm). Here, we express net SRET as [(long-wavelength emission)/(short-wavelength emission)]-Cf where Cf corresponds to [(long-wavelength emission)/(short-wavelength emission)] for cells expressing D_2LR^Rluc, mGlu₅R^GFP2 (Carriba et al., 2008). Curves were fitted as described above.

Results

Investigation of mGlu₅R, A_2AR ad D₂R oligomerization by classical biochemical approaches

The association of the D₂R with mGlu₅R and A_2AR was first studied by classical biochemical approaches, namely immunolabeling experiments and coimmunoprecipitation assays. Thus, by means of confocal microscopy analysis of HEK-293 cells transiently transfected with the cDNAs encoding for the D₂R, A_2AR and mGlu₅R some overlapping in the distribution of the three receptors was found (Fig. 1a). Interestingly, it is important to mention here that around 70% of the cells were simultaneously expressing all three receptors. Next, the existence of mGlu₅R/A_2AR/D₂R oligomers was subsequently assayed by co-immunoprecipitation experiments. From extracts of HEK-293 cells transfected with mGlu₅R^GFP2 (Fig. 1b, lane 1), D₂R (Fig. 1b, lane 2), A_2AR (Fig. 1b, lane 3) or with A_2AR plus D₂R plus mGlu₅R^GFP2 (Fig. 1b, lane 4) the mouse monoclonal antibody against GFP, which recognizes all GFP variants, immunoprecipitated a band of ~160 kDa, which corresponds to the mGlu₅R^GFP2 (Fig. 1b, IP: GFP, lanes 1 and 4, lower panel). This band did not appear in immunoprecipitates from cells transfected only with the cDNA encoding for D₂R (Fig. 1b, lane 2, lower panel) or A_2AR (Fig. 1b, lane 3, lower panel) or when an irrelevant antibody was used (data not shown). Interestingly, when immunoprecipitates from the triple-transfected cells were analyzed by immunoblot using an antibody against the A_2AR, a band of ~42 kDa, which corresponds to the A_2AR, was observed (Fig. 1b, IP:GFP, lane 4, upper panel), and similarly, when these immunoprecipitates were analyze by western-blot using an antibody against the D₂R, a band of ~90 kDa, which corresponds to the D₂R, was observed (Fig. 1b, IP:GFP, lane 4, middle ₂ panel). Importantly, similar results were obtained when A_2AR and D₂R were tagged with GFP and analogous co-immunoprecipitation experiments were performed (Fig. 1c and d). Overall, these results suggest that mGlu₅R, A_2AR and D₂R are expressed in the same membrane domain and might form receptor heteromers.

Fig. 1.

Heteromerization in HEK cells. Immunofluorescence localization of mGlu₅R, A_2AR and D₂R in transiently transfected HEK-293 cells (panel a). Cells were transiently cotransfected with the cDNA’s encoding mGlu₅R_, A_2AR and D₂R and processed for immunostaining (see Materials and methods) using a rabbit anti-mGlu₅R, a mouse anti-A_2AR and a goat anti- D₂R. The bound primary antibody was detected using donkey-anti-rabbit-Cy5 (1/500), donkey-anti-mouse-Cy3 (1/500) and donkey-anti-goat-Cy2 (1/500). Superimposition of images reveals a high receptor codistribution in white (merge). Scale bar: 10 μm. Coimmunoprecipitation assays in HEK-293 cells (panels b, c and d). (b) Cells were transiently transfected with mGlu₅R^GFP2 (lane 1), D₂R (lane 2), A_2AR (lane 3) or with all three receptors simultaneously (lane 4). (c) Cells were transiently transfected with mGlu₅R (lane 1), D₂R^GFP2 (lane 2), A_2AR (lane 3) or with all three receptors simultaneously (lane 4). (d) Cells were transiently transfected with mGlu₅R (lane 1), D₂R (lane 2), A_2AR^GFP2 (lane 3) or with all three receptors simultaneously (lane 4). Cells were washed, solubilized and processed for immunoprecipitation using anti-GFP monoclonal antibody (2 μg/ml; IP: GFP). Solubilized membranes (Crude; 20 μg) and immunoprecipitates (IP) were analyzed by SDS-PAGE and immunoblotted using rabbit anti-A_2AR whole serum (1/2000), rabbit anti-D₂R whole serum (1/2000) or rabbit anti-mGlu₅R polyclonal antibody (1 μg/ml) and horseradish-peroxidase (HRP)-conjugated goat anti-rabbit IgG as a secondary antibody. These blots are representative of four different experiments with similar qualitative results.

Study of A_2AR/D₂R, mGlu₅R/A_2AR and mGlu₅R/D₂R oligomers by BRET assays

Although the use of biochemical approaches to demonstrate receptor oligomerization has been widely used, it might have some disadvantages since the cellular structure is destroyed by detergent treatment. Thus, in order to assess if mGlu₅R, A_2AR and D₂R form oligomers in living cells, biophysical approaches were performed in cells transiently transfected with receptor constructs that carry appropriate fluorophore pairs in their C-terminus. Initially, by using a BRET approach, a positive and saturable BRET signal for the transfer of energy between D₂R^Rluc and A_2AR^YFP in cells co-transfected with a constant concentration of the D_2LR^Rluc and increasing concentrations of A_2AR^YFP was observed (Fig. 2a). As the control pair D_2LR^Rluc and CD₄R^YFP led to a quasi-linear curve, the specificity of the saturation (hyperbolic) assay for the D_2LR^Rluc and A_2AR^YFP pair could be established (Fig. 2a). Similarly, BRET² studies between A_2AR^Rluc and mGlu₅R^GFP2 and between D_2LR^Rluc and mGlu₅R ^GFP2 were performed and similar results obtained (data not shown). These results clearly corroborate previous studies about heterodimerization of A_2AR with D₂R and A_2AR with D₂R (Canals et al., 2003; Ciruela et al., 2004; Ferre et al., 2002), and demonstrate the ability of D₂R to heteromerize with mGlu₅R. Interestingly, when the adenosine A₁ receptor (A₁R) and the GABA_B2 receptor (GABA_B2R) were used as acceptors molecules (e.g. A₁R^YFP and GABA_B2R^YFP) in a BRET process with the D_2LR^Rluc a considerable lower BRET efficiency was observed (Fig. 2b). These results suggest that receptors with none predicted physiological heteromer complex formation with D_2LR (e.g. A₁R and GABA_B2R) might indeed be used as negative controls in BRET experiments.

Fig. 2.

BRET-based study of receptor heteromerization in HEK-293 cells. BRET saturation experiments between D₂R^Rluc and A_2AR^YFP (●) or D₂R^Rluc and CD₄R^YFP (△) (panel a). Both fluorescence and luminescence of each sample were measured prior to every experiment to confirm equal expression of Rluc construct while monitoring the increase of YFP expression. mBU is the BRET ratio x1000 (see Materials and methods). Error bars indicate SD of mean specific BRET ratio (mBU) values of five individual experiments grouped as a function of the amount of fluorescence of the acceptor. (b) Comparison of the BRET_max obtained for D₂R^Rluc/A_2AR^YFP (solid bar), D₂R^Rluc/A₁R^YFP (white bar) and D₂R^Rluc/GABA_B2R^YFP (grey bar) under similar experimental conditions. Results are expressed as a percentage of the BRET_max obtained for the D₂R^Rluc/A_2AR ^YFP pair. (c) Displacement study of a constant BRET ratio (~BRET₅₀) of A_2AR^Rluc/D₂R^YFP pair by increasing amounts of mGlu₅R (●) or CD₄R (△) constructs. Also, a constant BRET ratio (~BRET₅₀) of D₂R^Rluc/mGlu₅R^YFP pair was displaced by increasing amounts of A_2AR (■).

To initially test the ability of these receptors to form heterotrimers, we performed a set of BRET competition experiments. Thus, HEK-293 were transiently transfected with a constant ratioof A_2AR^Rluc and D₂R^YFP (2μg and 1μg, respectively), that was around the BRET₅₀ of this BRET pair, and increasing amounts of mGlu₅R or CD₄R (Fig. 2c). As shown in Fig. 2b, under these conditions mGlu₅R was able to reduce the BRET₅₀ efficiency of the A_2AR^Rluc/D₂R^YFP- BRET pair in a biphasic mode. Conversely, increasing amounts of A_2AR were able to reduce BRET₅₀ efficiency of the D₂R^Rluc/mGlu₅R^YFP-BRET pair in a biphasic fashion (Fig. 2c). It is important to mention here that co-transfection of a fixed amount of the A_2AR^Rluc/mGlu₅R^YFP-BRET pair with increasing amounts of D₂R produced a markedly alteration of the donor/acceptor (e.g. A_2AR _R^luc/mGlu₅R^YFP) ratio, thus precluding BRET efficiencies comparison (data not shown). Finally, no competition of CD₄R with the A_2AR^Rluc/D₂R^YFP-BRET pair was observed (Fig. 2c). Overall, these results suggest that both mGlu₅R and A_2AR can either compete for the same D₂R binding site with two different affinities or two different binding sites with different affinities exists within the D₂R structure.

A combined BRET/BiFC assay to determine the existence of higher-order mGlu₅R/A_2AR/D₂R oligomers

As shown above, the existence of mGlu₅R/A_2AR/D₂R oligomers can be indirectly shown by using a combination of biochemical and biophysical techniques, but these approaches do not directly prove the existence of receptor complexes containing more than two protomers. For that purpose, we used a new experimental approach recently described (Gandia et al., 2008a) which consists in the combination of BRET and BiFC techniques (Fig. 3a). The BiFC assay, a protein fragment complementation method (Kerppola, 2006a), is based on the ability to produce a fluorescent complex from non-fluorescent constituents if a protein-protein interaction occurs (Kerppola, 2006b). To this end, mGlu₅R and D₂R fused at the C terminal with N-YFP and C-YFP fragments were generated and transfected into HEK cells. Cells single transfected with receptors containing either the C-YFP or N-YFP fragments in their C-terminal tail did not provide a positive fluorescent signal, (data not shown). Since receptor heterodimerization caused YFP reconstitution, thus allowing fluorescence detection, we were able to visualize mGlu₅R/D₂R heterodimers in living cells (Fig. 3b). On the other hand, it has been previously shown that fluorescent protein fragments are able to complement with low efficiency, thus forming fluorescent complexes, even in the absence of a specific interaction (Kerppola, 2006b). Therefore, some control experiments were designed to ensure that receptor heterodimerization was not driven by spontaneous YFP complementation. The most appropriate negative controls for this technique are fusion proteins that are expressed in the same subcellular compartment than the proteins under study but that are not able to interact. In this study we have used other unrelated GPCRs as negative controls, namely the GABA_B1bR and GABA_B2R. Under the same experimental conditions, we were not able to detect fluorescence complementation between the mGlu₅R or D₂R and either the GABA_B1bR or GABA_B2R (data not shown). Overall, these results convincingly show for the first time that mGlu₅R and D₂R do heterodimerize and that the mGlu₅R/D₂R heterodimers are located at the plasma membrane level of living cells (Fig. 3b, right panel)

Fig. 3.

Detection of mGlu₅R/A_2AR/D₂R heteromeric complexes by combined BRET/BiFC assays. (a) Schematic representation of the combined BRET/BiFC assay. BRET signal is triggered by the oxidation of h-coelenterazine by the Rluc fused to the C-terminal tail of the A_2AR^Rluc construct. This enzymatic reaction produces light emission at 475 nm, which might excite a proper fluorophore acceptor found in close proximity (within 10 nm). As acceptor we use the mGlu₅R/D₂R heterodimer-mediated complemented-YFP (mGlu₅R^N-YFP/D₂R^C-YFP) that after excitation it emits at 527 nm. (b) Visualization of mGlu₅R/D₂R heterodimer by BiFC assays. HEK cells were transiently transfected with 5 μg of the cDNA encoding mGlu₅R^YFP, D₂R^YFP or mGlu₅R^N-YFP (5 μg) plus D₂R^C-YFP (5 μg) and processed for confocal microscopy imaging. Microscope observations were made with Olympus Fluoview 500 confocal scanning laser adapted to an inverted Olympus IX-70 microscope. Scale bar: 10 μm. (c) BRET saturation curve. BRET was measured in HEK cells coexpressing A_2AR^Rluc and mGlu₅R^N-YFP+D₂R^C-YFP (●) or A_2AR^Rluc and GABA_B2R^YFP (△) constructs. Co-transfections were performed with increasing amounts of plasmid DNA for the YFP construct whereas the DNA for the Rluc construct was maintained constant. Both fluorescence and luminescence of each sample were measured prior to every experiment to confirm equal expression of Rluc while monitoring the increase of YFP expression (see Materials and methods).

Under the same experimental conditions, we performed a BRET saturation curve in cells co-transfected with a constant amount of the A_2AR^Rluc construct and increasing concentrations of the mGlu₅R^N-YFP+D₂R^C-YFP plasmids or GABA_B2R^YFP. A positive BRET signal for the transfer of energy between A_2AR^Rluc and mGlu₅R^N-YFP +D₂R^C-YFP was obtained (Fig. 3c). The BRET signal increased as a hyperbolic function of the heterodimer-mediated complemented YFP (assessed by the fluorescence emitted upon direct excitation at 480 nm), thus strongly suggesting the formation of receptor oligomers containing more than two protomers, i.e. mGlu₅R/D₂R/A_2AR oligomers (Fig. 3c).

SRET in the study of higher-order mGlu₅R/A_2AR/D₂R oligomers

To pursue the direct detection of mGlu₅R/D₂R/A_2AR oligomer formation, we implemented a new experimental approach developed by our research group (Carriba et al., 2008). This technique, named SRET, engage a BRET² method with a FRET process (Carriba et al., 2008) (Fig. 4a). SRET is based on an adequate combination of donors and acceptors for BRET and for FRET and on the fact that different emission profiles result from the use of different Rluc substrates. Thus, DeepBlueC is used as a trigger agent for SRET instead of coelenterazine H. By the action of Rluc, oxidation of DeepBlueC is able to excite GFP², which is a so called BRET² signal. While returning to the ground state, excited GFP² emits energy that excites YFP, providing a FRET signal. For a specified level of expression of the receptor coupled to the luciferase (Rluc), the highest SRET efficiency is observed when the FRET donor is relatively low whereas the expression of the FRET acceptor is relatively high. Therefore, we performed a set of preliminary experiments to establish the optimal concentrations of receptor constructs that need to be transfected to provide a maximum BRET² and FRET efficiency with the minimum photophysic cross-talks (Supplementary Fig. 1) (Carriba et al., 2008; Ciruela, 2008). Once the optimal conditions were found, we performed a single SRET measurement to characterize the mGlu₅R/D₂R/A_2AR oligomer. Thus, by using the following fusion protein ratio: D_2LR^Rluc (2 μg), mGlu₅R^GFP2 (4μg) and A_2AR^YFP (8 μg), we obtained a positive and specific SRET signal (Fig. 4b). Next, we performed a SRET saturation curve. Thus, HEK-293 cells co-expressing a constant ratio of D_2LR^Rluc(2 μg) plus mGlu₅R^GFP2(4 μg) and increasing amounts of A_2AR^YFP or CD₄R^YFP (Fig. 4c). Again, we could establish the specificity of the mGlu₅R/D₂R/A_2AR oligomer formation, since we obtained a saturable (hyperbolic) curve when compared to the negative control (Fig. 4c)

Fig. 4.

Detection of mGlu₅R/A_2AR/D₂R heteromeric complexes by SRET assays. Schematic representation of the SRET assay (panel a). The initial BRET signal is triggered by the oxidation of DeepBlueC by the Rluc fused to the C-terminal tail of the D₂R^Rluc construct. This enzymatic reaction produces light emission at 400 nm, which might excite a proper fluorophore acceptor, for instance mGlu₅R^GFP2, if found in close proximity (within 10 nm). This fluorophore acceptor (mGlu₅R^GFP2)can engage now in a FRET process acting as a donor whose emission light at 510 nm can excite a suitable acceptor (A_2AR^YFP), thus ending in a net 530 nm fluorescence emission. (b) NetSRET was measured in HEK-293 cells co-expressing D₂R ^Rluc, mGlu₅R^GFP2 and A_2AR^YFP. Fluorescence and luminescence of each sample were measured prior to every experiment to confirm similar expression levels. Data, expressed as net SRET are means ± SEM of three independent experiments performed in triplicate. One-way ANOVA followed by Newman-Keuls test showed significant differences respect to both negative controls. ***p<0.001. (c) SRET saturation was measured in HEK-293 cells co-expressing D₂R ^Rluc, mGlu₅R ^GFP2 (in a constant ratio 2 μg and 4 μg, respectively) with increasing amounts of the A_2AR^YFP (●) or CD₄R^YFP (△). Net SRET was detected for D_2LR ^Rluc/mGlu₅R^GFP2/A_2AR^YFP compared with net SRET detected with cells expressing only D_2LR^Rluc plus mGlu₅R^GFP2.

mGlu₅R/D₂R/A_2AR interactions in rat striatum

The demonstration of mGlu₅R/D₂R/A_2AR oligomers in living cells provides for a morphological framework for the already known multiple functional interactions between these receptors in the brain (see Introduction). To asses the physiological significance of the mGlu₅R/D₂R/A_2AR interaction co-immunoprecipitation experiments were performed using rat striatum homogenates. Using soluble extracts from rat striatum the anti-A_2AR antibody was able to immunoprecipitate a band around 45 kDa which corresponds to the striatal A_2AR, as expected (Fig. 5a). Interestingly, this antibody was able to immunoprecipitate the D₂R (~90 kDa) and mGlu₅R (~130 kDa) from the same extracts (Fig. 5a, lane 2). These bands did not appear when an irrelevant rabbit IgG was used for immunoprecipitation (Fig. 5a, lane 1), showing that the reaction was specific. Conversely, the anti-D₂R antibody was able to immunoprecipitate the A_2AR and mGlu₅R and the anti-mGlu₅R antibody the A_2AR and D₂R, respectively (Fig. 5a). Overall, these results suggest that mGlu₅R, D₂R and A_2AR do oligomerize in native tissue, namely striatum, and that this oligomerization might be physiologically relevant.

Fig. 5.

Interaction of A_2AR, D₂R and mGlu₅R in rat striatum. (a) Co-immunoprecipitation of A_2AR, D₂R and mGlu₅R from rat striatum. Rat striatal membanes were solubilized and processed for immunoprecipitation using control rabbit IgG (5μg/ml; lane 1), rabbit anti-A_2AR polyclonal antibody (5μg/ml; lane 2), rabbit anti-D₂R polyclonal antibody (5μg/ml; lane 3) and rabbit anti-mGlu₅R polyclonal antibody (5μg/ml; lane 4). Immunoprecipitates were analyzed by SDS-PAGE and immunoblotted using rabbit anti-A_2AR whole serum (1/2000), rabbit anti-D₂R whole serum (1/2000) or rabbit anti-mGlu₅R polyclonal antibody (1 μg/ml) and HRP-conjugated anti-rabbit IgG TrueBlot^™ (1/1000) as a secondary antibody. These blots are representative of three different experiments with similar qualitative results. (b) Subcellular distribution of A_2AR, D₂R and mGlu₅R in rat striatum. Electron micrographs showing immunoreactivity for A_2AR, D₂R and mGlu₅R in rat striatum as revealed using a triple-labelling post-embedding immunogold technique. Immunoparticles for A_2AR (10 nm size, arrows), D₂R (15 nm size, crossed arrows) and mGlu₅R (20 nm size, arrowheads) were detected along the extrasynaptic and perisynaptic plasma membrane of the same dendritic spine (s) establishing excitatory synaptic contact with axon etrminals (b). Scale bar: 0.2 μm.

To provide the precise morphological evidence and to elucidate the subcellular localization of the mGlu₅R/D₂R/A_2AR oligomer in neurons of the rat striatum, we performed triple-labelling post-embedding immunogold techniques at electron microscopic level. Thus, we have showed here that mGlu₅R, D₂R and A_2AR co-distributed in postsynaptic structures along the extra-synaptic and peri-synaptic plasma membrane of spines, establishing asymmetrical, putative glutamatergic, synapses with axon terminals (Fig. 5b). This precise distribution of mGlu₅R, D₂R and A_2AR in striatal neurons resembles that previously described for these receptors (Uchigashima et al., 2007a)(Ciruela et al., 2006a). Overall, this is the first direct anatomic evidence for mGlu₅R, D₂R and A_2AR co-distribution in the same neuronal compartment and support the idea that the mGlu₅R/D₂R/A_2AR oligomer localized in the GABAergic striatopallidal neurons might play a key role in striatal function both in normal and pathological conditions.

Discussion

A key aim of post-genomic biomedical research is to systematically catalogue all molecules and their interactions in a living cell. Many fundamental cellular processes involve multiple interactions among proteins and other biomolecules, i.e. biomolecular interaction networks (Xia et al., 2004). Protein-protein physical interactions constitute an important group of biomolecular interaction networks (Xia et al., 2004), such as the neuronal horizontal molecular networks (Agnati et al., 2003; Bockaert et al., 2003; Franco et al., 2003). Horizontal molecular networks take place at the neuronal plasma membrane level, where specific GPCR interact and integrate the messages provided by a variety of neurotransmitters. Immunodetection and co-immunoprecipitation assays are being used successfully to draw a map of molecular networks involving protein-protein interactions of cytosolic proteins. In contrast, these techniques have limitations when analyzing heptaspanning membrane receptors. In the early eighties, and based on indirect functional evidences, it was proposed that GPCR receptors could interact at the level of the neuronal plasma membrane. In the early nineties, electrophoretic mobility and co-immunoprecipitation assays gave the first indication of GPCR homomerization. More recently, several experimental approaches have been used in the study of the quaternary structure of GPCRs, revealing the existence of receptor homo- and heteromers (Milligan and Bouvier, 2005). Among these approaches, fluorescence-based methods, as non-invasive techniques, have played a key role in the characterization of a large array of protein-protein interactions in general and in the study of GPCR oligomerization in particular (Gandia et al., 2008b). The use of the biophysical techniques BRET and FRET, allowed the demonstration of GPCR homodimerization and heterodimerization GPCRs in living cells (Agnati et al., 2003; Bouvier, 2001; Franco et al., 2003). We and others have recently introduced a set of new techniques based on the combination of RET-based methods together with Bimolecular Fuorescence/Luminescence Complementation techniques (BiFC and BiLC, respectively) that has made possible to detect GPCR heteromers constituted by more than two receptors (Carriba et al., 2008; Dupre et al., 2006; Gandia et al., 2008a; Guo et al., 2008; Lopez-Gimenez et al., 2007; Navarro et al., 2008; Vidi et al., 2008).

The existence of functional mGlu₅R/D₂R/A_2AR oligomers in the GABAergic striatopallidal neuron has often been discussed in the literature, based on the high and selective co-expression of mGlu₅R, D₂R and A_2AR in these particular cells (see Introduction), on the demonstration of A_2AR/D₂R and A_2AR/mGlu₅R heteromers (see Introduction) and on the existence of strong multiple interactions between the three receptors (see below). However, the demonstration of their simultaneous physical interaction had not yet been reported. By using immunodetection and co-immunoprecipitation, BRET competition, BRET/BiFc and SRET techniques, this study provides clear evidence for the existence of the mGlu₅R/D₂R/A_2AR oligomers in living cells and in rat striatum. Co-immunodetection and co-immunoprecipitation experiments demonstrated that when co-expressed in the same cell or in native tisue, mGlu₅R, D₂R and A_2AR are physically linked in the cell membrane. BRET competition experiments gave indirect evidence for the existence of mGlu₅R/D₂R/A_2AR oligomers, by demonstrating that co-expression of mGlu₅R significantly modified the BRET signal of the heteromer formed by A_2AR^Rluc and D₂R^YFP. Also, the direct proof for mGlu₅R/D₂R/A_2AR oligomerization came from experiments with BRET/BiFC and SRET. Finally, our immunogold experiments clearly demonstrated the specific postsynactic location of the mGlu₅R/D₂R/A_2AR oligomer in GABAergic striatopallidal neurons. However, it is important to mention here that apart of the assumed postsynaptic interaction describe here a putative presynaptic interaction for A_2AR/mGlu₅R and A_2AR/D₂R has been also showed (Rodrigues et al., 2005; Tozzi et al., 2007). To this regard we can not exclude the possibility that the mGlu₅R/D₂R/A_2AR oligomer might also exists presynaptically, thus controlling the excitability and firing of medium spiny neurons. This is an issue that deserves more future work.

Strong functional antagonistic interactions between A_2AR and D₂R were first described and recently reviewed (Ferre et al., 2008)) and, later on, antagonistic interactions between mGlu₅R and D₂R and synergistic interactions between A_2AR and mGlu₅R were also reported (Ferre et al., 2002; Popoli et al., 2001). In membrane preparations from rat striatum, stimulation of either A_2AR or mGlu₅R produces a decrease in the affinity of D₂R for agonists and a decrease in D₂R agonist-mediated motor activation (Ferre et al., 1991; Popoli et al., 2001). Interestingly, co-stimulation of A_2AR and mGlu₅R produces a synergistic antagonistic modulation of D₂R ligand binding and function that is significantly stronger than the reduction induced by stimulation of either receptor alone (Popoli et al., 2001). Thus, co-administration of A_2AR and mGlu₅R agonists inhibits motor activation induced by D₂R agonists (Ferre et al., 2002; Popoli et al., 2001). Also, central A_2AR and mGlu₅R agonist co-administration counteracts the inhibitory effect that endogenous dopamine exerts on c-fos expression by tonically stimulating striatal D₂R (Ferre et al., 2002). It was, therefore, hypothesized that the multiple mGlu₅R-D₂R-A_2AR interactions could provide a new therapeutic approach for some neuropsychiatric disorders, such as Parkinson’s disease (Fuxe et al., 2003; Popoli et al., 2001). In fact, A_2AR antagonists are being shown to be clinically useful as antiparkinsonian agents (Muller and Ferre, 2007). Also, there is preclinical evidence for the efficacy of mGlu₅R antagonists (Ossowska et al., 2001)(Breysse et al., 2002) and, more importantly, for a synergistic effect of A_2AR and mGlu₅R antagonists in animal models of Parkinson’s disease (Coccurello et al., 2004; Kachroo et al., 2005). Hyperactivity of the GABAergic striatopallidal neuron is a main pathophysiological mechanism responsible for hypokinesia in Parkinson’s disease (Obeso et al., 2000). The reported synergistic antiparkinsonian effect of A_2AR and mGlu₅R antagonists in animal models (Coccurello et al., 2004; Kachroo et al., 2005) suggests that the dopamine depletion-induced hyperactivity of striatopallidal neurons depends on a synergistic stimulatory endogenous tone of endogenous adenosine and glutamate on A_2AR and mGlu₅R, upon interruption of D₂R signaling.

In summary, in the present study we provide biochemical evidence that supports the existence of mGlu₅R-D₂R-A_2AR heteromers in living cells and strongly suggest their presence in the brain, in the GABAergic striatopallidal neurons, where they can provide important targets for the treatment of neuropsychiatric disorders.